UvA-DARE (Digital Academic Repository) Glycobiology in cardiometabolic homeostasis ... ·...

UvA-DARE is a service provided by the library of the University of Amsterdam (http://dare.uva.nl)

UvA-DARE (Digital Academic Repository)

Glycobiology in cardiometabolic homeostasis

Hassing, H.C.

Link to publication

Citation for published version (APA):Hassing, H. C. (2013). Glycobiology in cardiometabolic homeostasis.

General rightsIt is not permitted to download or to forward/distribute the text or part of it without the consent of the author(s) and/or copyright holder(s),other than for strictly personal, individual use, unless the work is under an open content license (like Creative Commons).

Disclaimer/Complaints regulationsIf you believe that digital publication of certain material infringes any of your rights or (privacy) interests, please let the Library know, statingyour reasons. In case of a legitimate complaint, the Library will make the material inaccessible and/or remove it from the website. Please Askthe Library: https://uba.uva.nl/en/contact, or a letter to: Library of the University of Amsterdam, Secretariat, Singel 425, 1012 WP Amsterdam,The Netherlands. You will be contacted as soon as possible.

Download date: 25 Mar 2020

https://dare.uva.nl/personal/pure/en/publications/glycobiology-in-cardiometabolic-homeostasis(8a2f3094-9bc1-46e2-87c0-d5235ecae932).html

GLYCOBIO

LOGY IN

CARDIOM

ETABOLIC HO

MEO

STASIS Carlijne Hassing

UITNODIGING

voor het bijwonen vande openbare verdediging van

het proefschrift

GLYCOBIOLOGY IN CARDIOMETABOLIC

HOMEOSTASIS

door

Carlijne Hassing

op vrijdag 31 mei 2013om 14:00 uur

in de AgnietenkapelOudezijds Voorburgwal 231

te Amsterdam

Receptie ter plaatse na afloop van de verdediging

Carlijne [email protected]

06-42133398

ParanimfenRobert-Jan Hassing

[email protected]

en

Sanneke [email protected]

06-41017655

GLYCOBIOLOGY IN CARDIOMETABOLIC HOMEOSTASIS

Carlijne Hassing


Carlijne Hassing

“Publication of this thesis was financially supported by Isis Pharmaceuticals Inc, MSD BV, Universiteit van Amsterdam, Stichting tot Steun Promovendi Vasculaire Geneeskunde, Genzyme Nederland, Novo Nordisk, Servier Nederland Farma BV, Boehringer Ingelheim BV and Chipsoft”

ISBN: 978-94-6191-735-5 78


ACADEMISCH PROEFSCHRIFT

ter verkrijging van de graad van doctoraan de Universiteit van Amsterdamop gezag van de Rector Magnificus

prof. dr. D.C. van den Boomten overstaan van een door het collegevoor promoties ingestelde commissie,

in het openbaar te verdedigen in de Agnietenkapelop vrijdag 31 mei 2013, te 14:00 uur

door

Henrieke Carlyne Hassinggeboren te Utrecht

PROMOTIECOMMISSIE

Promotor: Prof. Dr. E.S.G. Stroes

Co-promotores: Dr. M. Nieuwdorp Dr. G.M. Dallinga-Thie

Overige leden: Prof. dr. J.M.F.G. Aerts Prof. dr. J.J.P. Kastelein Prof. dr. R.J.G. Peters Prof. dr. J.B.L. Hoekstra Prof. dr. U.H.W. Beuers Prof. dr. J.M. Dekker

Faculteit der Geneeskunde

Het verschijnen van dit proefschrift werd mede mogelijk gemaakt door de steun van de Nederlandse Hartstichting

CONTENTS

CHAPTER 1 General introduction and outline of the thesis 7

PART I Causes of hypertriglyceridemia and novel treatment targets 17

CHAPTER 2 Pathophysiology of hypertriglyceridemia 19

CHAPTER 3 A multilocus genetic risk score for hypertriglyceridemia 37

CHAPTER 4 Intensive lipid-lowering therapy in subjects with low and high triglycerides plus low and high HDL-C - An analysis of the Treating to New Targets (TNT) trial

55

CHAPTER 5 The effect of a diiodothyronine mimetic on insulin sensitivity in male cardiometabolic patients: a double-blind randomized controlled trial

75

CHAPTER 6 Hypertriglyceridemia: The future of genetics to guide individualized therapeutic strategies

95

PART II Acquired and inborn errors of heparansulfates in hyperglycaemia and hypertriglyceridemia

111

CHAPTER 7 Genetic variations in heparan sulfates only modestly affect postprandial triglyceride clearance in humans

113

CHAPTER 8 Carriers of loss-of-function mutations in EXT display impaired pancreatic beta-cell reserve due to smaller pancreas volume

129

CHAPTER 9 Inhibition of hepatic Sulf2 in vivo: a novel strategy to correct diabetic dyslipidemia

147

CHAPTER 10 A genetic variant at the SULF2 locus associates with postprandial triglycerides in patients with type 2 diabetes mellitus

165

CHAPTER 11 A genetic variant in SULF2 and metabolic responses in a population-based cohort

185

CHAPTER 12 Summary and perspectives 199

NEDERLANDSE SAMENvATTING 207

AUTHORS AND AFFILIATIONS 213

DANkwOORD 219

CURRICULUM vITAE 225

1 GENERAL INTRODUCTION AND OUTLINE OF THE THESIS

8

Chapter 1

INTRODUCTIONCardiovascular disease and the cardiometabolic syndromeCardiovascular disease (CVD) remains a leading cause of death worldwide. With an estimated number of 17.5 million per year, it represents 30% of all global deaths. 1 Due to the emerging prevalence of obesity and type 2 diabetes mellitus (T2DM), in 2030 almost 25 million people will die from any form of CVD. The cardiometabolic syndrome is a cluster of medical disorders often seen in subjects with type 2 diabetes mellitus including obesity, impaired glucose tolerance, insulin resistance and metabolic dyslipidemia (elevated plasma triglycerides (TGs) and decreased plasma high-density lipoprotein cholesterol (HDL-C) levels). Treatment of the cardiometabolic syndrome is currently aimed at treating its individual components, but still a considerable residual risk remains in these subjects. Although hampered by its complex pathophysiology, development of new therapeutic strategies targeting the cardiometabolic syndrome is urgently needed. The focus of this thesis therefore lies on the origin and treatment modalities of triglycerides and glucose metabolism that accompany the cardiometabolic syndrome.

HypertriglyceridemiaTGs are required for energy storage. They are derived from both exogenous (dietary) and endogenous (liver) sources. 2 TGs are transported into chylomicrons and very low-density lipoprotein (VLDL) particles and hydrolyzed in muscle, heart and adipose tissue by lipoprotein lipase to release free fatty acids for uptake. Triglyceride-depleted remnant particles are then transported to the liver where they are taken up and cleared. Because of the smaller particle size, these remnant particles are able to penetrate the vessel wall, where they accelerate atherogenesis most likely via an inflammatory reaction. 3 As plasma TGs reflects daily dietary fat consumption, they are highly variable. Hypertriglyceridemia, defined as plasma triglycerides levels ≥1.7 mmol/L (≥150 mg/dL), often arises from overconsumption of lipid-rich diets, obesity, physical inactivity, insulin resistance and related conditions resulting in increased production or decreased uptake of TGs. 4 Despite its coherence with insulin resistance induced dysglycemia, low HDL-C plasma levels and increased small-dense low-density lipoprotein (LDL) levels; elevated (both fasting and non-fasting) TG levels are an independent risk factor for CVD. 5, 6 However, nowadays therapeutic interventions to lower plasma TGs including ezetimibe and fibrates have failed to consistently reduce cardiovascular risk in prospective randomized studies. 7, 8 Thus, novel partakers in human lipid metabolism need to be identified in order to provide novel therapeutic targets.

Heparan sulfate proteoglycansDiabetes-associated, metabolic dyslipidemia, characterized by increased triglyceride levels with concomitant small dense LDL-cholesterol, is a major contributor to diabetic macrovascular complications. 6 Current strategies comprising statin and fibrate therapy prevent only 25% of all cardiovascular events, highlighting the need for additional lipid-modulating strategies in T2DM. 9 Historically, hyperglycaemia-associated changes in heparan sulfate (HS)-synthesis have been put

General introduction

9

1

forward as causal factor for cardiovascular complications in T2DM. 10, 11 Heparansulfate proteoglycans (HSPG) play a role in many biological processes including fine-tuning most of the (patho)physiological processes such as organ development, inflammatory pathways and lipid metabolism. 11 HSPGs comprise a proteoglycan/glycoprotein backbone (syndecans and glypicans) on which sulfated polysaccharides (heparan sulfates) are attached as side-chains that are able to “catch” and then internalize plasma proteins like chylomicron/VLDL remnant particles. 11, 12 Synthesis of HSPGs results from the action of multiple enzymes, that built HS chains on proteoglycans, involved in chain initiation (xylosyl transferase or XT), elongation (exostosin or EXT) or sulfation (N-deacetylase/N-sulfotransferase or NDST and O-sulfotransferase or OST), see Figure 1, 13.

Figure 1 - Effects of different heparansulfate synthesis genes on chain architecture. (adapted from Forsberg E, J Clin Invest. 2001).

10

Chapter 1

With respect to dyslipidemia, pioneering work showed that these negatively-charged HS facilitate binding and uptake of triglyceride-rich lipoprotein remnants (TRL) in hepatocytes. 14, 15 Following lipoprotein lipase-mediated hydrolysis of TRLs, the ensuing chylomicron- and VLDL-remnant particles are cleared by the liver via three receptors: the LDL receptor (LDLR), LDLR-related protein 1 (LRP1) and HSPG, see Figure 2 and references 16, 17. Intrigued by the observation that reduction in the degree of hepatic HS-sulfation was associated with increased TG-levels 18, Esko unambiguously demonstrated that the proteoglycan syndecan-1 (SDC1) is of pivotal importance for hepatic remnant clearance in conditional knockout mice on a normal C57BL6 background. 19, 20

Figure 2 - Hepatic clearance of TRL. After lipolytic processing of lipoproteins in the circulation by lipoprotein lipase (Lpl; dark greytriangles), remnant lipoproteins enter the space of Disse through openings in the endothelium. Remnant lipoproteins are thought to be sequestered near the hepatocyte cell surface via apoE (black circles)–heparan sulfate binding or lipase–heparan sulfate bridging on heparan sulfate proteoglycans (HSPGs). Lipoproteins then are further processed in the space of Disse by transfer of soluble apoE (grey circles) and by hepatic lipase (light grey triangles) bound via heparan sulfate. Endocytosis of lipoprotein particles occurs by LDL receptor (LDLR), LDL-receptor related protein (LRP) or by proteoglycans. (Addaptem from Bishop, Curr Opin Lipidol 2008).

Recent animal studies have implicated that HSPG degradation in T2DM might attribute to its characteristic metabolic dyslipidemia. 21 Hepatic glucosamine-6-O-endosulfatase-2 (Sulf2), a HSPG degrading enzyme that selectively removes 6-O-sulfates from HS chains, was strongly over-expressed in livers of diabetic mice which coincided with a diminished TRL binding to primary hepatocytes and concomitant elevations in plasma TG. In addition, genome wide association studies have linked


11

1

HSPG genes to the development of T2DM.22 Human studies exploring the role of HSPG in glucose and lipid homeostasis are lacking.

OUTLINE OF THE THESISThis thesis explores the first steps in linking HSPG homeostasis to cardiometabolic diseases. Although it takes many steps for medical findings to develop from bench to bedside, we aim to address a number of steps to achieve translation insights: we studied in-vitro and animal models and translated these findings into specific patient groups as well as performing observational studies in large cohorts. Our results in cell, mice and men form part of the initial steps towards possible targets for therapy.

This thesis consists of two parts. The first part focuses on the causes of hypertriglyceridemia and novel targets in the cardiometabolic syndrome. In chapter 2 we review novel aspects in TG metabolism and the pathophysiology of hypertriglyceridemia. In chapter 3 we explore the value of a gene risk score for hypertriglyceridemia to improve cardiovascular risk prediction. To achieve this, we evaluated the ability of single nucleotide polymorphisms (SNPs) in TG modulating genes to predict plasma triglyceride levels as well as first cardiovascular event in a prospective case-control (EPIC) study. Chapter 4 describes the additional benefit of increasing statin dose in patients with elevated TGs on coronary artery disease. In this chapter we performed a post hoc analysis in a large randomized clinical statin trial in order to evaluate the additional benefit of atorvastatin 80 mg versus 10 mg in high risk patients stratified by HDL-C and TG levels. Another potential novel candidate, with promising results from animal studies targeting cardiometabolic sequelae, are the thyroid hormone mimetics. Since thyroid hormone mimetics are capable of uncoupling the beneficial metabolic effects of thyroid hormones from their deleterious effects on heart, bone and muscle, this class of drug is considered as adjacent therapeutics to weight-lowering strategies. In chapter 5 we performed a randomized, placebo-controlled, double-blind trial to investigate the effect of TRC150094, a thyroid hormone mimetic, on insulin sensitivity, liver fat content and lipid profile, as well as on safety markers in obese male subjects with an increased cardiometabolic risk. Chapter 6 describes the future of gene sequencing in therapeutic strategies aimed at lowering plasma triglyceride levels.

The second part of this thesis explores the role of HSPG in the development of hypertriglyceridemia and hyperglycaemia. Chapter 7 describes a study investigating the postprandial lipid handling by heparan sulfates in subjects with hereditary multiple exostosis (HME) and familial hypercholesterolaemia (FH). Given that subjects with HME suffer from an inborn error in HSPG synthesis (loss-of-function mutations in EXT) they form a unique model to evaluate the role of HSPGs in human lipid metabolism. In chapter 8 we used a translational approach to study the effect of EXT mutations on pancreas volume, insulin secretion capacity (betacell reserve) and subsequent glucose handling in both mice and humans (using heterozygous EXT mice and HME subjects with

12

Chapter 1

heterozygous EXT mutations). In chapter 9 we evaluate the effect of hepatic Sulf2 inhibition by antisense strategy on diabetic dyslipidemia in diabetic mice as previous data indicated that hepatic Sulf2 expression is strongly induced and linked to elevated plasma TG levels in these diabetic animals. In chapter 10 we aim to translate earlier in-vitro and animal findings linking acquired errors in Sulf2 regulation in diabetes and dyslipidemia to humans. In this chapter, we study the human hepatic SULF2 expression as well as the effect of genetic variation in SULF2 on fasting and postprandial lipid levels in subjects with T2DM and diabetic dyslipidemia. Furthermore, chapter 11 describes the effect of a predetermined common genetic variation in SULF2 on metabolic responses following oral glucose tolerance tests (OGTTs) and a meal tolerance test in a population-based non diabetic cohort. Finally, in chapter 12 we provide a summary and conclusion of this thesis.


13

1

REFERENCE LIST1. World Health Organisation. Cardiovascular diseases. 2013. http://www.who.int

2. Williams KJ. Molecular processes that handle - and mishandle - dietary lipids. J Clin Invest 2008;118:3247-3259.

3. Tabas I, Williams KJ, Boren J. Subendothelial lipoprotein retention as the initiating process in athero-sclerosis: update and therapeutic implications. Circulation 2007;116:1832-1844.

4. Hassing HC, Surendran RP, Mooij HL, Stroes ES, Nieuwdorp M, Dallinga-Thie GM. Pathophysiology of hypertriglyceridemia. Biochim Biophys Acta 2012;1821:826-832.

5. Bansal S, Buring JE, Rifai N, Mora S, Sacks FM, Ridker PM. Fasting compared with nonfasting triglyc-erides and risk of cardiovascular events in women. JAMA 2007;298:309-316.

6. Nordestgaard BG, Benn M, Schnohr P, Tybjaerg-Hansen A. Nonfasting triglycerides and risk of myo-cardial infarction, ischemic heart disease, and death in men and women. JAMA 2007;298:299-308.

7. Kastelein JJ, Akdim F, Stroes ES et al. Simvastatin with or without ezetimibe in familial hypercholes-terolemia. N Engl J Med 2008;358:1431-1443.

8. Ginsberg HN, Elam MB, Lovato LC et al. Effects of combination lipid therapy in type 2 diabetes mel-litus. N Engl J Med 2010;362:1563-1574.

9. Ahmed S, Cannon CP, Murphy SA, Braunwald E. Acute coronary syndromes and diabetes: Is intensive lipid lowering beneficial? Results of the PROVE IT-TIMI 22 trial. Eur Heart J 2006;27:2323-2329.

10. Kjellen L, Bielefeld D, Hook M. Reduced sulfation of liver heparan sulfate in experimentally diabetic rats. Diabetes 1983;32:337-342.

11. Bishop JR, Schuksz M, Esko JD. Heparan sulphate proteoglycans fine-tune mammalian physiology. Nature 2007;446:1030-1037.

12. Bishop JR, Stanford KI, Esko JD. Heparan sulfate proteoglycans and triglyceride-rich lipoprotein me-tabolism. Curr Opin Lipidol 2008;19:307-313.

13. Forsberg E, Kjellen L. Heparan sulfate: lessons from knockout mice. J Clin Invest 2001;108:175-180.

14. Ji ZS, Sanan DA, Mahley RW. Intravenous heparinase inhibits remnant lipoprotein clearance from the plasma and uptake by the liver: in vivo role of heparan sulfate proteoglycans. J Lipid Res 1995;36:583-592.

15. Fuki IV, Kuhn KM, Lomazov IR et al. The syndecan family of proteoglycans. Novel receptors mediating internalization of atherogenic lipoproteins in vitro. J Clin Invest 1997;100:1611-1622.

16. Yu KC, Chen W, Cooper AD. LDL receptor-related protein mediates cell-surface clustering and hepatic sequestration of chylomicron remnants in LDLR-deficient mice. J Clin Invest 2001;107:1387-1394.

17. Williams KJ, Chen K. Recent insights into factors affecting remnant lipoprotein uptake. Curr Opin Lipidol 2010;21:218-228.

18. Ebara T, Conde K, Kako Y et al. Delayed catabolism of apoB-48 lipoproteins due to decreased heparan sulfate proteoglycan production in diabetic mice. J Clin Invest 2000;105:1807-1818.

14

Chapter 1

19. MacArthur JM, Bishop JR, Stanford KI et al. Liver heparan sulfate proteoglycans mediate clear-ance of triglyceride-rich lipoproteins independently of LDL receptor family members. J Clin Invest 2007;117:153-164.

20. Stanford KI, Bishop JR, Foley EM et al. Syndecan-1 is the primary heparan sulfate proteoglycan me-diating hepatic clearance of triglyceride-rich lipoproteins in mice. J Clin Invest 2009;119:3236-3245.

21. Chen K, Liu ML, Schaffer L et al. Type 2 diabetes in mice induces hepatic overexpression of sulfatase 2, a novel factor that suppresses uptake of remnant lipoproteins. Hepatology 2010;52:1957-1967.

22. Sladek R, Rocheleau G, Rung J et al. A genome-wide association study identifies novel risk loci for type 2 diabetes. Nature 2007;445:881-885.

I PART I

CAUSES OF HYPERTRIGLYCERIDEMIA AND NOvEL TREATMENT TARGETS

Hassing HC, Surendran RP, Mooij HL, Stroes ESG, Nieuwdorp M, Dallinga-Thie GM

Biochim Biophys Acta. 2012;1821:826-32

2 PATHOPHYSIOLOGY OF HYPERTRIGLYCERIDEMIA

20

Chapter 2

ABSTRACTThe importance of triglycerides as risk factor for CVD is currently under debate. The international guidelines do not include TG into their risk calculator despite the recent observations that plasma TG is an independent risk factor for CVD. The understanding of the pathophysiology of triglycerides opens up avenues for development of new drug targets. Hypertriglyceridemia occurs through 1. Abnormalities in hepatic VLDL production, and intestinal chylomicron synthesis 2. Dysfunctional LPL-mediated lipolysis or 3. Impaired remnant clearance. The current review will discuss new aspects in lipolysis by discussing the role of GPIHBP1 and the involvement of apolipoproteins and in the process of hepatic remnant clearance with a focus upon the role of heparin sulfate proteoglycans. Finally we will shortly discuss future perspectives for novel therapies aiming at improving triglyceride homeostasis. This article is part of a Special Issue entitled Triglyceride Metabolism and Disease

PathoPhysiology of hyPertriglyceridemia

21

2

INTRODUCTIONLipoproteins are large macromolecular complexes of hydrophobic lipids and proteins designed to transport water insoluble lipids such as triglycerides and cholesteryl esters in body fluids.1 Lipoproteins contain a hydrophobic core of triglycerides and cholesterol ester enveloped by a monolayer of phospholipids, unesterified cholesterol and apolipoproteins. Triglyceride-rich lipoproteins (TRLs) originate from the intestine (chylomicrons) or from the liver (very-low-density lipoproteins, VLDL). After ingestion of a meal triglycerides are taken up in the enterocytes and packaged into large particles containing apoB48 as core protein. Upon secretion into the lymphatic system remodelling occurs before chylomicron remnants enter the systemic circulation.2 Endogenous synthesized TG in the liver is packaged into very low density lipoproteins (VLDL) containing apoB100 as core protein. TRLs play an essential role in delivering fatty acids (FFA) to tissues as source of energy (heart and skeletal muscle) or for storage (adipose tissue). Plasma TG levels are determined by several key metabolic pathways: Intestinal uptake from dietary fat, hepatic production, peripheral lipolysis induced TRL remodelling and hepatic removal of VLDL and chylomicron remnants will be discussed.3 Abnormalities in TG metabolism are a hallmark of a number of clinical disturbances including type 2 diabetes, familial combined hyperlipidemia, dysbetalipoproteinema and severe hypertriglyceridemia and are conferred to increased risk for CVD.

Recently, a scientific statement from the American Heart Association was issued to highlight the notion that plasma TG levels display a steady increase that contributes in a large extent to the continuously increasing cardiometabolic risk particularly.4

In the present review, recent insight into pathophysiology of hypertriglyceridemia and future developments in triglyceride-lowering therapies are discussed.

TG METaboLisM (FiGuRE 1-4)Dietary fat absorption and formation of chylomicrons in the intestineTriglycerides derived from dietary sources are hydrolysed in the intestine by pancreatic lipase in 2-monoacylglycerol (2-MG) and fatty acid (FA), which can be absorbed by the enterocytes by diffusion or specific transporters such as FAT/CD36.5 Within the enterocyte, 2-MG and FA are resynthesized into TGs by the enzyme acyl-CoA:diacylglycerol acyltransferase (DGAT).6 Subsequently, microsomal triglyceride transfer protein (MTP) in complex with protein disulphite isomerase (PDI) facilitates the lipidation of apolipoprotein B48 (apoB48), as a first step towards chylomicron formation. Epithelial COPII (Coatomere Protein II) transport carriers like SAR1a and SAR1b are essential for the transport of chylomicrons to the Golgi apparatus.7 Human relevance is underscored by the observation that chylomicron retention disease and Anderson disease are autosomal recessive disorders of severe fat malabsorption with a complete absence of circulating apoB48 particles due to a genetic defect in the COPII machinery.8 Nascent chylomicron particles are exocytosed from the basolateral membrane

22

Chapter 2

and enter the lymphatic compartment and eventually the systemic circulation.5 The intestine harbors the option to synthesize apoC-III and possibly apoA-V. Whether these apolipoproteins are secreted associated with chylomicron particles is still unknown. Upon entering the circulation direct apolipoprotein exchange occurs with HDL particles enriching the chylomicron particles with apoE and apoC-III. In the fasting state chylomicrons are small, whereas ingestion of a meal leads to an increase of chylomicron size rather than chylomicron particle number.

Figure 1 - intestinal chylomicron synthesis. Dietary lipids are taken up in the enterocytes, incorporated into chylomicron particles that are secreted into the lymphatic system. DGAT, the COPII machinery system and MTP are required for the intracellular processing of chylomicrons.

Hepatic VLDL productionTG is synthesized in the liver and packaged into VLDL particles, with apoB100 as the main protein. The required fatty acids are derived from de novo synthesis using glucose as substrate (DNL) or from lipolysis in adipose tissue by the action of hormone sensitive lipase and adipose tissue TG lipase. The two DGAT enzymes are responsible for the generation of TG stored in lipid droplets.6 MTP is essential for initial lipidation of apoB100, whereas the COPII machinery is responsible for the early translocation of small VLDL particles from the ER to Golgi apparatus where further lipidation occurs.9 Mature VLDL particles are then excreted by the hepatocytes.


23

2

The liver is the major organ for production of apoA-V and apoC-III, two proteins involved in TG homeostasis. In plasma, apoC-III and apoA-V are constituents of apoB-containing lipoprotein particles and HDL depending upon the plasma TG levels and are strongly associated with plasma TG levels.10,11 ApoC-III is a very abundant plasma apolipoprotein (4-10 mg/dl), whereas apoA-V levels are extremely low 50 – 400 μg/l. Interestingly, it has been postulated that apoC-III may contribute to VLDL production, at least in mouse models.12 ApoC-III overexpression coincides with increased VLDL secretion whereas apoC-III deficiency results in the opposite phenotype. ApoC-III has multiple isoforms due to the presence of 0,1 or 2 sialic acid residues as a result of O-linked glycosylation.13 In a recent genome-wide association study, polymorphisms in GalNac-T2 transferase (GALNT2) were associated with increased plasma TG levels and decreased HDL cholesterol levels.14 Liver-specific Galnt2 results in a decrease of HDLc, whereas the knockdown of mouse liver galnt2 results in the opposite phenotype. It is tempting to see whether mutations in GALNT2 will have an impact on TG metabolism through modulation of apoC-III glycosylation.

The role of apoA-V in facilitating VLDL production remains unclear, since in vitro experiments could not establish any involvement of apoA-V.15 In line, data from studies in different mouse models do not support a direct role of apoA-V in VLDL production.16 Interestingly, it has recently been shown that apoA-V in the liver is co-localized with lipid droplets and that increased levels of hepatic apoA-V coincides with increased liver TG storage in mice, suggesting a role for apoA-V in mobilization of TG for VLDL production.17 It remains, however, to be determined how apoA-V finds its way into the plasma compartment.

The availability of TG partly determines the fate of apoB and consequently the secretion rate of VLDL particles.18 Insulin plays an essential role in this process. On the one hand insulin resistance increases TG lipolysis in adipose tissue, leading to an increased flux of FFA to the liver and thus an increased availability of TG cargo.19 On the other hand, insulin activates the regulatory machinery required for apoB synthesis. In an insulin resistant state this regulation is lost. As a consequence, VLDL production will increase leading to the generation of TG-rich atherogenic remnant particles, small dense LDL particles and TG-enriched HDL. 20

24

Chapter 2

Lipoprotein lipase (LPL)-mediated peripheral lipolysisVLDL and chylomicron particles provide fatty acids to tissues for energy as well as for storage. Lipolysis of TG by lipoprotein lipase (LPL) occurs in small capillaries in tissues that require fatty acids for storage (adipose tissue) or energy (heart and skeletal muscle). Fatty acids are directly taken up by CD36, whereas the liver will eventually clear the remnant particles. LPL is synthesized in parenchymal cells in these tissues and lipase maturation factor 1 (LMF1) is essential for proper folding and assembly of LPL.21 Subsequently, LPL is transported to the endothelial cell surface where it binds to glycosyl-phosphatidyl-inositol anchored high-density lipoprotein binding protein 1 (GPIHBP1).22 GPIHBP1 belongs to the family of lymphocyte-6 (Ly6) domain proteins. Human GPIHBP1 contains a heavily negative charged N-terminal domain, the Ly-6 domain consisting of 10

Figure 2 - Hepatic VLDL synthesis. Hepatic VLDL synthesis involved a series of intracellular processes enabling the formation of lipid-enriched apoB100 particles that can be excreted.


25

2

cysteine residues that form 5 double bonds and the gpi-anchor at the C-terminal end of the protein that serves as the attachment site to the extracellular leaflet of the cell membrane.23,24 The carboxyl-terminal sequence is removed in the endoplasmic reticulum, where cleavage occurs at one of the predicted sites for attachment of the gpi anchor in human GPIHBP1: residues 159, 153 and 154. The exact cellular localization of GPIHBP1 is still unknown. Based on data of other gpi-anchor proteins it is predicted that GPIHBP1 is localized in the lipid raft domains.25 GPIHBP1 provides the platform for LPL and TG-rich lipoproteins to come into close proximity, resulting into the hydrolysis of core TG. It remains an intriguing question why GPIHBP1 has so much more potency to bind LPL than the heparin sulfate proteoglycans (HSPGs), a core protein with negatively charged polysaccharide chains, that are abundantly expressed on the cell surface of capillaries. In line, mice with a deficiency in sulfation of HSPG in endothelial cells, due to a deficiency of GlcNAc N-deacetylase/N-sulfotransferase (Ndst1), exhibit normal plasma TG levels.26 More importantly, loss of function mutations in GPIHBP1 (so far 8 different mutations have been published) results in a severe hypertriglyceridemic phenotype providing a proof of concept that GPIHBP1 is essential for lipolysis to take place.24,27

Interestingly, GPIHBP1 may have an essential function in the transport of LPL through the endothelial cell layer towards the cell surface of the capillaries.28 Originally it was suggested that HSPGs were involved in the immobilization of LPL, however recent studies were not able to confirm this.26 Interestingly, a basement membrane proteoglycan, collagen XVIII, however, may also be involved in LPL translocation as recently shown in col18−/− mice and humans with a mutation in COL18 (Knobloch Syndrome) who develop a mild hypertriglyceridemic phenotype with reduced LPL activity and mass. Although Gpihbp1−/− mice display severe chylomicronemia, the Gpihbp1−/−/Angtpl4−/− mice has normal plasma TG levels, illustrating that Angptl4 destabilizes LPL which in the absence of functional GPIHBP1 become an unstable dysfunctional protein. The interesting finding of this model is that the lack of Gpihbp1 can be rescued on the premise that LPL remains in a stable form. It also illustrates that stabilized mouse LPL can perform its action without the presence of functional GPIHBP1.29

LPL action is dependent upon various co-factors. ApoC-II is an essential co-factor for LPL activation, whereas apoC-III may inhibits lipolysis. Indeed, low levels of apoC-III, due to a loss of function mutation, results in rapid postprandial clearance due to efficient LPL-mediated hydrolysis.30 The role of apoAV remains to be elucidated. Based on numerous in vitro experiments and the use of different genetically engineered mouse models, including apoa5 over-expression or deficiency, it has been suggested that apoA-V is required for efficient LPL action. Interestingly, in cell-based assays, apoAV-phospholipids complexes, but not apoE or apoC-III phospholipid particles, avidly bind to the acidic domain of GPIHBP1, although apoA-V could not compete with LPL binding, which additionally requires the Ly-6 domain of GPIHBP1 for binding to take place.31 Moreover, chylomicron binding requires both the Ly-6 as well as the acidic domain. Thus, chylomicron binding to GPIHBP1 only occurs in the presence of LPL and apo A-V does not seem to play an essential role in this process.

26

Chapter 2

If apoA-V is not directly involved in GPIHBP1-LPL mediated lipolysis, why does one then consistently observes a strong positive correlation with plasma TG levels and why are genetic variations in APOAV consistently a strong marker for plasma TG?14,15,32 These are intriguing questions that will hopefully be answered in the near future and will provide us with more insight into the function of apoA-V.

Figure 3 - Peripheral Lipolysis. GPIHBP1 is involved in the transport of LPL through the endothelial cell layer to the cell surface and is required for stabilisation of LPL. At the endothelial cell surface GPIHBP1 forms the platform to allow TG hydrolysis.

Hepatic remnant clearanceThe process of hepatic remnant clearance is complex. Although the detailed mechanism of the receptor mediated remnant uptake by the liver is still unclear, several endocytic hepatic receptors, present at hepatocyte microvilli, have been studied during the past few decades.33-35 It involves interactions between proteins located on the surface of the remnant particle that serve as ligands for hepatic receptors i.e. low density lipoprotein receptor (LDLr), LDL receptor related protein 1 (LRP1) and HSPG. ApoE is essential for hepatic remnant clearance that is illustrated by the observation that apoe−/− mice have massive accumulation of remnant particles.36 ApoE contains positively


27

2

charged residues, which favorably binds to negative charged domains on the hepatic receptors, whereas hepatic lipase (HL) and LPL mediates particle-receptor interaction with the LRP1 and HSPG receptors. 3,33

The LDL receptor was originally reported in to contribute to TRL clearance by the liver.37 However, as patients with genetic alterations of the LDLR are not characterized by hypertriglyceridemia, the search for other receptors continued.38 A family member of the LDL receptor involved in remnant clearance is the LDL receptor related protein 1 (LRP1).39,40 Predominantly, the binding of apoE-enriched remnants to HSPGs in the space of Disse was shown to depend upon final internalization via LRP1.41 Since combined LDLr and LRP1 deficiency did not result in hypertriglyceridemia, other pathways had to be present. Evidence from the early nineties already indicated the potential involvement of HSPG in hepatic remnant clearance.36 In a pioneering study hepatic remnant uptake was enhanced by heparin, a highly sulphated glycosaminoglycan, and inhibited by administration of heparanase, an enzyme that cleaves heparan-sulfate polysaccharides. 42 In line, remnant uptake was abolished in HSPG deficient cells. Recent studies using syndecan-1 null (Sdc1−/−) mice revealed that syndecan-1 is an important HSPG core protein in the liver. Sdc1−/− mice display elevated plasma TG levels and have impaired postprandial remnant clearance which is in agreement with earlier data in syndecan-1 overexpressing cells, indicating that syndecan-1 mediates TRL internalization.43,44 Syndecan-1 is synthesized in hepatocytes and undergoes posttranslational modifications that results in specific GAG chain length and sulfation patterns.45 Liver HSPGs contain an extreme large proportion of highly sulfated heparin-like structures located in the distal part of the GAG chain that allows lipoprotein binding.46 More than 50 enzymes are involved in the processing and degradation of HSPGs. Thus, chain length and sulfation pattern determine the biological activity of HSPG as was shown in different genetically modified animal models. Addition of sulfate groups by N-deacetylase/N-sulfotransferase 1 (Ndst1) and different O-sulfotransferases (HS2ST1, HS3ST1 and HS6ST1) provide a specific functional signature to HSPG. Mice lacking Ndst1 have a 50% reduction of hepatic HSPG sulfation.47 Consequently, plasma triglycerides were two-fold increased due to a reduced hepatic clearance of TRL. Mutant mice with defects in hepatic Hs2st1 and Hs6st1 expression also have increased plasma TG levels.48 Interestingly, in a streptozotocin-induced diabetic mouse model, a reduced hepatic Ndst1 mRNA expression was associated with impaired hepatic remnant clearance thus underscoring the effect of hyperglycemia on HSPG remodelling.49 Two proteins, heparanase (HSPE) and sulfotransferase 2 (SULF2) are involved in the extracellular remodeling of HSPG. Differences in expression of these proteins may lead to impaired hepatic HSPG-mediated TRL clearance. SULF2 is secreted into the extracellular space where it selectively modifies the 6-O-sulfate esters of HSPG.50 Interestingly, hepatic Sulf2 expression is highly increased in db/db mice and correlated with elevated plasma TG levels and impaired hepatic TRL clearance.51 It will be of interest to investigate whether inhibition of hepatic Sulf2 expression leads to an improvement of the dyslipidemic phenotype. Interestingly, transgenic mice overexpressing human heparanase also have increased plasma TG levels, a reduced hepatic TRL clearance and increased formation of fatty

28

Chapter 2

streaks, whereas no difference was found in postheparin LPL activity indicating normal peripheral LPL-mediated lipolysis.52

Collectively, these studies indicate that hepatic remnant clearance is a complicated process involving 3 different receptor-mediated pathways. Studies in genetically engineered mouse models revealed that deficiency of apoE results in the most severe remnant phenotype since all remnant uptake pathways will be partially eliminated. Abnormalities in HSPG result in a mild hypertriglyceridemic phenotype due to delayed clearance of remnant particles.

In conclusion, the importance of HSPG for optimal hepatic clearance of TRL is obvious, but in contrast with the LDLr, the relevance of HSPG in human lipid metabolism remains to be established.

Figure 4 - Hepatic TRL clearance. TRL clearance involves 3 hepatic receptors: LDLr, LRP1 and HSPG. Sulf2 is an extracellular protein that modulates HSPG thereby influencing hepatic TRL clearance.


29

2

PATHOPHYSIOLOGY OF HUMAN HYPERTRIGLYCERIDEMIA (FiGuRE 5) A mild to severe elevation of plasma TG (2-5 mmol/l) is a common feature in subjects with obesity, the metabolic syndrome or type 2 diabetes mellitus. Lipid abnormalities are a consequence of metabolic dysregulation resulting in mild to severe hypertriglyceridemia due to enhanced VLDL production, a delayed hepatic remnant clearance and mild disturbances in peripheral lipolysis.

More severe elevations of TG (between > 5–10 mmol/l) are observed in individuals diagnosed with familial combined hyperlipidemia (FCH), and familial hypertriglyceridemia (FHTG). FCH is present in 1:300 subjects, which makes FCH the most common genetic lipid disorder associated with increased risk for CVD. The hyperlipidemic phenotype can, however, be mixed including the presence of hypercholesterolemia. Pathophysiological studies have revealed abnormalities in VLDL production as well as hepatic TRL clearance.53 However, the heterogeneity of the lipid profile makes it difficult to reliably identify true cases, which has severely hampered the elucidation of the underlying metabolic abnormalities.54,55

Figure 5 - a schematic view of the occurrence of elevated plasma TG levels in the population. The higher the plasma TG levels the more impact of the genetic background will be observed.

30

Chapter 2

Dysbetalipoproteinemia or remnant removal disease is characterized by increased levels of remnant particles due to impaired hepatic clearance. Patients are carriers of dysfunctional apoE2. However, the presence of homozygosity for APOE2 alone is not sufficient to explain the pathophysiology of dysbetalipoproteinemia, since only 4% of all homozygous APOE2 carriers will develop severe remnant accumulation.

Severe hyperTG (TG> 10 mmol/l) is a hallmark of rare genetic disorder caused by loss of function of LPL due to mutations in LPL, APOC2, APOA5, GPIHBP1 or LMF1 as has been extensively described.56

TRIGLYCERIDES AND CARDIOvASCULAR DISEASESElevated triglyceride levels have been associated with cardiovascular risk in general population as well as in subjects with type 2 diabetes mellitus.57 Conversely, hypotriglyceridemia, due to a null mutation in APOC3, has been associated with longevity.30 It is difficult, however, to show that TG is a truly independent risk factor since hypertriglyceridemia is often part of the ‘metabolic’ dyslipidemic profile comprising low HDL levels, elevated levels of small dense LDL particles, as well as obesity and insulin resistance. Adjustment for these factors often led to very small effect sizes in association studies with positive hazard ratios. Another complicating factor relates to fasting versus postprandial TGs. Recent studies have suggested that nonfasting plasma TGs may be a superior risk marker for CV-risk.58 Indeed, chylomicron and VLDL remnant particles are able to penetrate the vessel wall.59 At the other site of the spectrum, the very large TRL particles as observed in LPL deficiency, do not associate with atherogenesis, but lead to an increased risk of pancreatitis.

FUTURE PERSPECTIvESElevated levels of triglycerides and TRL remnants are independent risk factors for the development of CVD, which bears direct relevance in view of the pandemia of obesity and type 2DM. The attention for hyperTG has been limited in view of the absence of effective compounds selectively lowering fasting and postprandial TGs. Novel insight into the TG metabolism has led to several therapeutics being developed which selectively target proteins involved in LPL-mediated lipolysis (ANGPTL4 and apoCIII), interfere with hepatic production (MTP and DGAT1) and/or that increase hepatic TRL clearance ( apoCIII and SULF2). It will be a challenge to evaluate what the impact of these selective TG lowering strategies is on cardiometabolic risk.


31

2

REFERENCE LIST1. Rader DJ, Hobbs HH. Disorders of Lipoprotein Metabolism. In: Fauci AS, Kasper DL, Longo DL, Braun-

wald E, Hauser SL, Jameson JL, Loscalzo J eds. Harrison’s Principle of Internal Medicine. New York: McGraw Hill; 2008.

2. Twickler T, Dallinga-Thie GM, Chapman MJ, Cohn JS. Remnant lipoproteins and atherosclerosis. Curr Atherosclerosis Reports 2005;7:140-7.

3. Williams KJ. Molecular processes that handle -- and mishandle -- dietary lipids. J Clin Invest 2008;118:3247-59.

4. Miller M, Stone NJ, Ballantyne C, et al. Triglycerides and cardiovascular disease: a scientific state-ment from the American Heart Association. Circulation 2011;123:2292-333.

5. Iqbal J, Hussain MM. Intestinal lipid absorption. Am J Physiol Endocrinol Metab 2009;296:E1183-94.

6. Yen CL, Stone SJ, Koliwad S, Harris C, Farese RV, Jr. Thematic review series: glycerolipids. DGAT en-zymes and triacylglycerol biosynthesis. J Lipid Res 2008;49:2283-301.

7. Calandra S, Tarugi P, Speedy HE, Dean AF, Bertolini S, Shoulders CC. Mechanisms and genetic deter-minants regulating sterol absorption, circulating LDL levels, and sterol elimination: implications for classification and disease risk. J Lipid Res 2011;52:1885-926.

8. Cefalu AB, Calvo PL, Noto D, et al. Variable phenotypic expression of chylomicron retention disease in a kindred carrying a mutation of the Sara2 gene. Metabolism 2010;59:463-7.

9. Verges B. Abnormal hepatic apolipoprotein B metabolism in type 2 diabetes. Atherosclerosis 2010;211:353-60.

10. Dallinga-Thie GM, van Tol A, Hattori H, van Vark-van der Zee LC, Jansen H, Sijbrands EJ. Plasma apoli-poprotein A5 and triglycerides in type 2 diabetes. Diabetologia 2006;49:1505-11.

11. Dallinga-Thie GM, Berk-Planken II, Bootsma AH, Jansen H. Atorvastatin decreases apolipoprotein C-III in apolipoprotein B-containing lipoprotein and HDL in type 2 diabetes: a potential mechanism to lower plasma triglycerides. Diabetes Care 2004;27:1358-64.

12. Aalto-Setala K, Fisher EA, Chen X, et al. Mechanism of hypertriglyceridemia in human apolipoprotein (apo) CIII transgenic mice. Diminished very low density lipoprotein fractional catabolic rate associ-ated with increased apo CIII and reduced apo E on the particles. J Clin Invest 1992;90:1889-900.

13. J aeken J. Congenital disorders of glycosylation. Ann N Y Acad Sci 2010;1214:190-8.

14. Teslovich TM, Musunuru K, Smith AV, et al. Biological, clinical and population relevance of 95 loci for blood lipids. Nature 2010;466:707-13.

15. Nilsson SK, Heeren J, Olivecrona G, Merkel M. Apolipoprotein A-V; a potent triglyceride reducer. Atherosclerosis 2011.

16. Merkel M, Loeffler B, Kluger M, et al. Apolipoprotein AV accelerates plasma hydrolysis of triglyc-eride-rich lipoproteins by interaction with proteoglycan-bound lipoprotein lipase. J Biol Chem 2005;280:21553-60.

17. Shu X, Nelbach L, Ryan RO, Forte TM. Apolipoprotein A-V associates with intrahepatic lipid droplets and influences triglyceride accumulation. Biochim Biophys Acta 2010;1801:605-8.

18. Fisher EA, Khanna NA, McLeod RS. Ubiquitination regulates the assembly of VLDL in HepG2 cells and is the committing step of the apoB-100 ERAD pathway. J Lipid Res 2011;52:1170-80.

32

Chapter 2

19. Taskinen MR. Diabetic dyslipidemia. Atheroscler Suppl 2002;3:47-51.

20. Adiels M, Olofsson SO, Taskinen MR, Boren J. Overproduction of very low-density lipoproteins is the hallmark of the dyslipidemia in the metabolic syndrome. Arterioscler Thromb Vasc Biol 2008;28:1225-36.

21. Peterfy M, Ben-Zeev O, Mao HZ, et al. Mutations in LMF1 cause combined lipase deficiency and severe hypertriglyceridemia. Nat Genet 2007;39:1483-7.

22. Beigneux AP, Davies BS, Gin P, et al. Glycosylphosphatidylinositol-anchored high-density lipopro-tein-binding protein 1 plays a critical role in the lipolytic processing of chylomicrons. Cell Metab 2007;5:279-91.

23. Dallinga-Thie GM, Franssen R, Mooij HL, et al. The metabolism of triglyceride-rich lipoproteins revis-ited: new players, new insight. Atherosclerosis 2010;211:1-8.

24. Young SG, Davies BS, Voss CV, et al. GPIHBP1, an endothelial cell transporter for lipoprotein lipase. J Lipid Res 2011.

25. Simons K, Gerl MJ. Revitalizing membrane rafts: new tools and insights. Nat Rev Mol Cell Biol 2010;11:688-99.

26. Weinstein MM, Yin L, Beigneux AP, et al. Abnormal patterns of lipoprotein lipase release into the plasma in GPIHBP1-deficient mice. J Biol Chem 2008;283:34511-8.

27. Rios JJ, Shastry S, Jasso J, et al. Deletion of GPIHBP1 causing severe chylomicronemia. J Inherit Metab Dis 2011.

28. Davies BS, Beigneux AP, Barnes RH, 2nd, et al. GPIHBP1 is responsible for the entry of lipoprotein lipase into capillaries. Cell Metab 2010;12:42-52.

29. Sonnenburg WK, Yu D, Lee EC, et al. GPIHBP1 stabilizes lipoprotein lipase and prevents its inhibition by angiopoietin-like 3 and angiopoietin-like 4. J Lipid Res 2009;50:2421-9.

30. Pollin TI, Damcott CM, Shen H, et al. A null mutation in human APOC3 confers a favorable plasma lipid profile and apparent cardioprotection. Science 2008;322:1702-5.

31. Gin P, Beigneux AP, Voss C, et al. Binding preferences for GPIHBP1, a glycosylphosphatidylinositol-anchored protein of capillary endothelial cells. Arterioscler Thromb Vasc Biol 2011;31:176-82.

32. Wang J, Ban MR, Kennedy BA, et al. APOA5 genetic variants are markers for classic hyperlipopro-teinemia phenotypes and hypertriglyceridemia. Nat Clin Pract Cardiovasc Med 2008;5:730-7.

33. Havel RJ, Hamilton RL. Hepatic catabolism of remnant lipoproteins: where the action is. Arterioscler Thromb Vasc Biol 2004;24:213-5.


35. Mahley RW, Hui DY, Innerarity TL, Beisiegel U. Chylomicron remnant metabolism. Role of hepatic lipoprotein receptors in mediating uptake. Arteriosclerosis 1989;9:I14-8.

36. Mahley RW, Ji ZS. Remnant lipoprotein metabolism: key pathways involving cell-surface heparan sulfate proteoglycans and apolipoprotein E. J Lipid Res 1999;40:1-16.

37. Kita T, Brown MS, Bilheimer DW, Goldstein JL. Delayed clearance of very low density and intermedi-ate density lipoproteins with enhanced conversion to low density lipoprotein in WHHL rabbits. Proc Natl Acad Sci USA 1982;79:5693-7.


33

2

38. Rubinsztein DC, Cohen JC, Berger GM, van der Westhuyzen DR, Coetzee GA, Gevers W. Chylomicron remnant clearance from the plasma is normal in familial hypercholesterolemic homozygotes with defined receptor defects. J Clin Invest 1990;86:1306-12.

39. Beisiegel U, Weber W, Ihrke G, Herz J, Stanley KK. The LDL-receptor-related protein, LRP, is an apoli-poprotein E-binding protein. Nature 1989;341:162-4.

40. Rohlmann A, Gotthardt M, Hammer RE, Herz J. Inducible inactivation of hepatic LRP gene by cre-mediated recombination confirms role of LRP in clearance of chylomicron remnants. J Clin Invest 1998;101:689-95.

41. Mahley RW, Huang Y. Atherogenic remnant lipoproteins: role for proteoglycans in trapping, transfer-ring, and internalizing. J Clin Invest 2007;117:94-8.


43. Stanford KI, Bishop JR, Foley EM, et al. Syndecan-1 is the primary heparan sulfate proteoglycan me-diating hepatic clearance of triglyceride-rich lipoproteins in mice. J Clin Invest 2009;119:3236-45.

44. Fuki IV, Kuhn KM, Lomazov IR, et al. The syndecan family of proteoglycans. Novel receptors mediat-ing internalization of atherogenic lipoproteins in vitro. J Clin Invest 1997;100:1611-22.

45. Foley EM, Esko JD. Hepatic heparan sulfate proteoglycans and endocytic clearance of triglyceride-rich lipoproteins. Prog Mol Biol Transl Sci 2010;93:213-33.

46. Bishop JR, Foley E, Lawrence R, Esko JD. Insulin-dependent diabetes mellitus in mice does not alter liver heparan sulfate. J Biol Chem 2010;285:14658-62.

47. MacArthur JM, Bishop JR, Stanford KI, et al. Liver heparan sulfate proteoglycans mediate clear-ance of triglyceride-rich lipoproteins independently of LDL receptor family members. J Clin Invest 2007;117:153-64.

48. Stanford KI, Wang L, Castagnola J, et al. Heparan sulfate 2-O-sulfotransferase is required for triglyc-eride-rich lipoprotein clearance. J Biol Chem 2010;285:286-94.

49. Goldberg IJ, Hu Y, Noh HL, et al. Decreased lipoprotein clearance is responsible for increased choles-terol in LDL receptor knockout mice with streptozotocin-induced diabetes. Diabetes 2008;57:1674-82.

50. Morimoto-Tomita M, Uchimura K, Werb Z, Hemmerich S, Rosen SD. Cloning and characteriza-tion of two extracellular heparin-degrading endosulfatases in mice and humans. J Biol Chem 2002;277:49175-85.

51. Chen K, Liu ML, Schaffer L, et al. Type 2 diabetes in mice induces hepatic overexpression of sulfatase 2, a novel factor that suppresses uptake of remnant lipoproteins. Hepatology 2010;52:1957-67.

52. Planer D, Metzger S, Zcharia E, Wexler ID, Vlodavsky I, Chajek-Shaul T. Role of heparanase on hepatic uptake of intestinal derived lipoprotein and fatty streak formation in mice. PLoS One 2011;6:e18370.

53. Suviolahti E, Lilja HE, Pajukanta P. Unraveling the complex genetics of familial combined hyperlipi-demia. Ann Med 2006;38:337-51.

54. Horswell SD, Ringham HE, Shoulders CC. New technologies for delineating and characterizing the lipid exome: prospects for understanding familial combined hyperlipidemia. J Lipid Res 2009;50 Suppl:S370-5.

34

Chapter 2

55. Wierzbicki AS, Graham CA, Young IS, Nicholls DP. Familial combined hyperlipidaemia: under - defined and under - diagnosed? Curr Vasc Pharmacol 2008;6:13-22.

56. Johansen CT, Hegele RA. Genetic bases of hypertriglyceridemic phenotypes. Curr Opin Lipidol 2011;22:247-53.

57. Turner RC, Millns H, Neil HA, et al. Risk factors for coronary artery disease in non-insulin dependent diabetes mellitus: United Kingdom Prospective Diabetes Study (UKPDS: 23). Brit Med J 1998;316:823-8.


59. Proctor SD, Vine DF, Mamo JC. Arterial permeability and efflux of apolipoprotein B-containing li-poproteins assessed by in situ perfusion and three-dimensional quantitative confocal microscopy. Arterioscler Thromb Vasc Biol 2004;24:2162-7.

Hassing HC, Visser ME, Boekholdt SM, Wareham NJ, Khaw KT, Zwinderman AH, Stroes ESG, Dallinga-Thie GM

Manuscript in preparation

3 A MULTILOCUS GENETIC RISk SCORE FOR HYPERTRIGLYCERIDEMIA

38

Chapter 3

ABSTRACTintroduction - Plasma triglyceride (TG) concentration is independently associated with coronary artery disease (CAD). DNA sequence variations in TG-related genes have been shown to predict TG increase, but the use of TG-related single nucleotide polymorphisms (SNPs) in predicting future risk of CVD is equivocal. In the present study, we set out to evaluate whether a gene score composed of a minimal number of SNPs associated with plasma TG levels improves CAD risk prediction.

Methods - We assessed associations between 26 SNPs in TG-related genes and plasma TG levels in a prospective case-control study in the EPIC-Norfolk cohort comprising 2105 cases and 2131 controls. We combined those SNPs that were significantly associated with plasma TG into a single model and developed a gene risk score based on the TG-raising alleles of those SNPs driving the association between the combined model and plasma TG. Subsequently, we evaluated the association between this gene score and plasma lipids and CAD risk.

Results - A gene score composed of only three SNPs in LPL and APOA5 was linearly associated with plasma TG concentrations (+0.32 mmol/l [95% CI 0.257-0.379] per allele change, p<0.0001) and other lipid parameters representing an atherogenic lipid profile including decreased LDL size, increased LDL number, increased VLDL particle number and decreased HDL particle size. In line, we observed a trend towards a positive linear association between TG gene score and risk of CAD (odds ratio of 1.104 per allele increase in score, 95% CI 0.996-1.225). The risk of CAD was elevated in individuals with the highest gene score compared to those with lowest gene scores (odds ratio 1.88 [95% CI 1.11-3.18]; p=0.02).

Conclusion - A concise TG gene score predicts both elevated TG levels as well as future CAD risk. These data lend further support to a causal relation between TG elevation and CAD.

A genetic risk score for hypertriglyceridemiA

39

3

INTRODUCTIONAn increasing number of epidemiological studies report positive associations between - both fasting and nonfasting- plasma triglyceride (TG) concentration and risk of coronary artery disease (CAD) beyond low-density lipoprotein cholesterol (LDL-C). 1-3 Early identification of subjects at risk for developing hypertriglyceridemia (HTG) could reduce risk of CAD by inducing early lifestyle modification or pharmacological intervention. Since plasma TGs reflects daily dietary fat consumption, plasma TG measurements are prone to marked fluctuations. A DNA sequence variant could be an indicator for lifelong exposure to elevated plasma TG levels. Therefore DNA sequence variations may add predictive information beyond individual measurements of more volatile lipid measurements. Over the years, multiple single nucleotide polymorphisms (SNPs) with significant effects on TG concentrations have been identified including SNP’s in lipoprotein lipase (LPL) and APOA5. 4, 5 Recent genome wide association studies (GWAS) 6-14 were able to replicate these findings and additionally identified novel promising loci, such as in genes encoding for TRIB1, GCKR and ANGPTL4, associated with plasma TG levels. 15, 16 However the potential of genetic testing is limited by the fact that each sequence variant explains only a modest fraction of the variation in lipid levels 12. Therefore we hypothesized that a combination of different polymorphisms associated with plasma TG levels in a gene score improved risk prediction. We therefore evaluated, based on a panel of validated TG associated SNPs, the best combination of SNPs that predicted plasma triglyceride levels as well as a first cardiovascular event in a prospective case-control study in the EPIC-Norfolk cohort comprising 2105 cases and 2131 controls.

METHODSEPiC-Norfolk prospective case-control studyThe European Prospective Investigation of Cancer (EPIC)-Norfolk is a prospective population study of apparently healthy men and woman of which the design and characteristics have been described previously 17. For associations between genotypes and plasma triglyceride levels a maximum of 2105 cases and 2131 controls were available. Cases were identified as those having fatal or non fatal coronary artery disease (CAD) during follow-up. Participants who reported a history of heart attack or stroke at the baseline clinic visit were excluded. Controls were study participants who remained free of any cardiovascular disease.

sNP selection and GenotypingWe selected leading SNPs that were significantly associated with plasma TG levels in the large GWAS study 16 as well as SNPs originally published in single association studies (LPL, APOA5, APOA4, ANGPTL4, GALNT2, APOC3, GCKR, TRIB1, see supplemental Table 1). We also included tag SNPs for genes that have an essential role in TG metabolism such as APOC2, LMF1 and GPIHBP1. Patients with complete loss of function mutations in these genes suffer from severe hypertriglyceridemia 18 but no data so far have been generated showing the influence of genetic variation in these genes

40

Chapter 3

and the effect on plasma TG levels. Tagging SNPs were selected as described previously using the publicly available Hapman program. 19 Genotyping was performed using allelic discrimination with VIC- and FAM- labeled probes designed by Applied Biosystems. PCR conditions were denaturation for 10 min at 95°C, followed by 40 cycles (30 sec 92°C, 45 sec 60°C). Tagman PCR assay mix was obtained from Applied Biosystems. All SNPs were in Hardy Weinberg equilibrium with a probability value of >0.001.

Biochemical analysesNonfasting blood samples were taken by venipuncture at the baseline clinic visit. Blood samples were stored at -80 before analysis. Circulating levels of total cholesterol, HDL cholesterol, and triglycerides were measured on fresh samples with the RA 1000 (Bayer Diagnostics, Basingstoke, United Kingdom), and LDL cholesterol levels were calculated with the Friedewald formula. 20 Lipid particle number and size were measured with an automated nuclear magnetic resonance spectroscopic assay as described. 21 Other biochemical and hematological measurements involved

standard assays, as previously described 17. Samples were analyzed in random order to avoid systemic bias. Researchers and laboratory personnel had no access to identifiable information, and could identify samples by number only.

statistical analysesAll SNPs were initially determined in a random subset of the cohort (n=2,649) whereas those SNPs selected for final gene score were determined in the complete cohort (n=4,236). The relationship between single SNPs and plasma TG levels was determined by a linear regression model. Baseline characteristics were compared between cases and controls with a mixed-effect model for continuous variables and with conditional logistic regression for categorical variables. Variables with skewed distribution were log-transformed before being used as continuous variables in statistical analyses. Associations between genotype score and plasma lipid parameters were calculated using linear regression for continuous variables and the chi-square test for trend for categorical variables. To estimate the relative risk of CAD, conditional logistic regression was used to calculate odds ratios (ORs) and 95% confidence intervals in genetic risk score, using the lowest TG gene score as the reference category. A probability value of <0,05 was considered statistically significant. All statistical analyses were performed using SPSS version 16.0.2.

RESULTSbaseline characteristicsA complete dataset was available for 2105 cases and 2131 controls. Biochemical characteristics and cardiovascular risk factors in cases and controls are presented in supplemental Table 2. As expected individuals who developed CAD during follow up were more likely to have cardiovascular risk factors. Plasma cholesterol, LDL cholesterol, and triglyceride levels were significantly higher in cases compared to controls whereas plasma HDL cholesterol levels were significantly lower.


41

3

associations between sNPs and plasma triglycerides .We determined major GWAS SNPs in TRIB1, GCKR, APOA5, LPL, APOC2, APOC3, known missense variants in ANGPTL4 and APOA4 and tagSNPs in LMF1 and GPIHBP1, both playing a pivotal role in LPL function and thus TG metabolism, in a random subset of the EPIC cohort (n=2,649). Details of all SNPs are documented in supplemental Table 1. We did not find any associations between genetic variations in LMF1 or GPIHBP1 and plasma TG levels. Similarly no significant association could be found for SNPs in APOC2, APOA4, GCKR, APOC3 and GALNT2. The SNPs in ANGPTL4 (rs1044250, rs116843064), TRIB1 (rs2954029), LPL (rs328, rs12678919) and APOA5 (rs75423577, rs3135506, rs662799) were all significantly associated with plasma TG levels with P values varying from <0.05 - <0.001 (see supplemental Table 3). Based on this genotype-phenotype analysis we selected 8 SNPs that were associated with plasma triglyceride levels to develop a gene score including the number of unfavourable, TG raising, alleles carried by each subject for each of the selected SNPs. To explore which SNPs appeared to drive the association with plasma triglyceride levels, we used backward elimination (retention threshold P<0.05) in a model that included all 8 SNPs. Only three SNPs remained significant following backward elimination: the rs3135506, the rs662799 (both in APOA5) and rs328 (LPL). Because the association with triglyceride levels was driven byb these three SNPs, they were analyzed in the complete cohort (n=4,236) and subsequently used to develop the final gene risk score. In the complete cohort carriership of the rs3135506 G allele, the rs662799 G allele (both in APOAV) and the rs328 C allele (in LPL) was related to higher plasma TG concentrations (+0.35 mmol/l [95% CI 0.237-0.464] per rs3135506 G allele, +0.35 mmol/l [95% CI 0.235-0.463] per rs662799 G allele and +0.39 mmol/l [95% CI 0.138- 0.641] per rs328 C allele), see Table 1.

Table 1 - Plasma triglyceride concentration according to triglyceride related sNPs.

Gene SNP N GT TG (mmol/l) %Var P† P‡

APOA5 rs3135506 3392 CC 1.70 (1.20-2.50) 0.8 <0.001 <0.001

420 CG 1.90 (1.40-2.90)

22 GG 2.10 (1.50-3.50)

APOA5 rs662799 3473 AA 1.70 (1.20-2.50) 0.9 <0.001 <0.001

434 AG 1.90 (1.40-2.80)

25 GG 2.60 (1.60-4.30)

LPL rs328 3169 CC 1.80 (1.30-2.60) 1.3 <0.001 <0.001

724 CG 1.60 (1.10-2.20)

45 GG 1.40 (1.00-2.00)

Number of individuals (N) and triglyceride serum concentrations according to polymorphism are presented as median and 25th and 75th percentile. GT; genotype. P†= unadjusted p-value, P‡= p-value adjusted for age, sex, waist, body mass index and diabetes

42

Chapter 3

Gene risk score and plasma triglyceride-parameters.Our final gene risk score was based upon the presence of TG raising alleles for each of the three selected SNPs resulting in a score of 0-6. Genotype for all three SNPs was available for 3,787 individuals (89.4%). Due to a low number, we combined individuals with a score ≥4 into one group. Demographic and lifestyle parameters did not differ significantly by TG gene risk score (Table 2). A higher TG gene risk score was strongly associated with higher plasma TG concentrations in a stepwise manner (+0.32 mmol/l [95% CI 0.257-0.379] per allele change, p<0.0001) independent from other risk factors such as diabetes and BMI (Figure 1). In line, plasma apoB and apoAV levels were significantly higher in individuals with upper gene risk score whereas plasma LPL levels were significantly lower (p<0.001).

Table 2 - Demographic and lifestyle parameters of individuals according to triglyceride genotype score

0 1 2 3 ≥4 P

N 28 547 2540 608 64

Men, n (%) 18 (64%) 353 (65%) 1689

(66%) 407 (67) 40 (63%) 0.883

Age, years 65.1±8.1 65.2±8.2 65.1±7.8 64.9±8.0 65.3±7.4 0.978

Body Mass Index, kg/m2 26.7±4.2 26.5±3.8 26.8±3.7 26.8±3.7 27.4±4.0 0.263

Waist Circumference, cm 92.1±12.1 92.2±11.7 93.1±11.7 93.1±11.4 92.9±10.7 0.574

Current smoker n, (%) 2 (7.1) 63 (11.7) 287 (11.4) 60 (9.9) 6 (9.7) 0.767

Diabetes Mellitus n, (%) 1 (3.6) 21 (3.8) 105 (4.1) 19 (3.1) 4 (6.3) 0.693

Systolic blood pressure mmHg 142±16 140±18 141±18 140±18 145±20 0.258

Diastolic blood pressure, mmHg 84±9 84±11 85±11 84±12 87±12 0.205

Data are presented as number (percentage), mean ± SD or median and 25th and 75th percentile.

Gene risk score and other lipid parameters Associations between the gene risk score and lipid parameters are displayed in Table 3. Across the gene risk score categories we observe a gene score dependent increase towards an atherogenic lipoprotein profile. Plasma levels of total cholesterol and LDL cholesterol were significantly increased at higher gene risk score categories whereas HDL cholesterol levels were significantly lower (all p<0.001). Interestingly, a higher gene risk score coincided with the presence of smaller LDL particle size and higher LDL particle number (p<0.001). In line, VLDL particle number was increased


43

30

12

3>4

1.0

1.5

2.0

2.5

3.0

3.5

p <0

.001

r² 2

.9%

p‡<0

.001

Gen

esco

re

Triglycerides (mmol/L)

01

23

>410

0

150

200

250

300

350

p <0

.001

r² 3

.1%

p‡<0

.001

Gen

esco

re

APOAV (ng/ml)

01

23

>4050100

150

200

p <0

.001

r² 7

.8%

p‡<0

.001 G

enes

core

LPL (ng/ml)

01

23

>480100

120

140

160

180

p <0

.001

r² 1

.6%

p‡<0

.001

Gen

esco

reAPOB (mg/dl)

Figu

re 1

- P

lasm

a tr

igly

cerid

e-pa

ram

eter

s ac

cord

ing

to t

rigly

cerid

e ge

noty

pe s

core

. Pl

asm

a tr

igly

cerid

e, a

poAV

, LP

L an

d ap

oB

conc

entr

ation

(mea

n ±

95%

CI)

per t

rigly

cerid

e ge

ne sc

ore.

P‡

= p

valu

e fo

llow

ing

corr

ectio

n fo

r gen

der,

age,

bod

y-m

ass i

ndex

, wai

st,

LDL

and

diab

etes

.

44

Chapter 3

Tabl

e 3

- Dis

trib

ution

of l

ipid

s, li

pid

parti

cle

num

ber,

parti

cle

size

and

lipo

prot

eins

acc

ordi

ng to

gen

e sc

ore

01

23

>4R

P†P‡

Parti

cipa

nts,

n (%

)28

(0.7

)54

7 (1

4.4)

2540

(67.

1)60

8 (1

6.1)

64 (1

.7)

Trig

lyce

ride,

mm

ol/l

1.25

(1.0

0-19

0)1.

50 (1

.10-

2.20

)1.

80 (1

.20-

2.50

)2.

00 (1

.40-

2.90

)2.

10 (1

.50-

3.55

)0.

169

<0.0

01<0

.001

Tota

l cho

lest

erol

, mm

ol/L

5.99

±0.9

76.

22±1

.15

6.34

±1.1

56.

54±1

.31

6.74

±1.1

50.

089

<0.0

01<0

.001

LDL-

chol

este

rol,

mm

ol/L

3.67

±0.8

24.

05±1

.00

4.12

±1.0

04.

24±1

.08

4.30

±0.9

10.

063

<0.0

01<0

.001

LDL

parti

cle

size,

nm

21.0

±0.6

721

.1±0

.53

21.0

±0.6

020

.8±0

.67

20.8

±0.6

6-0

.108

<0.0

01<0

.001

LDL

parti

cle,

nm

ol/l

1492

±454

1598

±406

1640

±455

1759

±500

1878

±559

0.12

1<0

.001

<0.0

01

HDL-

chol

este

rol,

mm

ol/L

1.64

±0.4

31.

39±0

.41

1.33

±0.3

81.

26±0

.36

1.25

±0.3

9-0

.106

<0.0

01<0

.001

HDL-

parti

cle

size,

nm

9.04

±0.6

28.

94±0

.49

8.86

±0.4

68.

80±0

.44

8.75

±0.3

6-0

.097

<0.0

01<0

.001

HDL

parti

cle,

nm

ol/l

35.7

±6.7

33.7

±5.5

33.8

±5.5

33.6

±6.0

34.9

±5.6

-0.0

020.

925

0.97

3

VLDL

par

ticle

size

, nm

52.0

±7.3

50.7

±8.9

52.1

±8.8

52.5

±8.8

52.3

±9.4

0.05

0<0

.05

<0.0

5

VLDL

par

ticle

, nm

ol/l

84.4

±36.

291

.4±3

0.7

97.8

±32.

510

8.2±

34.5

116.

4±33

.60.

163

<0.0

01<0

.001

ApoA

5, n

g/m

l16

0±57

183±

7219

8±81

230±

149

291±

162

0.17

7<0

.001

<0.0

01

ApoB

, mg/

dL11

5±30

128±

3113

3±32

140±

3614

7±40

0.12

5<0

.001

<0.0

01

LPL,

ng/

ml

158±

5911

4±55

67±4

665

±41

65±3

0-0

.279

<0.0

01<0

.001

Data

are

pre

sent

ed a

s num

ber (

perc

enta

ge),

med

ian

and

25th

and

75th

per

centi

le o

r mea

n ±

SD. R

= Pe

arso

n or

Spe

arm

an co

rrel

ation

, P†=

una

djus

ted

p-va

lue,

P‡=

p-

valu

e ad

just

ed fo

r age

, sex

, dia

bete

s, w

aist

and

bod

y-m

ass i

ndex

.


45

3

significantly (p<0.001). Concomitantly HDL-C particle size was smaller but no difference in HDL particle number was observed.

Gene risk score and coronary artery diseaseWe found that the odds ratio for CAD in the highest (≥4) versus the lowest (≤1) gene risk score was 1.88 (95% CI 1.11-3.18; p=0.02) irrespectively of covariates (Table 4). Accordingly, we observed a trend towards a positive linear association between TG gene score and risk of CAD (odds ratio of 1.104 per allele increase in score, 95% CI 0.996-1.225).

Table 4 - Risk of future coronary artery disease per triglyceride genescore category.

Genescore ≤1 2 3 ≥4 P†

Cases (%)/controls 261 (45%)/ 301 1272 (50%) / 1268 298 (49%) / 310 39 (61%)/ 25

OR (95% CI) 1.00 1.21 (1.01-1.45) 1.16 (0.92-1.45) 1.88 (1.11-3.18) 0.057

p-value 0.043 0.212 0.020

OR‡ (95% CI) 1.00 1.22 (1.02-1.46) 1.14 (0.91-1.44) 1.79 (1.05-3.04) 0.066

p-value‡ 0.034 0.251 0.033

Odds radtio (OR) and 95% CI, relative to TG genescore ≤1, for each TG genescore category. † p value for linearity. ‡ OR adjusted for age, sex, body-mass index, diabetes and smoking with corresponding p-value.

DISCUSSIONIn the present study we composed a gene risk score based on only three TG modulating SNPs (APOA5 rs3135506, rs662799 and LPL rs328). A higher gene risk score was associated with increased plasma TG in a stepwise manner as well as other parameters representing an atherogenic lipid profile comprising LDL particle number and size. The risk of future CAD was elevated in individuals with the highest gene score compared to those with lowest gene scores.

In our gene risk score, the presence of each TG-raising allele is associated with an increase in plasma triglyceride concentration of 0.32 mmol/l per allele change. This effect was independent from other risk factors affecting plasma TG such as diabetes and BMI. For this TG gene risk score the odds ratio for CAD for the highest versus the lowest gene score was 1.88 (p=0.02) irrespective of covariates. In line, we found a trend toward linearity for CAD and gene risk score. These data provide support for the hypothesis that a TG gene risk score is a valid predictor not only for plasma TG levels but also for future CAD risk.

Previous GWAS findings have identified multiple genetic loci associated with plasma triglycerides which has been confirmed by a large meta-analysis in >100,000 subjects, identifying 32 loci

46

Chapter 3

harbouring common variants that contribute to variations in plasma TG concentration. 16 The effect sizes of most of these recently identified loci are however small. For clinical use the number of SNPs to be used in a gene risk score should restricted to a bear minimum for both practical and economical purposes. Therefore we set out to identify the optimal combination with maximal predictive ability using the lowest number of genetic variants. We observed that a robust predictive model can be build using only three SNPs in LPL and APOA5 that display the most vigorous effect.

The metabolism of triglyceride-rich lipoproteins, HDL and LDL are closely linked. Hence, the gene risk score also coincides with other features of the atherogenic lipid profile such as increased LDL particle number, smaller LDL particle size and increased VLDL particle number and decreased HDL particle size. These findings lend further support to the ability of the gene risk score to predict the presence of an adverse lipid profile. It is widely accepted that APOAV and LPL are involved in TG homeostasis 4, 22, 23 with severe, rare mutations leading to hypertriglyceridemic phenotypes. 18 Of interest, a higher gene risk score was also found to be associated with lower pre-heparin LPL plasma levels implying a reduced LPL availability. The high gene risk score group also had elevated plasma apoAV levels. In earlier association studies it has already been reported that plasma apoAV levels were positively associated with plasma TG levels 24 underscoring the need for more functional studies on apoAV in humans to understand its function.

Since human plasma triglyceride levels display a large diurnal variation, a TG gene score can be expected to offer a better reflection of prolonged exposure to elevated plasma TG levels. As known, TG levels are invariably associated to lower HDL-C cholesterol and the presence of small dense LDL-C, making it difficult to draw conclusions on direct causality of increased TG for increased CVD risk. Recent data, however, have corroborated that the association between elevated fasting as well as non-fasting TG levels and CVD is robust.

In conclusion, we show that a gene score combining only 3 SNPs in APOA5 and LPL, respectively, predicts both increased plasma triglyceride levels as well as a higher risk of CAD. These data provide lend further support to the concept that a TG gene score is a valid predictor for both elevated TG levels as well as future CAD risk.


47

3

REFERENCE LIST1. Hokanson JE, Austin MA. Plasma triglyceride level is a risk factor for cardiovascular disease inde-

pendent of high-density lipoprotein cholesterol level: a meta-analysis of population-based prospec-tive studies. J Cardiovasc Risk 1996;3:213-219.



4. Talmud PJ, Martin S, Taskinen MR et al. APOA5 gene variants, lipoprotein particle distribution, and progression of coronary heart disease: results from the LOCAT study. J Lipid Res 2004;45:750-756.

5. Vaessen SF, Schaap FG, Kuivenhoven JA et al. Apolipoprotein A-V, triglycerides and risk of coronary artery disease: the prospective Epic-Norfolk Population Study. J Lipid Res 2006;47:2064-2070.

6. Kathiresan S, Melander O, Guiducci C et al. Six new loci associated with blood low-density lipoprotein cholesterol, high-density lipoprotein cholesterol or triglycerides in humans. Nat Genet 2008;40:189-197.

7. Willer CJ, Sanna S, Jackson AU et al. Newly identified loci that influence lipid concentrations and risk of coronary artery disease. Nat Genet 2008;40:161-169.

8. Kooner JS, Chambers JC, Aguilar-Salinas CA et al. Genome-wide scan identifies variation in MLXIPL associated with plasma triglycerides. Nat Genet 2008;40:149-151.

9. Aulchenko YS, Ripatti S, Lindqvist I et al. Loci influencing lipid levels and coronary heart disease risk in 16 European population cohorts. Nat Genet 2009;41:47-55.

10. Sabatti C, Service SK, Hartikainen AL et al. Genome-wide association analysis of metabolic traits in a birth cohort from a founder population. Nat Genet 2009;41:35-46.

11. Kathiresan S, Willer CJ, Peloso GM et al. Common variants at 30 loci contribute to polygenic dyslipi-demia. Nat Genet 2009;41:56-65.

12. Hegele RA. Plasma lipoproteins: genetic influences and clinical implications. Nat Rev Genet 2009;10:109-121.

13. Johansen CT, Kathiresan S, Hegele RA. Genetic determinants of plasma triglycerides. J Lipid Res 2011;52:189-206.

14. Jorgensen AB, Frikke-Schmidt R, West AS, Grande P, Nordestgaard BG, Tybjaerg-Hansen A. Geneti-cally elevated non-fasting triglycerides and calculated remnant cholesterol as causal risk factors for myocardial infarction. Eur Heart J 2012.

15. Johansen CT, Wang J, Lanktree MB et al. Excess of rare variants in genes identified by genome-wide association study of hypertriglyceridemia. Nat Genet 2010;42:684-687.

16. Teslovich TM, Musunuru K, Smith AV et al. Biological, clinical and population relevance of 95 loci for blood lipids. Nature 2010;466:707-713.

17. y N, Oakes S, Luben R et al. EPIC-Norfolk: study design and characteristics of the cohort. European Prospective Investigation of Cancer. Br J Cancer 1999;80 Suppl 1:95-103.

18. Surendran RP, Visser ME, Heemelaar S et al. Mutations in LPL, APOC2, APOA5, GPIHBP1 and LMF1 in patients with severe hypertriglyceridaemia. J Intern Med 2012;272:185-196.

48

Chapter 3

19. Vergeer M, Boekholdt SM, Sandhu MS et al. Genetic variation at the phospholipid transfer protein locus affects its activity and high-density lipoprotein size and is a novel marker of cardiovascular disease susceptibility. Circulation 2010;122:470-477.

20. Friedewald WT, Levy RI, Fredrickson DS. Estimation of the concentration of low-density lipoprotein cholesterol in plasma, without use of the preparative ultracentrifuge. Clin Chem 1972;18:499-502.

21. El HK, Arsenault BJ, Franssen R et al. High-density lipoprotein particle size and concentration and coronary risk. Ann Intern Med 2009;150:84-93.

22. Blanchette-Mackie EJ, Scow RO. Effects of lipoprotein lipase on the structure of chylomicrons. J Cell Biol 1973;58:689-708.

23. Pennacchio LA, Olivier M, Hubacek JA et al. An apolipoprotein influencing triglycerides in humans and mice revealed by comparative sequencing. Science 2001;294:169-173.

24. Vaessen SF, Schaap FG, Kuivenhoven JA et al. Apolipoprotein A-V, triglycerides and risk of coronary artery disease: the prospective Epic-Norfolk Population Study. J Lipid Res 2006;47:2064-2070.


49

3

SUPPLEMENTAL MATERIAL

Supplemental Table 1 - SNPs genotyped in EPIC-Norfolk cohort

Gene name Number SNP Amino acid change Nucleotide change

LPL

APOA4

APOC2

LMF1

GPIHBP1

GCKR

GalNT2

APOC3

TRIB1

APOA5

ANGPTL4

NM_000237.2

NM_000482.2

NM_001646.2

NM_022773.2

NM_178172.3

NM_001486.3

NM_004481.3

NM_000040.1

NM_025195.2

NM_052968.4

NM_139314.1

rs268rs11542065rs12678919rs328

rs675

rs5157rs5158

rs760560rs1544800rs2076426rs4984982rs7187700rs6600235

rs11538388rs11538389rs61747644

rs1260326

rs2273970rs3748006

rs2854117

rs2954029

rs75423577rs3135506rs662799

rs116843064rs10444250

p.Asn318Serp.His71Gln

p.Ser474Ter

p.Thr467Ser

p.Val96Valp.Cys14Phep.Leu4Leu

p.Leu446Pro

p.Val554Metp.Asp275Asp

p.Ser19Trp

p.Glu40Lysp.Thr266Met

c.953A>G6

c.213C>G6

g.19844222A>G3,4

c.1421C>G6

c.1099A>T

c.77-845T>Cc.77-828C>T

c.664-5894G>Ac.504-1693C>Tc.730-273C>Gc.515-9519G>Tc.193+8038C>Tc.514+1081G>A

c.138G>T6

c.41G>T6

c.12C>T6

c.1337T>C3

c.1660G>Ac.825T>C

c.-46-482T>C

g.126490972A>T3

c.-73-1083T>C6

c.56C>G6

c.-73-571C>T5

c.118G>A1,2 c.797C>T2

1 Romeo et al Nat Genet 2007; 2 Talmud et al ATVB 2008; 3 Teslovich et al Nature 2010; 4 Johansen et al ATVB 2011; 5 Willer et al Nat Genet 2008; 6 Surendran et al J Int Med 2012.

50

Chapter 3

supplemental Table 2 - Characteristics of cases and controls

Controls(n=2131)

Cases(n=2105)

P

Age, years 64.9 ± 7.9 65.0 ± 7.9 0.657

Men, n (%) 1382 (64.9) 1368 (65.0) 0.949

Body Mass Index, kg/m2 26.3 ± 3.5 27.2 ± 3.9 <0.001

Waist Circumference, cm 91.4 ± 11.4 94.2 ± 11.7 <0.001

Current smoker n, (%) 167 (7.9) 301 (14.5) <0.001

Diabetes Mellitus n, (%) 40 (1.9) 118 (5.6) <0.001

Systolic blood pressure mmHg 138 ± 17 143 ± 19 <0.001

Diastolic blood pressure, mmHg 83 ± 11 86 ± 12 <0.001

Total cholesterol, mmol/L 6.26 ± 1.15 6.48 ± 1.21 <0.001

LDL cholesterol, mmol/L 4.04 ± 1.00 4.23 ± 1.04 <0.001

HDL cholesterol, mmol/L 1.38 ± 0.40 1.28 ± 0.37 <0.001

Triglycerides, mmol/L 1.60 (1.20-2.30) 1.90 (1.40-2.70) <0.001

Use of anti-hypertensive drugs, n (%) 327 (15.3) 776 (36.9) <0.001

Data are presented as mean ± standard deviation, number with the corresponding percentage or as median with the 25th and 75th percentile.


51

3

supp

lem

enta

l Tab

le 3

- Pl

asm

a tr

igly

cerid

e co

ncen

trati

on a

ccor

ding

to p

olym

orph

ism

in tr

igly

cerid

e-re

late

d ge

nes

Gene

SNP

NGT

TG (m

mol

/l)N

GTTG

(mm

ol/l)

NGT

TG (m

mol

/l)%

Var

P†P‡

APO

C2rs

5157

702

CC1.

80 (1

.20-

2.50

)12

45CT

1.80

(1.3

0-2.

60)

555

TT1.

70 (1

.30-

2.40

)0.

000.

958

0.57

9

rs51

5818

72CC

1.70

(1.2

0-2.

50)

586

CT1.

80 (1

.30-

2.50

)44

TT1.

65 (1

.03-

1.98

)0.

000.

796

0.27

8

LMF1

rs71

8770

014

88AA

1.80

(1.3

0-2.

50)

876

AG1.

80 (1

.30-

2.60

)11

1GG

1.80

(1.2

0-2.

60)

0.00

0.21

50.

400

rs49

8498

211

60AA

1.80

(1.3

0-2.

50)

1039

AC1.

70 (1

.20-

2.50

)24

0CC

1.70

(1.2

0-2.

50)

0.00

0.88

20.

813

rs66

0023

515

05CC

1.80

(1.3

0-2.

50)

848

CT1.

75 (1

.20-

2.50

)12

2TT

1.70

(1.4

0-2.

50)

0.00

0.96

80.

967

rs15

4480

012

59AA

1.70

(1.3

0-2.

50)

1019

AG1.

70 (1

.20-

2.50

)19

7GG

1.90

(1.3

0-2.

50)

0.00

0.35

60.

572

rs76

0560

1330

CC1.

80 (1

.30-

2.50

)96

7CT

1.70

(1.2

0-2.

50)

178

TT1.

70 (1

.10-

2.50

)0.

000.

103

0.12

3

rs20

7642

622

79CC

1.80

(1.2

0-2.

50)

158

CG1.

70 (1

.28-

2.40

)2

GG-

0.00

0.61

40.

682

GPI

HBP1

rs61

7476

4416

63CC

1.80

(1.2

0-2.

50)

750

CT1.

70 (1

.30-

2.50

)84

TT1.

70 (1

.20-

2.20

)0.

000.

303

0.25

2

rs11

5383

8876

9GG

1.70

(1.3

0-2.

40)

1262

GT1.

80 (1

.20-

2.60

)50

9TT

1.70

(1.2

0-2.

50)

0.00

0.49

80.

640

rs11

5383

8920

77GG

1.80

(1.2

0-2.

50)

390

GT1.

70 (1

.20-

2.40

)30

TT1.

70 (1

.20-

2.50

)0.

000.

487

0.52

4

APO

A5rs

7542

3577

2081

TT1.

70 (1

.20-

2.50

)27

1TC

2.00

(1.5

0-2.

90)

9CC

2.90

(2.0

0-4.

90)

1.5

<0.0

01<0

.001

rs31

3550

622

57CC

1.70

(1.2

0-2.

50)

269

CG2.

00 (1

.40-

2.90

)16

GG2.

20 (1

.55-

3.40

)0.

8<0

.001

<0.0

01

rs66

2799

2234

AA1.

70 (1

.20-

2.50

)29

3AG

1.90

(1.5

0-2.

90)

15GG

2.50

(1.8

0-4.

00)

0.9

<0.0

01<0

.001

LPL

rs11

5420

6525

64CC

1.80

(1.2

0-2.

50)

-CG

-GG

rs26

824

61AA

1.70

(1.2

0-2.

50)

95AG

1.90

(1.4

0-2.

60)

1GG

-0.

20.

051

0.04

8

rs32

8*20

94CC

1.80

(1.3

0-2.

50)

479

CG1.

60 (1

.10-

2.20

)27

GG1.

50 (1

.10-

2.00

)1.

3<0

.001

<0.0

01

rs12

6789

1920

40AA

1.80

(1.3

0-2.

50)

458

AG1.

60 (1

.10-

2.20

)21

GG1.

50 (1

.10-

2.00

)1.

1<0

.001

<0.0

01

52

Chapter 3

supp

lem

enta

l Tab

le 3

- Co

ntinu

ed

Gene

SNP

NGT

TG (m

mol

/l)N

GTTG

(mm

ol/l)

NGT

TG (m

mol

/l)%

Var

P†P‡

APO

A4rs

675

1496

TT1.

80 (1

.30-

2.50

)71

7TA

1.70

(1.2

0-2.

50)

89AA

1.60

(1.1

5-2.

40)

0.2

0.05

10.

055

ANG

PTL4

rs10

4425

012

76CC

1.80

(1.3

0-2.

60)

1066

CT1.

75 (1

.20-

2.40

)23

2TT

1.70

(1.1

3-2.

40)

0.2

<0.0

5<0

.05

rs11

6843

064

2526

GG1.

80 (1

.30-

2.50

)87

GA1.

40 (1

.10-

2.40

)1

AA-

0.2

<0.0

5<0

.01

GCK

R rs

1260

326

827

TT1.

70 (1

.20-

2.40

)11

52TC

1.80

(1.2

5-2.

50)

375

CC1.

80 (1

.30-

2.60

)0.

10.

072

0.05

7

TRIB

1rs

2954

029

662

AA1.

80 (1

.30-

2.60

)11

26AT

1.80

(1.3

0-2.

40)

442

TT1.

70 (1

.20-

2.40

)0.

6<0

.05

<0.0

5

APO

C3rs

2854

117

1343

TT1.

70 (1

.20-

2.50

)90

3TC

1.80

(1.3

0-2.

50)

150

CC1.

70 (1

.40-

2.63

)0.

10.

258

0.25

7

Gal

NT2

Rs37

4800

619

12TT

1.80

(1.2

0-2.

50)

158

TC1.

95 (1

.40-

2.70

)4

CC2.

05 (1

.15-

3.40

)0.

10.

130

0.22

4

Rs22

7397

019

56GG

1.80

(1.3

0-2.

50)

178

GA1.

90 (1

.30-

2.60

)7

AA1.

50 (1

.20-

3.40

)0.

10.

207

0.53

2

Num

ber

of in

divi

dual

s (N

) an

d tr

igly

cerid

e se

rum

con

cent

ratio

ns a

ccor

ding

to

poly

mor

phism

are

pre

sent

ed a

s m

edia

n an

d 25

th a

nd 7

5th

perc

entil

e. G

T; g

enot

ype.

P†=

unad

just

ed p

-val

ue, P

‡= p

-val

ue a

djus

ted

for a

ge, s

ex, w

aist

, bod

y m

ass i

ndex

and

dia

bete

s*

= SN

Ps se

lect

ed fo

r fina

l gen

e sc

ore

Hassing HC, Boekholdt SM, Breazna A, DeMicco DA, LaRosa JC, Kastelein JJP


4 iNTENsiVE LiPiD-LowERiNG THERaPy iN subjECTs wiTH Low aND HiGH TRiGLyCERiDEs PLus Low aND HiGH HDL-C

An analysis of the Treating to New Targets (TNT) trial

56

Chapter 4

ABSTRACTBackground - Patients with high plasma triglyceride (TG) levels and low high-density lipoprotein cholesterol (HDL-C) levels display a considerable residual cardiovascular risk despite of the level of low-density lipoprotein cholesterol (LDL-C). It is unclear whether intensified statin therapy can reduce cardiovascular morbidity and mortality in these patients. The objective of this post hoc analysis of the Treating to New Targets (TNT) study was to assess whether intensive lowering of low-density lipoprotein cholesterol (LDL-C) results in additional cardiovascular benefits in patients with high TG levels and low HDL-C levels.

Methods - The TNT study was a prospective, double blind trial including 10,001 patients with clinically evident coronary heart disease. The current analysis includes 9,994 patients of whom baseline TG and HDL-C levels where available. Following a run-in phase in which all patients received atorvastatin 10 mg, participants were randomly assigned to receive either atorvastatin 10 mg per day (n=5,004) or 80 mg per day (n=4,990). The primary outcome measure was time to first major cardiovascular event. Treatment effects where analysed based upon baseline TG levels (< or ≥ 150 mg/dL; 1.7 mmol/L) and baseline HDL-C (< or ≥ 40 mg/dL; 1.0 mmol/L) based upon ESC and AHA guidelines.

Results - At a median follow-up of 4.9 years, there was a incremental effect of atorvastatin 80 mg versus 10 mg in reducing the risk of major cardiovascular events in subjects with high TG levels and low HDL-C levels (hazard ratio 0.621 [CI 0.563-0.831]; p=0.0014). The treatment effect differed significantly across TG levels (HR 0.907 [CI 0.766-1.073) for TG <150 mg/dl, HR 0.651 [CI 0.538-0.788] for TG ≥ 150 mg/dL; p for interaction 0.0109). Treatment effect did not differ between HDL-C levels (HR 0.720 [CI 0.574-0.902) for HDL-C <40 mg/dl, HR 0.820 [CI 0.704-0.955] for HDL-C ≥ 40 mg/dL; p for interaction 0.35).

Conclusion - These results show that increasing atorvastatin dose from 10 mg to 80 mg results in significant incremental cardiovascular benefit. Since other therapeutic interventions including ezetimibe, fibrates, nicotinic acid and CETP inhibition have failed in reducing cardiovascular risk, intensive statin therapy is currently the only pharmacotherapeutic strategy to reduce cardiovascular risk in patients with metabolic dyslipidemia in clinical practice.

IntensIve lIpId-lowerIng therapy In metabolIc dyslIpIdemIa

57

4

INTRODUCTIONCardiovascular disease (CVD) remains a major cause of morbidity and mortality worldwide. As a result of the increasing rates of obesity and obesity-related diseases such as the metabolic syndrome and diabetes mellitus, the burden of CVD is expected to rise even further.1-3 Low-density lipoprotein cholesterol (LDL-C) is currently the primary target for lipid-lowering therapy to reduce cardiovascular risk. However, in patients with metabolic dyslipidemia characterized by elevated plasma triglyceride (TG) levels and low plasma levels of high-density lipoprotein cholesterol (HDL-C), a considerable residual risk remains, even at optimal LDL-C levels.4-6 The atherosclerotic phenotype of metabolic dyslipidemia results from the presence of triglyceride-rich lipoprotein remnant particles and small dense LDL-C particles contributing to intimal cholesterol deposition and oxidative stress.

Patients with the “lipid triad” i.e. elevated LDL-C levels, elevated TG levels and reduced HDL-C levels endure more benefit from simvastatin therapy than patients with isolated LDL-C elevation.7 In addition, patients with coronary heart disease (CHD) and diabetes mellitus or the metabolic syndrome have incremental benefit from high-dose compared to usual-dose atorvastatin therapy.8, 9

However, it remains unclear what the optimal treatment strategy is for patients who are treated with statin therapy and still have a metabolic lipid profile, and as a consequence current recommendations and treatment goals for these patients are not consistent.10, 11 The ACCORD trial showed that in patients with diabetes mellitus, fenofibrate versus placebo on top of simvastatin did not reduce the risk of cardiovascular events, but prespecified subgroup analyses suggested that those with hypertriglyceridemia and low HDL-C levels did benefit from add-on fenofibrate.12, 13 Whether increasing the statin dose is beneficial in these patients is unknown, largely because statin trials are usually designed to include participants based on off-statin lipid levels.

The Treating to New Targets (TNT) trial enrolled 10,001 patients with stable CHD who had LDL-C levels <130 mg/dL (3.4 mmol/L) after an open-label run-in period on atorvastatin 10 mg.14, 15 Patients were subsequently randomized to atorvastatin 10 mg or 80 mg per day for a median follow-up of about 5 years. The current post hoc analysis of the TNT trial investigates whether patients on usual-dose atorvastatin who still had hypertriglyceridemia (≥150 mg/dL; 1.7 mmol/L) or low HDL-C levels (<40 mg/dL; 1.0 mmol/L) gain cardiovascular benefits from switching to high-dose atorvastatin. Cut-off values were depicted based upon ESC/ AHA guidelines since small dense LDL-C particles, and thereby the atherogenic phenotype, are know to arise from these levels. 11, 16

METHODSThe TNT trial was a randomized, double-blind, placebo-controlled trial, which has been described in detail previously. 14, 15 Briefly, patients eligible for inclusion were men and women aged 35–75 years with clinically evident CHD, defined as previous myocardial infarction, previous or present angina

58

Chapter 4

with objective evidence of atherosclerotic CHD, or previous coronary revascularisation. Major exclusion criteria were statin hypersensitivity, current liver disease, nephrotic syndrome, pregnancy, uncontrolled cardiovascular risk factors, a CHD event or revascularisation within a month, congestive heart failure, unexplained creatine phosphokinase concentrations six or more times the upper limit of normal, life-threatening malignancy, or immunosuppressive or lipid-lowering drug treatment. Potentially eligible patients entered an 8-week open-label run-in period with atorvastatin 10 mg daily. At the end of the run-in phase, patients with LDL-C <130 mg/dl (3.4 mmol/l) were randomized to double-blind therapy with either atorvastatin 10 or 80 mg daily. The primary outcome measure was the time to first occurrence of a major cardiovascular event, defined as death from CHD, non-fatal non-procedure-related myocardial infarction, resuscitated cardiac arrest, or fatal or non-fatal stroke. Major coronary events were defined as CHD death, non-fatal non-procedure-related myocardial infarction, or resuscitated cardiac arrest. Major cerebrovascular events were defined as fatal or nonfatal stroke or transient ischemic attack. For the current post hoc analysis, we excluded patients with missing data on baseline triglycerides or high-density lipoprotein cholesterol.

statistical analysisPatients were divided into 4 groups based on low or high baseline plasma TG levels (TG < or ≥ 150 mg/dL; 1.7 mmol/L) and low or high baseline HDL-C (< or ≥ 40 mg/dL; 1.0 mmol/L). Differences in baseline characteristics between these 4 strata were calculated by one-way analysis of variance for continuous variables and chi-square test for categorical variables. Pair-wise comparisons for continuous variables were based on one-way analysis variance adjusting by Bonferroni test and logistic regression analysis based on reference design matrix coding for categorical variables. Within each group, Cox proportional hazards models were used to calculate unadjusted hazard ratios and corresponding 95% confidence intervals and c2 statistics between the atorvastatin 80 mg and 10 mg treatment groups. A likelihood ratio test was used to test for interaction between treatment effect and the subgroups of subjects cross-classified by baseline TG and/or HDL-C by Cox proportional hazard model. Analyses were performed by 4 groups (as defined above based on high and low levels of TG and HDL-C) and in addition by 2 groups (TG < versus ³ 150 mg/dL, and HDL-C < versus ³ 40 mg/dL). Two-sided P values <0.05 were regarded as significant.

RESULTSOf the 10,001 patients randomized, 7 were excluded because of missing data for baseline TG or HDL-C. Thus, a total of 9,994 (99.9%) patients were included in the current analysis. Of these, 5,004 patients were assigned to receive atorvastatin 10 mg per day and 4,990 patients to atorvastatin 80 mg per day. Division into 4 groups based on baseline TG and HDL levels yielded 4,984 subjects (49.9%) in group 1 (TG < 150 mg/dL and HDL-C ≥ 40 mg/dL), 984 subjects (9.8%) in group 2 (TG < 150 mg/dL and HDL-C < 40 mg/dL), 2,470 subjects (24.7%) in group 3 (TG ≥ 150 mg/dL and HDL-C ≥ 40 mg/dL) and 1,556 subjects (15.6%) in group 4 (TG ≥ 150 mg/dL and HDL-C < 40 mg/dL). All relevant baseline characteristics differed significantly between the 4 strata (Table 1). Patients in group 4


59

4

Tabl

e 1

- bas

elin

e Ch

arac

teris

tics b

y Le

vels

of T

rigly

cerid

es a

nd H

DL-C

Grou

p 1:

Grou

p 2:

Grou

p 3:

Grou

p 4:

Ove

rall

P-Va

lue

Base

line

Char

acte

ristic

sTG

< 1

50 m

g/dL

HDL-

C >=

40

mg/

dLTG

< 1

50 m

g/dL

HDL-

C <

40 m

g/dL

TG >

= 15

0 m

g/dL

HDL-

C >=

40

mg/

dLTG

>=

150

mg/

dLHD

L-C

< 40

mg/

dL

N49

8498

424

7015

56

Men

40

25 (8

0.76

%)

937

(95.

22%

)17

55 (7

1.05

%)

1376

(88.

43%

)<.

0001

*

Age,

yea

rs62

.10

± 8.

5859

.93

± 9.

0360

.58

± 8.

7558

.95

± 9.

12<.

0001

*

Curr

ent s

mok

er54

5 (1

0.93

%)

183

(18.

60%

)29

3 (1

1.86

%)

318

(20.

44%

)<.

0001

*

Body

mas

s ind

ex, k

g/m

227

.57

± 4.

2428

.78

± 4.

7029

.25

± 4.

5630

.31

± 4.

72<.

0001

*

Diab

etes

mel

litus

556

(11.

16%

)16

0 (1

6.26

%)

439

(17.

77%

)34

5 (2

2.17

%)

<.00

01*

Fasti

ng g

luco

se, M

G/dL

103.

58 ±

25.

7610

7.44

± 2

7.55

110.

38 ±

33.

0211

7.01

± 3

9.10

<.00

01*

Syst

olic

blo

od p

ress

ure,

mm

Hg13

0.8

± 17

.012

8.4

± 16

.413

1.9

± 16

.113

0.0

± 17

.1<.

0001

*

Dias

tolic

blo

od p

ress

ure,

mm

Hg77

.8 ±

9.5

76.8

± 9

.678

.6 ±

9.2

78.4

± 9

.6<.

0001

*

Tota

l cho

lest

erol

, mg/

dL17

0.53

± 2

0.98

152.

49 ±

17.

8419

0.21

± 2

2.17

177.

74 ±

22.

65<.

0001

*

LDL

chol

este

rol,

mg/

dL96

.96

± 17

.23

93.4

0 ±

16.5

010

0.41

± 1

8.05

97.1

9 ±

17.9

2<.

0001

*

HDL

chol

este

rol,

mg/

dL52

.67

± 10

.25

36.0

6 ±

2.99

48.6

3 ±

7.77

35.1

2 ±

3.26

<.00

01*

Trig

lyce

rides

, mg/

dL10

4.49

± 2

5.08

115.

21 ±

22.

4620

6.86

± 5

5.21

231.

37 ±

77.

05<.

0001

*

Non

-HDL

cho

lest

erol

, mg/

dL11

7.86

± 1

8.51

116.

44 ±

17.

5514

1.58

± 2

1.10

142.

62 ±

22.

31<.

0001

*

Apol

ipop

rote

in A

-I, m

g/dL

152.

94 ±

22.

5511

7.84

± 1

2.27

156.

82 ±

22.

0412

5.22

± 1

3.26

<.00

01*

Apol

ipop

rote

in B

, mg/

dL10

4.40

± 1

6.58

104.

50 ±

16.

1312

0.70

± 1

8.72

121.

56 ±

18.

78<.

0001

*

Data

are

pre

sent

ed a

s mea

n ±

stan

dard

dev

iatio

n or

num

ber (

perc

enta

ge).

Abbr

evia

tion:

TG

= Tr

igly

cerid

es, H

DL-C

= h

igh-

dens

ity li

popr

otei

n ch

oles

tero

l, LD

L-C

= lo

w-d

ensit

y lip

opro

tein

cho

lest

erol

.

60

Chapter 4

were younger, more overweight, and more likely to have diabetes mellitus, than those in the other strata. In addition, they were more likely to smoke. Pair-wise comparisons of baseline characteristics between groups are shown in supplementary Table 1.

Efficacy of high-dose versus usual dose atorvastatin by triglycerides and HDL-C levelsAfter a median follow-up of 4.9 years, a total of 980 primary events occurred. Among those in group 1 (low TG, high HDL-C), a major cardiovascular event occurred in 202 patients (8.14%) of those receiving atorvastatin 80 mg and 224 patients (8.95%) of those receiving atorvastatin 10 mg, yielding a hazard ratio of 0.907 (95% CI 0.750-1.096; p=0.312, table2). In group 2 (low TG, low HDL-C), a major cardiovascular event occurred in 56 patients (11.48%) receiving atorvastatin 80 mg and 62 patients (12.50%) receiving atorvastatin 10 mg, yielding a hazard ratio of 0.908 (95% CI 0.633-1.304; p=0.603). For those in group 3 (high TG, high HDL-C), the number of events in the 80 mg and 10 mg arms were 103 (8.14%) and 141 (11.70%) respectively (hazard ratio 0.682, 95% CI 0.529-0.879; p=0.0031). For those in group 4 (high TG, low HDL-C), the number of events were 72 (9.54%) and 120 (14.98%) for the atorvastatin 80 mg and 10 mg arms respectively (hazard ratio 0.621, 95% CI 0.463-0.831; p=0.0014).

For major coronary events, results were similar to those for major cardiovascular events. For cerebrovascular events, the absolute event rates were substantially lower than for major CV and coronary events, leading to a lower statistical power to detect differences. In all subgroups, treatment effects were directionally similar to those observed for major CV events, but interaction terms between groups were non-significant.

In the two groups with low TG combined (groups 1 and 2) there was no significant incremental effect of atorvastatin 80 mg versus 10 mg in reducing the risk of major CV events (hazard ratio 0.907; 95% CI 0.766-1.073; p=0.2537, supplementary Table 2). In the two groups with high TG levels combined (groups 3 and 4), atorvastatin 80 mg versus 10 mg resulted in a significant additional reduction of major CV events (hazard ratio 0.651, 95% CI 0.538-0.788; p<0.0001). The treatment effect differed significantly between the high TG and low TG groups (p for interaction 0.0109).

Conversely, in the two groups with high HDL-C combined (groups 1 and 3), the hazard ratio for major CV events was 0.820 (95% CI 0.704-0.955; p=0.0105, supplementary Table 2). In the combined groups with low HDL-C (groups 2 and 4) the hazard ratio was 0.720 (95% CI 0.574-0.902; p=0.0043). The treatment effect of atorvastatin 80 mg versus 10 mg did not differ significantly between the high and low HDL-C groups (p for interaction 0.3488).

Efficacy of high-dose versus usual dose atorvastatin by TG and HDL-C levels, diabetes mellitus and baseline LDL-C levelsSubgroup analyses among patients with diabetes mellitus versus patients without diabetes mellitus showed similar results for treatment effects on in groups with high TG and low HDL levels to those


61

4

Tabl

e 2

- Tre

atm

ent E

ffect

by

subg

roup

s bas

ed o

n TG

and

HDL

-C L

evel

s

Base

line

Lipi

d Pr

ofile

TG <

150

mg/

dLTG

< 1

50 m

g/dL

TG >

= 15

0 m

g/dL

TG >

= 15

0 m

g/dL

Inte

racti

onIn

tera

ction

HDL-

C >=

40

mg/

dLHD

L-C

< 40

mg/

dLHD

L-C

>= 4

0 m

g/dL

HDL-

C <

40 m

g/dL

Trea

tmen

t by

Trea

tmen

t by

(N =

498

4)(N

= 9

84)

(N =

247

0)(N

= 1

556)

TG <

or ³

15

0 m

g/dL

HD

L-C

< or

³ 40

mg/

dL

P-

valu

eP-

valu

e

Maj

or c

ardi

ovas

cula

r eve

nts

Even

t Rat

e: n

/ N

(%)

- Ato

rvas

tatin

80

mg

- Ato

rvas

tatin

10

mg

202

/ 248

2 (8

.14%

)22

4 / 2

502

(8.9

5%)

56 /

488

(11.

48%

)62

/ 49

6 (1

2.50

%)

103

/ 126

5 (8

.14%

)14

1 / 1

205

(11.

70%

)72

/ 75

5 (9

.54%

)12

0 / 8

01 (1

4.98

%)

Haza

rd ra

tio:

ator

vast

atin

80m

g v

10m

g0.

907

0.90

80.

682

0.62

1P

= 0.

0109

*P

= 0.

3488

95%

con

fiden

ce in

terv

al(0

.750

, 1.0

96)

(0.6

33, 1

.304

)(0

.529

, 0.8

79)

(0.4

63, 0

.831

)

c2 stati

stics

P-va

lue

c2 = 1

.023

P =

0.31

19c2 =

0.2

71P

= 0.

6026

c2 = 8

.743

P =

0.00

31*

c2 = 1

0.24

8P

= 0.

0014

*

Maj

or c

oron

ary

even

tsEv

ent R

ate:

n /

N (%

)- A

torv

asta

tin 8

0 m

g- A

torv

asta

tin 1

0 m

g

152

/ 248

2 (6

.12%

)16

3 / 2

502

(6.5

1%)

45 /

488

(9.2

2%)

48 /

496

(9.6

8%)

76

/ 126

5 (6

.01%

)10

7 / 1

205

(8.8

8%)

61 /

755

(8.0

8%)

99 /

801

(12.

36%

)

Haza

rd ra

tio:

ator

vast

atin

80m

g v

10m

g0.

939

0.94

60.

666

0.64

0P

= 0.

0125

*P

= 0.

4639

95%

con

fiden

ce in

terv

al(0

.753

, 1.1

71)

(0.6

30, 1

.420

)(0

.496

, 0.8

93)

(0.4

65, 0

.881

)

c2 stati

stics

P-va

lue

c2 = 0

.314

P =

0.57

52c2 =

0.0

73P

= 0.

7875

c2 = 7

.357

P =

0.00

67*

c2 = 7

.497

P =

0.00

62*

62

Chapter 4

Tabl

e 2

- Con

tinue

d

Base

line

Lipi

d Pr

ofile

TG <

150

mg/

dLTG

< 1

50 m

g/dL

TG >

= 15

0 m

g/dL

TG >

= 15

0 m

g/dL

Inte

racti

onIn

tera

ction

HDL-

C >=

40

mg/

dLHD

L-C

< 40

mg/

dLHD

L-C

>= 4

0 m

g/dL

HDL-

C <

40 m

g/dL

Trea

tmen

t by

Trea

tmen

t by

(N =

498

4)(N

= 9

84)

(N =

247

0)(N

= 1

556)

TG <

or ³

15

0 m

g/dL

HD

L-C

< or

³ 40

mg/

dL

P-

valu

eP-

valu

e

Cere

brov

ascu

lar e

vent

sEv

ent R

ate:

n /

N (%

)- A

torv

asta

tin 8

0 m

g- A

torv

asta

tin 1

0 m

g

89

/ 248

2 (3

.59%

)10

6 / 2

502

(4.2

4%)

19 /

488

(3.8

9%)

31 /

496

(6.2

5%)

56

/ 126

5 (4

.43%

)63

/ 12

05 (5

.23%

)31

/ 75

5 (4

.11%

)52

/ 80

1 (6

.49%

)

Haza

rd ra

tio:

ator

vast

atin

80m

g v

10m

g0.

845

0.61

80.

842

0.62

3P

= 0.

7387

P =

0.14

36

95%

con

fiden

ce in

terv

al(0

.637

, 1.1

20)

(0.3

49, 1

.094

)(0

.587

, 1.2

07)

(0.4

00, 0

.972

)

c2 stati

stics

P-va

lue

c2 = 1

.373

P =

0.24

13c2 =

2.7

33P

= 0.

0983

c2 = 0

.876

P =

0.34

92

c2 = 4

.339

P =

0.03

72*

Abbr

evia

tion:

TG

= Tr

igly

cerid

es, H

DL-C

= h

igh-

dens

ity li

popr

otei

n ch

oles

tero

l.


63

4

found for the overall population (supplemental Table 3). There was a significant additional benefit for atorvastatin 80 mg versus 10 mg in reducing the risk of major CV events in group 4 among patients with diabetes mellitus (hazard ratio 0.581, 95% CI 0.352-0.959; p=0.0337) as well as for patient without diabetes mellitus (hazard ratio 0.638; 95% CI 0.445-0.914; p=0.0143). Among patients with diabetes mellitus, the treatment effect of atorvastatin 80 mg versus 10 mg in reducing the risk of major CV event did neither differ significantly between the high and low TG groups (p for interaction 0.2052) nor between the high and low HDL-C groups (p for interaction 0.0597).

Among patients with baseline LDL-C < 100 mg/dL, those with high TG and low HDL-C had a substantial additional benefit from atorvastatin 80 mg versus 10 mg in reducing the risk of major CV event (hazard ratio 0.562, 95%CI 0.375-0.844; p=0.0055, supplemental Table 4). In the subgroup with high TG and high HDL-C the treatment effect was less pronounced (hazard ratio 0.943, 95%CI 0.636-1.399; p=0.7722).

Among those with LDL-C > 100 mg/dL at baseline, there was a large treatment effect in both groups with high TG, although not significantly in the group with high TG and low HDL-C due to low number of CV events in this subgroup.

DISCUSSIONIn the current analysis of the TNT trial, we show that CHD patients, who have high TG levels and low HDL-C levels on usual-dose statin therapy, benefit from switching to high-dose statin therapy as evidenced by a lower rate of major CV events during follow-up.

Interestingly, almost 16% of the total cohort had elevated TG levels as well as low HDL-C levels, an indication that a considerable part of these patients can be classified as having an elevated cardiovascular risk beyond the risk already conferred by the fact these patients suffer from coronary heart disease. In TNT, atorvastatin 80 mg versus atorvastatin 10 mg relatively reduced the risk of major cardiovascular event by 36% in patients with high TG and low HDL-C levels and 9% in patients with low TG and high HDL-C levels; a significant incremental benefit of 25%.

Patients with metabolic dyslipidemia are characterized by a residual cardiovascular risk, even at low LDL-C levels. Plasma TG levels represent both hepatic-derived very low-density lipoproteins (VLDL) predominately present in the fasting state as well as the postprandial intestine-derived chylomicrons and their remnant particles.17 Recent evidence indicates that these remodeled triglyceride-rich lipoprotein remnants have atherogenic properties because of their small particle size and enrichment in cholesterol which supports a causal association between triglyceride-rich lipoproteins and CVD risk.17-19 It is currently unknown whether higher statin doses are able to correct the residual CV risk in patients with elevated TG and low HDL-C levels. According to current guidelines, the LDL-C target for patient with a very high cardiovascular risk –including established CAD- are < 70 mg/dL (1.8

64

Chapter 4

mmol/L).20 However, LDL-C targets for patients with high TG and/or low HDL-C levels are insufficient to establish. A recent meta-analysis in 26 randomized trials revealed that any reduction in LDL-C is accompanied by a reduction in cardiovascular without any threshold within the studied LDL-C range.21 The CHD risk reduction per 1 mmol/L LDL-C reduction was independent of baseline HDL-C and TG levels 22. In the current analysis, cardiovascular event rate for atorvastatin 80 mg was 9.5% in subjects with high TG levels and low HDL-C levels, which was significantly lower than the 15.0% event rate in atorvastatin 10 mg. The study provides evidence to suggest that patients with CHD, high TG levels and low HDL-C levels might be good candidates for more intensive lipid-lowering therapy. In our study, patients with CHD, high TG and low HDL-C levels were at higher risk than those with low TG levels and/or high HDL-C levels because of the presence of multiple risk factors for CVD including a higher body-mass index, higher blood pressure and higher fasting glucose levels. Therefore also the absolute benefit will be greater in these subjects because of their higher absolute risk.

Some limitations warrant comment when interpreting the results of this analysis. First and most importantly, this was a non-prespecified post-hoc analysis with all its inherent limitations, and as such the results should be interpreted with caution. However, it should be noted that the results are in line with several previous analyses on the same topic. Second, the TNT trial enrolled patients with clinical evident coronary heart disease, and therefore, the additional benefit of high-dose statin in patients with high TG and low HDL levels observed in this subanalysis cannot be generalized to people without established coronary heart disease. Finally, despite the fact that TNT was a very large trial, the number of CV events in analyses by TG and HDL-C levels were small in the subgroups according to diabetes status and baseline LDL-C level. The results in these small subgroups should therefore be interpreted with caution.

In the past few years, efficacy of various drugs for treatment of hypertriglyceridemia in combination with statin therapy was evaluated in large clinical trials. As other therapeutic interventions including ezetimibe, fibrates, nicotinic acid, and CETP inhibition have failed in reducing cardiovascular risk on top of statin therapy 12, 23-25 intensive statin therapy is currently the only pharmacotherapeutic strategy to reduce cardiovascular risk in patients with metabolic dyslipidemia in clinical practice.

In summary, our results provide evidence that among CHD patients who have high TG levels and low HDL-C levels on usual-dose statin therapy, switching to high-dose statin therapy results in a lower risk of CVD events during follow-up. Therefore, statin therapy may be more potent to correct part of the residual CV risk in these patients. Whether other therapies beyond statin can also address this, is the objective of intense clinical research.


65

4

REFERENCE LIST1. Shaw JE, Sicree RA, Zimmet PZ. Global estimates of the prevalence of diabetes for 2010 and 2030.

Diabetes Res Clin Pract 2010;87:4-14.

2. Haffner SM, Lehto S, Ronnemaa T, Pyorala K, Laakso M. Mortality from coronary heart disease in subjects with type 2 diabetes and in nondiabetic subjects with and without prior myocardial infarc-tion. N Engl J Med 1998;339:229-234.

3. Buse JB, Ginsberg HN, Bakris GL et al. Primary prevention of cardiovascular diseases in people with diabetes mellitus: a scientific statement from the American Heart Association and the American Diabetes Association. Diabetes Care 2007;30:162-172.

4. Olsson AG, Schwartz GG, Szarek M et al. High-density lipoprotein, but not low-density lipoprotein cholesterol levels influence short-term prognosis after acute coronary syndrome: results from the MIRACL trial. Eur Heart J 2005;26:890-896.

5. Barter P, Gotto AM, LaRosa JC et al. HDL cholesterol, very low levels of LDL cholesterol, and cardio-vascular events. N Engl J Med 2007;357:1301-1310.

6. Miller M, Cannon CP, Murphy SA, Qin J, Ray KK, Braunwald E. Impact of triglyceride levels beyond low-density lipoprotein cholesterol after acute coronary syndrome in the PROVE IT-TIMI 22 trial. J Am Coll Cardiol 2008;51:724-730.

7. Ballantyne CM, Olsson AG, Cook TJ, Mercuri MF, Pedersen TR, Kjekshus J. Influence of low high-den-sity lipoprotein cholesterol and elevated triglyceride on coronary heart disease events and response to simvastatin therapy in 4S. Circulation 2001;104:3046-3051.

8. Shepherd J, Barter P, Carmena R et al. Effect of lowering LDL cholesterol substantially below currently recommended levels in patients with coronary heart disease and diabetes: the Treating to New Tar-gets (TNT) study. Diabetes Care 2006;29:1220-1226.

9. Deedwania P, Barter P, Carmena R et al. Reduction of low-density lipoprotein cholesterol in patients with coronary heart disease and metabolic syndrome: analysis of the Treating to New Targets study. Lancet 2006;368:919-928.

10. Reiner Z, Catapano AL, De BG et al. ESC/EAS Guidelines for the management of dyslipidaemias: the Task Force for the management of dyslipidaemias of the European Society of Cardiology (ESC) and the European Atherosclerosis Society (EAS). Eur Heart J 2011;32:1769-1818.

11. Chapman MJ, Ginsberg HN, Amarenco P et al. Triglyceride-rich lipoproteins and high-density lipopro-tein cholesterol in patients at high risk of cardiovascular disease: evidence and guidance for manage-ment. Eur Heart J 2011;32:1345-1361.


13. Sacks FM, Carey VJ, Fruchart JC. Combination lipid therapy in type 2 diabetes. N Engl J Med 2010;363:692-694.

14. Waters DD, Guyton JR, Herrington DM, McGowan MP, Wenger NK, Shear C. Treating to New Targets (TNT) Study: does lowering low-density lipoprotein cholesterol levels below currently recommended guidelines yield incremental clinical benefit? Am J Cardiol 2004;93:154-158.

15. LaRosa JC, Grundy SM, Waters DD et al. Intensive lipid lowering with atorvastatin in patients with stable coronary disease. N Engl J Med 2005;352:1425-1435.

66

Chapter 4

16. Miller M, Stone NJ, Ballantyne C et al. Triglycerides and cardiovascular disease: a scientific statement from the American Heart Association. Circulation 2011;123:2292-2333.


18. Ginsberg HN. New perspectives on atherogenesis: role of abnormal triglyceride-rich lipoprotein me-tabolism. Circulation 2002;106:2137-2142.

19. Twickler TB, Dallinga-Thie GM, Cohn JS, Chapman MJ. Elevated remnant-like particle cholester-ol concentration: a characteristic feature of the atherogenic lipoprotein phenotype. Circulation 2004;109:1918-1925.

20. Grundy SM, Cleeman JI, Merz CN et al. Implications of recent clinical trials for the National Choles-terol Education Program Adult Treatment Panel III guidelines. Circulation 2004;110:227-239.

21. Baigent C, Blackwell L, Emberson J et al. Efficacy and safety of more intensive lowering of LDL cholesterol: a meta-analysis of data from 170,000 participants in 26 randomised trials. Lancet 2010;376:1670-1681.

22. Baigent C, Keech A, Kearney PM et al. Efficacy and safety of cholesterol-lowering treatment: pro-spective meta-analysis of data from 90,056 participants in 14 randomised trials of statins. Lancet 2005;366:1267-1278.

23. Kastelein JJ, Akdim F, Stroes ES et al. N Engl J Med.

24. Keech A, Simes RJ, Barter P et al. Effects of long-term fenofibrate therapy on cardiovascular events in 9795 people with type 2 diabetes mellitus (the FIELD study): randomised controlled trial. Lancet 2005;366:1849-1861.

25. Barter PJ, Caulfield M, Eriksson M et al. Effects of torcetrapib in patients at high risk for coronary events. N Engl J Med 2007;357:2109-2122.


67

4

68

Chapter 4


supplementary Table 1 - baseline Characteristics by subjects with Low and High Triglycerides and HDL-C levels

Group 1: Group 2: Group 3: Group 4: Pair-Wise Comparison:

P-Value

Baseline Characteristics

TG < 150 mg/dLHDL >= 40 mg/dL

TG < 150 mg/dLHDL < 40 mg/dL

TG >= 150 mg/dLHDL >= 40 mg/dL

TG >= 150 mg/dLHDL < 40 mg/dL Group 1 vs. 2 Group 1 vs. 3 Group 1 vs. 4 Group 2 vs. 3 Group 2 vs. 4 Group 3 vs. 4

N 4984 984 2470 1556

Men 4025 (80.76%) 937 (95.22%) 1755 (71.05%) 1376 (88.43%) <.0001* <.0001* <.0001* <.0001* <.0001* <.0001*

Age, years 62.10 ± 8.58 59.93 ± 9.03 60.58 ± 8.75 58.95 ± 9.12 <.0001* <.0001* <.0001* 0.0496* 0.0060* <.0001*

Current smoker 545 (10.93%) 183 (18.60%) 293 (11.86%) 318 (20.44%) <.0001* 0.2330 <.0001* <.0001* 0.2565 <.0001*

Body mass index, kg/m2 27.57 ± 4.24 28.78 ± 4.70 29.25 ± 4.56 30.31 ± 4.72 <.0001* <.0001* <.0001* 0.0053* <.0001* <.0001*

Diabetes mellitus 556 (11.16%) 160 (16.26%) 439 (17.77%) 345 (22.17%) <.0001* <.0001* <.0001* 0.2892 0.0003* 0.0006*

Fasting glucose, mg/dL 103.58 ± 25.76 107.44 ± 27.55 110.38 ± 33.02 117.01 ± 39.10 0.0003* <.0001* <.0001* 0.0097* <.0001* <.0001*

Systolic blood pressure, mmHg 130.8 ± 17.0 128.4 ± 16.4 131.9 ± 16.1 130.0 ± 17.1 <.0001* 0.0134* 0.0699 <.0001* 0.0220* 0.0005*

Diastolic blood pressure, mmHg 77.8 ± 9.5 76.8 ± 9.6 78.6 ± 9.2 78.4 ± 9.6 0.0023* 0.0004* 0.0170* <.0001* <.0001* 0.5936

Total cholesterol, mg/dL 170.53 ± 20.98 152.49 ± 17.84 190.21 ± 22.17 177.74 ± 22.65 <.0001* <.0001* <.0001* <.0001* <.0001* <.0001*

LDL cholesterol, mg/dL 96.96 ± 17.23 93.40 ± 16.50 100.41 ± 18.05 97.19 ± 17.92 <.0001* <.0001* 0.6593 <.0001* <.0001* <.0001*

HDL cholesterol, mg/dL 52.67 ± 10.25 36.06 ± 2.99 48.63 ± 7.77 35.12 ± 3.26 <.0001* <.0001* <.0001* <.0001* 0.0061* <.0001*

Triglycerides, mg/dL 104.49 ± 25.08 115.21 ± 22.46 206.86 ± 55.21 231.37 ± 77.05 <.0001* <.0001* <.0001* <.0001* <.0001* <.0001*

Non-HDL cholesterol, mg/dL 117.86 ± 18.51 116.44 ± 17.55 141.58 ± 21.10 142.62 ± 22.31 0.0383* <.0001* <.0001* <.0001* <.0001* 0.1049

Apolipoprotein A-I, mg/dL 152.94 ± 22.55 117.84 ± 12.27 156.82 ± 22.04 125.22 ± 13.26 <.0001* <.0001* <.0001* <.0001* <.0001* <.0001*

Apolipoprotein B, mg/dL 104.40 ± 16.58 104.50 ± 16.13 120.70 ± 18.72 121.56 ± 18.78 0.8692 <.0001* <.0001* <.0001* <.0001* 0.1310

Data are presented as mean ± standard deviation or number (percentage). Abbreviation: TG = Triglycerides, HDL = high-density lipoprotein, LDL = low-density lipoprotein.


69

4


supplementary Table 1 - baseline Characteristics by subjects with Low and High Triglycerides and HDL-C levels

Group 1: Group 2: Group 3: Group 4: Pair-Wise Comparison:

P-Value

Baseline Characteristics

TG < 150 mg/dLHDL >= 40 mg/dL

TG < 150 mg/dLHDL < 40 mg/dL

TG >= 150 mg/dLHDL >= 40 mg/dL

TG >= 150 mg/dLHDL < 40 mg/dL Group 1 vs. 2 Group 1 vs. 3 Group 1 vs. 4 Group 2 vs. 3 Group 2 vs. 4 Group 3 vs. 4

N 4984 984 2470 1556

Men 4025 (80.76%) 937 (95.22%) 1755 (71.05%) 1376 (88.43%) <.0001* <.0001* <.0001* <.0001* <.0001* <.0001*

Age, years 62.10 ± 8.58 59.93 ± 9.03 60.58 ± 8.75 58.95 ± 9.12 <.0001* <.0001* <.0001* 0.0496* 0.0060* <.0001*

Current smoker 545 (10.93%) 183 (18.60%) 293 (11.86%) 318 (20.44%) <.0001* 0.2330 <.0001* <.0001* 0.2565 <.0001*

Body mass index, kg/m2 27.57 ± 4.24 28.78 ± 4.70 29.25 ± 4.56 30.31 ± 4.72 <.0001* <.0001* <.0001* 0.0053* <.0001* <.0001*

Diabetes mellitus 556 (11.16%) 160 (16.26%) 439 (17.77%) 345 (22.17%) <.0001* <.0001* <.0001* 0.2892 0.0003* 0.0006*

Fasting glucose, mg/dL 103.58 ± 25.76 107.44 ± 27.55 110.38 ± 33.02 117.01 ± 39.10 0.0003* <.0001* <.0001* 0.0097* <.0001* <.0001*

Systolic blood pressure, mmHg 130.8 ± 17.0 128.4 ± 16.4 131.9 ± 16.1 130.0 ± 17.1 <.0001* 0.0134* 0.0699 <.0001* 0.0220* 0.0005*

Diastolic blood pressure, mmHg 77.8 ± 9.5 76.8 ± 9.6 78.6 ± 9.2 78.4 ± 9.6 0.0023* 0.0004* 0.0170* <.0001* <.0001* 0.5936

Total cholesterol, mg/dL 170.53 ± 20.98 152.49 ± 17.84 190.21 ± 22.17 177.74 ± 22.65 <.0001* <.0001* <.0001* <.0001* <.0001* <.0001*

LDL cholesterol, mg/dL 96.96 ± 17.23 93.40 ± 16.50 100.41 ± 18.05 97.19 ± 17.92 <.0001* <.0001* 0.6593 <.0001* <.0001* <.0001*

HDL cholesterol, mg/dL 52.67 ± 10.25 36.06 ± 2.99 48.63 ± 7.77 35.12 ± 3.26 <.0001* <.0001* <.0001* <.0001* 0.0061* <.0001*

Triglycerides, mg/dL 104.49 ± 25.08 115.21 ± 22.46 206.86 ± 55.21 231.37 ± 77.05 <.0001* <.0001* <.0001* <.0001* <.0001* <.0001*

Non-HDL cholesterol, mg/dL 117.86 ± 18.51 116.44 ± 17.55 141.58 ± 21.10 142.62 ± 22.31 0.0383* <.0001* <.0001* <.0001* <.0001* 0.1049

Apolipoprotein A-I, mg/dL 152.94 ± 22.55 117.84 ± 12.27 156.82 ± 22.04 125.22 ± 13.26 <.0001* <.0001* <.0001* <.0001* <.0001* <.0001*

Apolipoprotein B, mg/dL 104.40 ± 16.58 104.50 ± 16.13 120.70 ± 18.72 121.56 ± 18.78 0.8692 <.0001* <.0001* <.0001* <.0001* 0.1310

Data are presented as mean ± standard deviation or number (percentage). Abbreviation: TG = Triglycerides, HDL = high-density lipoprotein, LDL = low-density lipoprotein.

70

Chapter 4

supp

lem

enta

ry T

able

2 -

Trea

tmen

t Effe

ct b

y su

bjec

ts w

ith L

ow a

nd H

igh

Trig

lyce

rides

and

by

subj

ects

with

Low

and

Hig

h HD

L-C

Inte

racti

onIn

tera

ction

Base

line

Lipi

d Pr

ofile

TG <

150

mg/

dLTG

>=

150

mg/

dLHD

L-C

< 40

mg/

dLHD

L-C

>= 4

0 m

g/dL

Trea

tmen

t by

Trea

tmen

t by

P-Va

lue

(N =

596

8)(N

= 4

026)

(N =

254

0)(N

= 7

454)

TG <

or ³

15

0 m

g/dL

HDL-

C <

or ³

40 m

g/dL

P-va

lue

P-va

lue

Maj

or ca

rdio

vasc

ular

eve

nts

Even

t Rat

e: n

/ N

(%)

- ato

rvas

tatin

80

mg

- ato

rvas

tatin

10

mg

258

/ 297

0 (8

.69%

)28

6 / 2

998

(9.5

4%)

175

/ 202

0 (8

.66%

)26

1 / 2

006

(13.

01%

)12

8 / 1

243

(10.

30%

)18

2 / 1

297

(14.

03%

)30

5 / 3

747

(8.1

4%)

365

/ 370

7 (9

.85%

)

Haza

rd ra

tio:

ator

vast

atin

80m

g v

10m

g0.

907

0.65

10.

720

0.82

0P

= 0.

0109

P =

0.34

88

95%

con

fiden

ce in

terv

al(0

.766

, 1.0

73)

(0.5

38, 0

.788

)(0

.574

, 0.9

02)

(0.7

04, 0

.955

)

c2 stati

stics

P-va

lue

c2 = 1

.303

P =

0.25

37c2 =

19.

303

P <

0.00

01*

c2 = 8

.136

P =

0.00

43*

c2 = 6

.549

P =

0.01

05*

Abbr

evia

tion:

TG

= Tr

igly

cerid

es, H

DL-C

= h

igh-

dens

ity li

popr

otei

n ch

oles

tero

l.


71

4

supp

lem

enta

ry T

able

3 -

Trea

tmen

t Effe

ct b

y su

bgro

ups b

ased

on

TG a

nd H

DL-C

Lev

els f

or th

ose

with

and

with

out D

iabe

tes M

ellit

us

Base

line

Lipi

d Pr

ofile

TG <

150

mg/

dLTG

< 1

50 m

g/dL

TG >

= 15

0 m

g/dL

TG >

= 15

0 m

g/dL

Inte

racti

onIn

tera

ction

HDL-

C >=

40

mg/

dLHD

L-C

< 40

mg/

dLHD

L-C

>= 4

0 m

g/dL

HDL-

C <

40 m

g/dL

Trea

tmen

t by

Trea

tmen

t by

TG <

or ³

150

m

g/dL

HD

L-C

< or

³ 40

mg/

dL

P-

valu

eP-

valu

e

Patie

nts w

ith D

iabe

tes M

ellit

us (T

otal

N =

150

0)

(N =

556

)(N

= 1

60)

(N =

439

)(N

= 3

45)

Maj

or ca

rdio

vasc

ular

eve

nts

Even

t Rat

e: n

/ N

(%)

- ato

rvas

tatin

80

mg

- ato

rvas

tatin

10

mg

38

/ 263

(14.

45%

)39

/ 29

3 (1

3.31

%)

9 / 7

7 (1

1.69

%)

19 /

83 (2

2.89

%)

32 /

242

(13.

22%

)35

/ 19

7 (1

7.77

%)

24 /

165

(14.

55%

)42

/ 18

0 (2

3.33

%)

Haza

rd ra

tio:

ator

vast

atin

80m

g v

10m

g1.

111

0.47

20.

721

0.58

1P

= 0.

2052

P

= 0.

0597

95%

con

fiden

ce in

terv

al(0

.711

, 1.7

37)

(0.2

14, 1

.044

)(0

.447

, 1.1

65)

(0.3

52, 0

.959

)

c2 stati

stics

P-va

lue

c2 = 0

.213

P =

0.64

45c2 =

3.4

35P

= 0.

0638

c2 = 1

.784

P =

0.18

17

c2 = 4

.512

P =

0.03

37*

Patie

nts w

ithou

t Dia

bete

s Mel

litus

(Tot

al N

=84

94 )

(N =

442

8)(N

= 8

24)

(N =

203

1)(N

= 1

211)

Maj

or ca

rdio

vasc

ular

eve

nts

Even

t Rat

e: n

/ N

(%)

- ato

rvas

tatin

80

mg

- ato

rvas

tatin

10

mg

164

/ 221

9 (7

.39%

)18

5 / 2

209

(8.3

7%)

47 /

411

(11.

44%

)43

/ 41

3 (1

0.41

%)

71/1

023

(6.9

4%)

106/

1008

(10.

52%

)48

/ 59

0 (8

.14%

) 7

8 / 6

21 (1

2.56

%)

Haza

rd ra

tio:

ator

vast

atin

80m

g v

10m

g0.

878

1.09

90.

647

0.63

8P

= 0.

0180

*P

= 0.

9332

95%

con

fiden

ce in

terv

al(0

.712

, 1.0

84)

(0.7

27, 1

.662

)(0

.479

, 0.8

74)

(0.4

45, 0

.914

)

c2 stati

stics

P-va

lue

c2 = 1

.465

P =

0.22

62c2 =

0.2

10P

= 0.

6540

c2 = 8

.057

P =

0.00

45*

c2 = 6

.000

P =

0.01

43*

Abbr

evia

tion:

TG

= Tr

igly

cerid

es, H

DL-C

= h

igh-

dens

ity li

popr

otei

n ch

oles

tero

l.

72

Chapter 4su

pple

men

tary

Tab

le 4

- Tr

eatm

ent E

ffect

by

subg

roup

s bas

ed o

n TG

and

HDL

-C L

evel

s, fo

r tho

se o

n an

d off

LDL

-C ta

rget

Base

line

Lipi

d Pr

ofile

TG <

150

mg/

dLTG

< 1

50 m

g/dL

TG >

= 15

0 m

g/dL

TG >

= 15

0 m

g/dL

Inte

racti

onIn

tera

ction

HDL-

C >=

40

mg/

dLHD

L-C

< 40

mg/

dLHD

L-C

>= 4

0 m

g/dL

HDL-

C <

40 m

g/dL

Trea

tmen

t by

Trea

tmen

t by

TG <

or ³

15

0 m

g/dL

HD

L-C

< or

³ 40

mg/

dL

P-

valu

eP-

valu

e

Patie

nts o

n LD

L-C

Targ

et: B

asel

ine

LDL-

C <

100

mg/

dL (T

otal

N =

560

7)

(N =

286

5)(N

= 6

56)

(N =

119

8)(N

= 8

88)

Maj

or ca

rdio

vasc

ular

eve

nts

Even

t Rat

e: n

/ N

(%)

- ato

rvas

tatin

80

mg

- ato

rvas

tatin

10

mg

114

/ 142

8 (7

.98%

)12

4 / 1

437

(8.6

3%)

33 /

323

(10.

22%

)39

/ 33

3 (1

1.71

%)

49

/ 610

(8.0

3%)

50 /

588

(8.5

0%)

36 /

432

(8.3

3%)

66

/ 456

(14.

47%

)

Haza

rd ra

tio:

ator

vast

atin

80m

g v

10m

g0.

924

0.85

40.

943

0.56

2P

= 0.

2207

P

= 0.

0874

95%

con

fiden

ce in

terv

al(0

.716

, 1.1

91)

(0.5

37, 1

.358

)(0

.636

, 1.3

99)

(0.3

75, 0

.844

)

c2 stati

stics

P-va

lue

c2 = 0

.376

P =

0.53

98c2 =

0.4

42P

= 0.

5060

c2 = 0

.084

P =

0.77

22

c2 = 7

.718

P =

0.00

55*

Patie

nts n

ot o

n LD

L-C

Targ

et: b

asel

ine

LDL-

C ³

100

mg/

dL (T

otal

N =

438

6 )

(N =

211

9)(N

= 3

28)

(N =

127

2)(N

= 6

67)

Maj

or ca

rdio

vasc

ular

eve

nts

Even

t Rat

e: n

/ N

(%)

- ato

rvas

tatin

80

mg

- ato

rvas

tatin

10

mg

88

/ 105

4 (8

.35%

)10

0 / 1

065

(9.3

9%)

23 /

165

(13.

94%

)23

/ 16

3 (1

4.11

%)

54

/ 655

(8.2

4%)

91 /

617

(14.

75%

)36

/ 32

3 (1

1.15

%)

54 /

344

(15.

70%

)

Haza

rd ra

tio:

ator

vast

atin

80m

g v

10m

g0.

886

0.99

70.

537

0.69

2P

= 0.

0213

*P

= 0.

6695

95%

con

fiden

ce in

terv

al(0

.666

, 1.1

80)

(0.5

59, 1

.777

)(0

.383

, 0.7

52)

(0.4

54, 1

.054

)

c2 stati

stics

P-va

lue

c2 = 0

.682

P =

0.40

90c2 =

0.0

00P

= 0.

9913

c2 = 1

3.09

9P

= 0.

0003

*c2 =

2.9

39P

= 0.

0865

Abbr

evia

tion:

TG

= Tr

igly

cerid

es, H

DL-C

= h

igh-

dens

ity li

popr

otei

n ch

oles

tero

l, LD

L-C

= lo

w-d

ensit

y lip

opro

tein

cho

lest

erol

.


73

4

van der Valk FM, Hassing HC, Visser ME, Thakkar P, Mohanan A, Pathak K, Dutt C, Chauthaiwale V, Ackermans MT, Serlie MJ, Nieuwdorp M, Stroes ESG

Submitted

5 THE EFFECT OF A DIIODOTHYRONINE MIMETIC ON INSULIN SENSITIvITY IN MALE CARDIOMETABOLIC PATIENTS: A DOUBLE-BLIND RANDOMIZED CONTROLLED TRIAL

76

Chapter 5

ABSTRACTContext Our world is under the spell of obesity and their associated cardiometabolic co-morbidities. Since thyroid hormone mimetics are capable of uncoupling the beneficial metabolic effects of thyroid hormones from their deleterious effects on heart, bone and muscle, this class of drug is considered as adjacent therapeutics to weight-lowering strategies.

objective - This study investigated the safety and efficacy of TRC150094, a thyroid hormone mimetic.

Design - This 4-week, randomized, placebo-controlled, double-blind trial was conducted in India and The Netherlands. Hyperinsulinemic euglycemic clamp and 1H-Magnetic Resonance Spectroscopy (MRS) were performed before and after treatment.

subjects - In total 40 male subjects aged 30 to 65 years and characterized by a metabolic syndrome were included of whom none withdrew from the study after randomization.

intervention - Subjects were randomized at a 1:1 ratio to receive either TRC150094 dosed at 50 mg or placebo once daily for 4 weeks.

Main Outcome Measure - Primary efficacy was assessed via the change in hepatic or peripheral insulin sensitivity from baseline to week 4.

Results - TRC150094 dosed 50mg once daily was safe and well tolerated, however, insulin sensitivity, hepatic fat content and lipid profiles did not improve following 4 weeks of TRC150094 administration.

Conclusions - Collectively, these data show that, in contrast to the potent metabolic effects in experimental models, TRC150094 at a dose of 50mg daily does not improve the metabolic homeostasis in subjects at an increased cardiometabolic risk. Further studies are needed to evaluate whether TRC150094 has beneficial effects in patients with more severe metabolic derangement, such as overt diabetes mellitus and hypertriglyceridemia.

T2 analogue in meTabolic men

77

5

INTRODUCTION Despite the growing awareness of the detrimental impact of obesity on global health, the pandemic still shows no signs of abating.1 Currently, two thirds of the world’s population lives in countries where obesity-associated co-morbidity is the leading cause of premature death.2 Hence, there is an immense, unmet medical need for safe and effective therapies aimed at preventing the cardiometabolic sequelae associated with central adiposity, which can be implemented on top of weight-lowering strategies. Among the potential candidates, thyroid hormones (TH) have been shown to increase basal energy expenditure and oxygen consumption leading to a reduction in body weight with concomitant favorable improvements in lipid and carbohydrate metabolism.3, 4 In a clinical setting, however, TH have failed predominantly due to cardiotoxicity, as well as bone and muscle toxicity.5 Subsequently, selective TH analogs were designed in an effort to retain the beneficial effects whilst avoiding the toxic side effects. Analogs of TH with a 22-fold higher affinity for the hepatic thyroid hormone receptor beta (TRβ) than the ‘ubiquitous’ TRα isoform were reported to lower low density lipoprotein cholesterol (LDLc) by approximately 30% without significant heart, muscle or bone toxicity.6 Though, the first data with this compound were promising6, the Eprotirome program had to be discontinued due to the observation of increased cartilage damage following prolonged exposure to Eprotirome in dogs. 7

More recently, a mimetic of diiodothyronine (T2) - TRC150094 – was studied in a phase I study (data not published). TRC150094 has a very low potency for both TR isoforms when compared to T3, the active form of TH. The mechanism of action of T2 has been attributed to a direct, receptor-independent interaction of T2 with mitochondria.8 In preclinical studies, TRC150094 was shown to stimulate mitochondrial fatty acid oxidation (FAO) which led to a reduction of visceral adiposity in Wistar rats.9 In line, TRC150094 improved glucose tolerance and hepatic steatosis in obese Zucker spontaneously hypertensive fatty (ZSF1) rats with a concomitant reduction in plasma cholesterol and triglycerides in ZSF1 rats.10, 11 Most importantly, TRC150094 was not associated with any adverse safety signal in experimental models up to 24 weeks.9, 10 In phase I clinical studies, once daily oral administration of TRC150094 at doses of 50mg and 150mg for 28 days were well tolerated without any adverse safety signals in the obese subjects.

In the present study, we set out to evaluate the effect of TRC150094 on insulin sensitivity, liver fat content and lipid profile, as well as on safety markers in obese male subjects with an increased cardiometabolic risk.

78

Chapter 5

METHODSStudy designThis randomized, placebo-controlled, double-blind trial was conducted at 2 sites and was approved by the local Institutional Review board at Veeda Clinical Research, India and at the Academic Medical Centre (AMC), The Netherlands. The trial was conducted according to the principles of the International Conference on Harmonisation–Good Clinical Practice guidelines, and externally monitored by an independent contract research organization and registered on clinicaltrials.gov (NCT01408667). All participants provided written informed consent. In total, 40 subjects were enrolled; 20 subjects at Veeda Clinical Research, Ahmedabad, India, and 20 subjects at AMC, Amsterdam, The Netherlands. Each subject attended the study center for 5 visits; 1 screening visit, 2 study visits (1 baseline and 1 end of treatment), 1 intermediate safety visit and 1 post-study follow-up visit. During all 5 visits physical examination, vital signs, safety biochemistry and laboratory investigations were performed and evaluated by a physician blinded for treatment allocation. Before and after treatment a hyperinsulinemic euglycemic clamp and 1H-Magnetic Resonance Spectroscopy (MRS) were performed (details are provided in the Supplementary Information 1). The primary efficacy variable was the change in insulin sensitivity from baseline to week 4. Secondary efficacy variables were changes in hepatic fat content (IHTG) and lipid profile. Safety assessments included documentation of adverse events, blood pressure, heart rate, body temperature, weight and laboratory tests, including thyroid and liver-function tests.

Patient selectionEligible subjects were male, aged 30 to 65 years, and characterized by a metabolic syndrome based on the following criteria: increased waist circumference (Indian ≥90cm, Caucasian ≥102 cm), blood pressure ≥130/85 mmHg or use of antihypertensive drugs, fasting glucose >5.5 mmol/l - 11.0 mmol/l and fasting insulin level ≥10 mU/mL. Subjects were considered not eligible in case of history of somatic illness, including neoplasm, endocrine or neurologic disorders, active infection, unstable weight 3 months prior to inclusion or recent surgical procedure within 3 months of the study initiation; respectively systolic and diastolic blood pressure of ≥160mmhg or ≥100mmHg, impaired kidney function (eGFR <60 mL/min/1.73m2 as evaluated by MDRD method) or impaired liver function (ALT or AST >3 x ULN) at screening. After screening, subjects were randomized at a 1:1 ratio to receive either TRC150094 dosed at 50 mg or placebo once daily for 28 days.

statistical analysisContinuous data were analysed with parametric or non-parametric tests depending on the data distribution verified by the Shapiro-Wilk test. Within-group comparisons of pre- and post-treatment values were performed using the paired samples Student t-test or Wilcoxon signed ranks test.


79

5

Between-group comparisons of the relative changes were performed using the unpaired samples t-test or Mann–Whitney U test. Data for qualitative variables are presented as incidence rates (N, number and percent). The data of continuous variables were summarized using measures of central tendency (i.e. mean, median) and dispersion (i.e. standard deviation, range). Statistical analysis was performed using SPSS 19.0 (SPSS, Chicago, IL, USA).

RESULTSbaseline characteristicsFrom November 2011 through May 2012 we randomly assigned 40 men to TRC150094 (n = 20) or placebo (n = 20), all of whom completed the study protocol (Figure S 1). At baseline, clinical characteristics were comparable between TRC150094 and placebo group (Table 1). Baseline characteristics were also comparable between Indian and Caucasian subjects except for BMI, fasting insulin and FFA (Table S 1).

Table 1 – Characteristics of study subjects at baseline

TRC150094 ( N = 20) Placebo ( N = 20)

Age, y 49 ± 11 50 ± 10

Weight, kg 102 ± 18 103 ± 19

Body mass index, kg/m2 33.3 ± 4.5 33.6 ± 4.9

Waist, cm 112.5 ± 12 114.7 ± 11,8

Fasting plasma glucose, mmol/L 5.6 ± 1 5.3 ± 0,6

Fasting plasma insulin, mU/L 12 ± 7 14 ± 8

HOMA-IR 2.9 ± 1.8 3.5 ± 2.4

Cholesterol, mmol/L 4.63 ± 1.03 4.90 ± 0.74

HDLc 0.94 ± 0.24 1.01 ± 0.32

LDLc 2.91 ± 0.82 3.13 ± 0.71

TG 1.60 ± 1.06 1.67 ± 0.72

Fasting free fatty acids, mmol/L 0.51 ± 0.10 0.52 ± 0.19

Systolic blood pressure, mm Hg 140 ± 8 137 ± 10

Diastolic blood pressure, mm Hg 88 ± 4 87 ± 7

NOTE. Values are expressed as mean ± standard deviation. No significant differences in clinical variables were found between TRC and Placebo group at baseline, p < 0.05. The body mass index is the weight in kilograms divided by the square of the height in meters. HDLc, high-density lipoprotein cholesterol; LDLc, low-density lipoprotein cholesterol; TG, triglycerides.

80

Chapter 5

Safety analyses No serious adverse events were reported and no subjects withdrew from the study after enrolment. The total number of adverse events during the study was similar among the study groups (8 adverse events in both groups). The majority of these events were mild (81%) or moderate (19%). Supplementary Table 2 lists the number, intensity, relationship to treatment and type of adverse events that occurred during the study. No changes in vital signs were observed; blood pressure, heart rate, body temperature and weight remained stable throughout the study (Table 2). Liver function tests including ALT, AST and GGT did not change after TRC150094 treatment (Table 2). A marginal increase in FT4 in the treatment arm was observed, however, there was no concomitant reduction in TSH (Table 2).

Table 2 – safety analyses of study subjects at baseline and after 4 weeks

TRC150094 ( N = 20) Placebo ( N = 20)

Baseline Week 4 Baseline Week 4

Systolic blood pressure, mm Hg 140 ± 8 140 ± 9 137 ± 10 133 ± 9

Diastolic blood pressure, mm Hg 88 ± 4 88 ± 5 87 ± 7 85 ± 8

Pulse, beats/min 76 ± 10 74 ± 9 87 ± 7 75 ± 7

Body temperature, oC 36.6 ± 0.5 36.4 ± 0.5 36.6 ± 0.4 36.4 ± 0.4

Body weight, kg 101 ± 18 103 ± 21 103 ± 19 102 ±15

Liver function

ALT, U/L 40 ± 22 39 ± 20 34 ± 10 34 ± 15

AST, U/L 29 ± 12 27 ± 11 29 ± 13 26 ± 9

GGT, IU/L 44 ± 31 43 ± 28 46 ± 31 47 ± 34

Thyroid function

FT3, pmol/L 4.27 ± 1.02 4.88 ± 1.29 4.45 ± 0.68 5.11 ± 1.03

FT4, pmol/L 11.70 ± 2.12 12.57 ± 1.96 * 12.1 ± 1.84 12.3 ± 1.94

TSH, mIU/L 2.10 ± 0.00 2.23 ± 1.14 1.86 ± 0.67 2.01 ± 1.06

NOTE. Values are expressed as mean ± standard deviation. * Nonparametric test show a significant increased T3 in TRC150094 group (p= 0.005) and significant decrease in placebo group (p =0.049). Also, a significant increase in FT4 after TRC150094 treatment (p=0.025) ALT, alanine aminotransferase; AST, aspartate aminotransferase; GGT, gamma glutamyltransferase; FT3, free trio-iodothyronine; FT4, free thyroxine; TSH, thyroid stimulating hormone (thyrotropin).


81

5

Efficacy analyses in subjects at increased cardiometabolic riskEffect of TRC150094 on insulin sensitivity At baseline, male subjects were characterized by markedly impaired hepatic and peripheral insulin sensitivity, compared to reference values observed in healthy, non-obese control subjects12-15 (Figure 1 A-B). Hepatic insulin sensitivity was expressed as the suppression of Endogenous Glucose Production (EGP). After TRC150094 administration there was no improvement in suppression of endogenous glucose production (mean EGP suppression from 59.5 to 62.1%; p = 0.477) (Figure 1 A), whereas peripheral insulin sensitivity (expressed as the rate of glucose disappearance (Rd)) was not altered upon TRC150094 administration (mean Rd from 28.8 to 26.4 μmo/kg-1min-1, p = 0.185) (Figure 1 B). Although T2’s mechanism of action is expected to stimulate lipolysis and FAO, TRC150094 administration did not result in differences in fasting plasma FFA (mean FFA from 0.51 to 0.51 mmol/L, p= 0.887) or in insulin-mediated suppression of lipolysis (lipolysis suppression from 57 to 54%, p = 0.102) (Figure 1 C). Overview of efficacy results in glucose kinetics, lipolysis and glucoregulatory hormones at baseline and after TRC administration are provided in Supplementary Table 3.

Effect of TRC150094 on IHTG content and lipid profileIntrahepatic triglyceride (IHTG) content was measured with 1H MRS. At baseline, mean IHTG content was 10.6% (± 6.4%) in the whole group. After 4-weeks of treatment, IHTG was unaltered in both the TRC150094 and the placebo group (Figure 1). Similarly, no change in lipid profile, i.e. total cholesterol, LDLc, HDLc or TG, was detected in the patients on either TRC150094 or placebo (Figure 1). Responses divided per treatment arm are provided in Supplementary Table 4.

subgroup analysis in subjects with severe metabolic derangementTo evaluate whether the response differed between subjects with mild and severe metabolic derangement, we analyzed the subjects with a mean TG above 1.64 mmol/L compared to those below the mean plasma TG. Subgroup analysis showed a numerical reduction of IHTG content in the highest TG group (absolute IHTG from 12.7% (± 3.9) to 11.8% (± 4.3) (p=0.378) with a relative IHTG change of – 6.3% (p=0.682), which did not reach statistical significance. In the subjects below mean TG, IHTG levels remained stable (absolute IHTG from 9.9% (± 6.9) to 10.4% (± 8.1), as reflected by a relative IHTG change of +3.6% (p=0.759). Changes in hepatic and peripheral insulin sensitivity were not significantly different between upper versus lower TG groups. Finally, TG levels decreased in the upper TG group following TRC150094 (from 2.66 ± 1.17 to 2.26 ± 1.31 mmol/L, p =0.012) whereas no change was observed in the lower TG group. See supplementary Table 5 for an overview of the subgroup analyses.

82

Chapter 5

Figure 1 – Efficacy data TRC150094 in males with increased cardiometabolic risk. a-D; Box plots of hepatic insulin sensitivity (suppression of EGP %), peripheral insulin sensitivity (Rd umol*kg-1min-1), hepatic fat content (IHTG %) and insulin mediated suppression of lipolysis (suppression of lipolysis %) before after TRC administration. Blue background depicts reference values in healthy population, based on historical data12-15. E; Bar graph of lipid profile showing no improvement after TRC150094 administration.


83

5

DISCUSSIONIn the present study we show that short-term administration of TRC150094 dosed 50mg once daily is safe and well tolerated. Neither hepatic, nor peripheral insulin sensitivity improved in subjects at an increased cardiometabolic risk. In line, IHTG content and plasma lipid profiles were not altered following 4 weeks of TRC150094 administration. A subgroup analysis in subjects with TG levels above the mean did reveal a significant reduction in TG levels but no changes in insulin sensitivity nor IHTG. Collectively, these data show that, in contrast to the potent metabolic effects in experimental models, TRC150094 at a dose of 50mg daily does not improve the metabolic homeostasis in subjects at an increased cardiometabolic risk. Further studies are needed to evaluate whether TRC150094 may have an effect in subjects with more severe metabolic derangement, such as overt diabetes mellitus and hypertriglyceridemia.

insulin sensitivityIn the present study all enrolled subjects were characterized by decreased EGP suppression as well as decreased peripheral glucose disposal rate, indicative of the presence of both hepatic and peripheral insulin resistance. Also, intrahepatic fat accumulation as assessed via IHTG content showed hepatic steatosis in all subjects. Following 4 weeks of TRC150094 administration at a dose of 50mg once daily, neither hepatic nor peripheral insulin sensitivity changed. In line, hepatic fat content and lipid profile were unaltered. This apparent discrepancy between the marked impact of TRC150094 on glycemic profile, hepatic fat accumulation and serum lipids in experimental protocols9, 10, 16, 17 and the absence of any change in the present clinical study may have several explanations, consisting of the mechanism of action, the dose and concentration of TRC150094.

First, the mechanism of action of di-iodothyronine (T2). The biologically active thyroid hormone tri-iodothyronine (T3) exerts its effects via specific nuclear receptors; namely TR α and β.3 T2 has a 50-400 times lower affinity for TR than T3, making it unlikely that TR activation contributes to the effects of T2.18 Extensive preclinical work, however, did substantiate a rapid effect of T2 on energy expenditure in rats, which were shown to be mediated by direct effects on mitochondria independent of classical nuclear thyroid receptors.19 The T2 mimetic TRC150094, which is also associated with minimal TR transcriptional activation,9 was observed to increase whole body mitochondrial fatty acid oxidation (FAO) and resting metabolic rate (RMR) in rats, leading to marked improvements of glucose and lipid homeostasis. 9, 10, 16, 17 In contrast, data on the (patho)physiological relevance of T2 in humans are absent.20 In fact, proof for a receptor-independent effect of T2 in humans is lacking altogether. In the present study, TRC150094 also failed to increase FAO in terms of decreased plasma FFA and less insulin mediated suppression of lipolysis. As this study was of shorter duration i.e. 28 days, we can not exclude the possibility that the receptor-independent activities of T2 may become apparent in humans after a longer duration. 18

84

Chapter 5

Secondly, in the current study 50 mg once daily dosed was selected based on drug exposure (AUC) as well as safety and tolerability data obtained from earlier animal models9-11 and a phase I Multiple Ascending Dose study in humans (data not published). In previous animal studies, the plasma exposure in which significant effects on insulin sensitivity and hepatic lipid content were observed ranged between 3.8 to 11.8 µg*h/ml. Steady state exposure of TRC150094 in humans was observed between 2.7 to 8.0 µg*h/ml after administration of 50mg once a day for 28 days. Nevertheless, the selected dose of TRC150094 50 mg once daily may have been insufficient. This could be explained by human equivalent dose (HED) calculation 21 from animal efficacy studies. The current dose of 0.5 mg/kg (mean weight approx 100kg) in humans may be at the lower end, since conversion for drug dosage between rats and humans indicates a HED of approximately 4 mg/kg.21 Thus efficacy of TRC150094 at a higher dose of 75 to 100 mg once daily needs further exploration in future studies.

IHTG and plasma TG changesFollowing 4 weeks of TRC150094 administration we did not observe changes in hepatic and serum lipids. The absence of a reduction in hepatic fat can in part explain the lack of an effect on insulin sensitivity, although the association between hepatic fat content and insulin sensitivity is ambiguous.22 Subgroup analyses included subjects with severe metabolic derangement,23 identified via plasma TG >1.64 mmol/l at baseline. Whereas the ‘high-TG’ subjects showed comparable hepatic fat content levels at baseline compared to the low TG subjects, following TRC150094 administration hepatic fat content following TRC150094 administration was reduced numerically by 6% in the high TG subjects versus no change in the subjects with lower TG levels. In line, hepatic and peripheral insulin sensitivity did show a trend towards improvement (respectively +2.69% and +3.61%). Thus, it cannot be ruled out that an effect may have been observed in case of selection of metabolic syndrome patients with markedly elevated TG levels.

Safety and tolerability Overall, TRC150094 administration was well tolerated and showed no safety concerns. The modest changes in FT4 were unexpected24, since the affinity of TRC150094 for TR α and β is extremely low. Most importantly, we can exclude any biological relevance since no decrease in TSH was observed. Besides, other clinical manifestations of TR activation were absent, such as changes in blood pressure, heart rate, body temperature and body weight following TRC150094 administration.

In conclusion, in the present phase 2, randomized double-blind controlled trial we show that TRC150094 did not improve insulin sensitivity and lipid metabolism or decrease hepatic steatosis in obese insulin resistant subjects with an increased cardiometabolic risk. Since subgroup analysis in subjects with high triglyceride levels provided a trend towards improvement, future studies should address the potential impact of TRC150094 administration at higher dose, particularly in patients at a high cardiometabolic risk with elevated TG levels.


85

5

REFERENCE LIST1. Ogden CL, Carroll MD, Kit BK, Flegal KM. Prevalence of obesity and trends in body mass index among

US children and adolescents, 1999-2010. JAMA 2012;307:483-490.

2. WHO. The World Health Statistics 2012. WHO 2012.

3. Cheng SY, Leonard JL, Davis PJ. Molecular aspects of thyroid hormone actions. Endocr Rev 2010;31:139-170.

4. Klieverik LP, Coomans CP, Endert E et al. Thyroid hormone effects on whole-body energy homeostasis and tissue-specific fatty acid uptake in vivo. Endocrinology 2009;150:5639-5648.

5. Klein I, Ojamaa K. Thyroid hormone and the cardiovascular system. N Engl J Med 2001;344:501-509.

6. Ladenson PW, Kristensen JD, Ridgway EC et al. Use of the thyroid hormone analogue eprotirome in statin-treated dyslipidemia. N Engl J Med 2010;362:906-916.

7. Per Bengtsson. KARO BIO Terminsates the Eprotirome Program. 2012

8. Moreno M, de LP, Lombardi A, Silvestri E, Lanni A, Goglia F. Metabolic effects of thyroid hormone derivatives. Thyroid 2008;18:239-253.

9. Cioffi F, Zambad SP, Chhipa L et al. TRC150094, a novel functional analog of iodothyronines, reduces adiposity by increasing energy expenditure and fatty acid oxidation in rats receiving a high-fat diet. FASEB J 2010;24:3451-3461.

10. Silvestri E, Glinni D, Cioffi F et al. Metabolic effects of the iodothyronine functional analogue TRC150094 on the liver and skeletal muscle of high-fat diet fed overweight rats: an integrated prot-eomic study. Mol Biosyst 2012;8:1987-2000.

11. Zambad SP, Munshi S, Dubey A et al. TRC150094 attenuates progression of nontraditional cardio-vascular risk factors associated with obesity and type 2 diabetes in obese ZSF1 rats. Diabetes Metab Syndr Obes 2011;4:5-16.

12. Soeters MR, Lammers NM, Dubbelhuis PF et al. Intermittent fasting does not affect whole-body glucose, lipid, or protein metabolism. Am J Clin Nutr 2009;90:1244-1251.

13. van Raalte DH, Brands M, van der Zijl NJ et al. Low-dose glucocorticoid treatment affects multiple aspects of intermediary metabolism in healthy humans: a randomised controlled trial. Diabetologia 2011;54:2103-2112.

14. Pasarica M, Rood J, Ravussin E, Schwarz JM, Smith SR, Redman LM. Reduced oxygenation in human obese adipose tissue is associated with impaired insulin suppression of lipolysis. J Clin Endocrinol Metab 2010;95:4052-4055.

15. Hickner RC, Racette SB, Binder EF, Fisher JS, Kohrt WM. Suppression of whole body and regional lipolysis by insulin: effects of obesity and exercise. J Clin Endocrinol Metab 1999;84:3886-3895.

16. Moreno M, Silvestri E, De MR et al. 3,5-Diiodo-L-thyronine prevents high-fat-diet-induced insulin re-sistance in rat skeletal muscle through metabolic and structural adaptations. FASEB J 2011;25:3312-3324.

17. de LP, Cioffi F, Senese R et al. Nonthyrotoxic prevention of diet-induced insulin resistance by 3,5-dii-odo-L-thyronine in rats. Diabetes 2011;60:2730-2739.

18. Ball SG, Sokolov J, Chin WW. 3,5-Diiodo-L-thyronine (T2) has selective thyromimetic effects in vivo and in vitro. J Mol Endocrinol 1997;19:137-147.

86

Chapter 5

19. Horst C, Rokos H, Seitz HJ. Rapid stimulation of hepatic oxygen consumption by 3,5-di-iodo-L-thyro-nine. Biochem J 1989;261:945-950.

20. Pinna G, Hiedra L, Meinhold H et al. 3,3’-Diiodothyronine concentrations in the sera of patients with nonthyroidal illnesses and brain tumors and of healthy subjects during acute stress. J Clin Endocrinol Metab 1998;83:3071-3077.

21. Reagan-Shaw S, Nihal M, Ahmad N. Dose translation from animal to human studies revisited. FASEB J 2008;22:659-661.

22. Visser ME, Lammers NM, Nederveen AJ et al. Hepatic steatosis does not cause insulin resistance in people with familial hypobetalipoproteinaemia. Diabetologia 2011;54:2113-2121.

23. Eckel RH. The complex metabolic mechanisms relating obesity to hypertriglyceridemia. Arterioscler Thromb Vasc Biol 2011;31:1946-1948.

24. Ribeiro MO. Effects of thyroid hormone analogs on lipid metabolism and thermogenesis. Thyroid 2008;18:197-203.

25. Finegood DT, Bergman RN, Vranic M. Estimation of endogenous glucose production during hyper-insulinemic-euglycemic glucose clamps. Comparison of unlabeled and labeled exogenous glucose infusates. Diabetes 1987;36:914-924.

26. Ackermans MT, Pereira Arias AM, Bisschop PH, Endert E, Sauerwein HP, Romijn JA. The quantification of gluconeogenesis in healthy men by (2)H2O and [2-(13)C]glycerol yields different results: rates of gluconeogenesis in healthy men measured with (2)H2O are higher than those measured with [2-(13)C]glycerol. J Clin Endocrinol Metab 2001;86:2220-2226.

27. Ackermans MT, Ruiter AF, Endert E. Determination of glycerol concentrations and glycerol iso-topic enrichments in human plasma by gas chromatography/mass spectrometry. Anal Biochem 1998;258:80-86.

28. Matthews DR, Hosker JP, Rudenski AS, Naylor BA, Treacher DF, Turner RC. Homeostasis model assess-ment: insulin resistance and beta-cell function from fasting plasma glucose and insulin concentra-tions in man. Diabetologia 1985;28:412-419.

29. Steele R. Influences of glucose loading and of injected insulin on hepatic glucose output. Ann N Y Acad Sci 1959;82:420-430.

30. Kotronen A, Vehkavaara S, Seppala-Lindroos A, Bergholm R, Yki-Jarvinen H. Effect of liver fat on insu-lin clearance. Am J Physiol Endocrinol Metab 2007;293:E1709-E1715.

31. Naressi A, Couturier C, Devos JM et al. Java-based graphical user interface for the MRUI quantitation package. MAGMA 2001;12:141-152.

32. de Bazelaire CM, Duhamel GD, Rofsky NM, Alsop DC. MR imaging relaxation times of abdominal and pelvic tissues measured in vivo at 3.0 T: preliminary results. Radiology 2004;230:652-659.

33. Szczepaniak LS, Babcock EE, Schick F et al. Measurement of intracellular triglyceride stores by H spectroscopy: validation in vivo. Am J Physiol 1999;276:E977-E989.


87

5

SUPPLEMENTAL MATERIALsupplementary information 1 – Methods A. Hyperinsulinemic euglycemic clampPrior to the study day, all subjects refrained form vigorous exercise for 48 hours. After an overnight fast, subjects were admitted to the metabolic ward of the study centre at 07:15 hours. A catheter was inserted in an antecubital vein for infusion of stable isotope tracers, insulin and glucose. Another catheter was inserted into a contralateral hand vein and kept in a thermoregulated (60°C) Plexiglas box for sampling of arterialised venous blood. Saline was infused as NaCl 0.9% at a rate of 50 ml/h to sustain catheter patency. [6,6-²H2]glucose and [1,1,2,3,3-2H5]glycerol were infused as tracers (>99% enriched; Cambridge Isotopes, Andover, MA, USA) to study glucose kinetics and lipolysis (total triacylglycerol hydrolysis), respectively. At time 0 (08:30 hours) blood samples were drawn for determination of background enrichments, where after a continuous infusion of isotopes was started ([6,6-²H2]glucose and [1,1,2,3,3-2H5]glycerol, both at a rate of 0.11 μmol*kg−1 min−1, with a priming dose equivalent to 80 min of infusion) and continued until the end of study. After an equilibration time of 150 min, three blood samples were taken for the measurement of isotope enrichments and one for the measurement of glucoregulatory hormones and NEFA. Thereafter, a two-step hyperinsulinaemic–euglycaemic clamp was started. A continuous infusion of insulin (Actrapid 100 U/ml; Novo Nordisk Farma, Alphen aan de Rijn, the Netherlands) was started for 130 min at the rate of 20 mU [m2 body surface area]−1min–1, followed by an infusion of insulin at a rate of 60 mU [m2 body surface area]−1min–1 for another 130 min. Plasma glucose levels were measured every 10 min at the bedside. Glucose was infused as 20% glucose at a variable rate, to maintain a plasma glucose concentration of 5.0 mmol/l. [6,6-²H2]glucose was added to the 20% glucose solution to achieve glucose enrichments of 1% to approximate the values for enrichment reached in plasma and thereby minimise changes in isotopic enrichment due to changes in the infusion rate of exogenous glucose.25 During the last 40 min of both hyperinsulinaemic periods, blood samples were drawn at 5 min intervals for determination of isotope enrichments and glucoregulatory hormones. During the study day, all subjects remained fasted but were allowed to drink water.

B. Glucose and glucoregulatory hormones measurementsPlasma glucose concentrations were measured with the glucose oxidase method using a YSI analyzer. [6,6-2H2]glucose enrichment (tracer-to-tracee ratio) was measured as reported earlier26 with an intra-assay variation of 0.5–1% and an inter-assay variation of 1% and a detection limit of 0.04%. [1,1,2,3,3-2H5]glycerol enrichment was determined with an intra-assay variation of 1–3% for glycerol and 4% for [1,1,2,3,3-2H5]glycerol, and inter-assay variation of 2–3% for glycerol and 7% for [1,1,2,3,3-2H5]glycerol, as reported earlier.27 Insulin was determined on an Immulite 2000 system (Diagnostic Products, Los Angeles, CA, USA). Insulin was measured with a chemiluminescent immunometric assay with intra-assay variation of 3–6%, inter-assay variation of 4–6% and detection limit of 15 pmol/l. Calculations and statistics HOMA of insulin resistance (HOMA-IR) was calculated using the formula described previously by Matthews et al.28 Endogenous glucose production (EGP)

88

Chapter 5

and peripheral glucose uptake (rate of disappearance [Rd]) were calculated using the modified forms of the Steele equations.29 EGP and Rd were expressed as μmol kg−1min−1. Insulin clearance was calculated as the rate of insulin infusion (mU [m2 body surface area]−1 min−1) divided by the mean plasma insulin concentration during the clamp.30 Lipolysis (glycerol turnover) was calculated using formulas for steady-state kinetics adapted for stable isotopes and was expressed as μmol kg−1min−1.27, 29 Lipolysis was assessed as percentage change in rate of appearance of glycerol from the basal to low dose insulin-stimulated state. Plasma FFA concentrations were measured with an enzymatic colorimetric method (NEFA-C test kit; Wako Chemicals GmbH, Neuss, Germany) (intra-assay variation 1%, total-assay variation 4-15%; detection limit 0.02 mmol/L)

C. 1H MRS¹H-MRS spectra were acquired using a 3.0 T Intera (Philips, Best, the Netherlands). During the measurements, subjects remained in the supine position within the MRI scanner. IHTG content was obtained using single-voxel ¹H-MRS, using a body array coil as the transmitter and phased surface coils as receivers. MRS measurements were acquired during breathhold, using single-voxel stimulated acquisition mode (TE/TR 20/3.000 ms, six acquisitions). Volumes of interest in the liver were located away from major vascular structures and bile ducts. Voxel size was 27 mm3. The water and fat resonance peaks, located at 4.65 and 1.3 ppm, were integrated using jMRUI software31, and relative fat content was expressed as the ratio of the fat peak area over the cumulative water and fat peak areas. Calculated peak areas of water and fat were corrected for T2 relaxation (T2water, 34 ms; T2fat, 68 ms32) and the percentage hepatic fat content was calculated.33


89

5

Clamp

1H MRS

TRC150094

(N=20)

Placebo

(N=20)

End of study (N=40)

36 screenfailures*

76 eligible subjects

40 subjects included

Safety

visit

Screening Randomisation Close out Follow up visit

Clamp

1H MRS

Baseline

<4 wks 28 days 35 days

supplementary Figure 1 - overview of study scheme. * Screenfailures in specific; due to malignancy in history, ECG abnormalities at screening, too low fasting insulin or glucose, haemoglobin, age and/or withdrawing of consent after screening.

90

Chapter 5

supplementary Table 1 – Characteristics of study subjects at baseline between Ethnicities

Indian ( N = 20) Caucasian ( N = 20)

TRC150094( N = 10 )

Placebo( N = 10 )

TRC150094( N = 10 )

Placebo( N = 10 )

Age, y 41 ± 7 43 ± 6 57 ± 6 57 ± 8

Weight, kg 87 ± 7 90 ± 11 116 ± 13 116 ± 16

Body mass index, kg/m2* 30.2 ± 2.7 31.1 ± 3.3 36.4 ± 3.7 36.0 ± 5.1

Waist, cm 105 ± 6 107 ± 7 120 ± 11 123 ± 11

Fasting plasma glucose, mmol/L 5.7 ± 1.3 5.3 ± 0.7 5.5 ± 0.7 5.3 ± 0.4

Fasting plasma insulin, mU/L * 8 ± 3 13 ± 9 17 ± 8 16 ± 5

HOMA-IR 2.0 ± 1.0 3.3 ± 2.8 3.7 ± 2.1 3.7 ± 2.0

Cholesterol, mmol/L 4.68 ± 1.09 4.93 ± 0.90 4.59 ± 1.02 4.87 ± 0.59

HDLc 1.02 ± 0.30 1.04 ± 0.42 0.87 ± 0.14 0.98 ± 0.21

LDLc 2.89 ± 0.81 3.03 ± 0.87 2.93 ± 0.87 3.24 ± 0.52

TG 1.50 ± 1.16 1.91 ± 0.91 1.72 ± 1.00 1.44 ± 0.39

Plasma free fatty acids, mmol/L* 0.52 ± 0.12 0.40 ± 0.09 0.50 ± 0.09 0.65 ± 0.18

Systolic blood pressure, mm Hg 139 ± 2 139 ± 2 145 ± 10 140 ± 11

Diastolic blood pressure, mm Hg 89 ± 1 89 ± 1 90 ± 4 89 ± 6

NOTE. Values are expressed as mean ± standard deviation. *Baseline characteristics were comparable between Indian and Caucasian subjects except for BMI, fasting insulin and FFA. The body mass index is the weight in kilograms divided by the square of the height in meters. HDLc, high-density lipoprotein cholesterol; LDLc, low-density lipoprotein cholesterol; TG, triglycerides.


91

5

supplementary Table 2 – Number of patients with adverse events

TRC150094 ( N = 20) Placebo ( N = 20)

Serious adverse event 0 0

Adverse event 8 8

Intensity

Mild 6 7

Moderate 2 1

Relationship to TRC150094/Placebo

Not likely 4 3

Possible 4 5

Probable 0 0

Event

Back pain 2 2

Blurred vision 1 1

Diarrhea 1 1

Dry mouth 0 1

Fatigue 1 0

Flu like symptoms 0 2

Headache 1 2

Heartburn 1 0

Hip pain 0 2

Increased appetite 1 0

Insomnia 0 1

Polyuria 1 0

Rash 1 0

92

Chapter 5

supplementary Table 3 – Glucose Kinetics, Glucoregulatory hormones in TRC and Placebo Group at baseline and after 4 weeks Treatment

TRC ( N = 20) Placebo ( N = 20 )

Baseline 4 weeks Baseline 4 weeks

Glucose, mmol/L

Basal 5.2 (4.1-8.4) 5.5 (4.5-8.9) 5.1 (4.4-6.5) 5.3 (4.6-7.8)

Step 1 5.1 (4.7-6.4) 5.1 (4.8-5.8) 5.0 (4.8-5.6) 5.1 (4.9-6.2)

Step 2 5.0 (4.8-5.5) 5.1 (4.6-5.4) 4.9 (4.4-5.4) 5.0 (4.6-5.8)

Insulin, mU/L

Basal 11 (2-28) 11 (4-29) 13 (4-37) 13 (2-24)

Step 1 32 (19-68) 30 (17-81) 39 (19-149) 40 (17-55)

Step 2 96 (62-225) 97 (54-197) 98 (58-145) 106 (51-162)

EGP, μmo/kg-1min-1

Basal 9.1 (7.5-13.4) 9.0 (7.4-13.4) 8.7 (7.3-13.1) 8.7 (6.8-11.9)

Step 1 3.9 (1.2-3.8) 2.7 (0.6-7.7) 3.2 (0.7-5.7) 2.4 (0-9.3)

EGP suppr, % 60 (52-67) 62 (55- 70) 67 (60-74) 61 (50-72)

Rd, μmo/kg-1min-1 27.5 (11.7-48.0) 24.2 (11.1-46.9) 27.3 (16.6-39.4) 29.9 (14.2-45.5)

Plasma FFA, mmol/L 0.51 (0.36-0.77) 0.51 (0.26-0.67) 0.52 (0.25-0.94) 0.47 (0.21-0.69)

Lipolysis, μmolk-1min-1

Basal 2.1 (1.2-3.2) 2.2 (1.3-4.6) 2.2 (1.1-3.3) 2.2 (1.4-4.1)

Step 1 0.9 (0.6-1.7) 1.0 (0.6-3.2) 1.0 (0.6–1.6) 0.9 (0.5-1.5)

Lipolysis suppr, % 57 (32-77) 54 (20-77) 53 (14-72) 56 (33-78)

NOTE. Values are expressed medians (minimum – maximum)Step 1 is measurements during low dose insulin infusion and step 2 is measurements during high dose insulin infusion during hyperinsulinemic euglycemic clamp.

supplementary Table 4 – iHTG content and lipid profiles in TRC and Placebo Group at baseline and week 4

TRC ( N = 20) Placebo ( N = 20 )

Baseline 4 weeks Baseline 4 weeks

Liver

IHTG content, % 10.8 ± 6.1 10.9 ± 6.9 10.4 ± 6.9 10.6 ± 7.2

Serum

Tot al cholesterol, mmol/L 4.63 ± 1.03 4.58 ± 1.02 4.90 ± 0.74 4.88 ± 0.79

LDLc 2.91 ± 0.82 2.81 ± 0.79 3.13 ± 0.71 3.00 ± 0.71

HDLc 0.94 ± 0.24 0.97 ± 0.21 1.01 ± 0.32 0.97 ±0.31

TG 1.61 ± 1.06 1.75 ± 1.00 1.67 ± 0.72 2.01 ± 0.94

NOTE. Values are expressed mean ± standard deviation.


93

5

supp

lem

enta

ry T

able

5 –

sub

grou

p an

alys

es in

sub

ject

s with

Hig

h TG

TRC

( N =

7)

Plac

ebo

( N =

8 )

Base

line

4 w

eeks

Rel c

hang

eP

Base

line

4 w

eeks

Rel

chan

geP

EGP

supp

r, %

56.2

± 1

0.5

58.3

± 1

5.4

+ 2.

690.

510

70.0

± 1

7.2

64.4

± 1

7.9

- 6.7

30.

294

Rd, μ

mol

/kg/

min

25.9

± 8

.127

.0 ±

11.

4+

3.61

0.71

226

.5 ±

5.9

25.5

± 5

.4- 0

.44

0.70

6

IHTG

, %12

.7 ±

3.9

11.8

± 4

.3- 6

.31

0.31

79.

3 ±

4.0

9.5

± 5.

4- 2

.42

0.80

6

Tota

l cho

lest

erol

, mm

ol/L

4.87

± 1

.33

4.60

± 1

.52

- 6.4

60.

237

5.13

± 0

.98

5.23

± 0

.76

+ 2.

930.

542

LDLc

2.79

± 1

.08

2.67

± 1

.14

- 4.6

00.

488

3.25

± 0

.81

3.29

± 0

.69

- 0.2

20.

805

HDLc

0.86

± 0

.12

0.90

± 0

.11

+ 5.

890.

144

0.82

± 0

.13

0.83

± 0

.19

+ 2.

440.

974

TG2.

66 ±

1.1

72.

26±

1.31

- 18.

60.

012

2.33

± 0.

662.

47 ±

0.6

1+

7.23

0.32

6

Subg

roup

ana

lyse

s of

sub

ject

s w

ith s

erum

TG

>1.6

4 m

mol

/L a

t bas

elin

e. N

OTE

. Val

ues

are

expr

esse

d m

ean

± st

anda

rd d

evia

tion.

Rel

ative

cha

nge

is th

e m

ean

rela

tive

chan

ge in

per

cent

age.

Bernelot Moens SJ, Hassing HC, Nieuwdorp M, Stroes ESG, Dallinga-Thie GM

Accepted Clinical Lipidology

6 HYPERTRIGLYCERIDEMIA: THE FUTURE OF GENETICS TO GUIDE INDIvIDUALIZED THERAPEUTIC STRATEGIES

96

Chapter 6

ABSTRACTThe use of genetic information to exploreand treat diseases is ever expanding, varying from the use of classical approaches for monogenetic disorders, to the growing genome wide association studies (GWAS) in understanding more complex traits. In hypertriglyceridemia, development has rapidly progressed. We now have seen the use of genetic information by treating monogenetic disorders with gene therapy for the first time being implemented successfully in human subjects. Also, ASO therapy in mice and very recently also in humans has been shown to lower triglyceride levels. In polygenetic disease, the use of large-scale GWAS studies has changed our perceptive of the underlying phenotypes, showing a large overlap in common genetic determinants. This information is translated to understandingreaction to drug therapy, but also in relation to environment interaction. Finally, the use of genetics in predicting risk of cardiovascular disease is continuously being studied, although application is still a far road ahead.

Gene therapy in hypertriGlyceridemia

97

6

ExECUTIvE SUMMARYClinical hypertriglyceridemiaHigh plasma triglyceride levels, both fasting and post-prandial, are associated with cardiovascular disease and acute pancreatitis

TG MetabolismTriglycerides are derived from dietary sources in the intestine or de novo hepatic synthesis. In the circulation, they are lipolysed by LPL and used for energy or storage, finally remnants are cleared by the liver.

Classification of hypertriglyceridemiaHypertriglyceridemia, defined as fasting plasma triglycerides > 200 mg/dl (or > 2.2 mmol/l) can be of secondary origin, due to diabetes or obesity, or primary causes due to genetic loss of LPL function.

Current treatmentA combination of diet, exercise and drugs is currently recommended as the first line treatment of hypertriglyceridemia. Drug classes used are fibrates, nicotinic acid and n3 fatty acids.

implications of genetics in hypertriglyceridemiaThe use of genetic information in hypertriglyceridemia comprises approaches to unravel and treat monogenetic disease, as well as the use of genome wide association studies (GWAS) to understand more complex traits. In the near future whole exome sequencing analysis will provide new insight in the genetic background of hypertriglyceridemia.

Genetic approaches in monogenetic disordersBoth the use of gene therapy and antisense therapy are currently undergoing swift development, leading to treatment options for monogenetic disease currently resistant to all known therapies

individualized therapy in polygenetic disordersThe use of genetic information in polygenetic disorders is complex, but has been aptly used to explain differences in reaction to drugs based on different phenotypes. Also, this information is at the base of developing genetic risk scores.

ConclusionsGenetic information in hypertriglyceridemia has been successfully used to explain monogenetic hyperTG disease, and gene therapy has for the first time been implemented in human subjects for this disease. Although the use in polygenetic disease is more complex, ever expanding information has led to a better understanding of these traits and will possibly lead to new approaches in prediction and treatment.

98

Chapter 6

CLINICAL HYPERTRIGLYCERIDEMIAHigh plasma levels of triglycerides (TG) have been recognized as an independent risk factor for cardiovascular disease (CVD) since 19981. This association holds true for fasting triglyceride levels2, but also for non-fasting levels, where the postprandial remnant lipoproteins are thought to contribute to atherogenesis3–5. An exception are the rare cases of high and severe high triglycerides in the familial chylomicronemia syndrome, where the extreme levels are not clearly linked to an increased risk for the development of CVD6,7, but their role in the onset of acute pancreatitis has since long been recognized8 and widely accepted9.

TG METABOLISM Triglyceride-rich lipoproteins originate from the intestine (apoB-48 containing chylomicrons), where they transport exogenous derived lipids into the circulation, or from the liver (apoB-100 containing very-low-density lipoprotein (VLDL)). In the intestine, TGs derived from dietary sources are hydrolyzed and the residues (2-monoacylglycerol (2-MG) and fatty acid (FA)) are taken up by enterocytes where they are re-synthesized into TGs by the enzyme acyl-CoA:diaglycerolacyltransferase 1 (DGAT1). Hereafter, lipidation of apoB48 through microsomal triglyceride transfer protein (MTTP) is the first step into chylomicron formation and eventually, the nascent chylomicrons reach the systemic circulation through the lymphatic system10,11. Triglycerides synthesized in the liver, deriving the required fatty acids from either de novo lipogenesis (DNL) or lipolysis in adipose tissue, are packaged into VLDL in the endoplasmic reticulum (ER), with apoB100 as their main apolipoprotein. Next they are transported to the Golgi apparatus for further lipidation, after which they are excreted by hepatocytes into the bloodstream. Circulating VLDL particles carry additional proteins such as apoCII12, apoAV13 and apoCIII14. These proteins function as co-factors for Lipoprotein Lipase (LPL), the protein responsible for the lipolysis that is required to generate free fatty acids (FFA) which can enter peripheral tissues as a source for energy and storage. LPL is synthesized in the parenchymal cells of tissues that require these fatty acids, and lipase maturation factor 1 (LMF1) is essential for folding it into a functional form15. LPL is then transported to the endothelial cell surface where glycosyl-phosphatidyl-inositol anchored high-density lipoprotein binding protein 1 (GPIHBP1) serves as a binding platform for LPL to come into proximity with TG particles16.

Finally, the tissues take up the newly freed fatty acids, whereas the liver eventually clears the remnant particles. Hepatic remnant clearance appears to be the result of several overlapping mechanisms of independent receptors, including the LDL receptor, the LDL receptor related protein 1(LRP1), and heparan sulfate proteoglycans (HSPGs)17,18.


99

6

CLASSIFICATION OF HYPERTRIGLYCERIDEMIA Hypertriglyceridemia, defined as fasting plasma triglycerides > 200 mg/dl (or > 2.2 mmol/l)2, can arise from increased TG production, reduced TG catabolism, or a combination hereof. A mild to moderate elevation (200-500 mg/dl or 2.2-5.5 mmol/l) can usually be attributed to secondary causes, including obesity, diabetes mellitus, pregnancy, alcohol and different drugs19, due to several different mechanisms. Adequately distinguishing these and other, primary, causes of high triglyceride plasma levels may prove difficult and understanding of the underlying mechanisms is crucial, as the origin will mainly determine the choice of treatment. For example, an increase in VLDL production may be found as a result of excess flux of fatty acids to the liver, by several causes. The hyperinsulimia in type 2 diabetes leads to increased de novo lipogenesis20, while hepatic insulin resistance leads to a loss of suppression of VLDL production21 and insulin resistance in adipose tissue leads to a mild disruption of LPL function22. Differentiating between the origins of the triglyceride rich particles is partially possible by determining apoB100 levels. An apoB level > 0,75 g/L indicates a degree of fatty acid delivery to the liver and thus, at least to an extent, functional LPL. The TG:apoB ratio can thus distinguish between chylomicrons and VLDL. Sniderman et al. provided an comprehensive algorithm, where based on apoB100 and TG levels a clear direction is given towards the underlying genetic or secondary cause23.

Traditionally, primary causes of HTG can be divided into several disorders. Already in the 1960s, Fredrickson designed a system in which hyperlipidemias were divided based on their lipoprotein particle24. These phenotypes comprise both the rare monogenetic disorders, where complete loss of function mutations lead to severe elevations in plasma TGs >1000mg/dl (type 1)25 and 4 polygenic familial phenotypes (type II-IV)26.

The use of the Fredrickson Classification has been widely accepted since it’s adoption by the WHO in 197027, and still applies to the cases where (severe) hypertriglyceridemia is caused by a mutation leading to the loss of LPL function. In these cases, a loss of function of LPL is caused by either a loss of function mutation in the LPL gene or a mutation in genes coding for proteins directly involved in LPL activity. Known mutations have been extensively described previously in publications28 (especially see supplementary data) and25. These mutations result in severely high plasma triglyceride levels and, as illustrated above, additional low apoB100, low LDL and HDL levels that are resistant to current lipid-lowering therapies29, with strong restriction of dietary fat as the remaining treatment option30. Those cases are rare, with an estimated prevalence of 1-2:106,31. Even in a cohort of pre-selected patients with severe HTG from our hospital, only 54% was identified as having a rare, monogenetic variant as a cause of the HTG, whereas in 26% only common variants were found and no mutation in 21% of the patients. In total, 19 known and 16 novel disease causing mutations were indentified in the LPL, APOC2, APOA5 and GPIHBP125. However, cumulative evidence suggest that most of the different polygenic phenotypes IIB-V (phenotype 2A representing the familial hypercholesterolemia syndrome, often caused by heterozygous LDLR mutations32) share numerous

100

Chapter 6

common genetic determinants33–36. In these patients, lifestyle (dietary intervention) changes remain the initial treatment and although several drugs can be used to lower serum triglyceride levels9, little evidence exists on the effectiveness of the different drug classes on reducing cardiovascular risk2. Moreover, in these patients differences in genetic background are believed to be associated with the fasting plasma triglyceride response to pharmacological treatment37 and also with the response to lifestyle interventions38.

CURRENT TREATMENT A recent statement from the Endocrine Society recommends the combination of diet, exercise and drugs as the first line treatment of hypertriglyceridemia9. Although no solid data exists on the effect of fibrate (peroxisome proliferator activated receptor agonists, PPAR-alfa) treatment on cardiovascular risk-reduction, these compounds are to be used as first-line treatment for those patients at risk of developing triglyceride induced pancreatitis9. Furthermore treatment with niacin and n-3 fatty acids or any combination can be used, although the effect on cardiovascular risk reduction remains unclear2.

IMPLICATIONS OF GENETICS IN HYPERTRIGLYCERIDEMIA The use of genetic information in explaining and treating diverse disease modalities is ever expanding, varying from the use of classical approaches for monogenetic disorders, to the growing genome wide association studies (GWAS) in understanding more complex traits39. The clinical relevance of large scale GWAS studies remains a topic of wide discussion and investigation, where some suggest that most common variants found in such studies will be of little biological interest because of their small effect40. Still, in 2010 Teslovich et al. discovered 59 new loci associated with blood lipids contributing to 10-12% of the total variance in blood lipids41, amongst which some with clear biological and clinical importance. The relevance of the ever expanding available genomic information to our biological understanding of diseases seems quite clear, however, the challenge remains to translate this epidemiological-genetic information to pathophysiology and subsequently the clinic in order to provide new therapeutic opportunities for an individual patient.

The assessment of common genetic variations on plasma TG levels is complicated by the role of gene-environment interactions (GxE). Yet, this may provide the bridge between genetic risk and actual clinical disease in many cases, where the genetic variation only becomes an actual risk when certain environmental or lifestyle factors, such as diet, pass a certain threshold42. A special remark should be made to non-fasting triglyceride levels, as this is the state in which humans remain the majority of time. Little information is available on the post-prandial responses and its association with common variations in genes involved in TG metabolism. Of more concern, little validation reports have been published43–45. Based on the available data one might predict that the effect size of genetic variation in the prediction of abnormal postprandial TG excursions will be small. More


101

6

research in this field is warranted to adapt approaches specifically to alterations in the post-prandial state.

GENETIC APPROACHES IN MONOGENETIC DISORDERS When thinking of using genetic information for therapeutic purposes, several approaches can be considered, depending on the nature of the genetic defect. As described previously, in the case of hypertriglyceridemia due to monogenetic loss of function of LPL, current therapies are deemed to be ineffective. Treatment for these subjects would obviously be found with the successful implementation of gene therapy: introducing a competent gene to supplement the dysfunctional one. This is not a novel concept, a symposium took place on May 26 1966, at Columbia University College of Physicians and Surgeons in New York City, entitled “Reflections on Research and the Future of Medicine”46. Here Edward Tanner already mentions the use of genetic engineering, possibly using viruses, a contribution he later published separately47. Though there have also been many setbacks since the first clinical trial in 199048, advances in this field have been great. In the case of HTG based on an LPL deficiency, an AAV 1LPLs447x vector has successfully been introduced in murine49 and feline50 models. In humans, the LPLs447x variant is associated with lower serum TGs, higher HDL-C levels and lower incidence of cardiovascualar disease, as well as enhanced ApoB48 clearance51. Subsequently, AAV 1LPLs447x genetherapy, using a plasmid-based52 or a baculovirus-based production53, was administrated in LPL deficient subjects with resident LPL mass. This therapy was shown to be well tolerated and (moderately) effective with regard to plasma triglyceride levels. However, even though the advances on biological level in this field are great, these invasive, therapies lack longterm beneficial clinical effects. Although self-reported, clinical improvement has been shown up to 2 years in subjects, the TG-lowering effect so far has not been reported after 12 weeks52,53. A longer lasting improvement in postprandial chylomicron metabolism has been shown, but small number of subjects makes it difficult to link these results to clinical parameters54. Also, it should be noted that LPL gene replacement is only effective for complete null mutations in LPL and can not be used for patients with mutations in other genes causing complete loss of function of LPL, such as APOC2 LMF1, APOA5 or GPIHBP155, and thus is only suitable for a very small group of patients.

With respect to gene therapy focusing on other causative genes, preliminary data show that ApoA5 null mice have elevated levels of triglycerides and that treatment with a sense adenoviral ApoA5 construct can rescue the hypertriglyceridemic phenotype56. Gene transfer of ApoA5 lead to a marked decrease of 50% in plasma triglycerides57. However, the mechanism by which ApoAV affects human triglyceride metabolism remains to be elucidated. SNPs in the APOA5 have been linked to hypertriglyceridemia26,34,58, but controversy remains over the association between ApoAV levels and plasma triglycerides, which have rendered both positive59 and inverse60 correlations. Interestingly, in type 2 diabetes a positive correlation was found between apoAV and plasma triglyceride levels, whereas atorvastatin treatment was actually able to reduce both plasma ApoAV levels and

102

Chapter 6

plasma triglycerides61. Moreover, analysis in the EPIC-Norfolk cohort showed a significant positive correlation62. The differences in these findings may be the direct result of gene-environment interactions, as was shown particularly for carriers of the APOA5 -1131T>C SNP, who have a strong inverse correlation between fat intake and TG levels as opposed to major allele carriers38. Thus, until the mechanism by which ApoAV influences human triglyceride metabolism has been further investigated, genetic approaches aiming at altering plasma triglyceride levels seem an unlikely option.

Finally, genetics can also be used to indirectly influence expression of a gene. This can be conceived either by using antisense oligonucleotides (ASOs) or RNAi (RNA interference)63. One of the targets pursued is ApoCIII, which is thought to delay clearance of TG-rich lipoproteins by inhibiting LPL. Heterozygous carriers of a null mutation in APOC3 a favorable lipid profile and lower subclinical atherosclerosis in a human population64. In mice, overexpression of ApoC3 indeed leads to increased levels of plasma triglycerides65, whereas silencing leads to a marked decrease66. Interestingly, a recent paper from our group demonstrated a link between GALNT2 mutations and posttranslational modification of ApoCIII, thereby improving LPL mediated lipolysis67. In line with these findings, preclinical data suggest that an APOC3 ASO can be used safely to reduce triglyceride levels in mice68. Notably, Graham et al. revealed that a second generation APOC3 ASO was capable of lowering plasma TG and VLDL levels in mice and non-human primates. Interestingly, in healthy human subjects APOC3 ASO administration was well-tolerated and showed a dose-dependent apoCIII reductions and a signficant lowering of TG levels69. It should be noted that normal LPL function is necessary when aiming at apoCIII reduction. In the case of chylomicron production, the intestinal cell would be the most likely target. Diaglycerol acyltransferase (DGAT) is responsible for the final catalyzation in triglyceride synthesis. In mice, DGAT1 inhibiton by oral gavage of DGAT1i (a selective inhibitor of human DGAT170) as well as DGAT1 knockout, will inhibit chylomicron formation and thus lower postprandial triglyceride levels71. Recently, the DGAT1 inhibitor LCQ908 was assessed in a very small group of LPL deficient patients, showing decreases in mean fasting triglycerides after 3 weeks of treatment72. Currently a randomized, double-blind, placebo trial is running to further evaluate these findings.

INDIvIDUALIZED THERAPY IN POLYGENETIC DISORDERS Using genetic information in the treatment of polygenetic hypertriglyceridemia seems much more complex. In this respect, two major approaches can be distinguished: Pharmacogenetics and Genetic Risk Prediction. The genetic basis of inter-individual differences in response to pharmaceutical treatment is a field of broad interest73,74. Dating back to the 1950s, when a role for genetics in adverse drug reactions was first suggested75. Brisson et al. showed that carriers of a polymorphism in APOE (ApoE2, leading to a dysfunctional ApoE protein) respond more pronounced to treatment with fibrates than non-carriers. Although the ApoE2 allele is mainly associated with type III hypertriglyceridemia, these findings were also reproduced in ApoE2 carriers of subtypes IIB, IV and V, thus suggesting that in these cases response to therapy could be dependent on the genetic


103

6

architecture29. Yet, the debated position of triglycerides in the prevention of cardiovascular disease and the lack of therapeutic options make it difficult to implicate that pharmacogenetic findings may lead to identification of novel pathways and possibly the development of more target-specific pharmaceutical products.

Finally, when thinking of individualized medicine, genetic risk scoring has to be evaluated. The concept of using genetic information to further enhance current risk scores is a field of broad discussion with up to date no certainty about it’s feasibility. The utility of genetic risk scores in cardiovascular diseases has been discussed elaborately and is excellently summarized by Thannasoulis and Vassan76. These authors propose that, at the current time, genetics are not ready to be used in risk prediction for cardiovascular diseases. This problem is partially driven by the fact that current major risk factors already may explain the major burden of CVD77,78 and that most likely, the risk factor ‘family history of CVD’ represents the net effect of many common risk variants76. Another difficulty when using combined genetic loci to explain cardiovascular disease is the notion that there may be additional effects of found variants on other causal pathways41.

In the case of TG level associated loci, another problem emerges. The extent to which elevated triglyceride levels are an active partaker in cardiovascular disease or merely ‘innocent bystanders’ still remains a topic of wide discussion. Analysis of current available data failed to identify a causal role of triglycerides in CVD79 and it has been argued that after adjustment for HDL levels, triglycerides do not significantly contribute to CVD risk80.Thus, the debated causal role of triglycerides in cardiovascular disease in conjunction with the lack of specific genetic risk factors minimize a possible role for TG specific genes in cardiovascular risk management. Moreover, it has been shown that the discrimination in genetic risk score between the high-TG population and the normal population is small though significant, and subsequently risk allele scores are currently mainly able to discriminate for extreme values31. For example, in hyperlipidemia, and more specific for severe hypertriglyceridemia, Teslovich et al. did show that subjects with an allelic dosage score (a score summarizing the number of TG raising alleles weighted by effect size) in the top quartile have a 44 fold increased likelihood to be hypertriglyceridemic41. This extreme increase, especially when compared to the likelihood for LDL-C and HDL-c (respectively 13 and 4), should however be interpreted with caution. As both TG levels and the odds ratio itself are widely spread, without further phenotype characterization of the subjects the clinical implication of this epidemiological-genetic finding seems limited.

CONCLUSIONS The ever-expanding genomic information has led to an explosion of novel approaches to unravel the pathophysiology of hypertriglyceridemia. We have highlighted the use of gene therapy, or much more pathway specific treatments in the case of monogenetic disease as well as the reclassification, novel approaches to risk assessment and treatment of polygenetic disease. Although in the case

104

Chapter 6

of monogenetic disease, these novel therapies are scarce, they do seem promising and approval of these therapies would mean an enormous improvement over the current options in treating monogenetic severe hypertriglyceridemia. In the polygenetic diseases, the future is less clear. Although evidence has been found that genetic differences can influence personal drug response, a translation to the use of this information to clinical practice is not yet available. Finally, the evidence for using genetics in cardiovascular risk assessment is too limited to allow application in current clinical practice.

At this time, the importance of genetic information in therapeutics for hypertriglyceridemia can be balanced between unravelling and understanding of pathophysiological pathways underlying this disease, redefining disease categories and to predict inter-individual response-differences to therapies.


105

6

REFERENCE LIST1. Austin MA, Hokanson JE, Edwards KL. Hypertriglyceridemia as a cardiovascular risk factor. Am. J.

Cardiol. 1998;81:7B–12B.

2. Miller M, Stone NJ, Ballantyne C, et al. Triglycerides and cardiovascular disease: a scientific state-ment from the American Heart Association. Circulation. 2011;123:2292–333.

3. Nordestgaard BG, Benn M, Schnohr P, Tybjaerg-Hansen A. Nonfasting triglycerides and risk of myo-cardial infarction, ischemic heart disease, and death in men and women. JAMA. 2007;298:299–308.

4. Nordestgaard BG, Freiberg JJ. Clinical relevance of non-fasting and postprandial hypertriglyceridemia and remnant cholesterol. Curr Vasc Pharmacol. 2011;9:281–6.

5. Zilversmit DB. Atherogenic nature of triglycerides, postprandial lipidemia, and triglyceride-rich rem-nant lipoproteins. Clin. Chem. 1995;41:153–8.

6. Benlian P, De Gennes JL, Foubert L, Zhang H, Gagné SE, Hayden M. Premature atherosclerosis in patients with familial chylomicronemia caused by mutations in the lipoprotein lipase gene. N. Engl. J. Med. 1996;335:848–54.

7. Nordestgaard BG, Abildgaard S, Wittrup HH, Steffensen R, Jensen G, Tybjaerg-Hansen A. Heterozy-gous lipoprotein lipase deficiency: frequency in the general population, effect on plasma lipid levels, and risk of ischemic heart disease. Circulation. 1997;96:1737–44.

8. Cameron JL, Capuzzi DM, Zuidema GD, Margolis S. Acute pancreatitis with hyperlipemia: the inci-dence of lipid abnormalities in acute pancreatitis. Ann. Surg. 1973;177:483–9.

9. Berglund L, Brunzell JD, Goldberg AC, et al. Evaluation and treatment of hypertriglyceridemia: an endocrine society clinical practice guideline. J. Clin. Endocrinol. Metab. 2012;97:2969–89.

10. Ros E. Intestinal absorption of triglyceride and cholesterol. Dietary and pharmacological inhibition to reduce cardiovascular risk. Atherosclerosis. 2000;151:357–79.

11. Iqbal J, Hussain MM. Intestinal lipid absorption. Am. J. Physiol. Endocrinol. Metab. 2009;296:E1183–94.

12. Olivecrona G, Beisiegel U. Lipid binding of apolipoprotein CII is required for stimulation of lipoprotein lipase activity against apolipoprotein CII-deficient chylomicrons. Arterioscler. Thromb. Vasc. Biol. 1997;17:1545–9.

13. Nilsson SK, Heeren J, Olivecrona G, Merkel M. Apolipoprotein A-V; a potent triglyceride reducer. Atherosclerosis. 2011;219:15–21.

14. Goldberg IJ, Scheraldi CA, Yacoub LK, Saxena U, Bisgaier CL. Lipoprotein ApoC-II activation of lipopro-tein lipase. Modulation by apolipoprotein A-IV. J. Biol. Chem. 1990;265:4266–72.

15. Péterfy M, Ben-Zeev O, Mao HZ, et al. Mutations in LMF1 cause combined lipase deficiency and severe hypertriglyceridemia. Nat. Genet. 2007;39:1483–7.

16. Young SG, Davies BSJ, Voss C V, et al. GPIHBP1, an endothelial cell transporter for lipoprotein lipase. J. Lipid Res. 2011;52:1869–84.

17. Hassing HC, Surendran RP, Mooij HL, Stroes ES, Nieuwdorp M, Dallinga-Thie GM. Pathophysiology of hypertriglyceridemia. Biochim. Biophys. Acta. 2012;1821:826–32.

18. Williams KJ. Molecular processes that handle -- and mishandle -- dietary lipids. J. Clin. Invest. 2008;118:3247–59.

106

Chapter 6

19. Ewald N, Hardt PD, Kloer H-U. Severe hypertriglyceridemia and pancreatitis: presentation and man-agement. Curr. Opin. Lipidol. 2009;20:497–504.

20. Koopmans SJ, Kushwaha RS, DeFronzo RA. Chronic physiologic hyperinsulinemia impairs suppres-sion of plasma free fatty acids and increases de novo lipogenesis but does not cause dyslipidemia in conscious normal rats. Metab. Clin. Exp. 1999;48:330–7.

21. Sparks JD, Sparks CE, Adeli K. Selective hepatic insulin resistance, VLDL overproduction, and hyper-triglyceridemia. Arterioscler. Thromb. Vasc. Biol. 2012;32:2104–12.

22. Preiss-Landl K, Zimmermann R, Hämmerle G, Zechner R. Lipoprotein lipase: the regulation of tissue specific expression and its role in lipid and energy metabolism. Curr. Opin. Lipidol. 2002;13:471–81.

23. Sniderman A, Couture P, De Graaf J. Diagnosis and treatment of apolipoprotein B dyslipoproteine-mias. Nat Rev Endocrinol. Nature Publishing Group; 2010;6:335–46.

24. Fredrickson DS, Lees RS. Editorial: A System for Phenotyping Hyperlipoproteinemia. Circulation. 1965;31:321–7.

25. Surendran RP, Visser ME, Heemelaar S, et al. Mutations in LPL, APOC2, APOA5, GPIHBP1 and LMF1 in patients with severe hypertriglyceridaemia. J. Intern. Med. 2012;272:185–96.

26. Johansen CT, Wang J, Lanktree MB, et al. An increased burden of common and rare lipid-associated risk alleles contributes to the phenotypic spectrum of hypertriglyceridemia. Arterioscler. Thromb. Vasc. Biol. 2011;31:1916–26.

27. Lewis B. Classification of hyperlipidaemias. Proc. R. Soc. Med. 1971;64:905–8.

28. Merkel M, Eckel RH, Goldberg IJ. Lipoprotein lipase: genetics, lipid uptake, and regulation. J. Lipid Res. 2002;43:1997–2006.

29. Brisson D, Méthot J, Tremblay K, Tremblay M, Perron P, Gaudet D. Comparison of the efficacy of fibrates on hypertriglyceridemic phenotypes with different genetic and clinical characteristics. Phar-macogenet. Genomics. 2010;20:742–7.

30. Gotoda T, Shirai K, Ohta T, et al. Diagnosis and management of type I and type V hyperlipoproteine-mia. J. Atheroscler. Thromb. 2012;19:1–12.

31. Johansen CT, Hegele RA. Genetic bases of hypertriglyceridemic phenotypes. Curr. Opin. Lipidol. 2011;22:247–53.

32. Rader DJ, Cohen J, Hobbs HH. Monogenic hypercholesterolemia: new insights in pathogenesis and treatment. J. Clin. Invest. 2003;111:1795–803.

33. Wang J, Cao H, Ban MR, et al. Resequencing genomic DNA of patients with severe hypertriglyceri-demia (MIM 144650). Arterioscler. Thromb. Vasc. Biol. 2007;27:2450–5.

34. Wang J, Ban MR, Kennedy BA, et al. APOA5 genetic variants are markers for classic hyperlipopro-teinemia phenotypes and hypertriglyceridemia. Nat Clin Pract Cardiovasc Med. 2008;5:730–7.

35. Wang J, Ban MR, Zou GY, et al. Polygenic determinants of severe hypertriglyceridemia. Hum. Mol. Genet. 2008;17:2894–9.

36. Hegele RA, Ban MR, Hsueh N, et al. A polygenic basis for four classical Fredrickson hyperlipoproteine-mia phenotypes that are characterized by hypertriglyceridemia. Hum. Mol. Genet. 2009;18:4189–94.


107

6

37. Smith JA, Arnett DK, Kelly RJ, et al. The genetic architecture of fasting plasma triglyceride response to fenofibrate treatment. Eur. J. Hum. Genet. 2008;16:603–13.

38. Sánchez-Moreno C, Ordovás JM, Smith CE, Baraza JC, Lee Y-C, Garaulet M. APOA5 gene variation interacts with dietary fat intake to modulate obesity and circulating triglycerides in a Mediterranean population. J. Nutr. 2011;141:380–5.

39. Hegele RA. Plasma lipoproteins: genetic influences and clinical implications. Nat. Rev. Genet. 2009;10:109–21.

40. Goldstein DB. Common genetic variation and human traits. N. Engl. J. Med. 2009;360:1696–8.

41. Teslovich TM, Musunuru K, Smith A V, et al. Biological, clinical and population relevance of 95 loci for blood lipids. Nature. 2010;466:707–13.

42. Lee Y-C, Lai C-Q, Ordovas JM, Parnell LD. A Database of Gene-Environment Interactions Pertaining to Blood Lipid Traits, Cardiovascular Disease and Type 2 Diabetes. J Data Mining Genomics Proteomics. 2011;2.

43. Delgado-Lista J, Perez-Jimenez F, Ruano J, et al. Effects of variations in the APOA1/C3/A4/A5 gene cluster on different parameters of postprandial lipid metabolism in healthy young men. J. Lipid Res. 2010;51:63–73.

44. Perez-Martinez P, Garcia-Rios A, Delgado-Lista J, Perez-Jimenez F, Lopez-Miranda J. Nutrigenetics of the postprandial lipoprotein metabolism: evidences from human intervention studies. Curr Vasc Pharmacol. 2011;9:287–91.

45. Jackson KG, Poppitt SD, Minihane AM. Postprandial lipemia and cardiovascular disease risk: Interre-lationships between dietary, physiological and genetic determinants. Atherosclerosis. 2012;220:22–33.

46. Campbell TL. Reflections on research and the future of medicine. Science. 1966;153:442–9.

47. Tatum EL. Molecular biology, nucleic acids, and the future of medicine. Perspect. Biol. Med. 1966;10:19–32.

48. Blaese RM, Culver KW, Miller AD, et al. T lymphocyte-directed gene therapy for ADA- SCID: initial trial results after 4 years. Science. 1995;270:475–80.

49. Nierman MC, Rip J, Twisk J, et al. Gene therapy for genetic lipoprotein lipase deficiency: from prom-ise to practice. Neth J Med. 2005;63:14–9.

50. Ross CJD, Twisk J, Bakker AC, et al. Correction of feline lipoprotein lipase deficiency with adeno-asso-ciated virus serotype 1-mediated gene transfer of the lipoprotein lipase S447X beneficial mutation. Hum. Gene Ther. 2006;17:487–99.

51. Nierman MC, Rip J, Kuivenhoven JA, et al. Enhanced apoB48 metabolism in lipoprotein lipase X447 homozygotes. Atherosclerosis. 2007;194:446–51.

52. Stroes ES, Nierman MC, Meulenberg JJ, et al. Intramuscular administration of AAV1-lipoprotein li-pase S447X lowers triglycerides in lipoprotein lipase-deficient patients. Arterioscler. Thromb. Vasc. Biol. 2008;28:2303–4.

53. Gaudet D, Méthot J, Déry S, et al. Efficacy and long-term safety of alipogene tiparvovec (AAV1-LPL(S447X)) gene therapy for lipoprotein lipase deficiency: an open-label trial. Gene Ther. 2012;1–9.

108

Chapter 6

54. Carpentier AC, Frisch F, Labbé SM, et al. Effect of alipogene tiparvovec (AAV1-LPL(S447X)) on post-prandial chylomicron metabolism in lipoprotein lipase-deficient patients. J. Clin. Endocrinol. Metab. 2012;97:1635–44.

55. Gaudet D, Méthot J, Kastelein J. Gene therapy for lipoprotein lipase deficiency. Curr. Opin. Lipidol. 2012;23:310–20.

56. Van der Vliet HN, Schaap FG, Levels JHM, et al. Adenoviral overexpression of apolipoprotein A-V reduces serum levels of triglycerides and cholesterol in mice. Biochem. Biophys. Res. Commun. 2002;295:1156–9.

57. Sharma V, Beckstead JA, Simonsen JB, et al. Gene Transfer of Apolipoprotein A-V Improves the Hy-pertriglyceridemic Phenotype of apoa5 (-/-) mice. Arterioscler. Thromb. Vasc. Biol. 2013;33:474–80.

58. Garelnabi M, Lor K, Jin J, Chai F, Santanam N. The paradox of ApoA5 modulation of triglycerides: Evidence from clinical and basic research. Clin. Biochem. 2013;46:12–9.

59. Schaap FG, Nierman MC, Berbée JFP, et al. Evidence for a complex relationship between apoA-V and apoC-III in patients with severe hypertriglyceridemia. J. Lipid Res. 2006;47:2333–9.

60. Hyun YJ, Jang Y, Chae JS, et al. Association of apolipoprotein A5 concentration with serum insulin and triglyceride levels and coronary artery disease in Korean men. Atherosclerosis. 2009;205:568–73.

61. Dallinga-Thie GM, Van Tol A, Hattori H, Van Vark-van der Zee LC, Jansen H, Sijbrands EJG. Plasma apolipoprotein A5 and triglycerides in type 2 diabetes. Diabetologia. 2006;49:1505–11.

62. Vaessen SFC, Schaap FG, Kuivenhoven J-A, et al. Apolipoprotein A-V, triglycerides and risk of coro-nary artery disease: the prospective Epic-Norfolk Population Study. J. Lipid Res. 2006;47:2064–70.

63. Kassim SH, Wilson JM, Rader DJ. Gene therapy for dyslipidemia: a review of gene replacement and gene inhibition strategies. Clin Lipidol. 2010;5:793–809.

64. Pollin TI, Damcott CM, Shen H, et al. A null mutation in human APOC3 confers a favorable plasma lipid profile and apparent cardioprotection. Science. 2008;322:1702–5.

65. Ito Y, Azrolan N, O’Connell A, Walsh A, Breslow JL. Hypertriglyceridemia as a result of human apo CIII gene expression in transgenic mice. Science. 1990;249:790–3.

66. Maeda N, Li H, Lee D, Oliver P, Quarfordt SH, Osada J. Targeted disruption of the apolipoprotein C-III gene in mice results in hypotriglyceridemia and protection from postprandial hypertriglyceridemia. J. Biol. Chem. 1994;269:23610–6.

67. Holleboom AG, Karlsson H, Lin R-S, et al. Heterozygosity for a loss-of-function mutation in GALNT2 improves plasma triglyceride clearance in man. Cell Metab. 2011;14:811–8.

68. Williams AJK, Thrower SL, Sequeiros IM, et al. Pancreatic volume is reduced in adult patients with recently diagnosed type 1 diabetes. J. Clin. Endocrinol. Metab. 2012;97:E2109–13.

69. Graham MJ, Lee RG, Bell TA, et al. Antisense Oligonucleotide Inhibition of Apolipoprotein C-III Re-duces Plasma Triglycerides in Rodents, Nonhuman Primates, and Humans. Circ Res. 2013;

70. Birch AM, Birtles S, Buckett LK, et al. Discovery of a potent, selective, and orally efficacious py-rimidinooxazinyl bicyclooctaneacetic acid diacylglycerol acyltransferase-1 inhibitor. J. Med. Chem. 2009;52:1558–68.

71. Ables GP, Yang KJZ, Vogel S, et al. Intestinal DGAT1 deficiency reduces postprandial triglyceride and retinyl ester excursions by inhibiting chylomicron secretion and delaying gastric emptying. J. Lipid Res. 2012;53:2364–79.


109

6

72. Meyers, Charles; Gaudet Daniel; Tremblay K, Amer, Ahmed; Chen, Jin; Aimin F. The DGAT1 Inhibitor LCQ908 Decreases Triglyceride Levels in Patients with the Familial Chylomicronemia Syndrome. J Clin Lipidol. 2012;6:266–7.

73. Marshall A. Laying the foundations for personalized medicines. Nat. Biotechnol. 1997;15:954–7.

74. Evans WE, Relling M V. Pharmacogenomics: translating functional genomics into rational therapeu-tics. Science. 1999;286:487–91.

75. Motulsky AG. Drug reactions enzymes, and biochemical genetics. J Am Med Assoc. 1957;165:835–7.

76. Thanassoulis G, Vasan RS. Genetic cardiovascular risk prediction: will we get there? Circulation. 2010;122:2323–34.

77. Khot UN, Khot MB, Bajzer CT, et al. Prevalence of conventional risk factors in patients with coronary heart disease. JAMA. 2003;290:898–904.

78. Yusuf S, Hawken S, Ounpuu S, et al. Effect of potentially modifiable risk factors associated with myo-cardial infarction in 52 countries (the INTERHEART study): case-control study. Lancet. 2004;364:937–52.

79. Lewington S, Bragg F, Clarke R. A review on metaanalysis of biomarkers: promises and pitfalls. Clin. Chem. 2012;58:1192–204.

80. Di Angelantonio E, Sarwar N, Perry P, et al. Major lipids, apolipoproteins, and risk of vascular disease. JAMA. 2009;302:1993–2000.

II PART II

ACqUIRED AND INBORN ERRORS OF HEPARAN SULFATES IN HYPERGLYCAEMIA AND HYPERTRIGLYCERIDEMIA

Mooij HL, Hassing HC, Bernelot Moens SJ, Esko JD, Tanck MWT, Stroes ESG, Dallinga-Thie GM, Nieuwdorp M


7 GENETIC vARIATIONS IN HEPARAN SULFATES ONLY MODESTLY AFFECT POSTPRANDIAL TRIGLYCERIDE CLEARANCE IN HUMANS

114

Chapter 7

ABSTRACT introduction - Elevated postprandial triglyceride rich lipoprotein (TG) remnants represent a risk factor for cardiovascular disease. Murine studies have indicated that both the hepatic LDL receptor as well as heparan sulfate proteoglycans (HSPGs) are involved in hepatic remnant clearance. To substantiate the relevance of these pathways in human TG homeostasis we evaluated TG clearance in hereditary multiple exostosis (HME) patients with a heterozygous mutation in exostosin (EXT) as well as patients with heterozygous Familial Hypercholesterolemia (FH) selected for genetic variants in HSPG-related genes.

Methods - Study population comprised of HME Subjects and matched unaffected controls (n=10 per group) and FH patients (n=16), who were screened for tagSNPs in HSPGs encoding genes potentially affecting plasma triglycerides (TG). Correlation coefficients were calculated for each SNP to generate a HSPG gene score. FH subjects with highest and lowest gene risk score (n=11 per group) were selected. All subjects underwent an oral fat tolerance test to study postprandial TG metabolism. Vitamin A was added to the fat bolus to allow for estimation of dietary lipid clearance.

Results - No differences in fasting TG (HME 0.9 [0.7 – 1.4 mmol/l] vs controls 0.8 [0.7-1.0] mmol/l, ns) and postprandial TG clearance (iAUC-TG HME: 4.29 ± 0.84 versus controls 3.56 ± 0.94 mmol/l.h-1; ns) were observed in the HME patients. For FH, no difference was observed between the highest versus the lowest HSPG gene score, neither for fasting plasma TG (1.3 [1.0 -1.9] and 1.1 [0.79 – 2.2] mmol/l; ns), nor for postprandial TG clearance (iAUC-TG: 7.02 ± 1.16 and 6.92 ± 1.66 mmol/l.h-1; ns) In contrast, retinyl palmitate excursions were significant higher in FH subjects with highest gene score (iAUC 12.77 [0.790-20.07] and 24.86 [12.48-28.42] ng.L-1.h-1 respectively (P=0.04).

Conclusion - Whereas postprandial TG clearance is delayed in subjects with LDLR mutations, genetic variation in HSPG genes offer at best a minor contribution. The discrepancy between human and murine data underscores the lack of translational relevance of murine studies for understanding TG homeostasis.

Heparan sulfates and triglyceride Handling in Humans

115

7

INTRODUCTIONHypertriglyceridemia affects 10–20% of the population in Western countries and increases the risk of coronary artery disease 1. The etiology of hypertriglyceridemia is complex, and current strategies to lower TG levels are limited. Hence, the need for identification and characterization of factors affecting the synthesis, transport and clearance of plasma triglyceride-rich lipoproteins (TRLs) is still on. Although plasma TG levels have usually been measured after an overnight fast, observational studies have identified nonfasting TG levels to be a superior predictor of CVD risk compared to fasting levels 1-5. Thus postprandial TG levels have been suggested to saturate hepatic clearance pathways thereby allowing for their uptake by macrophages 6. The most important players involved in hepatic uptake of TRL-remnants are the LDL receptor and the LDL Receptor Related Protein 1 (LRP1) 7. However, as previous studies have reported that heterozygous and even homozygous FH patients with deleterious mutation in the LDL receptor are characterized by modestly elevated fasting as well as postprandial remnant concentrations, other pathways have been hypothesized to be involved 7.

Heparan sulfate proteoglycans (HSPGs) have been identified as a class of receptors that mediate the clearance of triglyceride-rich lipoproteins in the liver in conjunction with the LDL receptor and the LRP1 receptors 8. Heparan sulfate proteoglycans (HSPGs) are expressed by endothelial cells lining the hepatic sinusoids, including the perisinusoidal space (space of Disse) and the underlying hepatocytes 9. These membrane-bound hepatic HSPGs consist of a core protein (predominantly syndecan1) with heparan sulfate chains attached to it. Several intracellular genes such as EXT1, EXT2, NDST or HS2ST1 are involved in the synthesis of the sulfate chains 10,11. The in vivo role of HSPGs in remnant clearance was first illustrated by injecting heparinase, a heparan sulfate degrading enzyme, into the murine portal vein, which resulted in decreased VLDL and chylomicron remnant clearance 12. In line, syndecan1 was shown to bind VLDL remnant particles, allowing for endocytosis and subsequent hepatic degradation 9,13. These findings were corroborated in mouse models with specific hepatic mutations in HSPG associated genes, which were invariably associated with delayed postprandial TG excursions. Collectively, this has validated the concept that syndecan1 and specific HSPG-sulfation patterns are critical determinants for hepatic TG-rich remnant clearance 9,14,15.

Yet, to date, human studies are lacking. In patients with Multiple Hereditary Exostosis (HME) the HS synthesis is severely disrupted leading to, amongst other growth of multiple bony tumors (eg. Exostoses or osteochondromas) 16. Heterozygous mutations in EXT1 and EXT2 are known to be involved in the development of HME 17. Metabolic derangements have, however, not been studied in these patients 18.

In the present study we set out to evaluate the role of HSPGs in postprandial TG clearance in subjects with a single loss-of-function mutation in EXT1 or EXT2. In order to address a potential additive effect of HSPG on the LDL receptor in human postprandial TG metabolism, we also studied the effect

116

Chapter 7

of multiple tagSNPs in HSPG genes, using a HSPG gene score, on postprandial lipid metabolism in subjects with heterozygous familial hypercholesterolemia (FH).

METHODSstudy population We enrolled subjects with clinically established diagnosis of HME 18 as well as unaffected, healthy family controls, who were recruited via the website of the Dutch HME foundation (www.hme-mo.nl). Exclusion criteria were current pregnancy, malignancy with limited lifespan and/or overt cardiovascular disease. If subjects were using any medication, they were asked to stop cardiovascular medication ATII/ACE inhibitors (5 days) and statin (4 weeks) before the lipid load.

Moreover, we used a cohort (GiRaFH) 19 including DNA of n=2345 FH patients characterized with a heterozygous mutation from 27 lipid clinics throughout the Netherlands. The FH diagnostic criteria were based on internationally established criteria 20. For 1600 FH patients we had the complete plasma lipid profiles and was DNA available for SNP genotyping. Exclusion criteria’s for participation in the postprandial sub-study were the use of any lipid lowering medication in 4 weeks preceding the fat tolerance test, clinical signs of malabsorption (eg diarrhoea) or exogenous insulin treatment. Both studies were approved by the Institutional Review Board and conducted at the Academic Medical Center Amsterdam in accordance with the Declaration of Helsinki (updated version 2008). Each patient provided written informed consent.

Oral Fat LoadParticipants were asked to refrain from alcohol intake the day before. Participants were admitted at 7:30 am after an overnight fast. Cream [consisting of 40% fat (wt/vol) with a polyunsaturated fat to saturated fat ratio of 0.06, 0.001% cholesterol (wt/vol), was administered in a dose of 35 gram fat per m2 body surface. The cream drink, supplemented 60.000 IU/m2 body surface of vitamin A (retinyl palmitate (RP) Department of Clinical Pharmacy, AMC) for specific labeling dietary derived lipids, was consumed within 2 minutes. Pre- and postprandial blood samples were drawn at 0, 2, 3, 4, 5, 6 and 8 hours. Subsequently, the catheter was removed and the subject was discharged. Venous blood was collected into EDTA containing tubes, which were placed on ice and protected from light. Plasma was separated within 30 minutes by centrifugation at 3000 rpm for 20 min. at 4°C. Aliquots of plasma were protected from light and frozen at -800C for subsequent analysis of TG and retinyl palmitate (RP). At the end of the fat load an intravenous injection of heparin (MW 6500; Leo Pharma; 50 IU/kg body weight) was given and a blood samples was taken after 15 minutes to analyse heparin-releasable LPL capacity by measuring the decrease in plasma TG.

Biochemical analysesTotal cholesterol, HDL-cholesterol, LDL-cholesterol, and TG were measured by standard enzymatic methods (Roche Diagnostics, Basel, Switzerland) on a COBAS MIRA automated spectrophotometric


117

7

analyzer (Roche Diagnostica, Basel, Switzerland). ApoB was analysed with a turbidimetric assay from Randox on a Cobas Mira autoanalyzer. Glucose was assessed using the hexokinase method (Gluco-quant, Hitachi 917; Hitachi). Plasma insulin was measured by an immunoluminimetric assay (Immulite insuline) on Immulite 2000 (Diagnostic Products). HbA1C was measured by HPLC (Reagens Bio-Rad Laboratories, Veenendaal, the Netherlands) on a Variant II (Bio-Rad Laboratories). Plasma retinyl palmitate was analyzed using reversed phase HPLC in 200 μl plasma after extraction of retinyl esters using chloroform/methanol/water as described 21,22. In short, retinyl propionate (Sigma Chemicals; St Louis, USA) was used as internal standard; methanol was used as mobile phase at a flow rate of 1 ml/min and the effluent is monitored at 330 nm. A standard curve of retinyl palmitate in pooled plasma was used as reference. Peak heights were measured and used for calculations of the absolute RE values.

selection of tagsNPsWe selected tagSNPs in genes involved in HSPG biosynthesis such as NDST, EXT1, EXT2, SDC1, HS2ST1 (see Table 2). Tagging SNPs were selected using the HAPMAP database and the TAGGER algorithm as presented earlier 23. We limited our search to SNPs with a minor allele frequency > 5%, using a pairwise tagging approach with an r2 cutoff level > 0.8. Genotyping of SNPs was carried out on an ABI 7900 system, using Assay by DesignTM assays (Applied Biosystems, Foster City, CA, USA). Allelic discrimination was performed using either FAM or VIC as fluorophore. PCR conditions were denaturation for 10 min at 95°C, followed by 40 cycles (30 sec 92°C, 45 sec 60°C). PCR assay mix was obtained from Applied Biosystems. All SNPs were in Hardy Weinberg equilibrium.

statistical analysisWe applied extreme phenotype sampling to obtain the two groups of GiRaFH patients subjected to the oral fat load test. All GiRaFH patients were genotyped for the selected tagSNPs (see above). Individual effects of these tagSNPs on (log transformed) triglyceride levels were tested using ANOVA. Subsequently, these tagSNPs were used as predictor variables in a multiple linear regression model with the observed (log-transformed) triglyceride levels as outcome. Individual triglyceride levels were predicted based on this model (i.e. a weighted genetic risk score). Patients from the upper and lower 10th percentile of these predicted triglyceride levels (i.e. low or high HSPG gene score) were selected to participate in the fat load test. Data are presented as mean ± SD or medians with interquartile range [IQR] unless stated otherwise. When normally distributed, baseline characteristics were compared using a student’s t test. Differences in TG levels and continuous outcome variables were assessed using the nonparametric Mann-Whitney U test. Log tranformed TG levels were used in the patient selection process (see above). Postprandial TG and RE were calculated as total area under the curve (AUC) calculated by the trapezoid rule. Incremental area under the curve (iAUC) was obtained by subtracting the fasting plasma level from each postprandial time point. Two-sided probability values of less than 0.05 were considered statistical significant. Statistical analyses were performed using SPSS version 18.0.

118

Chapter 7

RESULTSPatients with hereditary multiple exostosis (HME)Patients with HME have a known loss of function mutation in either EXT1 or EXT2 resulting in decreased HSPG and may serve as a model to show the association between heterozygous HSPG mutations and postprandial lipid handling. We selected 10 patients with HME and 10 gender-age matched controls (Table 1). HME was established by identification of the heterozygous mutation in EXT1 or EXT2. The patients and controls had normal BMI. Plasma glucose and insulin levels in HME patients were normal and not different from controls. Fasting plasma cholesterol, HDLc, and LDLc were within the normal range and no difference was found between HME patients and controls. Fasting plasma TG levels were not significantly different between patients and controls (HME: 0.82 [0.74 - 1.12] mmol/l and controls: 0.78 [0.65-1.04] mmol/l, ns). Similarly, HME subjects had comparable postprandial TG excursions (see Figure 1A). The incremental area under the curve (iAUC) in HME patients was 3.59 [2.0 – 5.4] and in controls 2.72 [1.9 – 4.3] mmol/l.h-1 (Figure 1b). No difference in heparin releasable LPL capacity was found.

Table 1 - baseline characteristics of the HME patients and the matched controls

Non carriers Carriers P-value

NAge (years)Gender (% male)BMICholesterol mmol/lHDLC mmol/lLDLC mmol/lTG mmol/lGlucose mmol/lHbA1CInsulin pmol/l

1041 ± 156 (40%)23.9 ± 2.54.5 ± 0.61.4 ± 0.32.3 ± 0.40.8 [0.7-1.0]4.9 ± 0.636 ± 2.344 ± 19

1041 ± 146 (40%)23.8 ± 2.94.8 ± 1.31.5 ± 0.32.5 ± 0.80.9 [0.7 – 1.4]4.8 ± 0.536 ± 3.730 ± 15

0.580.760.530.300.630.820.40

Data are means ± SD or median [IQR] Abbreviations: LDL = low-density lipoprotein; HDL = high-density lipoprotein; TG = triglycerides; BMI = body mass index. Differences were tested using T-test statistics and for plasma TG a Mann-Whitney test.


119

7

0 2 4 6 8 100.00.51.01.52.02.5

TG m

mol

/l

Time (h)

Control HME0

2

4

6

8

IAU

C T

G (m

mol

/l.h-

1)

Figure 1 - Figure 1A Postprandial TG excursions after an oral fat load of cream in HME patients () and controls (). Data are presented as mean ± sem. Figure 1B postprandial iAUC-TG in HME patients and controls. Data are presented as mean ± SD.

Patients with heterozygous Familial Hypercholesterolemia Selection of GiRaFH patientsBaseline characteristics of all GiRaFH patients are presented in supplementary Table 1. We selected tagSNPs in 5 genes involved in HSPG synthesis i.e. NDST1, EXT1, EXT2, SDC1, and HS2ST1 (for details see supplementary Table 2). We could only detect minor associations with plasma TG levels (supplementary Table 3). A weighted risk score, based on the total genetic variation in HSPG genes, was calculated. The average predicted plasma TG levels in the low and high HSPG gen score groups were 1.23 [0.96 – 1.75] and 1.75 [1.26 – 2.42] respectively, P< 0.001. We randomly selected 11 patients from each group for participation in an oral fat load test. The baseline characteristics of the patients in the high and low HSPG gene score group (n= 11 each) are shown in Table 2. The groups were comparable for age, BMI and fasting plasma TG levels.

120

Chapter 7

Oral fat load testOur first observation was that FH patients have an impaired postprandial TG clearance in comparison to the control subjects used in the HME study reflected by an increased iAUC-TG 6.1 [3.8 – 11.7] versus 2.7 [1.9 – 4.3] mmol/l.h-1. Second, there was no significant difference in postprandial plasma TG excursions between the two HSPG gene score groups (iAUC-TG was 6.93 [3.89 – 7.62] in low gene score group versus 4.39 [2.94 – 13.4] mmol/l.h-1 in the high gene score group (Figure 2)). Postprandial plasma retinyl palmitate excursions were significant higher in FH patients with a high gene score suggesting a delay in dietary-derived lipid clearance (Figure 2) (iAUC for retinyl palmitate excursions 14.3 ± 5.6 ng/ml.h-1 in the low gene score group and 20.9 ± 8.5 ng/ml.h-1 for the high gene score group, p<0.05). No difference in heparin releasable LPL capacity was found.

DISCUSSIONIn this study we evaluated the effect of genetic variation in genes involved in HSPG synthesis on postprandial TG and retinyl palmitate clearance in HME patients with heterozygous loss of function mutations in HSPG as well as in a large cohort of patients with heterozygous familial hypercholesterolemic (FH) stratified for a high or low HSPG gene score. In HME patients, a deleterious mutation in EXT1 or EXT2 was not associated with alterations in fasting or postprandial TG levels. In FH patients, the combination of an LDL receptor mutation and multiple SNPs in pivotal HSPG genes had no effect on fasting or postprandial TG levels. A modest increase in the postprandial retinyl palmitate levels in FH with high HSPG gene score was observed. These data imply that in humans HSPGs play, if any, a minor role in TG metabolism.

Postprandial remnant lipoprotein particles have to be considered important mediators in the atherogenic process and hepatic uptake of TG-rich remnant particles involves the participation of 3 different receptors 24. In this respect, LDL receptors are since long known to be involved. LDL

Table 2 - baseline Characteristics of the FH patients participating in the fat load study

Low gene score High gene score P-value

NGender (% male)Age (y)BMIHbA1CCholesterol (mmol/l)LDL chol (mmol/l)HDL chol (mmol/l)TG (mmol/l)Glucose (mmol/l)Insulin (pmol/l)

115 (50 %)26 ± 225.8 ± 3.637 ± 2.68.1 ± 1.45.8 ± 1.01.46 ± 0.451.3 [1.0 -1.9]5.1 ± 0.535 ± 32

115 (50%)26 ± 226.4 ± 4.038 ± 3.58.6 ± 1.46.2 ± 1.41.45 ± 0.451.1 [0.79 – 2.2]5.3 ± 0.740 ± 32

0.550.390.190.060.390.660.150.51

Data are presented as mean ± SD. TG are presented as median [interquartile range]. Differences were tested using T-test statistics and for plasma TG a Mann-Whitney test.


121

7

receptors have a high capacity for remnant clearance, yet they are not very abundantly expressed at the cell surface 7. Interestingly, the vast majority of FH patients are characterized by a partial loss of function of the LDL receptor, which can be expected to have an impact on remnant clearance in vivo. In support, we show that postprandial TG clearance is impaired in the FH population as compared to the control group. The observation that the increase in TG levels is only modest is substantiated by the reports of only minor TG elevations in FH homozygote patients 25. This has suggested the participation of other receptors in the uptake of remnant particles in the liver. Based on their hepatic localisation, HSPG have been implicated in this process, but their capacity to take up remnant particles is known to be low 26. Studies on the role of HSPGs in human lipoprotein metabolism has are scarce. Patients with Knobloch Syndrome, caused by a heterozygous loss of function mutation in vascular collagenXVIII (COL18), were reported to have increased fasting triglycerides but unfortunately no postprandial lipid loads were performed 27. CollagenXVIII is a proteoglycan that is present in the basement membrane and is involved the transport of lipoprotein lipase to the cell surface where it can bind to GPIHBP1, thus enabling TG lipolysis 28.

0 2 3 4 5 6 80

1

2

3

4

time (h)

TG (m

mol

/l)

low gene score high gene score02468

10

AU

C-T

G m

mol

/l.h

0 1 2 3 4 5 6 7 80

2

4

6

Time(h)

RE

(ng/

ml)

low gene score high gene score05

10152025

AU

C-R

E ng

/ml/h P<0.05

Figure 2 - Postprandial lipids after an oral fat load in selected FH with either a low gene score or a high gene score. A: postprandial plasma TG excursions. B. Postprandial retinyl ester excursions. () represents the low gene score group; () represents the high gene score group. Data are presented as mean ± SD. Data were tested using T-test statistics.

122

Chapter 7

In the present study we investigated two groups of patients with affected HSPGs. First: HME patients who have a diagnosed heterozygous loss of function mutation in either EXT1 or EXT2 and second: heterozygous FH patients with loss of function mutation in the LDLR and genetic variation in common SNPS in genes in HSPG homeostasis. EXT 1 and EXT2 are required for heparan sulfate chain elongation 11, thus loss of function mutations result in HSPGs with reduced chain length of the heparan sulfate chain yet normal sulfation pattern. Therefore, HME patients form a unique human model for HSPG chain length insufficiency. In an earlier study we showed that type 2 diabetic mice overexpress glucosamine-6-O-endosulfatase 2 (SULF2), an enzyme that removes the 6-O sulfate groups from HSPG. The animals have an impaired HSPG-mediated hepatic TRL clearance, which can be normalized by treating the mice with an antisense oligonucleotide targeting Sulf2. 29 Thus complete sulfation of HSPGs, in a murine model, seems to be essential for clearance of fasting and postprandial TG-rich lipoprotein remnants.

Studying postprandial lipid metabolism in FH subjects may highlight the role of other (non-LDLr) hepatic TG clearance pathways, since it is not expected that saturation of all involved pathways is likely to occur 8. We genotyped tag SNPs reflecting the total genetic variation in each tested HSPG gene and tested for the association with plasma TG levels 23. Only minor associations were found. We then generated a HSPG gene score based on the cumulated correlation coefficients of each HSPG SNP and its association with fasting plasma TG level. This provided us with a tool to categorize FH patients with a low burden in genetic variation in HSPG genes and those with a large number of genetic variants in HSPG genes. In a subset of FH patients with a low and high HSPG gene score we performed an oral fat load test. No differences were found in fasting or postprandial plasma TG levels were noted. Interestingly, postprandial retinyl palmitate levels did show an increase in FH subjects with the highest HSPG gene score implying a potential role for HSPGs particularly in chylomicron clearance.

study limitationsSeveral aspects need closer consideration when interpreting the results of this study. A major limitation of our study was the relatively small number of participants in postprandial lipid loading studies predominantly due to low number of HME patients in the Netherlands as well as the limited number of FH patients with a high HSPG gene score who meet the selection criteria. Second, although retinyl palmitate has been widely used in postprandial studies as marker for dietary derived TG, it has been suggested that retinyl palmitate may exchange very slowly between chylomicrons and VLDL 30. However detailed studies have revealed that only 5% of retinyl palmitate was associated with VLDL lipoproteins 5 h after a fat load. Thus it is not very likely that retinyl palmitate exchange between lipoprotein particles highly contribute to the differences in postprandial curves.

Conclusion We here show for the first time that heparan sulfate proteoglycans might be involved in the clearance of postprandial triglycerides in humans. In HME subjects, with impaired HS chain length;


123

7

there was a trend towards delayed hepatic clearance capacity of postprandial TG-rich lipoproteins. Furthermore, FH subjects stratified by HSPG gene score (thus partly bypassing the LDLr pathway) revealed a small difference in clearance of postprandial retinyl palmitate, suggesting only a minor contribution of HSPGs to human triglyceride clearance. Nevertheless, further studies are required to further elucidate the role of HSPG, with a focus on HS sulfation pattern, in human TG metabolism.

acknowledgmentsWe are grateful to all participating HME subjects and Jan de Lange from the Dutch HME Foundation (www.hme-mo.nl) for their help with inclusion and being able to perform this study.

124

Chapter 7

REFERENCE LIST1. Nordestgaard BG, Benn M, Schnohr P, Tybjaerg-Hansen A. Nonfasting triglycerides and risk of myo-

cardial infarction, ischemic heart disease, and death in men and women. JAMA 2007;298:299-308.

2. Kolovou GD, Mikhailidis DP, Kovar J, et al. Assessment and clinical relevance of non-fasting and post-prandial triglycerides: an expert panel statement. Curr Vasc Pharmacol 2011;9:258-70.

3. Miller M, Stone NJ, Ballantyne C, et al. Triglycerides and cardiovascular disease: a scientific state-ment from the American Heart Association. Circulation 2011;123:2292-333.


5. Jackson KG, Poppitt SD, Minihane AM. Postprandial lipemia and cardiovascular disease risk: Interre-lationships between dietary, physiological and genetic determinants. Atherosclerosis 2012;220:22-33.

6. Schwartz EA, Reaven PD. Lipolysis of triglyceride-rich lipoproteins, vascular inflammation, and ath-erosclerosis. Biochim Biophys Acta 2012;1821:858-66.

7. Williams KJ. Molecular processes that handle -- and mishandle -- dietary lipids. J Clin Invest 2008;118:3247-59.

8. Hassing HC, Surendran RP, Mooij HL, Stroes ES, Nieuwdorp M, Dallinga-Thie GM. Pathophysiology of hypertriglyceridemia. Biochim Biophys Acta 2012;1821:826-32.

9. Stanford KI, Bishop JR, Foley EM, et al. Syndecan-1 is the primary heparan sulfate proteoglycan me-diating hepatic clearance of triglyceride-rich lipoproteins in mice. J Clin Invest 2009;119:3236-45.


11. Busse M, Feta A, Presto J, et al. Contribution of EXT1, EXT2, and EXTL3 to heparan sulfate chain elon-gation. J Biol Chem 2007;282:32802-10.


13. Fuki IV, Kuhn KM, Lomazov IR, et al. The syndecan family of proteoglycans. Novel receptors mediat-ing internalization of atherogenic lipoproteins in vitro. J Clin Invest 1997;100:1611-22.

14. Stanford KI, Wang L, Castagnola J, et al. Heparan sulfate 2-O-sulfotransferase is required for triglyc-eride-rich lipoprotein clearance. J Biol Chem 2010;285:286-94.

15. Planer D, Metzger S, Zcharia E, Wexler ID, Vlodavsky I, Chajek-Shaul T. Role of heparanase on hepatic uptake of intestinal derived lipoprotein and fatty streak formation in mice. PLoS One 2011;6:e18370.

16. Solomon L. Hereditary Multiple Exostosis. Am J Hum Genet 1964;16:351-63.

17. Duncan G, McCormick C, Tufaro F. The link between heparan sulfate and hereditary bone dis-ease: finding a function for the EXT family of putative tumor suppressor proteins. J Clin Invest 2001;108:511-6.

18. Goud AL, de Lange J, Scholtes VA, Bulstra SK, Ham SJ. Pain, physical and social functioning, and qual-ity of life in individuals with multiple hereditary exostoses in The Netherlands: a national cohort study. J Bone Joint Surg Am 2012;94:1013-20.


125

7

19. Jansen AC, van Aalst-Cohen ES, Tanck MW, et al. The contribution of classical risk factors to cardiovas-cular disease in familial hypercholesterolaemia: data in 2400 patients. J Intern Med 2004;256:482-90.

20. Hovingh GK, Davidson MH, Kastelein JJ, O’Connor AM. Diagnosis and treatment of familial hypercho-lesterolaemia. Eur Heart J 2013.

21. Ruotolo G, Zhang H, Bentsianov V, Le NA. Protocol for the study of the metabolism of retinyl esters in plasma lipoproteins during postprandial lipemia. J Lipid Res 1992;33:1541-9.

22. Bligh EG, Dyer WJ. A rapid method of total lipid extraction and purification. Canadian journal of bio-chemistry and physiology 1959;37:911-7.

23. Vergeer M, Boekholdt SM, Sandhu MS, et al. Genetic variation at the phospholipid transfer protein locus affects its activity and high-density lipoprotein size and is a novel marker of cardiovascular disease susceptibility. Circulation 2010;122:470-7.

24. Lambert JE, Parks EJ. Postprandial metabolism of meal triglyceride in humans. Biochim Biophys Acta 2012;1821:721-6.

25. Raal FJ, Santos RD, Blom DJ, et al. Mipomersen, an apolipoprotein B synthesis inhibitor, for lowering of LDL cholesterol concentrations in patients with homozygous familial hypercholesterolaemia: a randomised, double-blind, placebo-controlled trial. Lancet 2010;375:998-1006.


27. Bishop JR, Passos-Bueno MR, Fong L, et al. Deletion of the basement membrane heparan sul-fate proteoglycan type XVIII collagen causes hypertriglyceridemia in mice and humans. PLoS One 2010;5:e13919.

28. Beigneux AP, Davies BS, Gin P, et al. Glycosylphosphatidylinositol-anchored high-density lipopro-tein-binding protein 1 plays a critical role in the lipolytic processing of chylomicrons. Cell Metab 2007;5:279-91.

29. Hassing HC, Mooij H, Guo S, et al. Inhibition of hepatic sulfatase-2 in vivo: a novel strategy to correct diabetic dyslipidemia. Hepatology 2012;55:1746-53.

30. Lemieux S, Fontani R, Uffelman KD, Lewis GF, Steiner G. Apolipoprotein B-48 and retinyl palmitate are not equivalent markers of postprandial intestinal lipoproteins. J Lipid Res 1998;39:1964-71.

126

Chapter 7


supplementary Table 1 - baseline characteristics of the FH cohort

Parameter

N

Age

BMI

Cholesterol (mmol/l)

HDLC (mmol/l)

LDLC (mmol/L)

Triglycerides (mmol/l)

2344

45 ± 12

25 ± 3.5

8.6 ± 1.9

1.24 ± 0.35

6.4 ± 1.8

1.69 [1.07]

Data are presented as mean ± SD. TG is given as median [IQR]

Supplementary Table 2 - Selected Tag SNPs in HSPG related genes

Gene rs number Genetic variation Genotype counts

NDST1

NM_001543.3

rs10074650

rs1290147

rs12516924

rs3846709

c.-387-3139T>G

c.1662G>T;p.T554T

c.513+505A>G

c.513+2530G>T

TT: 750

CC: 649

AA: 546

GG: 508

GT: 598

CG: 630

AG: 682

GT: 702

GG: 100

GG: 169

GG: 220

TT: 238

SDC1

NM_002997.4

rs11096648

rs11805809

rs3732165

rs3771240

rs4666298

g.4099116T>C

g.4069330G>A

c.148+63G>C

c.67-7501A>G

c.66+3831T>G

TT: 586

GG: 473

GG: 1167

AA: 580

TT: 901

TC: 655

GA: 731

GC: 265

AG: 658

TG: 474

CC: 207

AA: 244

CC: 16

GG: 210

GG: 73

EXT1

NM_000127.2

rs11562695

rs7010382

rs5001657

rs17479145

rs17429936

rs921957

c.962+6329A>G

c.962+14119C>T

c.2055+780T>C

c.1723-103C>G\

c.1723-755A>C

c.962+14890T>C

AA: 952

CC: 587

TT: 515

CC: 1205

AA: 839

TT: 867

AG: 449

CT: 678

TC: 690

CG: 234

AC: 520

TC: 503

GG: 47

TT: 183

CC: 243

GG: 9

CC: 89

CC: 78

EXT2

NM_000401.3

rs4447163

rs4551754

rs12280077

rs4755230

c.1273-8083C>A

c.1272+19768T>C

c.1595-3583G>C

c.1273-8817A>G

CC: 1212

TT: 672

GG: 523

AA: 787

CA: 223

TC: 616

GC: 671

AG: 562

AA: 13

CC: 160

CC: 254

GG: 99

HS2ST1

NM_012262.3

rs6702601

rs2764427

rs4610985

rs4655920

c.125-25059G>A

c.363+3463G>A

c.125-17647G>A

c.124+51630A>G

GG: 900

GG: 879

GG: 1237

AA: 1055

GA: 491

GA: 504

GA: 203

AG: 360

AA: 57

AA: 65

AA: 8

GG: 33


127

7

supplementary Table 3 - Frequencies of the sNPs and association of tagsNPs in HsPG associated genes and fasting plasma TG levels

SNP Homozygous major allele

Heterozygous Homozygous minor allele

P-value

NDST1 rs10074650 T>G

NDST1 rs1290147 C>G

NDST1 rs12516921 A>G

NDST1 rs3846709 G>T

SDC1 rs11805809 T>C

SDC1 rs11096648 G>A

SDC1 rs3732165 G>C

SDC1 rs3771240 A>G

SDC1 rs4666298 A>C

HS2ST1 rs4655920 A>G

HS2ST1 rs6702601 G>A



EXT1 rs11562695 A>G

EXT1 rs7010382 C>T

EXT1 rs5001657 T>C

EXT1 rs17479145 C>G

EXT1 rs17429936 A>C

EXT1 rs921957 T>C

EXT2 rs4447163 C>A

EXT2 rs4551754 T>C

EXT2 rs12280077 G>C

EXT2 rs475200 A>G

1.38 [1.02 – 2.09]

1.70 [1.13 – 2.20]

1.71 [1.12 – 2.29]

1.70 [1.12 – 2.26]

1.56 [1.12 – 2.12]

1.74 [1.14 – 2.17]

1.15 [0.87 – 2.52]

1.64 [1.16 – 2.17]

1.64 [1.10 – 2.20]

1.66 [1.12 – 2.16]

1.88 [1.08 – 2.54]

1.63 [1.01 – 2.53]

1.14 [0.79 - 1.87]

1.54 [1.21 – 2.33]

1.65 [1.10 – 2.11]

1.70 [1.12 – 2.23]

1.29 [0.99 – 1.94]

1.65 [1.10 – 2.20]

1.67 [1.11 – 2.15]

1.67 [1.10 – 2.16]

1.73 [1.13 – 2.18]

1.73 [1.16 – 2.16]

1.63 [1.08 – 2.19]

1.70 [1.10 – 2.11]

1.65 [1.09 – 2.16]

1.71 [1.12 – 2.13]

1.71 [1.10 – 2.14]

1.71 [1.07 – 2.17]

1.62 [1.10 – 2.14]

1.57 [1.11 – 2.19]

1.63 [1.10 – 2.15]

1.74 [1.12 – 2.11]

1.70 [1.07 – 2.20]

1.69 [1.09 – 2.10]

1.70 [1.10 – 2.10]

1.60 [1.12 – 2.11]

1.66 [1.09 – 2.19]

1.68 [1.13 – 2.20]

1.66 [1.10 – 2.11]

1.64 [1.15 – 2.28]

1.71 [1.13 – 2.18]

1.69 [1.08 – 2.20]

1.75 [1.12 – 2.20]

1.65 [1.09 – 2.22]

1.64 [1.08 – 2.23]

1.71 [1.12 – 2.20]

1.70 [1.12 – 2.26]

1.70 [1.06 – 2.16]

1.51 [1.06 – 2.04]

1.59 [1.10 – 2.09]

1.66 [1.13 – 2.20]

1.73 [1.12 – 2.23]

1.70 [1.10 – 2.18]

1.73 [1.12 – 2.22]

1.79 [1.16 – 2.37]

1.79 [1.10 – 2.25]

1.68 [1.12 – 2.20]

1.68 [1.12 – 2.20]

1.70 [1.10 – 2.19]

1.70 [1.10 – 2.16]

1.75 [1.05 – 2.24]

1.71 [1.09 – 2.15]

1.69 [1.10 – 2.17]

1.85 [1.01 – 1.99]

1.75 [1.12 – 2.21]

1.80 [1.09 – 2.59]

1.65 [1.08 – 2.05]

1.66 [1.09 – 2.11]

1.77 [1.22 – 2.09]

0.01

0.589

0.088

0.418

0.857

0.476

0.289

0.862

0.583

0.434

0.468

0.703

0.579

0.176

0.507

0.590

0.003

0.537

0.854

0.477

0.929

0.902

0.759

Data are expressed as median [IQR]. Differences between the genotype groups were tested using ANOVA (log transformed TG).

Mooij HL, Bernelot Moens SJ, Hassing HC, Kruit JK, Witjes JJ, van de Sande MAJ, Stoker J, Nederveen AJ, Xu D, Esko JD, Dallinga-Thie GM, Stroes ESG, Nieuwdorp M

8 CaRRiERs oF Loss-oF-FuNCTioN MuTaTioNs iN EXT DisPLay iMPaiRED PaNCREaTiC bETa-CELL REsERVE DuE To SMALLER PANCREAS vOLUME

130

Chapter 8

ABSTRACTAim - A genome wide association study identified exostosin 2 (EXT2) as a novel risk factor for the development of type 2 diabetes mellitus. As EXT genes, involved in the chain elongation step of heparan sulfate (or HSPG) biosynthesis, are intricately involved in organ development, we hypothesized that mutations in these genes might affect pancreatic islet mass and insulin secretion capacity. Here we used a translational approach to study the effect of EXT mutations on pancreatic development, insulin secretion and glucose metabolism in mice and humans with heterozygous EXT mutations causing hereditary multiple exostoses (HME).

Methods - In heterozygous Ext1 or Ext2 knock-out mice we performed oral glucose tolerance tests (OGTT), insulin tolerance tests (ITT) and harvested each mouse pancreas for extraction of islets (insulin secretion) and immunohistochemistry (beta cell mass). In HME subjects and family-based non-carriers (similar age, sex, and BMI) we repeated OGTT followed by hyperglycemic clamps to investigate first-phase insulin secretion (GSIS). Finally, abdominal MRI was assessed to quantify total pancreas volume.

Results - No differences in oral glucose tolerance and insulin resistance were found in mice and humans with EXT mutations compared to controls. No effects on insulin signalling were found in isolated islets challenged with hyperglycaemia. However, glucose stimulated insulin secretion (hyperglycaemic clamp) showed that HME subjects had a significantly altered GSIS as well as an impaired beta cell reserve (upon arginine bolus). In line with these finding, Magnetic Resonance Imaging showed a significantly smaller pancreas volume in HME subjects compared to controls. Conclusions - Carriers of loss-of-function mutations in EXT showed impaired GSIS without insulin resistance due to reduced functional beta cell mass and decreased anatomic pancreas volume. Our data provide evidence that heparansulfates are important for normal beta-cell function in humans.

EXT mutations and beta-cell function in humans

131

8

INTRODUCTIONAlthough the worldwide rising incidence of type 2 diabetes and obesity is largely attributed to environmental factors, the existence of a genetic predisposition has long been recognized1. In this respect, genome wide association studies (GWAS) have rendered conflicting results on the role of SNPs in EXT2 (exostosin 2) as a novel genetic risk factor for the development of type 2 diabetes mellitus (DM2)2–8. Also, SNPs in EXT2 have been associated with impaired glucose clearance in DM2 as assessed by an oral glucose tolerance test9. A recent meta-analysis of Liu et al showed a significant association between the EXT2 SNP variants and the risk of developing DM210. However, despite these epidemiological associations, little is known about the pathophysiological role of EXT in glucose metabolism.

EXT 1 and 2 genes encode for an endoplasmic reticulum-resident type II transmembrane glycosyltransferase involved in the chain elongation step of heparan sulfate biosynthesis11,12. Heparansulfate proteoglycans (HSPG) play a role in many biological processes including fine-tuning most of the (patho)physiological processes such as (fetal) organ development, lipid metabolism and inflammatory pathways13. More specific, a subset of the Ext genes, Ext2 and Extl3, were reported to be involved in pancreatic β cells development in mice14. Whereas in humans EXTL3 is not involved in the initiation of HS Biosynthesis, the role of EXT1 and EXT2 genes has been widely recognized15.

Heterozygous EXT mutations are known to be involved in the development of hereditary multiple exostoses (HME) syndrome16, a disorder in which disrupted HS synthesis induces growth of multiple bony tumors (eg. Exostoses or osteochondromas)17. However, metabolic derangements have never been studied in this patient group18. In the present study we designed a dedicated series of experiments in both mice and humans to unravel the effect of EXT mutations on beta cell mass and function, as well as insulin-glucose homeostasis in mice and humans with heterozygous EXT mutations.

METHODSOral glucose tolerance test and insulin tolerance tests in Ext1 and Ext2 heterozygous knockout mice and wild type miceAll animals were housed in barrier conditions in vivaria of the AMC-UvA and all protocols were approved by the Association for Assessment and Accreditation of Laboratory Animal Care. Moreover, standards and procedures approved by the local Institutional Animal Care and Use Committee were followed. Mice were weaned at 3 weeks, were maintained on a 12-hour light–dark cycle and fed water and standard rodent chow (Harlan-TekLad City Country) ad libitum. All animals were fully backcrossed on a C57Bl/6 background. Genotyping for Ext1 and Ext2 heterozygosity was performed using primers as described earlier19. All investigations were performed in 4h fasted male mice (25-30g, aged 9-12 weeks). After baseline blood sampling, glucose was administered as oral gavage (2g/

132

Chapter 8

kg body weight) followed by tailvein blood sampling at 15,30,45,60,90 and 120 min. For the insulin tolerance test (ITT) mice were fasted for 4 hours before the study. After baseline blood sampling, insulin (Actrapid 1U/kg bodyweight) was administered intraperitoneally and subsequent tailvein blood sampling was performed at 15,30,45,60,90 and 120 min. Plasma glucose concentrations were determined with a glucose meter (Lifescan One Touch, Johnson and Johnson Company).

insulin secretion in isolated mice isletsAt 12 weeks of age, mice were sacrificed using and pancreas was extracted after collagenase injection for islet isolation as described previously20. Islets were cultured overnight in RPMI 1640 medium (Sigma Aldrich). The following day, batches of 10 hand picked islets were preincubated in Krebs Ringer bicarbonate buffer (KRB) containing 1,67 mM glucose for 1 h at 37 °C in 95% O2/5% CO2, Thereafter, islets were incubated for 1 hour in 0.1 ml of fresh KRB containing either 1,67 mM or 20mM of glucose . Subsequently, media was removed and islets were lysed in 1M acidic acid. Insulin levels in both media and islets were determined using an ultrasensitive Mouse Insulin ELISA (Mercodia, Uppsala, Sweden).

Human studies We enrolled HME subjects and family-based non-carriers over 18 years of age, without pre-existent type 1 or 2 diabetes. As redundancy between EXT1 and EXT2 exists we tested both subjects with either EXT1 or EXT2 heterozygous mutations for alterations in glucose metabolism and pancreatic reserve. For all tests, subjects were requested to arrive in a fasting state and written informed consent was obtained after explanation of the study. The study was approved by the institutional review board of the Academic Medical Center of the University of Amsterdam and carried out according to the Declaration of Helsinki.

Oral Glucose Tolerance Test (OGTT) and hyperglycaemic normoinsulinemic clampAfter an overnight fast, a standardized OGTT was performed. After baseline venous sampling subjects were asked to ingest 75 g glucose. At t = 30, 60, 90 and 120 minutes an 4.5 ml blood sample was obtained for assessment of blood glucose, insulin and C-peptide. On a separate study day, a hyperglycemic clamp was performed. On the day of study antecubital veins of both arms were canulated for blood sampling and infusion of fluids. All bedside glucose measurements were performed using a bedside calibrated glucose sensor (YSI 2300 STAT S; YSI, Yellow Springs, OH). Based on the fasting plasma glucose level and the subject’s bodyweight first phase insulin secretion was determined using a 20% glucose bolus (weigt/70 x 10 – plasma glucose = millilitres required) given over 1 min, with the aim of reaching a plasma glucose level of 14 mmmol/L. Subsequently, blood glucose levels were kept at 14 mmol/L by continuous glucose infusion. Pump settings (glucose infusion rate) were adapted based on blood glucose levels at t= 0, 2.5, 5, 7.5, 10 and 20. Simultaneously, blood samples were collected for insulin and C-peptide determination. After 120 minutes an arginine bolus (5 gram arginine hydrochloride, 50 ml per 100mg/ml solution) was given, followed by measurement of plasma insulin levels at t= 125, 130,140 and 150 minutes. Basal


133

8

fasting glucose, HbA1c, total cholesterol, HDL, and LDL cholesterol and triglycerides were assessed in fasting plasma using standard laboratory procedures within 1 h after sampling. For the OGTT and hyperglycaemic clamps, samples were centrifuged at 4ºC, 3000 RPM for 20 minutes and plasma was stored at -80 ºC until analysis. Glucose was determined by the hexokinase method (Hitachi), Insulin was determined on an Immulite 2000 system (Diagnostic Products, Los Angeles, CA). C-peptide was measured by RIA (RIA-coat C-peptide; Byk-Sangtec Diagnostica, Dietzenbach, Germany). Homeostasis model assessment (HOMA) indexes were calculated for insulin sensitivity (HOMA-ir = insulin (picomoles)/6.945*glucose (millimoles)/22.5) and insulin secretion (HOMA-%β = 20* fasting insulin (picomoles)/6.954/glucose (millimoles-3.5). In line, in the OGTT insulin sensitivity was estimated using the metabolic clearance rate (MCR) of glucose and the insulin sensitivity index (ISI), both as described previously21. Overall glucose-stimulated insulin secretion was calculated as AUCinsulin/AUCglucose ratio.

MRIIn a subset of previous participants (both HME-subjects and healthy controls) we performed abdominal imaging preceding the OGTT, using a 3-T MR imaging unit (Intera, Philips Healthcare, Best, The Netherlands). A Single Shot Fast Spin Echo (SSFSE) sequence was performed and a T1-weighted high-resolution axial anatomical scan obtained in breath hold (matrix 320X320, field of-view [FOV] ((450x450mm) was used to investigate the morphology of the pancreas. Images were analyzed by 2 independent, blinded investigators using ITK Snap software version 2.4 (University of Pennsylvania). Pancreatic area was labelled and number of voxels in this area was determined, subsequently this number was transcribed to volume in cubic millimetres as previously published22. The mean area of separate measurements was used.

statistical analysisData are presented as mean ± SD or medians with interquartile range ([IQR] unless stated otherwise. When normally distributed baseline characteristics where compared using a student’s t test (all but triglycerides). Differences in triglyceride levels, known not to be normally distributed, and continues outcome variables were assessed using the nonparametric Mann-Whitney U test. For all outcomes P<0.05 was used to indicate significant differences. All analyses were performed with SPSS software version 19.0.0.1.

RESULTSMetabolic parameters and islets function in Ext1 and Ext2 heterozygous versus wild type mice To investigate the effects of Ext on pancreas anatomy and function (insulin secretion) we performed oral glucose tolerance test (OGTT) and intraperitoneal insulin tolerance tests (ITT) in mice with heterozygous mutations in Ext1+/- , Ext2+/- as well as wild type mice (Table 1A and Table 1B). As compared to wild type mice, heterozygous Ext1+/- and Ext2+/- mice were characterized by normal oral

134

Chapter 8

glucose disposal (IAUC Ext1+/- 579 [494-707], Ext2+/- 680 [636-756] and WT 688 [501-794] mmol·l-

1·min-1, ns, see Figure 1A). Moreover, insulin sensitivity as assessed by ITT was comparable between Ext1+/-, Ext2+/- and WT mice (IAUC Ext1+/- 472 [453-492], Ext2+/- 381 [316-453], WT 362 [345-361] mmol·l-1·min-1, ns, see Figure 1B). Next, we investigated islets insulin secretion in isolated islets from all phenotypes. Insulin secretion was not impaired in Ext1+/- or Ext2+/- mice, also total islet insulin content was comparable between all groups (Insulin secretion after 24mmol glucose stimulation for Ext1+/-: 0,23[0,19-0,37] μg/L, Ext2+/-: 0,30[0,17-0,48] μg/L and wild type mice 0,24[0,18-0,38]μg/L, ns see Figure 2).

beta cell function and glucose metabolism in human EXT carriers versus controlsAge, BMI, fasting glucose, HbA1C and insulin levels did not differ significantly between carriers and control subjects (see Table 2). Moreover, no differences between EXT carriers and controls were found upon ingestion of 75 g of glucose with respect to plasma glucose (iAUC: carriers; 233 [2174-299] vs controls; 183 [100-286] nmol·l-1·min-1 p=0,46) and plasma insulin (iAUC: carriers; 13,2 [8,5-18,6,0] vs controls; 16,5 [11,1-21,7] nmol·l-1·min-1 p=0,46) (see Figure 3A and 3B). Several markers of insulin resistance and beta cell function were calculated yet showed no significant differences between the two groups (summarized in Table 3).

To evaluate the potential effect on GSIS, we subsequently performed a hyperglycaemic normoinsulinemic clamp, followed by arginine infusion. During this hyperglycaemic clamp, first phase insulin response to a glucose bolus (as determined by incremental AUC) was lower in carriers than control subjects (0.72 [0.46-1.16] vs. 1.53 [0.69-3.36] nmol·l-1·min-1, P=0.046) (Figure 4A). In addition, C-peptide responses were also lower in carriers (3,57 [2,26-5,00] vs. 6.62 [4,48-9,84] nmol·l-1·min-1 p=0.006) (Figure 4B). Of note, in line with our HOMA findings, glucose infusion rates were similar between the groups (iAUC carriers vs. controls 11,1 [8,88-19,00] vs. 14,5 [11,98-23,98] mg·kg-1·min-1 p=0,3)(Figure 4C), suggesting that differences found in insulin and C-peptide secretion are not due to differences in glucose tolerance. Finally, upon intravenous arginine bolus the peak-insulin response was impaired in EXT carriers compared to controls (IAUC 7,14 [4,22-10,95] vs. 10,32 [7,91-12,70] nmol·l-1·min-1 p=0.041)(Figure 4D and E).

Based on these findings of functional decreased beta cell insulin secretory capacity we hypothesized whether a difference in anatomical pancreas volume could be detected in a subset of the previously tested EXT carriers compared to controls (see Table 4). Subsequent abdominal MRI imaging revealed a significantly smaller pancreas volume in EXT carriers compared to control subjects (74 [63-86] vs.87 [82-105] cm3 p=0,016)(Figure 5).


135

8

Table 1 - baseline mouse characteristics

Table 1A WT (N=8) Ext1+/- (N=8) Ext2+/- (N=8)

Age (weeks) 8.5±0.5 8.6±0.5 8.5±0.5

Bodyweight (gr) 27,8±1,1 26,8±1.3 27.8±1.6

Fasting triglycerides 1,1 [0,91-1,2] 1,0 [0,96-1,1] 1,1[1,1-1,2]

Table 1B WT (N=8) Ext2+/- (N=8) Ext2+/- (n=8)

Age (weeks) 10,3±0.5 10,0±0,0 10,4±0,5

Bodyweight (gr) 30,0±0.69 29,2±0,82 30,2±1,3

Baseline mouse carachteristics for (A) oral glucose tolerance test and (B) itraperitoneal tolerance test. Data are means ± SD or median [IQR].

Figure 1A: glucose levels during OGTT in mice

0 15 30 45 60 90 120

0

5

10

15

20

25WT

Ext1+/-

Ext2+/-

Time after gavage (min)

gluc

ose

(mm

ol/L

)

Figure 1B: glucose levels during IPITT in mice

0 15 30 45 60 90 120

0

2

4

6

8WT

Ext1+/-

Ext2+/-


gluc

ose

(mm

ol/L

)

Figure 1 - Glucose homeostasis in wT (▲) versus Ext1 (■) vs Ext2 (○) mice. (A) Plasma glucose after oral glucose tolerance test. (B) Plasma glucose after intraperitoneal insulin tolerance test

Figure 1A: glucose levels during OGTT in mice

0 15 30 45 60 90 120

0

5

10

15

20

25WT

Ext1+/-

Ext2+/-


gluc

ose

(mm

ol/L

)

Figure 1B: glucose levels during IPITT in mice

0 15 30 45 60 90 120

0

2

4

6

8WT

Ext1+/-

Ext2+/-


gluc

ose

(mm

ol/L

)

136

Chapter 8

Figure 2A: Islet insulin secretion

1,67 24 1,6

7 24 1,67 24

0

5

10

15

20WT

Ext1+/-

Ext2+/-

Insu

lin s

ecre

tion

(fol

d ch

ange

from

bas

al)

Figure 2B: Islet insulin content

WT +/-

Ext1+/-

Ext2

0

2000

4000

6000

Isle

t ins

ulin

con

tent

(g

/mg

prot

ein)

Figure 2 - No difference in insulin secretion in isolated islets from each genotype. (A) Insulin secretion from isolated islets, value represent data from 2 seperate experiments, each consisting of 3 mice per genotype. Values are expressed as percent of islet content relative to basal secretion (set at 1) (B) Total islets insulin content

Figure 2A: Islet insulin secretion

1,67 24 1,6

7 24 1,67 24

0

5

10

15

20WT

Ext1+/-

Ext2+/-

Insu

lin s

ecre

tion

(fol

d ch

ange

from

bas

al)

Figure 2B: Islet insulin content

WT +/-

Ext1+/-

Ext2

0

2000

4000

6000

Isle

t ins

ulin

con

tent

(g

/mg

prot

ein)

Table 2 - baseline characteristics of participants

Noncarriers (N=13)

Carriers (N=14) P-value

Age (years) 49±12 39±10 0,6

Men 8 (40) 7 (30)

BMI 25,6±3,4 25,8±4,9 0,18

BSA 1,8±0,16 1,8±0,19 0,16

Cholesterol (mmol/l)

Total 5,44±1,38 4,83±1,15 0,24

LDL 3,38±1,28 2,99±1,05 0,16

HDL 1,49±0,44 1,32±0,40 0,40

Triglycerides (mmol/l) 0,92[0,68-1,38] 0,87[0,56-1,32] 0,96

Fasting glucose (mmol/l) 4,9±0,60 4,8±0,53 0,63

Fasting insulin (pmol/l) 44±19 30±15 0,40


137

8

Table 3 - beta-cell function and insulin sensitivity parameters

Noncarriers Carriers P-value

Baseline HOMA index

HOMA-IR 1,2[0,79-1,79] 0,97[0,60-1,67] 0,30

HOMA-B 78[47-126] 78[50-144] 0,85

Insulinogenic index (pmol/mmol) 41,9[36,4-182] 62,6[32,2-107,5] 0,76

AUCinsulin/AUCglucose ratio (pmol/mmol) 86,2[71,5-254,8] 52,2[43,1-93,4] 0,23

ISIcomp(μmol/(kg min pmol L)) 24,5[22,5-48,3] 43,1[30,0-51,0] 0,23

MCR (ml/(min kg)) 9,9[9,4-10,6] 10,1[9,8-10,6] 0,40

Figure 3A: Glucose

0 50 100 150 2000

2

4

6

8

10CONTROLHME

Time after glucose ingestion

Figure3B: insulin

0 30 60 90 120

180

0

100

200

300ControlHME

Time after glucose ingestion (min)

Insu

lin (p

mol

/L)

Figure 3 - oGTT results in HME subjects (■) versus controls (○) Plasma glucose and insulin curves after 75g orally ingested glucose

Figure 3A: Glucose

0 50 100 150 2000

2

4

6

8

10CONTROLHME

Time after glucose ingestion

Figure3B: insulin

0 30 60 90 120

180

0

100

200

300ControlHME

Time after glucose ingestion (min)

Insu

lin (p

mol

/L)

Glu

cose

(mm

ol/l

)

138

Chapter 8

Figu

re 4

A: 1

st p

hase

insu

lin

0

2,5

5

7,5

10

0

100

200

300

400

500

HME

Con

trol

Tim

e (m

inut

es)

Insulin (pmol/L)Fi

gure

4B

: 1st

pha

se c

-pep

tide

0

2,5

5

7,5

10

0

500

1000

1500

2000

Con

trol

HME

Tim

e (m

inut

es)

C-peptide (pmol/L)

Figu

re 4

C: g

luco

se in

fusi

on ra

te

0

30

60

90

120

145

0510152025C

ontro

lHM

E

Tim

e (m

inut

es)

GIR (mg/kg/min)

Figu

re 4

D: i

nsul

in re

spon

se to

arg

inin

e

9010

011

012

012

513

014

015

0

0

1000

2000

3000

Con

trol

HME

Tim

e (m

inut

es)

Insulin (pmol/L)

Figu

re4E

: IAU

C in

sulin

resp

onse

to a

rgin

ine

Control

HME

0

5000

1000

0

1500

0

IAUC insulin (nmol·l-1·min-1)

Figu

re 4

- Fun

ction

al (G

sis)

pan

crea

s res

erve

in H

ME

subj

ects

(■) v

ersu

s con

trol

s (○)

. (A

and

B) T

he fi

rst-p

hase

insu

lin a

nd C

-pep

tide

resp

onse

to a

hyp

ergl

ycae

mic

cl

amp

was

low

er in

HM

E su

bjec

ts c

ompa

red

to c

ontr

ols.

(C) T

he G

luco

se in

fusio

n ra

te (G

IR),

an e

stim

ation

of t

he a

mou

nt o

f glu

cose

bei

ng m

etab

olize

d, w

as n

ot

diffe

rent

bet

wee

n gr

oups

. (D

and

E) In

sulin

secr

etion

afte

r an

intr

aven

ous b

olus

of a

rgin

ine

was

low

er in

car

riers

vs c

ontr

ols.


139

8

Table 4 - baseline characteristics of participants in MRi

Noncarriers (N=12) Carriers (N=8) P-value

Age (years) 42±14 41±10 0,95

Men 5 (41) 3 (37,5)

length (m) 1,72±0,10 1,72±0,10 0,9

Weight 69±9 78±12 0,1

BMI 23,8±1,9 22,9±10 0,75

BSA 1,8±0,16 1,8±0,19 0,16

Data are means ± SD, n (%), or median [IQR]

Figure 5: Pancreas Volume

Control

HME0

50

100

Panc

reas

vol

ume

(cm

3 )

Figure 5 - image of pancreas in HME subjects and controls en volume. Pancreas volume, assessed with 3T MRI, was smaller in HME subjects versus controls. (A) Labelled pancreatic area and 3D composition. (B) Pancreatic volume.

A

B

140

Chapter 8

DISCUSSIONFollowing the reports on genetic associations between EXT-SNPs and type 2 diabetes mellitus, we now provide the first evidence that carriers of loss-of-function mutations in EXT show an impaired GSIS in hyperglycaemic clamping studies..Combined without any evidence for beta cell dysfunction or peripheral insulin resistance in mice, the significant reduction in total pancreas volume implies a structural rather than a functional beta cell defect in carriers of loss-of-function mutations in EXT.

Functional and structural beta cell defects in heterozygous carriers of EXT mutations In contrast to the previous reports in type 2 diabetes subjects(9), both murine and human displayed no signs of hyperglycemia but a small (non significant) difference in insulin response upon an oral glucose challenge. Using hyperglycemic clamps, we subsequently found that EXT heterozygotes show a reduced first-phase insulin response to hyperglycemia (GSIS) and normal insulin sensitivity, resulting in a markedly decreased disposition index compared with noncarriers of similar age, sex, and BMI. Collectively, these findings point to a defect in β-cell function in EXT heterozygotes.

The specific means by which EXT dysfunction impairs β-cell function remain incompletely understood. Glucose enters the β-cell via GLUT, whereupon it is modified by glucokinase in the rate-limiting step in glucose sensing. The subsequent glucose metabolism pathway results in closing of the ATP-sensitive potassium channel, membrane depolarization, calcium influx into the cell via the L-type calcium channel, and exocytosis of insulin-containing granules23. This first-phase secretory response is augmented by a potassium channel-independent pathway, which is largely responsible for the second-phase insulin response. Arginine is known to stimulate insulin secretion by directly inducing membrane depolarization independent of potassium channels and thus largely independent of glucose sensing and glucose metabolism pathways24,25. In EXT heterozygotes, the first-phase insulin response and secretory response to arginine were significantly impaired, whereas insulin secretion during steady hyperglycemia (between t = 90 and t = 120) was not statistically different between groups. These findings are in contrast with our previous report in ABCA1 carriers showing decreased GSIS with an intact maximal insulin release capacity following arginine26. Indeed, MRI based pancreas volume measurements underscore the presence of an absolute diminished beta-cell mass suggesting that EXT dysfunction reflects a latent insulin deficiency due to an absolutely decreased beta cell mass.

Pathophysiological mechanisms linking EXT to beta cell mass and function After the first reports on EXT in GWAS studies performed in type 2 diabetes subjects, several studies have rendered conflicting results. Recently however, Liu et al. performed a pooled analyses using all existing GWAS data available for EXT2 gene and showed a small but significant effect (OR 1.06-1.07)10. Recent genetic studies in Drosophila and mice have provided compelling evidence that HS plays an essential role in embryonic development by interacting with several signalling molecules including Wnts, Hedgehogs, and fibroblast growth factors (FGFs)27.Our findings of normal insulin


141

8

sensitivity combined with an impaired pancreatic beta cell architecture in Ext heterozygous knockout mice thus provide a pathophysiological substrate attributing to the role for the EXT gene in the pathogenesis of type 2 diabetes mellitus. Most likely via decreased pancreas beta cell development, which corroborate with findings in conditional Extl3 knockout mice14. Previous data have suggested that HS plays an important role in organ development since HS is required as a co-receptor for FGF growth factor signalling in β-cells28. In this respect, it has been suggested that hedgehog (hh) signalling via heparansulfates is of pivotal importance in murine organ development and function in particular29,30. Moreover, it is reported that mutations in the EXT protein family in this gene lead to impaired distribution of hhs31 and hh signalling remains important throughout life in beta cell function32.

On the other hand, it has been recognized that GSIS reflects the available previously synthesized formed and stored insulin that can be secreted upon glucose stimulation33. In line, previous in vitro studies using both genetic and enzymatic approaches to induce decreased beta cell HS have all resulted in impaired GSIS14,34. In this regard, it has been reported that Wnts are involved in GSIS in adult mouse islets35. However as we found no effect on in vitro GSIS in isolated islets of Ext heterozygous mice, we believe that defective Wnt signalling does not provide an additional explanation for impaired GSIS reported in our study. In contrast, a recently published paper implicated that in mice pancreas HSPGs are involved in beta cell survival providing a buffer mechanism against reactive oxygen species (ROS)36. Thus HSPGs may have several roles in beta cell homeostasis via either regulation of postnatal islet and pancreas development14, whereas on the other hand HSPGs might protect the beta cell against destruction later in life37. Thus, the inadvertent depletion of pancreas heparansulfates in EXT heterozygous subjects might render (the already decreased amount of) beta cells vulnerable for exogenous pathogenic stimuli including obesity. Indeed, it has been previously noted that β-cell failure precedes the development of impaired glucose tolerance (IGT) in insulin resistant subjects38 due to ROS induced exhaustion of the normal beta cell capacity to adjust for increased insulin demand39.

study limitationsOur study also has several limitations. First, it should be noted that in contrast to the intronic SNPs in the EXT2 gene identified in the different GWAS studies that may have rather small effects on impaired (first phase) insulin secretion and smaller pancreas volume, the EXT1 and EXT2 genes encode homologous proteins with significant sequence identity and the majority of mutations are nonsense or frameshift, leading to complete loss of function of the protein40. Thus, further studies are needed to address whether a similar mechanism of decreased pancreas volume might be responsible for the genetic association between EXT2 and development of type 2 diabetes mellitus. Second, based on the small numbers of available carriers of loss-of-function mutations in EXT our study does not allow us to analyse the individual effects of these specific genes on beta cell function. Finally, as large well genotyped clinical cohort of HME subjects are not available, we are currently

142

Chapter 8

unable to investigate the incidence of overt type 2 diabetes in HME subjects with ranging BMI which could underscore clinical validity of our findings.

In conclusion, we now provide the first evidence on the relation between genetic defects in heparan sulfates and smaller pancreas anatomic volume with ensuing impaired betacell reserve capacity in human carriers of loss-of-function mutations in EXT. Our findings will hopefully provide novel pathophysiological clues and potential therapeutic targets to prevent betalcell failure in obese humans.

acknowledgmentsWe are grateful to all participating HME subjects and Jan de Lange from the Dutch HME Foundation (www.hme-mo.nl) for their help with inclusion and being able to perform this study.


143

8

REFERENCES1. Permutt MA, Wasson J, Cox N. Genetic epidemiology of diabetes. J. Clin. Invest. 2005;115:1431–9.

2. Sladek R, Rocheleau G, Rung J, et al. A genome-wide association study identifies novel risk loci for type 2 diabetes. Nature. 2007;445:881–5.

3. Herder C, Rathmann W, Strassburger K, et al. Variants of the PPARG, IGF2BP2, CDKAL1, HHEX, and TCF7L2 genes confer risk of type 2 diabetes independently of BMI in the German KORA studies. Horm Metab Res. 2008;40:722–6.

4. Omori S, Tanaka Y, Takahashi A, et al. Association of CDKAL1, IGF2BP2, CDKN2A/B, HHEX, SLC30A8, and KCNJ11 with susceptibility to type 2 diabetes in a Japanese population. Diabetes. 2008;57:791–5.

5. Kifagi C, Makni K, Boudawara M, et al. Association of genetic variations in TCF7L2, SLC30A8, HHEX, LOC387761, and EXT2 with Type 2 diabetes mellitus in Tunisia. Genet Test Mol Biomarkers. 2011;15:399–405.

6. Horikoshi M, Hara K, Ito C, et al. Variations in the HHEX gene are associated with increased risk of type 2 diabetes in the Japanese population. Diabetologia. 2007;50:2461–6.

7. Lewis JP, Palmer ND, Hicks PJ, et al. Association analysis in african americans of European-derived type 2 diabetes single nucleotide polymorphisms from whole-genome association studies. Diabetes. 2008;57:2220–5.

8. Wu Y, Li H, Loos RJF, et al. Common variants in CDKAL1, CDKN2A/B, IGF2BP2, SLC30A8, and HHEX/IDE genes are associated with type 2 diabetes and impaired fasting glucose in a Chinese Han population. Diabetes. 2008;57:2834–42.

9. Rong R, Hanson RL, Ortiz D, et al. Association analysis of variation in/near FTO, CDKAL1, SLC30A8, HHEX, EXT2, IGF2BP2, LOC387761, and CDKN2B with type 2 diabetes and related quantitative traits in Pima Indians. Diabetes. 2009;58:478–88.

10. Liu L, Yang X, Wang H, et al. Association between variants of EXT2 and type 2 diabetes: a replication and meta-analysis. Hum. Genet. 2013;132:139–45.

11. McCormick C, Leduc Y, Martindale D, et al. The putative tumour suppressor EXT1 alters the expres-sion of cell-surface heparan sulfate. Nat. Genet. 1998;19:158–61.

12. Lind T, Tufaro F, McCormick C, Lindahl U, Lidholt K. The putative tumor suppressors EXT1 and EXT2 are glycosyltransferases required for the biosynthesis of heparan sulfate. J. Biol. Chem. 1998;273:26265–8.

13. Bishop JR, Schuksz M, Esko JD. Heparan sulphate proteoglycans fine-tune mammalian physiology. Nature. 2007;446:1030–7.

14. Takahashi I, Noguchi N, Nata K, et al. Important role of heparan sulfate in postnatal islet growth and insulin secretion. Biochem. Biophys. Res. Commun. Elsevier Inc.; 2009;383:113–8.

15. Nadanaka S, Kitagawa H. Heparan sulphate biosynthesis and disease. J. Biochem. 2008;144:7–14.

16. Duncan G, McCormick C, Tufaro F. The link between heparan sulfate and hereditary bone dis-ease: finding a function for the EXT family of putative tumor suppressor proteins. J. Clin. Invest. 2001;108:511–6.

17. Solomon L. HEREDITARY MULTIPLE EXOSTOSIS. Am. J. Hum. Genet. 1964;16:351–63.

144

Chapter 8

18. Goud AL, De Lange J, Scholtes VAB, Bulstra SK, Ham SJ. Pain, physical and social functioning, and quality of life in individuals with multiple hereditary exostoses in The Netherlands: a national cohort study. J Bone Joint Surg Am. 2012;94:1013–20.

19. Lin X, Wei G, Shi Z, et al. Disruption of gastrulation and heparan sulfate biosynthesis in EXT1-deficient mice. Dev. Biol. 2000;224:299–311.

20. Li D-S, Yuan Y-H, Tu H-J, Liang Q-L, Dai L-J. A protocol for islet isolation from mouse pancreas. Nat Protoc. 2009;4:1649–52.

21. Stumvoll M, Mitrakou A, Pimenta W, et al. Use of the oral glucose tolerance test to assess insulin release and insulin sensitivity. Diabetes Care. 2000;23:295–301.

22. Bali MA, Metens T, Denolin V, De Maertelaer V, Devière J, Matos C. Pancreatic perfusion: noninvasive quantitative assessment with dynamic contrast-enhanced MR imaging without and with secretin stimulation in healthy volunteers--initial results. Radiology. 2008;247:115–21.

23. Straub SG, Sharp GWG. Hypothesis: one rate-limiting step controls the magnitude of both phases of glucose-stimulated insulin secretion. Am. J. Physiol., Cell Physiol. 2004;287:C565–71.

24. Fajans SS, Floyd JC, Knopf RF, et al. A difference in mechanism by which leucine and other amino acids induce insulin release. J. Clin. Endocrinol. Metab. 1967;27:1600–6.

25. Thams P, Capito K. L-arginine stimulation of glucose-induced insulin secretion through membrane depolarization and independent of nitric oxide. Eur. J. Endocrinol. 1999;140:87–93.

26. Vergeer M, Brunham LR, Koetsveld J, et al. Carriers of loss-of-function mutations in ABCA1 display pancreatic beta-cell dysfunction. Diabetes Care. 2010;33:869–74.

27. Lin X. Functions of heparan sulfate proteoglycans in cell signaling during development. Develop-ment. 2004;131:6009–21.

28. Bernfield M, Götte M, Park PW, et al. Functions of cell surface heparan sulfate proteoglycans. Annu. Rev. Biochem. 1999;68:729–77.

29. Apelqvist a, Ahlgren U, Edlund H. Sonic hedgehog directs specialised mesoderm differentiation in the intestine and pancreas. Curr. Biol. 1997;7:801–4.

30. Hebrok M. Hedgehog signaling in pancreas development. Mech. Dev. 2003;120:45–57.

31. Bellaiche Y, The I, Perrimon N. Tout-velu is a Drosophila homologue of the putative tumour suppres-sor EXT-1 and is needed for Hh diffusion. Nature. 1998;394:85–8.

32. Nybakken K, Perrimon N. Heparan sulfate proteoglycan modulation of developmental signaling in Drosophila. Biochim. Biophys. Acta. 2002;1573:280–91.

33. Straub SG, Sharp GWG. Glucose-stimulated signaling pathways in biphasic insulin secretion. Diabetes Metab. Res. Rev. 18:451–63.

34. Piquer S, Casas S, Quesada I, et al. Role of iduronate-2-sulfatase in glucose-stimulated insulin secre-tion by activation of exocytosis. Am. J. Physiol. Endocrinol. Metab. 2009;297:E793–801.

35. Fujino T, Asaba H, Kang M-J, et al. Low-density lipoprotein receptor-related protein 5 (LRP5) is es-sential for normal cholesterol metabolism and glucose-induced insulin secretion. Proc. Natl. Acad. Sci. U.S.A. 2003;100:229–34.

36. Ziolkowski AF, Popp SK, Freeman C, Parish CR, Simeonovic CJ. Heparan sulfate and heparanase play key roles in mouse β cell survival and autoimmune diabetes. J. Clin. Invest. 2012;122:132–41.


145

8

37. Raats CJ, Bakker MA, Van den Born J, Berden JH. Hydroxyl radicals depolymerize glomerular heparan sulfate in vitro and in experimental nephrotic syndrome. J. Biol. Chem. 1997;272:26734–41.

38. Kahn SE, Prigeon RL, McCulloch DK, et al. Quantification of the relationship between insulin sen-sitivity and beta-cell function in human subjects. Evidence for a hyperbolic function. Diabetes. 1993;42:1663–72.

39. Fridlyand LE, Philipson LH. Reactive species and early manifestation of insulin resistance in type 2 diabetes. Diabetes Obes Metab. 2006;8:136–45.

40. Wuyts W, Van Hul W. Molecular basis of multiple exostoses: mutations in the EXT1 and EXT2 genes. Hum. Mutat. 2000;15:220–7.

Hassing HC, Mooij H, Guo S, Monia BP, Chen K, Kulik W, Dallinga-Thie GM, Nieuwdorp M, Stroes ESG, Williams KJ

Hepatology 2012; 55: 1746-53.

9 INHIBITION OF HEPATIC SuLF2 IN vIvO: A NOvEL STRATEGY TO CORRECT DIABETIC DYSLIPIDEMIA

148

Chapter 9

ABSTRACTBackground - Type 2 diabetes mellitus (T2DM) impairs hepatic clearance of atherogenic postprandial triglyceride-rich lipoproteins (TRL). We recently reported that livers from T2DM db/db mice markedly overexpress glucosamine-6-O-endosulfatase-2 (SULF2), an enzyme that removes 6-O sulfate groups from heparan sulfate proteoglycans (HSPGs) and suppresses uptake of TRLs by cultured hepatocytes. In the present study, we evaluated whether Sulf2 inhibition in T2DM mice in vivo could correct their postprandial dyslipidemia.

Methods and results - Selective second-generation antisense oligonucleotides (ASOs) targeting Sulf2 were identified. Db/db mice were treated for 5 weeks with Sulf2 ASO (20 or 50 mg/kg per week), non-target ASO, or phosphate buffered saline (PBS). Administration of Sulf2 ASO to db/db mice suppressed hepatic Sulf2 mRNA expression by 70-80%, i.e., down to levels in non-diabetic db/m mice, and increased the ratio of tri- to di-sulfated disaccharides in hepatic HSPGs (p<0.05). Hepatocytes isolated from db/db mice on non-target ASO exhibited a significant impairment in VLDL binding that was entirely corrected in db/db mice on Sulf2 ASO. Sulf2 ASO lowered the random, non-fasting plasma triglyceride (TG) levels by 50%, achieving non-diabetic values. Most importantly, Sulf2 ASO treatment flattened the plasma TG excursions in db/db mice after corn-oil gavage (iAUC 1500±470 (mg/dL)·h for non-target ASO versus 160±40 (mg/dL)·h for Sulf2 ASO (p<0.01).

Conclusion - Despite extensive metabolic derangements in T2DM mice, inhibition of a single dysregulated molecule, SULF2, normalizes the VLDL-binding capacity of their hepatocytes and abolishes postprandial hypertriglyceridemia. These findings provide a key proof-of-concept in vivo to support Sulf2 inhibition as an attractive strategy to improve metabolic dyslipidemia.

InhIbItIon of hepatIc Sulf2 In type 2 dIabetes

149

9

INTRODUCTIONThe prevalence of type 2 diabetes mellitus (T2DM) and related syndromes is rising at an alarming pace worldwide, and the overwhelming majority of affected individuals die from accelerated atherosclerotic cardiovascular disease. 1,2 Atherosclerosis is exacerbated in these patients in large part from their characteristic dyslipidemia, which includes increased fasting levels of very low density lipoprotein (VLDL) and its major component triglyceride (TG), as well as impaired clearance of postprandial triglyceride-rich lipoprotein (TRL) remnants.3-5 Atherosclerosis arises from the subendothelial retention of these lipoproteins, and increased plasma levels of VLDL and particularly postprandial TRL-remnants have been linked to atherosclerotic cardiovascular events in human cohorts. 6-9

Unfortunately, current therapeutic strategies have shown limited success in lowering fasting or postprandial TRL concentrations as a way to reduce cardiovascular morbidity or mortality. A major step forward in atherosclerotic cardiovascular risk reduction has been achieved in T2DM by the introduction of statins, a class of medicines that lower plasma levels of LDL cholesterol. 10;11 Nonetheless, T2DM subjects treated with optimal statin therapy, exhibit considerable residual risk for cardiovascular disease, which may occur in part because statins lower TRL levels by only 10-25%. 12 Although fibrates are widely used in the treatment of hypertriglyceridemia, there is no definitive evidence that fenofibrate, when added to statin therapy, reduces the risk of coronary events in subjects with T2DM. 13;14 In addition, we lack therapeutic strategies that specifically restore postprandial remnant lipoprotein clearance to normal in T2DM.

Healthy metabolism of TRLs involves a series of steps that culminate in uptake of TRL-remnants by hepatocytes. 8;15;16 During the past decades, we 17 and others 18;19 have implicated hepatic heparan sulfate proteoglycans (HSPGs) in TRL removal, specifically, the syndecan-1 HSPG. 20-22 The syndecan-1 HSPG comprises a single-span transmembrane core protein that has three extracellular covalent attachment sites for heparan sulfate (HS), 23 which is an unbranched polysaccharide that captures lipoproteins. Roughly 50 genes are involved in HSPG assembly and disassembly, affecting core protein expression, HS side-chain length, epimerization of glucuronyl residues, and sites and extent of sulfation.24 To molecularly characterize HSPG defects in T2DM liver, we recently used a highly annotated glycomic microarray to compare hepatic expression profiles in obese, T2DM db/db (Leprdb/

db) mice versus lean, non-diabetic db/m controls. 25 Despite the complexity of HSPG biology, just one gene was identified whose dysregulation could impair syndecan-1 HSPG structure or function: the HS glucosamine-6-O-endosulfatase-2 (Sulf2). 25 This gene encodes an enzyme, SULF2, that removes 6-O sulfate groups from HSPGs. 26;27 Livers of obese T2DM mice were found to markedly overexpress SULF2, and SULF2 was shown to inhibit the catabolism of TRLs by cultured liver cells. 25 Moreover, hepatic Sulf2 mRNA expression was positively related to plasma TG levels.25 These experimental findings imply that SULF2-mediated disruption of hepatic HSPGs may contribute to impaired TRL clearance in T2DM.

150

Chapter 9

In the present study, we evaluated whether inhibition of this single overexpressed target, Sulf2, could correct the characteristic postprandial dyslipidemia of T2DM mice in vivo. To address this question, second-generation antisense oligonucleotides (ASOs) were identified that selectively inhibit hepatic Sulf2 mRNA expression. We studied the effects of Sulf2 inhibition in vivo on hepatic HSPG sulfation, binding of TRLs to isolated primary hepatocytes, and most importantly, plasma TG excursions following corn-oil gavage under diabetic conditions.

METHODSantisense oligonucleotidesAntisense therapy relies on base-pair hybridization through which ASOs selectively bind to their complementary mRNA target.28 This binding typically results in selective, catalytic degradation of the target mRNA by RNase H 29 and thereby reduces levels of the encoded protein. All ASOs used in these studies were 20 nucleotides in length and chemically modified with phosphorothioate in the backbone and 2’-O-methoxyethyl on the wings with a central deoxy gap (5-10-5 gapmer) 28. Oligonucleotides were synthesized using an Applied Biosystems 380B automated DNA synthesizer (PerkinElmer Life and Analytical Sciences-Applied Biosystems) and purified as previously described.29 To identify a potent Sulf2 ASO for experiments in mice, a series of ASOs was designed and tested in primary mouse hepatocytes for their relative abilities to suppress Sulf2 mRNA levels. From these experiments, the optimal Sulf2 ASO was selected, and its efficacy was then verified by its ability to suppress hepatic Sulf2 mRNA levels in wild-type C57BL6 mice. An oligonucleotide that is not complementary to any known murine RNA sequence was used as non-target ASO. In C57BL6 mice (Jackson Laboratory, Bar Harbor, ME, USA), Sulf2 ASO treatment for four weeks (described below) resulted in an 80% ± 3% reduction of hepatic Sulf2 mRNA levels compared to levels after administration of the non-target ASO (two-sided, unpaired Student’s t test, p<0.0001, n=4/group).

animals and oligonucleotide dosingSeven-week-old male T2DM db/db (Leprdb/db) mice and lean non-diabetic control db/m mice from the same colony on the C57BLKS background, were used (Jackson Laboratory, Bar Harbor, ME, USA). Animals were injected intraperitoneally twice weekly with Sulf2 ASO (10 or 25 mg/kg per dose, i.e., 20 or 50 mg/kg per week), non-target ASO (50 mg/kg per week), or PBS for 5 weeks. The animals were housed in micro-isolator cages on a constant 12-hour light-dark cycle with controlled temperature and humidity and were given access to food and water ad libitum (Purina LabDiet #5008). Two days after the final dose, mice were weighted, and plasma samples were taken for in-house assays of plasma glucose, insulin, and markers of liver function, as well as plasma lipids (Olympus Analyser). Plasma insulin levels were analyzed using a commercially available Elisa (Crystal Chem Inc, 90080). HOMA-IR was defined as [fasting plasma insulin (μU/mL) * fasting plasma glucose (mmol/L)] /22.5. All animal procedures were approved by the Institutional Animal Care and Use Committee.


151

9

Measurements of hepatic mRNa levelsMouse livers were homogenized in guanidine isothiocyanate solution (Invitrogen) supplemented with 8% 2-mercaptoethanol (Sigma). Total RNA was prepared using RNeasy mini kits (Qiagen) and reversed transcribed with cDNA synthesis kit (Bio-Rad). Quantitative real-time reverse transcription polymerase chain reaction (qRT-PCR) assays for Sulfs were performed using an ABI Prism 7700 sequence detector (Applied Biosystems). The sequences of primers and probe for mouse Sulf2 were: 5’-TGGACGGTGAGATATACCACGTA-3’ (forward), 5’-CAGTGCGGCTTGCTAAGGTT-3’ (reverse), and F-5’-CTTGGATACTGTGCCTCAGCCCCG-3’-Q (probe) (Integrated DNA Technologies). The primers for mouse Sulf1 were: 5’-TCATTCGTGGTCCAAGCATAGA-3’ (forward), 5’-TGGTAGGAGCTAGGTCGATGTTC-3’ (reverse) and F-5’-CCAGGGTCGATAGTCCCACAGATTGTTC-3’ (probe). 18S RNA was used to normalize gene expression, primers: 5’-Gcaattattccccatgaacg-3’ (forward) and 5’GGGACTTAATCAACGCAAGC-3’ (reverse) AND 5’-TTCCCAGT-3’ (probe).

Purification and analysis of heparan disaccharides from liverHeparan sulphate (HS) disaccharides from murine liver tissue were prepared and measured as described previously.30 Briefly, 50 mg of liver tissue was homogenized in 300 µl NH4Ac/Ca(Ac)2, pH 7, and digested by a mixture of recombinant heparinases I, II and III (5 IU each; kind gifts from Dr. Jian Liu, University of North Carolina, Chapel Hill, USA) for two hours at 37°C. Samples were heat inactivated and centrifuged (16,000 g for 5 min). The supernatant was transferred to an Amicon ultracentrifugal filter (Millipore) with a 5-kDa cut-off. The filtered samples containing heparan disaccharides were applied to an LC/MS/MS (Acquity UPLC®, Waters Inc.) and Quattro Premiere XE (Micromass) using multiple reactions monitoring in negative ion mode. Separation of HS disaccharides by LC was performed using a Hypercarb column (2.1 mm i.d. ×100 mm, 5 μm, Thermo Scientific) with a gradient elution (10 mM NH4HCO3, pH10, to 100% acetonitrile). HS disaccharides standards were purchased from Iduron (Manchester, UK). SULF2 activity in vivo was analysed as the ratio of trisulfated (D2S6) vs disulfated (D0S6 and D2S0) heparan disaccharides (see 31 for nomenclature). The two disulfated disaccharides could not be separated.

isolation and DyLight labelling of TRL fractionHuman TRLs (d < 1.006 g/ml) were isolated by density gradient ultracentrifugation (SW41 rotor; 19 h, 36,000 rpm, 10ºC) from serum obtained from fasting healthy volunteers. The TRL fraction was labelled with Dy-Light fluorophore (Amine-Reactive Fluors 488, Thermo-Scientific), which allows a high dye-to-protein ratio. The labelling was performed according to the manufacturer’s protocol.

Primary hepatocyte isolation and lipoprotein bindingFollowing administration of ASO or PBS to db/db and db/m mice, primary hepatocytes were isolated using collagenase perfusion as described previously. 32 Isolated hepatocytes were plated on Primaria multiwell plates (Becton Dickinson) using Williams’ Medium E 1x (GIBCO ® Invitrogen). After 3 hours,

152

Chapter 9

the original culture medium was replaced by serum-free Williams’ Medium E containing 1% BSA, followed by 6-hrs incubation at 37ºC. Several minutes before the binding experiments, cells were pre-chilled on ice followed by a wash with Medium E/1% BSA at 4ºC. Cells were incubated with a combination of DyLight-TRL (50 µg/ml) and bovine lipoprotein lipase (LPL 5 ug/ml; L2254, Sigma) for 30 minutes at 4°C. Cells were rinsed once with cold PBS and lysed in 200 µl RIPA buffer supplemented with protease inhibitors (Roche, Basel Switzerland). Cell lysates were collected and transferred into a black 384-well plate, and fluorescence was measured using the Cytofluor Multiwell plate Reader 4000 ( Biosystems, USA).

Postprandial fat tolerance testingA stock preparation of 1 ml corn oil (Sigma #C8267) was supplemented with 27 μCi of [11.12-3H]retinol (44.4 Ci/mmol; Perkin Elmer Life Sciences) in ethanol. Mice were fasted for 4 hours, after which each mouse received 10 μl of the corn oil/[3H]retinol mixture per gram of body weight by gastric gavage. Blood was sampled at the indicated times by submandibular bleeding. Triglycerides were measured on an Olympus Clinical Analyzer (Beckman Coulter) and [3H] was quantified by scintillation counting.

statistical analysesNormally distributed data are presented as mean ± SEM unless otherwise stated. For comparisons between a single treatment group and a control, the unpaired, two-tailed Student’s t test was used. For comparisons amongst several groups, analysis of variance (ANOVA) was initially used, followed by pairwise comparisons using the Student-Newman-Keuls q statistics. P-values less than 0.05 were considered significant. Data and graphics were analysed and constructed by GraphPad Prism, Version 5 for Windows.

RESULTSTreatment of db/db mice with sulf2 aso specifically restores hepatic expression of sulf2 to normalBy the end of the five-week treatment period, body weights, random non-fasting plasma glucose and insulin levels, and HOMA-IR values were significantly higher in PBS-treated db/db mice compared to PBS-treated db/m mice (Table 1). These parameters were not corrected by the non-target ASO or by either dose of the Sulf2 ASO in db/db mice. Markers of liver function were mildly elevated following non-target and Sulf2 ASO (Table 1). Liver, kidney and spleen histology did not show remarkable differences between saline and oligo-treated animals (data not shown).

Hepatic Sulf2 mRNA expression was strongly induced in PBS-treated db/db mice, to five times the levels in db/m mice (Figure. 1A, PBS-treated db/db vs. PBS-treated db/m, P<0.0001), consistent with our prior report. 25 Importantly, administration of Sulf2 ASO to db/db mice suppressed hepatic Sulf2 mRNA levels by 70-80%, and the higher dose restored hepatic Sulf2 mRNA to levels indistinguishable from db/m (Figure. 1A). The Sulf2 ASO had no effect on hepatic Sulf1 mRNA levels (Figure. 1B),


153

9

Table 1 - Characteristics of db/db mice following treatment with sulf2 aso

db/db db/m

PBSNT ASO

(50 mg/kg)Sulf2 ASO (20mg/kg)

Sulf2 ASO(50 mg/kg)

PBS p-value

Body weight (g) 43.9±1.0a 43.6±1.0a 47.9±1.0b 47.0±0.8b 28.6±0.5c <0.0001

Glucose (mM) 36.8±2.4a 32.6±1.8a 32.9±2.0a 29.1±2.4a 11.4±0.3b <0.0001

Insulin (μIU/ml) 172±46a,b 179±14a,b 253±63b 215±30b 56±5a <0.05

HOMA-IR 267±59a 259±28a 345±63a 249±39a 28±3b <0.0001

ALT (IU/L) 49±4a 56±2a 53±2a 75±9b 20±1c <0.0001

AST (IU/L) 127±12a 83±4b 65±3b 81±5b 58±4c <0.0001

Seven-week-old male db/db and db/m mice were given PBS, Sulf2 ASO, or non-target (NT) ASO for five weeks at indicated weekly doses. Two days after the final dose, body weight and random plasma levels of glucose, insulin, ALT and AST were measured. Displayed are means ± SEM, n=5-8 animals/group. * P-value by ANOVA; columns labelled with different lowercase letters (a,b,c) are statistically different from each other by the Student-Newman-Keuls test (p< 0.05).

indicating specificity. These data show that Sulf2 ASO effectively and selectively normalizes hepatic Sulf2 mRNA expression in db/db mice.

administration of sulf2 aso in vivo to db/db mice increases trisulfated heparan disaccharides in liver and completely restores the ability of primary hepatocytes to bind triglyceride-rich lipoproteins.To assess the effects of Sulf2 inhibition in vivo, we began by analyzing heparan sulfation in liver homogenates from db/db mice treated with Sulf2 vs. non-target ASO. Consistent with the pattern of hepatic Sulf2 expression in Figure 1, administration of low and high dose Sulf2 ASO to db/db mice significantly raised the ratio of tri- to di-sulfated heparan disaccharides in their livers, whereas the non-target ASO had no effect over PBS (0.73 ± 0.04 and 0.72 ± 0.06 vs 0.56 ± 0.08 respectively; p<0.05, Figure 2A). Next, we analysed the binding of DyLight-labeled VLDL to primary hepatocytes isolated from db/db mice following treatment with Sulf2 or non-target ASOs and from db/m mice following treatment with PBS. Compared to db/m hepatocytes, hepatocytes isolated from db/db mice after administration of the non-target ASO exhibited a significant impairment in VLDL binding that was completely corrected in hepatocytes from db/db mice following treatment with Sulf2 ASO (Figure 2B). These data collectively show that Sulf2 inhibition in vivo in db/db mice increases hepatic HSPG sulfation and restores hepatocyte binding of TRLs.

Treatment of db/db mice with sulf2 aso corrects their random non-fasting hypertriglyceridemia.Consistent with previous reports 25, PBS-treated db/db mice exhibited a significant non-fasting hypertriglyceridemia (Figure. 3A, PBS-treated db/db vs. PBS-treated db/m, P<0.05). Administration of the non-target ASO to db/db mice had no detectable effect on their non-fasting TG levels. In

154

Chapter 9

0

100

200

300

400

500

600

PBSSulf2 ASO(20 mg/kg)

Sulf2 ASO(50 mg/kg)

NT ASO(50 mg/kg)

PBS

db/db

Sulf2

db/m

a a

bc c

Nor

mal

ized

hep

atic

Sul

f2 m

RN

Ale

vels

(% re

lativ

e to

db/m

)

0

50

100

150

200Sulf1


Sulf2 ASO(50 mg/kg)

NT ASO(50 mg/kg)

PBS

db/db db/m

Nor

mal

ized

hep

atic

Sul

f1 m

RN

Ale

vels

(% re

lativ

e to

dm/m

)

A B

0

100

200

300

400

500

600


Sulf2 ASO(50 mg/kg)

NT ASO(50 mg/kg)

PBS

db/db

Sulf2

db/m

a a

bc c

Nor

mal

ized

hep

atic

Sul

f2 m

RN

Ale

vels

(% re

lativ

e to

db/m

)

0

50

100

150

200Sulf1


Sulf2 ASO(50 mg/kg)

NT ASO(50 mg/kg)

PBS

db/db db/m

Nor

mal

ized

hep

atic

Sul

f1 m

RN

Ale

vels

(% re

lativ

e to

dm/m

)

A B

Figure 1 - Treatment of db/db mice with sulf2 aso specifically restores hepatic expression of sulf2 to normal. Seven-week-old male db/m and db/db mice were given PBS, non-target (NT) ASO, or Sulf2 ASO for five weeks at indicated weekly doses (n=5-8 animals per group). Two days after the final dose, we harvested livers for RNA isolation. Levels of Sulf2 (panel A) and Sulf1 (panel B) mRNA were assessed by way of qRT-PCR, normalized to 18S RNA, and expressed relative to PBS-treated db/m mice. In panel A p < 0.0001 by ANOVA; columns labelled with different lowercase letters (a,b,c) are statistically significant different from each other by the Student-Newman-Keuls test (p< 0.05). In panel B, the p-value was not significant by ANOVA.


155

9

0.0

0.2

0.4

0.6

0.8

PBS Sulf2 ASO(20 mg/kg)

Sulf2 ASO(50 mg/kg)

NT ASO(50mg/kg)

db/db

ab b

a

Rat

io T

riS/ D

iS

A B

80

90

100

110

120

PBSNT ASO(50 mg/kg)

Sulf2 ASO(50 mg/kg)

db/mdb/db

*RFU

(%)

0.0

0.2

0.4

0.6

0.8

PBS Sulf2 ASO(20 mg/kg)

Sulf2 ASO(50 mg/kg)

NT ASO(50mg/kg)

db/db

ab b

a

Rat

io T

riS/ D

iS

A B

80

90

100

110

120

PBSNT ASO(50 mg/kg)

Sulf2 ASO(50 mg/kg)

db/mdb/db

*RFU

(%)

Figure 2 - administration of sulf2 aso in vivo to db/db mice increases trisulfated heparan disaccharides in liver and restores the ability of primary hepatocytes to bind triglyceride-rich lipoproteins. Panel A: db/db mice were given PBS or ASO as indicated for five weeks (n=4 per group). Sulfation of heparan disaccharides in liver homogenates was measured and expressed as the ratio of tri- to di-sulfated disaccharides (D2S6 vs D2S0 and D0S6 combined). P < 0.0001 by ANOVA; columns labelled with different lowercase letters (a,b) are statistically significant different from each other by the Student-Newman-Keuls test (p< 0.05). Panel B: Mice were given PBS or ASO as indicated for five weeks. Primary hepatocytes were isolated two days after the final dose (n=4 animals per group). Hepatocytes were cultured overnight at 37ºC and then incubated for 30 minutes at 4°C with DyLight-labelled VLDL (50 µg/ml) plus LPL (5 µg/ml). VLDL binding was assessed by measuring cell-associated fluorescence (RFU: relative fluorescence units). * p < 0.05 compared to PBS-treated db/m.

156

Chapter 9

0

50

100

150

200

250


Sulf2 ASO(50 mg/kg)

NT ASO(50 mg/kg)

PBS

db/db db/m

aa a

bb

Trig

lyce

rides

(mg/

dL)

0

50

100

150

200


Sulf2 ASO(50 mg/kg)

NT ASO(50 mg/kg)

PBS

db/db db/m

a a a a

b

Tota

l cho

lest

erol

(mg/

dL)

A

B

Figure 3 - Treatment of db/db mice with sulf2 aso corrects their random non-fasting hypertriglyceridemia. Plasma lipids were measured in the same mice as in Figure 1, two days after the final dose of PBS or ASO (n=5-8 per group). (A) Random, non-fasting plasma triglyceride levels. (B) Random, non-fasting plasma total cholesterol concentrations. P < 0.0001 by ANOVA; columns labelled with different lowercase letters (a,b) are statistically significant different from each other by the Student-Newman-Keuls test (p< 0.05).

contrast, the Sulf2 ASO caused a dose-dependent improvement in non-fasting hypertriglyceridemia, reaching a 50% reduction in non-fasting TG levels at the higher dose (Figure. 3A, 102 ± 8 mg/dl in db/db Sulf2 ASO 50 mg/kg vs. 171 ± 23 mg/dl in db/db non-target ASO and 212 ± 18 mg/dl in db/db PBS, p<0.05), thereby restoring this parameter to a level indistinguishable from PBS-treated db/m mice (125 ± 7 mg/dl. Fasting plasma TG levels (not shown) and non-fasting plasma total cholesterol


157

9

concentrations (Figure. 3B) were significantly higher in PBS-treated db/db mice compared to the db/m mice and were not corrected by either dose of Sulf2 ASO.

Treatment of db/db mice with sulf2 aso completely abolishes their postprandial dyslipidemia.After five weeks of treatment, db/db animals were fasted for 4h, then given a gavage of corn oil enriched with [³H]retinol. As shown in Figure 4A-B, Sulf2 ASO administration to db/db mice flattened their postprandial TG excursions. The iAUC was 1500 ± 470 (mg/dL)·h in db/db mice given the non-target ASO, which fell to just 160 ± 40 (mg/dL)·h in mice treated with the higher dose of Sulf2 ASO (Figure. 4A-B). Likewise, Sulf2 ASO lowered plasma [3H]retinol excursions by >50%, indicating a profound improvement in the clearance of chylomicron remnant particles (Figure. 4C-D).

Figure 4 - Treatment of db/db mice with sulf2 aso completely abolishes their postprandial dyslipidemia. db/db mice were given PBS or ASO as indicated for five weeks (n=4-6 animals per group). Two days after the final dose, the mice were fasted for 4h, then given a gastric gavage of corn oil enriched with [³H]retinol (10 µl corn oil per gram of body weight). ▼ = PBS; ○ = Non-target ASO; ♦ = Sulf2 ASO (20 mg/kg); = Sulf2 ASO 50 (mg/kg) (A) Postprandial excursions and (B) incremental Area Under the Curves (iAUC) of plasma triglycerides. (C) Postprandial excursions and (D) iAUC of plasma [³H]retinol concentrations. * p<0.05 compared to Sulf2 ASO (50mg/kg).

158

Chapter 9

DISCUSSIONIn the present study, we show that Sulf2 inhibition in T2DM db/db mice increases heparan sulfation, normalizes the ability of their hepatocytes to bind TRLs, substantially decreases non-fasting plasma TG levels, and abolishes postprandial hypertriglyceridemia. Thus, despite extensive, persistent metabolic derangements in these animals (Table 1), inhibition of a single overexpressed molecule, Sulf2, down to control levels normalizes the hepatic metabolism of atherogenic remnant lipoproteins. These findings provide the first proof-of-concept in vivo to support Sulf2 inhibition as an attractive strategy to improve metabolic dyslipoproteinemia. Moreover, our current results bolster the concept that diabetes dysregulates a surprisingly small number of key molecules involved in the function of hepatic syndecan-1 as a receptor for TRL remnants. 15;25;33

Following Sulf2 ASO administration, non-fasting plasma TG levels were decreased by 50%. Non-fasting TG levels closely reflect persistent postprandial TRL particles.8 Likewise, by examining plasma TG excursions following corn-oil gavage, we found robust improvement following Sulf2 ASO administration to db/db mice (>90% reduction in iAUC). The magnitude of this improvement vastly exceeds the effects on postprandial TG excursions of conventional lipid-lowering interventions, such as statins (10-15% reduction in iAUC) and fibrates (10-20% reduction in iAUC) 34-38 Unlike Sulf2 ASO, these conventional interventions fail to specifically target the key molecular derangement in T2DM liver. Although NT-ASO also produce a mild, non-specific effect, the effects of Sulf2 ASOs are most consistent, greater in magnitude and hence clearly directly related to SULF2 inhibition.

Clinical implicationsResidual atherosclerotic cardiovascular risk in T2DM patients remains substantial, even during maximal conventional treatment with currently available therapies. Recent work has implicated non-fasting TG levels, a marker of persistent remnant lipoproteins, as an independent risk factor for atherosclerotic cardiovascular disease, 7;9 but there have been no therapeutic strategies that selectively target persistent postprandial remnants. Our present findings demonstrate that hepatic Sulf2 inhibition in vivo corrects postprandial dyslipidemia in T2DM mice. Translation of these findings to the clinic will benefit from the relative maturity of ASO technology. In other circumstances, ASO administration has been selective and effective against hepatic targets.39;40 Importantly, ASOs have been reported to be safe and effective during short-term administration to humans.41 In our system, the Sulf2 ASO lowered abnormally high levels of Sulf2 mRNA in T2DM mouse livers to normal, but not below normal, which is highly desirable from the standpoint of safety. In conclusion, our work provides a key proof-of-concept in vivo for a novel therapeutic approach to improve metabolic dyslipidemia through restoration of hepatic HSPG function in diabetes.


159

9

acknowledgementsWe greatly acknowledge J.A. Sierts for excellent laboratory assistance as well as H. van Lenthe and L. IJlst for technical assistance and expertise in the disaccharide analysis.

DisclosuresE.S. Stroes has received consultancy fee for apoB antisense program from ISIS. B. Monia and S. Guo are employees of ISIS Pharmaceuticals

160

Chapter 9

REFERENCE LIST1. Haffner SM, Lehto S, Ronnemaa T, Pyorala K, Laakso M: Mortality from coronary heart disease in

subjects with type 2 diabetes and in nondiabetic subjects with and without prior myocardial infarc-tion. N Engl J Med 1998;339:229-34.

2. Buse JB, Ginsberg HN, Bakris GL, Clark NG, Costa F, Eckel R, Fonseca V, Gerstein HC, Grundy S, Nesto RW, Pignone MP, Plutzky J, Porte D, Redberg R, Stitzel KF, Stone NJ: Primary prevention of cardio-vascular diseases in people with diabetes mellitus: a scientific statement from the American Heart Association and the American Diabetes Association. Diabetes Care 2007;30:162-72.

3. Moreton JR: Atherosclerosis and Alimentary Hyperlipemia. Science 1947;106:190-91.

4. Taskinen MR: Diabetic dyslipidaemia: from basic research to clinical practice. Diabetologia 2003;46:733-49.

5. Chahil TJ, Ginsberg HN: Diabetic dyslipidemia. Endocrinol Metab Clin North Am 2006;35:491-98.

6. Williams KJ, Tabas I: Lipoprotein retention--and clues for atheroma regression. Arterioscler Thromb Vasc Biol 2005;25:1536-1540.

7. Bansal S, Buring JE, Rifai N, Mora S, Sacks FM, Ridker PM: Fasting compared with nonfasting triglyc-erides and risk of cardiovascular events in women. JAMA 2007;298:309-16.

8. Williams KJ: Molecular processes that handle -- and mishandle -- dietary lipids. J Clin Invest 2008;118:3247-59.

9. Freiberg JJ, Tybjaerg-Hansen A, Jensen JS, Nordestgaard BG: Nonfasting triglycerides and risk of is-chemic stroke in the general population. JAMA 2008;300:2142-52.

10. Collins R, Armitage J, Parish S, Sleigh P, Peto R: MRC/BHF Heart Protection Study of cholesterol-low-ering with simvastatin in 5963 people with diabetes: a randomised placebo-controlled trial. Lancet 2003;361:2005-16.

11. Shepherd J, Barter P, Carmena R, Deedwania P, Fruchart JC, Haffner S, Hsia J, Breazna A, LaRosa J, Grundy S, Waters D: Effect of lowering LDL cholesterol substantially below currently recommended levels in patients with coronary heart disease and diabetes: the Treating to New Targets (TNT) study. Diabetes Care 2006;29:1220-26.

12. Jones PH, Davidson MH, Stein EA, Bays HE, McKenney JM, Miller E, Cain VA, Blasetto JW: Comparison of the efficacy and safety of rosuvastatin versus atorvastatin, simvastatin, and pravastatin across doses (STELLAR* Trial). Am J Cardiol 2003;92:152-60.

13. Keech A, Simes RJ, Barter P, Best J, Scott R, Taskinen MR, Forder P, Pillai A, Davis T, Glasziou P, Drury P, Kesaniemi YA, Sullivan D, Hunt D, Colman P, d’Emden M, Whiting M, Ehnholm C, Laakso M: Effects of long-term fenofibrate therapy on cardiovascular events in 9795 people with type 2 diabetes mellitus (the FIELD study): randomised controlled trial. Lancet 2005;366:1849-61.

14. Ginsberg HN, Elam MB, Lovato LC, Crouse JR, III, Leiter LA, Linz P, Friedewald WT, Buse JB, Gerstein HC, Probstfield J, Grimm RH, Ismail-Beigi F, Bigger JT, Goff DC, Jr., Cushman WC, Simons-Morton DG, Byington RP: Effects of combination lipid therapy in type 2 diabetes mellitus. N Engl J Med 2010;362:1563-74.

15. Williams KJ, Chen K: Recent insights into factors affecting remnant lipoprotein uptake. Curr Opin Lipidol 2010;21:218-28.


161

9

16. Mahley RW, Huang Y: Atherogenic remnant lipoproteins: role for proteoglycans in trapping, transfer-ring, and internalizing. J Clin Invest 2007;117:94-98.

17. Williams KJ, Fless GM, Petrie KA, Snyder ML, Brocia RW, Swenson TL: Mechanisms by which lipopro-tein lipase alters cellular metabolism of lipoprotein(a), low density lipoprotein, and nascent lipopro-teins. Roles for low density lipoprotein receptors and heparan sulfate proteoglycans. J Biol Chem 1992;267:13284-92.

18. Ji ZS, Sanan DA, Mahley RW: Intravenous heparinase inhibits remnant lipoprotein clearance from the plasma and uptake by the liver: in vivo role of heparan sulfate proteoglycans. J Lipid Res 1995;36:583-92.

19. MacArthur JM, Bishop JR, Stanford KI, Wang L, Bensadoun A, Witztum JL, Esko JD: Liver heparan sul-fate proteoglycans mediate clearance of triglyceride-rich lipoproteins independently of LDL receptor family members. J Clin Invest 2007;117:153-64.

20. Fuki IV, Kuhn KM, Lomazov IR, Rothman VL, Tuszynski GP, Iozzo RV, Swenson TL, Fisher EA, Williams KJ: The syndecan family of proteoglycans. Novel receptors mediating internalization of atherogenic lipoproteins in vitro. J Clin Invest 1997;100:1611-22.

21. Fuki IV, Meyer ME, Williams KJ: Transmembrane and cytoplasmic domains of syndecan medi-ate a multi-step endocytic pathway involving detergent-insoluble membrane rafts. Biochem J 2000;351:607-12.

22. Stanford KI, Bishop JR, Foley EM, Gonzales JC, Niesman IR, Witztum JL, Esko JD: Syndecan-1 is the primary heparan sulfate proteoglycan mediating hepatic clearance of triglyceride-rich lipoproteins in mice. J Clin Invest 2009;119:3236-45.

23. Bernfield M, Kokenyesi R, Kato M, Hinkes MT, Spring J, Gallo RL, Lose EJ: Biology of the syndecans: a family of transmembrane heparan sulfate proteoglycans. Annu Rev Cell Biol 1992;8:365-93.

24. Bishop JR, Schuksz M, Esko JD: Heparan sulphate proteoglycans fine-tune mammalian physiology. Nature 2007;446:1030-37.

25. Chen K, Liu ML, Schaffer L, Li M, Boden G, Wu X, Williams KJ: Type 2 diabetes in mice induces hepatic overexpression of sulfatase 2, a novel factor that suppresses uptake of remnant lipoproteins. Hepa-tology 2010;52:1957-67.

26. Morimoto-Tomita M, Uchimura K, Werb Z, Hemmerich S, Rosen SD: Cloning and characteriza-tion of two extracellular heparin-degrading endosulfatases in mice and humans. J Biol Chem 2002;277:49175-85.

27. Rosen SD, Lemjabbar-Alaoui H: Sulf-2: an extracellular modulator of cell signaling and a cancer target candidate. Expert Opin Ther Targets 2010;14:935-49.

28. Bennett CF, Swayze EE: RNA targeting therapeutics: molecular mechanisms of antisense oligonucleo-tides as a therapeutic platform. Annu Rev Pharmacol Toxicol 2010;50:259-93.

29. Baker BF, Lot SS, Condon TP, Cheng-Flournoy S, Lesnik EA, Sasmor HM, Bennett CF: 2’-O-(2-Methoxy)ethyl-modified anti-intercellular adhesion molecule 1 (ICAM-1) oligonucleotides selectively increase the ICAM-1 mRNA level and inhibit formation of the ICAM-1 translation initiation complex in human umbilical vein endothelial cells. J Biol Chem 1997;272:11994-000.

162

Chapter 9

30. Oguma T, Tomatsu S, Montano AM, Okazaki O: Analytical method for the determination of disac-charides derived from keratan, heparan, and dermatan sulfates in human serum and plasma by high-performance liquid chromatography/turbo ionspray ionization tandem mass spectrometry. Anal Biochem 2007;368:79-86.

31. Lawrence R, Lu H, Rosenberg RD, Esko JD, Zhang L: Disaccharide structure code for the easy repre-sentation of constituent oligosaccharides from glycosaminoglycans. Nat Methods 2008;5:291-92.

32. Quistorff B, Dich J, Grunnet N: Preparation of isolated rat liver hepatocytes; in Pollard JW, Walker JM, (eds): Methods in Molecular Biology: Animal Cell Culture. Totowa, N.J., Humana Press, 1989, vol 5, pp 151-60.

33. Williams KJ, Liu ML, Zhu Y, Xu X, Davidson WR, McCue P, Sharma K: Loss of heparan N-sulfotrans-ferase in diabetic liver: role of angiotensin II. Diabetes 2005;54:1116-22.

34. Lee SH, Park S, Kang SM, Jang Y, Chung N, Choi D: Effect of Atorvastatin Monotherapy and Low-Dose Atorvastatin/Ezetimibe Combination on Fasting and Postprandial Triglycerides in Combined Hyperli-pedemia. J Cardiovasc Pharmacol Ther 2011; doi 10.1177/1074248411399762.

35. Castro Cabezas M, Erkelens DW, Kock LA, De Bruin TW: Postprandial apolipoprotein B100 and B48 metabolism in familial combined hyperlipidaemia before and after reduction of fasting plasma tri-glycerides. Eur J Clin Invest 1994;24:669-78.

36. van Wijk JP, Buirma R, van TA, Halkes CJ, De Jaegere PP, Plokker HW, van der Helm YJ, Castro Cabezas M: Effects of increasing doses of simvastatin on fasting lipoprotein subfractions, and the effect of high-dose simvastatin on postprandial chylomicron remnant clearance in normotriglyceridemic pa-tients with premature coronary sclerosis. Atherosclerosis 2005;178:147-55.

37. Twickler TB, Dallinga-Thie GM, de Valk HW, Schreuder PC, Jansen H, Castro Cabezas M, Erkelens DW: High dose of simvastatin normalizes postprandial remnant-like particle response in patients with heterozygous familial hypercholesterolemia. Arterioscler Thromb Vasc Biol 2000;20:2422-27.

38. Wilmink HW, Twickler MB, Banga JD, Dallinga-Thie GM, Eeltink H, Erkelens DW, Rabelink TJ, Stroes ES: Effect of statin versus fibrate on postprandial endothelial dysfunction: role of remnant-like parti-cles. Cardiovasc Res 2001;50:577-82.

39. Butler M, McKay RA, Popoff IJ, Gaarde WA, Witchell D, Murray SF, Dean NM, Bhanot S, Monia BP: Spe-cific inhibition of PTEN expression reverses hyperglycemia in diabetic mice. Diabetes 2002;51:1028-34.

40. Zhang H, Lowenberg EC, Crosby JR, MacLeod AR, Zhao C, Gao D, Black C, Revenko AS, Meijers JC, Stroes ES, Levi M, Monia BP: Inhibition of the intrinsic coagulation pathway factor XI by antisense oli-gonucleotides: a novel antithrombotic strategy with lowered bleeding risk. Blood 2010;116:4684-92.

41. Kastelein JJ, Wedel MK, Baker BF, Su J, Bradley JD, Yu RZ, Chuang E, Graham MJ, Crooke RM: Potent reduction of apolipoprotein B and low-density lipoprotein cholesterol by short-term administration of an antisense inhibitor of apolipoprotein B. Circulation 2006;114:1729-35.

Hassing HC, Mooij HL, Bernelot Moens SJ, Surendran RP, Verrijken A, Williams KJ, ’t Hart LM, Nijpels G., van Gaal L, Staels B, Dekker JM, Stroes ESG, Dallinga-Thie GM, Nieuwdorp M

Submitted

10 A GENETIC vARIANT AT THE SuLF2 LOCUS ASSOCIATES wITH POSTPRANDIAL TRIGLYCERIDES IN PATIENTS wITH TYPE 2 DIABETES MELLITUS

166

Chapter 10

ABSTRACTintroduction - Type 2 diabetes mellitus (T2DM) is characterized by elevated plasma levels of atherogenic triglyceride-rich lipoproteins (TRL). Recent data revealed that hepatic heparansulfate proteoglycans (HSPGs) contribute to the clearance of TRL. Hepatic glucosamine-6-O-endosulfatase-2 (SULF2), a HSPG remodelling enzyme, is strongly associated with triglyceride levels in db/db mice, however data in humans are lacking. We thus investigated the role of SULF2 on TRL metabolism in patients with T2DM.

Methods - Human liver biopsies were analyzed for relation between SULF2 expression and triglycerides. Moreover, associations between seven SULF2 tagging single nucleotide polymorphisms (SNPs) and plasma TG levels were determined in two cohorts of T2DM subjects (combined n=1,540) with diabetic dyslipidemia. Postprandial TRL clearance was evaluated following an oral fat tolerance test in T2DM subjects (n=29) stratified by SULF2 genotypes.

Results - Liver SULF2 expression was significantly associated with fasting plasma triglycerides (r = 0.271; p=0.003). Genetic variation at the SULF2 locus is reproducible associated with lower fasting plasma TG levels in T2DM subjects with metabolic dyslipidemia (p<0.05). Carrierschip of the SULF2 rs2281279 minor G allele was associated with lower levels of postprandial plasma TG and RE levels in a stepwise manner (TG AUC: AA carriers: 29±10, AG Carriers: 21±7 and GG carriers: 17+8 [mmol/L]*h respectively, p<0.05; retinylesters AUC AA carriers 101±43, AG carriers 63±40 and GG carriers 15± 6 [mg/l]*h respectively, p<0.001).

Conclusion - A genetic variant in SULF2 rs2281279 predisposes to lower fasting and postprandial TRL levels in T2DM patients. These findings may implicate SULF2 as therapeutic target in diabetic dyslipidemia.

Genetic variation in SULF2 in type 2 diabetes

167

10

INTRODUCTIONThe prevalence of type 2 diabetes mellitus (T2DM) and its sequelae, including cardiovascular events, is increasing worldwide. (1, 2) The accelerated risk for atherosclerosis in these subjects results from their atherogenic dyslipidemia profile, which includes increased fasting levels of very low density lipoprotein (VLDL) and its major component triglyceride (TG) in combination with impaired clearance of postprandial triglyceride-rich lipoprotein (TRL) remnants. (3-5) In particular, subendothelial retention of VLDL and particularly postprandial TRL-remnants have been linked to cardiovascular events in these individuals. (6-9)

TRL metabolism involves a series of steps that culminate in the uptake of TRL-remnants by hepatocytes. (8, 10, 11) During the past decades, hepatic heparan sulfate proteoglycans (HSPGs) (12-14) have been implicated in TRL removal. (15-17) HSPGs are cell membrane bound core proteins to which 2-3 sulfated HS chains are attached. They can bind various ligands including TRL resulting in hepatic uptake and clearance of these particles. About 50 genes are involved in HSPG assembly and disassembly, affecting core protein expression, HS side-chain length, epimerization of glucuronyl residues or sulfation patterns. (18, 19) It was recently shown that hepatic HS glucosamine-6-O-endosulfatase-2 (Sulf2 encoding for sulfatase 2; a HSPG degrading enzyme) expression was increased in db/db mice. In these mice, hepatic Sulf2 mRNA expression was strongly and positively related to plasma TG levels. (20) Recently, we showed that inhibition of hepatic Sulf2 expression by targeted allele-specific antisense administration resulted in normalization of plasma TG levels and reduced postprandial hypertriglyceridemia.(21) In addition, Genome-Wide Association Studies (GWAS) have associated the SULF2 locus (20q13.1) with T2DM.(22) However, human data on the biology of hepatic SULF2 and its role in triglyceride metabolism remain to be elucidated.

In the present study, we assessed the expression of SULF2 in human liver tissue and additionally evaluated whether genetic variation at the human SULF2 locus is associated with postprandial TG clearance in subjects with type 2 diabetes mellitus. We used a population-based approach, which was complemented with an oral fat tolerance test in a subgroup of patients selected for the presence of genetic variation in SULF2.

RESEARCH DESIGN AND METHODSStudies were approved by Institutional Review Boards in accordance with the Declaration of Helsinki. Written informed consent was obtained from all subjects. The study was registered at the Dutch Trial Register (NTR 2641).

Hepatic suLF2 expressionHuman liver biopsies of T2DM subjects were analysed for SULF2 expression and results were validated in the Antwerp Human Liver Cohort.

168

Chapter 10

Human liver specimen (n=7 T2DM patients, not using insulin at time of biopsy, and n=7 controls) was obtained via an open liver biopsy during abdominal surgery. Biopsy material was frozen instantaneously in liquid nitrogen and stored at -800C. Hepatic tissue was homogenized in TriZol using Magna Lyzer beads in the Magna Lyser (Roche, USA). RNA was isolated according to manufacturer’s instructions. Quality of RNA was verified with an agilent Bioanalyzer (Agilent Technologies, Palo Alto, CA, USA). We used the human GLYCOv4 oligonucleotide array, a custom AffymetrixGeneChip designed for the Consortium for Functional Glycomics (23) (see http://www.functionalglycomics.org/static/consortium/resources.shtml ). Hybridisation and scanning were performed at Scipps Institute, San Diego USA using the GeneChip Scanner 3000 (Affymetrix, Santa Clara, USA). Arrays had a background of <50 intensity units and a glyceraldehyde 3-phosphate dehydrogenase (GAPDH) 3’/5’ ratio below 2.0.

Human Liver Replication Cohort (antwerp Cohort):

Liver biopsies (n=121) were obtained from patients visiting the obesity clinic of the Antwerp University Hospital, as recently described. (24) Liver material was immediately snap-frozen and stored at -80oC for RNA isolation. RNA was isolated from human liver by guanidinium thiocyanate/phenol/chloroform extraction. (25) Reverse transcription was performed using the High Capacity Reverse Transcription kit (Applied Biosystems, Life Technologies, Carlsbad, USA). PCR was performed with Brilliant II SYBR Green QPCR Master Mix (Agilent Technologies, Santa Clara, USA) on a Stratagene Mx3005P system (Agilent Technologies) using specific primers. mRNA levels were subsequently normalized to those of cyclophilin and fold induction was calculated using the ∆Ct method. (26) SULF2 RT-PCR primers were designed using Primer3 software (5’-CCT TTG CCG TGT ACC TCA AT-3’ and 5’-GCA CGT AGG AGC CGT TGT AT-3’). mRNA levels were normalized to those of transcription factor IIb (TFIIb) and was calculated using the ΔCt method.

suLF2 sNPs and plasma triglycerides in T2DM patients Associations between SULF2 SNPs and plasma triglyceride levels were assessed in the Diabetes Atorvastatin Lipid Intervention (DALI) study and subsequently validated in the Diabetes Care System.

Diabetes atorvastatin Lipid intervention (DaLi) studyThis double-blind randomized placebo-controlled multicenter study evaluated the effect of atorvastatin 10 mg versus 80 mg on lipid metabolism in 217 unrelated Dutch men and woman with T2DM. and diabetic dyslipidemia (fasting triglyceride levels between 1.5 and 6.0 mmol/l) (27)

Diabetes Care SystemPatients were included from this prospective dynamic patient cohort of patients with T2DM followed since 1998 in the Netherlands. (28) For the present study we selected 1,323 T2DM patients who had agreed upon participation in genetic analyses, an elevated fasting plasma triglyceride levels (≥1.7 mmol/l or ≥ 150 mg/dL) and the availibility of a DNA sample.


169

10

GenotypingTagging SNPs for SULF2 were selected with the HapMap database (29) and the TAGGER algorithm using a pairwise tagging approach. (30) Briefly, we selected from HapMap all common genetic variants (minor allele frequency >0.1) in the SULF2 locus in a population of European ancestry. A tagging SNP approach uses the knowledge of associations between genetic variants (linkage disequilibrium [LD] structure) to limit the number of SNPs that needs to be genotyped. TagSNPs are those SNPs that most effectively represent (or “tag”) all the SNPs in a particular locus. We selected tagSNPs using an r2 cutoff level >0.8 (For the genomic context of SULF2 see supplementary Figure 1).

Oral fat tolerance testIn a randomly selected subset of T2DM subjects, stratified by SULF2 genotype (rs2281279 c.2494+267 A>G) AA; n=11, AG; n=11 GG; n=7, an oral fat tolerance test was performed to study postprandial triglyceride metabolism. Selection was performed based upon the following inclusion criteria: male or post-menopausal female patients with type 2 diabetes mellitus, aged 45 to 75 years, without a history of manifest coronary artery disease and HbA1c <10% (<86 mmol/mol). The lipid inclusion criteria were: total cholesterol (TC) between 4.0 and 8.0 mmol/L (155 – 309 mg/dL) and fasting triglycerides (TG) < 6.0 mmol/L (<532 mg/dL). Exclusion criteria’s for participation in the postprandial sub-study were the use of any lipid lowering medication in 8 weeks preceding the fat tolerance test, clinical signs of malabsorption (eg diarrhoea) or exogenous insulin treatment. Participants were asked to refrain from alcohol intake the day before. Participants were admitted at 7:30 am after an overnight fast. Cream [consisting of 40% fat (wt/vol) with a polyunsaturated fat to saturated fat ratio of 0.06, 0.001% cholesterol (wt/vol), and carbohydrates were administered in a dose of 35 gram fat and 50 gram carbohydrates per m2 body surface. This mixture was supplemented with 150 ml water and 100,000 IU of vitamin A (Retinyl palmitate, AMC Clinical Pharmacy). The cream drink was consumed within 10 minutes. Postprandial blood samples were drawn at 0, 1, 2, 3, 4, 5 and 7 hours. Venous blood was collected into EDTA containing tubes, which were placed on ice and protected from light. Plasma was separated within 30 minutes by centrifugation at 3000 rpm for 20 min. at 4°C. Aliquots of plasma were frozen at -800C for subsequent analysis of TG and retinyl esters (RE).

Biochemical analysesTotal cholesterol, HDL-cholesterol, LDL-cholesterol and triglycerides were measured by standard enzymatic methods on a Cobas Mira system (Roche Diagnostics, Basel, Switzerland). Glucose was assessed using the hexokinase method (Gluco-quant, Hitachi 917; Hitachi). Plasma insulin was measured by an immunoluminimetric assay (Immulite insuline) on Immulite 2000 (Diagnostic Products). HbA1C was measured by HPLC (Reagens Bio-Rad Laboratories, Veenendaal, the Netherlands) on a Variant II (Bio-Rad Laboratories). Homeostatic model assessment for insulin resistance (HOMA-IR) was calculated by the following formula: [glucose (mmol/L) * insulin (mU/L)]/22.5. Retinyl esters (RE) were analysed in 200 μl plasma after extraction of retinyl esters using chloroform/methanol/water as described.(31, 32) In short, retinyl propionate (Sigma

170

Chapter 10

Chemicals; St Louis, USA) was used as internal standard; methanol was used as mobile phase at a flow rate of 1 ml/min and the effluent is monitored at 330 nm. A standard curve of retinyl palmitate in pooled plasma was used as reference. Peak heights were measured and used for calculations of the absolute RE values.

statistical analysesRaw expression values from the microarrays were normalized using the Robust Multi- chip Average expression summary (http://rmaexpress. bmbolstad.com). Differential expression was determined using the open-source, open-development Limma package in the R statistical package, available under the terms of the Free Software Foundation’s GNU General Public License. Comparisons between two or more experimental conditions were made by calculating the fold change in expression levels and the adjusted P value. The adjusted P value for the GLYCOv4 array corrects for multiple testing using Benjamini and Hoch- berg’s method to control the false discovery rate.

Clinical parameters are expressed as mean ± standard deviation unless stated otherwise. Effects of single SNPs on triglyceride levels were examined using one way analysis of variances (ANOVA). Linear regression analysis was used to correct for co-variables. Variables with skewed distribution were log-transformed before being used as continuous variables in statistical analyses. Effects of SNPs on baseline characteristics were analysed by ANOVA for continuous variables and with chi-square for categorical variables. Postprandial TG and RE were calculated as total area under the curve (AUC) calculated by the trapezoid rule. Area under the incremental curve (iAUC) was obtained by subtracting the fasting plasma level from each postprandial time point. Comparisons between groups were analysed using one-way ANOVA. Two-sided probability values of less than 0.05 were considered statistical significant. Statistical analyses were performed using SPSS version 18.0.

RESULTSHuman hepatic suLF2 expressionClinical data of both T2DM subjects and controls from whom liver tissue was obtained are depicted in supplementary Table 1. As expected, patients with T2DM had a higher BMI, higher fasting glucose levels, higher HbA1c and higher fasting triglyceride levels compared to non-diabetic controls. Human GLYCOv4 oligonucleotide array analyses revealed a trend towards significant higher hepatic SULF2 mRNA expression in diabetics versus controls (8.3 ± 0.4 versus 6.9 ± 0.9, unadjusted p<0.001, p=0.06 adjusted; see Figure 1). Replication of this finding was performed in a large consecutively recruited cohort of Caucasian obese patients (Antwerp Cohort) undergoing a liver biopsy before bariatric surgery. Main characteristics are presented in supplementary Table 2. Patients were characterized by similar BMI and fasting plasma triglyceride levels but lower HbA1c when compared to our smaller DM2 liverbiopsy cohort. Liver SULF2 expression was significantly associated with fasting plasma triglycerides (r = 0.271; p=0.003), fasting glucose (r = 0.252; p=0.005) and with HOMA-IR (r = 0.186; p=0.043).


171

10

Controls Diabetes0

2

4

6

8

10 *

Rel

ativ

e SU

LF2

expr

essi

on (g

lyco

chip

)

Figure 1 - suLF2 expression in human liver samples. Human hepatic SULF2 expression in diabetes versus controls (n=7 versus n=7) measured by glycochip (AffymetrixGeneChip designed for the Consortium for Functional Glycomics). ** p=0.001 unadjusted, p=0.06 adjusted.

Genetic variants in suLF2To investigate whether genetic variation in SULF2 is associated with plasma TG levels, we selected seven tagSNPs (rs2281279 [c.2494+267A>G), rs6090717 [c.416-3507T>C], rs6090714 [c.567+4691G>A], rs6094818 [-100-7141T>C], rs13044051 [c.416-8793G>C], rs6122615 [c.101+1307T>C] and rs2235734 [2227+48T>G]). All r2 values were below 0.8 confirming their non-redundant status as tagSNPs. All SNPs were in Hardy-Weinberg equilibrium (LD plots are given in supplementary Figure 2).

Common genetic variants in suLF2 and triglyceride levelsRelationships between SULF2 tagSNPs and plasma TG levels in the DALI study and Diabetes Care System are shown in Table 1. We observed a significant association between fasting plasma TG levels and SULF2 SNP rs2281279 in both the DALI study (homozygous carriers of the major allele (AA; n=90) TG 2.63 mmol/l (IQR 2.00-3.51), heterozygous (AG; n=95) TG 2.55 mmol/l (IQR 2.09-3.52 mmol/l) and homozygote carriers of the minor allele (GG; n=25) TG 2.20 mmol/l (IQR 1.89-2.83 mmol/l, p=0.031) and the Diabetes care System (homozygous carriers of the major allele (AA; n= 623) TG 2.30 mmol/l (IQR 1.90-2.83), heterozygous (AG; n=546) TG 2.40 mmol/l (IQR 2.00-3.03 mmol/l) and homozygote carriers of the minor allele (GG; n= 120) TG 2.19 mmol/l (IQR 1.90-3.00 mmol/l, p=0.049). Associations between SULF2 rs2281279 genotypes and other clinical parameters in the Diabetes Care System are presented in supplemental Table 3.

Genetic variation in suLF2 rs2281279 and postprandial responsesSince carriership of the rs2281279 GG allele displayed a significant association with lower fasting TG levels in patients with T2DM and mild hypertriglyceridemia, we selected 29 T2DM patients stratified

172

Chapter 10

Figure 2 - Postprandial triglyceride-rich lipoproteins by suLF2 genotype. Postprandial triglyceride curves presented as mean ± S.E.(A), AUC of triglycerides (B) and iAUC (C) of triglycerides as well as postprandial retinyl ester (RE) curves (D) and AUC of RE (E) according to rs2281279 genotypes AA, AG and GG.


173

10

Tabl

e 1

- ass

ocia

tion

of s

uLF

2 ta

gsN

Ps w

ith tr

igly

cerid

e le

vels

DA

LI

Di

abet

es C

are

Syst

em

SNP

(n=2

17)

p-va

lue

p-va

lue*

(n=1

323)

p-va

lue

p-va

lue*

G

enot

ypes

Gen

otyp

es

rs22

8127

9AA

AGGG

AAAG

GG

N90

9525

629

557

122

Trig

lyce

rides

2.63

(2.0

0-3.

51)

2.55

(2.0

9-3.

52)

2.20

(1.8

9-2.

83)

0.03

1†0.

048†

2.30

(1.9

0-2.

83)

2.40

(2.0

0-3.

03)

2.19

(1.9

0-3.

00)

0.04

9†0.

463

rs60

9071

7TT

TCCC

TTTC

CC

N71

9428

433

641

235

Trig

lyce

rides

2.60

(2.0

3-3.

57)

2.37

(2.0

0-2.

95)

3.09

(2.1

9-4.

24)

0.13

80.

087

2.30

(1.9

0-2.

90)

2.30

(1.9

0-2.

90)

2.30

(1.9

0-3.

00)

0.63

50.

161

rs60

9071

4GG

GAAA

GGGA

AA

N57

102

4637

666

026

9

Trig

lyce

rides

2.72

(2.1

0-3.

68)

2.40

(1.9

0-3.

13)

2.60

(2.0

3-3.

63)

0.93

30.

781

2.20

(1.9

0-2.

83)

2.37

(1.9

0-3.

00)

2.39

(2.0

0-3.

01)

0.22

60.

037

rs60

9481

8TT

TCCC

TTTC

CC

N12

256

1684

935

798

Trig

lyce

rides

2.57

(2.0

6-3.

43)

2.29

(1.9

0-3.

35)

2.66

(2.2

3-3.

67)

0.42

80.

412

2.30

(1.9

0-2.

90)

2.30

(1.9

7-3.

00)

2.20

(1.9

0-2.

77)

0.59

60.

295

174

Chapter 10

Tabl

e 1

- Con

tinue

d

DA

LI

Di

abet

es C

are

Syst

em

SNP

(n=2

17)

p-va

lue

p-va

lue*

(n=1

323)

p-va

lue

p-va

lue*

G

enot

ypes

Gen

otyp

es

rs13

0440

51GG

GCCC

GGGC

CC

N13

165

1378

647

450

Trig

lyce

rides

2.50

(2.0

6-3.

40)

2.55

(2.0

0-3.

62)

2.86

(2.2

4-3.

64)

0.46

50.

547

2.30

(1.9

0-3.

00)

2.35

(1.9

0-2.

90)

2.50

(1.9

0-3.

20)

0.97

10.

924

rs61

2261

5TT

TCCC

TTTC

CC

N14

160

785

741

046

Trig

lyce

rides

2.60

(2.0

1-3.

41)

2.35

(1.9

9-3.

58)

3.20

(2.3

0-4.

25)

0.37

30,

379

2.30

(1.9

0-2.

90)

2.34

(2.0

0-2.

90)

2.20

(1.9

0-2.

60)

0.43

50.

143

rs22

3573

4TT

TGGG

TTTG

GG

N81

9727

559

586

167

Trig

lyce

rides

2.60

(1.9

2-3.

58)

2.61

(2.0

9-3.

50)

2.29

(1.7

0-2.

83)

0.07

10.

052

2.30

(1.9

0-2.

90)

2.30

(1.9

0-2.

90)

2.40

(2.0

0-3.

10)

0.63

90.

925

Valu

es a

re g

iven

as

num

ber (

n) o

r med

ian

(inte

r qua

rtile

rang

e). T

rigly

cerid

es a

re p

rese

nted

as

mm

ol/L

. *P-

valu

es c

orre

cted

for a

ge, g

ende

r, BM

I, gl

ucos

e an

d Hb

A1c.

† not s

igni

fican

t fol

low

ing

Bonf

erro

ni p

ost h

oc te

st.


175

10

according to their rs2281279 genotype (AA; n=11, AG; n=11 and GG; n=7) for participation in the oral fat tolerance test to evaluate the effect of SULF2 rs2281279 G allele carriership on postprandial lipid metabolism. Baseline characteristics of T2DM patients participating in the oral fat tolerance test are listed in Table 2. Patients with homozygous carriership of the minor allele (GG) generally were older, had lower fasting glucose, HbA1c and fasting TG levels and had a lower diastolic blood pressure. HOMA-IR values as marker of insulin resistance were comparable between the three genotype groups. Following the oral fat tolerance test, postprandial TG levels were lower in AG carriers and GG carriers compared to AA carriers (see Figure 2A). The area under the curve (AUC) of plasma TG was 29±10 (mmol/l)*h in AA carriers, 21±7 (mmol/l)*h in AG carriers and 17±8 (mmol/l)*h in GG carriers (Figure 2B, p= <0.05). After adjustment for baseline TG levels, the incremental AUC (iAUC) remained significant smaller for both AG and GG carriers [TG iAUC of 6.4±3.8 (mmol/l)*h in AA, 2.9±2.9 (mmol/l)*h in AG and 4.1±1.9 (mmol/l)*h in GG, p= <0.05, Figure 2C]. With respect to postprandial plasma RE excursions, RE- AUCs were significantly decreased in both AG and GG carriers [101±43 (mg/l)*h for AA, 63±40 (mg/l)*h for AG and 15± 6 (mg/l)*h for GG, p<0.001, Figure 2D-E] again eluting to the improved TG clearance capacity of G allele carriers.

Table 2 - baseline characteristics of patients participating in the oral fat tolerance test.

rs2281279 genotypes

P-valueAA(n=11)

AG(n=11)

GG(n=7)

Male sex, n (%) 5 (45%) 6 (55%) 3 (43%) 0.870

Age, years 60.6±6.9 57.1±7.6 67.6±5.0 0.014

BMI, kg/m2 30.4±4.7 32.2±4.6 29.7±4.8 0,486

Systolic bloodpressure mmHg 149±17 147±19 144±15 0.845

Diastolic bloodpressure mmHg 91±12 90±13 73±7 0.005

Fasting glucose, mmol/L 9.2±1.6 11.2±3.4 7.5±1.3 0.012

Hba1C mmol/mol 64±8 71±12 52±11 0.003

HOMA-IR 7.65±5.23 7.98±6.32 7.33±12.83 0.986

Total cholesterol, mmol/L 6.14±0.69 6.00±0.86 5.81±0.91 0.716

LDL cholesterol, mmol/L 3.44±0.86 3.57±0.93 3.68±0.61 0.860

HDL cholesterol, mmol/L 0.96±0.20 1.05±0.17 1.20±0.36 0.133

Triglycerides, mmol/L 3.49 (3.26-5.19) 2.67 (2.44-3.96) 1.54 (1.30-3.66) 0.039

Data are presented as mean (±SD), number (percentage) or median (IQR). P-value for gender was calculated by Pearson chi-square. P-value for triglycerides was calculated using Kruskal-Wallis test. Other p-values are calculated by one-way ANOVA. Abbreviations: HOMA-IR = homeostasis model assessment of insulin resistance; LDL = low-density lipoprotein; HDL = high-density lipoprotein; BMI = body mass index.

176

Chapter 10

DISCUSSIONIn the present study we investigated the human hepatic SULF2 expression as well as the effect of genetic variation in SULF2 on fasting and postprandial triglycerides in a large cohort of T2DM. We show that hepatic SULF2 expression is higher in humans with diabetes. In the Antwerp cohort we subsequently showed that hepatic SULF2 expression is significantly associated with plasma TG. We also demonstrate that genetic variation at the SULF2 locus (rs2281279) is reproducibly associated with lower fasting plasma triglycerides in T2DM subjects with diabetic dyslipidemia in two cohorts. In line we found that carrierschip of the SULF2 rs2281279 minor G allele was associated with lower levels of postprandial triglycerides, underscoring the potential role of SULF2 on postprandial triglyceride handling in humans.

SULF2 in human T2DM liver samplesIn the present study we found that T2DM liver tissue was characterised by an increased SULF2 expression. These findings are the first in-human data on hepatic SULF2 expression and are in line with previously published data in livers of db/db mice. (20, 21) Consecutively, we found a significant positive relation between hepatic SULF2 expression and fasting plasma TG levels in the Antwerp cohort. These results lend further support to a potential role of SULF2 in liver pathology and subsequent lipid handling.

SULF2 and triglyceride metabolism in T2DM Seven SULF2 tagSNPs were genotyped in two Dutch cohorts of Caucasian T2DM subjects with diabetic dyslipidemia to evaluate the effect on fasting plasma triglyceride levels. Among the 7 SULF2 tagSNPs, only the rs2281279 was associated with fasting triglycerides. Moreover, homozygous carriership of the minor G allele of rs2281279 was inversely correlated to fasting plasma triglyceride levels. Following an oral fat tolerance test in a random selection of participants stratified by genetic variants of SULF2 rs2281279, postprandial triglycerides and retinyl esters were significantly lower in heterozygous and homozygous G (minor) allele carriers in a stepwise manner.

Potential biological mechanism and clinical implicationsThe SULF2 gene encodes for the heparan sulfate glucosamine-6-O-endosulfatase (SULF2); an enzyme that removes 6-O sulfate groups from HSPGs. (33) Syndecan-1 HSPGs are important mediators of hepatic uptake of TG-rich lipoproteins.(14, 15, 17, 34) Previously, we showed that db/db mice have highly increased hepatic Sulf2 expression leading to increased plasma TG levels and abnormal hepatic TG-rich lipoprotein clearance. (21) Inhibition of hepatic Sulf2 expression by targeted allele-specific oligo antisense (ASO) treatment resulted in normalization of (postprandial) plasma TG levels, which coincided with increased binding of TG-rich lipoproteins to isolated hepatocytes from Sulf2-ASO treated animals.(21) These interesting observations underscore the important role of Sulf2 in regulating the proteoglycan mediated hepatic remnant clearance capacity in mice. Our study is the first to address the question whether SULF2 has a comparable function in human TG


177

10

homeostasis. In the present study, a functional role of SULF2 in postprandial clearance of TG-rich lipoproteins in humans is supported by the observed correlation between a genetic alteration in SULF2 and postprandial triglycerides in T2DM subjects.

SULF2 preferentially modifies the sulfation grade of the HSPG chain and affects its affinity for binding TG-rich lipoproteins. (21) Increased hepatic Sulf2 expression, which is found in diabetic mouse models, is significantly associated with higher plasma TG levels. Interestingly adiponectin and advanced glycation endproducts are strong effectors regulating Sulf2 expression. (20) To further explore the important role of HSPG in human hepatic remnant clearance, attention has to be focused on the role of the core HSPG proteins syndecan or glypican-1 in humans, which may have a more deleterious effect.

In our study, homozygous carriership for the minor G allele of SULF2 rs2281279 was characterized by lower HbA1c levels although no difference in BMI and HOMA-IR was observed between the groups. Interestingly, hepatic SULF2 expression in the Antwerp Cohort was significantly associated with fasting plasma glucose. implicating that SULF2 might play a role in glycemic regulation. This is in line with our previous finding in the murine studies, where the highest dose of Sulf2 antisense in db/db mice was associated with an improvement in fasting glucose levels and subsequent HOMA-IR(21) Further evaluation on the pathophysiology of this intriguing finding is warranted and currently performed at our department.

study limitationsFirst hepatic SULF2 gene expression was only determined in specimen from relative obese subjects, as liver biopsies in healthy subjects is not ethical. Second, we used common tagSNPs, located in intronic regions to ascertain the relation between SULF2 and plasma TG levels. We are therefore not able to draw conclusions on protein expression of SULF2 as an ELISA is not available to evaluate plasma SULF2 concentrations. Therefore, we are not able to show a relationship between plasma SULF2 levels stratified by genotype and plasma TG. Nevertheless, plasma values are unlikely to reflect functionality of SULF2 in the liver. Finally, we measured plasma RE and TG excursions as surrogate markers for postprandial lipid handling and not plasma apoB48. In this respect, plasma apoB48 only measures the number of intestinal-derived TG-rich lipoprotein particles during the postprandial phase but not its lipid composition. Plasma RE measurement, on the other hand, better reflects the lipid excursions during the postprandial phase, although it can be disputed that in the later time points also exchange of RE will occur from apoB48- containing lipoprotein particles towards apoB-100 containing lipoproteins. (35)

In conclusion, our results provide the first human evidence for a novel mechanism, in which SULF2 is involved in T2DM (postprandial) dyslipidemia. In conjunction with our data on Sulf2 antisense treatment in diabetic mice, these findings may indicate that SULF2 inhibition may be an attractive target to normalize dyslipidemia in T2DM patients.

178

Chapter 10

REFERENCE LIST1. Haffner SM, Lehto S, Ronnemaa T, Pyorala K, Laakso M. Mortality from coronary heart disease in

subjects with type 2 diabetes and in nondiabetic subjects with and without prior myocardial infarc-tion. N Engl J Med 1998;339:229-234.

2. Buse JB, Ginsberg HN, Bakris GL, Clark NG, Costa F, Eckel R, Fonseca V, Gerstein HC, Grundy S, Nesto RW, Pignone MP, Plutzky J, Porte D, Redberg R, Stitzel KF, Stone NJ. Primary prevention of cardio-vascular diseases in people with diabetes mellitus: a scientific statement from the American Heart Association and the American Diabetes Association. Diabetes Care 2007;30:162-172.

3. Moreton JR. Atherosclerosis and alimentary hyperlipemia. Science 1947;106:190-191.

4. Taskinen MR. Diabetic dyslipidaemia: from basic research to clinical practice. Diabetologia 2003;46:733-749.

5. Chahil TJ, Ginsberg HN. Diabetic dyslipidemia. Endocrinol Metab Clin North Am 2006;35:491-viii.

6. Williams KJ, Tabas I. Lipoprotein retention-and clues for atheroma regression. Arterioscler Thromb Vasc Biol 2005;25:1536-1540.



9. Freiberg JJ, Tybjaerg-Hansen A, Jensen JS, Nordestgaard BG. Nonfasting triglycerides and risk of is-chemic stroke in the general population. JAMA 2008;300:2142-2152.


11. Mahley RW, Huang Y. Atherogenic remnant lipoproteins: role for proteoglycans in trapping, transfer-ring, and internalizing. J Clin Invest 2007;117:94-98.

12. Williams KJ, Fless GM, Petrie KA, Snyder ML, Brocia RW, Swenson TL. Mechanisms by which lipopro-tein lipase alters cellular metabolism of lipoprotein(a), low density lipoprotein, and nascent lipopro-teins. Roles for low density lipoprotein receptors and heparan sulfate proteoglycans. J Biol Chem 1992;267:13284-13292.


14. MacArthur JM, Bishop JR, Stanford KI, Wang L, Bensadoun A, Witztum JL, Esko JD. Liver heparan sul-fate proteoglycans mediate clearance of triglyceride-rich lipoproteins independently of LDL receptor family members. J Clin Invest 2007;117:153-164.

15. Fuki IV, Kuhn KM, Lomazov IR, Rothman VL, Tuszynski GP, Iozzo RV, Swenson TL, Fisher EA, Williams KJ. The syndecan family of proteoglycans. Novel receptors mediating internalization of atherogenic lipoproteins in vitro. J Clin Invest 1997;100:1611-1622.

16. Fuki IV, Meyer ME, Williams KJ. Transmembrane and cytoplasmic domains of syndecan mediate a multi-step endocytic pathway involving detergent-insoluble membrane rafts. Biochem J 2000;351 Pt 3:607-612.


179

10

17. Stanford KI, Bishop JR, Foley EM, Gonzales JC, Niesman IR, Witztum JL, Esko JD. Syndecan-1 is the primary heparan sulfate proteoglycan mediating hepatic clearance of triglyceride-rich lipoproteins in mice. J Clin Invest 2009;119:3236-3245.


19. Bernfield M, Kokenyesi R, Kato M, Hinkes MT, Spring J, Gallo RL, Lose EJ. Biology of the syndecans: a family of transmembrane heparan sulfate proteoglycans. Annu Rev Cell Biol 1992;8:365-393.

20. Chen K, Liu ML, Schaffer L, Li M, Boden G, Wu X, Williams KJ. Type 2 diabetes in mice induces hepatic overexpression of sulfatase 2, a novel factor that suppresses uptake of remnant lipoproteins. Hepa-tology 2010;52:1957-1967.

21. Hassing HC, Mooij H, Guo S, Monia BP, Chen K, Kulik W, Dallinga-Thie GM, Nieuwdorp M, Stroes ES, Williams KJ. Inhibition of hepatic sulfatase-2 In Vivo: A novel strategy to correct diabetic dyslipi-demia. Hepatology 2012;55:1746-1753.

22. Bento JL, Palmer ND, Zhong M, Roh B, Lewis JP, Wing MR, Pandya H, Freedman BI, Langefeld CD, Rich SS, Bowden DW, Mychaleckyj JC. Heterogeneity in gene loci associated with type 2 diabetes on hu-man chromosome 20q13.1. Genomics 2008;92:226-234.

23. Comelli EM, Head SR, Gilmartin T, Whisenant T, Haslam SM, North SJ, Wong NK, Kudo T, Narimatsu H, Esko JD, Drickamer K, Dell A, Paulson JC. A focused microarray approach to functional glycomics: transcriptional regulation of the glycome. Glycobiology 2006;16:117-131.

24. Francque SM, Verrijken A, Mertens I, Hubens G, Van ME, Pelckmans P, Michielsen P, Van GL. Nonin-vasive assessment of nonalcoholic fatty liver disease in obese or overweight patients. Clin Gastroen-terol Hepatol 2012;10:1162-1168.

25. Chomczynski P, Sacchi N. Single-step method of RNA isolation by acid guanidinium thiocyanate-phe-nol-chloroform extraction. Anal Biochem 1987;162:156-159.

26. Verrijken A, Beckers S, Francque S, Hilden H, Caron S, Zegers D, Ruppert M, Hubens G, Van ME, Michielsen P, Staels B, Taskinen MR, Van HW, Van GL. A gene variant of PNPLA3, but not of APOC3, is associated with histological parameters of NAFLD in an obese population. Obesity (Silver Spring) 2013.

27. The effect of aggressive versus standard lipid lowering by atorvastatin on diabetic dyslipidemia: the DALI study: a double-blind, randomized, placebo-controlled trial in patients with type 2 diabetes and diabetic dyslipidemia. Diabetes Care 2001;24:1335-1341.

28. Welschen LM, van OP, Dekker JM, Bouter LM, Stalman WA, Nijpels G. The effectiveness of adding cognitive behavioural therapy aimed at changing lifestyle to managed diabetes care for patients with type 2 diabetes: design of a randomised controlled trial. BMC Public Health 2007;7:74.

29. The International HapMap Project. Nature 2003;426:789-796.

30. de Bakker PI, Yelensky R, Pe’er I, Gabriel SB, Daly MJ, Altshuler D. Efficiency and power in genetic as-sociation studies. Nat Genet 2005;37:1217-1223.

31. BLIGH EG, DYER WJ. A rapid method of total lipid extraction and purification. Can J Biochem Physiol 1959;37:911-917.

180

Chapter 10

32. DeRuyter MG, De Leenheer AP. Simultaneous determination of retinol and retinyl esters in serum or plasma by reversed-phase high-performance liquid chromatography. Clin Chem 1978;24:1920-1923.

33. Rosen SD, Lemjabbar-Alaoui H. Sulf-2: an extracellular modulator of cell signaling and a cancer target candidate. Expert Opin Ther Targets 2010;14:935-949.


35. Cohn JS, Johnson EJ, Millar JS, Cohn SD, Milne RW, Marcel YL, Russell RM, Schaefer EJ. Contribution of apoB-48 and apoB-100 triglyceride-rich lipoproteins (TRL) to postprandial increases in the plasma concentration of TRL triglycerides and retinyl esters. J Lipid Res 1993;34:2033-2040.


181

10

supplementary Figure 1 - Genomic context of suLF2. suLF2 is located on chromosome 20q12-q13.2 and contains 21 exons. The location of the seven selected tag SNPs is depicted in the Figure.

supplementary Figure 2 - Linkage disequilibrium plot of selected tagsNPs. The relationship between the seven selected tagSNPs is shown in LD plot. The values in the plot represent r2 between the appropriate SNPs (x10-2) and were calculated using Haploview. All are below 0.8, confirming their non-redundancy.


182

Chapter 10

supplementary Table 1 - baseline characteristics of liver biopsy patients

Controls(n=7)

Diabetes(n=7) P-value

Male sex, n (%) 3 (42%) 4 (57%) 0.5

Age, years 57.9±15.2 47.4±7,4 0.13

BMI, kg/m2 25.5±2.3 41.7±3.6 <0.001

Systolic blood pressure, mmHg 124±13 131±19 0.47

Diastolic blood pressure, mmHg 78±6 80±8 0.48

Fasting glucose, mmol/L 6.1±1.0 9.5±2.8 0.02

Hba1C mmol/mol 36±5 54±15 0.04

Total cholesterol, mmol/L 3.77±0.77 4.22±1.51 0.53

LDL cholesterol mmol/l 2.38±0.63 2.40±1.47 0.97

HDL cholesterol mmol/l 0.96±0.27 0.90±0.45 0.77

Triglycerides, mmol/L 1.15 (0.90-1.32) 1.71 (1.54-2.64) 0.008

Data are presented as mean (±SD), number (percentage) or median (IQR). P-value for gender was calculated by Pearson chi-square. P-value for triglycerides was calculated using Kruskal-Wallis test. Other p-values are calculated by one-way ANOVA. Abbreviations: LDL = low-density lipoprotein; HDL = high-density lipoprotein; BMI = body mass index.

supplementary Table 2 - baseline characteristics of the antwerp Cohort

Parameters

NFemale/maleAge (years)BMI kg/m2

Cholesterol mmol/LHDL cholesterol mmol/LLDL cholesterol mmol/LTriglycerides mmol/LGlucose mmol/LInsulin mIU/LHbA1C %AST U/LALT U/L

12183/3845 ± 1238.7 ± 6.75.25 ± 0.961.29 ± 0.363.16 ± 0.881.69 ± 0.954.6 (4.3 – 5.05)14 (10 – 21.5)5.6 (5.3 – 5.8)28 (24 – 36)40 (32 – 53)

Data are presented as mean ± SD for normally distributed variables or as median (interquartile range) when distribution of variable is skewed. LDL cholesterol was calculated by the Friedewald formula (cholesterol – (triglycerides/5)).


183

10

supplementary Table 3 - baseline characteristics by suLF2 rs2281279 genotype in Diabetes Care system

rs228127 genotypes

AA AG GG p-value

N=633 N=561 N=122

Male sex, n (%) 364 (58%) 273 (49%) 71 (58%) <0.001

Age, years 59.6±10.1 58.9±9.9 58.9±10.3 0.659

BMI, kg/m2 31.0±5.4 31.3±5.3 30.9±4.7 0.440

Systolic bloodpressure mmHg 144±20 144±20 146±20 0.539

Diastolic bloodpressure mmHg 82±11 83±11 84±11 0.316

Fasting glucose, mmol/L 8.6±2.3 8.9±4.5 8.7±2.5 0.231

Hba1C mmol/mol 7.3±1.6 7.5±1.7 7.2±1.6 0.038

Total cholesterol, mmol/L 5.40±1.17 5.49±1.15 5.42±1.21 0.044

LDL cholesterol, mmol/L 3.12±1.13 3.17±1.03 3.13±1.12 0.644

HDL cholesterol, mmol/L 1.11±0.24 1.11±0.44 1.09±0.26 0.729

Data are presented as mean (±SD), number (percentage) or median (IQR). Abbreviations: LDL = low-density lipoprotein; HDL = high-density lipoprotein; BMI = body mass index.

Hassing HC, Bernelot Moens SJ, Mooij HL, Dekker JM, Stroes ESG, Dallinga-Thie GM, Nieuwdorp M


11 A GENETIC vARIANT IN SuLF2 AND METABOLIC REsPoNsEs iN a PoPuLaTioN-basED CoHoRT

186

Chapter 11

ABSTRACTBackground - Genome-Wide Association Studies have associated the SULF2 locus with the presence of type 2 diabetes mellitus (T2DM), whereas in T2DM the SULF2 rs2281279 has also been associated with lower fasting glucose and postprandial plasma triglycerides levels. The exact relation between SULF2 and glucose homeostasis or lipid metabolism is, however, to date unclear. Therefore, we evaluated the impact of the SULF2 rs2281279 on metabolic responses in non-diabetic healthy individuals.

Methods - A 75g oral glucose tolerance test (OGTT) and a standardized meal tolerance test (MTT) was performed in 165 non-diabetic individuals stratified according to SULF2 rs2281279.

Results - SULF2 rs2281279 genotype was associated with a significant decrease in postprandial glycaemic excursions following OGTT in a stepwise manner with highest excursion in the homozygote carriers of the major allele (AA) and the lowest excursions in the homozygote carriers of the minor allele (GG)(p<0.05). Insulin sensitivity also revealed a stepwise improvement from AA to GG phenotype (p<0.05). Postprandial triglyceride levels were similar between the groups.

Conclusion - These results imply that SULF2 is involved in insulin-glucose homeostasis in humans. Further studies are needed to unravel the mechanism by which SULF2 affects insulin sensitivity.

SULF2 in a population-based cohort

187

11

INTRODUCTIONThe prevalence of the metabolic syndrome and type 2 diabetes mellitus (T2DM) is increasing worldwide.(1) This cardiometabolic syndrome is characterized by impaired glucose metabolism and metabolic dyslipidemia, including delayed clearance of postprandial triglyceride-rich lipoprotein (TRL) remnants. Heparan sulfate proteoglycans (HSPGs) are cell-membrane bound core proteins with 2-3 sulfated HS chains attached enabling the binding of various ligands, including lipoproteins and growth hormones.(2) HSPGs are involved in the regulation of cell growth, apoptosis and lipid metabolism. Syndecan-1, the primary form of HSPGs in the liver, has been implicated in type 2 diabetes mellitus (T2DM) as well as diabetes associated hypertriglyceridemia.(3, 4) In T2DM, the HSPGs are characterized by a decreased negative charge, most likely reflecting decreased heparan sulfation, leading to impaired binding of remnant lipoproteins (4, 7). Thus, perturbation of HSPG metabolism may contribute to the development of the metabolic syndrome.

Genome-Wide Association Studies (GWAS) have associated the SULF2 locus (20q13.1) with the presence of T2DM. (8) SULF2, a member of the Sulfatase family, encodes for glucosamine-6-O-endosulfatase-2, known to be a HSPG degrading enzyme. We previously showed in murine diabetic model that hepatic Sulf2 expression was positively related to plasma TG levels. Inhibition of hepatic Sulf2 expression by targeted allele-specific antisense administration resulted in normalization of plasma TG levels, reduced postprandial hypertriglyceridemia and lower fasting glucose levels in db/db mice. (7) In humans, genetic variation at the SULF2 locus was associated with lower fasting glucose and plasma TG levels in T2DM as well as accelerated postprandial hepatic TG clearance (Hassing, submitted). These data suggest a direct effect of SULF2 on both triglyceride metabolism and glycemic control per se in T2DM. The relative contribution of SULF2 on postprandial glucose and lipid metabolism in non-diabetic individuals, however, has not been reported yet.

In the present study we evaluated the metabolic effects of genetic variation in SULF2 in a cohort of non-diabetic otherwise healthy individuals.

METHODSStudy designPatients were included from a population based-cohort drawn from the municipal registry of Hoorn consisting of 208 subjects for which in- and exclusion criteria have been described previously. (9) For the present analyses, we included patients of whom DNA was available. Patients with known diabetes mellitus type 2 (defined as those using oral antihyperglycaemic agents) were excluded. Participants received a 75g-OGTT and a standardized mixed meal test after a 10h-overnight fast, on separate days, in random order, within 2 weeks. To reduce the impact of diurnal variation, all tests started between 7:30 and 9:00 a.m. Apart from the OGTT or test meal and small amounts of water, participants refrained from food, drinks and physical activity during the test. The study was

188

Chapter 11

approved by the Institutional Review Board of the VU University Medical Center in accordance with the Declaration of Helsinki (updated version 2008).

GenotypingWe genotyped a single predetermined rs2281279 [c.2494+267A>G] by allelic discrimination using Tagman mastermix on a BioRad CFX system.

Oral glucose tolerance test (OGTT)Blood samples were drawn from the antecubital vein in the fasting state and at 15, 30, 60, 90, and 120 min following 75 gram glucose ingestion. Prior to the test, blood pressure (Collin Press-mate BP-8800, Colin, Komaki-City, Japan), weight, height, waist and hip circumference were measured.

Meal Tolerance Test (MTT)Subjects were served a standardized mixed breakfast consisting of 2 croissants (90 g), 10 g butter, 40 g cheese, 150 g full-fat milk, and 100 g yoghurt drink enriched with 10 g of soluble carbohydrates (maltose). The approximate total nutrient content was 3487 kJ (74 g [36 Energy%] carbohydrates, 49 g [52 Energy%] fat of which 28.2 g was saturated and 24 g [12 Energy%] proteins). Blood samples were collected in fasting state and at 15, 30, 60, 90, 120, 180, and 240 min after meal ingestion.

Laboratory analysisPlasma glucose levels were determined with a glucose hexokinase method (Gluco-quant; Roche Diagnostics, Mannheim, Germany); serum insulin and C-peptide, with immunometric assays (ACS Centaur; Bayer Diagnostics, Mijdrecht, the Netherlands), and TG, total cholesterol, and high-density lipoprotein cholesterol, with enzymatic colorimetric assays (Roche, Basel, Switzerland). Low-density lipoprotein cholesterol was calculated according to the Friedewald-formula except when fasting TG levels exceeded 5.0 mmol/L. Free fatty acid was measured by enzymatic colorimetric assays (WAKO Chemicals, Neuss, Germany). Homeostatic model assessment for insulin resistance (HOMA-IR) was calculated by the following formula: [glucose (mmol/L) * insulin (mU/L)]/22.5.

beta-cell function parameters and insulin sensitivity during oGTTThe insulinogenic index (as an estimation of early insulin secretion) was calculated by dividing the increment in insulin during the first 30 min by the increment in glucose over the same period (∆I30/∆G30). Negative or infinite insulinogenic indexes were excluded (n=10). Overall glucose-stimulated insulin secretion was calculated as AUCinsulin/AUCglucose ratio.

Insulin sensitivity was estimated from glucose and insulin values according to methods described by Mari et al. (OGIS: oral glucose insulin sensitivity) (10), Matsuda and DeFronso (ISIcomp: index of composite whole-body insulin sensitivity) (11) and Stumvoll et al. (MCRest: metabolic clearance rate of glucose) (12).


189

11

statistical analysisDescriptive statistics are presented as mean ± standard deviation (SD) or median and interquartile range. Differences between genotypes are calculated by Pearson chi-square for categorical variables. Continues variables are calculated by one-way analysis of variances (ANOVA) with Bonferroni post hoc test for normally distributed variables. Natural log-transformed variables were used in case of non-normally distributions. Total areas under the curve (AUC) of insulin, glucose, c-peptide and triglycerides were calculated by the trapezoid rule. Area under the incremental curve (iAUC) was obtained by subtracting the baseline (t=0) levels from each point. Comparisons between groups were analysed using univariate ANOVA. Two-sided probability values of less than 0.05 were considered statistical significant.

RESULTSPopulation characteristicsOf the 208 patients, 28 patients were excluded because of missing DNA analysis and additionally 15 patients were excluded because of known type 2 diabetes mellitus. Thus, a total of 165 patients were included in the current analysis. Stratification according SULF2 rs2281279 genotype, resulted in 87 (53%) homozygous carriers of the major allele (AA), 63 (38%) heterozygous carriers (AG) and 15 (9%) homozygous carriers of the minor allele (GG). Carriers of the minor G allele were more often female, displayed a lower waist-hip ratio, a lower HOMA-IR and had higher HDL-cholesterol levels. Other baseline characteristics did not differ significantly between the genotypes (Table 1).

Metabolic responses following oGTT and MTTThe presence of SULF2 rs2281279 genotype was associated with a significant decrease in glycaemic excursions following OGTT in a stepwise manner with highest excursion in the homozygote carriers of the major allele (AA) and the lowest excursions in the homozygote carriers of the minor allele (GG) (Figure 1). Areas under the glucose curves differed significantly between genotypes [AUC 804 (689-941) (mmol/L)*min, 793 (669-875) (mmol/L)*min and 650 (618-771) (mmol/L)*min for AA, AG and GG respectively, p<0.05] which remained significant following correction for baseline values [(iAUC 147 (53-274) (mmol/L)*min for AA, 148 (40-228) (mmol/L)*min for AG and 71 (0-159) for GG (mmol/L)*min, p<0.05]. In line, insulin and C-peptide responses (Figure.1) revealed a comparable stepwise trend by SULF2 genotype. Glucose, insulin, C-peptide and triglyceride excursions following mixed meal tolerance test are displayed in Figure 2. There were no significant differences in postprandial triglyceride values.

190

Chapter 11

Table 1 - baseline characteristics

Rs2281279 genotypes

P-valueAA(n=87)

AG(n=63)

GG(n=15)

Male sex, n (%) 50 (57%) 25 (40%) 3 (20%) 0.008

Age, years 52.9 (±6.5) 54.6 (±6.6) 50.4 (±5.9) 0.061

BMI, kg/m2 27.5 (±4.1) 26.5 (±3.5) 25.8 (±4.0) 0.153

Waist-hip ratio 0.93 (±0.08) 0.89 (±0.07) 0.86 (±0.07) <0.001¹

Systolic blood pressure mmHg 133.9 (±14.0) 136.6 (±17.3) 126.8 (±16.4) 0.087

Diastolic blood pressure mmHg 76.9 (±9.1) 77.5 (±10.5) 72.3 (±9.3) 0.172

Fasting glucose, mmol/L 4.8 (±0.6) 4.9 (±1.0) 4.7 (±0.3) 0.587

HOMA-IR 2.08 (±1.88) 1.44 (±0.89) 1.69 (±1.27) 0.0392

Total cholesterol, mmol/L 5.05 (±0.98) 5.16 (0.85) 5.05 (±1.08) 0.825

LDL cholesterol, mmol/L 3.09 (±0.89) 3.12 (0.81) 2.87 (0.68) 0.592

HDL cholesterol, mmol/L 1.33 (±0.35) 1.46 (±0.37) 1.52 (±0.48) 0.0323

Triglycerides, mmol/L 1.20 (0.80-1.70) 1.10 (0.90-1.50) 1.10 (0.70-1.40) 0.525

Data are presented as mean (±SD), number (percentage) or median (IQR). Abbreviations: LDL = low-density lipoprotein; HDL = high-density lipoprotein; BMI = body mass index. ¹Significant P-values for Bonferroni post hoc tests: AA vs AG p=0.005, AA vs GG p=0.002. 2Significant P-values for Bonferroni post hoc tests: AA vs AG p=0.0353Not significant following Bonferroni post hoc tests.


191

11

0 50 100 1502

4

6

8

10AAAGGG

p = <0.05p*= <0.05

time (min)

gluc

ose

(mm

ol/l)

0 50 100 1500

200

400

600AAAGGG

p = 0.067p*= 0.096

time (min)

insu

lin (p

mol

/l)

0 50 100 1500

1

2

3AAAGGG

p = 0.074p*= 0.100

time (min)

C-p

eptid

e (n

mol

/l)

A

B

C

Figure 1 - oral Glucose Tolerance Tests. Glucose, insulin and c-peptide responses (mean ± S.E.) following OGTT in subjects with different SULF2 Rs2281279 genotypes (AA, AG or GG). * = p value for incremental AUC.

192

Chapter 11

0 100 200 300 4001.0

1.5

2.0

2.5

3.0AAAGGG

time (min)

trig

lyce

rides

(mm

ol/l) p = 0.234

p*= 0.996

0 100 200 3000

200

400

600AAAGGG

time (min)

insu

lin (p

mol

/l)

p = <0.05p*= <0.05

0 100 200 3000

1

2

3AAAGGG

time (min)

C-p

eptid

e (n

mol

/l)

p = <0.05p*= 0.064

0 50 100 150 200 2504

5

6

7

8AAAGGG

time (min)

gluc

ose

(mm

ol/l)

p = 0.258p*= 0.626

Figure 2 - Meal Tolerance Tests. Glucose, insulin, c-peptide and triglyceride responses (mean ± S.E.) following a standardized mixed meal in subjects with different SULF2 Rs2281279 genotypes (AA, AG or GG). * = p value for incremental AUC

beta-cell function parameters and insulin sensitivityFollowing OGTT there were no differences in beta-cell function as measured by insulinogenic index (early insulin secretion) nor by AUCinsulin/AUCglucose ratio (overall glucose-stimulated insulin secretion, Table 4). Interestingly, insulin sensitivity differed significantly between genotypes by all three estimates of insulin sensitivity used (OGIS, ISIcomp and MCRest) in these non-diabetic subjects. Subject with heterozygote or homozygote carriership of the minor G allele displayed significant higher insulin sensitivity in a stepwise manner compare to homozygote carriers of the major allele (Table 2).

DISCUSSIONThis study evaluated the effect of genetic variation in SULF2 on metabolic responses following OGTT and MTT in a healthy non-diabetic population. We found that carriership of the rs2281279 minor G allele is already associated with higher insulin sensitivity Further studies are needed to dissect the underlying mechanisms by which SULF2 regulates insulin-glucose homeostasis and whether this can be pursued as therapeutic target for glucose control.


193

11

It has previously been shown that SULF2 is associated with elevated plasma triglyceride concentration and an impaired postprandial clearance of triglyceride-rich lipoprotein remnants in T2DM. (4, 7) In contrast to T2DM subjects, we did not find any effects of the rs2281279 genotype on postprandial triglyceride levels. It is recognized that SULF2 modifies the sulfation grade of the HSPG chain thereby affecting its affinity for binding of TG-rich lipoproteins. However, other ligand receptors beyond the HSPG system are present at the hepatocyte membrane, including the LDL-receptor, which also binds TRLs. (13, 14) The present study was performed in healthy normo-triglyceridemic subjects. In these subjects a normal capacity of the liver for TRL clearance is expected to suffice. Thus other receptors can compensate for the loss of HSPG capacity in the clearance of TRLs. In line, it was previously shown that both baseline TG levels and the presence of T2DM are independent predictors of postprandial TG curves. (15, 16) Therefore, in a diabetic and hypertriglyceridemic state this capacity might become limited leading to impaired TRL clearance due to a lower HSPG clearance capacity as suggested by our experimental models. (7)

Interestingly, we did find significant higher insulin sensitivity in non-diabetic subjects who were carriers of the rs2281279 minor G allele. Also, the SULF2 locus (20q13.1) was identified previously in a GWAS study to be associated with the presence of T2DM. (8) Although proteoglycans and the endothelial glycocalyx are recognized as being essential for optimal endothelial function, the role of SULF2 in peripheral insulin sensitivity is unknown. Interestingly, glypican-4 (Gpc4), which belongs to the HSPG family, acts as an insulin sensitizer. (17) In this study, Gpc4 was found to interact with the insulin receptor, enhance insulin receptor signalling and enhance adipocyte differentiation. Furthermore, serum Gpc4 levels were positively correlated with body fat content,

Table 2 - beta-cell function parameters and insulin sensitivity during oGTT

Rs2281279 genotypes P-value

AA (n=87) AG (n=63) GG (n=15)

Insulinogenic index (pmol/mmol) 105.4 (63.2-181.1) 117 (63.7-158.2) 114,8 (65.2-198.3) 0.677

AUCinsulin/AUCglucose ratio (pmol/mmol) 41.1 (30.1-63.2) 37.2 (28.2-53.2) 31.0 (30.1-60.7) 0.338

OGIS (ml/(min m²)) 408 (357-450) 422 (380-456) 471 (418-509) <0.051

ISIcomp (μmol/(kg min pmol L)) 16.4 (11.1-25.7) 21.4 (14.5-28.9) 27.3 (12.7-32.7) <0.052

MCR index (ml/(min kg)) 8.9 (7.6-10.1) 9.2 (8.5-10.1) 9.8 (8.3-11.2) <0.053

Values are presented as median (interquartile range). AUC = area under the curve, OGIS = oral glucose insulin sensitivity, ISIcomp = index of composite whole-body insulin sensitivity, MCRest = metabolic clearance rate of glucose. 1Significant p=values for Bonferroni post hoc test: AA vs GG; p=0.007. 2Not significant following Bonferroni post hoc tests. 3No t significant following Bonferroni post hoc tests

194

Chapter 11

insulin resistance and BMI. For the latter this was also true for non-diabetic subjects. In line, our findings imply a protective effect of SULF2 against impaired insulin sensitivity. As Sulf2 is strongly upregulated in diabetic liver tissue (4), it is tempting to speculate on a potential role in hepatic gluconeogenesis. Since these data are currently lacking, further studies on the role of SULF2 on hepatic gluconeogenesis and peripheral insulin sensitivity are warranted.

In summary, the present data provide the first evidence that SULF2 might be involved in human glycaemic control. Regarding the GWAS association of SULF2 with T2DM, previous experimental findings linking SULF2 to diabetic dyslipidemia and current results implicating an insulin sensitizing role for SULF2 in a pre-diabetic state, these finding indicate that SULF2 is involved in insulin-glucose metabolism. Although not significant in healthy subjects, SULF2 might also be pivotal for lipid metabolism in metabolically challenged subjects. Therefore, SULF2 might play an essential role in several processes at the core of the metabolic syndrome. Future studies are warranted to evaluate whether HSPG can serve as a target in metabolic disturbances.


195

11

REFERENCE LIST1. Buse JB, Ginsberg HN, Bakris GL, Clark NG, Costa F, Eckel R, Fonseca V, Gerstein HC, Grundy S, Nesto

RW, Pignone MP, Plutzky J, Porte D, Redberg R, Stitzel KF, Stone NJ. Primary prevention of cardio-vascular diseases in people with diabetes mellitus: a scientific statement from the American Heart Association and the American Diabetes Association. Diabetes Care 2007;30:162-172.


3. Stanford KI, Bishop JR, Foley EM, Gonzales JC, Niesman IR, Witztum JL, Esko JD. Syndecan-1 is the primary heparan sulfate proteoglycan mediating hepatic clearance of triglyceride-rich lipoproteins in mice. J Clin Invest 2009;119:3236-3245.

4. Chen K, Liu ML, Schaffer L, Li M, Boden G, Wu X, Williams KJ. Type 2 diabetes in mice induces hepatic overexpression of sulfatase 2, a novel factor that suppresses uptake of remnant lipoproteins. Hepa-tology 2010;52:1957-1967.

5. Olsson U, Egnell AC, Lee MR, Lunden GO, Lorentzon M, Salmivirta M, Bondjers G, Camejo G. Changes in matrix proteoglycans induced by insulin and fatty acids in hepatic cells may contribute to dyslipi-demia of insulin resistance. Diabetes 2001;50:2126-2132.

6. Rohrbach DH, Hassell JR, Kleinman HK, Martin GR. Alterations in the basement membrane (heparan sulfate) proteoglycan in diabetic mice. Diabetes 1982;31:185-188.

7. Hassing HC, Mooij H, Guo S, Monia BP, Chen K, Kulik W, Dallinga-Thie GM, Nieuwdorp M, Stroes ES, Williams KJ. Inhibition of hepatic sulfatase-2 In Vivo: A novel strategy to correct diabetic dyslipi-demia. Hepatology 2012;55:1746-1753.

8. Bento JL, Palmer ND, Zhong M, Roh B, Lewis JP, Wing MR, Pandya H, Freedman BI, Langefeld CD, Rich SS, Bowden DW, Mychaleckyj JC. Heterogeneity in gene loci associated with type 2 diabetes on hu-man chromosome 20q13.1. Genomics 2008;92:226-234.

9. Rijkelijkhuizen JM, Girman CJ, Mari A, Alssema M, Rhodes T, Nijpels G, Kostense PJ, Stein PP, Eekhoff EM, Heine RJ, Dekker JM. Classical and model-based estimates of beta-cell function during a mixed meal vs. an OGTT in a population-based cohort. Diabetes Res Clin Pract 2009;83:280-288.

10. Mari A, Pacini G, Murphy E, Ludvik B, Nolan JJ. A model-based method for assessing insulin sensitiv-ity from the oral glucose tolerance test. Diabetes Care 2001;24:539-548.

11. Matsuda M, DeFronzo RA. Insulin sensitivity indices obtained from oral glucose tolerance testing: comparison with the euglycemic insulin clamp. Diabetes Care 1999;22:1462-1470.

12. Stumvoll M, Mitrakou A, Pimenta W, Jenssen T, Yki-Jarvinen H, Van HT, Haring H, Fritsche A, Gerich J. Assessment of insulin secretion from the oral glucose tolerance test in white patients with type 2 diabetes. Diabetes Care 2000;23:1440-1441.



15. Twickler TB, Dallinga-Thie GM, Cohn JS, Chapman MJ. Elevated remnant-like particle cholester-ol concentration: a characteristic feature of the atherogenic lipoprotein phenotype. Circulation 2004;109:1918-1925.

196

Chapter 11

16. Ai M, Tanaka A, Ogita K, Sekinc M, Numano F, Numano F, Reaven GM. Relationship between plasma insulin concentration and plasma remnant lipoprotein response to an oral fat load in patients with type 2 diabetes. J Am Coll Cardiol 2001;38:1628-1632.

17. Ussar S, Bezy O, Bluher M, Kahn CR. Glypican-4 enhances insulin signaling via interaction with the insulin receptor and serves as a novel adipokine. Diabetes 2012;61:2289-2298.

12 SUMMARY AND PERSPECTIvES

200

Chapter 12

SUMMARYThe research presented in this thesis addresses several pathophysiological pathways in glucose and lipid metabolism, two major parameters of the cardiometabolic syndrome. The main findings reported in this thesis are presented below. Whereas the first part of the thesis deals with the cause of hypertriglyceridemia and novel treatment candidates targeting the cardiometabolic syndrome, the second part of this thesis explores the role of heparan sulfates in hyperglycaemia and hypertriglyceridemia. Moreover, using a translational approach we show that defects in heparan sulfate proteoglycan (HSPG) synthesis affects lipid and glucose metabolism.

Part I: Causes of hypertriglyceridemia and novel treatment targets Fasting and non-fasting triglyceride (TG) levels are recognized as independent risk factors for cardiovascular disease (CVD) 1, a finding which bears direct relevance in view of the pandemic of obesity and type 2 diabetes mellitus (T2DM). However, current risk prediction models do not include TG levels and therapeutic strategies to lower the CVD risk in cardiometabolic subjects are limited. In chapter 2 we review novel aspects in TG metabolism including the role of heparan sulfate proteoglycans (HSPGs) in hepatic lipid remnant clearance. Atherogenic TG remnant particles bind to hepatic HSPGs after which they are internalized and cleared by the liver. 2 In murine models, abnormalities in HSPG result in a mild hypertriglyceridemic (HTG) phenotype due to delayed clearance of these TG-rich lipoprotein remnant particles. 3 Therefore, HSPG modulation might be an interesting target for future therapeutic targets in humans. 4

Early identification of subjects at risk for HTG could reduce their risk of coronary artery disease by inducing early interventions. A DNA sequence variant could be such an early indicator for future risk of HTG and coronary artery disease. In chapter 3, we hypothesized that the combination of several polymorphisms associated with plasma TG levels improves CVD risk prediction. Accordingly we evaluated if a panel of validated single nucleotide polymorphisms (SNPs) in TG-modulating genes predicts plasma triglyceride levels as well as first cardiovascular event in a prospective case-control study in the EPIC-Norfolk cohort. This study shows that a gene score composed of three TG modulating SNPs in APOA5 and LPL was linearly associated with plasma TG concentrations (+0.32 mmol/l [95% CI 0.257-0.379] per allele change, p<0.0001) and other lipid parameters representing an atherogenic lipid profile including decreased low-density lipoprotein (LDL) size, increased LDL number, increased very low-density lipoprotein (VLDL) particle number and decreased high-density lipoprotein (HDL) particle size. In line, the risk of future coronary artery disease was elevated in individuals with the highest gene score compared to those with lowest gene scores (odds ratio 1.88 [95% CI 1.11-3.18]; p=0.02).

Nowadays, lifestyle interventions and statin and fibrate based therapy are the primary treatment steps in hypertriglyceridemic patients. However, a residual cardiovascular risk remains in these subjects, even when LDL-cholesterol target levels are achieved. 5 It is unclear what the optimal

Summary

201

12

treatment strategy is for patients who are treated with statin therapy and still display a metabolic lipid profile. In chapter 4 we investigated whether high-risk patients with metabolic dyslipidemia gain additional benefit from switching to high dose statin therapy. To address this question we performed a post-hoc analysis in the Treating to New Targets (TNT) trial including 9,994 patients with coronary artery disease who were randomly assigned to receive either atorvastatin 10 mg or 80 mg per day following a run-in phase on atorvastatin 10 mg. We found that increasing atorvastatin dose from 10 mg to 80 mg results in significant incremental cardiovascular benefit in patients with high TG levels and low HDL-cholesterol levels.

Another potential novel candidate, targeting cardiometabolic sequelae are the thyroid hormone mimetics. Since thyroid hormone mimetics are capable of uncoupling the beneficial metabolic effects of thyroid hormones from their deleterious effects on heart, bone and muscle, this class of drug is considered as adjacent therapeutic to weight-lowering strategies. In chapter 5 we performed a randomized, placebo-controlled, double-blind trial to investigate the safety and efficacy of TRC150094, a thyroid hormone mimetic, in male cardiometabolic subjects. In this study, TRC150094 dosed 50 mg once daily was safe and well tolerated, however, insulin sensitivity, hepatic fat content and lipid profiles did not improve following 4 weeks of administration. However, subgroup analysis indicated that TRC150094 might have beneficial effects in patients with more severe metabolic derangements (including as overt diabetes mellitus and hypertriglyceridemia), suggesting beneficial effects in specific patients groups.

Finally, the ever-expanding genomic information has led to an explosion of novel approaches to unravel the pathophysiology of dyslipidemia, varying from the use of classical approaches for monogenetic disorders, to the growing amount of genome wide association studies that have increased our understanding of the complex processes involved in (postprandial) lipid metabolism. 6 In chapter 6 we review the therapeutic use of genomic sequencing in monogenetic and polygenetic hypertriglyceridemia.

Part II: Acquired and inborn errors of heparansulfates in hyperglycaemia and hypertriglyceridemiaAs mentioned previously, experimental data indicates that hepatic HSPGs are involved in the clearance of atherogenic remnant lipoproteins by the liver. In chapter 7 we investigate the postprandial lipid handling by heparan sulfates following both a monogenetic (subjects with hereditary multiple exostosis or HME) and SNP based approach in patients with familial hypercholesterolemia (FH). HME subjects are characterized by loss-of-function mutations in exostin (EXT), involved in HS chain elongation, and therefore display inborn defects in HSPGs. FH subjects are already characterized by a defect in another hepatic remnant receptor namely LDL receptor, and therefore their lipid clearance might be more HSPG-dependent. We observed only a trend towards small changes in postprandial lipid handling in carriers of loss-of-function EXT mutations (HME subjects), indicating a negligible contribution of HS chain length to lipid metabolism. Using a HSPG SNP based approach in heterozygous FH subjects we did find small differences in postprandial retinyl palmitate. These

202

Chapter 12

findings suggest that, although modest compared to previous murine findings, heparan sulfates may indeed contribute in human lipid metabolism.

Several genome wide association studies have identified exostosin 2 (EXT2) as a novel risk factor for the development of type 2 diabetes mellitus. 7 As EXT genes (involved in the chain elongation step of heparan sulfate biosynthesis) are intricately involved in organ development, we hypothesized that mutations in these genes might affect pancreatic volume and subsequent insulin secretion/beta cell reserve capacity. In chapter 8 we used a translational approach showing an effect of EXT mutations on pancreatic volume, glucose stimulated insulin secretion and beta-cell function (by hyperglycaemic clamp) in mice and humans with heterozygous EXT mutations underlying HME syndrome. Although no differences in oral glucose tolerance and insulin sensitivity were found in both mice and men with EXT mutations when compared to controls, we did find a significantly reduced glucose stimulated first phase insulin secretion as well as a decreased absolute insulin secretion capacity (upon arginine stimulation) in HME subjects. In line with these findings, abdominal Magnetic Resonance Imaging revealed a significantly smaller pancreas volume in HME subjects compared to controls providing the first evidence that heparan sulfates are indeed important for normal pancreas development and subsequent beta-cell function in humans.

Postprandial hepatic clearance of triglyceride-rich lipoprotein (TRL) remnants is known to be impaired in T2DM and thereby attributing to its metabolic dyslipidemic phenotype. It was previously shown that glucosamine-6-O-endosulfatase-2 (Sulf2), a HSPG desulfase enzyme, is highly expressed in livers of diabetic mice. Increased levels of Sulf2 result in decreased sulfation of HSPG, a diminished negative charge and subsequently a decreased binding of TRLs in cultured hepatocytes. 8 In chapter 9 we show that inhibition of hepatic Sulf2, by a targeted allele-specific antisense approach, restores the TRL-binding capacity of primary hepatocytes, normalizes non-fasting plasma TG levels and reduced postprandial hypertriglyceridemia in db/db mice. These findings provide an in vivo proof-of-concept for Sulf2 inhibition as potential target to improve metabolic dyslipidemia.

In chapter 10 we aimed to translate previous findings implicating a role of SULF2 in diabetic dyslipidemia to humans. We investigated whether genetic variation in SULF2 associates with plasma triglyceride levels in T2DM subjects. This study shows that carriership of the SULF2 rs2281279 minor G allele predisposes to lower postprandial TRL levels in dyslipidemic T2DM patients, underscoring the relevance of our Sulf2 antisense findings in mice. Chapter 11 further illuminates the impact of the SULF2 rs2281279 on metabolic responses in non-diabetic individuals. Although we were not able to find any effect on postprandial plasma TG levels in these non-diabetic subjects, we did find that the SULF2 rs2281279 genotype was associated with a significant decrease in glycaemic excursions following an OGTT in a stepwise manner, which was accompanied by improved insulin sensitivity in these genotypes. This intriguing finding suggests a potential role of SULF2 in peripheral or hepatic insulin sensitivity in humans and further studies are warranted to study this potential mechanism.

Summary

203

12

PERSPECTIvESAs stated previously, the prevalence of the cardiometabolic syndrome and its sequelae is expanding worldwide. Hence, there is an immense, unmet medical need for approaches to unravel its complex pathophysiology and the identification of novel therapeutic targets to reduce morbidity and mortality rates.

In this thesis, we suggested a gene score based on the presence of TG-raising polymorphisms which predicts plasma TG levels as well as future coronary artery diseased risk. The use of a mini-TG gene chip might thus bolster future individualized patient therapeutic approaches aimed at improved cardiovascular risk prediction and/or therapeutic efficacy in guiding choice of TG lowering therapy and therapeutic rigor of treating hypertriglyceridemia. In the past few years, the efficacy of various drugs for treatment of hypertriglyceridemia including fibrates combination with statins was evaluated in large prospective randomized clinical trials showing negative outcomes. 9, 10 Unfortunately, also novel therapeutic interventions including nicotinic acid, 11 ezetimibe 12 and CETP inhibition 13 have failed in consistently reducing cardiovascular end-points on top of statin therapy. Thus, despite extensive clinical research efforts, intensive statin therapy remains the only evidence based pharmacotherapeutic strategy to reduce cardiovascular risk in patients with metabolic dyslipidemia in clinical practice. Fortunately, novel approaches keep entering the cardiometabolic arena including antisense therapy (i.e. antisense APOC3) 14 and monoclonal antibodies (i.e. PCSK9 antibody). 15 Large, long-term randomised trials with predefined clinical cardiovascular endpoints are now eagerly awaited to substantiate the promise of these new drugs to establish safety, tolerability, acceptability, and cost-effectiveness in dyslipidemic patients. One of these promising novel potential candidates in our battle against the cardiometabolic syndrome is the thyroid hormone mimetics. Thyroid analogues increase basal energy expenditure and oxygen consumption leading to a reduction in body weight with concomitant favourable improvements in lipid and carbohydrate metabolism. 16 Indeed, 3 months treatment with the thyroid hormone analogue Eprotirome was associated with decreases in levels of atherogenic lipoproteins in patients receiving treatment with statins. 17 However, reports on toxicity warrant caution and call for further research to identify compounds that are able to increase basal metabolism without toxicity.

This thesis opens a new research area focussing on the role of HSPGs in human glucose and lipid metabolism. For example, soon after the report on a role of hepatic over-expression of Sulf2 in diabetic dyslipidemia we were able to report that restoration of HSPG derangements in diabetic mice corrected their (postprandial) dyslipidemic phenotype. 8 Although the effect of HSPG genes, including SULF2, on postprandial TG levels seems rather modest in human lipid metabolism, it is still tempting to speculate about targeted (antisense based) therapeutic intervention aimed at normalizing hepatic SULF2 in human diabetic subjects. This based on the selectivity of antisense aimed at hepatic targets on the one hand, whereas redundancy of HSPG genes on the other hand

204

Chapter 12

will prevent a complete loss of HSPG function. 18 As minor genetic disturbances in HSPGs, results in human metabolic derangements varying from glucose-insulin metabolism, beta-cell mass and hepatic TG clearance, it is tempting to speculate on the lipid- and glucose lowering effects of HS mimetics 19 or modulation of HSPG core proteins in type 2 diabetes mellitus subjects with hypertriglyceridemia. With so many unknown, unidentified and mystifying HSPG targets in human biology, the research field linking (hepatic) glycobiology and cardiometabolic homeostasis could be a fruitful area for future attempts targeting HSPG biosysnthesis in a safe and effective manner in order to reduce cardiometabolic sequelae.

Summary

205

12

REFERENCE LIST1. Bansal S, Buring JE, Rifai N, Mora S, Sacks FM, Ridker PM. Fasting compared with nonfasting triglyc-

erides and risk of cardiovascular events in women. JAMA 2007;298:309-316.


3. Stanford KI, Bishop JR, Foley EM et al. Syndecan-1 is the primary heparan sulfate proteoglycan me-diating hepatic clearance of triglyceride-rich lipoproteins in mice. J Clin Invest 2009;119:3236-3245.


5. Ahmed S, Cannon CP, Murphy SA, Braunwald E. Acute coronary syndromes and diabetes: Is intensive lipid lowering beneficial? Results of the PROVE IT-TIMI 22 trial. Eur Heart J 2006;27:2323-2329.

6. Teslovich TM, Musunuru K, Smith AV et al. Biological, clinical and population relevance of 95 loci for blood lipids. Nature 2010;466:707-713.

7. Sladek R, Rocheleau G, Rung J et al. A genome-wide association study identifies novel risk loci for type 2 diabetes. Nature 2007;445:881-885.

8. Chen K, Liu ML, Schaffer L et al. Type 2 diabetes in mice induces hepatic overexpression of sulfatase 2, a novel factor that suppresses uptake of remnant lipoproteins. Hepatology 2010;52:1957-1967.


10. Keech A, Simes RJ, Barter P et al. Effects of long-term fenofibrate therapy on cardiovascular events in 9795 people with type 2 diabetes mellitus (the FIELD study): randomised controlled trial. Lancet 2005;366:1849-1861.

11. Boden WE, Probstfield JL, Anderson T et al. Niacin in patients with low HDL cholesterol levels receiv-ing intensive statin therapy. N Engl J Med 2011;365:2255-2267.

12. Taylor AJ, Villines TC, Stanek EJ et al. Extended-release niacin or ezetimibe and carotid intima-media thickness. N Engl J Med 2009;361:2113-2122.

13. Barter PJ, Caulfield M, Eriksson M et al. Effects of torcetrapib in patients at high risk for coronary events. N Engl J Med 2007;357:2109-2122.

14. Visser ME, Witztum JL, Stroes ES, Kastelein JJ. Antisense oligonucleotides for the treatment of dys-lipidaemia. Eur Heart J 2012;33:1451-1458.

15. Roth EM, McKenney JM, Hanotin C, Asset G, Stein EA. Atorvastatin with or without an antibody to PCSK9 in primary hypercholesterolemia. N Engl J Med 2012;367:1891-1900.

16. Pearce EN. Thyroid hormone and obesity. Curr Opin Endocrinol Diabetes Obes 2012;19:408-413.

17. Ladenson PW, Kristensen JD, Ridgway EC et al. Use of the thyroid hormone analogue eprotirome in statin-treated dyslipidemia. N Engl J Med 2010;362:906-916.

18. Presto J, Thuveson M, Carlsson P et al. Heparan sulfate biosynthesis enzymes EXT1 and EXT2 affect NDST1 expression and heparan sulfate sulfation. Proc Natl Acad Sci U S A 2008;105:4751-4756.

19. Ferro V, Dredge K, Liu L et al. PI-88 and novel heparan sulfate mimetics inhibit angiogenesis. Semin

Thromb Hemost 2007;33:557-568.

NEDERLANDSE SAMENvATTING

208

NEDERLANDSE SAMENVATTING

NEDERLANDSE SAMENvATTINGHet cardiometabool syndroom kenmerkt zich door een cluster van medische aandoeningen die veelvuldig voorkomen bij type 2 diabetes zoals overgewicht, een gestoorde glucose tolerantie, insuline resistentie en metabole dyslipidemie (hoog plasma triglyceriden en laag plasma high-density lipoprotein (HDL) cholesterol). De prevalentie van het cardiometabool syndroom en de gevolgen hiervan, zoals hart- en vaatziekten (HVZ), breiden zich wereldwijd nog steeds uit. Het is daarom van groot belang om de complexe pathofysiologie van het metabool syndroom te ontrafelen en zo tot nieuwe inzichten te komen die mogelijk kunnen leiden tot nieuwe aangrijpingspunten in de behandeling.

Uit experimenteel onderzoek blijkt dat heparan sulfaat proteoglycanen (HSPGs) mogelijk betrokken zijn bij type 2 diabetes en diabetes-geassocieerde ziekten zoals endotheeldysfunctie en dyslipidemie. Heparan sulfaten zijn suikerketens die gebonden aan eiwitten (proteoglycanen) op de celmembraan bijdragen aan allerlei (patho)fysiologische processen zoals orgaan ontwikkeling, ontstekingsprocessen en lipiden metabolisme. Eerdere preklinische studies in muizen tonen dat deze HSPGs belangrijk zijn voor het klaren van schadelijke triglyceriden uit het bloed door de lever. In het geval van type 2 diabetes blijken heparan sulfaten overmatig te worden afgebroken hetgeen een mogelijke verklaring levert voor de gestoorde lipiden klaring bij patiënten met type 2 diabetes. Concluderend zijn er uit dierexperimentele studies in het verleden aanwijzingen gekomen dat HSPGs een belangrijke rol in het lipiden metabolisme spelen. Echter, over de rol van HSPGs in het menselijke glucose en lipidenmetabolisme was tot voor kort weinig bekend.

Dit proefschrift beschrijft enkele nieuwe pathofysiologische mechanismen in het glucose en triglyceriden metabolisme, twee belangrijke parameters van het cardiometabool syndroom. Het eerste deel van dit proefschrift beschrijft de oorzaak van hypertriglyceridemie (verhoogde triglyceriden waarden) en potentiele nieuwe strategieën in de behandeling van het metabool syndroom. In het tweede deel van dit proefschrift onderzoeken we de rol van heparan sulfaten in het ontstaan van hyperglycaemie en hypertriglyceridemie.

DEEL I Zowel verhoogde nuchtere als niet-nuchtere triglyceriden waarden in het bloed worden beschouwd als een onafhankelijke risicofactor voor HVZ. Huidige risico-predictie modellen nemen echter de triglyceriden waarden niet mee omdat er in deze waarden een grote variatie over de dag bestaat en triglyceriden waarden tevens nauw samenhangen met andere variabelen zoals gewicht, HDL-cholesterol en type 2 diabetes. Daarnaast bestaan er momenteel weinig therapeutische opties om het risico op HVZ te verlagen bij personen met hypertriglyceridemie.


209

In hoofdstuk 2 beschrijven we de huidige inzichten omtrent het triglyceriden metabolisme waaronder de rol van HSPGs in de hepatische klaring van triglyceride-deeltjes. Triglyceriden binden aan de LDL-receptor, de LDL-receptor related protein en HSPGs op hepatocyten (levercellen) waarna ze door de lever worden opgenomen en kunnen worden geklaard. In muismodellen resulteren afwijkingen in hepatische HSPGs in een fenotype dat gekenmerkt wordt door vertraagde triglyceridenklaring. Modulatie van lever HSPGs zou dus mogelijk een aantrekkelijk aangrijpingspunt kunnen vormen voor toekomstige therapieën in de behandeling van hypertriglyceridemie.

Vroege identificatie van personen die een hoog risico lopen op het ontwikkelen van hypertriglyceridemie zou deze patiëntengroep, door middel van tijdige interventie, kunnen beschermen tegen het ontwikkelen van HVZ. Bepaalde DNA sequenties die geassocieerd zijn met hoge plasma triglyceriden zouden een dergelijke indicator kunnen vormen. In hoofdstuk 3 laten we zien dat een risico score bestaande uit de combinatie van slechts drie veel voorkomen genetische variaties in APOA5 en LPL (genen betrokken bij het triglyceriden metabolisme) een goede indicatie kan geven van plasma triglyceriden waarden en een toekomstig risico op HVZ.

Ondanks de huidige behandelingsstrategie, die voornamelijk berust op statine therapie, bestaat er nog steeds een residueel verhoogd risico op HVZ bij personen met hypertriglyceridemie. Hoofdstuk 4 beschrijft de resultaten van een post-hoc analyse in de Treating to New Targets (TNT) trial naar het additionele effect van het ophogen van de dagelijkse statine behandeling bij personen met coronair lijden en metabole dyslipidemie die reeds werden behandeld met een lage dosis statine. Uit deze studie onder 9994 personen blijkt dat het ophogen van atorvastatine van 10 mg naar 80 mg leidt tot een additionele risicoreductie op HVZ bij personen met een hoog plasma triglyceriden en laag plasma HDL-cholesterol.

Een veelbelovende kandidaat in de behandeling van het cardiometabool syndroom zijn de schildklierhormoon mimetica. Dit vanwege het gunstige effect van het schildklierhormoon op het metabolisme. In hoofdstuk 5 voeren we een gerandomiseerde, placebo-gecontroleerde, dubbelblinde studie uit naar de veiligheid en effectiviteit van TRC150094, een schildklierhormoon mimeticum, in mannen met het metabool syndroom. Ondanks de veelbelovende resultaten uit experimentele studies konden wij echter geen effect van TRC150094 aantonen op insuline gevoeligheid, het gehalte aan vet in de lever alsook het lipiden profiel. Echter, subgroep analyse laat zien dat er mogelijke wel gunstige effecten verwacht mogen worden in personen met een meer uitgesproken metabole ontregeling zoals type 2 diabetes of hypertriglyceridemie.

Tot slot beschouwt hoofdstuk 6 de mogelijkheden die onze huidige kennis over het genoom biedt voor toekomstige therapeutische opties in hypertriglyceridemie.

210


DEEL IIZoals eerder genoemd zijn HSPGs betrokken bij de hepatische klaring van triglyceriden. In hoofdstuk 7 onderzoeken wij het effect van HSPGs op postprandiale lipiden door middel van zowel een mono-genetisch model (personen met hereditaire multipele exostosen ofwel HME) en een poli-genetisch model in patiënten met familiaire hypercholesterolemie (FH). Personen met HME zijn geboren met mutaties in een exostosin (EXT) gen, dat betrokken is bij het verlengen van heparan sulfaat ketens. Deze personen kenmerken zich dus door aangeboren afwijkingen in HSPGs en kunnen ons daarom helpen de rol van HSPG in lipiden metabolisme te onderzoeken. Personen met FH hebben een genetisch defect in een andere lever receptor betrokken bij triglyceriden klaring, namelijk de low-density lipoprotein (LDL) receptor. Het lipiden metabolisme is bij deze personen derhalve mogelijk meer afhankelijk van HSPGs. Wij onderzochten daarom in deze personen het effect van veelvoorkomende genetische variaties in HSPG-modulerende genen. In deze studie vonden wij geen afwijkende lipiden klaring in personen met HME in vergelijking tot niet-aangedane familie leden na een vetbelastingtest. Het effect van een kortere HS-keten lengte lijkt dus mogelijk minder van belang voor het lipiden metabolisme. In personen met FH en frequente variaties in HSPG genen tonen we, in tegenstelling tot experimentele data, slechts kleine (mogelijk klinisch relevante) verschillen in postprandiale triglyceriden klaring aan.

GWAS (genome wide association studies) hebben exostosin 2 (EXT2) herhaaldelijk geïdentificeerd als risico factor voor type 2 diabetes. Aangezien EXT (betrokken bij de verlenging van heparan sulfaten) betrokken is bij de orgaan ontwikkeling, zouden mutaties in deze genen mogelijk de grootte van de pancreas kunnen aantasten en daardoor kunnen resulteren in verminderde insuline secretie capaciteit. In hoofdstuk 8 bestuderen we zowel muizen als mensen met EXT mutaties. In vergelijking tot gezonde controles laten we geen verschillen in glucose tolerantie en insuline gevoeligheid zien. Wel vinden we een verminderde insuline secretie en een kleinere pancreas grootte in personen met HME. Deze resultaten zouden kunnen betekenen dat HSPGs belangrijk zijn voor normale pancreas ontwikkeling en functie, en dat mensen met kleiner pancreas volume minder goed bestand zijn tegen verhoogde vraag naar insuline (bv als gevolg van insuline resistentie bij overgewicht).

Personen met type 2 diabetes hebben vaak een gestoorde hepatische postprandiale triglyceriden klaring. Uit eerder onderzoek blijkt dat Sulf2, een enzym dat sulfaat groepen verwijderd van HSPGs, sterk verhoogd tot expressie komt in levers van type 2 diabetische (db/db) muizen en resulteert in een verminderde hepatische bindingscapaciteit van triglyceriden. Hoofdstuk 9 beschrijft de resultaten van hepatische Sulf2 inhibitie in een preklinische studie. In een model voor type 2 diabetes tonen we aan dat hepatische Sulf2 inhibitie de bindingscapaciteit van levercellen voor triglyceriden herstelt hetgeen resulteert in een normalisatie van niet-nuchtere en postprandiale triglyceriden. Deze resultaten impliceren dat Sulf2 inhibitie mogelijk een aangrijpingspunt vormt voor de behandeling van diabetes-geassocieerde dyslipidemie. In hoofdstuk 10 onderzoeken we derhalve of SULF2 ook een belangrijke rol speelt in de ontwikkeling van diabetische dyslipidemie


211

bij de mens. We tonen aan dat een genetische variant in SULF2 inderdaad geassocieerd lijkt te zijn met lagere (postprandiale) triglyceriden waarden in patiënten met type 2 diabetes. Hoofdstuk 11 beschrijft de impact van genetische variatie in SULF2 in personen zonder type 2 diabetes. In deze personen vinden we geen effect op triglyceriden maar wel associaties tussen genetische variatie in SULF2 en insuline gevoeligheid.

AUTHORS AND AFFILIATIONS

214


AUTHORS AND AFFILIATIONSAckermans MTDepartment of Clinical Chemistry, Laboratory of Endocrinology, Academic Medical Center, Amsterdam, the Netherlands

bernelot Moens sjDepartment of Vascular Medicine, Academic Medical Center, Amsterdam, the Netherlands

Boekholdt SMDepartment of Cardiology, Academic Medical Center, Amsterdam, the Netherlands

breazna aPfizer, Inc., New York, New York, USA

Chauthaiwale VClinical Research Department, Torrent Pharmaceuticals Limited, Village-Bhat, Dist. Gandhinagar, India

Chen kSection of Endocrinology, Diabetes and Metabolism, Temple University School of Medicine, Philadelphia, PA USA

Dallinga-Thie GMDepartment of Vascular Medicine and Department of Experimental Vascular Medicine, Academic Medical Center, Amsterdam, The Netherlands.

Dekker jMDepartment of Epidemiology and Biostatistics, EMGO Institute for Health and Care Research, VU University Medical Center, Amsterdam, the Netherlands

DeMicco DAPfizer, Inc., New York, New York, USA

Dutt CClinical Research Department, Torrent Pharmaceuticals Limited, Village-Bhat, Dist. Gandhinagar, India


215

Esko jDDepartment of Cellular and Molecular Medicine, UC San Diego, San Diego, USA

Guo SDepartment of Antisense Drug Discovery, Isis Pharmaceuticals Inc, Carlsbad, CA USA

’t Hart LMDepartment of Molecular Epidemiology, Leiden University Medical Center, Leiden, the Netherlands

Kastelein jjPDepartment of Vascular Medicine, Academic Medical Center, Amsterdam, the Netherlands

Khaw KTDepartment of Public Health and Primary Care, Strangeways Research Laboratory, University of Cambridge, Cambridge, United Kingdom

Kruit jKDepartment of Pediatrics, Center for Liver, Digestive, and Metabolic Diseases, University Hospital Groningen, Groningen, the Netherlands

kulik wLaboratory Genetic Metabolic Diseases, Academic Medical Center, Amsterdam, the Netherlands

LaRosa jCState University of New York Health Science Center, New York, New York, USA

Mohanan AClinical Research Department, Torrent Pharmaceuticals Limited, Village-Bhat, Dist. Gandhinagar, India

Monia BPDepartment of Antisense Drug Discovery, Isis Pharmaceuticals Inc, Carlsbad, CA USA

Mooij HLDepartment of Vascular Medicine, Academic Medical Center, Amsterdam, the Netherlands

216


Nederveen ajDepartment of Radiology, Academic Medical Center, Amsterdam, the Netherlands

Nieuwdorp MDepartment of Vascular Medicine, Academic Medical Center, Amsterdam, the Netherlands

Nijpels GDepartment of General Practice, EMGO institute, VU University Medical Center, Amsterdam, the Netherlands

Pathak kClinical Research Department, Torrent Pharmaceuticals Limited, Village-Bhat, Dist. Gandhinagar, India

van de sande MajDepartment of Orthopedics, Leiden University Medical Center, Leiden, the Netherlands

serlie MjDepartment of Endocrinology, Academic Medical Center, Amsterdam, the Netherlands

stoker jDepartment of Radiology, Academic Medical Center, Amsterdam, the Netherlands

Stroes ESGDepartment of Vascular Medicine, Academic Medical Center, Amsterdam, the Netherlands

Surendran RPDepartment of Experimental Vascular Medicine, Academic Medical Center, Amsterdam, the Netherlands

Tanck MwDepartment of Biostatistics, Academic Medical Center, Amsterdam, the Netherlands

Thakkar PClinical Research Department, Torrent Pharmaceuticals Limited, Village-Bhat, Dist. Gandhinagar, India

van der valk FMDepartment of Vascular Medicine, Academic Medical Center, Amsterdam, the Netherlands


217

visser MEDepartment of Vascular Medicine, Academic Medical Center, Amsterdam, the Netherlands

wareham NjMedical Research Council Epidemiology Unit, Institute of Metabolic Science, Addenbrooke’s Hospital, Cambridge, United Kingdom

williams KjSection of Endocrinology, Diabetes and Metabolism, Temple University School of Medicine, Philadelphia, PA USA

witjes jjDepartment of Vascular Medicine, Academic Medical Center, Amsterdam, the Netherlands

xu D Department of Cellular and Molecular Medicine, UC San Diego, San Diego, USA.

Zwinderman aHDepartment of clinical epidemiology, biostatistics, bioinformatics, Academic Medical Center, Amsterdam, The Netherlands

DANkwOORD

220

dankwoord

DANkwOORD

“Als je iets wilt, spant de hele wereld samen om je daarbij te helpen. – Paulo Coelho, the Alchemist.”

Dat bijna 3 jaar onderzoek op de afdeling vasculaire geneeskunde heeft geresulteerd in het proefschrift dat thans in uw handen ligt is te danken aan hulp van velen. Tijdens dit project ben ik meer dan eens verrast door de enthousiaste en niet zelden onverwachte hulp die mij vanuit verscheidene kanten werd geboden. Graag wil ik hierbij al diegenen hartelijk bedanken, waarvan enkelen met een persoonlijk dankwoord.

Allereerst wil ik alle studiedeelnemers hartelijk bedanken. Zonder hun medewerking kan wetenschappelijk onderzoek niet plaatsvinden en mijn dank en waardering is dan ook groot.

Mijn promotor, professor dr. E.S.G. Stroes. Beste Erik, jouw enthousiasme, positiviteit en energie zijn ongeëvenaard. Jouw vermogen om mensen “van enthousiasme door elkaar schuddend” te stimuleren is groot. Ontelbare keren ben ik overtuigd van het ene jouw kamer binnengelopen om geïnspireerd door het andere jouw kamer weer te verlaten. Mijn dank hiervoor.

Mijn co-promotores, dr. M. Nieuwdorp en dr. G.M. Dallinga-Thie. Beste Max, wat ooit begon als een borrel in de sky-bar op Genève Aéroport heeft geresulteerd in jouw betrokken begeleiding van dit proefschrift. Jouw brede visie, gedrevenheid, positieve benadering en vooruitziende blik zijn voor mij een bron van inspiratie. Ik hoop nog veel mooie wijn met je te mogen drinken. Beste Geesje, jouw fundamentele lipiden kennis en contacten in lipiden-land waren onmisbaar voor dit proefschrift. Daarnaast ken ik niemand die zo snel loopt en die zo hard werkt als jij (en dat louter op ontelbare koppen koffie). Ik wil je bedanken voor alle tijd en energie die je in dit project hebt gestoken.

De leden van de promotiecommissie, prof. dr. J.J.P. Kastelein, prof. dr. J.M.F.G. Aerts, prof. dr. R.J.G. Peters, prof. dr. J.B.L. Hoekstra, prof. dr. U.H.W. Beuers en prof. dr. J.M. Dekker ben ik zeer erkentelijk voor de inzet van hun deskundigheid en hun waardevolle tijd om mijn proefschrift te beoordelen en hun bereidheid zitting te nemen in de commissie. Beste John, al vroeg tijdens mijn promotietraject heb jij het stokje overgedragen een Erik. Maar niet zonder op een vaderlijke wijze beschikbaar te blijven voor adviezen of het doornemen van manuscripten waarvoor mijn dank.

dankwoord

221

Prof. dr. H.R. Büller, prof. dr. S. Middeldorp en dr. M.D. Trip. Harry, wiens deur altijd openstond voor goede adviezen, of dit nu het presenteren van resultaten of iets heel anders betrof. Saskia, dank voor je kritische blik en sportieve input. Beste Mieke, als “moeder” van de afdeling zette jij je immer in voor het welzijn van eenieder op F4. Ik wil je bedanken voor je steun.

Mijn dank gaat verder uit naar alle medewerkers van het laboratorium (Alinda in het bijzonder) die hebben geholpen met de proeven in dit proefschrift. Ook dank aan de medewerkers van het trialbureau voor hun hulp bij alle klinische studies.

Graag wil ik alle co-auteurs van de hoofdstukken in dit proefschrift bedanken voor hun bijdrage en een goede samenwerking.

Mijn collega’s van de afdeling Cardiologie en de Interne Geneeskunde van het AMC Amsterdam wil ik bedanken voor een warm welkom in de kliniek en de ruimte die mij werd gegeven dit proefschrift af te ronden. Renée van den Brink wil ik bedanken voor het in mij gestelde vertrouwen.

Mijn collega’s van de Vasculaire Geneeskunde op F4 met wie ik vele congressen, skiweekenden, borrels en koffiemomenten doorbracht. Dat collegialiteit verder gaat dan alleen gezelligheid bleek op momenten waarop ik weer eens gezonde proefpersonen nodig had en ik nooit verder hoefde te zoeken dan onze eigen gang. Joyce, van onschatbare waarde voor de afdeling, behalve daarvoor, ben ik je dankbaar voor je vriendschap, Fatima, Karim, Lysette, Frederiek, Geerte, Menno, Hans Avis, Raphael, Danny, Olaf, Onno, Remco, Ester, Roeland, Renée, Nanne, Bas, Hans “hands on” Mooij (dank voor de samenwerking), Diederik, Lily, Maartje, Brigitte, Suthesh (Londen is wel hipper dan Almere), Ankie, Sara, Aart (the atherosclerotic-plank), Anne (altijd een luisterend oor), Katrijn (immer in voor een break), Meeike, Corien: roomy, met jouw gevoel voor humor en inspiratie bleef werken op een kamer zonder daglicht van 2,5m² toch altijd gezellig, veel dank hiervoor, Maayke, Elise, Barbara, Danka (Twente goes Africa!), Inge, Ties, Daan, Fouad, Ruud (no guts, no glory), Marjet, Andrea, Mandy, Josien (dank voor het delen van de eindsprint), Maurits, Loek, Sophie (ik kon geen betere SULF ambassadeur wensen), Paulien (altijd een goed advies) en Fleur (wat een bakken met energie).

Mijn vrienden, vriendinnen en studiegenoten wil ik bedanken voor alle steun. Mijn clubgenoten, Alian, Christine, Eveline, Felisa, Ilse, Jeanine, Merije en Sanneke: inmiddels uitgewaaierd over heel Europa en toch altijd in voor gezelligheid, bedankt voor de welverdiende etentjes en afleiding. Sanneke, na de middelbare school, onze studenten

222

dankwoord

tijd, vakanties en nog veel meer gedeeld te hebben, ben ik blij dat jij ook nu naast mij staat.

Fokke, Lammy, Catharina en Dennis. Dank voor jullie betrokkenheid, oprechte interesse en een motto dat meer dan eens van toepassing bleek: “As `t net kin sa`t it moat, dan moat it mar sa`t it kin”.

Lieve papa en mama, dank voor het liefdevolle nest waarin jullie ons hebben opgevoed. Al pretendeerde ik vroeger nog wel eens het allemaal zelf wel te kunnen: jullie nimmer aflatende steun, betrokkenheid en vertrouwen zijn voor mij van onschatbare waarde op zowel moeilijke als mooie momenten. Dit boekje is voor jullie. Robert-Jan, mijn grote broer, en samen met Laura een stille kracht en rots in de branding. Ik ben blij dat jij vandaag naast me staat. Mijn zus Anne-Roos en Daan, altijd in voor gezelligheid of een design advies. Je weet niet hoe fijn ik het vind om mijn grote zus weer zo dichtbij te hebben. Mijn zusje Rianne en Thom, dank voor jullie liefdevolle interesse, daadkracht, betrokkenheid en fantastische diners. Jorick en Jasmijn, de kleintjes van het stel en ook de sfeermakers. Bedankt voor jullie humor, steun en relativeringsvermogen.

Lieve Abel, dank voor je liefde en je pogingen om voortdurend het beste in mij naar boven te halen. Jij bent me vele malen liever dan alle 221 bladzijden hiervoor.

CURRICULUM vITAE

CURRICULUM VITAE

227

CURRICULUM vITAECarlijne Hassing was born on the 8th of April 1983 in Utrecht, the Netherlands. After graduating from secondary school at the “st. Thomas College” in Venlo in 2001, she was admitted to the University of Utrecht, the Netherlands, to study medicine. After completing medical school in 2007, Carlijne started working at the Department of Cardiology at the University Medical Center in Utrecht. In 2008 she entered her PhD program under supervision of prof. dr. J.J.P. Kastelein and prof. dr. E.S.G. Stroes at the Department of Vascular Medicine of the Academic Medical Center in Amsterdam. This translational research focused on glycobiology in dyslipidemia and hyperglycaemia and resulted in this thesis: ‘Glycobiology in cardiometabolic homeostasis’. In 2011, Carlijne initiated her Cardiology training under supervision of dr. R.B.A. van den Brink at the Department of Cardiology of the Academic Medical Center in Amsterdam.

UvA-DARE (Digital Academic Repository) Glycobiology in cardiometabolic homeostasis ... ·...

Documents

Transcript of UvA-DARE (Digital Academic Repository) Glycobiology in cardiometabolic homeostasis ... ·...