023 1842142291 lipid and asteroclorosis

Advances in Translational Medical Science

Lipids and Atherosclerosis

Prelims 14/7/05 4:44 pm Page i

Prelims 14/7/05 4:44 pm Page ii

Advances in Translational Medical Science

Lipids and AtherosclerosisEdited by

C.J. Packard, BSc, PhD, FRCPath, DSc

Department of Vascular BiochemistryGlasgow Royal InfirmaryScotland

and

D.J. Rader, MDDepartment of Medicine and Institute for Translational Medicines and TherapeuticsUniversity of Pennsylvania School of MedicinePhiladelphia, PAUSA

Foreword by

V. Fuster, MD, PhD

President, World Heart FederationDirector, Cardiovascular InstituteMount Sinai School of MedicineNew York, NYUSA

Prelims 14/7/05 4:44 pm Page iii

© 2006 Taylor & Francis, an imprint of the Taylor & Francis Group

First published in the United Kingdom in 2006by Taylor & Francis, an imprint of the Taylor & Francis Group, 2 Park Square, Milton Park, Abingdon, Oxon OX14 4RN

Tel.: +44 (0) 207 017 6000Fax.: +44 (0) 207 017 6699E-mail: [email protected]: http://www.tandf.co.uk/medicine

All rights reserved. No part of this publication may be reproduced, stored in a retrieval system, ortransmitted, in any form or by any means, electronic, mechanical, photocopying, recording, orotherwise, without the prior permission of the publisher or in accordance with the provisions of theCopyright, Designs and Patents Act 1988 or under the terms of any licence permitting limited copyingissued by the Copyright Licensing Agency, 90 Tottenham Court Road, London W1P 0LP.

Although every effort has been made to ensure that all owners of copyright material have beenacknowledged in this publication, we would be glad to acknowledge in subsequent reprints oreditions any omissions brought to our attention.

Although every effort has been made to ensure that drug doses and other information are presentedaccurately in this publication, the ultimate responsibility rests with the prescribing physician. Neither thepublishers nor the authors can be held responsible for errors or for any consequences arising from the useof information contained herein. For detailed prescribing information or instructions on the use of anyproduct or procedure discussed herein, please consult the prescribing information or instructionalmaterial issued by the manufacturer.

A CIP record for this book is available from the British Library

Library of Congress Cataloging-in-Publication DataData available on application

ISBN 1-84214-229-1

Distributed in North and South America by

Taylor & Francis2000 NW Corporate BlvdBoca Raton, FL 33431, USA

Within Continental USATel.: 800 272 7737; Fax.: 800 374 3401Outside Continental USATel.: 561 994 0555; Fax.: 561 361 6018E-mail: [email protected]

Distributed in the rest of the world byThomson Publishing ServicesCheriton HouseNorth WayAndover, Hampshire SP10 5BE, UKTel.: +44 (0) 1264 332424E-mail: [email protected]

Composition by Scribe Design Ltd, Ashford, Kent, UKPrinted and bound in Spain by Grafos SA

Prelims 14/7/05 4:44 pm Page iv

List of contributors vii

Foreword xi

Preface xiii

Acknowledgements xiv

1. The assembly of very low density lipoproteins in the liver 1S-O. Olofsson and L. Asp

2. Cardiovascular implications of partial, tissue-specific silencing of MTP 15C.C. Shoulders

3. Acyl CoA: Diacylglycerol acyltransferases (DGATs) as therapeutic targets for cardiovascular disease 31J.S. Millar and J.T. Billheimer

4. The biochemical and physiological roles of ACAT1 and ACAT2 in cholesterol homeostasis and atherosclerosis 41R.G. Lee and L.L. Rudel

5. Overview of intestinal lipid metabolism 55T.A. Miettinen and H. Gylling

6. The role of non-cholesterol sterols in the pathogenesis of atherosclerosis and their modulation by the sitosterolaemia locus 63E.L. Klett and S.B. Patel

7. FXR: the molecular link between bile acid and lipid metabolism 69T. Claudel, E. Sturm, B. Staels and F. Kuipers

8. Overview of HDL and reverse cholesterol transport 81P. Barter

9. LXR as a therapeutic target for atherosclerosis 93I.G. Schulman and R.A. Heyman

10. Endothelial lipase and the regulation of HDL metabolism 101K.O. Badellino and W. Jin

11. Cholesteryl ester transfer protein (CETP) inhibition as a therapeutic strategy for raising HDL-cholesterol levels and reducing atherosclerosis 111D.J. Rader

Contents

Prelims 14/7/05 4:44 pm Page v

12. Overview of insulin resistance and the metabolic syndrome 119N. Sattar and B. Mukhopadhyay

13. Peroxisome proliferator activated receptors and energy metabolism 137P. Gervois, J-C. Fruchart and B. Staels

14. Adipocytes and their secretory products 153J.B. Prins

15. Regulation of adipocyte triglyceride storage 165F. Karpe and G.D. Tan

16. Role of innate immunity in atherosclerosis: immune recognition, immune activation and the atherosclerotic process 175J-M. Fernández-Real

17. Role of CD44 in atherogenesis and its potential role as a therapeutic target 191E. Puré

18. Lipoxygenases: potential therapeutic target in atherosclerosis 207L. Zhao and C.D. Funk

19. Role of secretory phospholipase A2 isozymes (Lp-PLA2) 219C.H. Macphee

Index 231

vi CONTENTS

Prelims 14/7/05 4:44 pm Page vi

L. AspDepartment of Medical BiochemistryThe Wallenberg Laboratory forCardiovascular ResearchGöteborg UniversitySweden

K.O. Badellino Schools of Medicine and NursingUniversity of PennsylvaniaPhiladelphia, PAUSA

P. BarterThe Heart Research InstituteSydney, NSWAustralia

J.T. BillheimerThe Department of MedicineUniversity of PennsylvaniaPhiladelphia, PAUSA

T. ClaudelCenter for Liver, Digestive and Metabolic

DiseasesLaboratory of PediatricsAcademic Hospital GroningenGroningenThe Netherlands

J-M. Fernández-RealSection of Diabetes, Endocrinology and

NutritionUniversity Hospital of Girona ‘Dr JosepTrueta’GironaSpain

J-C. FruchartUnité de Recherche 545Institute National de la Santé et de la

Recherche MédicaleDepartement d’AthéroscléroseInstitut Pasteur de LilleUniversité de Lille 2LilleFrance

C.D. Funk Departments of Physiology and BiochemistryQueen’s UniversityKingston, ONCanada

P. GervoisUnité de Recherche 545Institute National de la Santé et de la

Recherche MédicaleDépartment d’AthéroscléroseInstitut Pasteur de LilleUniversité de Lille 2LilleFrance

H. GyllingDepartment of Clinical NutritionUniversity of KuopioKuopioFinland

R.A. HeymanExelixis Inc.San Diego, CAUSA

List of contributors

Prelims 14/7/05 4:44 pm Page vii

W. JinSchool of MedicineUniversity of PennsylvaniaPhiladelphia, PAUSA

F. KarpeOxford Centre for Diabetes, Endocrinology

and MetabolismNuffield Department of Clinical MedicineUniversity of OxfordOxford UK

E.L. KlettDivision of Endocrinology, Diabetes and

Medical GeneticsMedical University of South CarolinaCharleston, SCUSA

F. KuipersCenter for Liver, Digestive and Metabolic


R.G. LeeDepartment of PathologyWake Forest University School of MedicineWinston-Salem, NCUSA

C.H. MacpheeDepartment of Vascular Biology and

ThrombosisGlaxoSmithKlineKing of PrussiaPhiladelphia, PAUSA

T.A. Miettinen Biomedicum HelsinkiDivision of Internal MedicineHelsinkiFinland

J.S. MillarThe Department of PharmacologyUniversity of PennsylvaniaPhiladelphia, PAUSA

B. MukhopadhyayDepartment of Vascular Biochemistry and

DiabetesGlasgow Royal InfirmaryGlasgowUK

S-O. OlofssonWallenberg LaboratorySahlgrenska University HospitalGöteborgSweden

S.B. PatelDivision of Endocrinology and MetabolismMedical College of WisconsinMilwaukee, WIUSA

J.B. Prins Department of Diabetes and EndocrinologyPrincess Alexandra HospitalWoolloongabba, QLDAustralia

E. Puré The Wistar InstitutePhiladelphia, PAUSA

D.J. Rader Department of Medicine and Institute for

Translational Medicine and TherapeuticsUniversity of Pennsylvania School of MedicinePhiladelphia, PAUSA

L.L. RudelDepartment of PathologyWake Forest University School of MedicineWinston-Salem, NCUSA

viii LIST OF CONTRIBUTORS

Prelims 14/7/05 4:44 pm Page viii

N. SattarDepartment of Vascular BiochemistryGlasgow Royal InfirmaryGlasgowUK

I.G. SchulmanExelixis Inc.San Diego, CAUSA

C.C. ShouldersMRC Clinical Sciences CentreHammersmith HospitalDu Cane RoadLondon UK

B. StaelsUnité de Recherche 545Institut National de la Santé et de la

Recherche MédicaleDepartment d’AthéroscléroseInstitut Pasteur de LilleUniversité de Lille 2LilleFrance

E. SturmCenter for Liver, Digestive and Metabolic


G.D. TanOxford Centre for Diabetes, Endocrinology

and MetabolismNuffield Department of Clinical MedicineUniversity of OxfordOxford UK

L. ZhaoDepartment of Pharmacology and Institute for

Translational Medicine and TherapeuticsUniversity of Pennsylvania Philadelphia, PAUSA

LIST OF CONTRIBUTORS ix

Prelims 14/7/05 4:44 pm Page ix

Prelims 14/7/05 4:44 pm Page x

Recent insights into the aetiology ofatherothrombosis, the leading cause ofmorbidity and mortality in Western society,have opened the way to a more strategicapproach to treatment and prevention. It isrecognised now that the composition of theplaque as well as its size is crucial in determin-ing the propensity to rupture and to initiate alife threatening thrombotic event. Key featuresassociated with plaque fragility are: a large lipidcore, an abundance of tissue factor and inflam-matory cells including macrophages, vasovaso-rum, and a thin fibrous cap, the structuralintegrity of which is compromised by thepresence of matrix degrading metallopro-teinases.

Atherogenic lipoproteins play a central rolein the initiation and growth of atheroscleroticplaques. The ‘lipid hypothesis’ is now dogmawith the repeated demonstration that loweringthe level of plasma low density lipoprotein(LDL) cholesterol reduces the risk of onset ofrecurrent myocardial infarction and manyother cardiovascular complications. Statins,inhibitors of the rate-limiting enzyme in choles-terol synthesis, have been shown to preventcoronary heart disease and stroke, and are oneof the success stories of modern medicine.However, many patients cannot adequatelyreduce their levels of atherogenic lipoproteinswith statins alone. Therefore, new approachesto reducing LDL and other atherogeniclipoproteins are still needed. Furthermore,even in statin treated patients there is consider-able residual risk of cardiovascular disease.Therefore, it is imperative to ask what newapproaches to the control of other features ofatherothrombosis will ultimately yield furtherrisk reduction. Major areas of new therapeutic

targeting include intestinal and hepatic lipidmetabolism, HDL metabolism and reversecholesterol transport, and inflammatoryprocesses within the vascular wall and beyond.The tremendous advances in basic science willbe translated into therapeutic advances. Withinthis context, translational research is a termthat encapsulates the idea of ‘bench’ to‘bedside’ translation of etiological concepts andtherapeutic targets. The range of technologiesavailable now, from genome searches to molec-ular and clinical imaging of atherothromboticplaques permits novel insights into one arena tobe tested quickly in another, thereby develop-ing an integrated picture of the potential ofnew treatment modalities.

This book will be of interest to allresearchers in the field of atherothrombosis, tothose seeking further understanding of lipidand lipoprotein metabolism, to those seekingnew targets for drug development and to thoseevaluating the potential of novel biomarkersfor risk assessment. Packard and Rader haveassembled contributions from the world’sexperts in lipid and lipoprotein metabolismand atherogenesis. The chapters offer up-to-date overviews of hot topics and criticalappraisal of the potential of new approaches tothe prevention of vascular disease.

This book on the translation of science intonew therapies for atherothrombosis is timelyand will be of interest to a wide variety ofreaders.

Valentin Fuster, MD, PhDPast President, American Heart Association

President, World Heart FederationDirector, Cardiovascular Institute,

Mount Sinai School of Medicine

Foreword

Prelims 14/7/05 4:44 pm Page xi

Prelims 14/7/05 4:44 pm Page xii

We are pleased to present a new book entitled‘Lipids and Atherosclerosis’. Over the last twodecades, there have been major advances inour understanding of the molecular regulationof lipid and lipoprotein metabolism as well asof the pathogenesis of atherosclerosis. Thenext step is to translate that molecular knowl-edge into new therapeutic approaches to lipiddisorders and atherosclerosis. Therefore, wefelt it appropriate to focus a book on the inter-face between the basic science and new thera-peutic development in this critically importantarea of biology and medicine.

Chapters 1–4 focus on the molecular path-ways by which the liver assembles and secretesvery low density lipoproteins (VLDL). Themicrosomal transfer protein, the acyl CoA:diacylglycerol acyltransferase, and the ACATenzymes all play an important role in thisprocess, and each is represented by a chapterwritten by experts in these areas. In addition, anoverview on the mechanisms of VLDL assemblyin the liver is provided by Professor Olofsson, aninternationally recognized pioneer in this area.

The next 3 chapters examine intestinal lipidmetabolism and bioacid metabolism. Chapterson ABCG5 and ABCG8 as well as on the role ofFXR in bioacid metabolism are included. Anoverview on intestinal lipid metabolism isprovided, written by Professors Miettinen andGylling.

Then follows 4 chapters on: HDL metabolismand reverse cholesterol transport; LXR andregulation of cholesterol efflux; endotheliallipase and the regulation of HDL metabolism;and the role of CETP and its potential as atarget for the development of new therapies.An overview is written by Professor Barter,

internationally recognized for his work in thisarea.

The adipocyte and its role in energy andlipid metabolism is the basis of chapters 12–15.PPARs, the secretory products of adipocytes,and the regulation of adipocyte triglyceridestorage are all discussed in depth. An overviewof insulin resistance in metabolic syndrome isprovided by Professor Sattar, widely known forhis work in this area.

Finally, chapters 16–19 explore the role ofinflammatory pathways in atherosclerosis.Topics included in these chapters are therespective roles of CD44, lipoxygenases, andLp-PLA2 in atherogenesis. An overview of therole of innate immunity in atherosclerosis isprovided by Professor Real, who has long beenrecognized as a key opinion leader in this field.

Our goal has been to encourage authors topresent evidence and to speculate on the inter-face between new knowledge of molecularmechanisms and the potential to translatethose mechanisms into new therapeuticapproaches. Several chapters specifically dis-cuss novel therapies that are currently in clini-cal development or expected to enter theclinical development arena in the near future.Other contributions discuss the potential ofnew molecular targets. We recognize that theareas of lipid metabolism and atherosclerosisare intensively investigated and dynamic andoften changing, and every effort has beenmade to have the information provided as upto date as possible. Our hope is that this bookwill be of interest to individuals interested inthe developments in the areas of lipid metabo-lism and atherosclerosis, and the potential fornew therapies that will result.

Chris J. Packard and Daniel J. Rader

Preface

Prelims 14/7/05 4:44 pm Page xiii

The Editors would like to thank sincerely Linda Watts and Shelley Wilkie for all their help inmanaging the co-ordination of this book with our publishers, Taylor and Francis, and chapterauthors.

Acknowledgements

Prelims 14/7/05 4:44 pm Page xiv

OVERVIEW OF THE SECRETORYPATHWAY

Secretory proteins and integral membraneproteins are synthesized on ribosomes attached

to the surface of the endoplasmic reticulum(Figure 1). During the biosynthesis, the‘nascent’ polypeptide is translocated through achannel starting at the site of synthesis in theribosome and proceeding through a

The assembly of very low densitylipoproteins in the liverS-O. Olofsson and L. Asp

1

Proteasomal degradation

ER exit sites

ARF 1/ COP Ivesicles

ERGIC

ER

Secretion

Plasmamembrane

Golgi complex

Trans-Golginetwork

Sar1/Cop IIvesicles

Figure 1 Overview of the secretory pathway. Secretory proteins are translated on ribosomes attached to the endoplasmicreticulum (ER) and translocated to the lumen of this organelle. After being folded, the proteins leave the ER at exitsites via COP II vesicles. The GTPase Sar1 is involved in the formation of these vesicles. Misfolded proteins areretracted to the cytosol and sorted for proteasomal degradation. The COP II vesicles fuse to form the ERGIC (ER-Golgi intermediate compartment; also referred to as VTC (vesicular tubular cluster)). COP I vesicles, assembledunder the influence of ARF1, are involved in sorting processes in the ERGIC as well as in the next compartment, thecis-Golgi. Well-established COP I-dependent sorting processes are the return of proteins from the later part of thesecretory pathway to the ER. ERGIC fuse with the cis-Golgi, and the proteins are transferred though the differentlevels of the Golgi apparatus to be finally sorted to secretion in the trans-Golgi network

ch01 14/7/05 4:44 pm Page 1

membrane pore to the lumen of the endoplas-mic reticulum (ER). In this way, the nascentchain can reach the lumen of the ER withoutbeing exposed to the cytosol. The membranepore consists mainly of two components, thetrimeric complex Sec 61 (with its α, β and γchains) and Tram. The pore undergoes impor-tant interactions with the nascent polypeptideand influences its correct translocation, itsfolding and its targeting. The pore is gated intwo directions: from the cytosol to the lumenand in the plane of the membrane. The firstgate opens when the nascent chain starts totranslocate to the ER lumen, while the secondgate is involved in selectively allowing themembrane-spanning domains of membraneproteins to enter the hydrophobic portion ofthe membrane. (The translocon and itsfunction have been reviewed elsewhere; see,for example, references 1, 2.)

A chaperone-aided process allows theprotein to fold in the ER and the foldedproteins are sorted into exit sites and allowedto leave the ER by transport vesicles (Figure 1).Misfolded proteins, on the other hand, areretained in the ER and eventually retractedthrough the membrane channel (i.e. Sec 61and Tram) and sorted to proteasomal degrada-tion (Figure 1).1–7

The formation of the transport vesicles andthe correct transport from the ER to the Golgiapparatus are dependent on a series of proteins(these proteins and the transport processeshave been reviewed extensively, and the readeris referred to such reviews for more detailedinformation; see, for example, references 4,8–11). We shall restrict ourselves to a shortdescription of SAR 1 and ARF 1 (Figure 2), aswell as the coat proteins present on the trans-port vesicles that bud under the influence ofthese two proteins. ARF 1 and SAR 1 are smallGTPases which function as ‘switches’ in intra-cellular processes. The proteins are activatedby the exchange of GDP for GTP, a processcatalysed by a GEP (guanine nucleotideexchange protein). Several such GEPs havebeen identified for ARF 1 (for review, see refer-ence 8). These are soluble cytosolic proteins.SAR 1 is activated by one GEP (Sec 12p) which

is an integral membrane protein of the ER (forreview, see reference 8). In the case of ARF 1,the exchange reaction targets the protein tothe microsomal membrane, a process whichinvolves a change in the structure of ARF 1allowing an N-terminally linked myristic acid toleave a hydrophobic pocket and anchor theprotein in the membrane;12 see also reference8 for a review (Figure 2). SAR 1 is not acylatedand does not appear to contain any other lipid-anchoring structure; even so, a large portion ofthe SAR 1 pool is membrane associated (forreview, see reference 8).

There are different so-called coat proteinswhich are involved in the budding of transportvesicles. The assembly of the coat is determinedby SAR1 and ARF 1. Thus, after activation ARF1 (Figure 2) triggers the assembly of the Cop Icoat consisting of several proteins (α-COP, β'-COP, ε-COP, γ1- or γ2-COP, β-COP, δ-COP, ζ1- orζ2-COP), while SAR 1 influences the assembly ofthe Cop II coat, which consists of theSec23/24p and the Sec 13/31p complexes (see,for example, references 4, 8, 9, 11).

Hydrolysis of the GTP bound to ARF 1 andSAR 1 is activated by GAPs (GTPase activatingproteins). The importance of this hydrolysis forthe disassembly of the coat proteins and thesorting and collection of cargo to the vesicleshas been reviewed elsewhere.8,9

Secretory proteins leave the ER at so-calledexit sites, where they are sorted into COP IIvesicles (Figure 1). The process is driven bySAR 1. The COP II vesicles stay close to the ERand fuse to form the ER-Golgi intermediatecompartment (ERGIC; also referred to as thevesicular tubular cluster (VTC)) (Figure 1).ARF 1 and COP I are involved in importantsorting processes in the ERGIC, processeswhich involve the sorting of proteins thatshould be returned to the ER. The ERGIC istransported to the cis-Golgi on microtubulesand fuses with this compartment. It has beendemonstrated that the activity of ARF 1 is ofimportance for the structure of ERGIC and forits ability to mature into the cis-Golgi. ARF 1COP I vesicles are also involved in sortingproteins from the cis-Golgi and back to the ER(for review, see references 4, 13) (Figure 1).

2 LIPIDS AND ATHEROSCLEROSIS

ch01 14/7/05 4:44 pm Page 2

The secretory proteins are exposed to differ-ent levels of the Golgi apparatus (cis-, medialand trans-Golgi) during transfer through thesecretory pathway (Figure 1). The predomi-nant opinion seems to be that, to a largeextent, this occurs through the process ofmaturation of the Golgi cisterns, by which thecis-Golgi cisterns mature into the medial Golgi,which in turn becomes the trans-Golgi (forreview, see reference 14). The proteins arefinally transported from the trans-Golginetwork to the plasma membrane to besecreted (for review, see, for example, refer-ence 4, 14). During exposure to the differentGolgi cisterns, the N-linked carbohydrate(added during the translocation into the ERlumen) is processed and O-linked glycosylationoccurs.

OVERVIEW OF THE BIOSYNTHESISOF apoB-100 AND THE ASSEMBLY OFVERY LOW DENSITY LIPOPROTEINS

apoB-100 is a secretory protein and, as such,follows the secretory pathway. In order to besecreted, however, the protein must be incor-porated into the Very Low DensityLipoproteins (VLDLs). These consist of a coreof neutral lipids (triglycerides and cholesterolesters) surrounded by a monolayer of amphi-pathic structures (phospholipids, unesterifiedcholesterol) to which apoB-100 is bound. It isnow well established that this structure isassembled in two steps (Figure 3), of which thefirst occurs during the translation and translo-cation of apoB-100 to the lumen of theendoplasmic reticulum. In the second step,

THE ASSEMBLY OF VERY LOW DENSITY LIPOPROTEINS IN THE LIVER 3

αβ’ε-COPActivePLD1

GTP

GEP

GTP

ARF1GDP

COP Iproteins

Budding of transport vesicles

PLD 1

Phosphatidylcholine

Phosphatidic acid

Recruitment of proteins

Influencingmembranestructure

ARF1GTP

ARF1GTPARF1

GTPδ-COP ζ-COPγ-COPβ-COP

Figure 2 The activation and functions of ARF 1. ARF 1 is activated by a GEP which exchanges GDP for GTP. Thisleads to a change in the structure of ARF 1 that exposes a myristic acid, which anchors the protein to themicrosomal membrane. In this position, ARF 1 can recruit COP I proteins, thus promoting the budding of transportvesicles. ARF 1 can also activate phospholipase D1 (PLD 1), which converts phosphatidylcholine to phosphatidic acid.It has been proposed that the phosphatidic acid formed recruits proteins involved in, for example, the formation oftransport vesicles or influences the membrane in such a way that budding of transport vesicles is promoted

ch01 14/7/05 4:44 pm Page 3

which occurs outside the rough ER, apoB-100associates with the major proportion of lipids,forming a bona fide VLDL.

The first step: a co-translational partiallipidation of apoB-100 (for reviews, seereferences 15–17)

The first step occurs during the biosynthesis ofapoB-100 and its translocation through thetranslocon to the lumen of the ER (Figure 3).18

During this process, apoB-100 is partiallylipidated, forming a primordial lipoprotein – apre-VLDL. This lipoprotein was isolated fromthe endoplasmic reticulum and shown to be a100 Å particle with the density of HDL.19–21 Theprimordial particle with apoB-100 is retained inthe cell (for discussion, see below), while thatwith apoB-48 is allowed to leave the cell to besecreted.22 The retention is dependent on thestructure between aa 3266 and 4082 in apoB-100 and the interaction with chaperons (seebelow).23 The pre-VLDL contains both triglyc-erides and phospholipids19,21 and is relativelytightly associated with the membrane of theendoplasmic reticulum.24

The microsomal triglyceride transferprotein and the first step in VLDLassembly (for reviews, see references 25,26)

The co-translational lipidation of apoB-100 iscatalysed by a transfer protein which is referredto as the microsomal triglyceride transferprotein (MTP). The importance of MTP forVLDL assembly is illustrated by the observationthat MTP is the gene for abeta-lipopro-teinaemia,27,28 that is, the inability to assembleapoB-containing lipoproteins.

The structure of MTP has been modelled onthe structure of lamprey lipovitellin29 andproposed to contain a hydrophobic pocketwhich is most likely involved in the transfer oflipids.29 Details of the structure of MTP and itsinteraction with apoB-100 will be reviewedelsewhere. The protein has been demonstratedto interact with apoB, and a possible model forthe role of MTP in the co-translation lipidation

of apoB-100 has been proposed.30,31 Thus, MTPhas been proposed to interact with apoB toform lipid-binding pockets which involve theamphipathic β-sheet structures in apoB-10031

(Figure 3). These lipid-binding pockets arethought to acquire the lipids delivered by thetransfer protein (details of these models arepresented in references 26, 30, 31).

ApoB-100 is a very amphipathic molecule,with large regions of amphipathic β-sheet aswell as regions of amphipathic α-helix (for areview of apoB structure, see reference 32). Aco-translational lipidation may be necessary toallow the protein to fold in its correct position,i.e. in a lipid–water interface, when it entersthe lumen of the ER.

The second step in the VLDL assemblyoccurs outside rough ER (for reviews, see references 15–17)

Evidence for a second step in VLDL assemblywas first obtained by immuno-electronmicroscopy,33 where the presence of a non-VLDL form of apoB in the rough ER and thepresence of apoB free ‘lipid droplets’ in thesmooth ER were demonstrated. ApoB-contain-ing VLDL appeared at the junction betweenrough and smooth ER. Our group providedbiochemical18–21,34 and kinetic22 evidence for amodel in which a lipid-poor lipoprotein wasformed during one step and converted to alipid-rich structure in a second step. Recently,it has been possible to verify the presence ofthe two steps in vivo in humans by turnoverstudies based on stable isotopes.35

Very recent studies have demonstrated thatthe second step can be divided in two alterna-tive steps; one leading to VLDL 2 and the otherto VLDL 1.23 The formation of VLDL 2 frompre-VLDL is due to a size dependent lipidation,i.e. the triglycerides were added in proportionto the length of apoB. Based on these observa-tions, we concluded that the dense apoB-48particle was an analogue to VLDL 2 rather thanto the apoB-100 pre-VLDL. This explained whythe apoB-48 particle could be secreted.23 VLDL1 was assembled from VLDL 2 by the additionof a major load of lipid. This process required


ch01 14/7/05 4:44 pm Page 4

that apoB had reached a length of apoB-48 butwas otherwise independent of the length ofapoB.23

ApoB-48 has the advantage that it iscompletely dependent on oleic acid for theassembly of VLDL1 in the rat hepatoma cellline McARH7777. Thus, in the absence ofoleic acid the protein assembles only the denseVLDL2 analogue, while after a short incuba-

tion (15 minutes) with oleic acid the cells startto produce apoB-48 VLDL1. Stillemark et al21

took advantage of this to demonstrate that theassembly of VLDL1 occurred outside the roughER in a smooth membrane compartment thatbanded with the Golgi apparatus upon subcel-lular fractionation. Localization of the secondstep to the Golgi apparatus has been indicatedby the work of Swift and his co-workers.36


Rough

ER

Sar 1/Cop II

ARF 1 dependent

Second step compartment

1 2 3 4

6

5

7

11

8 10

9

VLDL

Figure 3 Overview of the assembly of very low density lipoproteins. ApoB is co-translationally translocated to the lumen ofthe endoplasmic reticulum (ER) (1). During this translocation, apoB interacts with MTP (shown in dark grey infigure) to form lipid-binding pockets (2) which accept the lipids from MTP (3). The lipid-binding pocket is thoughtto enlarge to accept more lipids (4). Thus, at the end of the biosynthesis of apoB a 10 nm primordial particle (apre-VLDL) is formed (5). This pre-VLDL is relatively strongly associated with the ER membrane and leaves the ERwith Sar1/Cop II vesicles (6 and 7) to eventually end up in a post-ER compartment. This compartment is dependenton ARF 1 to be intact (i.e. Golgi or ERGIC). In this ‘second step compartment’, lipid droplets (10) bud from themembrane (9) under the influence of a phospholipase D (see Figure 2). Pre-VLDL (8) fuses (11) with the luminallipid droplet (10). This fusion, the second step, gives rise to the bona fide VLDL

ch01 14/7/05 4:44 pm Page 5

Moreover, an assembly of VLDL outside therough ER is supported by results from Fisher’sgroup37 showing that apoB in a non-VLDLform is present in SAR 1 COP II vesicles thatbud from the ER (Figure 3). As discussedabove, such vesicles are known to be involvedin the exit of secretory and membraneproteins from the ER.

The second step and the formation oflipid droplets

Electron microscopy studies have indicatedthat a fusion between pre-VLDL and a lipid-free droplet may occur in the secretorypathway (Figure 3). The mechanism for theformation of the lipid droplet has still not beenelucidated in detail. We set out to investigatethis process by analysing the formation ofcytosolic lipid droplets. To do this, weconstructed a cell-free system38 that assembled100–400 nm large lipid droplets with triglyc-erides as major lipid and caveolin, vimentinADRP and GRP 78 as major proteins, i.e.proteins that have been reported to be ondroplets in intact cells.39–42

We observed that the formation of the lipiddroplets is dependent on a phospholipase D(PLD) activity.38 We also identified a cytosolicactivator for the formation of lipid droplets,which also activated PLD.

The formation of the droplets was alsodependent on the rate of triglyceride biosyn-thesis;38 however, the rate of triglyceridebiosynthesis was not influenced by the assemblyof lipid droplets. Thus, the assembly of lipid

droplets could be inhibited by omitting theactivator (see above) without any major effecton the rate of the biosynthesis of triglycerides.Instead, the triglycerides remained with themicrosomes (Marchesan et al, work inprogress). These observations indicate thatnewly formed triglycerides can be stored inassociation with the ER/Golgi membrane as analternative to being released as lipid droplets.This lends support to the model for dropletformation that has been put forward by severalauthors.42–44

A model for the formation of cytosolic lipiddroplets is shown in Figure 4A. According to thismodel, the triglycerides that are synthesized inthe cytosolic leaflet of the membrane exceed itssolubility in this amphipathic structure and ‘oilout’ as a lens in the hydrophobic portion of themembrane. There is one important prerequisitefor this to happen. The concentration of trigly-cerides in the leaflet must become sufficientlyhigh. This, in turn, requires a mechanismpreventing the free lateral diffusion of thetriglycerides in the ER membrane. Thus, theremust be local foci of triglyceride accumulation,and the molecule must be prevented fromleaving these foci by lateral diffusion. In thisrespect, it is interesting to note that caveolin isassociated with the lipid droplets and thatcaveolin is also found in regions of membraneswith modified structure. The ‘oiled out’ trigly-cerides bud from the microsomal membraneunder the influence of PLD and the formationof phosphatidic acid (Figure 4A). The role ofPLD and phosphatidic acid in this processremains to be elucidated. Phosphatidic acidappears to be involved in other budding


Figure 4 A tentative model for the assembly of cytosolic (A) and luminal (B) lipid droplets. (1) Triglycerides are synthesizedfrom diacylglycerol and acyl CoA at regions in the microsomal membrane which limit the lateral diffusion of theformed triglyceride. We speculate that caveolin (‘V’ in figure) may be involved in this limitation of lateral diffusion.Due to the restriction in lateral diffusion, the concentration of triglycerides in the luminal leaflet exceeds itssolubility and they oil out as a hydrophobic phase in the fatty acid portion of the membrane, forming a ‘membrane-associated droplet’. Proteins that are of importance for the direction of the budding of the droplet are recruited tothese ‘membrane-associated droplets’. ADRP and vimentin (cytosolic droplets; Figure 4A), and also MTP (luminaldroplets; Figure 4B) are tentative candidates. The size of the ‘membrane-associated droplet’ increases with ongoingtriglyceride biosynthesis (2) and finally phospholipase D is activated by the activator PLD Act. The formation of PAtriggers budding of the droplet from the membrane to the cytosol (Figure 4A) or lumen (Figure 4B). The luminaldroplet fuses with pre-VLDL to form VLDL (see also Figure 3).

ch01 14/7/05 4:44 pm Page 6


Diacylglycerol acyl-transterase

CoA

CoA

ADRPADRP

ADRPADRP

ADRPVimentin

Phospholipase D

Cytosolic activator to the release of lipid droplets and to PLD

Caveolin

GRP 78

Lumen

Cytosol

DGAcylCoa

DGAcylCoa

PCPA

TG

12

4 3

TG

DGAT

DG

AT

DG

AT

VIM

VIM

VIM

VIM

PLD

PLD PLD

GRP 78

GRP 78GRP 78

GRP 78

Diacylglycerol acyl-transterase

ADRP

Phospholipase D

ApoB in pre-VLDL

ApoB in VLDL

Cytosolic activator to the release of lipid droplets and to PLD

Un unknown protein perhaps MTP involved in bringing the lipid droplet into the lumen of the secretory pathway

Caveolin

Lumen

Cytosol

DGAcylCoa

DGAcylCoa PC

PA

VLDL Luminal lipiddroplet

TG1

2

3

4

TG

Pre-VLDL

ADRP

ADRP

UnknownMTP ?

UnknownMTP ?

UnknownMTP ?

UnknownMTP ?

UnknownMTP ?

ADRP

DGAT

DG

AT

DG

AT

ApoB

ApoB

CoA

CoA

PLD

PLD

(A)

(B)

ch01 14/7/05 4:44 pm Page 7

processes such as the generation of transportvesicles.45–48 This has been questioned, however(for review, see reference 49).

More than one explanation for the role ofphosphatidic acid in the budding reaction hasbeen discussed. First, it has been proposed thatthe shape of the phosphatidic acid moleculecould be beneficial for the bending of themembrane during the budding process (see,for example, reference 50). Second, it has beensuggested that phosphatidic acid is involved inthe recruitment of proteins needed for vesicu-lar trafficking.51 Indeed, the concept of recruit-ment of proteins by phosphatidic acid has beenextended to other pathways that involve PLDactivation.52 It is tempting to suggest that themechanisms discussed above are also involvedin the budding of lipid droplets.

The observation that PLD is involved in theformation of lipid droplets is interesting, sincewe have obtained results suggesting that PLD 1is also involved in the assembly of VLDL.34

Since cytosolic lipid droplets have a structurevery similar to that of lipoproteins (i.e. ahydrophobic core surrounded by a monolayerof amphipathic structures), it is not unlikelythat the mechanism involved in the assembly ofcytosolic lipid droplets is the same as thatinvolved in the formation of lipid droplets thatbud into the lumen of the secretory pathway(Figure 4B; see also references 43, 44). Indeed,the only difference may be the availability ofproteins that determine to which side of theER/Golgi membrane the droplet should bud.Another possibility is that the localization ofthe DGAT reaction, i.e. in the cytosolic orluminal leaflet,53,54 is of importance for thedirection of the budding.

It has been pointed out that MTP promotesthe entry of triglycerides into the secretorypathway.55,56 Thus, MTP may be a possiblecandidate for promoting the budding of lipiddroplets into the secretory pathway. The MTPactivity is required for 15–30 minutes after thetranslation of apoB-100 has been completed, inorder for VLDL to be formed.24 One interpre-tation is that the assembled pre-VLDL needs toaccumulate lipid post-translationally. However,the observations would also be compatible with

a role of MTP in the second step. One prereq-uisite is that the second step should be ratelimiting in the secretion process. This appearsto be the case, since the primordial particle isproduced in excess and a fusion with the lipidsin the second step is a prerequisite for secre-tion of apoB-100 (i.e. the lipid droplets formeddo not accumulate but are constantly secretedas VLDL). A role for MTP in the second step isalso suggested by the observation that theprotein is localized in the Golgi appara-tus,36,57,58 as well as by recent results fromBjörkegren and his group.59

The major proportion of the triglyceridesfound in VLDL is derived from triglyceridespresent in cytosolic lipid droplets. These tri-glycerides are hydrolysed and the fatty acidsare re-esterified into new triglycerides beforebeing incorporated into VLDL.60–62 Inter-estingly, Lehner and Vance and co-workershave identified a triglyceride hydrolase63,64 thatappears to be coupled to the rate of assemblyof VLDL. Thus, overexpression of the enzymeincreases the assembly, while inhibitiondecreases it. Also, Pease and his co-workershave identified a triglyceride hydrolase thatseems to influence the assembly of VLDL.65

The importance of small GTPases forVLDL assembly

SAR 1 appears to be the gene responsible forfailure to assemble the triglyceride-richchylomicrons in the intestine.66 Interestingly,Fisher and his co-workers37 demonstrated thatapoB was sorted into unique SAR 1 Cop IIvesicles that did not contain any other secre-tory proteins. Together, these results may pointto the possibility that there is a unique trans-port system that allows apoB to leave the ERand reach the second-step compartment.

We demonstrated that ARF 1 (Figure 2) isinvolved in the assembly of VLDL.34 Thus, adominant negative mutant of ARF1 (T31N)inhibited the formation of VLDL under condi-tions in which the production of apoB-100 wasonly slightly decreased. How could ARF 1 influ-ence the second step in the assembly? There is


ch01 14/7/05 4:44 pm Page 8

more than one possibility. First, ARF 1 could act as a switch to turn on enzymes ofimportance for the assembly process. One suchenzyme that is activated by ARF 1 is PLD 1 (seeabove; Figure 2). Thus, one possibility is thatARF 1 influences the assembly of VLDL byactivating PLD 1.34

Another possible role of ARF 1 in the assem-bly process could involve its function in sortingprocesses (Figure 2), processes that are essen-tial for the integrity of organelles in the secre-tory pathway. Thus it is well known that thedominant negative mutant of ARF1, T31N,gives rise to disintegration of the Golgi appara-tus and a redistribution of components fromthis organelle into the ER, ER exit sites and thecytoplasm.67 It is therefore possible that inhibi-tion of ARF 1 results in a loss of the compart-ment for the second step. Indeed recentresults68 demonstrate that the importance ofARF 1 in the assembly of VLDL 1 can beexplained by the role of ARF 1 in the transportbetween ERGIC and cis-Golgi. Thus a decreasein the assembly coincided with a decreasedtransport between ERGIC and cis-Golgi. Thiswould fit with the observation that the secondstep occurs in a compartment outside therough ER21 and that apoB in a non-VLDL formleaves the ER with COP II vesicles.37 Thus thesecond step in the assembly of VLDL 1 appearsto occur in the Golgi apparatus.68

CO- AND POST-TRANSLATIONALDEGRADATION OF apoB-100

It has long been known that newly synthesizedapoB-100 undergoes intracellular degradation(for review, see references 15, 17, 69). This isless pronounced in primary hepatocytes,70 butis an impressive phenomenon in some hepa-toma cell lines such as Hep G2 cells, wherealmost 80% of the newly synthesized apoB-100can be removed by such degradation.71 Thedegradation was dramatically reduced (from80% to 20%) when oleic acid was included inthe culture medium.20,71 The intracellulardegradation of apoB-100 occurs at three differ-ent levels:72,73

(1) Co-translationally by a mechanism thatinvolves retraction through the translocon,ubiquitination and proteasomal degrada-tion;72,74–76

(2) Post-translationally by an unknown mecha-nism that could be promoted by culturingthe cells in the presence of polyunsatu-rated fatty acids.72 This degradation seemsto occur in a compartment separated fromrough ER and has therefore been referredto as post-ER pre-secretory proteolysis(PERPP);72

(3) A reuptake from the unstirred water layeraround the outside of the plasma mem-brane77 via the LDL-receptor.

The LDL-receptor has been shown to have animportant role in regulation of the secretion ofapoB-100-containing lipoproteins.78,79 There iseven evidence that the effect of the receptor isnot only due to the interaction with apoB-100on the cell surface, but that the receptor andapoB-100 interact early in the secretorypathway and that this interaction is of impor-tance for the post-translational degradation ofapoB-100.59,79,80

It is well known that secretory proteins whichmisfold undergo proteasomal degradation (seeabove). Such proteins interact with chaperoneproteins and are unfolded and retractedthrough sec 61. In the cytosol, the retractedproteins interact with cytosolic chaperones andthe proteins are conjugated with ubiquitin andsorted to proteasomal degradation. In the caseof apoB-100, the sorting to proteasomal degra-dation seems to occur co-translationally, i.e.the apoB-100 nascent chain is exposed to thecytosol, unbiquitinated and sorted to proteaso-mal degradation (for review, see reference 73).The co-translational degradation is influencedby the availability of lipids and the activity ofMTP. Thus, when the availability of lipids orMTP activity is limiting, apoB-100 remainsassociated with the translocon and is sorted toproteasomal degradation.

PERPP is less well understood, thus nothing isknown about the enzyme systems involved orthe sorting of apoB-100 for this degradation. Itappears that the process is promoted by ω-3 fatty


ch01 14/7/05 4:44 pm Page 9

acids. It has also been demonstrated that apoB-100 interacts with the protease/chaperone ER60,81,82 and that this interaction is linked to theintracellular degradation of newly synthesizedapoB-100. The role of this interaction for thePERPP has not been elucidated.

Another potentially important molecule isthe LDL-receptor. Thus, Attie and his co-workers have demonstrated that the LDL-receptor interacts with apoB in the secretorypathway.79,80 In elegant studies, the ligandbinding domain of the receptor was fused to aKDEL sequence that retained the construct inthe ER. This construct, as well as other‘naturally occurring’ mutants that retained thereceptor in the ER, promoted the pre-secretorydegradation of apoB-100.80

The apoB-100 primordial particle (pre-VLDL)is not secreted from the cell to any significantdegree but appears to be retained and degraded.On the contrary, the dense apoB-48 particle issecreted from the cell as a lipoprotein within theHDL density region. The retention started whenapoB reached a size between 70 and 80% of thesize of apoB-100 (i.e. between apoB-70 and apoB-80).23 This is the region of the molecule thatcontains the LDL-receptor binding site.However, the LDL-receptor was not involved inthe retention of apoB-100 pre-VLDL. The reten-tion of the pre-VLDL particle coincided with thebinding of BiP (binding proteins) and PDI(protein disulphide isomerase) to the particle.One possibility is that the C-terminus of apoB-100 misfolds on the pre-VLDL, which leads tointeraction with BiP and retention in the ER.23

The pre-secretory degradation of apoB-100pre-VLDL points to the existence of a degrada-tional pathway that can cope with bothproteins and lipids. One candidate is thelysosomes, and it has been shown that lysoso-mal inhibitors prevent the intracellular degra-dation of newly formed apoB-100.83

In summary, there are several levels in thesecretory pathway where apoB-100 and apoB-100 pre-VLDL are sorted to degradation. Thisdegradation is important for the secretion ofVLDL, but it is not yet clear whether the degra-dation per se is regulated or if it is a result offailure in the assembly process.

THE EFFECT OF PEROXISOMEPROLIFERATOR ACTIVATEDRECEPTOR ALPHA AGONIST ONTHE ASSEMBLY OF VLDL: ASURPRISING MECHANISM LEADINGTO SMALL VLDL

Several clinical studies have demonstrated thatfibrates (i.e. proliferator activated receptor α(PPAR-α) agonists) give rise to a decrease inboth the amount of large VLDL (VLDL1) andin the amount of small dense LDL (for a recentreview, see reference 84). Part of the mecha-nism was revealed by investigating the role ofPPAR-α agonists in primary rat hepatocytes.70

The studies revealed that the agonists actuallyincrease biosynthesis and secretion of apoB-100, while no changes were seen in the biosyn-thesis and secretion of apoB-48 or albumin.The mechanism behind the increased produc-tion of apoB-100 was an inhibition of the co-translational degradation. This decrease in theco-translational degradation could not be dueto changes in lipid metabolism, since theagonists decreased the triglyceride biosynthe-sis. Instead it was due to an increase in theamount of MTP due to an increased transcrip-tion of the gene.85

The increased production of apoB-100 iscoupled to a decreased rate of triglyceridebiosynthesis induced by the PPAR-α agonist.70

Thus, the secreted apoB-100-containinglipoproteins become smaller, and thereforehave the capacity to be converted to large LDLwith a much shorter residence time than smalldense LDL (for reviews, see references 84,86–88). Thus, treatment with PPAR-α agonistswill lead to decreased levels of plasma apoB-100, in spite of the increased production ofapoB-100. This process is accentuated by thedecrease in apoC-III induced by the agonists(for reviews, see references 84, 89).

DRUG TARGETS IN VLDL ASSEMBLY

When identifying targets for drug therapy inthe VLDL assembly pathway, it should be keptin mind that the assembly of VLDL is linked to


ch01 14/7/05 4:44 pm Page 10

the removal of triglycerides from the liver.Thus, interference may result in steatosisunless the triglyceride biosynthesis is decreasedand/or the oxidation of fatty acids is increased.One example of complications due to the viola-tion of this basic principle comes from thetreatment with MTP inhibitors. Obviously,MTP is a prime drug target in the assemblyprocess. Inhibition of MTP increases theremoval of nascent apoB-100, thereby decreas-ing the assembly of VLDL. Moreover, it is possi-ble that inhibition of MTP prevents the entryof lipid droplets into the secretory pathway,thereby inhibiting the second step. However,MTP inhibitors do not interfere with trigly-ceride biosynthesis and fatty acid oxidation andtherefore the accumulation of triglycerides inthe liver cell increases, leading to steatosis. Thisis one of the reasons why no MTP inhibitor hasappeared on the market.

A more appealing strategy would be tomanipulate the structure of VLDL, allowingthe removal of triglycerides from the cells bysmall VLDLs. Such VLDLs have a more rapidturnover in plasma and give rise to large LDLswith short residence time in plasma. Thus,provided that the LDL-receptor is intact, thechange from large to small VLDL decreasesplasma apoB-100 and both triglycerides andLDL cholesterol. To achieve this, PPAR-αappears to be an important target (see discus-sion above). PPAR-α also has the advantagethat it inhibits triglyceride biosynthesis andincreases the β-oxidation of fatty acids.

The second step in the assembly of VLDL isan interesting target, since prevention of thelipidation of apoB-100 pre-VLDL leads to acomplete retention and intracellular degrada-tion of this primordial particle. However, steato-sis is once again a threat. Moreover, a great dealof basic information will be needed before arational approach to identifying specific drugtargets can be taken. Neither Sar 1 nor ARF 1appear to be useful targets for therapy, since

both are switches in essential processes in thecell. An exception may be if a GEP or GAP thatis specific for the apoB-100 related function ofARF1 or Sar 1 can be identified. Such aGEP/GAP may be a potential drug target. Also,the COP proteins and PLD carry out essentialfunctions in the cell and are therefore unlikelydrug targets for VLDL assembly. An interestingtarget would be the activator for the release oflipid droplets to the cytosol (and PLD) which wehave identified. Inhibitors of this activator havethe potential to inhibit not only the formationof cytosolic lipid droplets, but also the secondstep in the VLDL assembly. However, it is ofcourse essential to first clarify the relationshipbetween the budding of lipid droplets to thecytosol and to the lumen of the ER, and theimportance of these processes for the secondstep. Moreover, the activator must be identifiedand its role in metabolism and other functionsin the cell still has to be elucidated. Finally, werequire more information about the fate oftriglycerides and fatty acids under conditions inwhich lipid droplets are not formed. Interestingtargets that should be explored are enzymes intriglyceride biosynthesis (such as glycerol-3-phosphate acyl-transferase (GPAT) and diacyl-glycerol acyltransferase (DGAT)).

In summary, we believe that PPAR-α is a well-established, but still promising drug target inthe short term. In the longer term, interfer-ence with the second step in VLDL assemblyshould be explored.

ACKNOWLEDGEMENTS

This study was supported by grant No 7142from the Swedish Medical Research Counciland by the Swedish Heart and LungFoundation, Novo Nordic Foundation, theSöderberg Foundation and the SwedishStrategic Funds (National Network andGraduate School for Cardiovascular Research).


ch01 14/7/05 4:44 pm Page 11

1. Johnson AE, van Waes MA. The translocon: adynamic gateway at the ER membrane. Annu Rev CellDev Biol 1999; 15:799–842

2. Johnson AE, Haigh NG. The ER translocon and retro-translocation: is the shift into reverse manual orautomatic? Cell 2000; 102:709–12

3. Ellgaard L, Molinari M, Helenius A. Setting thestandards: quality control in the secretory pathway.Science 1999; 286:1882–8

4. Lippincott-Schwartz J, Roberts TH, Hirschberg K.Secretory protein trafficking and organelle dynamicsin living cells. Annu Rev Cell Dev Biol 2000;16:557–89

5. Ellgaard L, Helenius A. ER quality control: towards anunderstanding at the molecular level. Curr Opin CellBiol 2001; 13:431–7

6. Ellgaard L, Helenius A. Quality control in theendoplasmic reticulum. Nat Rev Mol Cell Biol 2003;4:181–91

7. Kostova Z, Wolf DH. For whom the bell tolls: proteinquality control of the endoplasmic reticulum and theubiquitin-proteasome connection. Embo J 2003;22:2309–17

8. Barlowe C. Traffic COPs of the early secretorypathway. Traffic 2000; 1:371–7

9. Spang A. ARF1 regulatory factors and COPI vesicleformation. Curr Opin Cell Biol 2002; 14:423–7

10. Bonifacino JS, Lippincott-Schwartz J. Coat proteins:shaping membrane transport. Nat Rev Mol Cell Biol2003; 4:409–14

11. Haucke V. Vesicle budding: a coat for the COPs.Trends Cell Biol 2003; 13:59–60

12. Goldberg J. Structural basis for activation of ARFGTPase: mechanisms of guanine nucleotideexchange and GTP-myristoyl switching. Cell 1998;95:237–48

13. Gorelick FS, Shugrue C. Exiting the endoplasmicreticulum. Mol Cell Endocrinol 2001; 177:13–8

14. Storrie B, Nilsson T. The Golgi apparatus: balancingnew with old. Traffic 2002; 3:521–9

15. Olofsson S-O, Asp L, Borén J. The assembly and secre-tion of apolipoprotein B-containing lipoproteins.Curr Opin Lipidol 1999; 10:341–6

16. Olofsson S-O, Stillemark-Billton P, Asp L. The intra-cellular assembly of VLDL – a process that consists oftwo major steps that occur in separate cell compart-ments. Trends Cardiovasc Med 2000; 10:338–45

17. Shelness GS, Sellers JA. Very-low-density lipoproteinassembly and secretion. Curr Opin Lipidol 2001;12:151–7

18. Borén J, Graham L, Wettesten M, et al. The assemblyand secretion of apoB 100-containing lipoproteins inHep G2 cells. ApoB 100 is cotranslationally integratedinto lipoproteins. J Biol Chem 1992; 267:9858–67

19. Boström K, Borén J, Wettesten M, et al. Studies on the

assembly of apo B-100-containing lipoproteins inHepG2 cells. J Biol Chem 1988; 263:4434–42

20. Borén J, Rustaeus S, Wettesten M, et al. Influence oftriacylglycerol biosynthesis rate on the assembly ofapoB-100-containing lipoproteins in Hep G2 cells.Arterioscler Thromb 1993; 13:1743–54

21. Stillemark P, Borén J, Andersson M, et al. The assem-bly and secretion of apolipoprotein-B48-containingvery low density lipoproteins in McA-RH7777 cells. JBiol Chem 2000; 275:10506–13

22. Borén J, Rustaeus S, Olofsson S-O. Studies on theassembly of apolipoprotein B-100- and B-48-contain-ing very low density lipoproteins in McA-RH7777cells. J Biol Chem 1994; 269:25879–88

23. Stillemark-Billton P, Beck C, Boren J, et al. Relation ofthe size and intracellular sorting of apoB to theformation of VLDL 1 and VLDL 2. J Lipid Res 2005;46:104–14

24. Rustaeus S, Stillemark P, Lindberg K, et al. Themicrosomal triglyceride transfer protein catalyzes thepost-translational assembly of apolipoprotein B-100very low density lipoprotein in McA-RH7777 cells. JBiol Chem 1998; 273:5196–203

25. Gordon DA, Jamil H. Progress towards understandingthe role of microsomal triglyceride transfer protein inapolipoprotein-B lipoprotein assembly. BiochimBiophys Acta 2000; 1486:72–83

26. Hussain MM, Shi J, Dreizen P. Microsomal triglyc-eride transfer protein and its role in apoB-lipoproteinassembly. J Lipid Res 2003; 44:22–32

27. Wetterau JR, Aggerbeck LP, Bouma M-E, et al.Abscence of microsomal triglyceride transfer proteinin indivals with abetalipoproteinemia. Nature 1992;258:999–1001

28. Gregg RE, Wetterau JR. The molecular basis of abetal-ipoproteinemia. Curr Opin Lipidol 1994; 5:81–6

29. Read J, Anderson TA, Ritchie PJ, et al. A mechanismof membrane neutral lipid acquisition by the micro-somal triglyceride transfer protein. J Biol Chem 2000;275:30372–7

30. Segrest JP, Jones MK, Dashti N. N-terminal domain ofapolipoproteinsB has structural homology tolipovitellin and microsomal triglyceride transferprotein: a ‘lipid pocket’ model for self-assembly ofapoB-containing lipoprotein particles. J Lipid Res1999; 40:1401–16

31. Dashti N, Gandhi M, Liu X, et al. The N-terminal1000 residues of apolipoprotein B associate withmicrosomal triglyceride transfer protein to create alipid transfer pocket required for lipoprotein assem-bly. Biochemistry 2002; 41:6978–87

32. Segrest JP, Jones MK, De Loof H, et al. Structure ofapolipoprotein B-100 in low density lipoproteins. JLipid Res 2001; 42:1346–67

33. Alexander CA, Hamilton RL, Havel RJ. Subcellular


References

ch01 14/7/05 4:44 pm Page 12

localization of B apoprotein of plasma lipoproteins inrat liver. J Cell Biol 1976; 69:241–63

34. Asp L, Claesson C, Borén J, et al. ADP-ribosylationfactor 1 and its activation of phospholipase D areimportant for the assembly of very low densitylipoproteins. J Biol Chem 2000; 275:26285–92

35. Adiels M, Packard CJ, Caslake MJ, et al. A newcombined multicompartmental model forapolipoprotein B100 and triglyceride metabolism inVLDL subfractions. Lipid Res 2005; 46:58–67

36. Valyi-Nagy K, Harris C, Swift LL. The assembly ofhepatic very low density lipoproteins: evidence of arole for the Golgi apparatus. Lipids 2002; 37:879–84

37. Gusarova V, Brodsky JL, Fisher EA. ApolipoproteinB100 exit from the ER is COPII dependent and itslipidation to very low density lipoprotein occurs post-ER. J Biol Chem 2003; 278:48051–8

38. Marchesan D, Rutberg M, Andersson L, et al. Aphospholipase D-dependent process forms lipiddroplets containing caveolin, adipocyte differentia-tion-related protein, and vimentin in a cell-freesystem. J Biol Chem 2003; 278:27293–300

39. Franke WW, Hergt M, Grund C. Rearragement ofvimentin cytoskeleton durin adipose conversion:formation of an intermediate filament cage aroundlipid globules. Cell 1987; 49:131–41

40. Londos C, Brasaemle DL, Schultz CJ, et al. Perilipins,ADRP, and other proteins that associate with intracel-lular neutral lipid droplets in animal cells. Semin CellDev Biol 1999; 10:51–8

41. Prattes S, Horl G, Hammer A, et al. Intracellulardistribution and mobilization of unesterified choles-terol in adipocytes: triglyceride droplets aresurrounded by cholesterol-rich ER-like surface layerstructures. J Cell Sci 2000; 113 (Pt 17):2977–89

42. Brown DA. Lipid droplets: Proteins floating on a poolof fat. Curr Biol 2001; 11:R446–9

43. Olofsson S-O, Bjursell G, Boström K, et al.Apolipoprotein B: structure, biosynthesis and role inthe lipoprotein assembly process. Atherosclerosis1987; 68:1–17

44. Murphy DJ, Vance J. Mechanisms of lipid-body forma-tion. Trends Biochem Sci 1999; 24:109–15

45. Bi K, Roth MG, Ktistakis T. Phosphatidic acid forma-tion by phospholipase D is required for transportfrom the endoplasmic reticulum to the Golgicomplex. Curr Biol 1997; 7:301–7

46. Humeau Y, Vitale N, Chasserot-Golaz S, et al. A rolefor phospholipase D1 in neurotransmitter release.Proc Natl Acad Sci USA 2001; 98:15300–5

47. Vitale N, Caumont AS, Chasserot-Golaz S, et al.Phospholipase D1: a key factor for the exocytoticmachinery in neuroendocrine cells. Embo J 2001;20:2424–34

48. Pathre P, Shome K, Blumental-Perry A, et al.Activation of phospholipase D by the small GTPaseSar1p is required to support COPII assembly and ERexport. Embo J 2003; 22:4059–69

49. Jones D, Morgan C, Cockcroft S. Phospholipase Dand membrane traffic. Potential roles in regulatedexocytosis, membrane delivery and vesicle budding.Biochim Biophys Acta 1999; 1439:229–44

50. Kooijman EE, Chupin V, de Kruijff B, et al.Modulation of membrane curvature by phosphatidicacid and lysophosphatidic acid. Traffic 2003;4:162–74

51. Manifava M, Thuring JW, Lim ZY, et al. Differentialbinding of traffic-related proteins to phosphatidicacid- or phosphatidylinositol (4,5)- bisphosphate-coupled affinity reagents. J Biol Chem 2001;276:8987–94

52. Ktistakis NT, Delon C, Manifava M, et al.Phospholipase D1 and potential targets of its hydroly-sis product, phosphatidic acid. Biochem Soc Trans2003; 31:94–7

53. Owen MR, Corstorphine CC, Zammit VA. Overt andlatant activities of diacylglycerol acyltransfrase in ratliver microsomes: possible roles in very-low densitylipoprotein triglycerol secretion. Biochem J 1997;323:17–21

54. Owen M, Zammit VA. Evidence for overt and latentforms of DGAT in rat liver microsomes. Implicationsfor the pathways of triacylglycerol incorporation intoVLDL. Biochem Soc Trans 1997; 25:21S

55. Raabe M, Véniant MM, Sullivan MA, et al. Analysingthe role of microsomal triglyceride transfer protein inthe liver with tissue-specific knockout mice. J ClinInvest 1999; 103:1287–98

56. Wang Y, Tran K, Yao Z. The activity of microsomaltriglyceride transfer protein is essential for accumula-tion of triglyceride within microsomes in McA-RH7777 cells. A unified model for the assembly ofvery low density lipoproteins. J Biol Chem 1999;274:27793–800

57. Levy E, Stan S, Delvin E, et al. Localization of micro-somal triglyceride transfer protein in the Golgi: possi-ble role in the assembly of chylomicrons. J Biol Chem2002; 277:16470–7

58. Swift LL, Zhu MY, Kakkad B, et al. Subcellular local-ization of microsomal triglyceride transfer protein. JLipid Res 2003; 44:1841–9

59. Larsson SL, Skogsberg J, Bjorkegren J. The lowdensity lipoprotein receptor prevents secretion ofdense apoB100–containing lipoproteins from theliver. J Biol Chem 2003; 279:831–6

60. Wiggins D, Gibbons GF. The lipolysis/esterificationcycle of hepatic triacylglyecerol. Its role in the secre-tion of very-low-density lipoprotein and its response tohormones and sulphonylureas. Biochem J 1992;284:457–62

61. Salter AM, Wiggins D, Sessions VA, et al. The intra-cellular triacylglycerol/fatty acid cycle: a comparisonof its activity in hepatocytes which secrete exclusivelyapolipoprotein (apo) B100 very-low-density lipopro-tein (VLDL) and in those which secrete predomi-nantly apoB48 VLDL. Biochem J 1998; 332 (Pt 3):667–72


ch01 14/7/05 4:44 pm Page 13

62. Gibbons GF, Islam K, Pease RJ. Mobilisation of triacyl-glycerol stores. Biochim Biophys Acta 2000;1483:37–57

63. Lehner R, Vance DE. Cloning and expression of acDNA encoding a hepatic microsomal lipase thatmobilizes stored triacylglycerol. Biochem J 1999; 343(Pt 1):1–10

64. Gilham D, Ho S, Rasouli M, et al. Inhibitors of hepaticmicrosomal triacylglycerol hydrolase decrease verylow density lipoprotein secretion. Faseb J 2003;17:1685–7

65. Trickett JI, Patel DD, Knight BL, et al.Characterization of the rodent genes for arylac-etamide deacetylase, a putative microsomal lipase,and evidence for transcriptional regulation. J BiolChem 2001; 276:39522–32

66. Jones B, Jones EL, Bonney SA, et al. Mutations in aSar1 GTPase of COPII vesicles are associated withlipid absorption disorders. Nat Genet 2003; 34:29–31

67. Ward TH, Polishchuk RS, Caplan S, et al.Maintenance of Golgi structure and functiondepends on the integrity of ER export. J Cell Biol2001; 155:557–70

68. Asp L, Magnusson B, Rutberg M, et al. Role of ADPribosylation factor 1 in the assembly and secretion ofApoB-100-containing lipoproteins. ArteriosclerThromb Vasc Biol 2005; 25:566–70

69. Davidson NO, Shelness GS. Apolipoprotein B: mRNAediting, lipoprotein assembly, and presecretory degra-dation. Ann Rev Nutr 2000; 20:169–93

70. Lindén D, Lindberg K, Oscarsson J, et al. Influence ofagonists to the peroxisome proliferator activatedreceptor a (PPARa) on the intracellular turnover andsecretion of apoB-100 and apoB-48. J Biol Chem 2002;277:23044–53

71. Boström K, Wettesten M, Borén J, et al. Pulse-chasestudies of the synthesis and intracellular transport ofapolipoprotein B-100 in Hep G2 cells. J Biol Chem1986; 261:13800–6

72. Fisher EA, Pan M, Chen X, et al. The triple threat tonascent apolipoprotein B. Evidence for multiple,distinct degradative pathways. J Biol Chem 2001;276:27855–63

73. Fisher EA, Ginsberg HN. Complexity in the secretorypathway: the assembly and secretion of apolipopro-tein B-containing lipoproteins. J Biol Chem 2002;277:17377–80

74. Mitchell DM, Zhou M, Pariyarath R, et al. ApoproteinB 100 has a prolonged interaction with the transloconduring which its lipidation and translocation changefrom dependence on the microsomal triglyceridetransfer protein to independence. Proc Natl Acad SciUSA 1998; 95:14733–8

75. Liang J-S, Wu X, Fisher EA, et al. The amino-terminaldomain of apolipoprotein B does not undergo retro-grade translocation from the endoplasmic reticulumto the cytosol. Proteasomal degradation of nascentapolipoprotein B begins at the carboxyl terminus ofthe protein, while apolipoprotein B is still in its

original translocon. J Biol Chem 2000; 275: 32003–10

76. Pariyarath R, Wang H, Aitchison JD, et al. Co-transla-tional interactions of apoprotein B with the ribosomeand translocon during lipoprotein assembly ortargeting to the proteasome. J Biol Chem 2001;276:541–50

77. Williams KJ, Brocia RW, Fisher EA. The unstirredwater layer as a site of control of apolipoprotein Bsecretion. J Biol Chem 1990; 265:16741–4

78. Horton JD, Shimano H, Hamilton RL, et al.Disruption of LDL receptor gene in transgenicSREBP-1a mice unmask hyperlipidemia resultingfrom production of lipid-rich VLDL. J Clin Invest1999; 103:1067–76

79. Twisk J, Gillian-Daniel DL, Tebon A, et al. The role ofthe LDL receptor in apolipoprotein B secretion. JClin Invest 2000; 105:521–32

80. Gillian-Daniel DL, Bates PW, Tebon A, et al.Endoplasmic reticulum localization of the low densitylipoprotein receptor mediates presecretory degrada-tion of apolipoprotein B. Proc Natl Acad Sci USA2002; 99:4337–42

81. Adeli K, Macri J, Mohammadi A, et al. ApolipoproteinB is intracellularly associated with an ER-60 proteasehomologue in HepG2 cells. J Biol Chem 1997;272:22489–94

82. Taghibiglou C, Rashid-Kolvear F, Van Iderstine SC, etal. Hepatic very low density lipoprotein-ApoB overpro-duction is associated with attenuated hepatic insulinsignaling and overexpression of protein-tyrosinephosphatase 1B in a fructose-fed hamster model ofinsulin resistance. J Biol Chem 2002; 277:793–803

83. Fleming JF, Spitsen GM, Hui TY, et al. Chinesehamster ovary cells require the coexpressionof micro-somal triglyceride transfer protein and cholesterol7alfa hydroxylase for the assembly and secretion ofapolipoprotein B-containing lipoproteins. J BiolChem. 1999; 274:9509–14

84. Packard CJ. Triacylglycerol-rich lipoproteins and thegeneration of small, dense low-density lipoprotein.Biochem Soc Trans 2003; 31:1066–9

85. Kang S, Spann NJ, Hui TY, et al. ARP-1/COUP-TF IIdetermines hepatoma phenotype by acting as both atranscriptional repressor of microsomal triglyceridetransfer protein and an inducer of CYP7A1. J BiolChem 2003; 278:30478–86

86. Berneis KK, Krauss RM. Metabolic origins and clinicalsignificance of LDL heterogeneity. J Lipid Res 2002;43:1363–79

87. Kwiterovich PO, Jr. Clinical relevance of the biochem-ical, metabolic, and genetic factors that influence low-density lipoprotein heterogeneity. Am J Cardiol 2002;90:30i–47i

88. Taskinen MR. Diabetic dyslipidemia. AtherosclerSuppl 2002; 3:47–51

89. Staels B, Dallongeville J, Auwerx J, et al. Mechanismof action of fibrates on lipid and lipoprotein metabo-lism. Circulation 1998; 98:2088–93


ch01 14/7/05 4:44 pm Page 14

BACKGROUND

Dietary fat provides a major source of nutri-tion, but may in excess lead to obesity, insulinresistance, high blood lipid levels and athero-sclerosis. In this chapter, I review the role ofthe microsomal triglyceride transfer protein(MTP) in co-ordinating whole-body lipidhomeostasis, and address the prospects forreducing blood lipid levels through a tissue-specific approach that reduces MTP mass. Thereason for this focus is because the past 5 yearshave seen remarkable progress in our under-standing of the function of MTP, and of itspartner protein disulphide isomerase (PDI), inthe assembly of chylomicrons (Cms) in thesmall intestine, very low density lipoproteins(VLDLs) in the liver, and of low densitylipoprotein (LDL)-like particles in cardiacmyocytes. The reader is referred to Chapters 1and 5 of this book, and other excellentreviews1–8 for further information onapolipoprotein (apo)B, the obligatory proteincomponent of Cm, VLDL and LDL (Figure 1).

In humans, serum cholesterol and trigly-ceride levels are primarily determined by aseries of metabolic pathways, ligands andreceptors that operate in the small intestine,liver, adipose tissue, skeletal muscle andcardiac myocytes (Figure 1). Dietary lipidsinitially enter the circulation in Cms, wherethey may provide peripheral tissues, includingthe heart, with an important source of energythrough the β-oxidation of fatty acids. In thepost-prandial period, a proportion of these

dietary lipids re-enter the circulation for re-distribution in VLDL particles,9,10 which areassembled in the liver (Figure 1). VLDL alsotransports non-dietary lipids formed from thecatabolism of dietary carbohydrate, therecycling of cellular membranes and the ester-ification of free fatty acids, which may derivefrom adipose tissue.11–13 Similarly, the heartsecretes LDL-sized lipoproteins14,15 to removeexcess lipids from this organ, which mightotherwise adversely affect contractility andconductivity.16–19

Recognition that MTP represents a molecu-lar target for lowering blood lipid levels arosefrom the discovery that this protein promotesthe assembly of apoB-containing lipopro-teins,20–22 and that mutations of MTP are themajor, if not sole, cause of abetalipopro-teinaemia.23–28 Abetalipoproteinaemia typicallypresents in infancy with failure to thrive due tosevere fat malabsorption or, in older children,with an atypical pigmented retinopathy associ-ated with spinocerebellar degeneration.29 Thecondition is biochemically characterized by avirtual absence of apoB-containing lipopro-teins in blood and deficiencies of the fat-soluble vitamins A, K and E. Once absorbedfrom the intestine, these vitamins are normallytransported to the liver via a Cm pathway. Fromthe liver, vitamin E is secreted on VLDL,30,31

whereas vitamins A and K have the capacity touse different transport systems.32,33 In theabsence of treatment, patients with abeta-lipoproteinaemia develop a range of debilitat-ing clinical symptoms (e.g. severe generalized

Cardiovascular implications of partial,tissue-specific silencing of MTPC.C. Shoulders

2

ch02 14/7/05 4:45 pm Page 15

weakness, kyphoscoliosis and lordosis), but itremains unclear whether these derive exclu-sively from a dependency on MTP function inaffected tissues or whether they are solely dueto deficiencies of fat-soluble vitamins, or acombination of both. In some patients, highdoses of the fat-soluble vitamins ameliorate theneurological and ophthamological manifesta-tions of the condition.29,30,34–36

THE CASE FOR PARTIAL, TISSUE-SPECIFIC SILENCING OF MTP

All of the MTP inhibitors that have beenstudied to date are conceptually the same: theytarget MTP protein, and possibly apoB, tolower the production of potentially athero-genic lipoproteins, and most probably LDL-sized lipoproteins from the heart. However, itis still not known whether long-term treatment

of hyperlipidaemia with MTP inhibitors willunmask adverse side-effects related to the total-ity of MTP function. Such effects could includethe development of pathologies in the kidney,testis, ovary and pancreas, organs in whichMTP mRNA is present at significant levels, andtherefore presumably contributing to normalfunction in, as yet, unknown ways.16,24,37 A priori, the nature and severity of side-effectsresulting from MTP inhibition may welldepend on the particular class of inhibitordeployed, as each is likely to affect a differentaspect of MTP functionality by virtue ofbinding to different sites on the MTP–PDIheterodimer. Evidence for this has alreadyemerged from two studies.38,39 In the first,Bakillah et al described a small (602 Da)organic molecule that blocked the secretion ofapoB-containing lipoproteins from HepG2cells, despite having no effect on MTP-mediated lipid transfer activity in vitro.38


Figure 1 Cholesterol and triglyceride transport in humans, viewed in the context of MTP expression in the smallintestine, liver and heart. For simplicity the transport of lipids other than cholesterol and triglyceride is not shown.CmR, chylomicron remnant particle; FFA, non-esterified fatty acids; MTP, microsomal triglyceride transfer protein.Adapted from Shoulders et al.7

ch02 14/7/05 4:45 pm Page 16

Instead, the inhibitor profoundly interferedwith the interaction of MTP with apoB (Figure2), and presumably subsequent loading ofapoB with lipid.22,38,40–42 In future studies, itmight prove instructive to establish whetherthe Bakillah inhibitor actually binds to apoB,MTP or the MTP–PDI complex. In the secondinhibitor study, Sellers et al compared theimpact of two MTP inhibitors, BMS-200150 andBMS-197636 (Figure 3), on the secretion ofapoB41-containing lipoproteins from aheterologous cell system expressing eitherhuman or Drosophila MTP.39 MTP inhibitorBMS-200150 decreased apoB41 lipoproteinsecretion from both cell systems, whereas thesecond (BMS-197636), and more potent, MTPinhibitor only blocked the secretion of apoB41lipoproteins from cells expressing humanMTP. Even at concentrations that were 100-fold higher than that used to inhibit human

CARDIOVASCULAR IMPLICATIONS OF PARTIAL, TISSUE-SPECIFIC SILENCING OF MTP 17

Figure 2 Model representing the Co-translationalloading of ApoB with lipid. The extreme amino-terminalof apoB interacts with the amino-terminal β-barrel ofMTP. This serves to align the predicted α-helical domainof apoB with the homologous domain in MTP. Thefigure indicates that the interaction of apoB with MTPperturbs MTP and PDI binding. Arrows indicate thetransfer of lipid from their sites of synthesis in themembrane of endoplasmic reticulum to MTP and apoB,and from MTP to apoB during the initial stages of theVLDL–apoB assembly process Figure 3 Chemical structures of representative MTP

inhibitors. Compounds inhibit: (1) the secretion ofapoB, but not apoA1, from HepG2 cells and (2) MTP-mediated transfer of radiolabelled triolein fromphospholipid donor liposomes to acceptor liposomes

ch02 14/7/05 4:45 pm Page 17

MTP, the BMS-197636 inhibitor had no impacton the secretion of apoB41-lipoproteins fromcells expressing Drosophila MTP.

The totality of MTP–PDI function may relyon different lipid transport roles, but this hasyet to be elucidated. The MTP–PDIheterodimer in vitro, has a capacity to transporta range of lipids such as triglyceride, choles-teryl ester, phosphatidylcholine andphosphatidylethanolamine,43 all of which areneeded for the production of VLDL andCm.1,44,45 The MTP–PDI heterodimer also hasthe ability to transfer diacylglycerol, lysophos-phatidic acid, phosphatidic acid, squalene andphosphatidylinositol between membranes invitro, but the physiological significance of this isuncertain. Read et al46 have considered howMTP might capture triglyceride monomersfrom lipid membranes, the rate-limiting step ofthe lipid transfer process.43,47,48 The datasuggest that MTP has a fusogenic helix at themouth of its predicted lipid-binding cavity,which interacts with a lipid membrane toperturb the arrangement of the acyl chains ofphospholipid molecules. This event is envis-aged to expose the acyl chains of neutral lipidand phospholipid monomers for interactionwith a second helix, which acts in concert withthe fusogenic helix, to ensure transfer of lipidmoieties to the hydrophobic lipid-bindingcavity of MTP, apoB or an acceptor membrane.Atzel and Wetterau have further suggested thatMTP contains two different classes of lipid-binding sites.49 The first is proposed to rapidlytransport neutral lipid and phospholipidbetween membranes, while the second appearsto selectively support the slow transport ofphospholipid.

The structures of the MTP–PDI complex andapoB are unknown, which leaves unresolved anumber of important questions regarding themechanism of action of inhibitor compounds,such as BMS-197636 and BMS-200150 (Figure3). From modelled structures and biochemicalanalyses it appears that MTP comprises threedomains: an amino terminal β-barrel (aminoacids 18–297), a central α-helical domain(amino acids 300–600) and a lipid-bindingcavity (amino acids 601–894) that may bind up

to five molecules of lipid.48,50–52 Moreover, itappears that the major binding site on MTP forPDI is formed by a series of residues residingon the outer surface of helices 13–17 of thelarge central α-helical domain of MTP (Figure2). To initiate the lipoprotein assemblyprocess, MTP interacts with apoB through twobinding sites (amino acids 22–303 and517–603), the second of which resides close tothe MTP–PDI interface.42,50–52 The proximity ofa binding site(s) on MTP for apoB and PDIsuggests that apoB might displace PDI fromMTP during the lipoprotein assembly process(Figure 2), only to be replaced again by PDIonce the newly assembled apoB lipoproteindissociates from MTP to exit the ER by a COPII(COat Protein)-dependent mechanism.53,54

Indirect evidence for this mechanism mayderive from a more in-depth analysis of thenon-synonymous, non-truncating APOB muta-tion (i.e. R463W) found in an extended ChristianLebanese family with the co-dominant disor-der, hypobetalipoproteinaemia.55 In vitroAPOB R463W interacted with MTP, but failedto support the subsequent production of apoB-containing lipoproteins. Moreover, expressionof APOB R463W in a stable cell line markedlyreduced the secretion of endogenous ratapoB100, leading the authors to speculate thatthe dominant-negative effect of APOB R463Wexpression on endogenous apoB secretionmight be attributable to enhanced binding ofthe mutant APOB R463W protein to MTP.

Although inhibitors of MTP action havealready produced promising reductions inblood lipid levels in phase I/II clinical trials,there is a real concern that blockage of intesti-nal fat absorption and hepatic lipid secretioncould compromise their long-term usage insome cases, as discussed in the excellent reviewof Harwood and colleagues.56 In the past yearsomewhat encouraging results have emergedfor the MTP inhibitor CP-346086, which maybe attributable to special properties of the drugitself (Figure 3), a novel dosing regime or acombination of both.57 Eight healthy menreceiving a nightly dose of the drug for 14 daysexperienced only occasional gastrointestinalsymptoms, such as mild diarrhoea and


ch02 14/7/05 4:45 pm Page 18

flatulence, and these were self-limiting.57

Moreover, magnetic resonance imaging of theliver at the end of the course provided noevidence of fatty infiltration. In these subjects,total cholesterol, LDL-cholesterol and apoBlevels dropped steadily for 13, 13 and 8 days,respectively. Compared to pre-dose levels,plateau values were reduced by 47%, 68% and52%. Importantly, high density lipoprotein(HDL) cholesterol levels remained within 10%of pre-dosing values at all time points. Serumtriglyceride levels also fell by up to 75%, in atime-dependent manner. The maximum dropwas observed 4 to 8 hours after drug adminis-tration, with serum triglyceride levels returningto baseline just prior to the next treatment.Although study participants had unchangedserum levels of vitamin A, vitamin E levels werelower by 43 ± 10% at the end of the study.

Therapeutic strategies that specificallyreduce MTP expression in the small intestineand liver would be expected to markedlyreduce the production of Cm and VLDL,without compromising the essential lipid trans-port function(s) of MTP in the heart.14–19,58 Assuch, a tissue-specific approach promises toprovide an efficacious therapeutic strategy toprevent the development of atheroscleroticcardiovascular disease attributable to increasedproduction of Cm remnant particles, and ofLDL, the main cholesterol-carrying particle inblood.59,60 In principle, the numerically impor-tant treatment groups would include patientswith prolonged and exaggerated post-prandialhyperlipidaemia,59–64 combined hyperlipi-daemia (i.e. raised serum levels of both choles-terol and triglyceride)65–70 and the metabolicsyndrome.7,71–74 By analogy with mice, a thera-peutic procedure that effectively inactivatedone MTP allele would be expected to reducetotal serum cholesterol, intermediate densitylipoproteins/LDL-cholesterol and apoB100levels.75 For example, in mice, a 50% reductionin intestinal and hepatic MTP mRNA levelslowers serum cholesterol levels by around24%.37 This may also be the case in theheterozygote parents and siblings of patientswith abetalipoproteinaemia,25 but this has beendifficult to establish definitively because of the

wide range of cholesterol levels in the generalpopulation. Notwithstanding, a 24% drop inserum cholesterol levels would be expected totranslate into a massive reduction in theincidence of cardiovascular disease, which is ontrack to become the world’s most commoncause of disease-related disability and death bythe year 2020.76,77

INSIGHTS INTO THE REGULATIONOF INTESTINAL MTP FROM MODELSYSTEMS

Indirect evidence suggests that the specificinhibition of intestinal MTP gene expressionmight provide a therapeutic means for loweringblood lipid levels. An early study showed thatthe hamster duodenum–jejunum contained2.5-fold higher levels of MTP mRNA levels thanthe ileum, and that in these tissues there was amonotonic relationship between MTP mRNAand protein levels.78 As importantly, the studyalso demonstrated that a high-fat diet compris-ing 20% w/w hydrogenated coconut oilincreased MTP mRNA levels by ~3-fold in theduodenum–jejunum preparations compared to~1.5-fold in the ileum. In these animals, intesti-nal MTP mRNA levels were strongly correlatedwith serum levels of total cholesterol, non-HDL-cholesterol and triglyceride.

Other workers have also reported strongcorrelations between intestinal MTP mRNAlevels and serum lipids in diabetic and insulin-resistant animal models,79–81 suggesting thatincreased intestinal MTP mRNA levels inhumans might causally contribute to the dyslip-idaemia of the metabolic syndrome. In onestreptozotocin-induced rat model of diabetes,for example, intestine preparations contained4-fold higher levels of MTP mRNA thansamples from non-induced animals.79 Therewas a strong correlation between MTP mRNAlevels and lymph Cm-triglyceride, and a moder-ate correlation with lymph Cm-cholesterol. Inan alloxan monohydrate-induced rabbit modelof diabetes, intestines contained 3- and 6-foldincreased levels of MTP mRNA and MTP-mediated lipid transfer activity, respectively.82


ch02 14/7/05 4:45 pm Page 19

Combining the diabetic and control groups,there was a positive correlation between intesti-nal MTP mRNA and both lymph apoB48- andB100-containing Cm. In the insulin-resistantZucker rat model of obesity, intestinal MTPmRNA levels were also increased by ~4-fold,and the release of Cm into the lacteal vessels by~2-fold.80 By contrast, the fructose-fed hamster

model of insulin resistance,81 characterized byoverproduction of intestinally derived apoB48-lipoproteins in both the fasting and post-prandial state, displayed a more modest 40%increase in MTP protein levels under basalconditions (Figure 4A–C). Thus, notwithstand-ing quantitative differences in MTP mRNA andprotein levels, in four different animal models


Secreted ApoB 48

Chow-fed

Fructose-fed

0 20 40 60 80 100

Chase time (min)

Labe

lled

Apo

B 4

8: %

of 0

min

100

90

80

70

60

50

40

30

20

10

0Chow-fed Fructose-fed

MTP protein mass

**

Per

cent

age

of c

ontr

ol

160

140

120

80

60

40

20

0

**

**

**

TC CE TG

Secreted lipid

TCCETG

– Cholesterol– Cholesterol ester– Triglyceride

Per

cent

age

of c

ontr

ol

800

750

700

650

300

250

200

150

100

50

0

** * *

BMS-197636 (µM)

0 0.5 1 1.5 2 2.5 3 3.5

120

100

80

60

40

20

0Apo

B 4

8 se

cret

ion

(% o

f con

trol

) MTP inhibition

Figure 4 Intestinal apoB48 and lipid secretion from thefructose-fed model of insulin resistance, and the impactof MTP inhibition. Asterisks indicate significantdifferences between chow- and fructose-fed animals.Data are adapted from Haidari et al81

(A) (B)

(C) (D)

ch02 14/7/05 4:45 pm Page 20

of perturbed insulin metabolism, intestinalMTP mRNA levels, lymph Cm concentrationsand serum lipid levels varied according to theglycaemic state of the animal. Thus, by extrap-olation, a capacity to develop high levels ofintestinal MTP protein may contribute to theexaggerated and prolonged post-prandialhyperlipidaemia observed in insulin-resistantand/or type 2 patients.59,60,83,84 Conversely, areduction in MTP protein levels in the post-prandial state would be expected to reduce therisk of atherosclerotic cardiovascular diseaseattributable to the generation of highly athero-genic Cm remnant particles.62,85–88

MTP inhibitor data from the hamster modelof insulin resistance suggest that in some situa-tions MTP protein may be present in excess ofthat needed to support the secretion of apoB-containing lipoproteins.81 In this animalmodel, for example, the MTP inhibitor, BMS-197636, had minimal impact on the secretionof Cm (Figure 4D), even at high concentra-tions. This finding may partly relate to higherintestinal MTP protein levels, increased lipidavailability and/or differences in the composi-tion of lipids within the membranes of the ER.In vitro, both the concentration of lipidmoieties within donor membranes and specificmembrane composition are major determi-nants influencing the rate of MTP-mediatedlipid transfer.49 For example, in one in vitrodata set, the specific activity of MTP increasedin proportion to the mass of triglyceride withindonor membranes, and varied inversely to theconcentration of negatively charged lipids (e.g.cardiolipin and phosphatidic acid).Significantly, Mensenkamp et al have alsoreported that the MTP inhibitor, BMS 197636,had minimal impact on the secretion ofhepatic triglyceride from Apoe–/– mice.89 Thesemice normally secrete around 50% less VLDL-triglyceride compared to control animals, withmuch of the additional triglyceride beingretained within the secretory pathway of thehepatocyte. Importantly, hepatic MTP proteinlevels in the Apoe–/– mice are comparable to thecontrol mice, suggesting that in conditions ofneutral lipid excess, MTP may have a higherspecific lipid transfer activity and/or be

prevented from binding to the MTP inhibitor,BMS 197636.

Crucially, no study has yet examinedwhether a reduction in intestinal MTP geneexpression actually leads to a reduction in thenumber of Cms entering the circulation. Lin etal reported that fresh garlic extracts lowerintestinal MTP mRNA,90 and suggested thatthis may account for the hypolipidaemic effectattributed to this additive.91 Additionally,Raabe et al have shown that the small intestineof Mttp–/+ mice contains ~50% lower levels ofMTP mRNA and protein than Mttp+/+ mice.37 Inother words, there was no evidence thattranscription from the remaining wild-typeMttp allele was upregulated to compensate forthe inactivated allele. However, because threegroups have shown that the MTP inhibitors CP-10447, BMS-200150 and an analogue of CP-10447 have remarkably little impact on the invitro secretion of apoB48-containing lipopro-teins (i.e. Cm in humans92–94), it will be impor-tant in future studies to titrate the impact ofMTP silencing on Cm assembly. The demon-stration that MTP inhibition virtually abolishespost-prandial hyperlipidaemia in rats (Novartiscompound 8aR), dogs (Novartis compound8aR and 19aR) and humans (Bay 13-9952)provides a major incentive to do so.95–97

INSIGHTS INTO REGULATION OFHEPATIC MTP FROM MODELSYSTEMS

The analysis of mouse models also providessome evidence that liver-specific silencing ofMTP might represent a useful therapeutic toolfor lowering blood lipid levels, and in turnreduce the risk of atherosclerotic cardiovascu-lar disease. In detail, in two different liver-specific (L)Mttp–/– models, mice had serumcholesterol levels that were ~50% lower thanwild-type animals.75,98 Serum triglyceride levelswere also lower in both strains of LMttp–/– mice,but to varying extents.75,98 Pertinent to a humanRNA-based therapy, liver microsomes fromLMttp–/+ mice contained on average 49% lowerlevels of MTP-mediated lipid transfer activity


ch02 14/7/05 4:45 pm Page 21

than microsomes from wild-type mice,98 againsuggesting that transcription from the wild-type allele was not upregulated to compensatefor the inactivated allele. These mice had 20%lower total serum cholesterol levels comparedto wild-type animals.

Importantly, when Young and colleaguesexamined whether MTP deficiency wouldeliminate the heightened susceptibility of‘apoB100-only’/LDL-receptor-deficient (i.eApob100/100:LDLr–/–) mice to atherosclerosis,the answer was affirmative.99 In outline, MTP-deficient Apob100/100:LDLr–/– mice haddramatically lower serum cholesterol andtriglyceride levels than Apob100/100:LDLr–/–

mice (i.e. 525.7 ± 32.24 mg/dl and 110.33 ±6.45 mg/dl versus 100.6 ± 14.33 mg/dl and23.4 ± 4.56 mg/dl). Additionally, by 20 weeksof age, atherosclerotic lesions covered 5.2 ±0.8% of the surface of the aorta of Apob100/100:LDLr–/– mice, whereas no such lesions werevisible in the LMttp–/+:Apob100/100:LDLr–/–

mice. In other words, the abolition of hepaticapoB100-containing lipoprotein assemblycompletely prevented the development ofatheroma in mice that were highly susceptibleto atherosclerosis. Similarly, for Apoe–/– mice,the MTP inhibitor BAY 13-9952 blocks theformation of atherosclerotic plaque in a dose-dependent manner.100,101

The study of genetic mouse models thatproduce markedly different amounts ofapoB100 in the setting of half-normal levels ofMTP also provides evidence that liver-specificsilencing of MTP might provide an effectivetherapy for treating genetic forms of athero-genic lipoprotein profiles attributable toincreased production of VLDL-apoB, such asfamilial combined hyperlipidaemia.7,65,102,103 Indetail, Young and colleagues compared serumapoB100 levels of wild-type and LMttp–/+ miceexpressing low, normal and high levels ofapoB100 mRNA.104 At each level of apoB100expression, half-normal levels of hepatic MTPwere associated with a reduction in serumapoB levels, average values falling by ~20–37%of control. Therefore, MTP deficiency had thelargest absolute impact on lowering serumapoB levels in mice that expressed the highest

level of apoB100. In the six different groups ofmice, hepatic triglyceride levels were lowest inthe transgenic mice expressing high levels ofapoB100 and wild-type levels of MTP, andhighest in LMttp–/+ mice expressing low levelsof apoB100 (i.e. apoB–/+). Pertinent for thehuman situation, the livers of LMttp–/+ micecontained 3-fold lower levels of triglyceriderelative to apoB100–/+ mice, presumablybecause they secreted more apoB100-contain-ing lipoproteins than apoB100–/+ mice. Thus, infuture studies it would be instructive to deter-mine the frequency and severity of non-alcoholic fatty liver disease in the parents ofpatients with abetalipoproteinaemia. In the co-dominant disorder familial hypobetalipopro-teinaemia, around 60% of patients haveevidence of fatty liver, with the percentages ofliver fat correlating positively with body massindex, waist circumference and impairedglucose tolerance.105 Because this ‘disorder’ isassociated with longevity and a very low risk ofpremature coronary disease, partial silencingof MTP could mimic this condition.

What is currently missing from mouse datasets is a description of the impact that insulinresistance, obesity and dietary determinants,such as alcohol and high-fat diets, would haveon both hepatic and blood lipid levels in thesetting of reduced hepatic MTP mass.Extrapolating from hamster models,106,107

insulin resistance would be predicted toincrease hepatic MTP mRNA levels, and thusthe secretion of VLDL-triglyceride. Carpentieret al, for example, have shown that ameliorat-ing insulin resistance in the fructose-fed modelof insulin resistance helps normalize hepaticMTP mRNA levels and reduce VLDL-trigly-ceride secretion.107 In all likelihood, thenegative insulin response element within thepromoter of the hamster MTP gene,108 which isconserved in humans, provides a mechanismfor the lowering of MTP mRNA levels in theseanimals.

In obesity, Bartel et al have shown that thelivers of young leptin-deficient (i.e. ob/ob)mice contained around 12-fold higher levels oftriglyceride than ob/+ mice.109 This wasmatched with a modest 45% increase in MTP


ch02 14/7/05 4:45 pm Page 22

mRNA levels and a 70% increase in VLDL-triglyceride secretion. However, in a secondstudy, ob/ob and wild-type mice had compara-ble levels of hepatic MTP mRNA and similarproduction rates of VLDL-triglyceride.110

Notwithstanding, serum cholesterol andtriglyceride levels of these mice were increasedby around 2-fold relative to control animals,and their livers combined, around 5- and 15-fold higher levels of hepatic triglyceride andcholesteryl ester. In a second animal model ofobesity, characterized by hypertriglyceri-daemia, hepatic MTP mRNA levels were raisedby around 20% relative to thin animals.111

Thus, a therapeutic strategy that reducedhepatic MTP protein levels in humans may wellprovide a powerful mechanism for reducingserum lipid levels in obese subjects, but thiscould incur a high cost in terms of promotingthe initiation and/or progression of non-alcoholic fatty liver disease.112

Alcohol abuse would almost certainly be acontraindication for utilizing an MTP silencingapproach to lower high blood triglyceridelevels.113,114 Lin et al have shown that thehuman MTP gene contains a negative ethanolresponsive element (i.e. –612 to –142 upstreamof transcription start site) and that rats gavagedwith ethanol for 3 hours had 32% and 18%lower levels of hepatic and intestinal MTPmRNA levels, respectively, than controlanimals.113 Moreover, these authors establishedthat a human hepatoma cell line exposed toethanol for 24 hours contained 50% lowerlevels of MTP mRNA levels than control cells.Likewise, Sugimoto et al showed that the liversof Sprague-Dawley rats fed an ethanol-contain-ing diet for 37 days contained 60% lower levelsof hepatic MTP mRNA than livers from controlanimals fed an isocaloric, alcohol-free diet, andrather predictably that this reduction in MTPmRNA levels was associated with dramaticallyincreased levels of hepatic cholesterol andtriglyceride.114

Our understanding of the potential impact ofa high-fat diet on human MTP gene expressionlargely derives from studies performed inhamsters.78,115,116 In one study, the feeding ofthese animals with increasing dietary fat

concentration from 11.7 energy % to 46.8energy % for one month increased hepaticMTP mRNA levels (r = 0.69, p = 0.0023), in adose-dependent manner by approximately60%.115 At the end of the study, MTP mRNAlevels were strongly correlated with total serumcholesterol (r = 0.74, p <0.001), VLDL-choles-terol (r = 0.60, p <0.01), LDL-cholesterol (r =0.55 p <0.05) and HDL-cholesterol (r = 0.70, p <0.001), but not with fasting serum triglyc-eride levels. In a second dataset, Bennett et alshowed that hepatic MTP gene expression mayalso depend on the type of fatty acid incorpo-rated into the dietary triglyceride; dietsenriched in either trimyristin or tripalmitindisplayed marked increases in MTP mRNAconcentrations relative to triolein, trilinolein ortristearin-rich diets.115 Puzzlingly, in a subse-quent study, modest amounts of dietary choles-terol (0.005–0.24 w/w%) abolished this effectof specific fatty acid on MTP mRNA level.116

Irrespective of triglyceride fatty acid type,cholesterol feeding increased MTP mRNA in adose-dependent manner, and at the end of 28days there were significant correlations betweenMTP mRNA, serum VLDL-triglyceride (r = 0.35,p = 0.011) and VLDL-cholesterol levels (r = 0.49,p = 0.0002). These data accord with in vitroinvestigations which have established that thehuman MTP promoter contains a sterol respon-sive element (i.e. –124 to –109 upstream oftranscription start site), and moreover that thiselement has the capacity to bind the sterolregulatory element-binding protein 2.117

Remarkably, an Affymetrix microarrayexperiment indicates that a deficiency of MTPin the liver has no major impact on the expres-sion of most genes involved in hepatic lipidmetabolism, despite the development of mildto moderate hepatic steatosis in LMttp–/–

mice.118 A notable exception was stearyl CoAdesaturase 1 mRNA, which was reduced, afinding that led Young and colleagues tofurther discoveries. First, the livers of LMttp–/–

mice contain lower levels of sterol regulatoryelement-binding protein 1c and the polyunsat-urated fatty acid, linoleic acid, relative tocontrol livers. Second, the animals had around45% lower levels of plasma insulin, which may,


ch02 14/7/05 4:45 pm Page 23

via a transcriptional mechanism, contribute tothe reduced levels of the sterol-regulatoryelement-binding protein 1c in these animals.119

Consistent with histological examination, themicroarray experiment also found no evidencefor an active inflammatory response in thelivers of LMttp–/– mice. Expression levels ofinflammation (e.g. macrophage inflammatoryproteins-1α, 1β, 2; a range of interleukins,interferon-α-β, γ and TNF-α) and apoptosis-related genes (e.g. bax, bcl-2, c-myc, c-jun andfas) were either comparable to control mice orbelow the level of detection.

INSIGHTS FROM HUMAN GENETICAND GENE EXPRESSION STUDIES

Arguably, the best evidence that long-termtissue-specific silencing of MTP expression mayprovide a powerful mechanism for reducingthe risk of cardiovascular disease derives fromgenetic comparisons of two types of people:those who have reached an exceptional longage and signally escaped age-associateddiseases, such as cardiovascular disease, andthose who have had a coronary heart diseaseevent at an early age. For the first group, Pucaand colleagues reported data showing that themuch-studied MTP–493T allele120–129 was under-represented in a group of long-livedAmericans, leading them to suggest that thisallele (or an allele in linkage disequilibrium)increases the risk of mortality.130 For thesecond, Karpe and colleagues58 reported thatthe MTP–493T allele increased the risk of acoronary heart disease event in a group ofmiddle-aged men (i.e. 45–64 years) with serumcholesterol levels of 7.0 ± 0.6 mmol/l, and that

this was likely to be attributable to reducedtranscriptional activity of this allele (or allele inlinkage disequilibrium) in cardiac myocytes.123

CONCLUDING COMMENT

The assembly and secretion of apoB-containinglipoproteins are crucial for survival in infancyand adulthood, but the intracellular processesthat regulate their production are sensitive tometabolic and genetic disease. In the future,the prevention and/or treatment of disordersattributable to increased production of Cmand VLDL should be feasible through anantisense or RNA silencing approach131–137 thatreduce hepatic and/or intestinal MTP mRNAlevels. Partial and tissue-specific silencing ofMTP promises to overcome a number of thepotential adverse side-effects associated withlong-term global inhibition of MTP.

ACKNOWLEDGEMENTS

I am grateful to the British Medical Researchand the British Heart Foundation, London,UK, who have supported most of the MTPresearch from my group. I also thank all co-investigators, especially Professors James Scottand Maryvonne Rosseneu, Leonard Banaszakand Drs Christopher J Mann, TimothyAnderson, Arjen R Mensenkamp andJacqueline Read, and Drs Andrew Dean andRossi Naoumova for critical reading of themanuscript, and Rocio Lale-Montes for excel-lent secretarial assistance. I also thankProfessor Khosrow Adeli for permission toreproduce the data in Figure 4.


1. Mason TM. The role of factors that regulate thesynthesis and secretion of very-low-density lipoproteinby hepatocytes. Crit Rev Clin Lab Sci 1998; 35:461–87

2. Olofsson SO, Stillemark-Billton P, Asp L. Intracellularassembly of VLDL: two major steps in separate cell

compartments. Trends Cardiovasc Med 2000;10:338–45

3. Gordon DA, Jamil H. Progress towards understandingthe role of microsomal triglyceride transfer protein inapolipoprotein-B lipoprotein assembly. BiochimBiophys Acta 2000; 1486:72–83

References

ch02 14/7/05 4:45 pm Page 24

4. Shelness GS, Sellers JA. Very-low-density lipoproteinassembly and secretion. Curr Opin Lipidol 2001;12:151–7

5. Fisher EA, Ginsberg HN. Complexity in the secretorypathway: the assembly and secretion of apolipopro-tein B-containing lipoproteins. J Biol Chem 2002;277:17377–80


7. Shoulders CC, Jones EL, Naoumova RP. Genetics offamilial combined hyperlipidemia and risk ofcoronary heart disease. Hum Mol Genet 2004; 13(Suppl 1):R149–60

8. Segrest JP, Jones MK, De Loof H, et al. Structure ofapolipoprotein B-100 in low density lipoproteins. JLipid Res 2001; 42:1346–67

9. Heath RB, Karpe F, Milne RW, et al. Selective partition-ing of dietary fatty acids into the VLDL-TG pool in theearly postprandial period. J Lipid Res 2003; 44:2065–72

10. Teusink B, Voshol PJ, Dahlmans VE, et al. Contri-bution of fatty acids released from lipolysis of plasmatriglycerides to total plasma fatty acid flux and tissue-specific fatty acid uptake. Diabetes 2003; 52:614–20

11. Wiggins D, Gibbons GF. Origin of hepatic very-low-density lipoprotein triacylglycerol: the contribution ofcellular phospholipid. Biochem J 1996; 320 (Pt2):673–9

12. Parks EJ, Hellerstein MK. Carbohydrate-inducedhypertriacylglycerolemia: historical perspective andreview of biological mechanisms. Am J Clin Nutr2000; 71:412–33

13. Haemmerle G, Zimmermann R, Zechner R. Lettinglipids go: hormone-sensitive lipase. Curr Opin Lipidol2003; 14:289–97

14. Boren J, Veniant MM, Young SG. Apo B100-contain-ing lipoproteins are secreted by the heart. J ClinInvest 1998; 101:1197–202

15. Nielsen LB, Veniant M, Boren J, et al. Genes forapolipoprotein B and microsomal triglyceride trans-fer protein are expressed in the heart: evidence thatthe heart has the capacity to synthesize and secretelipoproteins. Circulation 1998; 98:13–16

16. Bjorkegren J, Veniant M, Kim SK, et al. Lipoproteinsecretion and triglyceride stores in the heart. J BiolChem 2001; 276:38511–17

17. Nielsen LB, Bartels ED, Bollano E. Overexpression ofapolipoprotein B in the heart impedes cardiactriglyceride accumulation and development ofcardiac dysfunction in diabetic mice. J Biol Chem2002; 277:27014–20

18. Christoffersen C, Bollano E, Lindegaard ML, et al.Cardiac lipid accumulation associated with diastolicdysfunction in obese mice. Endocrinology 2003;144:3483–90

19. Yokoyama M, Yagyu H, Hu Y, et al. Apolipoprotein Bproduction reduces lipotoxic cardiomyopathy:studies in heart-specific lipoprotein lipase transgenicmouse. J Biol Chem 2004; 279:4204–11

20. Gordon DA, Jamil H, Sharp D, et al. Secretion ofapolipoprotein B-containing lipoproteins from HeLacells is dependent on expression of the microsomaltriglyceride transfer protein and is regulated by lipidavailability. Proc Natl Acad Sci USA 1994; 91:7628–32

21. Leiper JM, Bayliss JD, Pease RJ, et al. Microsomaltriglyceride transfer protein, the abetalipoproteine-mia gene product, mediates the secretion ofapolipoprotein B-containing lipoproteins fromheterologous cells. J Biol Chem 1994; 269:21951–4

22. Ingram MF, Shelness GS. Folding of the amino-termi-nal domain of apolipoprotein B initiates microsomaltriglyceride transfer protein-dependent lipid transferto nascent very low density lipoprotein. J Biol Chem1997; 272:10279–86

23. Sharp D, Blinderman L, Combs KA, et al. Cloningand gene defects in microsomal triglyceride transferprotein associated with abetalipoproteinaemia.Nature 1993; 365:65–9

24. Shoulders CC, Brett DJ, Bayliss JD, et al. Abetalipo-proteinemia is caused by defects of the gene encod-ing the 97 kDa subunit of a microsomal triglyceridetransfer protein. Hum Mol Genet 1993; 2:2109–16

25. Narcisi TM, Shoulders CC, Chester SA, et al.Mutations of the microsomal triglyceride-transfer-protein gene in abetalipoproteinemia. Am J HumGenet 1995; 57:1298–310

26. Ricci B, Sharp D, O’Rourke E, et al. A 30–amino acidtruncation of the microsomal triglyceride transferprotein large subunit disrupts its interaction withprotein disulfide-isomerase and causes abetalipopro-teinemia. J Biol Chem 1995; 270:14281–5

27. Rehberg EF, Samson-Bouma ME, Kienzle B, et al. Anovel abetalipoproteinemia genotype. Identificationof a missense mutation in the 97–kDa subunit of themicrosomal triglyceride transfer protein that preventscomplex formation with protein disulfide isomerase. JBiol Chem 1996; 271:29945–52

28. Berthier MT, Couture P, Houde A, et al. Thec.419–420insA in the MTP gene is associated withabetalipoproteinemia among French-Canadians. MolGenet Metab 2004; 81:140–3

29. Berriot-Varoqueaux N, Aggerbeck LP, Samson-Bouma M, et al. The role of the microsomaltriglygeride transfer protein in abetalipoproteinemia.Annu Rev Nutr 2000; 20:663–97

30. Rader DJ, Brewer HB, Jr. Abetalipoproteinemia. Newinsights into lipoprotein assembly and vitamin Emetabolism from a rare genetic disease. JAMA 1993;270:865–9

31. Wang X, Quinn PJ. Vitamin E and its function inmembranes. Prog Lipid Res 1999; 38:309–36

32. Blomhoff R, Green MH, Berg T, et al. Transport andstorage of vitamin A. Science 1990; 250:399–404

33. Schurgers LJ, Vermeer C. Differential lipoproteintransport pathways of K-vitamins in healthy subjects.Biochim Biophys Acta 2002; 1570:27–32

34. Illingworth DR, Connor WE, Miller RG.


ch02 14/7/05 4:45 pm Page 25

Abetalipoproteinemia. Report of two cases and reviewof therapy. Arch Neurol 1980; 37:659–62

35. Wang J, Hegele RA. Microsomal triglyceride transferprotein (MTP) gene mutations in Canadian subjectswith abetalipoproteinemia. Hum Mutat 2000;15:294–5

36. Chowers I, Banin E, Merin S, et al. Long-term assess-ment of combined vitamin A and E treatment for theprevention of retinal degeneration in abetalipopro-teinaemia and hypobetalipoproteinaemia patients.Eye 2001; 15(Pt 4):525–30

37. Raabe M, Kim E, Veniant M, et al. Using geneticallyengineered mice to understand apolipoprotein-Bdeficiency syndromes in humans. Proc Assoc AmPhysicians 1998; 110:521–30

38. Bakillah A, Nayak N, Saxena U, et al. Decreased secre-tion of ApoB follows inhibition of ApoB–MTP bindingby a novel antagonist. Biochemistry 2000; 39:4892–9

39. Sellers JA, Hou L, Athar H, et al. A Drosophila micro-somal triglyceride transfer protein homolog promotesthe assembly and secretion of human apolipoproteinB. Implications for human and insect transport andmetabolism. J Biol Chem 2003; 278:20367–73

40. Hussain MM, Bakillah A, Nayak N, et al. Amino acids430–570 in apolipoprotein B are critical for itsbinding to microsomal triglyceride transfer protein. JBiol Chem 1998; 273:25612–15

41. Wu DA, Bu X, Warden CH, et al. Quantitative traitlocus mapping of human blood pressure to a geneticregion at or near the lipoprotein lipase gene locus onchromosome 8p22. J Clin Invest 1996; 97:2111–18

42. Dashti N, Gandhi M, Liu X, et al. The N-terminal1000 residues of apolipoprotein B associate withmicrosomal triglyceride transfer protein to create alipid transfer pocket required for lipoprotein assem-bly. Biochemistry 2002; 41:6978–87

43. Jamil H, Dickson JK, Jr, Chu CH, et al. Microsomaltriglyceride transfer protein. Specificity of lipidbinding and transport. J Biol Chem 1995; 270:6549–54

44. Nishimaki-Mogami T, Yao Z, Fujimori K. Inhibition ofphosphatidylcholine synthesis via the phosphatidyl-ethanolamine methylation pathway impairs incorpo-ration of bulk lipids into VLDL in cultured rathepatocytes. J Lipid Res 2002; 43:1035–45

45. Noga AA, Zhao Y, Vance DE. An unexpected require-ment for phosphatidylethanolamine N-methyltrans-ferase in the secretion of very low densitylipoproteins. J Biol Chem 2002; 277:42358–65

46. Read J, Anderson TA, Ritchie PJ, et al. A mechanismof membrane neutral lipid acquisition by the micro-somal triglyceride transfer protein. J Biol Chem 2000;275:30372–7

47. Atzel A, Wetterau JR. Mechanism of microsomaltriglyceride transfer protein catalyzed lipid transport.Biochemistry 1993; 32:10444–50

48. Wetterau JR, Lin MC, Jamil H. Microsomal triglyc-eride transfer protein. Biochim Biophys Acta 1997;1345:136–50

49. Atzel A, Wetterau JR. Identification of two classes oflipid molecule binding sites on the microsomaltriglyceride transfer protein. Biochemistry 1994;33:15382–8

50. Mann CJ, Anderson TA, Read J, et al. The structure ofvitellogenin provides a molecular model for theassembly and secretion of atherogenic lipoproteins. JMol Biol 1999; 285:391–408

51. Bradbury P, Mann CJ, Kochl S, et al. A commonbinding site on the microsomal triglyceride transferprotein for apolipoprotein B and protein disulfideisomerase. J Biol Chem 1999; 274:3159–64

52. Segrest JP, Jones MK, Dashti N. N-terminal domain ofapolipoprotein B has structural homology tolipovitellin and microsomal triglyceride transferprotein: a ‘lipid pocket’ model for self-assembly ofapob-containing lipoprotein particles. J Lipid Res1999; 40:1401–16

53. Jones B, Jones EL, Bonney SA, et al. Mutations in aSar1 GTPase of COPII vesicles are associated withlipid absorption disorders. Nat Genet 2003; 34:29–31

54. Gusarova V, Brodsky JL, Fisher EA. ApolipoproteinB100 exit from the ER is COPII dependent and itslipidation to very low density lipoprotein occurs post-ER. J Biol Chem 2003; 278:48051–8

55. Burnett JR, Shan J, Miskie BA, et al. A novel nontrun-cating APOB gene mutation, R463W, causes familialhypobetalipoproteinemia. J Biol Chem 2003;278:13442–52

56. Chang G, Ruggeri RB, Harwood HJ, Jr. Microsomaltriglyceride transfer protein (MTP) inhibitors: discov-ery of clinically active inhibitors using high-through-put screening and parallel synthesis paradigms. CurrOpin Drug Discov Devel 2002; 5:562–70

57. Chandler CE, Wilder DE, Pettini JL, et al. CP-346086:an MTP inhibitor that lowers plasma cholesterol andtriglycerides in experimental animals and in humans.J Lipid Res 2003; 44:1887–901

58. Ledmyr H, McMahon AD, Ehrenborg E, et al. Themicrosomal triglyceride transfer protein gene -493Tvariant lowers cholesterol but increases the risk of coro-nary heart disease. Circulation. 2004; 109:2279–84

59. Karpe F, Hellenius ML, Hamsten A. Differences inpostprandial concentrations of very-low-densitylipoprotein and chylomicron remnants betweennormotriglyceridemic and hypertriglyceridemic menwith and without coronary heart disease. Metabolism1999; 48:301–7

60. Goldberg IJ, Kako Y, Lutz EP. Responses to eating:lipoproteins, lipolytic products and atherosclerosis.Curr Opin Lipidol 2000; 11:235–41

61. Demacker PN, Hectors MP, Stalenhoef AF.Chylomicron processing in familial dysbetalipopro-teinemia and familial combined hyperlipidemiastudied with vitamin A and E as markers: a new physi-ological concept. Atherosclerosis 2000; 149:169–80

62. Mero N, Malmstrom R, Steiner G, et al. Postprandialmetabolism of apolipoprotein B-48- and B-100-containing particles in type 2 diabetes mellitus:


ch02 14/7/05 4:45 pm Page 26

relations to angiographically verified severity ofcoronary artery disease. Atherosclerosis 2000; 150:167–77

63. Sharrett AR, Heiss G, Chambless LE, et al. Metabolicand lifestyle determinants of postprandial lipemiadiffer from those of fasting triglycerides: TheAtherosclerosis Risk In Communities (ARIC) study.Arterioscler Thromb Vasc Biol 2001; 21:275–81

64. Castro CM. Postprandial lipaemia in familialcombined hyperlipidaemia. Biochem Soc Trans 2003;31(Pt 5):1090–3

65. Venkatesan S, Cullen P, Pacy P, et al. Stable isotopesshow a direct relation between VLDL apoB overpro-duction and serum triglyceride levels and indicate ametabolically and biochemically coherent basis forfamilial combined hyperlipidemia. ArteriosclerThromb 1993; 13:1110–18

66. Austin MA, McKnight B, Edwards KL, et al.Cardiovascular disease mortality in familial forms ofhypertriglyceridemia: a 20-year prospective study.Circulation 2000; 101:2777–82

67. Voors-Pette C, de Bruin TW. Excess coronary heartdisease in Familial Combined Hyperlipidemia, inrelation to genetic factors and central obesity.Atherosclerosis 2001; 157:481–9

68. Hopkins PN, Heiss G, Ellison RC, et al. Coronaryartery disease risk in familial combined hyperlipi-demia and familial hypertriglyceridemia: a case-control comparison from the National Heart, Lung,and Blood Institute Family Heart Study. Circulation2003; 108:519–23

69. McNeely MJ, Edwards KL, Marcovina SM, et al.Lipoprotein and apolipoprotein abnormalities infamilial combined hyperlipidemia: a 20-year prospec-tive study. Atherosclerosis 2001; 159:471–81

70. Ayyobi AF, Brunzell JD. Lipoprotein distribution inthe metabolic syndrome, type 2 diabetes mellitus, andfamilial combined hyperlipidemia. Am J Cardiol2003; 92:27J–33J

71. World Health Organisation. Diabetes Mellitus:Report of a WHO Study Group. WHO, Geneva 1985

72. Expert Panel on Detection, Evaluation, andTreatment of High Blood Cholesterol in Adults.Executive Summary of The Third Report of TheNational Cholesterol Education Program (NCEP)Expert Panel on Detection, Evaluation, AndTreatment of High Blood Cholesterol In Adults(Adult Treatment Panel III). JAMA 2001; 285:2486–97

73. Sattar N, Gaw A, Scherbakova O, et al. Metabolicsyndrome with and without C-reactive protein as apredictor of coronary heart disease and diabetes inthe West of Scotland Coronary Prevention Study.Circulation 2003; 108:414–19

74. Ninomiya JK, L’Italien G, Criqui MH, et al.Association of the metabolic syndrome with history ofmyocardial infarction and stroke in the third nationalhealth and nutrition examination survey. Circulation2004; 109:42–6

75. Raabe M, Veniant MM, Sullivan MA, et al. Analysis of

the role of microsomal triglyceride transfer protein inthe liver of tissue-specific knockout mice. J Clin Invest1999; 103:1287–98

76. Murray CJ, Lopez AD. Alternative projections ofmortality and disability by cause 1990–2020: GlobalBurden of Disease Study. Lancet 1997; 349:1498–504

77. Fuster V. Epidemic of cardiovascular disease andstroke: the three main challenges. Presented at the 71stscientific sessions of the American Heart Association.Dallas, Texas. Circulation 1999; 99:1132–7

78. Lin MC, Arbeeny C, Bergquist K, et al. Cloning andregulation of hamster microsomal triglyceride trans-fer protein. The regulation is independent from thatof other hepatic and intestinal proteins which partici-pate in the transport of fatty acids and triglycerides. JBiol Chem 1994; 269:29138–45

79. Gleeson A, Anderton K, Owens D, et al. The role ofmicrosomal triglyceride transfer protein and dietarycholesterol in chylomicron production in diabetes.Diabetologia 1999; 42:944–8

80. Phillips C, Owens D, Collins P, et al. Microsomaltriglyceride transfer protein: does insulin resistanceplay a role in the regulation of chylomicron assembly?Atherosclerosis 2002; 160:355–60

81. Haidari M, Leung N, Mahbub F, et al. Fasting andpostprandial overproduction of intestinally derivedlipoproteins in an animal model of insulin resistance.Evidence that chronic fructose feeding in the hamsteris accompanied by enhanced intestinal de novolipogenesis and ApoB48-containing lipoproteinoverproduction. J Biol Chem 2002; 277:31646–55

82. Phillips C, Bennett A, Anderton K, et al. Intestinalrather than hepatic microsomal triglyceride transferprotein as a cause of postprandial dyslipidemia indiabetes. Metabolism 2002; 51:847–52

83. Boquist S, Hamsten A, Karpe F, et al. Insulin and non-esterified fatty acid relations to alimentary lipaemiaand plasma concentrations of postprandial triglyc-eride-rich lipoproteins in healthy middle-aged men.Diabetologia 2000; 43:185–93

84. Owens D. The extended postprandial phase indiabetes. Biochem Soc Trans 2003; 31(Pt 5):1085–9

85. Groot PH, van Stiphout WA, Krauss XH, et al.Postprandial lipoprotein metabolism in normolipi-demic men with and without coronary artery disease.Arterioscler Thromb 1991; 11:653–62

86. McNamara JR, Shah PK, Nakajima K, et al. Remnant-like particle (RLP) cholesterol is an independentcardiovascular disease risk factor in women: resultsfrom the Framingham Heart Study. Atherosclerosis2001; 154:229–36

87. Meyer E, Westerveld HT, Ruyter-Meijstek FC, et al.Abnormal postprandial apolipoprotein B-48 andtriglyceride responses in normolipidemic women withgreater than 70% stenotic coronary artery disease: acase-control study. Atherosclerosis 1996; 124:221–35

88. Boquist S, Ruotolo G, Tang R, et al. Alimentarylipemia, postprandial triglyceride-rich lipoproteins,and common carotid intima-media thickness in


ch02 14/7/05 4:45 pm Page 27

healthy, middle-aged men. Circulation 1999;100:723–8

89. Mensenkamp AR, Teusink B, Havinga R, et al.Involvement of apolipoprotein E in triacylglycerolincorporation into Very Low Density Lipoproteinparticles. J Hepatol 2004; 40:599–606

90. Lin MC, Wang EJ, Lee C, et al. Garlic inhibits micro-somal triglyceride transfer protein gene expression inhuman liver and intestinal cell lines and in rat intes-tine. J Nutr 2002; 132:1165–8

91. Ackermann RT, Mulrow CD, Ramirez G, et al. Garlicshows promise for improving some cardiovascular riskfactors. Arch Intern Med 2001; 161:813–24

92. Haghpassand M, Wilder D, Moberly JB. Inhibition ofapolipoprotein B and triglyceride secretion in humanhepatoma cells (HepG2). J Lipid Res 1996; 37:1468–80

93. van Greevenbroek MM, Robertus-Teunissen MG, etal. Participation of the microsomal triglyceride trans-fer protein in lipoprotein assembly in Caco-2 cells:interaction with saturated and unsaturated dietaryfatty acids. J Lipid Res 1998; 39:173–85

94. Nicodeme E, Benoist F, McLeod R, et al.Identification of domains in apolipoprotein B100 thatconfer a high requirement for the microsomal triglyc-eride transfer protein. J Biol Chem 1999; 274:1986–93

95. Ksander GM, deJesus R, Yuan A, et al.Diaminoindanes as microsomal triglyceride transferprotein inhibitors. J Med Chem 2001; 44:4677–87

96. Sorbera LA, Martin L, Silvestre J, et al. Implitapide.Drugs of the Future 2000; 25:1138–44

97. Stein EA, Arnes SA, Moore LJ, et al. Inhibition ofpost-prandial fat absorption with the MTP inhibitorBAY 13–9952. Circulation 2000; 102 (Supp II):2913

98. Chang BH, Liao W, Li L, et al. Liver-specific inactiva-tion of the abetalipoproteinemia gene completelyabrogates very low density lipoprotein/low densitylipoprotein production in a viable conditional knock-out mouse. J Biol Chem 1999; 274:6051–5

99. Lieu HD, Withycombe SK, Walker Q, et al.Eliminating atherogenesis in mice by switching offhepatic lipoprotein secretion. Circulation 2003;107:1315–21

100. Zaiss S, Gruetzmann R, Ullrich M. BAY 13–9952, aninhibitor of the microsomal triglyceride transferprotein (MTP) dose-dependently blocks the forma-tion of atherosclerotic plaques and renders themmore stable in apoE knockout mice. Circulation 1999;100 (Suppl I):1343

101. Zaiss S, Sander E. BAY 13–9952 (Implitapide), aninhibitor of the microsomal triglyceride transferprotein (MTP), inhibits atherosclerosis and prolongslifetime in apoE knockout mice. Eur Heart J 2000; 21(Suppl):16

102. Cortner JA, Coates PM, Bennett MJ, et al. Familialcombined hyperlipidaemia: use of stable isotopes todemonstrate overproduction of very low-densitylipoprotein apolipoprotein B by the liver. J InheritMetab Dis 1991; 14:915–22

103. Castro CM, de Bruin TW, de Valk HW, et al. Impairedfatty acid metabolism in familial combined hyperlipi-demia. A mechanism associating hepatic apolipopro-tein B overproduction and insulin resistance. J ClinInvest 1993; 92:160–8

104. Leung GK, Veniant MM, Kim SK, et al. A deficiency ofmicrosomal triglyceride transfer protein reducesapolipoprotein B secretion. J Biol Chem 2000;275:7515–20

105. Schonfeld G, Patterson BW, Yablonskiy DA, et al.Fatty liver in familial hypobetalipoproteinemia:triglyceride assembly into VLDL particles is affectedby the extent of hepatic steatosis. J Lipid Res 2003;44:470–8

106. Taghibiglou C, Carpentier A, Van Iderstine SC, et al.Mechanisms of hepatic very low density lipoproteinoverproduction in insulin resistance. Evidence forenhanced lipoprotein assembly, reduced intracellularApoB degradation, and increased microsomal triglyc-eride transfer protein in a fructose-fed hamstermodel. J Biol Chem 2000; 275:8416–25

107. Carpentier A, Taghibiglou C, Leung N, et al.Ameliorated hepatic insulin resistance is associatedwith normalization of microsomal triglyceride trans-fer protein expression and reduction in very lowdensity lipoprotein assembly and secretion in thefructose-fed hamster. J Biol Chem 2002; 277:28795–802

108. Au WS, Kung HF, Lin MC. Regulation of microsomaltriglyceride transfer protein gene by insulin in HepG2cells: roles of MAPKerk and MAPKp38. Diabetes2003; 52:1073–80

109. Bartels ED, Lauritsen M, Nielsen LB. Hepatic expres-sion of microsomal triglyceride transfer protein andin vivo secretion of triglyceride-rich lipoproteins areincreased in obese diabetic mice. Diabetes 2002;51:1233–9

110. Wiegman CH, Bandsma RH, Ouwens M, et al.Hepatic VLDL production in ob/ob mice is notstimulated by massive de novo lipogenesis but is lesssensitive to the suppressive effects of insulin. Diabetes2003; 52:1081–9

111. Kuriyama H, Yamashita S, Shimomura I, et al.Enhanced expression of hepatic acyl-coenzyme Asynthetase and microsomal triglyceride transferprotein messenger RNAs in the obese and hyper-triglyceridemic rat with visceral fat accumulation.Hepatology 1998; 27:557–62

112. Angulo P. Nonalcoholic fatty liver disease. N Engl JMed 2002; 346:1221–31

113. Lin MC, Li JJ, Wang EJ, et al. Ethanol down-regulatesthe transcription of microsomal triglyceride transferprotein gene. FASEB J 1997; 11:1145–52

114. Sugimoto T, Yamashita S, Ishigami M, et al.Decreased microsomal triglyceride transfer proteinactivity contributes to initiation of alcoholic liversteatosis in rats. J Hepatol 2002; 36:157–62

115. Bennett AJ, Billett MA, Salter AM, et al. Regulation ofhamster hepatic microsomal triglyceride transfer


ch02 14/7/05 4:45 pm Page 28

protein mRNA levels by dietary fats. Biochem BiophysRes Commun 1995; 212:473–8

116. Bennett AJ, Bruce JS, Salter AM, et al. Hepatic micro-somal triglyceride transfer protein messenger RNAconcentrations are increased by dietary cholesterol inhamsters. FEBS Lett 1996; 394:247–50

117. Sato R, Miyamoto W, Inoue J, et al. Sterol regulatoryelement-binding protein negatively regulates micro-somal triglyceride transfer protein gene transcription.J Biol Chem 1999; 274:24714–20

118. Bjorkegren J, Beigneux A, Bergo MO, et al. Blockingthe secretion of hepatic very low density lipoproteinsrenders the liver more susceptible to toxin-inducedinjury. J Biol Chem 2002; 277:5476–83

119. Shimomura I, Shimano H, Korn BS, et al. Nuclearsterol regulatory element-binding proteins activategenes responsible for the entire program of unsatu-rated fatty acid biosynthesis in transgenic mouse liver.J Biol Chem 1998; 273:35299–306

120. Couture P, Otvos JD, Cupples LA, et al. Absence ofassociation between genetic variation in the promoterof the microsomal triglyceride transfer protein geneand plasma lipoproteins in the FraminghamOffspring Study. Atherosclerosis 2000; 148:337–43

121. Talmud PJ, Palmen J, Miller G, et al. Effect of micro-somal triglyceride transfer protein gene variants (-493G > T, Q95H and H297Q) on plasma lipid levelsin healthy middle-aged UK men. Ann Hum Genet2000; 64(Pt 4):269–76

122. Watts GF, Riches FM, Humphries SE, et al. Genotypicassociations of the hepatic secretion of VLDLapolipoprotein B-100 in obesity. J Lipid Res 2000;41:481–8

123. Ledmyr H, Karpe F, Lundahl B, et al. Variants of themicrosomal triglyceride transfer protein gene areassociated with plasma cholesterol levels and bodymass index. J Lipid Res 2002; 43:51–8

124. Juo SH, Han Z, Smith JD, et al. Common polymor-phism in promoter of microsomal triglyceride trans-fer protein gene influences cholesterol, ApoB, andtriglyceride levels in young african american men:results from the coronary artery risk development inyoung adults (CARDIA) study. Arterioscler ThrombVasc Biol 2000; 20:1316–22

125. Vincent S, Planells R, Defoort C, et al. Geneticpolymorphisms and lipoprotein responses to diets.Proc Nutr Soc 2002; 61:427–34

126. Bjorn L, Leren TP, Ose L, et al. A functional polymor-

phism in the promoter region of the microsomaltriglyceride transfer protein (MTP -493G/T) influ-ences lipoprotein phenotype in familial hypercholes-terolemia. Arterioscler Thromb Vasc Biol 2000;20:1784–8

127. Yanagisawa Y, Kawabata T, Tanaka O, et al.Improvement in blood lipid levels by dietary sn-1,3–diacylglycerol in young women with variants oflipid transporters 54T-FABP2 and -493g-MTP.Biochem Biophys Res Commun 2003; 302:743–50

128. Bernard S, Touzet S, Personne I, et al. Associationbetween microsomal triglyceride transfer proteingene polymorphism and the biological features ofliver steatosis in patients with type II diabetes.Diabetologia 2000; 43:995–9

129. St Pierre J, Lemieux I, Miller-Felix I, et al. Visceralobesity and hyperinsulinemia modulate the impact ofthe microsomal triglyceride transfer protein -493G/Tpolymorphism on plasma lipoprotein levels in men.Atherosclerosis 2002; 160:317–24

130. Geesaman BJ, Benson E, Brewster SJ, et al. Haplotype-based identification of a microsomal transfer proteinmarker associated with the human lifespan. Proc NatlAcad Sci U S A 2003; 100:14115–20

131. Bi F, Liu N, Fan D. Small interfering RNA: a new toolfor gene therapy. Curr Gene Ther 2003; 3:411–17

132. Biroccio A, Leonetti C, Zupi G. The future ofantisense therapy: combination with anticancer treat-ments. Oncogene 2003; 22:6579–88

133. Lavery KS, King TH. Antisense and RNAi: powerfultools in drug target discovery and validation. CurrOpin Drug Discov Devel 2003; 6:561–9

134. Gonzalez FM, Crooke RM, Tillman L, et al. Stability ofpolycationic complexes of an antisense oligonu-cleotide in rat small intestine homogenates. Eur JPharm Biopharm 2003; 55:19–26

135. Xing HR, Cordon-Cardo C, Deng X, et al.Pharmacologic inactivation of kinase suppressor ofras-1 abrogates Ras-mediated pancreatic cancer. NatMed 2003; 9:1267–8

136. Wang L, Prakash RK, Stein CA, et al. Progress in thedelivery of therapeutic oligonucleotides: organ/cellu-lar distribution and targeted delivery of oligonu-cleotides in vivo. Antisense Nucleic Acid Drug Dev2003; 13:169–89

137. Wood MJ, Trulzsch B, Abdelgany A, et al. Therapeuticgene silencing in the nervous system. Hum Mol Genet2003; 12 Spec No 2:R279–84


ch02 14/7/05 4:45 pm Page 29

ch02 14/7/05 4:45 pm Page 30

INTRODUCTION

Coronary heart disease (CHD) is a major causeof death in Westernized societies and isprojected to be the leading cause of death wellinto the millennium.1 Over the last 30 years therole of plasma cholesterol and hypercholes-terolaemia in the development of CHD hasbeen well established. Recently the role of asecond plasma lipid, triglyceride (TG), in thedevelopment of CHD has come to promi-nence. Three CHD risk factors, hypertriglyceri-daemia, obesity and insulin resistance, areassociated with elevated tissue and plasmalevels of TG and they are among the metabolicabnormalities whose co-existence definessyndrome X or the metabolic syndrome.2

Several recent studies have demonstrated anassociation of elevated plasma TG with CHD.This association has led the Adult TreatmentPanel III to designate hypertriglyceridaemia asan independent risk factor for CHD.3

Hypertriglyceridaemia can act to increase thedevelopment of CHD in at least two ways. First,increased levels of TG-rich particles in plasmafrequently lead to increased production ofremnant lipoproteins.4 These remnant lipopro-teins are prone to oxidation, which leads totheir uptake by macrophages. Macrophagesthat have exceeded their capacity to processlipids transform into foam cells that initiate aseries of steps, leading to the deposition oflipid in the arterial wall.5 The second way thathypertriglyceridaemia can act to increase thedevelopment of CHD is through the exchange

of TG in TG-rich apoB-containing particles forcholesteryl ester in high density lipoprotein(HDL).6 This leads to a lowering of HDL levelsand reduced reverse cholesterol transportfrom peripheral tissues to liver, where choles-terol is excreted into bile.7

In addition to the role of plasma TG in thedevelopment of CHD, TG accumulation intissues, including adipose tissue, can havepathological consequences. According to theWorld Health Organization (WHO), morethan 1 billion people are recognized as beingclinically obese due, to a large extent, to theincreased consumption of energy-dense foodsand a sedentary lifestyle. This is anticipated tohave adverse consequences regarding thedevelopment of CHD since obese individualsare more likely than lean individuals todevelop CHD. The prevalence of CHD inmales almost doubles and in women triples forindividuals with a BMI >40 kg/m2 compared toa normal BMI (<25 kg/m2).8

Adipose tissue plays a direct role in mediat-ing plasma lipid concentrations through theuptake of plasma TG and the release ofadipose-derived fatty acids into plasma.Adipose tissue also plays an endocrine role andsecretes a number of adipocytokines whichaffect food intake, energy expenditure andlipid metabolism.9,10

Several studies have demonstrated the associ-ation between non-adipose tissue TG levels andinsulin resistance resulting from impairedinsulin signalling.11–14 While it is not known ifcellular TG directly impairs insulin signalling14

Acyl CoA: Diacylglycerol acyltransferases(DGATs) as therapeutic targets forcardiovascular diseaseJ.S. Millar and J.T. Billheimer

3

ch03 14/7/05 4:45 pm Page 31

or if it is the fatty acid derived from TG,15 it isknown that improvements in insulin signallingresulting from the use of insulin-sensitizingagents reduce tissue TG levels.16 Thus, it seemsreasonable to conclude that a means ofpreventing tissue TG accumulation would havea beneficial effect on insulin signalling.

Pharmacological interventions designed totreat hypertriglyceridaemia, obesity or type 2diabetes often have a specific effect on one ofthe above morbidities that may sometimes haveundesirable consequences on another riskfactor. For example, peroxisome proliferatoractivated receptor gamma agonists canimprove insulin sensitivity but frequentlyincrease plasma lipid levels and cause weightgain while sibutramine, a compound thatpromotes satiety, may also elevate bloodpressure, itself a risk factor.17,18 However,recently described enzymes, acyl CoA:diacyl-glycerol acyltransferase (DGAT1 and DGAT2),that catalyse the last step in triglyceride (TG)synthesis in tissues, have become attractivetargets for pharmacotherapy since they havethe potential to have a more generalizedbeneficial effect on hypertriglyceridaemia,obesity and insulin resistance.19–21 Theseenzymes have been associated with a resistanceto diet-induced weight gain,19 improved insulinsensitivity20 and lowering of plasma lipid levels,presumably resulting from reduced availabilityof TG for lipoprotein synthesis.21 While thereare adverse effects that need to be dealt withduring the design phase, the potential to havea beneficial impact simultaneously on hyper-lipidaemia, diet-induced obesity and type 2diabetes by inhibiting a single enzymatic stepmakes DGAT an attractive target for drugdiscovery.

STRUCTURE AND PROPERTIES OFDGAT1 AND DGAT2

Phospholipids, triglycerides and glycerolipidsshare common biosynthetic pathways. Theterminal and only committed step of the TGsynthetic pathway involves the fatty acid acyla-tion of diacylglycerol catalysed by DGAT.

DGAT activity is found intracellularly in the ERmembrane where lipid synthesis occurs.22 Thefirst isolation of a protein with DGAT activitycame from the work of Andersson et al.23 Thisprotein has never been fully characterized, butsubsequently two mammalian enzymes withDGAT activity have been cloned (DGAT1 andDGAT2).24,25 DGATs are membrane bound,presumably within the ER, and their activityhas been detected on both the cytosolic andlumenal sides of the ER.26 The presence ofDGAT activity on both sides of the ERmembrane suggests that these enzymes may bedifferentially involved in the synthesis of TG forstorage in cytoplasm and for use in lipoproteinsynthesis within the ER lumen similar toseparate roles that ACAT 1 and 2 play in thesynthesis of cholesteryl esters targeted forstorage or secretion.27

DGAT1

DGAT1 is a member of the membrane-boundO-acyltransferase (MBOAT) family thatincludes acyl CoA: cholesterol acyltransferases1 and 2 responsible for the intracellular synthe-sis of cholesterol esters.28 Members of thisenzyme family share a series of membrane-spanning regions (6–12 in the case of DGAT1)and catalyse the transfer of organic acids, suchas fatty acids, onto membrane-bound hydroxy-lated targeted molecules, such as cholesteroland diacylglycerol.28 The enzyme is active as ahomotetramer,29 although the significance ofthis is unknown. The orientation of the activesite is unknown, although the role of DGAT1in mammary and white adipose tissue24 andapparent lack of a role in lipoprotein secre-tion30 suggests the role of the enzyme is synthe-sizing cytoplasmic fat stores.

DGAT1 has been shown to be expressed inall tissues examined, with expression in theliver and small intestine being the greatest.25

Since these tissues synthesize lipoproteins, itwas anticipated that DGAT1 was involved insynthesizing TG for lipoprotein assembly. TheDgat1–/– (knockout) mouse, however, hadnormal plasma levels of TG, demonstratingthat there was no strict requirement of Dgat1


ch03 14/7/05 4:45 pm Page 32

expression for lipoprotein production,although there was an effect of DGAT1deficiency on the rate of intestinal fat absorp-tion. Thus a second TG-synthesizing enzyme,presumably DGAT2, is required for TGtargeted for secretion or can compensate forlack of DGAT1. We found that hepatic overex-pression of DGAT1 in wild-type mice, whichwould be expected to increase hepatic TGsynthesis, resulted in TG accumulation in liverbut did not have an effect on VLDL TG orapoB production.30 This indicates that thepresence of TG is necessary but not sufficientfor VLDL secretion.

In addition to decreased hepatic TGcontent, the Dgat1 knockout mouse displays adecreased TG content within many othertissues, including white adipose tissue.19

Interestingly, this mouse model of Dgat1deficiency has increased sensitivity to bothleptin and insulin and is resistant to diet-induced obesity due to increased energyexpenditure.20 Post-partum-Dgat1 knockoutmice were defective in their lactation andunable to support pups. The lactation defecthas recently been determined to be probablydue to developmental defects resulting fromdisrupted cellular lipid signalling pathways.31

Dgat1 knockout mice of both sexes alsodisplayed skin and hair follicle abnormalities,probably related to the lack of triglyceride as acomponent in skin oil.32 DGAT1 overexpres-sion in an adipocyte cell line resulted inincreased cytoplasmic fat stores, consistent witha role in regulating cytoplasmic triglyceridestores.33 DGAT1 expression is stimulated byglucose,34 which would act to increase fatdeposition in adipose tissue in response tofeeding, although the specific glucose-sensitiveresponse element has not been identified. Itwould be of interest to determine whetherthere is an inverse regulation of DGAT1expression in hepatic and non-hepatic tissuessince hepatic triglyceride levels are greatestduring prolonged fasting and lowest duringfed conditions. A polymorphism in the DGAT1gene (K232A) has been reported in the bovinehomologue of DGAT1.35 This polymorphismaffects enzyme activity by reducing the Vmax and

consequently results in a lower milk yield fromcarriers of this polymorphism. The lysine atthis position is conserved in the human enzymeand thus may influence DGAT1 activity inhumans. Taken together, these results indicatea role of DGAT1 in synthesis of cytoplasmic TGfor storage or export in milk with little or norole in regulating the production rate oflipoproteins in liver.

DGAT2

DGAT2, structurally unrelated to DGAT1,25 is asecond enzyme that catalyses the acylation ofdiacylglycerol using a fatty acyl CoA substrate.DGAT2 is also membrane-bound within the ERand has two to four putative membrane-spanning regions.25 DGAT2 is a member of the DGAT2/monoacylglycerol: acyltransferase(MOGAT) family that includes monoacylglyc-erol and wax ester acyltransferase members.36

Like members of the MBOAT family thatincludes DGAT1, the DGAT2-related enzymescatalyse the acylation of hydroxylated acceptormolecules using a fatty acyl CoA substrate.25

Similar to DGAT1, DGAT2 has been shown tobe expressed in all tissues examined, with theexpression in liver, white adipose tissue,mammary tissue and peripheral leukocytesbeing the greatest.25 Relative to DGAT1,DGAT2 is more sensitive to magnesiumconcentration, which allows one to estimateindividual in vitro DGAT activity in tissues.25

The high expression in liver indicated a poten-tial role for DGAT2 in regulating lipoproteinproduction. DGAT2 overexpression in mouseliver increases hepatic TG content to a greaterextent than DGAT1 and appears to have asmall effect on VLDL TG but not apoB produc-tion (Millar, unpublished). The decreasedplasma lipid levels seen in Dgat2-deficientmouse pups21 indicate that plasma lipoproteinproduction can be lowered through DGAT2inhibition by reducing hepatic TG levels to theextent that lipid becomes rate-limiting forVLDL production. However, increasinghepatic TG stores above normal levels hasminimal impact on VLDL production. Underthese conditions, there are sufficient lipid

ACYL CoA: DIACYLGLYCEROL ACYLTRANSFERASES AS THERAPEUTIC TARGETS FOR CVD 33

ch03 14/7/05 4:45 pm Page 33

levels for lipoprotein production, indicatingfactors other than triglyceride (such as micro-somal triglyceride transfer protein) are rate-limiting for VLDL production.37 The Dgat2knockout mouse, although not viable for morethan 24 hours following birth, shows amarkedly decreased TG content of tissue,including liver and adipose tissue.21 The rapidonset of death in these mice following birthappears to be, in part, a result of skin barrierabnormalities that cause rapid dehydrationand energy depletion due to decreased fatstores. However, rearing in a humidifiedenvironment combined with saline injectionsto prevent dehydration only prolongs the life-span by a few hours, suggesting that otherfactors contribute to their premature demise.

INVESTIGATIONS INVOLVING GENEKNOCKOUTS, POLYMORPHISMS ANDKNOWN INHIBITORS OF DGAT1 ANDDGAT2

The inhibition of one or both DGAT enzymes,and thereby the major pathway of TG synthesis,is a possible pharmacological strategy for treat-ment of hyperlipidaemia, obesity and diabetes.Much of the evidence supporting DGAT1 as atherapeutic target was learned from the disrup-tion of the Dgat1 gene.19 Unexpectedly, on achow diet, the DGAT1 knockout mouse hadnormal plasma TG levels and near-normaltissue TG levels, indicating other pathways ofTG synthesis. The normal plasma TG led inves-tigators to search for and find a second TGsynthesizing enzyme, DGAT2,25 to be integralfor supplying TG for lipoprotein production(see below). However, the Dgat1–/– mice, whenfed a high-fat diet, were resistant to dietary-induced obesity.19 A 40% decrease in total bodyTG accounted for the weight loss with noreduction in total body protein mass. Thiseffect did not appear to be due to fat malab-sorption but due, in part, to an increase inmetabolic rate resulting from increased activ-ity. It should be noted that subsequent studiesrevealed a delayed fat absorption, such that adecrease in post-prandial triglyceride response

was observed in the Dgat1–/– mice.38 Elevatedlevels of chylomicron remnants in plasma areknown to be highly atherogenic and DGAT1inhibition may therefore be expected todecrease the peak plasma concentration ofchylomicrons and their remnants.4

In addition to observations regarding thelack of changes in lipoprotein levels, the Dgat1knockout mouse was reported to have a low TGcontent of the adipose tissue due to a decreasein adipocyte size, not reduced number.19

Increased adipocyte size has been associatedwith insulin resistance and, indeed, the Dgat1knockout mice demonstrate an increase ininsulin sensitivity.20 Recent studies suggest thatthe endocrine function of white adipose tissueplays a role in the increased insulin sensitivityas well as the increased energy expenditure.39

The transplantation of Dgat1-deficient whiteadipose tissue into wild-type mice enhancesglucose disposal.40 A two-fold increase inadiponectin, an adipokine that increases fattyacid oxidation and insulin sensitivity, wasobserved in these mice.40 In contrast, overex-pression of DGAT1 in rat pancreatic cellsresulted in an increase in cellular TG contentand reduced insulin secretion in response toglucose.41 The Dgat1–/– mice also show anincrease in leptin sensitivity.20 Leptin, anadipokine synthesized by adipose tissue, inter-acts with satiety centres in the brain and isinvolved in the maintenance of whole-body TGstores42 and, hence, body weight. Much of theoverall effect on whole-body TG metabolismobserved in Dgat1–/– mice appears to relate tochanges in adipose metabolism.

While DGAT1 deficiency has not beendescribed in humans, polymorphisms of theDGAT1 gene and promoter have been identi-fied.43–45 As previously mentioned, one suchpolymorphism (K232A) has been shown toreduce the Vmax of the enzyme and has a signif-icant effect on milk fat yield in dairy cattle,43

with carriers having a relatively low content ofmilk fat. Comparison of the bovine sequence tothe human sequence shows the wild-typeresidue in the human sequence as lysine. Thismakes this an attractive polymorphism to testfor in human studies. Ludwig et al identified a


ch03 14/7/05 4:45 pm Page 34

DGAT1 promoter polymorphism (C→T atbase –79) in the Turkish population whichshowed reduced promoter activity in culturedcells.44 In Turkish women (but not men) thispolymorphism was associated with lower bodymass index (BMI). However, when obeseFrench subjects were analysed in a similarstudy, there was no association with bodyweight.45

While reduced DGAT1 activity has thepotential benefit of reducing body weight, atotal absence of DGAT1 activity has someundesirable consequences. Female Dgat1knockout mice do not produce milk and subse-quent studies have demonstrated that DGAT1activity is necessary for normal development ofmammary tissue.24,31 Also, starting at puberty,the hair of the Dgat1 knockout animals has adry appearance and subsequently falls out,beginning on the dorsal surface of the neck.24

Similarly, the skin of older knockout mice hasatrophied sebaceous glands, oil-producingglands within hair follicles, and abnormal lipidcomposition of fur oils.

The fact that the Dgat1 knockout micemaintained normal plasma TG levels led to thediscovery of a second enzyme, DGAT2, capableof synthesizing TG from diacylglycerol.Recently, a Dgat2 knockout mouse wasdescribed.21 The knockout of the DGAT2 geneled to more severe consequences than didknocking out Dgat1. The Dgat2–/– mice arelipopenic, have severe skin abnormalities anddie within 24 hours of birth. The plasma TGand FA content in the newborn Dgat2 knock-out mice was decreased by 70–90%. Totalcarcass and hepatic TG was also decreased by90%, indicating that DGAT1 is unable tocompensate for DGAT2 deficiency. TG is anormal component of skin and is one of thelipids which act as permeability barrier.DGAT2 activity is required for proper barrierfunction because there was a rapid loss ofweight in the newborn mice that suggesteddehydration due to increased transdermalwater loss. It is worth noting that DGAT1 isundetectable in dermis and epidermis ofneonatal Dgat2 knockout mice. The Dgat2knockout newborns are much smaller than

their litter mates and knockout embryos hadan 86% decrease in tissue TG, suggesting thatone cause of death may be related to improperfetal development. Studies with liver-specificknockout or knockdown of Dgat2 would clarifythe picture as to the role of hepatic DGAT2 inthe adult animal. The role of DGAT2 in skin isalso evident in humans; a recent study foundthat DGAT2 mRNA was decreased by two-thirds in psoriatic skin compared to normalskin.46 Partial liver-specific inhibition ofDGAT2 may lead to a beneficial profile regard-ing lipoprotein metabolism, while preventingadverse effects of DGAT2 inhibition in othertissues. While single nucleotide polymorphismsof the DGAT2 gene have been reported, theeffects of these polymorphisms on DGAT2activity have not been investigated.

Several naturally occurring compounds,including chalcone,47 roselipins,48 prenylflavo-noids,49 polyacetylenes,50 amidepsines51 andtashinones,52 and synthetic compounds, ben-zoxiperones53 and N-(7,10-dimethyl-11-oxo-10,11-dihydro-dibenzo[b,f][1,4]oxazepin-2-yl-4-hydroxy-benzamide,54 have been shown to benon-specific inhibitors of DGAT activity. Noneare very active, having IC50s in the micromolarrange. One of these, xanthumol, has beenassociated with a reduction in TG and apoBsecretion in HepG2 cells,55 although microso-mal triglyceride transfer protein levels, whichwere also reduced, may be responsible for thischange. Most compounds were identified inmicrosomal systems containing both DGAT1and DGAT2 isozymes and specificity towardsindividual isotypes has not been tested.Because the two isozymes are from differentgene families, some specificity is expected. A96-well plate assay has been developed whichmay aid in the identification of the potent andspecific inhibitors needed for proof of princi-ple studies.54

CURRENT THERAPY

There are three compound classes (niacin,fibrates and statins) for the treatment of hyper-triglyceridaemia. Niacin (nicotinic acid) has


ch03 14/7/05 4:45 pm Page 35

been shown to lower plasma TG levels 20 to50% in several clinical trials (see review 56).Niacin is not well tolerated but the newerextended-release formulation, Niaspan®, mayresult in fewer side-effects and better compli-ance. The exact mechanism of action of niacinis not well understood but is thought toinvolve, in part, a decrease in TG synthesis. Arecent article demonstrates that niacin is aninhibitor (albeit a weak one) of DGAT2, whichcould explain the decrease in TG synthesis.57

The fibrate class of therapeutic agents,which act to reduce plasma TG and increaseHDL-C with minimal effect on LDL-C, havebeen shown to decrease disease progression.58

Fibrates are ligands of peroxisome proliferatoractivated receptor alpha (PPAR-α) whoseactivation induces enzymes in fatty acid oxida-tion. Oxidation of fatty acids would decreasetheir availability for TG synthesis and secre-tion.59

The recently available, more potent statinshave been shown to reduce plasma TG levels.The extent of plasma TG lowering is depen-dent on the baseline TG level and the % LDLcholesterol lowering achieved.60 Statin therapyde-represses/induces the LDL-R which recog-nizes both apoB- and apoE-containing lipopro-

teins. This could result in increased removal ofTG-containing VLDL remnants throughbinding of apoE to LDL-R. The three classes ofTG-lowering compounds have individual side-effects and may be more beneficial for specificpatient populations.

Presently there are two types of anti-obesitydrugs, those that affect fat absorption (pancre-atic lipase inhibitors) and centrally actingappetite suppressants such as sibutramine.Neither is especially effective and the uses ofboth are limited by side-effects.61

THE THERAPEUTIC POTENTIAL OFDGAT INHIBITION

The potential of therapeutic agents that caninhibit or reduce the expression of DGAT1 andDGAT2 is promising. These compounds havethe potential to simultaneously reduce diet-induced obesity, increase insulin sensitivity andnormalize plasma lipid levels (summarized inTable 1). An additional potential benefit thatmight be observed is a reduction of hepaticlipotoxicity observed in steatosis. Treatment ofpatients with hepatic steatosis may reduce theirrisk of developing hepatitis resulting from


Table 1 The cytotoxic effects of triglyceride (TG) excess in various tissues and the expected benefits or adverseeffects of decreasing the tissue TG content through DGAT1 and DGAT2 inhibition

Tissue Role of TG Effect of excess triglyceride Expected effect of DGAT inhibition

Liver Lipoprotein synthesis Lipotoxicity Reduced VLDL productionEnergy storage Insulin resistance

Inflammation

Adipose Energy storage Obesity Decreased obesity andSecondary insulin resistance secondary sequelae

Intestine Lipoprotein synthesis Increased chylomicron Reduced chylomicron productionproduction Decreased post-prandial

triglyceridaemia

Pancreas Energy source Lipotoxicity leading to Normalized insulin secretionSource of lipid for signalling decreased insulin secretion

Lipoapoptosis

Muscle Energy source Insulin resistance Improved insulin sensitivity

Mammary Milk component Fat-enriched milk? Decreased milk fat content

Skin Water barrier component Cytotoxicity? Psoriasis and other skin barrier defects

ch03 14/7/05 4:45 pm Page 36

cytotoxic effects, although this remains to betested.

Despite these likely desirable effects ofDGAT1 and DGAT2 inhibition, the utility ofsuch inhibition may be limited by adverseeffects of reducing activity and/or expressionof these enzymes. Obviously, decreasing thesynthesis of total TG to the extent that thesupply of energy to tissues for basal metabolismis compromised, similar to that observed in theDgat2 knockout mice, would be deleterious.The extent of DGAT inhibition required for abeneficial effect versus that which shows toxic-ity (therapeutic index) is not known. Little isknown about the potential harmful effectswhich may occur due to increased levels of theTG precursors, diacylglycerol (DAG) and fattyacid due to DGAT inhibition. DAG is an activa-tor of protein kinase C (PKC) and PKC activa-tion has been associated with the developmentof diabetic complications.62 The accumulationof intracellular saturated fatty acids in severalcell lines has been shown to be lipotoxic and,in fact, under these conditions increasing TGsynthesis would reduce the level of saturatedfatty acid and improve tissue function.63

Whether DAG or fatty acids would accumulate

or be shunted into other pathways such asphospholipid synthesis and fatty acid oxidationis unknown. Interestingly, in the Dgat1 knock-out mouse, the concentration of DAG in whiteadipose tissue and skeletal muscle isunchanged and the level is actually decreasedin liver.20 While much has been learned aboutDGAT1 and DGAT2 since their initial descrip-tions less than 5 years ago, some basicquestions remain to be answered. Do DGAT1and DGAT2 have separate and/or overlappingfunctions? Are there tissue-specific functions?Is the triglyceride produced by DGAT1metabolically distinct from that produced byDGAT2? Can the two DGATs compensate foreach other? Are potential adverse effectsspecific to the individual DGAT or seen withinhibition of either enzyme? The answer to thisfinal question would determine whether ageneral DGAT inhibitor is sufficient or if thereis a need for specific inhibitors. Answeringthese questions will probably require tissue-specific knockout or knockdown of DGAT1and DGAT2 and/or the identification of apotent specific DGAT inhibitor. While DGATinhibition looks promising, additional researchis needed.


1. Murray D, Lopez A. Alternative projections of mortal-ity and disability by cause 1990–2020: global burdenof disease study. Lancet 1997; 349:1498–504

2. Grundy S. Hypertriglyceridemia, insulin resistanceand the metabolic syndrome. Am J Cardiol 1999;83:25F–9F

3. Expert Panel on Detection, Evaluation, andTreatment of High Blood Cholesterol in Adults,Executive Summary of the Third Report of theNational Cholesterol Education Program (NCEP)Expert Panel on Detection, Evaluation, andTreatment of High Blood Cholesterol in Adults(Adult Treatment Panel III). JAMA 2001;285:2486–97

4. Zilversmit DB. Atherogenic nature of triglycerides,postprandial lipidemia, and triglyceride-rich remnantlipoproteins. Clin Chem 1995; 41:153–8

5. Gianturco SH, Bradley WA. Pathophysiology oftriglyceride-rich lipoproteins in atherothrombosis:cellular aspects. Clin Cardiol 1999; 22:117–14

6. Tall A. Plasma lipid transfer proteins. Ann RevBiochem 1995; 64:235–57

7. Rader DJ. Regulation of reverse cholesterol transportand clinical Implications. Am J Cardiol 2003; 92:42J–9J

8. Must A, Spadno J, Coakley, et al. The disease burdenassociated with overweight and obesity. JAMA 1999;282:1523–9

9. Shirai K. Obesity as the core of the metabolicsyndrome and the management of coronary heartdisease. Curr Med Res Opin 2004; 20:295–304

10. Spiegelman B, Choy L, Hotamisligil G, et al.Regulation of adipocyte gene expression in differen-tiation and syndromes of obesity/diabetes. J BiolChem 1993; 268:6823–6

11. Perseghin G, Scifo P, De Cobelli F, et al.Intramyocellular triglyceride content is a determinantof in vivo insulin resistance in humans: a 1H–13Cnuclear magnetic resonance spectroscopy assessmentin offspring of type 2 diabetic parents. Diabetes 1999;48:1600–6

References

ch03 14/7/05 4:45 pm Page 37

12. Finck BN, Han X, Courtois M, et al. A critical role forPPARalpha-mediated lipotoxicity in the pathogenesisof diabetic cardiomyopathy: modulation by dietary fatcontent. Proc Natl Acad Sci USA 2003; 100:1226–31

13. Lupi R, Del Guerra S, Fierabracci V, et al. Lipotoxicityin human pancreatic islets and the protective effect ofmetformin. Diabetes 2002; 51:S134–7

14. Wanless IR, Shiota K. The pathogenesis of nonalco-holic steatohepatitis and other fatty liver diseases: afour-step model including the role of lipid releaseand hepatic venular obstruction in the progression tocirrhosis. Semin Liver Dis 2004; 24:99–106

15. Hulver MW, Lynis Dohm G. The molecular mecha-nism linking muscle fat accumulation to insulin resis-tance. Proc Nutr Soc 2004; 63:375–80

16. Mayerson AB, Hundal RS, Dufour S, et al. The effectsof rosiglitazone on insulin sensitivity, lipolysis, andhepatic and skeletal muscle triglyceride content inpatients with type 2 diabetes. Diabetes 2002;51:797–802

17. van Wijk JP, de Koning EJ, Martens EP, et al.Thiazolidinediones and blood lipids in type 2diabetes. Arterioscler Thromb Vasc Biol 2003;23:1744–9

18. Poston, W, Foreyt, J. Sibutramine and the manage-ment of obesity. Exp Opin Pharmacother 2004;5:633–42

19. Smith SJ, Cases S, Jensen DR, et al. Obesity resistanceand multiple mechanisms of triglyceride synthesis inmice lacking Dgat. Nat Genet 2000; 25:87–90

20. Chen HC, Smith SJ, Ladha Z, et al. Increased insulinand leptin sensitivity in mice lacking acyl CoA: diacyl-glycerol acyltransferase 1. J Clin Invest 2002;109:1049–55

21. Stone SJ, Myers HM, Watkins SM, et al. Lipopenia andskin barrier abnormalities in DGAT2–deficient mice.J Biol Chem 2003; 279:11767–76

22. Hamilton RL, Moorehouse A, Lear SR, et al. A rapidcalcium precipitation method of recovering largeamounts of highly pure hepatocyte rough endoplas-mic reticulum. J Lipid Res 1999; 40:1140–7

23. Andersson M, Wettesten M, Boren J, et al.Purification of diacylglycerol: acyltransferase from ratliver to near homogeneity. J Lipid Res 1994;35:535–45

24. Cases S, Smith SJ, Zheng YW, et al. Identification of agene encoding an acyl CoA:diacylglycerol acyltrans-ferase, a key enzyme in triacylglycerol synthesis. ProcNatl Acad Sci USA 1998; 95:13018–23

25. Cases S, Stone SJ, Zhou P, et al. Cloning of DGAT2, asecond mammalian diacylglycerol acyltransferase,and related family members. J Biol Chem 2001;276:38870–6

26. Owen MR, Corstorphine CC, Zammit VA. Overt andlatent activities of diacylglycerol acytransferase in ratliver microsomes: possible roles in very-low-densitylipoprotein triacylglycerol secretion. Biochem J 1997;323:17–21

27. Lee R, Willingham M, Davis M, et al. Differential

expression of ACAT1 and ACAT2 among cells withinliver, intestine, kidney and adrenal of nonhumanprimates. J Lipid Res 2000; 41:1991–2001

28. Hofmann K. A superfamily of membrane-bound O-acyltransferases with implications for wnt signaling.Trends Biochem Sci 2000; 25:111–2

29. Cheng D, Meegalla RL, He B, et al. Human acyl-CoA:diacylglycerol acyltransferase is a tetramericprotein. Biochem J 2001; 359:707–14

30. Millar JS, Tow B, Young SG, et al. Hepatic overex-pression of murine acyl CoA :diacylglycerol acyltrans-ferase (DGAT) has no effect on VLDL triglyceride orapoB production rate in vivo. Arterioscler ThrombVasc Biol 2001; 21:78 (abstract)

31. Cases S, Zhou P, Shillingford JM, et al. Developmentof the mammary gland requires DGAT1 expression instromal and epithelial tissues. Development 2004;113:3047–55

32. Chen HC, Smith SJ, Tow B, et al. Leptin modulatesthe effects of acyl CoA:diacylglycerol acyltransferasedeficiency on murine fur and sebaceous glands. J ClinInvest 2002; 109:175–81

33. Yu YH, Zhang Y, Oelkers P, et al. Posttranscriptionalcontrol of the expression and function of diacylglyc-erol acyltransferase-1 in mouse adipocytes. J BiolChem 2002; 277:50876–84

34. Meegalla RL, Billheimer JT, Cheng D. Concertedelevation of acyl-coenzyme A:diacylglycerol acyltrans-ferase (DGAT) activity through independent stimula-tion of mRNA expression of DGAT1 and DGAT2 bycarbohydrate and insulin. Biochem Biophys ResCommun 2002; 298:317–23

35. Winter A, Kramer W, Werner FA, et al. Association ofa lysine-232/alanine polymorphism in a bovine geneencoding acyl-CoA:diacylglycerol acyltransferase(DGAT1) with variation at a quantitative trait locusfor milk fat content. Proc Natl Acad Sci USA 2002;99:9300–5

36. Winter A, van Eckeveld M, Bininda-Emonds OR, et al.Genomic organization of the DGAT2/MOGAT genefamily in cattle (Bos taurus) and other mammals.Cytogenet Genome Res 2003; 102:42–7


38. Buhman KK, Smith SJ, Stone SJ, et al. DGAT1 is notessential for intestinal triacylglycerol absorption orchylomicron synthesis. J Biol Chem 2002; 277:25474–9

39. Guerre-Millo M. Adipose tissue and adipokines: forbetter or worse. Diabetes Metab 2004; 30:13–9

40. Chen HC, Jensen DR, Myers HM, et al. Obesity resis-tance and enhanced glucose metabolism in micetransplanted with white adipose tissue lacking acylCoA:diacylglycerol acyltransferase 1. J Clin Invest2003; 111:1715–22

41. Kelpe CL, Johnson LM, Poitout V. Increasing triglyc-eride synthesis inhibits glucose-induced insulin secre-tion in isolated rat islets of langerhans: a study using


ch03 14/7/05 4:45 pm Page 38

adenoviral expression of diacylglycerol acyltrans-ferase. Endocrinology 2002; 143:3326–32

42. Ahima RS, Osei SY. Leptin signaling. Physiol Behav2004; 81:223–41

43. Grisart B, Coppieters W, Farnir F, et al. Positionalcandidate cloning of a QTL in dairy cattle: identifica-tion of a missense mutation in the bovine DGAT1gene with major effect on milk yield and composition.Genome Res 2002; 12:222–31

44. Ludwig EH, Mahley RW, Palaoglu E, et al. DGAT1promoter polymorphism associated with alterationsin body mass index, high density lipoprotein levelsand blood pressure in Turkish women. Clin Genet2002; 62:68–73

45. Coudreau SK, Tounian P, Bonhomme G, et al. Roleof the DGAT gene C79T single-nucleotide polymor-phism in French obese subjects. Obes Res 2003;11:1163–7

46. Wakimoto K, Chiba H, Michibata H, et al. A noveldiacylglycerol acyltransferase (DGAT2) is decreased inhuman psoriatic skin and increased in diabetic mice.Biochem Biophys Res Commun 2003; 310:296–302

47. Tabata N, Ito M, Tomoda H, et al. Xanthohumols,diacylglycerol acyltransferase inhibitors, formHumulus lupulus. Phytochem 1997; 46:683–7

48. Tomoda H, Ohyama Y, Abe T, et al. Roselipins,inhibitors of diacylglycerol scyltransferase, producedby Gliocladium roseum KF-1040. J Antibiotics 1999;52:689–94

49. Chung M, Rho M, Ko K, et al. In vitro inhibition ofdiacylglycerol acyl-transferase by prenylflavonoidsfrom Sophora flavescens. Planta Med 2004; 70:258–60

50. Lee SW, Kim K, Rho M, et al. New polyacetylenes,DGAT inhibitors from the roots Panax ginseng. PlantaMed 2004; 70:197–200

51. Tomoda H, Ito M, Tabata N, et al. Amidepsines,inhibitors of diacylglycerol acyltransferase producedby Numicola sp. FO-2942. J Antibiotics 1995; 48:937–41

52. Ko J, Ryu S, Kim Y, et al. Inhibitory activity of diacyl-glycerol acyltransferase by tashinones from the root ofSalvia miltiorrhiza. Arch Pharm Res 2002; 25: 446–8

53. Burrows J, Block M, Burckett L, et al. Novel benzox-azepinone enzyme inhibitors of diacylglycerol acyltransferase. Nat Med Chem Symp 26th Virginia C-221998

54. Ramharack R, Spahr M. Diacylglycerol acyltransferase(DGAT) assay. US Patents 6607893. August 19, 2003.

55. Casaschi A, Maiyoh G, Rubio B, et al. The chalconexanthohumol inhibits triglyceride and apolipopro-tein B secretion in Hep G2 cells. J Nutr 2004; 134:1340–6

56. McKenney J. New perspectives on the use of niacin inthe treatment of lipid disorders. Arch Intern Med2004; 164:697–703

57. Ganji S, Tavintharan S, Zhu D, et al. Niacin non-competitively inhibits diacylglycerol acyltransferase-2(DGAT2) but not DGAT1 activity in HepG2 cells. JLipid Res 2004; 45:1835–45

58. Ginsburg H. Hypertriglyceridemia: new insights andnew approaches to pharmacologic therapy. Am JCardiol 2001; 87:1174–9

59. Francis G, Fayrd E, Picard F, et al. Nuclear receptorsand the control of metabolism. Annu Rev Physiol2003; 65:261–311

60. Stein E, Lane M, Laskarzewski P. Comparison ofstatins in hypertriglyceridemia. Am J Cardiol 1998;81:66B–9B

61. Kopelman P, Grace C. New thoughts on managingobesity. Gut 2004; 53:1044–53

62. Koya D, King G. Protein kinase C activation and thedevelopment of diabetic complications. Diabetes1998; 47:859–66

63. Listenberger LL, Han X, Lewis SE, et al. Triglycerideaccumulation protects against fatty acid-inducedlipotoxicity. Proc Natl Acad Sci USA 2003;100:3077–82


ch03 14/7/05 4:45 pm Page 39

ch03 14/7/05 4:45 pm Page 40

INTRODUCTION TO ACAT

Two isoforms of the enzyme known as acylcoenzyme A:cholesterol acyltransferase(ACAT, E.C. 2.3.1.26) have been identified.The official name of these enzymes is sterol o-acyltransferase (SOAT), but the enzymes aremuch more widely known as ACAT and thisname will be used throughout this chapter.Both isoforms, ACAT1 and ACAT2, areintegral membrane proteins localized to theendoplasmic reticulum and catalyse thereaction in which cholesterol and long-chainfatty acyl CoA molecules are converted into acholesteryl ester molecule.1 This conversion tocholesteryl ester limits the solubility of choles-terol in a phospholipid bilayer from a 1:1molar ratio to a 1:50 cholesteryl ester tophospholipid molar ratio.2,3 This decrease insolubility facilitates the storage of cholesterylesters in lipid droplets and prevents cytotoxic-ity due to build up of cholesterol inmembranes, making the ACAT reaction essen-tial in maintaining intracellular cholesterolbalance.4 A key event in the pathology of ather-osclerosis is the accumulation of cholesterylesters in the arterial intima, often first appear-ing in macrophage-derived foam cells.Accordingly, the inhibition of ACAT, particu-larly in arterial macrophages, has long beenconsidered a pharmaceutical target. However,the following discussion will provide evidencethat the most effective ACAT inhibition toprevent atherosclerosis is probably not in the

macrophage, a site of ACAT1. Rather, the genedeletion of ACAT2, the isoform primarilyexpressed in hepatocytes and enterocytes, hasbeen found to be more effective in limitingatherogenesis in mouse models, a findingrecommending ACAT2 as a preferred target.

IDENTIFICATION OF THEACYLTRANSFERASE GENE FAMILY

While the ACAT reaction was first described inliver in 1957,5 the inability to purify an activeform of ACAT from membranes has been amajor hurdle for studies of the enzyme for over35 years.6 This obstacle was partially alleviatedin 1993 when TY Chang and colleagues usedsomatic cell genetics to clone ACAT1 by trans-fecting human macrophage genomic DNA intoACAT-deficient CHO (AC29) cells.7 Cells werethen screened for cholesterol esterificationactivity, and the source DNA for the enzymaticactivity was identified. The Chang group wasable to isolate a 4 kb gene that, when trans-fected into AC29 cells, increased cholesterolesterification activity by 20-fold.8 This seminalwork in the Chang laboratory provided theessential tool, the ACAT1 cDNA sequence, forthe many subsequent molecular studies thathave contributed to our understanding of thestructure and function of ACAT enzymes.

After the identification of ACAT1, RobertFarese Jr and colleagues9,10 generated andcharacterized an ACAT1 knockout mouse, and

The biochemical and physiological rolesof ACAT1 and ACAT2 in cholesterolhomeostasis and atherosclerosisR.G. Lee and L.L. Rudel

4

ch04 14/7/05 4:45 pm Page 41

unexpectedly found that while cholesterolesterification was ablated in the adrenal of theACAT–/– mice, it was essentially unchanged inthe liver and small intestine, suggesting that atleast one additional ACAT enzyme was presentin these tissues.9,10 The cloning of two ACATisoforms from the yeast genome at about thesame time by Yang et al11further supported thepossibility that more than one ACAT isoformexisted. To follow up on the experiments donein yeast, these workers identified two expressedsequence tags that exhibited sequence similar-ity to regions within the human ACAT1 gene.The first sequence was subsequently identifiedas a part of the acyl coenzyme A:diacylglycerolacyltransferase (DGAT) gene, an enzyme thatcatalyses the reaction in which an acyl CoA isesterified to a diacylglcyerol molecule, result-ing in a triacylglycerol product.12 The secondsequence was later identified as a part of the ACAT2 gene in monkeys, mice andhumans.13–15 The ACAT2 gene productincreased cholesterol esterification activity byalmost 200-fold when transfected into AC29cells.13 A key difference between the ACAT1and ACAT2 isoforms was discovered whenNorthern blots from 17 non-human primatetissues were probed with isoform-specific radio-probes. ACAT1 mRNA was present in all tissuesexamined, but ACAT2 mRNA was presentprimarily in the liver and small intestine.13

Despite the sequence differences in the threeenzymes, the shared ability to transfer acylchains from coenzyme A to an acceptorsubstrate permitted the ACAT1, ACAT2 andDGAT1 enzymes to be identified collectively asmembers of the acyltransferase gene family.16

CHROMOSOMAL LOCATION ANDGENE STRUCTURE OF ACAT1 ANDACAT2

Chang and colleagues mapped the humanACAT1 gene and concluded that the gene ismade up of 17 exons found on two differentchromosomes.17 Exon 1, which makes up 90%of the 5' untranslated region (UTR), is foundon chromosome 7, while the rest of the gene is

found on chromosome 1, band q25. Theyconcluded that the mRNAs produced from thetwo chromosomes then combine to form thefull-length mRNA by a novel, as yet undefinedtrans-splicing event. This full-length mRNAconsists of a 1396 bp 5' UTR, a 1.65 kb codingregion and a 963 bp 3' UTR that codes for aprotein consisting of 550 amino acids. UponNorthern analysis, a heterogeneous bandingpattern for ACAT1 mRNA is typically found inalmost all tissues.13 The four-band pattern issimilar in most tissues, with bands at approxi-mately 2.0, 2.6, 3.0 and 3.6 kb. The significanceof the presence of the first exon on a differentchromosome from the rest of the gene remainsunknown.

The sequence of the human ACAT2 gene hasbeen determined by two different groups.18,19

Both groups suggest a more traditional genemap for ACAT2 made up of 15 exons spanningover 18–21 kb of genomic DNA on chromo-some 12. The full-length 2040 bp hACAT2mRNA is made of a 51 bp 5' UTR, a 1.569 kbcoding region and a 420 bp 3' UTR thatencodes a protein of 522 amino acids. Theposition of the ACAT2 gene on chromosome 12places it only about 1300 nucleotide residuesdownstream of the insulin-like growth factorbinding protein-6 structural gene, so that the 5'non-coding region is relatively truncated.19 Inthis region, the presence of Cdx-2 bindingregions has been identified20 which couldconfer the intestine-specific expression that hasbeen noted for ACAT2. Also present in thepromoter region HNF-1α and C/EBP-βtranscriptional elements and at least the formeris likely to be responsible for expression of thisenzyme in hepatocytes (Pramfalk et al,manuscript in revision). In a limited epidemio-logical study of the hACAT2 gene, 91 dyslipi-daemic patients were screened for mutations inthe ACAT2 gene.18 Two mutations were found(E14G, T254I) in the coding region. The onlyphenotypic change observed in the patientspossessing the T254I mutation was an increasein apoC-III levels, which has been shown to bean inhibitor of lipoprotein lipase.21 Eventhough these elevated apoC-III levels did notresult in significant changes in plasma


ch04 14/7/05 4:45 pm Page 42

triglyceride levels, further investigation usinglarger cohorts is required to determine whethermutations in the ACAT2 gene alter plasma lipidand lipoprotein parameters.

The chromosomal location of ACAT1,ACAT2 and DGAT1 in the genome of threenon-human primate species has recently beendescribed by in situ hybridization.22 In theAfrican green monkey, the cynomolgusmonkey and the squirrel monkey, each enzymewas found to be in chromosomal regionshomologous to those in humans.

STRUCTURAL STUDIES OF ACAT1AND ACAT2

The comparison of the primary amino acidsequences of primate ACAT1 and ACAT2proteins has yielded several important insightsinto the enzyme’s structure and function.Comparison of the full-length proteins revealedthat the N-terminal 100 amino acids had nosequence similarity, while the remaining C-terminal amino acids had 63% sequencesimilarity. It has been hypothesized that theunique N-terminal amino acids could be impor-tant in the distinct properties of the twoisoforms.15 In an effort to elucidate some ofthese unique properties, the N-terminal 34amino acids were deleted from the ARE-2enzyme, the yeast analogue of ACAT.23

Significantly decreased activity levels resultedfor the truncated protein compared to the wild-type enzyme and this finding led the investiga-tors to speculate that this portion of the proteinmay play a regulatory role, although furtherinvestigation is necessary. In an effort to learnmore about the location of the active sites ofACAT1 and ACAT2, Sturley and colleagues23

compared the amino acid sequences ofacyltransferases from several different speciesthat esterify oleoyl-CoA to either sterol (ACAT-like reaction) or diacylglycerol (DGAT-likereaction). All of the acyltransferases had onehighly conserved motif consisting of a FY ×DWWN heptapeptide that was hypothesized tofunction in the binding of the oleoyl-CoAmolecule. All of the sterol acyltransferases had

another highly conserved motif consisting of anH(Y)SF tripeptide that could function in thebinding of the sterol acceptor molecule.Mutation of either of the two conserved motifsin ACAT1 or ACAT2 led to ablation of activity,with no change in protein expression, support-ing the hypothesized role of these sites inenzyme activity.

When ACAT1 and ACAT2 were originallycloned, their amino acid sequences wereanalysed by computer software to determinethe probable membrane topology of the twoenzymes based on the grouping of hydrophobicamino acids.8,13 Based on computer predictionsthat both enzymes had seven to eight trans-membrane domains, two different groups haveexperimentally determined the topology of thetwo enzymes. Using truncation mutantscontaining reporter glycosylation sequencesinserted after predicted transmembranedomains, Joyce et al24 found that both enzymeshad five transmembrane domains, with the N-terminus of both enzymes in the cytosol of thecell and the C-terminus in the lumen. TheDWWN sites were found on the cytoplasmicside for both enzyme models and the H(Y)SFmotif was found on the cytoplasmic side of theACAT1 topology model and on the luminal sideof the ACAT2 topology model. The functionalsignificance of this putative active site being onopposite sides of the ER membrane for the twoisoforms has yet to be addressed experimen-tally. Lin et al25,26 used epitopes inserted afterthe putative transmembrane domains toconstruct an entirely different topology modelwith ACAT1 possessing seven transmembranedomains and ACAT2 possessing only two trans-membrane domains. In the Lin topologymodels, the DWWN and H(Y)SF regions wereon the cytoplasmic side of the ER membranefor both enzymes. The experimental methodsused by both groups have been used effectivelyby others,27–30 making it difficult to determinewhether either group’s topology model iscorrect. Any discrepancy in the number and/orlocations of utilized transmembrane domainsbetween ACAT1 vs ACAT2 is curious given suchhigh sequence homology (60%) in thepredicted transmembrane regions of the two

BIOCHEMICAL AND PHYSIOLOGICAL ROLES OF ACAT1 AND ACAT2 43

ch04 14/7/05 4:45 pm Page 43

proteins. However, in at least one of thestudies24 the topology for either enzyme wasanalysed simultaneously with several methodsand differences in the utilization of some of thetransmembrane domains by ACAT1 andACAT2 were still predicted.

In the early 1990s, two groups developeddata to suggest that the minimum molecularmass of ACAT in rat liver microsomes wasbetween 170 and 224 kDa.31,32 These datacontrast to the apparent molecular weightsbased on a primary sequence of about66 000 kDa for ACAT1 and 63 000 kDa forACAT2, and the apparent molecular weights asseen upon SDS gel electrophoresis of about50 000 kDa for ACAT1 and 47 000 kDa forACAT2.33 The enzymes are not apparentlyglycosylated,13 so the reason for the smallerestimate from SDS electrophoresis is unknown.Despite the fact that the experiments did notdiscern between ACAT1 and ACAT2, it didpoint to the possibility that one or both of theenzymes existed in an oligomeric state. Toinvestigate the oligomerization of ACAT1,Chang and colleagues performed sedimenta-tion and cross-linking studies on detergent-solubilized enzyme and concluded that ACAT1existed as a tetramer.34 Although no studieshave examined the oligomeric state of ACAT2,immunoprecipitation studies showed thatACAT1 and ACAT2 did not co-immunoprecip-itate, suggesting that the two enzymes do notassociate in a hetero-oligomeric state.35

Experiments in which mutation of prolines inthe N-terminal amino acids of the hACAT1protein led to changes in the oligmerizationstate brought out the possibility that theisoform-specific oligomerization is due to theinvolvement of the unique N-terminal aminoacids in the self-association of the two isoforms,and that the self-association may be involved inregulation of enzyme activity.36

STUDIES SUGGESTING FUNCTIONALDIFFERENCES OF ACAT1 AND ACAT2

The localization of ACAT1 and ACAT2 mRNAin non-human primate tissues suggested that

the two isoforms perform unique physiologicalfunctions. To further our knowledge of thelocation and associated function of theisoforms, antibodies specific for the N-terminalregion of ACAT1 and ACAT2 were used toimmunohistologically examine the adrenal,kidney, liver and small intestine of Africangreen monkeys.33 ACAT1 was found in theadrenal cortex, a finding supported by theobservation that the adrenals of ACAT1 knock-out mice are depleted of cholesteryl ester(CE).37 In the kidney, ACAT1 was localized tothe distal tubules and podocytes of the kidney,and even though its function in this tissue is notwell understood, studies have shown CE enrich-ment of the tubules during renal injury.38

ACAT1 was localized within the Kupffer cells ofthe liver, and goblet cells, interstitialmacrophages and Paneth cells of the smallintestine, but was not found in the parenchymalcells. ACAT2 localization was limited to thehepatocytes in the liver and the apical portionof the enterocytes in the small intestine, whichare both sites of assembly and secretion ofapoB-containing lipoprotein particles.39,40

In spite of the lack of evidence that bothenzymes exist in the same cell types in non-human primates, determination of the locationof the isoforms in human liver has met withconflicting results. Immunolocalization experi-ments in human tissue taken 2 to 24 hourspost-mortem found that ACAT1 was localizedto the hepatocytes and Kupffer cells of theliver, while ACAT2 was found in fetal, but notadult, hepatocytes.35,41 In these experiments itwas difficult to ascertain the intracellularlocation of ACAT1 in liver due to low magnifi-cation and staining intensity of the images.When similar experiments were carried out inliver biopsies removed surgically from Swedishgallstone patients, the localization of ACAT1and ACAT2 was identical to that found in non-human primate liver, i.e. ACAT2 was localizedto the ER of the hepatocytes and ACAT1 to theER of the Kupffer cells.42 Assays with humanliver microsomes were used to determine theportion of total ACAT activity that was dueeither to ACAT1 or to ACAT2. Incubation withthe ACAT2–specific inhibitor, pyripyropene


ch04 14/7/05 4:45 pm Page 44

A,43 showed a 20–70% decrease in ACAT activ-ity in human liver versus >90% inhibition inAfrican green monkey liver microsomes. Thedata suggested that ACAT2 was responsible forthe majority of ACAT activity in almost all liversamples, although relative amounts of ACAT2activity were lower in human liver microsomesthan in those of monkeys. The relativeamounts of ACAT1 activity were the same forlivers from humans and monkeys. ACAT1 andACAT2 mRNA levels were quantified in thehuman livers by real-time PCR. The ratio ofACAT1 mRNA to ACAT2 mRNA was high inliver of both humans and monkeys, althoughmonkeys had relatively higher ACAT2 mRNAlevels. Recent studies in which ACAT mRNAwas quantified in various human tissues by real-time PCR demonstrated that hepatic ACATmRNA was 90% ACAT1 mRNA and 10%ACAT2 mRNA.44 The data seem consistent inshowing a relatively higher mRNA abundancefor ACAT1 than ACAT2 and this difference islarger than the differences in activity andestimated protein mass for the two enzymes. Intransfection studies using AC29 cells (R Temeland L Rudel, unpublished), the ACAT1protein was found to be 5- to 7-fold morerapidly degraded than the ACAT2 protein. If asimilar difference exists in the liver in vivo,then perhaps the relatively higher level ofACAT1 mRNA is related.

REGULATION OF ACAT1 AND ACAT2

Until recently, it was thought that neitherACAT gene was under transcriptional regula-tion. Despite the absence of detectable SRE45,46

and LXR elements47–49 in the 5' flankingregions there were several potential cis actingelements that may play a role in sterol depen-dent transcriptional regulation of the genes.Nevertheless, the only evidence for transcrip-tional regulation of the ACAT1 mRNA hasbeen upregulation by interferon-γ inmacrophages that might provide a partialexplanation for some of the pro-atherogeniceffects attributed to this cytokine.50,51 Evidence

of transcriptional regulation of ACAT2 was firstobserved in vitro where HepG2 cells adminis-tered citrus flavonoids showed decreases incholesterol esterification activity.52 While theflavonoids inhibited both ACAT isoforms tothe same degree, mRNA quantitation showedsignificant decreases of ACAT2 mRNA with nochanges in ACAT1 mRNA. Further evidencefor transcriptional regulation of ACAT2 wasobserved in recent and as yet unpublishedstudies (Parini, Angelin and Rudel, manuscriptin preparation) in which, for 30 days, Swedishpatients were administered either atorvastatin,a HMG CoA reductase inhibitor that loweredplasma cholesterol, or placebo. A liver biopsywas then surgically collected and RT-PCRmeasurements showed a 50% lower level ofACAT2 mRNA in patients receiving theatorvastatin when compared to controls, whileACAT1 mRNA levels were not different.

Additional evidence for transcriptionalregulation of ACAT2 was generated whencultured rat hepatocytes were incubated withchylomicron remnants enriched in n-3 polyun-saturated fatty acids. Expression of ACAT2mRNA dropped when compared to that inhepatocytes incubated with corn-oil-enrichedchylomicron remnants.53 Finally, it was shownthat cynomolgus monkeys, a primate specieshighly responsive to dietary cholesterolchallenge, had increased hepatic ACAT2mRNA when fed a cholesterol-enriched vs low-cholesterol diet.54 This evidence taken togethersuggests that ACAT2 may undergo somedegree of transcriptional regulation. However,further investigation is required to define themechanism and extent of this regulation.

Despite the recent evidence supporting ameasure of transcriptional regulation ofACAT2, post-transcriptional regulation at boththe protein and activity levels appears to be thepredominant mechanism of sterol-dependentregulation for both ACAT isoforms. Evidencefor post-transcriptional regulation at the activ-ity level derives from studies of the kineticsshowing that the cholesterol substrate satura-tion curves of both enzymes were found to besigmoidal, raising the strong possibility thatcholesterol is an allosteric activator of the


ch04 14/7/05 4:45 pm Page 45

ACAT enzymes,55 where binding of cholesterolcauses a conformational/structural changethat converts the enzyme from an inactive formto an active form. Studies by Cheng et al56

supported this conclusion when they expressedhuman ACAT1 in Sf9 insect cells and showedthat both cholesterol and 25-hydroxycholes-terol acted as an ACAT activator in intact cellsas well as in vitro.

Post-transcriptional regulation of ACAT1 andACAT2 at the protein level was recentlysuggested by Rudel et al54 when the livers fromtwo species of non-human primate wereanalysed for total ACAT activity, mRNAabundance for ACAT1 and ACAT2, and ACAT1and ACAT2 protein mass by quantitative westernblotting. Highly dietary-cholesterol-sensitivecynomolgus monkeys and the less responsiveAfrican green monkeys were compared whenfed either a cholesterol-enriched or low-choles-terol diet. ACAT activity was 75% higher whencholesterol was fed to cynos, but no differenceoccurred in greens. Given that over 90% ofACAT activity in monkey liver has been found tobe due to ACAT2,43 the increase in total ACATactivity induced by dietary cholesterol incynomolgus monkeys was probably due to anincrease in ACAT2 activity, and a high correla-tion was found when ACAT activity was corre-lated to ACAT2 protein mass.54 No increase inhepatic ACAT2 message or protein in livers ofthe green monkeys occurred upon challengewith the cholesterol-enriched diet. In contrast,the livers of the cynomolgus monkeys showed anapproximate 20% increase in ACAT2 mRNAand a 3-fold increase in ACAT2 protein whenthe cholesterol-enriched diet was fed. Thedisparity showing greater increases in proteinthan in mRNA led to the suggestion that muchof the cholesterol-dependent regulation of theprotein was post-transcriptional. The observa-tion was made that the relative level of hepaticACAT2 expression in humans and non-humanprimates (cynomolgus>African green>human)mirrors the degree of dietary cholesterolresponsiveness. This finding has led to specula-tion that ACAT2 is potentially an importantfactor in the genetic sensitivity to diet-inducedhypercholesterolaemia.

ACAT AND CHOLESTEROLABSORPTION IN THE SMALLINTESTINE

For many years, it has been recognized that themajority of newly absorbed cholesterol trans-ported in chylomicrons is esterified.57 ACAThas long been hypothesized to participate inthe esterification of cholesterol associated withintestinal cholesterol absorption and enzym-atic activity was first demonstrated in rat intesti-nal mucosa in 1976 by Haugen and Norum.58

The interest in cholesterol absorption and anyrole for ACAT is further supported by thecorrelation that exists between plasma choles-terol concentrations and percentage absorp-tion.59–62 Development of ACAT inhibitors hasbeen attempted in the pharmaceutical industryfor over 20 years.63 Data suggesting a role forACAT in cholesterol absorption have beengenerated by the administration of variousACAT inhibitors to different animal models.Many ACAT inhibitors have decreased intesti-nal cholesterol absorption, although such aneffect has not always been seen in humans andin non-human primates (see review by Sliskovicet al63,64).

Most of this work was done before it wasknown that there are two isoforms of ACATthat must be considered. Linkage analysis ofcrosses between low and high cholesterol-absorbing mouse strains revealed seven traitloci that influence cholesterol absorption.65

One locus mapped to the identical region ofchromosome 15 where the ACAT2 gene isfound, suggesting that regulation of ACAT2may be involved in the cholesterol absorptionpathway. Cholesterol absorption in ACAT1knockout mice fed either chow or a high-cholesterol diet was not different from wild-type mice, arguing against a role for ACAT1 incholesterol absorption.9 However, whenACAT2 knockout mice were challenged with ahigh-fat, high-cholesterol diet, absorption ofcholesterol decreased by 85% when comparedto wild-type animals, although cholesterolabsorption in chow-fed ACAT2 knockout micedid not show a decrease in cholesterol absorp-tion.66 This led to the conclusion that mice


ch04 14/7/05 4:45 pm Page 46

possess compensatory mechanisms thatmaintain basal levels of cholesterol absorptionin the absence of ACAT2, but are unable tocompensate when the animal is challengedwith elevated levels of dietary cholesterol. Insum, the strength of the evidence for a role ofACAT2 in regulating intestinal cholesterolabsorption is suggestive but not definitive.

Although potentially attractive, intestinalACAT inhibition as a pharmacological target forplasma cholesterol lowering in humans has yetto be fully exploited. While the mechanism ofintestinal cholesterol absorption is not fullyunderstood at this time, a current model is thatunesterified plant and animal sterols are trans-ported in bile micelles across the unstirred waterlayer where they interact with the enterocytes ofthe duodenum and jejunum. Absorption occursvia a saturable process that is inhibited byezetimibe (Zetia™), presumably through itseffects to limit uptake into the enterocyte via theNiemann-Pick C1 Like 1 (NPC1L1) protein.67 Inthe apical portion of the enterocyte, ACAT2 canreadily esterify absorbed cholesterol for incor-poration into the core of chylomicron particlesduring MTP-facilitated particle assembly. Theselipoproteins are subsequently secreted into thelymph and travel through the blood until thecholesteryl esters are taken up by the liver.68

Significant plasma LDL lowering with ezetimibehas been found which is additive to that seenwith statin administration.69 Efficacy for ezetim-ibe is not through any effect on ACAT, but thefindings establish the attractiveness of choles-terol absorption inhibition as a target for plasmaLDL cholesterol lowering, and ACAT2 repre-sents such a target.

Recent in vitro studies done in AC29 cellsstably expressing either ACAT1 or ACAT2provided evidence of another important roleof ACAT2 that may occur during the intestinalabsorption of sterols.70 In these experiments,microsomes isolated from either ACAT1- orACAT2-expressing cells were loaded witheither sitosterol or cholesterol using β-hydroxypropyl cyclodextrin. The mass ratio ofesterified cholesterol to esterified sitosterolsynthesized by ACAT1 was 1.6 in microsomesfrom ACAT1 cells and 7.2 in microsomes from

ACAT2 cells, indicating that ACAT2 exhibitssignificantly more selectivity than ACAT1 andfavours the esterification of cholesterol. Theauthors suggested that ACAT2 could act as a‘gatekeeper’ in the intestine that allows choles-terol to enter the circulation in the core ofchylomicron particles, while the plant sterolsthat are not efficiently esterified by ACAT2 areeffluxed back into the intestinal lumen via theABCG5/G8 transporter.71,72 The hypothesizedrole of the ABCG5/G8 transporter arises fromthe finding that mutations in the transporterslead to sitosterolaemia, a disease in whichpatients have elevated plasma concentrationsof plant sterols.73

ACAT AND HEPATIC LIPOPROTEINSECRETION

Over the last 10 years, several groups haveinvestigated the role of ACAT in both CEenrichment and secretion of apoB-containinglipoprotein particles. Both cell culture and invivo evidence suggest that CE availability mayplay a regulatory role in the secretion of apoBfrom cells. Several studies in HepG2 cells, ahuman hepatoma cell line, as well as in primaryhepatocytes, show that chemical inhibition ofACAT decreases the secretion of apoB fromthe cells.52,74–78 In vivo administration of ACATinhibitors to pigs fed a high-cholesterol diet ledto a 40% decrease in hepatic VLDL secretion,resulting in a 30% decrease in plasma VLDLcholesterol.79 Selected ACAT inhibitors admin-istered during perfusion of monkey livers ledto significant decreases in both CE and apoBaccumulation rates, with some decrease intriglyceride accumulation rates as well,although the patterns of effects on perfusateaccumulation rates were specific to individualACAT inhibitors.80 Interestingly, none of thethree different ACAT inhibitors was limited tospecific effects only on CE secretion. A strongassociation across all three inhibitors wasidentified when the per cent decrease in apoBsecretion was correlated to the per centdecrease in CE secretion. This findingsupported a hypothesized role for ACAT2 in


ch04 14/7/05 4:45 pm Page 47

coupling CE enrichment with whole-particlesecretion of apoB-containing lipoproteins.

Identification of pro-atherogenic propertiesof hepatic ACAT occurred when lipoproteinCE secretion was monitored for isolated,perfused livers of monkeys that had been fedvarious fatty-acid-enriched atherogenic dietsfor a period of 5 years.81 Liver perfusate CEaccumulation rates had highly significantpositive correlations (r ≥0.8) to coronary arteryatherosclerosis extent. This suggested that theenzyme responsible for hepatic CE synthesis,now known to be ACAT2, was important infacilitating the progression of atherosclerosis.The association between hepatic CE secretionwas equally strong in monkeys fed saturatedand mono-unsaturated fatty acids, and thisoccurred in spite of the LDL-cholesterol lower-ing and higher HDL/LDL-cholesterol ratio inthe mono-unsaturated fat group. The secretionof cholesteryl oleate as the primary CE waspromoted by dietary mono-unsaturated fat andenrichment of plasma lipoproteins with choles-teryl oleate appeared to be a factor in theatherosclerosis outcome.82 Studies in coronaryheart disease (CHD) patients and controlshave been published where the degree ofcholesteryl linoleate enrichment in plasma hasbeen found to be higher in controls than inpatients.83–87 The percentage of cholesteryloleate was inversely associated with thepercentage of cholesteryl linoleate, and thismay signify that similar effects occur in humansand monkeys, i.e. when hepatic ACAT-derivedcholesteryl oleate secretion is higher, morecoronary heart disease is found.

ATHEROSCLEROSIS AND ACAT INMACROPHAGES

A role for the cholesterol esterification reactionin the developing atherosclerotic lesion haslong been known. In fact, the most prominentcomponent of the atheromatous gruel fromwhich atherosclerosis derives its name is choles-teryl ester. The original isolation of a cDNA forACAT1 was from DNA derived from humanmacrophages.8 In almost all studies where it has

been examined, ACAT1 has been found to bethe enzyme present in macrophages includingthose associated with the artery wall.13–15 Theprocess of cholesterol accumulation in thearterial intima appears to follow from the infil-tration of plasma LDL, with higher concentra-tions of plasma LDL promoting more arterialaccumulation.88 The data indicate that thecholesteryl linoleate of LDL is hydrolysed andresynthesized and cholesteryl oleate becomesthe predominant CE in lesions. In this hydroly-sis, resynthesis appears to involve ACAT1 inmacrophages, with the esterification possiblyrepresenting conversion of excess amounts ofcholesterol into cholesteryl ester, the physicalform less damaging to cell membranes. Theaccumulation of higher melting cholesterylesters (e.g. enrichment with cholesteryl oleate)in lesions would appear to result in greateraccumulation through decreased mobiliza-tion.82,89,90 The isoform of ACAT that partici-pates in CE accumulation in atherogenesis hasnot been carefully documented in most studies,but since ACAT1 is found in almost all cell typesin the artery wall, the evidence favours thisenzyme isoform as the predominant player. Inmost cell types studied in tissue culture, ACAT1is the isoform expressed. One report describingthe appearance of ACAT2 in some cells ofatherosclerotic lesions has appeared,91 but theextent to which this was quantitatively signifi-cant was undefined. The factors in the promot-ers of the ACAT genes that determine tissueand cell type expression remain largely unstud-ied. It is of interest that, in vivo in most tissues,any one cell type appears to express onlyACAT1 or ACAT2, and in some tissues such asthe liver, for example, hepatocytes expressACAT2 while the adjacent Kupffer cells expressACAT1.33 However, when hepatoma cells, suchas HepG2 cells, are studied in tissue culture,both ACAT1 and ACAT2 are expressed.77

Clearly, the factors regulating expression of theACAT isoforms need to be identified to help usunderstand the physiological roles of theseenzymes.

While it has long been accepted that the intra-cellular location of the majority of the ACATenzymes is the membranes of the endoplasmic


ch04 14/7/05 4:45 pm Page 48

reticulum, Khelef et al92 have shown that a smallamount of ACAT1 is found in a paranuclearregion of macrophages that does not co-localizewith resident ER protein. Further characteriza-tion of the paranuclear ACAT showed that it waslocated in a region proximal to the trans-Golgiand the endocytic recycling compartment.93

Because these two organelles are important inthe trafficking of internalized cholesterol, it isconceivable that paranuclear ACAT1 may playan important role in the esterification of inter-nalized cholesterol and therefore the develop-ment of foam cells. More information is neededto define this relationship.

ATHEROSCLEROSIS IN ACAT1 ANDACAT2 KNOCKOUT MICE

The pathophysiological significance of ACAT1and ACAT2 in atherogenesis has been mostconvincingly demonstrated in the genedeletion studies done in mouse models ofatherosclerosis where the relative contribu-tions of the two enzymes to the progression ofthe disease have been studied. After theACAT1 knockout mouse had been character-ized,9,10 the ACAT1 gene deletion was bred intomice with LDL receptor or apoE genedeletions by Accad et al94 and by Yagyu et al95 toexamine the effect of the loss of ACAT1 on thedevelopment of atherosclerosis. One group feda 0.15% cholesterol diet to the ACAT1–/–,apoE–/– mice and a 1.25% cholesterol diet tothe ACAT1–/–, LDLr–/– mice for 90 days.95 Atthe end of the study, total plasma cholesteroldid not change significantly in the ACAT1–/–,LDLr–/– mice when compared to ACAT1+/+,LDLr–/– controls, but plasma cholesterol in theACAT1–/– apoE–/– dropped by ~40% whencompared to chow-fed animals. When athero-sclerosis was measured as aortic CE content theACAT1–/–, LDLr–/– and ACAT1–/–, apoE–/–

showed decreases of 2-fold and 3-fold, respec-tively, when compared to controls.

However, the beneficial effects on athero-sclerosis were counterbalanced by the dry eyesyndrome and cutaneous xanthomatosisobserved in the mice lacking ACAT1.

Histological examination of the eye revealedthat atrophy of the meibomian glands, amodified sebaceous gland, caused the dry eyein the mice. That fact that ACAT1 is highlyexpressed in the sebaceous gland suggests thatloss of the enzyme led to the inability of themeibomian gland to promote tearing. Thecutaneous xanthomatosis was also a seriousproblem for the mice, causing loss of hair andlesions on the skin. Biochemical and histologi-cal examination of the skin revealed that 6- to7-fold higher skin cholesterol levels hadresulted in cholesterol crystal deposition andacute inflammation in the skin.94 These investi-gators described the truncal skin as beingessentially one massive cholesterol xanthoma.

In an effort to avoid the severe skin pathologyobserved in the ACAT1 knockout mice, Fazio etal96 transplanted ACAT1–/– bone marrow intoLDLr–/– mice. This procedure generated amouse with ACAT1-deficient macrophages, butnormal ACAT1 expression in other cells. Theseanimals were then placed on a 0.15% choles-terol diet for 12 weeks, and atherosclerosis wasmeasured as the percent of the aorta covered inlesion. Even though the mice with the ACAT1–/–

macrophages did not exhibit the skin pathologyor dry eye observed in the ACAT1–/– mice, theyhad 2- to 3-fold more plaque involvement thancontrol mice. Staining of the aortas formacrophages showed that the mice withACAT1–/– macrophages had fewer macro-phages in their lesions than the control mice. Ithad been shown in tissue culture that inhibitionof ACAT1 in macrophages exposed to choles-terol-rich lipoproteins can be cytotoxic due tothe cell’s inability to store cholesterol in lipiddroplets.4,97,98 To address the possibility that thiscould occur in vivo, the extent of apoptosis andnecrosis in the aorta was examined by TUNELstaining. The mice with ACAT1–/– macrophageshad 3-fold more apoptosis than the othermice.99 Premature death of macrophages wascited by the authors as probably being theprincipal cause of the increased atherosclerosisin the mice.

Taken together, the results of these studiesargue that while modest decreases in athero-sclerosis associated with ACAT1 gene deletion


ch04 14/7/05 4:45 pm Page 49

were sometimes observed, the noteworthyaccompanying negative side-effects in the eyeand skin create considerable doubt as to theefficacy of ACAT1 as a pharmaceutical target.The data suggest that even the macrophage-specific ACAT1 deficiency may not lead todesirable decreases in foam cell accumulationand atherosclerosis.

By contrast, the results of studies in theACAT2–/– mouse indicate that the gene deletionof the ACAT2 isoform is an effective anti-ather-osclerotic strategy in mice with minor side-effects. Characterization of the ACAT2–/– mouseshowed that ACAT activity in the liver and thesmall intestine was decreased by greater than90%.66 When a chow diet was fed for 3 weeks,plasma cholesterol values for the ACAT2–/– micewere unaffected. However, when these micewere fed the cholesterol, fat and cholate-richPaigen diet for 3 weeks, plasma cholesterol was56% less than in ACAT2+/+ control mice. Theauthors speculated that the resistance to dietarycholesterol challenge was related to the >80%decrease in cholesterol absorption observed inthe Paigen-diet-fed mice. Adverse side-effectswere not observed in ACAT2-deficient mice fedeither chow or Paigen diets. The effect of theACAT2 gene deletion on atherosclerosis wasexamined by crossing it into the apoE–/–

mouse.100 Absence of ACAT2 in the apoE–/–

mice led to a 60% decrease in plasma choles-terol and a 70% decrease in plasma CE whencompared to ACAT2+/+, apoE–/– controls. After27 weeks on a chow diet the ACAT2+/+, apoE–/–

mice had 4.8 ± 2.9% of their aorta surface areacovered in lesions compared to 0.1 ± 0.2% in theACAT2–/–, apoE–/– mice. The authors speculatedthat the loss of ACAT2 was athero-protective due

to (1) the replacement of cholesteryl ester withtriglyceride in the core of apoB-containing parti-cles and (2) the depletion of pro-atherogeniccholesteryl esters, e.g. cholesteryl oleate, result-ing in an enrichment of LCAT-derived CE, e.g.cholesteryl linoleate and cholesteryl arachido-nate. Additional studies have recently beencompleted in LDLr–/–, ACAT2–/– mice.101 In thismodel, as in the apoE–/–, ACAT2–/– mouse, thedeficiency of ACAT2 protected the mice againstthe development of atherosclerosis (aorticlesion surface areas of 1 vs 5% and CE concen-trations of 3 vs 25 mg/g protein in ACAT2–/– vsACAT2+/+, LDLr–/– mice, respectively).

CONCLUDING REMARKS

The discovery of the first ACAT gene sequenceoccurred only about 10 years ago, but theexplosion of new knowledge about ACAT1 andACAT2 in the intervening period of time hasbeen significant. We have tried to discuss manyof the discoveries that have contributed to ourpresent understanding of the biochemical andphysiological roles of these enzymes in choles-terol homeostasis. In most cases, more infor-mation is needed for us to be certain aboutisoform specificity and associated molecularmechanisms surrounding ACAT-mediatedcholesterol esterification as it affects a physio-logical or pathophysiological process, e.g.intestinal cholesterol absorption and athero-sclerosis. Nevertheless, the information that isnow available has brought us to a higher levelof understanding and has provided a soliddatabase upon which to develop hypotheses forfuture experimentation.


1. Chang TY, Chang CCY, Cheng D. Acyl-coenzyme A:Cholesterol acyltransferase. Annu Rev Biochem 1997;66:613–38

2. Hamilton JA, Small DM. Solubilization and localiza-tion of cholesteryl oleate in egg phosphatidylcholinevesicles. A carbon 13 NMR study. J Biol Chem 1982;257:7318–21

3. Small DM, Shipley GG. Physical-chemical basis oflipid deposition in atherosclerosis. Science 1974;185:222–9

4. Warner GJ, Stoudt G, Bamberger M, et al. Cell toxic-ity induced by inhibition of acyl coenzyme A:choles-terol acyltransferase and accumulation of unesterifiedcholesterol. J Biol Chem 1995; 270:5772–8

References

ch04 14/7/05 4:45 pm Page 50

5. Goodman DS, Deykin D, Shiratori T. The formationof cholesterol esters with rat liver enzymes. J BiolChem 1964; 239:1335–45

6. Doolittle GM, Chang TY. Solubilization, partial purifi-cation, and reconstitution in phosphatidylcholine-cholesterol liposomes of acyl-CoA:cholesterolacyltransferase. Biochemistry 1982; 21:674–9

7. Cadigan KM, Heider JG, Chang T-Y. Isolation andcharacterization of Chinese hamster ovary cellmutants deficient in acyl-coenzyme A:cholesterolacyltransferase activity. J Biol Chem 1988; 263:274–82

8. Chang CCY, Huh HY, Cadigan KM, Chang TY.Molecular cloning and functional expression ofhuman acyl-coenzyme A:cholesterol acyltransferasecDNA in mutant Chinese hamster ovary cells. J BiolChem 1993; 268:20747–55

9. Meiner V, Tam C, Gunn MD, et al. Tissue expressionstudies on the mouse acyl-CoA:cholesterol acyltrans-ferase gene (Acact): findings supporting the existenceof multiple cholesterol esterification enzymes inmice. J Lipid Res 1997; 38:1928–33

10. Meiner VL, Cases S, Myers HM, et al. Disruption ofthe acyl-CoA:cholesterol acyltransferase gene in mice:evidence suggesting multiple cholesterol esterifica-tion enzymes in mammals. Proc Natl Acad Sci USA1996; 93:14041–6

11. Yang H, Bard M, Bruner DA, et al. Sterol esterifica-tion in yeast: a two-gene process. Science 1996;272:1353–6

12. Cases S, Smith SJ, Zheng YW, et al. Identification of agene encoding an acyl CoA:diacylglycerol acyltrans-ferase, a key enzyme in triacylglycerol synthesis. ProcNatl Acad Sci USA 1998; 95:13018–23

13. Anderson RA, Joyce C, Davis M, et al. Identification ofa form of acyl-CoA:cholesterol acyltransferase specificto liver and intestine in nonhuman primates. J BiolChem 1998; 273:26747–54

14. Cases S, Novak S, Zheng Y-W, et al. ACAT-2, a secondmammalian acyl-CoA:cholesterol acyltransferase. Itscloning, expression, and characterization. J BiolChem 1998; 273:26755–64

15. Oelkers P, Behari A, Cromley D, et al.Characterization of two human genes encoding acylcoenzyme A:cholesterol acyltransferase-relatedenzymes. J Biol Chem 1998; 273:26765–71

16. Buhman KF, Accad M, Farese RV Jr. Mammalian acyl-CoA: cholesterol acyltransferases. Biochim BiophysActa Mol Cell Biol Lipids 2000; 1529:142–54

17. Li BL, Li XL, Duan ZJ, Lee O, et al. Human acyl-CoA:cholesterol acyltransferase-1 (ACAT-1) geneorganization and evidence that the 4.3-kilobaseACAT-1 mRNA is produced from two differentchromosomes. J Biol Chem 1999; 274:11060–71

18. Katsuren K, Tamura T, Arashiro R, et al. Structure ofthe human acyl-CoA:cholesterol acyltransferase-2(ACAT-2) gene and its relation to dyslipidemia.Biochim Biophys Acta Mol Cell Biol Lipids 2001;1531:230–40

19. Song BL, Qi W, Yang XY, et al. Organization of

human ACAT-2 gene and its cell-type-specificpromoter activity. Biochem Biophys Res Commun2001; 282:580–8

20. Song BL, Qi W, Wang CH, et al. Preparation of ananti-Cdx-2 antibody for analysis of different speciesCdx-2 binding to acat2 promoter. Biochim BiophysActa Sinica 2003; 35:6–12

21. Wang CS, McConathy WJ, Kloer HU, et al.Modulation of lipoprotein lipase activity byapolipoproteins: effect of apolipoprotein C-III. J ClinInvest 1984; 75:384–90

22. von Kap-herr C, Cockman T, Rudel L, et al. Assignmentof acyl-CoA:cholesterol acyltransferase 1 and 2 ((S)AT1,SOAT2)) and diacylglycerol O-acyltransferase 1(DGAT1) to M. fascicularis chromosome band 1p32,12q13, 8qter; C. aethiops sabaeus 13q22, 3q12, 1qter; S.sciureus 19q22, 15q21, 16qter by in situ hybridization.Gytogenet Genome Res 2003; 103:203

23. Guo ZM, Cromley D, Billheimer JT, et al.Identification of potential substrate-binding sites inyeast and human acyl-CoA sterol acyltransferases bymutagenesis of conserved sequences. J Lipid Res2001; 42:1282–91

24. Joyce CW, Shelness GS, Davis MA, et al. ACAT1 andACAT 2 membrane topology segregates a serineresidue essential for activity to opposite sides of theendoplasmic reticulum membrane. Mol Biol Cell2000; 11:3675–87

25. Lin S, Lu X, Chang CCY, et al. Human acyl-coenzymeA:cholesterol acyltransferase expressed in Chinesehamster ovary cells: membrane topology and activesite location. Mol Biol Cell 2003; 14:2447–60

26. Lin S, Cheng D, Liu MS, et al. Human acyl-CoA:cholesterol acyltransferase-1 in the endoplasmicreticulum contains seven transmembrane domains. JBiol Chem 1999; 274:23276–85

27. Galvan DL, Borrego-Diaz E, Perez PJ, et al. Subunitoligomerization, and topology of the inositol 1,4,5–trisphosphate receptor. J Biol Chem 1999;274:29483–92

28. Mo C, Holland TC. Determination of the transmem-brane topology of herpes simplex virus type 1 glyco-protein K. J Biol Chem 1997; 272:33305–11

29. Heymann JA, Subramaniam S. Integration of deletionmutants of bovine rhodopsin into the membrane ofthe endoplasmic reticulum. Mol Membr Biol 2000;17:165–74

30. Manoil C, Traxler B. Insertion of in-frame sequencetags into proteins using transposons. Methods 2000;20:55–61

31. Billheimer JT, Cromley DA, Kempner ES. Thefunctional size of acyl-coenzyme A (CoA):cholesterolacyltransferase and acyl-CoA hydrolase as determinedby radiation inactivation. J Biol Chem 1990;265:8632–5

32. Erickson SK, Lear SR, McCreery MJ. Functional sizesof hepatic enzymes of cholesteryl ester metabolismdetermined by radiation inactivation. J Lipid Res1994; 35:763–9


ch04 14/7/05 4:45 pm Page 51

33. Lee RG, Willingham MC, Davis MA, et al. Differentialexpression of ACAT1 and ACAT2 among cells withinliver, intestine, kidney, and adrenal of nonhumanprimates. J Lipid Res 2000; 41:1991–2001

34. Yu CJ, Chen J, Lin S, et al. Human acyl-CoA:cholesterolacyltransferase-1 is a homotetrameric enzyme in intactcells and in vitro. J Biol Chem 1999; 274:36139–45

35. Chang CCY, Sakashita N, Ornvold K, et al.Immunological quantitation and localization ofACAT-1 and ACAT-2 in human liver and small intes-tine. J Biol Chem 2000; 275:28083–92

36. Yu CJ, Zhang Y, Lu XH, et al. Role of the N-terminalhydrophilic domain of acyl-coenzyme A:cholesterolacyltransferase I on the enzyme’s quaternary structureand catalytic efficiency. Biochemistry 2002; 41:3762–9

37. Meiner VL, Welch CL, Cases S, et al. Adrenocorticallipid depletion gene (ald) in AKR mice is associatedwith an acyl-CoA:cholesterol acyltransferase (ACAT)mutation. J Biol Chem 1998; 273:1064–9

38. Zager RA, Kalhorn TF. Changes in free and esterifiedcholesterol. Am J Pathol 2000; 157:1007–16

39. Alexander CA, Hamilton RL, Havel RJ. Subcellularlocalization of B apoprotein of plasma lipoproteins inrat liver. J Cell Biol 1983; 69:241–63

40. Cardell RR, Jr, Badenhausen S, Porter KR. Intestinaltriglyceride absorption in the rat. An electron micro-scopical study. J Cell Biol 1967; 34:123–55

41. Sakashita N, Miyazaki A, Takeya M, et al. Localizationof human acyl-coenzyme A:cholesterol acyltrans-ferase-1 (ACAT-1) in macrophages and in varioustissues. Am J Pathol 2000; 156:227–36

42. Parini P, Davis M, Lada AT, et al. ACAT2 is localized tohepatocytes and is the major cholesterol-esterifyingenzyme in human liver. Circulation 2004; 110:2017–23

43. Lada AT, Davis M, Kent C, et al. Identification ofACAT1– and ACAT2-specific inhibitors using a novel,cell-based fluorescent assay: individual ACAT unique-ness. J Lipid Res 2004; 45:378–86

44. Smith JL, Rangaraj K, Simpson R, et al. Quantitativeanalysis of expression of ACAT genes in humantissues by real-time PCR. J Lipid Res 2004; 45:686–96

45. Brown MS, Goldstein JL. The SREBP pathway: regula-tion of cholesterol metabolism by proteolysis of amembrane-bound transcription factor. Cell 1997;89:331–40

46. Horton JD, Goldstein JL, Brown MS. SREBPs: activa-tors of the complete program of cholesterol and fattyacid synthesis in the liver. J Clin Invest 2002;109:1125–31

47. Edwards PA, Kast HR, Anisfeld AM. BAREing it all:the adoption of LXR and FXR and their roles in lipidhomeostasis. J Lipid Res 2002; 43:2–12

48. Lehmann JM, Kliewer SA, Moore LB, et al. Activationof the nuclear receptor LXR by oxysterols defines anew hormone response pathway. J Biol Chem 1997;272:3137–40

49. Janowski BA, Willy PJ, Devi TR, et al. An oxysterolsignaling pathway mediated by the nuclear receptorLXRα. Nature 1996; 383:728–31

50. Yang JB, Duan ZJ, Yao W, et al. Synergistic transcrip-tional activation of human acyl-coenzyme A:cholesterolacyltransterase-1 gene by interferon-gamma and all-trans-retinoic acid THP-1 cells. J Biol Chem 2001;276:20989–98

51. Panousis CG, Zuckerman SH. Regulation of choles-terol distribution in macrophage-derived foam cellsby interferon-gamma. J Lipid Res 2000; 41:75–83

52. Wilcox LJ, Borradaile NM, de Dreu LE, et al. Thesecretion of hepatocyte apoB is inhibited by theflavonoids, naringenin and hesperetin, via reducedactivity and expression of ACAT2 and MTP. J LipidRes 2001; 42:725–34

53. Botham KM, Zheng XZ, Napolitano M, et al. Theeffects of dietary n-3 polyunsaturated fatty acids deliv-ered in chylomicron remnants on the transcription ofgenes regulating synthesis and secretion of very-low-density lipoprotein by the liver: modulation by cellu-lar oxidative state. Exp Biol Med. 2003; 228:143–51

54. Rudel LL, Davis M, Sawyer J, et al. Primates highlyresponsive to dietary cholesterol upregulate hepaticACAT2 while less responsive primates do not. J BiolChem 2002; 277:31401–6

55. Chang CCY, Lee C-YC, Chang ET, et al. Recombinantacyl-CoA:cholesterol acyltransferase-1 (ACAT-1)purified to essential homogeneity utilizes cholesterolin mixed micelles or in vesicles in a highly cooperativemanner. J Biol Chem 1998; 273:35132–41

56. Cheng D, Chang CCY, Qu X, et al. Activation of acyl-coenzyme A:cholesterol acyltransferase by cholesterolor by oxysterol in a cell-free system. J Biol Chem 1995;270:685–95

57. Rudel LL, Morris MD, Felts JM. The transport ofexogenous cholesterol in the rabbit. I. Role of choles-terol ester of lymph chylomicron and lymph very lowdensity lipoproteins in absorption. J Clin Invest 1972;51:2686–92

58. Haugen R, Norum KR. Coenzyme-A-dependent ester-ification of cholesterol in rat intestinal mucosa. ScandJ Gastroenterol 1976; 11:615–21

59. Miettinen TA, Kesaniemi A. Cholesterol absorption:regulation of cholesterol synthesis and eliminationand within-population variations of serum cholesterollevels. Am J Clin Nutr 1989; 49:629–35

60. Kesaniemi YA, Miettinen TA. Cholesterol absorptionefficiency regulates plasma cholesterol level in theFinnish population. Eur J Clin Invest 1987; 17:391–5

61. Miettinen TA, Gylling H. Cholesterol absorptionefficiency and sterol metabolism in obesity.Atherosclerosis 2000; 153:241–8

62. Rudel L, Deckelman C, Wilson M, et al. Dietarycholesterol and downregulation of cholesterol 7α-hydroxylase and cholesterol absorption in Africangreen monkeys. J Clin Invest 1994; 93:2463–72

63. Sliskovic DR, Picard JA, Krause BR. ACAT inhibitors:the search for a novel and effective treatment ofhypercholesterolemia and atherosclerosis. ProgMedicinal Chem 2002; 39:121–71

64. Wrenn SM Jr, Parks JS, Immermann FW, et al. ACAT


ch04 14/7/05 4:45 pm Page 52

inhibitors CL 283,546 and CL 283,796 reduce LDLcholesterol without affecting cholesterol absorptionin African green monkeys. J Lipid Res 1995;36:1199–210

65. Schwarz M, Davis DLVBR, Russell DW. Genetic analy-sis of intestinal cholesterol absorption in inbred mice.J Lipid Res 2001; 42:1801–11

66. Buhman KK, Accad M, Novak S, et al. Resistance todiet-induced hypercholesterolemia and gallstoneformation in ACAT2–deficient mice. Nature Med2000; 6:1341–7

67. Altmann SW, Davis HR Jr, Zhu L, et al. Niemann-PickC1 like 1 protein is critical for intestinal cholesterolabsorption. Science 2004; 303:1201–4

68. Goodman DS. The metabolism of chylomicroncholesteryl ester in the rat. J Clin Invest 1962;41:1886–96

69. Bruckert E, Giral P, Tellier P. Perspectives in choles-terol-lowering therapy – The role of ezetimibe, a newselective inhibitor of intestinal cholesterol absorp-tion. Circulation 2003; 107:3124–8

70. Temel RE, Gebre AK, Parks JS, et al. Compared withacyl-CoA:cholesterol O-acyltransferase (ACAT)1 andlecithin:cholesterol acyltransferase, ACAT2 displaysthe greatest capacity to differentiate cholesterol fromsitosterol. J Biol Chem 2003; 278:47594–601

71. Yu LQ, Hammer RE, Li-Hawkins J, et al. Disruption ofAbcg5 and Abcg8 in mice reveals their crucial role inbiliary cholesterol secretion. Proc Natl Acad Sci USA2002; 99:16237–42

72. Yu L, Li-Hawkins J, Hammer RE, et al. Overexpressionof ABCG5 and ABCG8 promotes biliary cholesterolsecretion and reduces fractional absorption of dietarycholesterol. J Clin Invest 2002; 110:671–80

73. Lee MH, Lu K, Patel SB. Genetic basis of sitos-terolemia. Curr Opin Lipidol 2001; 12:141–9

74. Avramoglu RK, Cianflone K, Sniderman AD. Role ofthe neutral lipid accessible pool in the regulation ofsecretion of apoB-100 lipoprotein particles by HepG2cells. J Lipid Res 1995; 36:2513–28

75. Cianflone KM, Yasruel Z, Rodriquez MA, et al.Regulation of apoB secretion from HepG2 cells:evidence for a critical role for cholesteryl ester synthe-sis in the response to a fatty acid challenge. J LipidRes 1990; 31:2045–55

76. Sniderman AD, Zhang Z, Genest J, et al. Effects onapoB-100 secretion and bile acid synthesis by redirect-ing cholesterol efflux from HepG2 cells. J Lipid Res2003; 44:527–32

77. Wilcox LJ, Barrett PHR, Newton RS, et al. ApoB100secretion from HepG2 cells is decreased by the ACATinhibitor CI-1011 – an effect associated withenhanced intracellular degradation of ApoB.Arterioscler Thromb Vasc Biol 1999; 19:939–49

78. Zhang ZJ, Cianflone K, Sniderman AD. Role ofcholesterol ester mass in regulation of secretion ofApoB100 lipoprotein particles by hamster hepato-cytes and effects of statins on that relationship.Arterioscler Thromb Vasc Biol 1999; 19:743–52

79. Huff MW, Telford DE, Barrett PHR, et al. Inhibitionof hepatic ACAT decreases apoB secretion in minia-ture pigs fed a cholesterol-free diet. ArteriosclerThromb 1994; 14:1498–508

80. Carr TP, Hamilton RL Jr, Rudel LL. ACAT inhibitorsdecrease secretion of cholesteryl esters andapolipoprotein B by perfused livers of African greenmonkeys. J Lipid Res 1995; 36:25–36

81. Rudel LL, Haines J, Sawyer JK, et al. Hepatic origin ofcholesteryl oleate in coronary artery atherosclerosis inAfrican green monkeys. Enrichment by dietarymonounsaturated fat. J Clin Invest 1997; 100:74–83

82. Lada AT, Rudel LL, St Clair RW. Effects of LDLenriched with different dietary fatty acids on choles-teryl ester accumulation and turnover in THP-1macrophages. J Lipid Res 2003; 44:770–9

83. Logan RL, Thomson M, Riemersma RA, et al. Riskfactors for ischaemic heart-disease in normal menaged 40. Lancet 1978; i:949–55

84. Lewis B. Composition of plasma cholesterol ester inrelation to coronary-artery disease and dietary fat.Lancet 1958; ii:71–3

85. Kingsbury KJ, Brett C, Stovold R, et al. Abnormal fattyacid composition and human atherosclerosis.Postgrad Med J 1974; 50:425–40

86. Kingsbury KJ, Morgan DM, Stovold R, et al.Polyunsaturated fatty acids and myocardial infarction.Follow-up of patients with aortoiliac and femoro-popliteal atherosclerosis. Lancet 1969; ii:1325–9

87. Lawrie TDV, McAlpine SG, Rifkind BM, et al. Serumfatty-acid patterns in coronary-artery disease. Lancet1961; 14:421–4

88. Smith EB. The relationship between plasma andtissue lipids in human atherosclerosis. Adv Lipid Res1974; 12:1–49

89. Mahlberg FH, Glick JM, Jerome WG, et al.Metabolism of cholesteryl ester lipid droplets in aJ774 macrophage foam cell model. Biochim BiophysActa 1990; 1045:291–8

90. Glick JM, Adelman SJ, Phillips MC, et al. Cellularcholesteryl ester clearance. Relationship to the physi-cal state of cholesteryl ester inclusions. J Biol Chem1983; 258:13425–30

91. Sakashita N, Miyazaki A, Chang CCY, et al. Acyl-coenzyme A:cholesterol acyltransferase 2 (ACAT2) isinduced in monocyte-derived macrophages: in vivoand in vitro studies. Lab Invest 2003; 83:1569–81

92. Khelef N, Buton X, Beatini N, et al. Immuno-localization of acyl-coenzyme A:cholesterol O-acyltransferase in macrophages. J Biol Chem 1998;273:11218–24

93. Khelef N, Soe TT, Quehenberger O, et al. Enrich-ment of acyl coenzyme A:cholesterol o-acyltransferasenear trans-Golgi network and endocytic recyclingcompartment. Arterioscler Thromb Vasc Biol 2000;20:1769–76

94. Accad M, Smith SJ, Newland DL, et al. Massivexanthomatosis and altered composition of atheroscle-rotic lesions in hyperlipidemic mice lacking acyl


ch04 14/7/05 4:45 pm Page 53

CoA:cholesterol acyltransferase 1. J Clin Invest 2000;105:711–19

95. Yagyu H, Kitamine T, Osuga J, et al. Absence ofACAT-1 attenuates atherosclerosis but causes dry eyeand cutaneous xanthomatosis in mice with congenitalhyperlipidemia. J Biol Chem 2000; 275:21324–30

96. Fazio S, Liu L, Major AS, et al. Accelerated athero-sclerosis in LDL receptor null mice reconstituted withACAT negative macrophage. Circulation 1999; 100:1-613

97. Kellner-Weibel G, Jerome WG, Small DM, et al.Effects of intracellular free cholesterol accumulationon macrophage viability: a model for foam cell death.Arterioscler Thromb Vasc Biol 1998; 18:423–31

98. Maccarrone M, Bellincampi L, Melino G, et al.

Cholesterol, but not its esters, triggers programmedcell death in human erythroleukemia K562 cells. EurJ Biochem 1998; 253:107–13

99. Fazio S, Major AS, Swift LL, et al. Increased athero-sclerosis in LDL receptor-null mice lacking ACAT1 inmacrophages. J Clin Invest 2001; 107:163–71

100. Willner EL, Tow B, Buhman KK, et al. Deficiency ofacyl CoA:cholesterol acyltransferase 2 prevents ather-osclerosis in apolipoprotein E-deficient mice. ProcNatl Acad Sci USA 2003; 100:1262–7

101. Lee RG, Kelley KL, Sawyer JK, et al. Plasma cholesterylesters provided by lecithin:cholesterol acyltransferaseand acyl-Coenzyme A:cholesterol acyltransferase 2have opposite atherogenic potential. Circ Res 2004;95:998–1004


ch04 14/7/05 4:45 pm Page 54

INTRODUCTION

One of the most important problems inmodern Western society is increasing bodyweight and obesity. This is due to increasedabsolute amount of calorie consumption inrelation to decreased energy utilization bylimited physical activity. Even though theexcess calorie intake is not solely due toenhanced consumption of fat-containing food,weight reduction of obese subjects has renewedthe idea of using low-carbohydrate diet, notonly for acute weight control but also for long-term maintenance of body weight.1 The result-ing relative increase in fat intake involvesmainly less-saturated vegetable oil productsand also fish oil products.

Vegetable oils are the major sources ofdietary phytosterols (plant sterols: mainlycampesterol and sitosterol and their 5α-saturated stanols). Plant sterols from thenatural plant products and especially fromartificial functional food preparations inhibitcholesterol absorption and lower serum choles-terol levels. For that purpose, poorly solubleplant stanols were converted to fat-solubleesters and used as a plant stanol estermargarine2 and later a plant sterol ester spread3

was developed for serum cholesterol lowering.Dramatic reduction of cardiovascular risk withlong-term statin treatment, even in subjectswith a relatively low baseline cholesterol level,4

has increased interest in lowering serum choles-terol concentration through the long-termconsumption of diets enriched with phyto-

sterols.5 Fat-soluble plant stanol or sterol estersare the major products used for that purpose.6

Their increased use has raised several interest-ing scientific and also practical questionsconcerning the role of the matrix of the fat inthe diet, mechanism of cholesterol absorptionand inhibition, regulation of plant sterolabsorption and interference of saturated plantstanols in sterol absorption. Recent findings onthe role of the enterocyte ATP-binding cassettehalf-transporters ABC G5 and G8 gene expres-sions in sterol absorption,7 and the introduc-tion of a specific inhibitor of cholesterolabsorption, ezetimibe,8 have widened theunderstanding of intestinal sterol metabolism.Recently the Niemann-Pick C1 like 1 proteinhas been introduced as a putative cholesterol.9

However, the exact mechanisms of intestinalsterol metabolism at the molecular level are stillpoorly understood. This chapter will review theprinciples of fat, acidic and, especially, neutralsterol metabolism primarily in the small intes-tine of man, not dealing specifically withchylomicron formation in enterocytes or bacte-rial metabolic phenomena in the large bowel.Several review articles on different aspects ofthe intestinal lipid metabolism have recentlybeen published.6,10–13

PRINCIPLES OF DIETARY FATMETABOLISM

The largest component of dietary lipidsconsists of triglycerides (about 60–80 g/day),

Overview of intestinal lipid metabolismT.A. Miettinen and H. Gylling

5

ch05 14/7/05 4:45 pm Page 55

the actual fats, followed by phospholipids(about 5 g/day) and cholesterol (about200–500 mg/day). In addition to dietaryintake, lipids enter the gastrointestinal tractfrom desquamated epithelial cells, and choles-terol and phospholipids are also delivered tothe intestine by bile. Even though the molecu-lar structures of lipids differ markedly fromeach other, they have one feature in common:they are all hydrophobic compounds, whichmakes their digestion, absorption and trans-port in aqueous medium complicated.Phospholipids, derived from both diet andbile, have both hydrophilic and hydrophobicgroups in their structure, which enables themto act as bridges between hydrophobic andhydrophilic milieu. The following section dealsmainly with fats, even though cholesterol andphospholipids are also mentioned.

Emulsification and lipolysis of fats in thestomach

Digestion of dietary fats consists of emulsifica-tion of fats in the stomach, lipolysis by lipasesin the stomach and especially in the duode-num, solubilization with bile salts in the duode-num and absorption in the small intestine.Digestion starts in the oral cavity, where saliva-tion and mastication start to separate fats fromfood and disperse the large fat droplets intosmaller ones. The stomach contributes to fatdigestion by both emulsification and hydrolysisof triglycerides by gastric lipase. The fats arewarmed to body temperature and then enter aliquid phase. The grinding action of theantrum contributes to the emulsification, yield-ing smaller and smaller fat droplets and agreater surface area, which enhances lipolysis.

Gastric lipase is the main enzyme performingintragastric lipolysis, whereas the importance oflingual lipase is minimal in humans.14 Gastriclipase is secreted by principal cells of thestomach, and it acts in an acidic milieu. It issubstrate-specific and hydrolyses only trigly-cerides and only the fatty acids in the sn-3position. Accordingly, the major products ofgastric lipase action are short-, medium- orlong-chain free fatty acids and diglycerides.

However, intragastric hydrolysis is importantfor triglyceride cleavage, because it is assumedthat about one-third of fat lipolysis takes placein the stomach.14 This is the reason why patientslacking pancreatic lipase can escape massivesteatorrhoea. Mastication, antral grinding andlipolysis lead to an unstable gastric emulsioncalled chyme, which is released into the duode-num in a volume of 3 ml twice a minute.

The release of long-chain fatty acids in thestomach stimulates the release of cholecys-tokinin, which in turn stimulates the secretionof pancreatic lipase. Accordingly, gastric lipoly-sis prepares the intraduodenal circumstancesfor fat digestion. In addition, a complicatedinterplay regulates the motility of the gastroin-testinal tract in order to secure as complete fatdigestion as possible. For example, dietary fatsregulate the emptying rate of the stomach. Inthe duodenal mucosa there are lipid-sensitivereceptors, which, in connection with neuralpathways, are capable of slowing gastric empty-ing in order to potentiate gastric emulsificationand lipolysis. A similar ‘brake’ system operatesin the terminal ileum and proximal colon.Intraluminal fats, bile salts and complex carbo-hydrates arriving to these parts of the gastro-intestinal tract activate an endocrine–neurohumoral cascade, which decreases gastricemptying and small bowel transit.

Lipolysis in the small intestine

The entry of chyme into the duodenum stimu-lates the marked release of cholecystokinin,which inhibits gut mobility and enhances secre-tion of bile acids from the gall bladder andpancreatic juice with several lipases. Of theselipolytic enzymes, the most important ones forfat digestion are colipase and pancreatic lipase.The pancreatic lipase is an interfacial enzyme,which works best in a heterogeneous milieu ofwater–oil interface. To optimize the lipolyticeffect, emulsified fats are solubilized by bilesalts and phospholipids to form aggregatescalled mixed micelles. The mixed micelles alsocontain free fatty acids, cholesterol and othersterols and fat-soluble vitamins. In bile, whenthe concentration of conjugated bile acids


ch05 14/7/05 4:45 pm Page 56

exceeds 2 mmol/l, they aggregate and formmicelles with cholesterol and phospholipids.The concentration of bile salts needed formicelle formation is called the critical micellarconcentration. However, if the concentrationof bile salts exceeds the critical micellarconcentration, they inhibit the lipase activity bya physical competition for the lipid interface.

Pancreatic lipase hydrolyses fatty acids fromthe glyceride backbone at the sn-1 and sn-3positions, releasing free fatty acids andmonoglycerides. Pancreatic lipase is activatedafter colipase anchors it to the substrate inter-face, after which lipolysis can be started. Inaddition, colipase can overcome the competi-tion between surplus bile salts and lipase activ-ity. Two other lipases secreted from thepancreas are phospholipase A2, which releasesfatty acid from phospholipid, and cholesterolesterase (also called bile-salt-stimulated lipase),which hydrolyses cholesterol esters and fat-soluble vitamin esters.

Fat absorption

The products of fat hydrolysis, i.e. fatty acidsand mono- or diglycerides together with freecholesterol, phospholipids and fat-solublevitamins, are carried in mixed micelles throughthe unstirred water layer to the apicalmembrane of enterocytes. The absorption oflong-chain fatty acids and monoglyceridesacross the cell membrane to the enterocyte iscompleted by receptor uptake and also by adiffusion gradient. The receptor is called fattyacid binding protein, and it is located in thecell membrane. In the enterocyte, the long-chain fatty acids are transported protein-boundto the endoplasmic reticulum, where, followingre-esterification, they are incorporated intochylomicrons by microsomal transfer protein(MTP). Short- and medium-chain fatty acidscan be absorbed readily into enterocytes evenwithout micellar transport, and they can betransported to the portal system withoutassembly to chylomicrons. The absorption oftriglycerides (and probably also that ofphospholipids) is very effective so that over90% of triglycerides are absorbed.

PRINCIPLES OF CHOLESTEROLABSORPTION

Hydrolysis of dietary cholesterol

Of dietary cholesterol roughly 20% is esteri-fied. As described above, dietary lipids, includ-ing cholesterol and plant sterols, are alreadymixed in the mouth to a crude emulsion andfurther processed in the stomach whichcontains lingual and especially gastric lipases.Cell-membrane cholesterol from food, e.g.from animal or fish meat, should be releasedfrom the membranes before emulsification.Pancreatic juice contains several enzymesincluding cholesterol esterase (E.C. 3.1.1.13),which hydrolyses the small amount of dietarycholesterol esters. However, hydrolysis doesnot seem to be complete in the small intestine,because in colectomy patients, small amountsof ester cholesterol are found in ileal excreta.15

Hydrolysis of esterified cholesterol seems to benecessary for cholesterol absorption in experi-mental animals.16 Preliminary studies in manhave shown that a cholesterol esteraseinhibitor did not change the percentage ofcholesterol absorption or serum cholesterollevel;17 thus, cholesterol esterase does notappear to be a cholesterol transporter intoenterocytes in man.

Micelle formation with biliary lipids

Bile tranports from 800 mg to 1200 mg ofcholesterol per day mainly in micellar forminto the intestinal lumen, indicating that thedaily cholesterol load for intestinal absorptioncan be up to1700 mg.18 During intestinal diges-tion, bile acids, free fatty acids, monoglyc-erides, phospholipids and lysophospholipidsform with cholesterol mixed micelles whichcan easily dissolve dietary water-insolublecompounds, including fat-soluble vitamins andsterols. Dietary cholesterol is incorporated intothese micelles and absorbed, apparently inincreasing proportion to bile cholesterol, whenmoved more distally in the intestine. Theabsorption rate of cholesterol ester may beslower, depending on its hydrolysis. Inter-estingly, the amount of cholesterol ester may

OVERVIEW OF INTESTINAL LIPID METABOLISM 57

ch05 14/7/05 4:45 pm Page 57

actually increase in the duodenum, suggestingthat under some conditions, e.g. in thepresence of other sterol esters and fat, hydroly-sis of dietary cholesterol esters may be inhib-ited, and cholesterol esterase could even formcholesterol esters retarding cholesterol absorp-tion.15,19 Bile cholesterol may be absorbedpreferentially in the upper part of the smallintestine owing to its micelle form in bile, andits absorption decreases when it moves to amore caudal direction, while absorption ofdietary cholesterol could increase progressivelywhen passing through the upper small intesti-nal lumen. Transfer of cholesterol to the micel-lar phase depends on many factors, includingthe amount of bile acids (critical micellarconcentration), the amount and type ofphospholipids and hydrolysis of triglycerides. Alarge oil phase tends to trap cholesterol andcompetes in general for micellar solubility ofsterols. Thus, pancreatic insufficiency, use of alipase inhibitor or olestra20 and the presence ofother sterols, mainly plant sterols and stanolesters in large amounts,21 can retard transfer ofcholesterol to micelles. The role of plantsterols/stanols will be dealt with later. Also,increased secretion of bile cholesterol, as inobesity or in type 2 diabetes, could supersatu-rate bile and prevent the transfer of dietarycholesterol to the micellar phase. In this case,the fractional absorption of dietary cholesterolis lowered despite normal or even increasedtotal cholesterol absorption. In fact, in obesitythe percentage absorption of dietary choles-terol is negatively, but not significantly, relatedto intestinal cholesterol flux, whereas it issignificantly negatively related to biliary choles-terol concentration.22 Accordingly, supersatu-rated bile cannot dissolve dietary cholesteroluntil ‘extra’ cholesterol is absorbed in theupper small intestine from the micellar phase.In general, cholesterol not entering the mi-cellar phase is not absorbed and is finallyexcreted in the colon. Colonic bacterial actionsaturates the ∆5 double bond of cholesterol tothe 5β position, causing formation ofcoprostanol and 3-ketocoprostanone, in whichform up to 85% of cholesterol is eliminatedfrom the body.

Cholesterol transfer to enterocytes

Micellar solubilization of cholesterol in theupper duodenum is considered to be essentialfor its transfer to the enterocyte through theunstirred water layer to the intestinal brush-border of the enterocyte membrane.13

Apparently micellar structure is disaggregatedin the membrane and cholesterol is taken upby the enterocyte through mechanism(s) notcurrently understood. The extracellular vsintracellular concentration gradient of freecholesterol may partly facilitate its enterocyteentry, even though current opinions considerthat several membrane-located proteins,cholesterol transporters, are responsible forthis transfer. Adenosine-triphosphate-bindingcassette (ABC) transporters, ABCA1 and ABCG5/G8, appear to regulate the intracellular vsextracellular transport of cholesterol, and thescavenger receptor class B1 type (SR-B1) andanother scavenger receptor, CD36, mayenhance cholesterol uptake. A new candidatereceptor protein has recently been intro-duced.9 Experiments with gene-modifiedanimals have raised questions about the possi-ble role of ABCA1, SR-B1 and CD36 as specificcholesterol transporters in enterocytes.23,24

Mutation of ABC G5/G8 genes, half trans-porters of sterols (sterol pumps) in enterocytesand hepatocytes, results in sitosterolaemiabecause of enhanced sterol uptake by entero-cytes and decreased secretion by hepatocytes.7

Ezetimibe, a specific inhibitor of cholesterolabsorption, seems to inhibit cholesterol andplant sterol absorption25 in sitosterolaemicpatients, also lowering their serum LDLcholesterol and plant sterol levels.26 In micewith high cholesterol absorption, an ezetimibederivative reduces ABC G5/G8 expression inenterocytes for animals on both low- and high-cholesterol diets. The findings have beeninterpreted to indicate that cholesterol andplant sterols are taken up by the enterocytethrough a common permease blocked byezetimibe independently of the ABC-trans-porter family.13 The amount of intracellularsterols would then regulate expression of theABC G5/G8 proteins.


ch05 14/7/05 4:45 pm Page 58

After absorption, cholesterol is partially ester-ified by acyl CoA:cholesterol acyltransferase-2(ACAT 2) in the enterocyte. Esterificationreduces free cholesterol and changes the intra-cellular vs extracellular gradient to one morefavourable for additional absorption. Esterifiedcholesterol is incorporated into chylomicronparticles, formation of which is facilitated byMTP, apoprotein B48 and triglycerides. Therole of ACAT-2 and MTP in cholesterol absorp-tion in man seems to be limited, and inhibitionof the latter may result in side-effects.27

Chylomicrons, or remnants thereof, arefinally taken up by the liver. It has been calcu-lated that about 1.2–1.6 g of cholesterol(200–500 mg from diet and 900–1100 mg frombile and intestinal mucosa) enters the intestinallumen daily. Since the mean absorption ofdietary cholesterol, and probably also of bilecholesterol, is about 50% in a random malepopulation,28 and in other populations,25,29 up to1 g of intestinal cholesterol can enter the liverdaily. Thus, inhibition of cholesterol absorption,e.g. by plant stanols/sterols or ezetimibe, caneasily affect the hepatic cholesterol balance, alsochanging lipoprotein and cholesterol synthesis,with a final lowering of LDL-cholesterol.

INTESTINAL METABOLISM OFPLANT STEROLS/STANOLS ANDTHEIR CONTRIBUTION TOCHOLESTEROL ABSORPTION

Dietary plant sterols and stanols

Our daily food contains about 250–350 mg/dayof plant sterols, consisting of a large number ofphytosterols (over 200), mainly sitostosterol,campesterol, stigmasterol and avenasterol (forexample, see references 10–12). Their chemicalstructure differs from that of cholesterol mainlyby the presence of either a methyl or ethylgroup, with or without an additional doublebond in the side chain. In vegetarian food theamounts can be higher, because any dietaryplant material also contains plant sterols. Sincesmall amounts of plant sterols are also present inmeat and fish, it is actually difficult to have a dietfree from plant sterols. In fact, like cholesterol

in the mammalian organism, plant sterols arenecessary for normal cellular function, beingstructural components of plant cell membranes.The highest amounts of plant sterols areobtained from vegetable oils and oil products,especially from corn and rapeseed oils, howevercereal grains, cereal products and nuts are alsorich in these sterols. The ∆5 double bond ofplant sterols can be saturated in nature to formthe respective plant stanols, of which sitostanoland campestanol are the most important. About10% of dietary plant sterols are plant stanolsoriginating from, e.g., cereals, especially rye andwheat, corn and some corn products. Dietaryplant sterols can be free, esterified at 3β-OH toa fatty acid or hydroxycinnamic acid or thehydroxyl group can be glycosylated with ahexose or a fatty acid–hexose. Thus, the conju-gated forms of plant sterols are more compli-cated than that of mammalian cholesterol,which is only esterified with different fatty acids.

Hydrolysis of dietary sterols

Dietary phytosterols lower serum cholesterol byinhibiting intestinal cholesterol absorption. Itis generally considered that they compete withcholesterol for incorporation into mixedmicelles, resulting in reduced transfer ofcholesterol into enterocytes.30 However,micelle solubility requires free sterols, indicat-ing that dietary natural phytosterol conjugatesor phytosterol esters of functional foods shouldfirst be hydrolysed. Free plant sterols are appar-ently poorly incorporated into micelles,indicating that they should be presented in fat-soluble form or esterified, but even then thefood matrix should contain some fat.31–35 Asnoted above, dietary plant sterols, togetherwith food fat and cholesterol, form a crudeemulsion in the mouth and stomach, and afiner emulsion in the upper intestine.However, owing to rough plant material,sterols incorporated into cells in constrainedplant tissues may not be released as easily asmammalian cholesterol to emulsion droplets,retarding hydrolysis of several types of conju-gated plant sterols. On the other hand, most ofthese conjugates, especially plant sterol esters,


ch05 14/7/05 4:45 pm Page 59

appear to be already hydrolysed within a shortsegment of the upper small intestine, eventhough hydrolysis of glycosylated sterols is lessknown. Hydrolysis of plant sterols from fineemulsions favours transfer of free sterols tomicelles. Intubation studies have shown15,19

that, at low plant sterol and stanol concentra-tions, with high respective ester percentages ofinfused sterols, the amounts of ester choles-terol, cholestanol and their esterificationpercentages increased and those of plantsterols and stanols decreased in the uppersmall intestine when the infusate had passed60 cm more distally in the duodenum.Hydrolysis of plant sterols was 60–80% and thatof plant stanols about 55%. After a largeamount of plant stanol esters had been addedto the infusate, including a small amount ofplant sterol esters (ester percentage of non-cholesterol sterols 56–96%), the amount ofjejunal ester cholesterol was further increased,but hydrolysis of the respective plant sterolsand stanols, including cholestanol, was about40%. This finding suggests that, at low esterpercentages, large amounts of free sterols inthe intestinal contents might even increaseester formation, while markedly increasedsterol esters, shown in fact after plant stanolester infusion, could overload the hydrolysisprocess and even increase the amount of estersof the sterols with a low baseline ester percent-age like cholesterol. The increase in choles-terol esters may explain why plant stanol esters,perhaps also plant sterol esters, retard choles-terol absorption. The increase in cholesterolesters may be temporary, indicating that, in themore distal intestine, hydrolysis may proceedafter further absorption of free cholesterol. Infact, virtually no increase was seen in choles-terol esters, and only a small increase insitostanol esters of ileal excreta of colectomypatients after consumption of large amounts ofplant stanol esters.

Role of plant sterols in micelleformation

Infusion of a vegetable oil/egg yolk/glycerol/mono-oleate/water mixture (experiments

shown above15) in the upper duodenumshowed that campesterol, sitosterol and therespective stanols after hydrolysis were incor-porated up to 70% into mixed micelles withfree cholesterol and cholestanol (70–80%).Addition of a large amount of stanol esters(about 150-fold) with slightly increased plantsterol esters to the infusion mixture revealed,as shown above, a rapid hydrolysis. Subse-quently, free intestinal plant stanols were trans-ferred into the mixed micellar phase, such thatthe absolute concentration of campestanol andsitostanol were increased 100–200 times.Respective infusate increased unsaturatedplant sterols only by up to 45%, while that ofcholesterol decreased by one-third and alsothat of cholestanol tended to decrease in themicellar phase. No precipitation of sterols wasfound when the intestinal contents moved60 cm more distally in the upper small intes-tine. The small quantities of plant sterols in the infusion material indicated that largeramounts might have similar effects on theirmicellar solubility with plant stanols, but nodata in man are available.

Plant sterol/stanol absorption

Micelle transfer through the enterocytemembrane by permease would allow furthermetabolism of sterols. However, recent devel-opments in the understanding of plant sterolabsorption in sitosterolaemia have indicatedthat the sterol pump proteins, expression ofwhich is regulated by the ABC G5/G8 genes,enhance the output of absorbed plant sterolsand cholesterol from the enterocyte back tothe intestinal lumen (see reference 7).Accordingly, specificity of this pumping outmay limit plant sterol absorption to less thanone-fifth of that for cholesterol, to less than10% vs about 50% for cholesterol. However,even with the markedly high micelle concen-tration of plant stanols, as in the perfusionstudy15 cited above, their absorption is low,<1%, suggesting that their transfer to entero-cytes and further to blood and tissues islimited. Thus, consumption of plant stanolesters increases serum plant stanol


ch05 14/7/05 4:45 pm Page 60

concentrations only slightly.36,37 A high dietaryplant sterol intake, on the other hand, mostlikely increases plant sterols in micelles andsubsequently enhances their bypass of thesterol pump in the enterocytes, resultingultimately in increased serum plant sterollevels. The fact that serum levels of bothcholestanol and plant stanols are alsoincreased in sitosterolaemia, and that some ofthe dietary plant stanols are even absorbednormally, indicate that their absorption is alsoregulated by the sterol pump system.Competition of plant stanols with cholesterol,cholestanol and plant sterols for intestinalabsorption includes enhanced ester formationand reduced micellar solubility of cholesterol,some reduction of cholestanol concentrationbut only a relative reduction of micellar plant

sterols in the total intestinal sterols. The speci-ficity of sterol output from enterocytes back tothe intestinal lumen may effectively regulatesterol absorption according to the size of sterolmolecules and their saturation.

Experiments with caco-2 cells in vitro haveshown that mixed micelles with sitostanol orcholesterol plus sitostanol induced expressionof ABCA1. Thus, increased enterocyte plantstanols, and possibly also plant sterols, mightincrease the ABCA1-mediated sterol effluxback into the intestinal lumen and the LXRpathway may regulate intestinal lipid metabo-lism.38 Stimulation of ABCA1 gene expressionby plant sterols may explain the finding thatone daily dose of plant stanols or sterols in manlowers serum cholesterol by a similar extent asthree doses.34,39


1. Foster GD, Wyatt HR, Hill JO, et al. A randomizedtrial of a low-carbohydrate diet for obesity. New EnglJ Med 2003; 348:2082–90

2. Miettinen TA, Puska P, Gylling H, et al. Reduction ofserum cholesterol with sitostanol-ester margarine in amildly hypercholesterolemic population. N Engl JMed 1995; 333:1308–12

3. Weststrate JA, Meijer GW. Plant sterol-enrichedmargarines and reduction of plasma total- and LDL-cholesterol concentrations in normocholesterolaemicand mildly hypercholesterolaemic subjects. Eur J ClinNutr 1998; 52:334–43

4. Heart Protection Study Collaborative Group.MRC/BHF Heart Protection Study of cholesterollowering with simvastatin in 20536 high-risk individu-als: a randomised placebo-controlled trial. Lancet2002; 361:2005–16

5. Miettinen TA, Gylling H. Plant stanol and sterol estersin prevention of cardiovascular disease. Ann Med2004; 36(2):126–34

6. Katan MB, Grundy SM, Jones P, et al. for the StresaWorkshop Participants. Efficacy and safety of plantstanols and sterols in the management of bloodcholesterol levels. Mayo Clin Proc 2003; 78:965–78

7. Berge KE. Sitosterolemia: a gateway to new knowl-edge about cholesterol metabolism. Ann Med 2003;35: 502–11

8. Darkes MJM, Poole RM, Goa KL. Ezetimibe. Am JCardiovasc Drugs 2003; 3:67–76

9. Altmann SW, Davis Jr HR, Zhu L-j, et al. Niemann-

Pick C1 like 1 protein is critical for intestinal choles-terol absorption. Science 2004; 303:1201–4

10. Piironen V, Lindsay DG, Miettinen TA, et al. Plantsterols: biosynthesis, biological function and theirimportance to human nutrition. J Sci Food Agric2000; 80:939–66

11. Moreau RA, Whitaker BD, Hicks KB. Phytosterols,phytostanols, and their conjugates in foods: structuraldiversity, quantitative analysis, and health-promotinguses. Progr Lipid Res 2002; 41:457–500

12. Ostlund RE Jr. Phytosterols in human nutrition.Annu Rev Nutr 2002; 22:533–49

13. Turley SD, Dietschy JM. Sterol absorption by the smallintestine. Curr Opin Lipidol 2003; 14:233–40

14. Hamosh M. Preduodenal fat digestion. In ChristopheAB, de Vriese S, eds. Fat Digestion and Absorption.Champaign, Ill: AOCS Press, 2000:1–12

15. Nissinen M, Gylling H, Vuoristo M, et al. Micellardistribution of cholesterol and phytosterols afterduodenal plant sterol ester infusion. Am J Physiol2002; 282:G1009–15

16. Howles PN, Carter CP, Hui DY. Dietary free and ester-ified cholesterol absorption in cholesterol esterase(bile salt-stimulated lipase) gene-targeted mice. J BiolChem 1996; 271:7196–202

17. Bosner MS, Wolff AA, Ostlund RE Jr. Lack of effect ofcholesterol esterase inhibitor CVT-1 on cholesterolabsorption and LDL cholesterol in humans.Cardiovasc Drugs Ther 1999; 13:449–54

18. Mitchell JC, Stone BG, Logan GM, et al. Role of

References

ch05 14/7/05 4:45 pm Page 61

cholesterol synthesis in regulation of bile acid synthe-sis and biliary cholesterol secretion in humans. J LipidRes 1991; 32:1143–9

19. Miettinen TA, Siurala M. Bile salts, sterols, sterolesters, glycerides and fatty acids in micellar and oilphases of intestinal contents during fat digestion inman. Z Klin Chem Klin Biochem 1971; 9:47–52

20. Mittendorfer B, Ostlund RE Jr, Patterson BW, et al.Orlistat inhibits dietary cholesterol absorption. ObesRes 2001; 9:599–604

21. Miettinen TA. Cholesterol absorption inhibition: astrategy for cholesterol lowering therapy. Int J ClinPract 2001; 55:599–604

22. Miettinen TA, Gylling H. Cholesterol absorptionefficiency and sterol metabolism in obesity.Atherosclerosis 2000; 153:241–8

23. Dropnik W, Lindenthal B, Lieser B, et al. ATP-binding cassette transporter A 1 (ABCA1) affects totalbody cholesterol. Gastroenterology 2001;120:1203–11

24. Mardones P, Quinones V, Amigo L, et al. Hepaticcholesterol and bile acid metabolism and intestinalcholesterol absorption in scavenger receptor class Btype I-deficient mice. J Lipid Res 2001; 42:170–80

25. Sudhop T, Lütjohann D, Kodal A, et al. Inhibition ofintestinal cholesterol absorption by ezetimibe inhumans. Circulation 2002; 106:1943–8

26. Salen G, von Bergmann K, Kwiterovich P, et al.Ezetimibe is an effective treatment for homozygoussitosterolemia. Circulation 2002; 106 (Suppl): 929

27. Sudhop T, von Bergmann K. Cholesterol absorptioninhibitors for the treatment of hyperlipidemia. Drugs2002; 62:2333–47

28. Miettinen TA, Kesäniemi YA. Cholesterol absorption:regulation of cholesterol synthesis and eliminationand within-population variations of serum cholesterollevels. Am J Clin Nutr 1989; 49:629–35

29. Bosner MS, Lange LG, Stenson WE, et al. Percentcholesterol absorption in normal women and menquantified with dual stable isotopic tracers andnegative ion mass spectrometry. J Lipid Res 1999;40:302–8

30. Ikeda I, Tanabe Y, Sugano M. Effects of sitosterol and

sitostanol on micellar solubility of cholesterol. J NutrSci Vitaminol 1989; 35:361–9

31. Denke MA. Lack of efficacy with low-dose sitostanoltherapy as an adjunct to a cholesterol-lowering diet inmen with moderate hypercholesterolemia. Am J ClinNutr 1995; 61:392–6

32. Plat J, van Onselen ENM, van Heugten MMA, et al.Effects on serum lipids, lipoproteins and fat-solubleantioxidant concentrations of consumptionfrequency of margarines and shortenings enrichedwith plant stanol esters. Eur J Clin Nutr 2000; 54:671–7

33. Nestel P, Cehun M, Pomeroy S, et al. Cholesterol-lowering effects of plant sterol esters and non-esteri-fied plant stanols in margarine, butter and low-fatfoods. Eur J Clin Nutr 2001; 55:1084–90

34. Mensink RP, Ebbing S, Lindhout M, et al. Effects ofplant stanol ester supplied in low-fat yoghurt onserum lipids and lipoproteins, non-cholesterol sterolsand fat soluble antioxidant concentrations.Atherosclerosis 2002; 160:205–13

35. Jones PJH, Vanstone CA, Raeini-Sarjaz M, et al.Phytosterols in low- and nonfat beverages as part of acontrolled diet fail to lower lipid levels. J Lipid Res2003; 44:1713–9

36. Miettinen TA, Vuoristo M, Nissinen M, et al. Serum,biliary and fecal cholesterol and plant sterols in colec-tomized patients before and during consumption ofstanol ester margarine. Am J Clin Nutr 2000;71:1095–102

37. Ostlund RE Jr, McGill JB, Zeng C-M, et al.Gastrointestinal absorption and plasma kinetics of soydelta(5)-phytosterols and phytostanols in humans.Am J Physiol Endocrinol Metab 2002; 282:E911–6

38. Plat J, Mensink RP. Increased intestinal ABCA1expression contributes to the decrease in cholesterolabsorption after plant stanol consumption. FASEB J2002; 16:1248–53

39. Matvienco OA, Lewish DE, Swanson M, et al. A singledaily dose of soybean phytosterols in ground beefdecreases serum total cholesterol and LDL choles-terol in young, mildly hypercholesterolemic men. AmJ Clin Nutr 2002; 76:57–64


ch05 14/7/05 4:45 pm Page 62

INTRODUCTION

Elevated blood cholesterol levels have beenshown to be one of the major risk factors forthe development of cardiovascular disease.While a great deal is known about how choles-terol is transported in the body via lipopro-teins, very little is known about the molecularmechanisms by which cholesterol exerts itspotential metabolic effects. Additionally, itremains to be shown that cholesterol itselffunctions as a bioactive molecule resultingdirectly in atherosclerosis. Despite these reser-vations, one cannot overlook the importantlink between cholesterol and cardiovasculardisease. This review will focus on the pathwaysof dietary cholesterol and non-cholesterolentry and excretion by the body. In particular,the role of the newly described transporters,ABCG5 and ABCG8, in these processes will behighlighted. Further, we draw attention to thedifference between bioactive molecules,versus biomarkers of disease. Cholesterol isclearly an important biomarker, but how goodis it as a bioactive molecule and if it is not abioactive molecule with respect to the patho-genesis of atherosclerosis, which sterolmolecules are? This chapter will focus only onthe sterol molecules and other chapters inthis book highlight the roles of other potentmolecules.

Sterol homeostasis is a tightly regulatedprocess; mechanisms of dietary sterol absorp-tion, endogenous de novo synthesis and excre-tion and/or breakdown of sterols ensure

balance is maintained. For mammals, choles-terol is the major sterol; it is synthesizedendogenously, is not a dietary necessity andexcess cholesterol is metabolized by the bodyand excreted as biliary cholesterol or bile acids.The use of inhibitors of de novo cholesterolsynthesis, namely 3-hydroxy-3-methyl glutarylcoenzyme A reductase inhibitors or statins, hashad a significant impact on reducing the risk ofcardiovascular events as shown by multipletrials (4S, CARE, LIPID, HPS, etc.).1–4 Basedupon present knowledge, statin efficacy oncardiovascular disease is based upon the reduc-tion in serum cholesterol levels as well aspotential ‘pleiotropic’ effects. Irrespective ofthese mechanisms and despite the availabilityof very potent statins, the risk reduction withappropriate use of statins is only ~20–30% inmany clinical studies.1–4 The major mechanismwhereby statins are likely to exert their effectsis the increased clearance of blood lipoproteinparticles by the liver (upregulation of the LDLreceptors) that results in a reduction of choles-terol as well as any bioactive sterols (oxysterolas well as non-cholesterol sterols) carried inthese particles. Note that, on statin therapy,whole-body sterol balance does not changeand, at the cellular level, the free cholesterolcontent is also unlikely to be reduced, butstored pools of cholesterol (as cholesterolesters) are reduced, especially atheromatousplaques as they regress. The thesis of thistreatise is to explore the hypothesis that thebioactive molecule(s) track with cholesterol,use pathways that cholesterol may use and that

The role of non-cholesterol sterols in thepathogenesis of atherosclerosis and theirmodulation by the sitosterolaemia locusE.L. Klett and S.B. Patel

6

ch06 14/7/05 4:46 pm Page 63

the proteins ABCG5 and ABCG8 may play a keyrole in keeping these molecules from accumu-lating.

A typical Western diet contains ~400 mg ofcholesterol per day, derived from animalsources, and about 200–400 mg of non-choles-terol sterols, derived mostly from plants.5,6 Onaverage, about 55% of the dietary cholesterol isabsorbed and retained on a daily basis, butalmost none of the non-cholesterol sterols(such as the plant sterols – sitosterol, campe-sterol, brassicasterol, etc.) are retained.7

Additionally, cooked foods contribute oxidizedsterols to the dietary intake pool of which themetabolic fate is relatively unknown. Whilemost of the total body pool of cholesterol isderived from de novo cholesterol synthesis, it isnow clear that dietary sources play a crucialrole in maintaining total body sterol balance.Influx of cholesterol from the diet regulatesthe amount of de novo synthesis. The liver is thecentral organ in maintaining this balance.Excess body cholesterol, derived from endoge-nous synthesis or from dietary absorption, isexcreted exclusively by the liver, either bydirect excretion as free cholesterol into bile, orby breakdown to bile acids and excretion asbile acid conjugates into bile. A small amountof dietary non-cholesterol also enters the body.However, the liver rapidly excretes these sterolsinto bile, thus resulting in a very low net dailyabsorption.8 Why have mechanism(s) evolvedthat regulate dietary non-cholesterol (ordietary cholesterol), if these non-cholesterolsterols are thought not to be biologically activein our bodies? Our group and others havehypothesized that the body has evolved veryefficient mechanisms that allow for the separa-tion and/or exclusion of potentially deleteri-ous dietary compounds since within the bulk ofthese dietary sterols there are potent andbioactive sterols or sterol derivatives that havepotentially deleterious effects in our bodies. Itis only when these mechanisms are perturbedthat atherosclerosis develops. This review willfocus on the potential role of two ABC half-transporters, ABCG5 and ABCG8, in thisprocess of abrogating dietary-induced athero-sclerosis.

CLUES FROM A RARE GENETICDISORDER, SITOSTEROLAEMIA

The identification of a rare autosomal reces-sive disorder of sterol absorption/excretiontermed sitosterolaemia (also known as phyto-sterolaemia, MIM 210250), has helped to shedlight on the molecular mechanisms by whichthe intestinal cells traffic sterols and has led tothe identification of the elusive hepatobiliarycholesterol exporter(s).9,10 Patients with thisdisease can present with tendon xanthomas(usually involving the Achilles tendon),haemolytic episodes, arthritis/arthralgias and,most strikingly, accelerated atherosclerosis.The key feature of this disease is the disrup-tion of the normal pathways that prevent theaccumulation of non-cholesterol sterols, suchas plant sterols.9,11,12 The gene(s) causing sito-sterolaemia was mapped to the STSL locus onhuman chromosome 2p21.13–15 Elucidation ofthe molecular defects of sitosterolaemiashowed that the STSL locus comprises twogenes, ABCG5 and ABCG8, encoding sterolin-1and sterolin-2, respectively, and mutation ineither one of these genes is sufficient to causesitosterolaemia.13 In the absence of function-ing sterolin-1 (ABCG5) or sterolin-2 (ABCG8),patients hyper-absorb all non-cholesterolsterols (and cholesterol), retain these in thebody and fail to excrete sterols into bile.8,16–19

Based on this it would be expected that thesepatients would have tremendously high serumcholesterols, but paradoxically they havenormal to only slightly elevated serum choles-terol.11 Since patients with sitosterolaemiahave accelerated atherosclerosis and since thedefective genes alter non-cholesterol sterollevels (all of which are of dietary origin), thisraises the possibility that ABCG5/ABCG8evolved to keep bioactive non-cholesterolsterols out of the body. The STSL locus ishighly conserved and these genes are found inanimals as disparate as fish and mammals, andsince one important role for these proteins isthe biliary excretion of sterols, the possibilitythat these proteins may also be responsible forclearance of bioactive sterols from the bodyhas also been raised.


ch06 14/7/05 4:46 pm Page 64

Genetics of ABCG5 and ABCG8

Genetic analyses of sitosterolaemia pedigreesallowed the mapping of the STSL locus tohuman chromosome 2p21, between D2S2294and D2S2298.20,21 By using positional cloningprocedures or by screening for genes inducedby exposure to LXR agonists, two groupsidentified not one but two genes, ABCG5 andABCG8, that encode the proteins sterolin-1 andsterolin-2, mutations of which cause sitostero-laemia.14,15 To date, no patient with sitostero-laemia has been identified with mutations inboth genes. ABCG5 and ABCG8 are organizedin a head-to-head configuration with the twoinitiator ATGs of the two genes separated byonly 340 bp.13 The bi-directional promoter forthe two genes has not been definitivelydefined, but ~140 bp of DNA separates the twotranscription start sites. Each gene is made upof 13 exons and 13 introns. Although LXRagonists have been used to stimulate themRNA expression from the STSL locus, nomotifs suggestive of LXR recognitionsequences in the intergenic region have beenidentified.22 Recently, regulation of this locusthrough LRH has been reported and suggeststhat LXR modulation of this locus may beindirect.23

The STSL locus is polymorphic but, unusu-ally, while ABCG8 is highly polymorphic,ABCG5 seems almost invariant in humans.13

Despite the closeness of ABCG5 and ABCG8,the increased polymorphic frequency ofABCG8 or the relative ‘conservation’ of ABCG5suggests that there may be a biological pressurefor these genetic changes. Many of thepolymorphic changes detected at the STSLlocus are in linkage disequilibrium, suggestingmany of these changes are relatively new.

Since STSL locus defects cause sitostero-laemia, the heritability of plasma plant sterollevels has been examined as a prelude towhether natural variation at this locus is astrong determinant of these levels. In bothnuclear families, as well as twin studies, plantsterols are highly heritable.24 Preliminaryevidence suggests an association with certainpolymorphic variations and plant sterol levels,

but a study to formally examine this has notbeen performed.

ARE NON-CHOLESTEROL STEROLSINVOLVED IN ATHEROGENESIS?

Although Bhattacharyya and Connor were thefirst to propose that sitosterol may be a factorin causing atherosclerosis,9 Glueck andcolleagues were the first to show that plasmaplant sterols are present in atheromatousplaques and subsequently showed that theseare elevated in patients with proven heartdisease and may be genetically determined.25,26

Miettinen and colleagues,27 as well as vonBergmann and colleagues,28 have also reportedcase-controlled studies showing the associationof plant sterols with proven coronary heartdisease. Data from the PROCAM study,presented at the American Heart AssociationScientific Sessions 2003, also suggests plasmasitosterol levels as independent predictors offuture CVD.29 More recently, associationstudies between the responsiveness to statindrugs and genetic variations at the STSL locushave been reported, although it should bepointed out that this locus shows significantlinkage disequilibrium amongst the multiplepolymorphisms thus-far identified.30–32

Oxysterols are products of cholesterol oxida-tion and have been implicated in the initiationand/or development of atherosclerosis.33,34

However, there is no direct evidence inhumans that oxysterols contribute to the devel-opment of atherosclerosis, despite observa-tional reports such as detection ofnon-enzymatically derived oxysterols in athero-sclerotic plaques (such as 7-ketocholesterol) orlimited epidemiological studies.35–39 A possibleconnection between cardiovascular disease anddiet may be the dietary-derived oxidativelymodified sterols. The identification of nuclearreceptors LXRα and LXRβ, which bind withhigh affinity to oxidized sterols, as regulators ofcholesterol homeostasis has implicatedoxysterols as important modulators of thesepathways. To date, the endogenous physiologi-cal ligand for LXR has not been identified,

THE ROLE OF NON-CHOLESTEROL STEROLS IN THE PATHOGENESIS OF ATHEROSCLEROSIS 65

ch06 14/7/05 4:46 pm Page 65

although a number of potential candidateshave been proposed, a major one being 27-OHcholesterol. LXRα strongly modulates thetranscription of several genes involved in themetabolism and transport of sterols includingCyp7a1, SREBP-1c, CETP, ApoE, ABCA1, ABCG1and ABCG5/ABCG8.40,41 Two pools of oxysterolsexist, those derived endogenously and thosethat arise from the diet. Dietary oxysterols areproduced as a result of heating in air (cooking)or prolonged storage of foods.33 Absorption ofdietary oxysterols, as determined by lymphaticcannulation, has been shown in rats to rangefrom 6 to 30%.42,43 Once dietary-derivedoxysterols enter the circulation their fate is notknown. Given their capacity to alter geneexpression and their direct toxic effects on thevasculature,34 it is likely that oxysterols mayutilize the same cholesterol transport andclearance mechanisms for rapid removal, andit remains to be seen whether ABCG5/ABCG8play a major role. Current dogma suggests thatmost oxysterols are rapidly taken up by theliver and metabolized via the bile acid synthesispathways, and some may be excreted directlyinto bile. Given the high biological activity andrelatively low levels, it is apparent from a teleo-logical perspective that organisms have evolvedcertain protective mechanisms to specificallyexclude and/or excrete these potentiallydeleterious compounds.

DO ABCG5/ABCG8 PLAY A CRUCIALROLE IN CARDIOVASCULAR RISKAND ARE THESE PROTEINS ORTHEIR ACTIVITIES ‘TARGETABLE’FOR THERAPEUTIC INTERVENTION?

At present, there are no data that directly linkthe STSL locus and a predisposition to athero-matous cardiovascular disease. Interestingly,statin responsiveness in humans has beenlinked in some preliminary analyses, but thebest study to date has not reproduced thesefindings. There are lines of investigation thatare indirect. For example, sitosterolaemic ratsprone to hypertension (SHRSP) showincreased morbidity and mortality when fed

diets rich in plant sterols (although dietarystudies are known for not controlling for thedietary constituents).44 Feeding high quantitiesof plant sterols to such hypertensive rats alsoraised their blood pressure even higher. And,as stated above, plant sterol levels in the plasmaseem to act as a biomarker of predisposition tocardiovascular disease. We have proposed amodel whereby the dietary entry of sterols(cholesterol and non-cholesterol sterols) maybe initially regulated by the activity of NPC1L1and subsequently by activity of ABCG5/ABCG8in the intestine.45–47 On the other hand, therates of sterol excretion via the hepatobiliarysystem may be primarily determined byABCG5/ABCG8 activity. Currently, we have nofurther biochemical characterization of howABCG5/ABCG8 act to ‘pump’ sterols out. It isnot clear whether their activity is increased byany allosteric interactions and thus directmodulation by small molecules seems unlikely.In contrast, transcriptional activation by thera-peutic ligands seems feasible. To date, only twotranscriptional factors are known, LXR andLRH, and both of these have a multitude ofeffects and a specific targeted ABCG5/ABCG8activation does not seem to be possible at thistime.

CONCLUSIONS

The identification of the molecular defectsunderlying sitosterolaemia has affordedconsiderable insight into the molecular mecha-nisms by which dietary cholesterol is absorbed,how non-cholesterol sterols may be preventedfrom entry and how sterols are excreted intothe biliary system. The renewed interest in non-cholesterol sterols has also led to a renewedinterest in defining the bioactive sterol orsterol metabolite that is likely to play a key rolein atherogenic cardiovascular disease and mayfinally lead to a molecular basis for the connec-tion between diet and heart disease. Currently,the lack of biochemical characterization of howABCG5/ABCG8 may act to pump sterols out,the lack of clear understanding of how theSTSL locus is regulated and the fact that any


ch06 14/7/05 4:46 pm Page 66

therapeutic intervention will need to activate,rather than inactivate, this pathway suggeststhat these proteins are unlikely to be‘targetable’. However, the continued investiga-tion of these pathways has shed considerablelight onto the role of sterol trafficking and this,

in itself, has already led to significant advance-ment in our understanding of these physiolog-ical processes. Perhaps this insight may allow usto better identify the bioactive sterols and moveaway from the biomarkers, with newer thera-pies targeted at these molecules.

THE ROLE OF NON-CHOLESTEROL STEROLS IN THE PATHOGENESIS OF ATHEROSCLEROSIS 67

1. Randomised trial of cholesterol lowering in 4444patients with coronary heart disease: theScandinavian Simvastatin Survival Study (4S). Lancet1994; 344:1383–9

2. Prevention of cardiovascular events and death withpravastatin in patients with coronary heart diseaseand a broad range of initial cholesterol levels. TheLong-Term Intervention with Pravastatin inIschaemic Disease (LIPID) Study Group. N Engl JMed 1998; 339:1349–57

3. Sacks FM, Pfeffer MA, Moye LA, et al. The effect ofpravastatin on coronary events after myocardialinfarction in patients with average cholesterol levels.Cholesterol and Recurrent Events Trial investigators.N Engl J Med 1996; 335:1001–9

4. Shepherd J, Cobbe SM, Ford I, et al. Prevention ofcoronary heart disease with pravastatin in men withhypercholesterolemia. West of Scotland CoronaryPrevention Study Group. N Engl J Med 1995;333:1301–7

5. Weihrauch JL, Gardner JM. Sterol content of foods ofplant origin. J Am Diet Assoc 1978; 73:39–47

6. Nair PP, Turjman N, Kessie G, et al. Diet, nutritionintake, and metabolism in populations at high andlow risk for colon cancer. Dietary cholesterol, beta-sitosterol, and stigmasterol. Am J Clin Nutr 1984; 40(4Suppl):927–30

7. Wilson MD, Rudel LL. Review of cholesterol absorp-tion with emphasis on dietary and biliary cholesterol.J Lipid Res 1994; 35:943–55

8. Salen G, Tint GS, Shefer S, et al. Increased sitosterolabsorption is offset by rapid elimination to preventaccumulation in heterozygotes with sitosterolemia.Arterioscler Thromb 1992; 12:563–8

9. Bhattacharyya AK, Connor WE. Beta-sitosterolemiaand xanthomatosis. A newly described lipid storagedisease in two sisters. J Clin Invest 1974; 53:1033–43

10. Rao MK, Perkins EG, Connor WE, et al. Identificationof beta-sitosterol, campesterol, and stigmasterol inhuman serum. Lipids 1975; 10:566–8

11. Salen G, Shefer S, Nguyen L, et al. Sitosterolemia. JLipid Res 1992; 33:945–55

12. Salen G, Patel SB, Batta AK. Sitosterolemia.Cardiovasc Drug Rev 2002; 20:255–70

13. Lu K, Lee M-H, Hazard S, et al. Two genes that mapto the STSL locus cause sitosterolemia: genomic struc-ture and spectrum of mutations involving sterolin-1and sterolin-2, encoded by ABCG5 and ABCG8 respec-tively. Am J Hum Genet 2001; 69: 278–90

14. Berge KE, Tian H, Graf GA, et al. Accumulation ofdietary cholesterol in sitosterolemia caused bymutations in adjacent ABC transporters. Science2000; 290:1771–5

15. Lee M-H, Lu K, Hazard S, et al. Identification of agene, ABCG5, important in the regulation of dietarycholesterol absorption. Nat Genet 2001; 27:79–83

16. Miettinen TA. Phytosterolaemia, xanthomatosis andpremature atherosclerotic arterial disease: a case withhigh plant sterol absorption, impaired sterol elimina-tion and low cholesterol synthesis. Eur J Clin Invest1980; 10:27–35

17. Bhattacharyya AK, Connor WE, Lin DS, et al. Sluggishsitosterol turnover and hepatic failure to excretesitosterol into bile cause expansion of body pool ofsitosterol in patients with sitosterolemia and xantho-matosis. Arterioscler Thromb 1991; 11:1287–94

18. Gregg RE, Connor WE, Lin DS, et al. Abnormalmetabolism of shellfish sterols in a patient with sitos-terolemia and xanthomatosis. J Clin Invest 1986;77:1864–72

19. Salen G, Shore V, Tint GS, et al. Increased sitosterolabsorption, decreased removal, and expanded bodypools compensate for reduced cholesterol synthesisin sitosterolemia with xanthomatosis. J Lipid Res1989; 30:1319–30

20. Lee M-H, Gordon D, Ott J, et al. Fine mapping of agene responsible for regulating dietary cholesterolabsorption; founder effects underlie cases of phytos-terolemia in multiple communities. Eur J Hum Genet2001; 9:375–84

21. Lu K, Lee M-H, Carpten JD, et al. High-resolutionphysical and transcript map of human chromosome2p21 containing the sitosterolemia locus. Eur J HumGenet 2001; 9:364–74

22. Remaley AT, Bark S, Walts AD, et al. Comparativegenome analysis of potential regulatory elements inthe ABCG5–ABCG8 gene cluster. Biochem BiophysRes Commun 2002; 295:276–82

References

ch06 14/7/05 4:46 pm Page 67

23. Freeman LA, Kennedy A, Wu J, et al. The orphannuclear receptor LRH-1 activates the ABCG5/ABCG8intergenic promoter. J Lipid Res 2004; 45:1197–206

24. Berge KE, von Bergmann K, Lutjohann D, et al.Heritability of plasma noncholesterol sterols andrelationship to DNA sequence polymorphism inABCG5 and ABCG8. J Lipid Res 2002; 43:486–94

25. Mellies MJ, Ishikawa TT, Glueck CJ, et al. Phytosterolsin aortic tissue in adults and infants. J Lab Clin Med1976; 88:914–21

26. Glueck CJ, Speirs J, Tracy T, et al. Relationships ofserum plant sterols (phytosterols) and cholesterol in595 hypercholesterolemic subjects, and familialaggregation of phytosterols, cholesterol, and prema-ture coronary heart disease in hyperphytosterolemicprobands and their first-degree relatives. Metabolism1991; 40:842–8

27. Rajaratnam RA, Gylling H, Miettinen TA. Inde-pendent association of serum squalene and noncho-lesterol sterols with coronary artery disease inpostmenopausal women. J Am Coll Cardiol 2000;35:1185–91

28. Sudhop T, Gottwald BM, von Bergmann K. Serumplant sterols as a potential risk factor for coronaryheart disease. Metabolism 2002; 51:1519–21

29. Assman G, Cullen P, Erbey JR, et al. Elevation in plasmasitosterol concentration is associated with an increasedrisk for coronary events in the PROCAM study.American Heart Association Scientific Sessions 2003,Abstract 3300. Circulation 2003; 108(Suppl 4):730

30. Miettinen TA, Gylling H, Lindbohm N, et al. Serumnoncholesterol sterols during inhibition of choles-terol synthesis by statins. J Lab Clin Med 2003;141:131–7

31. Kajinami K, Brousseau ME, Nartsupha C, et al. ATPbinding cassette transporter G5 and G8 genotypesand plasma lipoprotein levels before and after treat-ment with atorvastatin. J Lipid Res 2004; 45:653–6

32. Gylling H, Hallikainen M, Pihlajamaki J, et al.Polymorphisms in the ABCG5 and ABCG8 genesassociate with cholesterol absorption and insulinsensitivity. J Lipid Res 2004; 45:1660–5

33. Brown AJ, Jessup W. Oxysterols and atherosclerosis.Atherosclerosis 1999; 142:1–28

34. Garcia-Cruset S, Carpenter K, Codony R, et al.Cholesterol oxidation products and atherosclerosis.In: Guardiola F, Dutta P, Codony R, Savage G, eds.Cholesterol and Phytosterol Oxidation Products:Analysis, Occurrence, and Biological Effects.Champaign, Illinois: AOCS Press, 2002: 241–77

35. Jacobson MS. Cholesterol oxides in Indian ghee:possible cause of unexplained high risk of atheroscle-rosis in Indian immigrant populations. Lancet 1987;2:656–8

36. Brown AJ, Leong SL, Dean RT, et al. 7-Hydroperoxy-cholesterol and its products in oxidized low densitylipoprotein and human atherosclerotic plaque. JLipid Res 1997; 38:1730–45

37. Bjorkhem I, Andersson O, Diczfalusy U, et al.Atherosclerosis and sterol 27–hydroxylase: evidencefor a role of this enzyme in elimination of cholesterolfrom human macrophages. Proc Natl Acad Sci USA1994; 91:8592–6

38. Crisby M, Nilsson J, Kostulas V, et al. Localization ofsterol 27-hydroxylase immuno-reactivity in humanatherosclerotic plaques. Biochim Biophys Acta 1997;1344:278–85

39. Suarna C, Dean RT, May J, et al. Human atheroscle-rotic plaque contains both oxidized lipids andrelatively large amounts of alpha-tocopherol andascorbate. Arterioscler Thromb Vasc Biol 1995;15:1616–24

40. Lu TT, Repa JJ, Mangelsdorf DJ. Orphan nuclearreceptors as eLiXiRs and FiXeRs of sterol metabo-lism. J Biol Chem 2001; 276:37735–8

41. Repa JJ, Berge KE, Pomajzl C, et al. Regulation ofATP-binding cassette sterol transporters ABCG5 andABCG8 by the liver X receptors alpha and beta. J BiolChem 2002; 277:18793–800

42. Osada K, Sasaki E, Sugano M. Lymphatic absorptionof oxidized cholesterol in rats. Lipids 1994; 29:555–9

43. Vine DF, Croft KD, Beilin LJ, et al. Absorption ofdietary cholesterol oxidation products and incorpora-tion into rat lymph chylomicrons. Lipids 1997;32:887–93

44. Ratnayake WM, Plouffe L, Hollywood R, et al.Influence of sources of dietary oils on the life span ofstroke-prone spontaneously hypertensive rats. Lipids2000; 35:409–20

45. Klett EL, Patel SB. Biomedicine. Will the real choles-terol transporter please stand up. Science 2004;303:1149–50

46. Altmann SW, Davis HR Jr, Zhu LJ, et al. Niemann-PickC1 Like 1 protein is critical for intestinal cholesterolabsorption. Science 2004; 303:1201–4

47. Davis HR Jr, Zhu LJ, Hoos LM, et al. Niemann-Pick C1Like 1 (NPC1L1) is the intestinal phytosterol andcholesterol transporter and a key modulator of whole-body cholesterol homeostasis. J Biol Chem 2004;279:33586–92


ch06 14/7/05 4:46 pm Page 68

INTRODUCTION

Bile acids (BAs) are synthesized from choles-terol exclusively by the liver. The biosyntheticsteps that collectively accomplish the conver-sion of hydrophobic, water-insoluble choles-terol molecules into more water-solublecompounds also confer detergent properties tothe BA that are crucial for their physiologicalfunctions in bile formation and intestinal fatabsorption. Bile acids are actively secreted bythe hepatocytes into the bile canaliculi thatdrain into intrahepatic bile ducts, stored in thegallbladder and expelled into the intestinallumen in response to a fatty meal. In the smallintestine, BAs act as detergents to emulsify andfacilitate the absorption of dietary fats andlipid-soluble vitamins. Subsequently, BAs arereabsorbed from the terminal ileum by theactions of specific transporter proteins: about95% returns to the liver to be secreted againinto the bile, completing the so-called entero-hepatic circulation, whereas 5% escapesreabsorption and is lost via the faeces. Thefraction of BAs that are lost per cycle iscompensated for by hepatic synthesis fromcholesterol, which maintains BA pool size.Although the fractional loss of BAs per cycle isrelatively small, daily BA synthesis in adulthumans amounts up to ~500 mg, whichaccounts for about 90% of the cholesterol thatis actively metabolized in the body.

The detergent properties of BAs, deter-mined in part by the number of hydroxylgroups present at the steroid moiety, arecrucial for most of their biological functions,

i.e. in generating bile flow, promoting biliaryexcretion of hydrophobic compounds andfacilitating intestinal fat absorption. The physi-cal characteristics of BAs, which allow them toform micelles, also impose a certain risk to cellsthat are exposed to high concentrations ofthese natural detergents. When present at highconcentrations, BAs may become cytotoxic.Hepatocytes and bile duct cells, in particular,are at risk, for instance in conditions ofdisturbed bile formation or stasis of bile in theductular system (cholestasis), and protectivemechanisms appear to become active whenintracellular BA concentrations are elevated.Obviously, both the maintenance of physiolog-ical control of the enterohepatic circulationand the initiation of cell-protective reactionsrequire a mode of ‘bile acid sensing’ in partic-ular cells. Recently, it became clear that BAsthemselves are directly involved in regulationof gene expression in liver and intestine viainteraction with a nuclear receptor calledfarnesoid X receptor (FXR), which providessuch a sensor function. Nuclear receptors areligand-activated transcription factors that, ingeneral, bind small molecules of endogenousor dietary origin, such as oxysterols, fatty acids,vitamins or certain drugs. FXR is involved inthe regulation of several steps of BA biology,including synthesis, detoxification and trans-port. Moreover, a direct link has recently beenestablished between BA–FXR activation andvarious aspects of lipid metabolism. Therefore,FXR modulation may constitute a new pharma-ceutical target with potential application in thetreatment of lipid disorders.

FXR: the molecular link between bile acidand lipid metabolismT. Claudel, E. Sturm, B. Staels and F. Kuipers

7

ch07 14/7/05 4:46 pm Page 69

THE BIOLOGY OF FXR

FXR (or NR1H4) was cloned independently bytwo groups using different strategies in 1995.In a quest for new retinoid X receptor (RXR)partners, the RXR interacting protein (RIP14)was isolated by Seol et al. Two isoforms weresubsequently cloned in mouse, that wereexpressed in the liver and in the intestine.1

Simultaneously, Forman et al cloned a newsequence of a nuclear receptor by liver cDNAbank screening.2 Since it was originally shownto be activated by the isoprenoid farnesol, thisreceptor was designated farnesoid X receptor.2

FXR expression was detected in the liver, theintestine, the kidney and in the adrenals ofadult rats and similar expression patterns havebeen reported in mice.3 Being closely relatedto the insect ecdysone receptor (EcR), it wasshown that FXR, after heterodimerization withits partner RXR, also binds to the EcR responseelement inverted repeat-1 (IR-1).

Analysis of the FXR locus demonstrated theexpression of four splice variants with differentN-terminal domains (FXRα or FXRβ) and afour-amino-acid insertion between the DNA-and the ligand-binding domains (FXRα1,FXRα2, FXRβ1, FXRβ2). Two promoters drivethe expression of the two FXR isoforms that areexpressed in a developmental and tissue-specificmanner with distinct transcriptional activities.4,5

Relatively little is known about the regula-tion of FXR expression. Very recently, we wereable to show that the expression of the FXRgene is suppressed in rat models of diabetes.Using rat hepatocytes in culture, it was demon-strated that glucose positively regulates FXRgene expression in a dose- and time-dependentmanner, probably by acting at the transcrip-tional level through the actions of metabolitesof the pentose-phosphate shunt.6

In 1999, three groups identified BAs asnatural FXR ligands.7–9 Interestingly, FXR canbe activated by several primary and secondaryBA species conjugated to either taurine orglycine (Figure 1) with an affinity in the micro-molar range.

To evaluate the potential beneficial effects ofpharmacological FXR activation, several groups

started the search for synthetic, non-steroidalFXR agonists. The first compound isolated wasthe potent FXR agonist GW4064, an isoxazolederivative, which is often used as a ‘chemicaltool’ to show that BA-target genes are regulatedin an FXR-specific manner10 since BAs are alsoable to activate FXR-independent pathways.11

Soon thereafter, several new syntheticcompounds were described. By modification ofthe BA backbone to increase its affinity for FXR,Pellicciari et al produced 6-ethyl CDCA, a morepotent FXR agonist than the natural BAchenodeoxycholic acid (CDCA, see further) inin vitro experiments that conferred protectionagainst cholestasis induced by the toxic BAspecies lithocholic acid in rats.12 Analysis byDownes et al of the properties of anotheragonist, FeXaRamine, in comparison to thesynthetic agonist GW4064 and the naturalagonist CDCA, showed that these compoundsdo not regulate the same genes.13 In addition,Dussault et al characterized the mixed proper-ties of AGN34, a molecule structurally derivedfrom a relatively weak and non-specific FXRagonist, which was shown to act either as anagonist, an antagonist or neutral on differentFXR target genes.14 In analogy with the oestro-gen receptor (ER), for which different ligandscalled selective oestrogen receptor modulators,or SERMs, were shown to induce distinctbiological responses depending on theresponse element structure in the promoter ofthe target gene, the co-factor recruitment and


C

OH R1

O

R2NHCH2COOH

NHCH2CH2SO3H

Glyco -conjugate

Tauro -conjugateor

Figure 1 FXR is a bile acid receptor. LCA, lithocholicacid; CDCA, chenodeoxycholic acid; DCA, deoxycholicacid; CA, cholic acid; UDCA, ursodeoxycholic acid. The+/– indicate the relative potencies of the bile acids toactivate the human FXR receptor

Bile acids R1 R2 hFXRLCA H H +CDCA αOH H +++DCA H αOH ++CA αOH αOH ++UDCA βOH H –

ch07 14/7/05 4:46 pm Page 70

the target tissue, these observations led to thenew concept of selective Bile Acid ReceptorModulator or BARM. The BARM concept willbe of crucial importance for the developmentof potential therapeutic applications of FXRmodulators.

THE ROLE OF FXR IN THECONTROL OF BILE ACIDMETABOLISM

Bile acids are amphipathic molecules withdetergent-like properties that are essential fortheir physiological functions. As a BA receptor,FXR controls several of the adaptive changesthat are required to maintain optimal BAconcentrations at the sites of their physiologicalactions, i.e. in the biliary tree to generate bileflow and in the intestinal lumen to facilitate fatabsorption. At the same time, FXR is crucial forthe protection of cells contributing to theenterohepatic circulation from BA-inducedtoxicity, for instance when the capacity of livercells to secrete BAs into the bile is perturbed incholestatic liver disease. During the past few

years, it has become clear that FXR has a crucialrole in control of BA synthesis and transport aswell as in their ‘detoxification’ when intracellu-lar concentrations exceed a certain threshold.

Bile acid synthesis

Bile acids are endproducts of a series of choles-terol-converting enzymatic reactions.Chenodeoxycholic acid (CDCA) and cholicacid (CA) are the major primary BAs inhumans that can be synthesized by two differ-ent pathways: the neutral pathway producesboth CDCA and CA, whereas the acidicpathway gives rise to CDCA only (Figure 2).

Microsomal CYP7A1 is the rate-controllingenzyme of the neutral pathway. It has beenknown for decades that BAs exert a negativefeedback regulation on their own synthesis.Recently, FXR was identified as a key player inBA-induced downregulation of CYP7A1 expres-sion, through an indirect mechanism.15 Theoriginal observation that SHP (smallheterodimer partner or NR0B2) expression wasreduced in FXR-deficient mice whereasCYP7A1 expression was increased, combined

FXR: THE MOLECULAR LINK BETWEEN BILE ACID AND LIPID METABOLISM 71

Neutral pathway

Acidic pathway

Cholesterol

CYP7A1

CYP27A

CDCA

CACYP8B1

CYP27A

CDCA

CA

CYP27A

SHP

FXR activation

Bile acids

FXR

SHP/LRH-1 SHP/HNF4

FXR

SHP/LRH-1

Figure 2 FXR controls the bile acid biosynthetic pathways. CYP7A1 is the key enzyme of the neutral pathway,whereas CYP27A is the key enzyme of the acidic pathway. CYP7A1 and CYP8B1 are negatively regulated by FXR (greyboxes) after induction of SHP. The inactive SHP/LRH-1 complex will subsequently impair CYP7A1, whereasSHP/HNF4 will impair CYP8B1 promoter trans-activation. Moreover, SHP/LRH-1 will also turn off SHP inductionand thus the repression mechanism. The normal arrows show activation, the block arrows inhibition

ch07 14/7/05 4:46 pm Page 71

with the fact that the atypical nuclear receptorSHP was devoid of a DNA-binding domain andclosely related to another orphan nuclear Dax-1 (dosage-sensitive sex reversal – congenitaladrenal hypoplasia or NR0B1), led to a newhypothesis based on the model of the Dax-1/SF-1 (steroidogenic factor 1 or NR5A2) inter-action. FXR activation was found to increasegene expression of SHP. Subsequently, theinteraction between SHP and LRH-1 (liverreceptor homologue-1 or NR5A2) inhibitsLRH-1 trans-activation of CYP7A116,17 (Figure2). Moreover, the basal gene expression of SHPis also activated by LRH-1, therefore theinactive SHP/LRH-1 complex reduces SHPexpression and turns off the negative feedbacksignal (Figure 2). FXR also induces the expres-sion of a secreted growth factor that activatesthe FGF4 receptor isotype in hepatocytes:fibroblast growth factor-19 (FGF-19).18 FGF-19represses CYP7A1 expression in human andmouse hepatocytes via a signalling cascadeinvolving the c-Jun N-terminal kinase pathway.Interestingly, FGFR4-deficient mice display anincreased BA pool size and higher hepatobiliaryBA excretion rate19 and CYP7A1 expression isunresponsive to dietary cholesterol, but can berepressed by BA in these mice.

However, CYP7A1 regulation is extremelycomplex and several other, FXR-independentmechanisms have been invoked in the negativefeedback regulation exerted by circulatingBA.11,20 Discussion of these mechanisms isbeyond the scope of this chapter, but can befound in excellent recent reviews.21,22

The CYP8B1 enzyme controls the hydropho-bicity of the BA pool by modulating the relativeamount of CA synthesized. Einarrsson et aldemonstrated that CYP8B1 expression is upreg-ulated after ileal resection, suggesting alleviationof negative feedback control by BA.23 Underlyingmechanisms may involve a decreased expressionof HNF4α and/or a decreased trans-activation ofthe CYP8B1 promoter by HNF4α due to SHPinduction24 (Figure 2).

Finally, it is important to know that theacidic pathway is not under the control of BA,since the key enzyme CYP27 is not regulated byBA nor changed in FXR-deficient mice.25

Bile acid transport within theenterohepatic circulation

After conjugation with either taurine or glycineon their side chain, a process also regulated atthe transcriptional level by FXR,26 BAs becomemore hydrophilic and require a transporternetwork to cycle between liver and intestine.

During their enterohepatic cycle, BAs aretaken up by the liver via the Na+ taurocholateco-transporting polypeptide (NTCP),27–29

expressed exclusively at the basolateral plasmamembrane of hepatocytes,30,31 which representsalmost 75% of BA uptake (Figure 3). NTCPexpression is downregulated by FXR since it isreduced in wild-type but not in FXR-deficientmice upon CA treatment,25 but its basal expres-sion is not different between wild-type andFXR-deficient mice.25,32 FXR acts, via SHPinduction, to inhibit RXR/RAR trans-activationof the NTCP promoter, at least in rats.33 Giventhe fact that the RXR/RAR site is lacking inhuman and mouse NTCP promoters, it isunlikely that a negative transcriptional regula-tion by BA will occur in these species.34

A Na+-independent BA uptake is mediated bythe organic anion transporter polypeptides(OATPs). In humans, OATP-C is the mostabundant Na+-independent BA transporter local-ized at the basolateral membrane35–37 (Figure 3).In mice, OATP1 expression is either repressedor not affected by BA treatment.25,38 Importantly,FXR deficiency did not change OATP1 geneexpression at the basal level in mice.25,32

Within the hepatocytes, BAs traffic from thebasolateral membrane across the cells to thecanalicular pole prior to their secretion intobile. The mechanism(s) of trans-cellular BAtransport remain poorly understood and it isunclear whether FXR is functionally involved inthe regulation of this process. Several proteinswere identified as potential intracellular BAcarriers, like the liver fatty-acid-binding protein(L-FABP), which is induced by CA treatment inwild-type but not in FXR-deficient mice.25

BA secretion at the canalicular membrane isan ATP-dependent process mediated by thebile salt export pump (BSEP or ABCB11).39–43

BSEP mutations underlie the bile secretion


ch07 14/7/05 4:46 pm Page 72

failure in patients with progressive familialintrahepatic cholestasis type II (PFIC 2).44

Inactivation of BSEP in mice results in a lesssevere phenotype than in humans,45 which maybe due to differences in BA metabolismbetween the species. BA–FXR activationincreases BSEP expression in rodents as well asin human liver cells46,47 (Figure 3).

Taken together, FXR activation decreasesBA uptake by NTCP downregulation andincreases BA excretion via BSEP induction, amechanism that will protect the hepatocytesfrom a potentially toxic BA overload. SinceFXR-deficient mice display low BSEP expres-sion but have normal biliary BA excretion,32

BSEP expression is not limiting for biliary BAsecretion at the basal level. This suggests that

the overabundant expression of BSEP undernormal conditions provides the hepatocyteswith a safety valve to deal with situations inwhich the liver is exposed to a high BA influx,e.g. in the post-prandial period.

Bile acid absorption at the intestinal leveloccurs with an extremely high efficiency(~95%). BAs cross the enterocyte membranesvia the action of the apical sodium-dependentbile acid transporter (ASBT) protein.48–52 BAdoes not regulate rat ASBT gene expression,53

but FXR influences the transcription factornetwork involved in the BA response54,55 (Figure3). In the enterocytes, FXR induces the expres-sion of IBABP (ileal bile acid binding protein),a small cytosolic protein.56 Until recently, it wasbelieved that IBABP acts as a shuttle deliveringBA from the apical to the basolateralmembrane.57 Yet, the observation that IBABPexpression is strongly decreased in FXR-deficient mice while BA absorption is enhancedindicates that the exact physiological role ofIBABP remains to be defined.32 It is plausible tosuggest that IBABP may act as a ‘BA sink’ toprotect enterocytes from BA toxicity whenexposed to high concentrations. Bile acids maycross the basolateral membrane of enterocytesto enter the portal blood stream using the MRP3transporter58 or the tABST protein59 (Figure 3).

Bile acid detoxification

As mentioned previously, relatively hydropho-bic BA species, as exemplified by CDCA and itsbacterial metabolite lithocholic acid, arepotentially toxic agents and readily inducecholestasis when administered to rodents.Sulphation and glucuronidation are commonphase II detoxification reactions to facilitateremoval of poorly water-soluble endo- andxenobiotics from the body. It has been knownfor more than 30 years that these reactions alsooccur with BA, especially in conditions associ-ated with cholestatic liver disease. Dehydroepi-androsterone– sulphotransferase SULT2A1 is acytosolic enzyme that promotes 3α-OHsulphation of BAs.60 SULT2A1 gene expressionis induced by FXR.61 The human uridineglucuronosyltransferase 2B4, that converts BAs


Bile acids

Intestine

Liver

IBAB-P

BSEP

NTCP

MRP3

ASBT

Bile acids

Cyp7A1

Cholesterol

Bile acids

FXR

FXR

Bile acids

UGT2B4SULT2A1

tASBT

FXR

OATP

?

Figure 3 FXR controls the bile acid enterohepaticcycle. At the hepatic level, FXR represses the expressionof NTCP and CYP7A1. Simultaneously, FXR induces theexpression of the detoxification enzymes SULT2A1 andUGT2B4 and promotes BA clearance by the inductionof BSEP. At the intestinal level, FXR represses ilealApical Sodium-dependent Bile acid Transporter (ASBT)expression and induces the expression of the putativeintracellular transporter IBABP. BAs will subsequently goto the portal circulation after excretion by MRP3 ortASBT. The normal arrows show activation, the blockarrows inhibition

ch07 14/7/05 4:46 pm Page 73

into more hydrophilic glucuronide deriva-tives,62,63 is positively regulated by FXR bindingto a monomeric FXRE localized in its pro-moter.64 Since glucuronidated and sulphatedBAs are usually more water-soluble, FXR mayreduce BA toxicity as well by enhancing theirclearance via bile and/or urine. Thus, a majorphysiological role of the BA receptor FXR is toprotect cells from the deleterious effects of BAoverload by decreasing endogenous BAproduction and promoting BA biotransforma-tion and clearance.

AN EMERGING ROLE FOR FXR INTHE CONTROL OF LIPIDMETABOLISM

In the past few years, several studies havedemonstrated that BAs, via FXR, are potentmodulators of lipid and lipoprotein metabo-lism in a way that goes far beyond their estab-lished functions as ‘catabolites’ of cholesteroland as ‘soaps’ that facilitate the absorption offat and cholesterol from the intestine. Thisnewly established physiological function of BAis of potentially great interest for the develop-ment of alternative strategies for the treatmentof hyperlipidaemias.

FXR and HDL

The resins cholestyramine, colestipol andcolesevelam are non-absorbable drugs withhigh affinity for BAs in the intestine. Afterbinding to the BAs, the resin/BA complex iseliminated via the faeces. As a consequence ofthe interruption of the enterohepatic circula-tion, CY7A1 expression in the liver will beinduced which, in turn, will increase theconversion of cholesterol into BA. The ensuinghepatic cholesterol depletion will then lead toan increase in surface-active LDL receptors andtherefore to a fall in low density lipoprotein(LDL)-cholesterol levels due to increasedclearance of apoB-100- and apoE-containinglipoproteins. A side-effect of this modificationin hepatic cholesterol homeostasis is a rise inhepatic HMG-CoA reductase activity in

conjunction with an increased very low densitylipoprotein (VLDL) triglyceride production(see below).

Interestingly, treatment with BA-bindingresins also leads to an increase in high densitylipoprotein (HDL)-cholesterol levels,65,66

thereby promoting the reverse cholesteroltransport pathway (Figure 4). Intriguingly,patients with gallstones and cerebrotendinousxanthomatosis (CTX) treated with BAs displaylower levels of HDL-cholesterol.67,68 Moreover,patients with progressive familial intrahepaticcholestasis (PFIC), a disease characterized byintrahepatic accumulation of BA, have lowserum levels of apoA-I, the major HDLapolipoprotein.69 The molecular basis for thisinteraction between BA and HDL metabolismhas remained enigmatic for many years.Recently, we were able to demonstrate thatboth the human and mouse apo A-I genes arenegatively regulated by FXR, that binds as amonomer to a negative response element inthe apo A-I promoter.69 In line with this, chow-fed FXR-deficient mice displayed higher HDL-cholesterol levels than wild-type mice.25 Inaddition, FXR was also shown to induce theexpression of the phospholipid transferprotein (PLTP),70 an enzyme involved in HDLremodelling (Figure 4).

FXR and triglyceride metabolism

As already described previously, disruption ofthe enterohepatic cycle of BAs in humansusing either sequestrants71 or ileal bypasssurgery72 results in an increase in serum triglyc-eride levels.73 Furthermore, hypertriglyceri-daemia in type IV hyperlipoproteinaemiapatients is associated with impaired absorptionof BA74 due to diminished ASBT expression.73

Conversely, triglyceride levels decrease ingallstone patients as well as in hyperlipaemicpatients treated with BA.75–78 Finally, FXR-deficient mice have higher triglyceride levelsthan wild-type mice, mainly confined to theVLDL-sized lipoprotein fractions.25 Takentogether, these results suggest that FXRactivation by BA also modulates triglyceridemetabolism.


ch07 14/7/05 4:46 pm Page 74

A key player in the control of plasma triglyc-eride metabolism is the lipoprotein lipase (LPL),an enzyme that mediates lipolysis of triglyceride-rich lipoproteins. ApoC-II and apoA-V are activa-tors, whereas apoC-III is an inhibitor of LPLactivity. Since FXR activation increases apoC-IIexpression79 and simultaneously inhibits apoC-III gene expression in the liver80 and since,although not confirmed by in vivo data, FXRtrans-activated the promoter of the human apoA-V gene,81 the overall effect of FXR activation willbe to increase the LPL activity and therefore toaccelerate triglyceride clearance.

The peroxisome proliferator-activated recep-tor α (PPAR-α) is a nuclear receptor activatedby fatty acids82 and the synthetic drugs fibrates83

that also regulates BA synthesis and poolcomposition.84,85 Moreover, PPAR-α activation

increases LPL activity and subsequently lowerstriglyceride levels. PPAR-α is a species-specificFXR-target gene, i.e. in humans, but not inmice, FXR increases PPAR-α gene expression.86

The existence of such a cross-talk betweenthese nuclear receptors suggests the existenceof a delicate co-ordination between fatty acidand BA metabolism.

The peroxisome proliferator activated recep-tor gamma coactivator 1α (PGC-1α) is anuclear receptor co-factor involved in thecontrol of adaptive thermogenesis and glucosemetabolism as well as CYP7A1 expression.87

Under fasting conditions the expression of theFXRβ isoform is induced via a mechanismimplicating PGC-1α and PGC-1α was alsoidentified as a potential FXR cofactor involvedin the increase in triglyceride clearance and


Extrahepatictissues

Cholesterol

BA

SR-BI

Liver Bile acids

Cholesterol

CholesterolPL

ApoA-I HDL2

LCAT

ApoA-I

BileFXR

Intestine

Apo CIII

ApoC-II

TG

Remnants

LPL

Peripheral tissues

PLTP

PLTPLPL

Remnants

ApoC-III

Figure 4 FXR controls plasma HDL and triglyceride metabolism. HDL-cholesterol promotes cholesterol excretionfrom the peripheral tissues to the liver. FXR represses apoA-I and induces PLTP gene expression, respectively.Simultaneously, FXR induces apoC-II (LPL co-factor) and represses apoC-III (LPL inhibitor) gene expression,respectively, and therefore increases LPL activity and triglyceride clearance. The normal arrows show activation, theblock arrows inhibition. PL, phospholipid; LCAT, lecithin:cholesterol acyltransferase; PLTP, phospholipid transferprotein; LPL, lipoprotein lipase; TG, triglycerides

ch07 14/7/05 4:46 pm Page 75

the lowering of triglyceride synthesis mediatedby FXR.88

Finally, the trans-membrane heparansulphate proteoglycan syndecan 1 is expressedin the liver and binds several lipoproteins andproteins, such as LPL.89 Remnant lipoproteinparticles are believed to bind to syndecan 1prior to their transfer to receptors, such as theLDL-receptor, which increases lipoproteincatabolism. FXR increases syndecan 1 geneexpression in an isoform-dependent manner,i.e. FXR-α2 and FXR-β2 are involved in thetrans-activation promoter.90

Taken together, it is evident that BAactivated-FXR modulates several aspects oftriglyceride metabolism. Therefore, it is attrac-tive to speculate that stimulation of the entero-hepatic BA flux upon ingestion of a (fatty) mealprovides a signal that prepares the body toadequately handle the ‘upcoming’ fat load andthat blunts the post-prandial hyperlipidaemia.

NATURAL NON-BILE ACID FXRMODULATORS: THEGUGGULSTERONE STORY

Guggulipid, a resin extract of the treeCommiphora mukul, is a traditional medicinalherb used in India to treat obesity and lipiddisorders.91,92 Given the therapeutic interest inFXR modulators, several recent studies havefocused on the action mechanism of E- and Z-guggulsterone, which are believed to be theactive components of guggulipid.93 Urizar et aldemonstrated that guggulsterone acted as FXRantagonist and pregnane X receptor (PXR)partial agonist.94 Moreover, in vivo experimentsusing mice on a high-cholesterol diet showedthat guggulsterone treatment decreased the

hepatic cholesterol content in an FXR-depen-dent manner.94 Nevertheless, guggulsteroneseems to act both as FXR and PXR modula-tor.95 It was subsequently proposed that theFXR modulation properties of the drugexplained its hypolipidaemic effect. Untilrecently, appropriately well-designed humanstudies on the effects of the guggulsteroneswere lacking, especially with patients fromWestern countries. In 2003, Szapary et aldemonstrated a complete failure of guggul-sterone to decrease the cholesterol levels inhypercholesterolaemic patients.96 Therefore,despite a plausible mechanism of action viaFXR, guggulsterone seems not to be ‘the’appropriate FXR modulator suitable for thetreatment of obesity or hyperlipidaemia. Thequest for such a compound is still open.

CONCLUSION

FXR is a nuclear receptor that acts at the cross-road of BA and lipid metabolism. It is evidentthat FXR constitutes a potential attractivetarget for treatment of cholestatic liver diseasesas well as of (specific forms of) hyperlipi-daemia. From its multitude of actions, it is clearthat selective FXR modulators are required tolimit potential undesirable side-effects. Sincesuch selective modulators have already beengenerated, exciting results from this area ofresearch can be expected in the coming years.

ACKNOWLEDGEMENTS

Thierry Claudel was supported by Grantnumber 2002B017 from the NederlandseHartstichting.


1. Seol W, Choi HS, Moore DD. Isolation of proteinsthat interact specifically with the retinoid X receptor:two novel orphan receptors. Mol Endocrinol 1995;9:72–85

2. Forman BM, Goode E, Chen J, et al. Identification ofa nuclear receptor that is activated by farnesolmetabolites. Cell 1995; 81:687–93

3. Repa JJ, Mangelsdorf DJ. Nuclear receptor regulation

References

ch07 14/7/05 4:46 pm Page 76

of cholesterol and bile acid metabolism. Curr OpinBiotechnol 1999; 10:557–63

4. Huber RM, Murphy K, Miao B, et al. Generation ofmultiple farnesoid-X-receptor isoforms through theuse of alternative promoters. Gene 2002; 290:35–43

5. Zhang Y, Kast-Woelbern HR, Edwards PA. Naturalstructural variants of the nuclear receptor farnesoid Xreceptor affect transcriptional activation. J Biol Chem2003; 278:104–10

6. Duran-Sandoval D, Mautino G, Martin G, et al.Glucose regulates the expression of the farnesoid Xreceptor in liver. Diabetes 2004; 53:890–8

7. Wang H, Chen J, Hollister K, et al. Endogenous bileacids are ligands for the nuclear receptor FXR/BAR.Mol Cell 1999; 3:543–53

8. Makishima M, Okamoto AY, Repa JJ, et al.Identification of a nuclear receptor for bile acids.Science 1999; 284:1362–5

9. Parks DJ, Blanchard SG, Bledsoe RK , et al. Bile acids:natural ligands for an orphan nuclear receptor.Science 1999; 284:1365–8

10. Maloney PR, Parks DJ, Haffner CD, et al.Identification of a chemical tool for the orphannuclear receptor FXR. J Med Chem 2000; 43:2971–4

11. Gupta S, Stravitz RT, Dent P, et al. Down-regulation ofcholesterol 7alpha-hydroxylase (CYP7A1) geneexpression by bile acids in primary rat hepatocytes ismediated by the c-Jun N-terminal kinase pathway. JBiol Chem 2001; 276:15816–22

12. Pellicciari R, Fiorucci S, Camaioni E, et al. 6alpha-ethyl-chenodeoxycholic acid (6α-ECDCA), a potentand selective FXR agonist endowed with anticholesta-tic activity. J Med Chem 2002; 45:3569–72

13. Downes M, Verdecia MA, Roecker AJ, et al. A chemi-cal, genetic, and structural analysis of the nuclear bileacid receptor FXR. Mol Cell 2003; 11:1079–92

14. Dussault I, Beard R, Lin M, et al. Identification ofgene-selective modulators of the bile acid receptorFXR. J Biol Chem 2003; 278:7027–33

15. Chiang JY, Kimmel R, Weinberger C, et al. FarnesoidX receptor responds to bile acids and repressescholesterol 7alpha-hydroxylase gene (CYP7A1)transcription. J Biol Chem 2000; 275:10918–24

16. Goodwin B, Jones SA, Price RR, et al. A regulatorycascade of the nuclear receptors FXR, SHP-1, andLRH-1 represses bile acid biosynthesis. Mol Cell 2000;6:517–26

17. Lu TT, Makishima M, Repa JJ, et al. Molecular basisfor feedback regulation of bile acid synthesis bynuclear receptors. Mol Cell 2000; 6:507–15

18. Holt JA, Luo G, Billin AN, et al. Definition of a novelgrowth factor-dependent signal cascade for thesuppression of bile acid biosynthesis. Genes Dev 2003;17:1581–91

19. Yu C, Wang F, Kan M, et al. Elevated cholesterolmetabolism and bile acid synthesis in mice lackingmembrane tyrosine kinase receptor FGFR4. J BiolChem 2000; 275:15482–9

20. Miyake JH, Wang SL, Davis RA. Bile acid induction of

cytokine expression by macrophages correlates withrepression of hepatic cholesterol 7alpha-hydroxylase.J Biol Chem 2000; 275:21805–8

21. Russell DW. The enzymes, regulation, and genetics ofbile acid synthesis. Annu Rev Biochem 2003;72:137–74

22. Chiang JY. Regulation of bile acid synthesis: pathways,nuclear receptors, and mechanisms. J Hepatol 2004;40:539–51

23. Einarsson K, Akerlund JE, Reihner E, et al. 12 alpha-hydroxylase activity in human liver and its relation tocholesterol 7 alpha-hydroxylase activity. J Lipid Res1992; 33:1591–5

24. Zhang M, Chiang JY. Transcriptional regulation ofthe human sterol 12alpha-hydroxylase gene(CYP8B1). Roles of hepatocyte nuclear factor 4alphain mediating bile acid repression. J Biol Chem 2001;276:41690–9

25. Sinal CJ, Tohkin M, Miyata M, et al. Targeted disrup-tion of the nuclear receptor FXR/BAR impairs bileacid and lipid homeostasis. Cell 2000; 102:731–44

26. Pircher PC, Kitto JL, Petrowski ML, et al. Farnesoid Xreceptor regulates bile acid-amino acid conjugation. JBiol Chem 2003; 278:27703–11

27. Hagenbuch B, Stieger B, Foguet M, et al. Functionalexpression cloning and characterization of thehepatocyte Na+/bile acid cotransport system. ProcNatl Acad Sci USA 1991; 88:10629–33

28. Hagenbuch B, Meier PJ. Molecular cloning, chromo-somal localization, and functional characterization ofa human liver Na+/bile acid cotransporter. J ClinInvest 1994; 93:1326–31

29. Cattori V, Eckhardt U, Hagenbuch B. Molecularcloning and functional characterization of two alter-natively spliced Ntcp isoforms from mouse liver1.Biochim Biophys Acta 1999; 1445:154–9

30. Stieger B, Hagenbuch B, Landmann L, et al. In situlocalization of the hepatocytic Na+/taurocholatecotransporting polypeptide in rat liver. Gastro-enterology 1994; 107:1781–7

31. Ananthanarayanan M, Ng OC, Boyer JL, et al.Characterization of cloned rat liver Na(+)-bile acidcotransporter using peptide and fusion proteinantibodies. Am J Physiol 1994; 267:G637–43

32. Kok T, Hulzebos CV, Wolters H, et al. Enterohepaticcirculation of bile salts in farnesoid X receptor-deficient mice: efficient intestinal bile salt absorptionin the absence of ileal bile acid-binding protein. J BiolChem 2003; 278:41930–7

33. Denson LA, Sturm E, Echevarria W, et al. The orphannuclear receptor, shp, mediates bile acid-inducedinhibition of the rat bile acid transporter, ntcp.Gastroenterology 2001; 121:140–7

34. Jung D, Hagenbuch B, Fried M, et al. Role of liver-enriched transcription factors and nuclear receptorsin regulating the human, mouse, and rat NTCP gene.Am J Physiol Gastrointest Liver Physiol 2004;286:G752–61

35. Abe T, Kakyo M, Tokui T, et al. Identification of a


ch07 14/7/05 4:46 pm Page 77

novel gene family encoding human liver-specificorganic anion transporter LST-1. J Biol Chem 1999;274:17159–63

36. Hsiang B, Zhu Y, Wang Z, et al. A novel humanhepatic organic anion transporting polypeptide(OATP2). Identification of a liver-specific humanorganic anion transporting polypeptide and identifi-cation of rat and human hydroxymethylglutaryl-CoAreductase inhibitor transporters. J Biol Chem 1999;274:37161–8

37. Konig J, Cui Y, Nies AT, et al. A novel human organicanion transporting polypeptide localized to thebasolateral hepatocyte membrane. Am J PhysiolGastrointest Liver Physiol 2000; 278:G156–64

38. Fickert P, Zollner G, Fuchsbichler A, et al. Effects ofursodeoxycholic and cholic acid feeding on hepato-cellular transporter expression in mouse liver.Gastroenterology 2001; 121:170–83

39. Adachi Y, Kamisako T, Yamamoto T. The effects oftemporary occlusion of the superior mesenteric veinor splenic vein on biliary bilirubin and bile acid excre-tion in rats. J Lab Clin Med 1991; 118:261–8

40. Muller M, Ishikawa T, Berger U, et al. ATP-dependenttransport of taurocholate across the hepatocytecanalicular membrane mediated by a 110-kDa glyco-protein binding ATP and bile salt. J Biol Chem 1991;266:18920–6

41. Nishida T, Gatmaitan Z, Che M, et al. Rat livercanalicular membrane vesicles contain an ATP-dependent bile acid transport system. Proc Natl AcadSci USA 1991; 88:6590–4

42. Gerloff T, Stieger B, Hagenbuch B, et al. The sister ofP-glycoprotein represents the canalicular bile saltexport pump of mammalian liver. J Biol Chem 1998;273:10046–50

43. Green RM, Hoda F, Ward KL. Molecular cloning andcharacterization of the murine bile salt export pump.Gene 2000; 241:117–23

44. Jansen PL, Strautnieks SS, Jacquemin E, et al.Hepatocanalicular bile salt export pump deficiency inpatients with progressive familial intrahepaticcholestasis. Gastroenterology 1999; 117:1370–9

45. Wang R, Salem M, Yousef IM, et al. Targeted inactiva-tion of sister of P-glycoprotein gene (spgp) in miceresults in nonprogressive but persistent intrahepaticcholestasis. Proc Natl Acad Sci USA 2001; 98:2011–16

46. Ananthanarayanan M, Balasubramanian N,Makishima M, et al. Human bile salt export pumppromoter is transactivated by the farnesoid X recep-tor/bile acid receptor. J Biol Chem 2001;276:28857–65

47. Plass JR, Mol O, Heegsma J, et al. Farnesoid X recep-tor and bile salts are involved in transcriptionalregulation of the gene encoding the human bile saltexport pump. Hepatology 2002; 35:589–96

48. Wong MH, Oelkers P, Craddock AL, Dawson PA.Expression cloning and characterization of thehamster ileal sodium-dependent bile acid trans-porter. J Biol Chem 1994; 269:1340–7

49. Wong MH, Oelkers P, Dawson PA. Identification of amutation in the ileal sodium-dependent bile acidtransporter gene that abolishes transport activity. JBiol Chem 1995; 270:27228–34

50. Shneider BL, Dawson PA, Christie DM, et al. Cloningand molecular characterization of the ontogeny of arat ileal sodium-dependent bile acid transporter. JClin Invest 1995; 95:745–54

51. Saeki T, Matoba K, Furukawa H, et al.Characterization, cDNA cloning, and functionalexpression of mouse ileal sodium-dependent bile acidtransporter. J Biochem (Tokyo) 1999; 125:846–51

52. Dawson PA, Haywood J, Craddock AL, et al. Targeteddeletion of the ileal bile acid transporter eliminatesenterohepatic cycling of bile acids in mice. J BiolChem 2003; 278:33920–7

53. Arrese M, Trauner M, Sacchiero RJ, et al. Neitherintestinal sequestration of bile acids nor common bileduct ligation modulate the expression and functionof the rat ileal bile acid transporter. Hepatology 1998;28:1081–7

54. Chen F, Ma L, Dawson PA, et al. Liver receptorhomologue-1 mediates species- and cell line-specificbile acid-dependent negative feedback regulation ofthe apical sodium-dependent bile acid transporter. JBiol Chem 2003; 278:19909–16

55. Neimark E, Chen F, Li X, et al. Bile acid-inducednegative feedback regulation of the human ileal bileacid transporter. Hepatology 2004; 40:149–56

56. Grober J, Zaghini I, Fujii H, et al. Identification of abile acid-responsive element in the human ileal bileacid-binding protein gene. Involvement of the farne-soid X receptor/9–cis-retinoic acid receptorheterodimer. J Biol Chem 1999; 274:29749–54

57. Kramer W, Corsiero D, Friedrich M, et al. Intestinalabsorption of bile acids: paradoxical behaviour of the14 kDa ileal lipid-binding protein in differentialphotoaffinity labelling. Biochem J 1998; 333 (Pt2):335–41

58. Inokuchi A, Hinoshita E, Iwamoto Y, et al. Enhancedexpression of the human multidrug resistanceprotein 3 by bile salt in human enterocytes. Atranscriptional control of a plausible bile acid trans-porter. J Biol Chem 2001; 276:46822–9

59. Lazaridis KN, Tietz P, Wu T, et al. Alternative splicingof the rat sodium/bile acid transporter changes itscellular localization and transport properties. ProcNatl Acad Sci USA 2000; 97:11092–7

60. Barnes S, Buchina ES, King RJ, et al. Bile acid sulfo-transferase I from rat liver sulfates bile acids and3–hydroxy steroids: purification, N-terminal aminoacid sequence, and kinetic properties. J Lipid Res1989; 30:529–40

61. Song CS, Echchgadda I, Baek BS, et al. Dehydro-epiandrosterone sulfotransferase gene induction bybile acid activated farnesoid x receptor. J Biol Chem2001; 276:42549–56

62. Pillot T, Ouzzine M, Fournel-Gigleux S, et al.Glucuronidation of hyodeoxycholic acid in human


ch07 14/7/05 4:46 pm Page 78

liver. Evidence for a selective role of UDP-glucurono-syltransferase 2B4. J Biol Chem 1993; 268:25636–42

63. Monaghan G, Burchell B, Boxer M. Structure of thehuman UGT2B4 gene encoding a bile acid UDP-glucuronosyltransferase. Mamm Genome 1997;8:692–4

64. Barbier O, Torra IP, Sirvent A, et al. FXR induces theUGT2B4 enzyme in hepatocytes: a potential mecha-nism of negative feedback control of FXR activity.Gastroenterology 2003; 124:1926–40

65. Bard JM, Parra HJ, Douste-Blazy P, et al. Effect ofpravastatin, an HMG CoA reductase inhibitor, andcholestyramine, a bile acid sequestrant, on lipopro-tein particles defined by their apolipoprotein compo-sition. Metabolism 1990; 39:269–73

66. Hagen E, Istad H, Ose L, et al. Fluvastatin efficacy andtolerability in comparison and in combination withcholestyramine. Eur J Clin Pharmacol 1994; 46:445–9

67. Leiss O, von Bergmann K. Different effects ofchenodeoxycholic acid and ursodeoxycholic acid onserum lipoprotein concentrations in patients withradiolucent gallstones. Scand J Gastroenterol 1982;17:587–92

68. Kuriyama M, Tokimura Y, Fujiyama J, et al. Treatmentof cerebrotendinous xanthomatosis: effects ofchenodeoxycholic acid, pravastatin, and combineduse. J Neurol Sci 1994; 125:22–8

69. Claudel T, Sturm E, Duez H, et al. Bile acid-activatednuclear receptor FXR suppresses apolipoprotein A-Itranscription via a negative FXR response element. JClin Invest 2002; 109:961–71

70. Urizar NL, Dowhan DH, Moore DD. The farnesoid X-activated receptor mediates bile acid activation ofphospholipid transfer protein gene expression. J BiolChem 2000; 275:39313–17

71. Molgaard J, von Schenck H, Olsson AG. Comparativeeffects of simvastatin and cholestyramine in treatmentof patients with hypercholesterolemia. Eur J ClinPharmacol 1989; 36:455–60

72. Buchwald H, Varco RL, Matts JP, et al. Effect of partialileal bypass surgery on mortality and morbidity fromcoronary heart disease in patients with hypercholes-terolemia. Report of the Program on the SurgicalControl of the Hyperlipidemias (POSCH). N Engl JMed 1990; 323:946–55

73. Duane WC, Hartich LA, Bartman AE, et al.Diminished gene expression of ileal apical sodiumbile acid transporter explains impaired absorption ofbile acid in patients with hypertriglyceridemia. J LipidRes 2000; 41:1384–9

74. Angelin B, Hershon KS, Brunzell JD. Bile acid metab-olism in hereditary forms of hypertriglyceridemia:evidence for an increased synthesis rate in monogenicfamilial hypertriglyceridemia. Proc Natl Acad Sci USA1987; 84:5434–8

75. Miller NE, Nestel PJ. Triglyceride-lowering effect ofchenodeoxycholic acid in patients with endogenoushypertriglyceridaemia. Lancet 1974; 2:929–31

76. Begemann F. Influence of chenodeoxycholic acid on

the kinetics of endogenous triglyceride transport inman. Eur J Clin Invest 1978; 8:283–8

77. Angelin B, Einarsson K, Hellstrom K, et al. Effects ofcholestyramine and chenodeoxycholic acid on themetabolism of endogenous triglyceride in hyper-lipoproteinemia. J Lipid Res 1978; 19:1017–24

78. Bateson MC, Maclean D, Evans JR, et al.Chenodeoxycholic acid therapy for hypertriglyceri-daemia in men. Br J Clin Pharmacol 1978; 5:249–54

79. Kast HR, Nguyen CM, Sinal CJ, et al. Farnesoid X-activated receptor induces apolipoprotein C-IItranscription: a molecular mechanism linking plasmatriglyceride levels to bile acids. Mol Endocrinol 2001;15:1720–8

80. Claudel T, Inoue Y, Barbier O, et al. Farnesoid Xreceptor agonists suppress hepatic apolipoproteinCIII expression. Gastroenterology 2003; 125:544–55

81. Prieur X, Coste H, Rodriguez JC. The humanapolipoprotein AV gene is regulated by peroxisomeproliferator-activated receptor-alpha and contains anovel farnesoid X-activated receptor responseelement. J Biol Chem 2003; 278:25468–80

82. Willson TM, Brown PJ, Sternbach DD, et al. ThePPARs: from orphan receptors to drug discovery. JMed Chem 2000; 43:527–50


84. Patel DD, Knight BL, Soutar AK, et al. The effect ofperoxisome-proliferator-activated receptor-alpha onthe activity of the cholesterol 7 alpha-hydroxylasegene. Biochem J 2000; 351 (Pt 3):747–53

85. Post SM, Duez H, Gervois PP, et al. Fibrates suppressbile acid synthesis via peroxisome proliferator-activated receptor-alpha-mediated downregulation ofcholesterol 7alpha-hydroxylase and sterol 27-hydroxy-lase expression. Arterioscler Thromb Vasc Biol 2001;21:1840–5

86. Pineda Torra I, Claudel T, Duval C, et al. Bile acidsinduce the expression of the human peroxisomeproliferator-activated receptor alpha gene via activa-tion of the farnesoid X receptor. Mol Endocrinol2003; 17:259–72

87. De Fabiani E, Mitro N, Gilardi F, et al. Coordinatedcontrol of cholesterol catabolism to bile acids and ofgluconeogenesis via a novel mechanism of transcrip-tion regulation linked to the fasted-to-fed cycle. J BiolChem 2003; 278:39124–32

88. Zhang Y, Castellani LW, Sinal CJ, et al. Peroxisomeproliferator-activated receptor-gamma coactivator1alpha (PGC-1alpha) regulates triglyceride metabo-lism by activation of the nuclear receptor FXR. GenesDev 2004; 18:157–69

89. Williams KJ, Fless GM, Petrie KA, et al. Mechanisms bywhich lipoprotein lipase alters cellular metabolism oflipoprotein(a), low density lipoprotein, and nascentlipoproteins. Roles for low density lipoprotein recep-tors and heparan sulfate proteoglycans. J Biol Chem1992; 267:13284–92


ch07 14/7/05 4:46 pm Page 79

90. Anisfeld AM, Kast-Woelbern HR, Meyer ME, et al.Syndecan-1 expression is regulated in an isoformspecific manner by the farnesoid-X receptor. J BiolChem 2003; 26:26

91. Dev S. Ethnotherapeutics and modern drug develop-ment the potential of Ayurveda. Curr Sci 1997;73:909–28

92. Satyavati GV, Dwarakanath C, Tripathi SN.Experimental studies on the hypocholesterolemiceffect of Commiphora mukul. Engl. (Guggul). Indian JMed Res 1969; 57:1950–62

93. Nityanand S, Kapoor NK. Cholesterol lowering activ-

ity of the various fractions of guggul. Indian J Exp Biol1973; 11:395–8

94. Urizar NL, Liverman AB, Dodds DT, et al. A naturalproduct that lowers cholesterol as an antagonistligand for FXR. Science 2002; 296:1703–6

95. Cui J, Huang L, Zhao A, et al. Guggulsterone is afarnesoid X receptor antagonist in coactivator associ-ation assays but acts to enhance transcription of bilesalt export pump. J Biol Chem 2003; 278:10214–20

96. Szapary PO, Wolfe ML, Bloedon LT, et al. Guggulipidfor the treatment of hypercholesterolemia: a random-ized controlled trial. JAMA 2003; 290:765–72


ch07 14/7/05 4:46 pm Page 80

INTRODUCTION

The concentration of cholesterol in highdensity lipoproteins (HDLs) has long beenknown to correlate inversely with the risk ofdeveloping premature coronary heartdisease.1–8 In animal studies this relationshiphas been shown to be one of cause and effect.Increasing the concentration of HDLs inseveral animal models leads to an inhibition ofatherosclerosis.9–11 The mechanism by whichHDLs protect is still uncertain, although severalof the known functions of these lipoproteinshave anti-atherogenic potential. Thesefunctions include the ability of HDL particles toact as extracellular acceptors of cholesterol inthe first step of the reverse cholesterol transport(RCT) pathway as well as anti-oxidant, anti-inflammatory and anti-thrombotic properties ofthese lipoproteins. The best-documentedfunction relates to the role of HDLs in RCT.

Reverse cholesterol transport is the termused to describe the transport of cholesterolfrom extrahepatic tissues to the liver, either forrecycling or for elimination from the bodythrough bile.12,13 Not only is this pathway funda-mental to normal cell function, but there is alsocompelling evidence that it is anti-atherogenic.The RCT pathway is facilitated and regulated byseveral factors that operate in plasma. Some ofthese factors are potential targets for therapiesdesigned to prevent atherosclerosis. In order toexploit this potential to its fullest, it is necessaryto understand how RCT operates, how it isregulated and which plasma factors should betargeted therapeutically.

This chapter provides an overview of HDLsand the pathway of RCT and identifies possibletargets for therapeutic intervention.

COMPOSITION AND STRUCTURE OFHDLS

High density lipoproteins are the smallest anddensest of the plasma lipoproteins. As with otherplasma lipoproteins, most of the HDLs in plasmaconsist of a hydrophobic core (mainly choles-teryl esters plus a small amount of triglyceride)surrounded by a surface monolayer consisting ofphospholipids, unesterified cholesterol andapolipoproteins. There are two main HDLapolipoproteins, apoA-I and apoA-II, that collec-tively account for about 90% of total. HumanHDLs also contain several other minorapolipoproteins, including apoA-IV, apoA-V, theC apolipoproteins, apoD, apoE, apoJ and apoL.In addition, HDLs act as a transport vehicle forseveral other proteins that are involved in plasmalipid metabolism and RCT. These includecholesteryl ester transfer protein (CETP),lecithin:cholesterol acyltransferase (LCAT) andphospholipid transfer protein (PLTP).

HDL SUBPOPULATIONS

The HDL fraction in human plasma is hetero-geneous, consisting of several distinct subpop-ulations of particles that differ in shape, size,density, composition and surface charge(Figure 1).

Overview of HDL and reverse cholesterol transportP. Barter

8

ch08 14/7/05 4:46 pm Page 81

Most of the HDLs circulating in normalhuman plasma are spherical, with discoidalparticles accounting for a very small percent-age of the total. When isolated on the basis ofdensity by ultracentrifugation, human HDLsseparate into two major subfractions, HDL2

(1.063 < d < 1.125 g/ml) and HDL3 (1.125 < d< 1.21 g/ml), and one minor subfraction, veryhigh density lipoproteins (VHDL) (1.21 < d <1.25 g/ml). Non-denaturing polyacrylamidegradient gel electrophoresis separates HDLson the basis of particle size into at least fivedistinct subpopulations of particles with diame-ters ranging from 7.6 to 10.6 nm.14

HDLs can also be divided into two mainsubpopulations on the basis of theirapolipoprotein composition. One subpopula-tion comprises HDLs containing apoA-I but noapoA-II (A-I HDLs), while another comprisesparticles containing both apoA-I and apoA-II(A-I/A-II HDLs).15,16 ApoA-I is distributedapproximately equally between A-I HDLs andA-I/A-II HDLs in most subjects, while almost allof the apoA-II is in A-I/A-II HDLs.16 Most of the

A-I/A-II HDLs are found in the HDL3 densityrange, while A-I HDLs are prominent compo-nents of both HDL2 and HDL3.

16

The HDL fraction is also heterogeneous interms of its electrophoretic mobility. Whensubjected to agarose gel electrophoresis, thehuman HDLs include subpopulations withalpha, pre-alpha, pre-beta and gamma migra-tion.17–20 Most human plasma HDLs are spheri-cal particles with an alpha mobility. The minorpool of pre-beta-migrating HDLs consistsmainly of discoidal apoA-I-containing particles.Gamma-migrating HDLs are discoidal, apoE-containing particles.

These HDL subpopulations are closely inter-related and are potentially interconvertible byfactors acting in the vascular compartment.

The functional implications of HDL hetero-geneity are still uncertain. There is evidencethat the preferred extracellular acceptor of cellcholesterol in the process mediated by theATP-binding cassette A1 (ABCA1) is a minorsubfraction of lipid-poor (or even lipid-free)apoA-I.21 There is also evidence that discoidalHDLs are the preferred substrates for LCAT,22

while larger, spherical HDLs are the preferredacceptors of cell cholesterol in the effluxprocess mediated by the scavenger receptortype B1 (SR-B1).23 But, overall, there is poorunderstanding of either functional or anti-atherogenic differences between the variousHDL subpopulations.

METABOLISM OF HDLS

Formation of HDLs

apoA-I and apoA-II originate mainly in the liver.Once in plasma they are rapidly lipidated. Aproportion of the lipid-free apoA-I and apoA-IIinteracts with ABCA1 in cell membranes,24 in aprocess that generates discoidal HDLs contain-ing apolipoproteins, phospholipids and a smallamount of unesterified cholesterol. DiscoidalHDLs are also generated from the interactionof lipid-poor apoA-I with redundant surfacecomponents that are shed from triglyceride-rich lipoproteins following hydrolysis of theirtriglyceride by lipoprotein lipase.25,26


HDLHDL2b2b,, HDLHDL2a2a,, HDLHDL3a3a,, HDLHDL3b3b,, HDLHDL3c3c, lipid-free apoA-I, lipid-free apoA-ISize and densitySize and density

Apolipoprotein compositionApolipoprotein composition

ShapeShape

apoA-I containing HDLsapoA-I containing HDLsapoA-I/apoA-II containing HDLsapoA-I/apoA-II containing HDLs

apoE containing HDLsapoE containing HDLs

Spherical HDLs Spherical HDLs Discoidal HDLsDiscoidal HDLs

Electrophoretic mobilityElectrophoretic mobilityAlpha-migrating HDLsAlpha-migrating HDLs

Pre-beta-migrating HDLsPre-beta-migrating HDLsGamma-migrating HDLsGamma-migrating HDLs

Pre-alpha-migrating HDLsPre-alpha-migrating HDLs

HDL HeterogeneityHDL heterogeneity

Figure 1 High density lipoproteins (HDLs) in humanplasma are heterogeneous in terms of shape, size,density, apolipoprotein composition and electrophoreticmobility

ch08 14/7/05 4:46 pm Page 82

Once formed, discoidal HDLs acquireadditional unesterified cholesterol that ispassively transferred either from other plasmalipoproteins or from cell membranes. Theunesterified cholesterol in A-I HDL discs isthen rapidly esterified by LCAT in a reactionthat provides the particle with a core of choles-teryl esters and converts the disc into a smallsphere.22 The rapidity with which LCAT cataly-ses cholesterol esterification in discoidal HDLsexplains why most of the HDLs in plasma arespherical rather than discoidal.

LCAT is also involved in the formation of A-I/A-II HDLs.27 It is probable that apoA-I andapoA-II are secreted into plasma separately andare assembled into A-I/A-II HDLs only aftertheir entry into the circulation. As with apoA-I,apoA-II is synthesized mainly (or exclusively) inthe liver and is most likely secreted in a lipid-free/lipid-poor form. Once in the extracellularspace, this hydrophobic protein acquiresphospholipids and unesterified cholesterolfrom cell membranes, probably in an ABCA1-dependent process, to form discoidal com-plexes containing apoA-II, phospholipids andunesterified cholesterol. However, unlikediscoidal A-I HDLs, discoidal A-II HDLs arenon-reactive with LCAT and are thus notconverted into spherical particles. Thepresence of apoA-II in spherical A-I/A-II HDLsdepends on a process in which discoidal A-IIHDLs fuse with A-I HDLs in a reaction cata-lysed by LCAT.27

The initial product of an interaction ofLCAT with discoidal A-I HDLs is a small spher-ical A-I HDL particle that contains the samenumber of apoA-I molecules (two) as theprecursor discoidal particles. Continuing activ-ity of LCAT results in an increase in the choles-teryl ester content of these small, spherical A-IHDLs. To accommodate the expanding parti-cle core, the HDL particle must acquireadditional apolipoproteins in the surfacemonolayer. This may be achieved in two ways.One mechanism involves fusion of the expand-ing particle with discoidal A-I HDLs to form alarger spherical A-I HDL in which the numberof molecules of apoA-I is also increased.28 Asecond mechanism involves fusion of the

expanding spherical A-I HDLs with discoidal A-II HDLs to form spherical A-I/A-II HDLs.27

Catabolism of HDL

Once formed, the subsequent metabolism ofspherical HDLs is complex. Most of the HDLconstituents are metabolized as discrete entitiesrather than being removed from the circulationas an uptake of intact HDL particles. Forexample, HDL-cholesteryl esters are eithertransferred to very low density lipoproteins(VLDLs) and low density lipoproteins (LDLs)by CETP29,30 or are selectively taken up by theliver in a process dependent on the binding ofHDLs to SR-B1 (also known as CLA-1).31 HDLtriglyceride and phospholipids are removedfrom the particle by hydrolysis in reactionscatalysed by lipases, including hepatic lipase(HL), endothelial lipase (EL) and secretoryphospholipase A2 (sPLA2). The apoA-I in HDLsmay also be independently metabolized

OVERVIEW OF HDL AND REVERSE CHOLESTEROL TRANSPORT 83

Lipid-poorLipid-poorapoA-IapoA-I

CETP

PLTP

HL

LCAT

DiscoidalDiscoidalHDLHDL

Excretion through kidneyExcretion through kidneyLarge sphericalLarge sphericalHDLHDL

Small sphericalSmall sphericalHDLHDL

LCAT

Large sphericalLarge sphericalHDLHDL

ABCA1

CECE

CE

Figure 2 Cycling of apolipoprotein (apo)A-I betweenlipid-poor and lipid-rich pools. Large, spherical highdensity lipoproteins (HDLs) are reduced in size bycholesteryl ester transfer protein (CETP), hepatic lipase(HL) and phospholipid transfer protein (PLTP) in aprocess that results in the dissociation of lipid-poor/lipid-free apoA-I from the particle. The dissociatedapoA-I may be excreted in urine, relipidated by theATP-binding cassette A1 (ABCA1) to form discoidalHDLs or re-incorporated into pre-existing HDLs as theyare increased in size by lecithin:cholesterolacyltransferase (LCAT)

ch08 14/7/05 4:46 pm Page 83

following its dissociation from the particleduring HDL remodelling.32,33 The presence ofapoA-II in the particle (e.g. in A-I/A-II HDLs)reduces this dissociation of apoA-I.34 Once disso-ciated from HDLs, lipid-poor/lipid-free apoA-Imay be excreted in urine35 or it may be recycledinto the HDL fraction as part of the continualremodelling of HDLs in plasma (Figure 2).

REMODELLING OF HDLS IN PLASMA

As outlined above, discoidal HDLs are rapidlyconverted into spherical particles by the actionof LCAT. Once formed, spherical HDLs arealso subject to continuous remodelling duringtheir circulation in plasma.

Factors that increase the size of sphericalHDLs

Small, spherical A-I HDLs retain a degree ofreactivity with LCAT and continue to acquirecholesteryl esters in an expanding particlecore. Additional surface constituents (includ-ing apolipoproteins) are then required toaccommodate the expanded lipoprotein core.Additional apoA-I is provided, either in theform of lipid-free apolipoprotein33 or in aprocess involving the fusion of the particle withan A-I HDL disc.28 Additional apolipoproteinmay also be provided in the form of apoA-II byfusion of the expanding particle with an A-IIHDL disc.27 Additional phospholipids andunesterified cholesterol are acquired as trans-fers from other plasma lipoproteins or as diffu-sion from cell membranes.

The size of spherical, A-I HDLs may also beincreased by PLTP in a process of particlefusion that is accompanied by dissociation oflipid-poor apoA-I from the fusion product.36,37

Remodelling of HDLs by PLTP is enhanced bythe presence of triglyceride in the particle.38

Factors that decrease the size ofspherical HDL

The combined activities of CETP and HL inthe presence of triglyceride-rich lipoproteinsare especially effective in reducing the particle

size of HDLs.39 CETP promotes a net masstransfer of cholesteryl esters from HDLs totriglyceride-rich lipoproteins in exchange fortriglyceride that is transferred into HDLs. Thisexchange results in formation of HDLs that aredepleted of cholesteryl esters and enriched intriglyceride. The triglyceride enrichment notonly enhances the interaction of HDLs withPLTP but also provides the particles with thepreferred substrate for HL. When HL hydroly-ses the newly acquired HDL triglyceride, theconsequent reduction in HDL core volume isaccompanied by a decrease in HDL particlesize and dissociation of lipid-free/lipid-poorapoA-I from the particle39 (Figure 2).

Metabolism of lipid-free/lipid-poorapoA-I in plasma

As outlined above, one byproduct of theremodelling of spherical HDLs by PLTP, CETPand HL is a pool of lipid-poor (or lipid-free)apoA-I. A small proportion of this newly gener-ated apoA-I may be excreted in urine,35

although most appears to be re-incorporatedinto the HDL fraction. There are at least threemechanisms by which lipid-poor apoA-I returnsto the HDL fraction. The simplest involvesincorporation directly into pre-existing spheri-cal HDLs that are expanding as a consequenceof an interaction with LCAT.33 ApoA-I may alsobe relipidated by acquiring phospholipids andunesterified cholesterol from cells in theABCA1-mediated process21 to form newdiscoidal HDL particles. Third, apoA-I mayform discoidal complexes with phospholipidsreleased as redundant surface constituentsfrom triglyceride-rich lipoproteins followinghydrolysis of their triglyceride by lipoproteinlipase.40 Of all these potential metabolic fatesof lipid-poor apoA-I, the most important physi-ologically is probably the interaction withABCA1 in the first step of RCT.

HDL AND REVERSE CHOLESTEROLTRANSPORT

Reverse cholesterol transport involves severaldiscrete steps:


ch08 14/7/05 4:46 pm Page 84

(1) Unesterified cholesterol in cell mem-branes is transferred to acceptors in theextracellular space where it becomes incor-porated into HDL particles;

(2) The unesterified cholesterol in HDL isthen either delivered directly to the liver(and possibly other tissues) or is convertedinto cholesteryl esters by LCAT;

(3) The newly formed HDL cholesteryl estersmay then either be delivered directly to theliver (and steroidogenic tissues) in aprocess dependent on SR-B1 (CLA-1) orbe transferred by CETP to VLDL/LDL;

(4) The cholesteryl esters in VLDL/LDL aredelivered to the liver as a component ofthe receptor-mediated uptake of LDL.

Efflux of cellular unesterified cholesterolto extracellular acceptors

There are at least three distinct processes thatpromote the efflux of cholesterol from cellmembranes to acceptors in the extracellularspace. One involves ABCA1, another ABCG1, athird involves SR-B1 and a fourth involvespassive diffusion (Figure 3).

ABCA1 translocates phospholipids andcholesterol from the inner to the outer leafletsof cell membranes where they are picked up byapoA-I (or apoA-II) in the extracellularspace.21,41,42 This interaction is limited toapolipoproteins that contain no or very littlelipid. Precisely how lipid-poor apolipoproteinsremove cholesterol from cells is not known.According to one view, there is a simultaneoustransfer of both cellular phospholipids andcholesterol to the apolipoproteins in a processthat results in formation of discoidalcomplexes containing apolipoproteins, phos-pholipids and unesterified cholesterol. Analternative view holds that the lipid-poorapolipoproteins first acquire phospholipidsfrom cells in an ABCA1-mediated process thatgenerates discoidal complexes of apolipopro-tein and phospholipid. These discoidal parti-cles are then able to accept unesterifiedcholesterol from cell membranes in a diffusionprocess that may not require ABCA1.44 Onceformed by either mechanism, the discoidal

complexes of apolipoprotein, phospholipidsand unesterified cholesterol (nascent HDLs)are subsequently transported, via lymphatics,to the plasma compartment where they arefurther metabolised as outlined above.

ABCG1 promotes the transfer of cholesterolfrom cells, including macrophages, topreformed, spherical HDL in the extracellularspace.45,46

SR-BI is an HDL receptor.47 Its best knownfunction is to promote the selective hepaticuptake of HDL cholesteryl esters.31 However, ithas also been implicated in the efflux ofunesterified cholesterol from cells to HDLs.48,49

In contrast to the efflux promoted by ABCA1,SR-B1 mediates the bidirectional transfer ofunesterified cholesterol between cells andHDLs and promotes a net efflux only if there isa concentration gradient of unesterifiedcholesterol from the donor cell to the acceptorHDL. Larger, spherical HDLs are preferred as


Discoidal HDL

Cell membrane

Small spherical HDL

Larger spherical HDL UCDiffusion

LCAT

LCAT

Lipid-poor apoA-I

SR-B1

Extracellular space

UCDiffusion

SR-B1

UCDiffusion

SR-B1

UCABCA1

CE

CE

Figure 3 Efflux of unesterified cholesterol (UC) fromcells. UC in cell membranes is transferred (withphospholipids) to lipid-free apolipoprotein (apoA)-I inthe extracellular space to form discoidal high densitylipoproteins (HDLs). The discoidal HDLs aresequentially converted into small and then largerspherical HDLs by lecithin:cholesterol acyltransferase(LCAT). The discoidal and the spherical HDLs acquireadditional UC from cells either by passive diffusion orby diffusion facilitated by scavenger receptor type B1(SR-B1). ABCG1 promotes the active transfer of cellcholesterol to spherical HDLs.

ch08 14/7/05 4:46 pm Page 85

acceptors in the efflux promoted by SR-B1.23

The contribution made by SR-B1 to overallRCT is uncertain.

Mature, spherical HDLs in plasma alsoaccept unesterified cholesterol efflux fromcells in a process of passive aqueous diffusionthat requires neither ABCA1 nor SR-B1.50

Unesterified cholesterol in cell membranes isspontaneously released into the aqueous,extracellular space where it collides with andincorporates into any preformed HDL particlesthat are present. This is a bidirectional processin which unesterified cholesterol exchangesbetween HDLs and cell membranes. However,a net transfer of cholesterol into HDLs may beachieved by the formation of a concentrationgradient generated by LCAT-mediated esterifi-cation of cholesterol on the HDL surface. Therelative contribution made by these threeefflux processes to overall RCT is not known.

Once cell cholesterol has been transferred toextracellular acceptors and ultimately incorpo-rated into HDLs, it may be transported to theliver for elimination from the body by severalpathways.

Delivery of HDL-cholesterol to the liver(Figure 4)

The unesterified cholesterol incorporated intoHDLs may be taken up directly by the liver in aprocess that probably involves binding of HDLsto SR-B1 (pathway 1 in Figure 4). Alternatively, itmay be converted by LCAT into cholesterylesters that are then delivered to the liver byeither a direct (pathway 2 in Figure 4) or indirect(pathway 3 in Figure 4) pathway. The directpathway involves binding of HDL to hepatic SR-B1 (CLA-1).47 Binding to SR-B1 results in a selec-tive uptake of cholesteryl esters from HDLs,31

leaving core-lipid-depleted particles to return tothe circulation where they are available to partic-ipate further in the RCT process. In the indirectpathway, HDL cholesteryl esters are transferredby CETP to the VLDL/LDL fraction and thendelivered to the liver as a consequence of thereceptor-mediated uptake of LDLs.

The relative contribution made by each ofthese various pathways is not known. On the basis

of mathematical modelling following injectionsinto humans of HDL radiolabelled in theunesterified cholesterol moiety, it has beenconcluded that an hepatic uptake of HDLunesterified cholesterol represents the majorpathway by which peripheral cholesterol isreturned to the liver.50 Another view holds thatthe predominant fate of unesterified cholesterolin HDL is to be esterified by LCAT, with a subse-quent delivery to the liver being achieved as anuptake of cholesteryl esters. In this latter case, it isnot known what proportion of the HDL choles-teryl esters is subsequently taken up directly bythe liver via the SR-B1 pathway and what propor-tion is first transferred by CETP to other lipopro-teins before being taken up by the liver.

It is important to discover the contributionof each of these pathways to overall RCT. If, forexample, the predominant pathway is viaCETP-mediated transfers of cholesteryl estersfrom HDLs to the VLDL/LDL pool, it followsthat therapeutic inhibition of CETP maycompromise RCT. If, on the other hand, the


HDL

VLDL/LDL

CECE

Liver

Bile

CECE

CETP

SR-B1

LDL-RLDL-R

CE

UCUC

UCUC

(1)

(3)

(3)

SR-B1

LCAT

Extrahepatic

tissues

UC

(2)

Figure 4 Pathways for delivery of cholesterol fromextrahepatic tissues to the liver. Following theincorporation of cellular unesterified cholesterol (UC)into HDL, it may be delivered directly to the liver in aprocess involving binding of HDLs to hepatic scavengerreceptor type B1 (SR-B1) (pathway 1) or it may beesterified by lecithin:cholesterol acyltransferase (LCAT)to form cholesteryl esters (CE). The HDL CE may thenbe delivered directly to the liver, again in a processinvolving binding of HDLs to hepatic SR-B1 (pathway2), or it may first be transferred by cholesteryl estertransfer protein (CETP) to the VLDL/LDL fraction andthen delivered to the liver as a component of the LDLreceptor-mediated uptake of LDLs (pathway 3)

ch08 14/7/05 4:46 pm Page 86

major pathway involves the direct SR-B1-mediated hepatic uptake of either HDLunesterified cholesterol or HDL cholesterylesters, inhibition of CETP would be predictedto be anti-atherogenic by virtue of partitioningcholesterol in HDLs at the expense of LDLswithout compromising RCT. Further researchin this area is clearly required.

POTENTIAL MECHANISMS BYWHICH HDLS PROTECT AGAINSTATHEROSCLEROSIS

HDL-mediated cell cholesterol efflux

There is compelling evidence that HDLscounteract the effects of elevated levels of LDLsby removing cholesterol from cells in the arterialintima. In cultured fibroblasts harvested fromhumans with mutations of ABCA1, ABCA1-mediated efflux of cholesterol correlatespositively and significantly with the concentra-tion of HDL-cholesterol in the plasma of thefibroblast donor.52 Another study of apoA-I andHDL-mediated efflux in carriers of ABCA1mutations revealed a positive correlation ofefflux with both HDL-cholesterol levels inplasma and carotid intima-media thickness.53

These observations suggest a direct link betweencholesterol efflux, HDL levels and the develop-ment of atherosclerosis. Further direct evidencethat cholesterol efflux is anti-atherogenic hasbeen provided in mice in studies in which selec-tive expression of ABCA1 in macrophages wasfound to be anti-atherogenic, in this case withoutany change in HDL concentration in plasma.54

Anti-oxidant properties of HDLs

LDL oxidation within the arterial intima is anearly event in atherogenesis.55 The ability ofHDLs to inhibit this oxidation is thus potentiallyanti-atherogenic. HDLs inhibit the transmigra-tion of monocytes induced by oxidized LDLs,56

the cytotoxicity induced by oxidized LDLs57 andthe oxidized LDL-induced adhesion ofmonocytes to endothelial cells.58 HDLs alsoinhibit LDL oxidation in vitro59,60 and in vivo.61

Paraoxonase may be involved in this effect of

HDLs,62–64 although apoA-I and apoA-II have alsobeen shown to have anti-oxidant properties.65

The importance of these anti-oxidant propertiesof HDLs to their anti-atherogenic function invivo in humans remains to be determined.

HDL-mediated inhibition of adhesionmolecule expression

HDLs inhibit the cytokine-induced expression ofVCAM-1, ICAM-1 and E-selectin in endothelialcells growing in tissue culture in a concentration-dependent manner66 in a process that mayinvolve an HDL-mediated inhibition of endothe-lial cell sphingosine kinase.67 HDL-mediatedinhibition of VCAM-I and E-selectin proteinexpression is paralleled by significant reductionsin the steady-state mRNA levels of these adhesionmolecules.66 The in vivo significance of thesefindings has been supported by the finding of asignificant reduction of plasma sICAM-1 and sE-selectin concentrations in patients whose HDLlevels were increased by treatment with fenofi-brate.68 It remains to be determined whether thisanti-inflammatory property of HDLs contributesto their ability to protect against atherosclerosis.

Stimulation of endothelial NOproduction

Another potential mechanism by which HDLsprotect relates to their ability to modulateendothelial function.69 The probable mecha-nism of this effect is a stimulation of theproduction of endothelial NO, a known anti-atherogenic mediator.70 An effect of HDL onendothelial function has also been demon-strated in vivo in human studies in which short-term intravenous infusion of reconstitutedHDLs (rHDLs) containing apoA-I complexedwith phospholipids normalized endothelial-dependent vasodilatation.71

INTERVENTION STUDIES WITHHDLS

Experimental animals

There is clear evidence of a direct anti-athero-genic of HDLs in experimental animals.


ch08 14/7/05 4:46 pm Page 87

Weekly infusions of HDLs into cholesterol-fedrabbits have been shown to reduce aortic lipiddeposition and promote regression of aorticatherosclerosis.72 A similar beneficial effect hasbeen observed in rabbits infused with reconsti-tuted HDL (rHDL) containing apoA-IMilano

complexed with phospholipid.73

Additional compelling evidence of a directcardio-protective effect of HDLs has beenprovided by studies of genetically modifiedanimals. Overexpression of the human apoA-Igene in rabbits9 and mice10,11 increases theconcentration of HDL-cholesterol and protectsagainst diet-induced atherosclerosis10 and thespontaneous atherosclerosis that develops inLDL-receptor-deficient,74 apoE-deficient11 andLP(a)75 transgenic animals. This effect is notconfined to overexpression of apoA-I. Miceoverexpressing the combination of human apoA-I and apoA-II76 or human apoA-II alone77 are alsoprotected. In contrast, mice overexpressingmurine (as distinct from human) apoA-II have anincreased susceptibility to atherosclerosis.78

Human studies

Until recently, the evidence that HDL inhib-ited atherosclerosis in humans was circumstan-tial, depending on the analysis of results fromintervention trials in which an increase in HDLconcentration appeared to be a predictor ofbenefit that was independent of effects onother lipoprotein fractions.79–81 However, thereis now compelling direct evidence that HDL isanti-atherogenic in humans. Human subjectswith coronary atherosclerosis received weeklyintravenous injections of rHDL (complexes ofapoA-IMilano and phospholipid) for 5 weeks.Coronary atherosclerosis was assessed byintravascular ultrasound. In subjects receivingthe infusions of rHDL there was a statisticallysignificant reduction in the atherosclerosisburden after only 5 weeks of treatment.82

There is now a clear need to repeat this studyin a larger cohort of subjects. Confirmation ofthe result will provide the stimulus for amassive research effort directed towards thedevelopment of new strategies designed toincrease HDL concentration.

POTENTIAL NEW THERAPIES

Current therapies for raising the HDL concen-tration are, at best, only moderately effective.PPAR-α activists such as fibrates raise HDL-cholesterol levels by 5–25%. This is achieved byat least two mechanisms: directly by stimulatingthe synthesis of apoA-I and apoA-II andindirectly by increasing lipoprotein lipase.83,84

Statins also increase HDL-cholesterol, both byincreasing apoA-I synthesis85 and by inhibitingCETP,86 although the effect is relatively smallwith increases tending to be less than 10%.Nicotinic acid is the most effective of currentlyavailable agents, raising HDL levels by up to30% by a mechanism that remains to be deter-mined.

The mounting evidence in both animal andhuman studies that raising HDL levels retardsor even reverses the progression of atheroscle-rosis has stimulated great interest in develop-ing new therapies designed to increase theconcentration of HDLs and to enhance thepathway of RCT. Potential approaches toraising HDLs include:

(1) More potent PPAR-α agonists and otheragents designed to increase the synthesis ofapoA-I;

(2) Therapies with the potential to increasethe activities of ABCA1, LCAT and lipopro-tein lipase and

(3) Therapies that inhibit CETP.

It should be noted, however, that an elevationof HDL concentration might not automaticallytranslate into an increased protection againstatherosclerosis.

It may also be possible to increase RCT bystrategies that could be accompanied by aparadoxical decrease in the concentration ofHDL. One such approach would be activationof SR-B1 in a process that simultaneouslyincreases the delivery of HDL-cholesterol tothe liver and reduces the concentration ofHDL-cholesterol in plasma. This latterapproach is underpinned by the concept thatHDL function is more important than HDL concentration in protecting against


ch08 14/7/05 4:46 pm Page 88

atherosclerosis. The problem here is that it isstill not known which of the documentedfunctions of HDLs is responsible for theirability to inhibit atherosclerosis.

CONCLUSIONS

With a growing understanding of the metab-olism and regulation of HDLs and their

involvement in several potential anti-athero-genic processes, there has been a majorinterest in exploiting the knowledge with aview to developing novel approaches for theprevention of atherosclerosis. The potentialof these new approaches is considerable, butonly time will tell whether the potential will translate into safe and effective newtherapies.


1. Gordon DJ, Knoke J, Probstfield JL, et al. High-density lipoprotein cholesterol and coronary heartdisease in hypercholesterolemic men: the LipidResearch Clinics Coronary Primary Prevention Trial.Circulation 1986; 74:1217–25

2. Enger SC, Hjermann I, Foss OP, et al. High densitylipoprotein cholesterol and myocardial infarction orsudden coronary death: a prospective case-controlstudy in middle-aged men of the Oslo study. Artery1979; 5:170–81

3. Miller NE, Thelle DS, Forde OH, et al. The Tromsoheart-study. High-density lipoprotein and coronaryheart-disease: a prospective case-control study. Lancet1977; 1:965–8

4. Goldbourt U, Medalie JH. High density lipoproteincholesterol and incidence of coronary heart disease –the Israeli Ischemic Heart Disease Study. Am JEpidemiol 1979; 109:296–308

5. Gordon T, Castelli WP, Hjortland MC, et al. Highdensity lipoprotein as a protective factor againstcoronary heart disease. The Framingham Study. Am JMed 1977; 62:707–14

6. Jacobs DR Jr, Mebane IL, Bangdiwala SI, et al. Highdensity lipoprotein cholesterol as a predictor ofcardiovascular disease mortality in men and women:the follow-up study of the Lipid Research ClinicsPrevalence Study. Am J Epidemiol 1990; 131:32–47

7. Miller M, Seidler A, Kwiterovich PO, et al. Long-termpredictors of subsequent cardiovascular events withcoronary artery disease and ‘desirable’ levels ofplasma total cholesterol. Circulation 1992;86:1165–70

8. Pekkanen J, Linn S, Heiss G, et al. Ten-year mortalityfrom cardiovascular disease in relation to cholesterollevel among men with and without preexisting cardio-vascular disease. N Engl J Med 1990; 322:1700–7

9. Duverger N, Kruth H, Emmanuel F, et al. Inhibitionof atherosclerosis development in cholesterol-fedhuman apolipoprotein A-I-transgenic rabbits.Circulation 1996; 94:713–17

10. Rubin EM, Krauss RM, Spangler EA, et al. Inhibition

of early atherogenesis in transgenic mice by humanapolipoprotein AI. Nature 1991; 353:265–7

11. Plump AS, Scott CJ, Breslow JL. Human apolipopro-tein A-I gene expression increases high densitylipoprotein and suppresses atherosclerosis in theapolipoprotein E-deficient mouse. Proc Natl Acad SciUSA 1994; 91:9607–11

12. Barter PJ. HDL and reverse cholesterol transport.Curr Opin Lipidol 1993; 4:210–17

13. Fielding CJ, Fielding PE. Molecular physiology ofreverse cholesterol transport. J Lipid Res 1995;36:211–28

14. Blanche PJ, Gong EL, Forte TM, et al. Characteriza-tion of human high-density lipoproteins by gradientgel electrophoresis. Biochim Biophys Acta 1981;665:408–19

15. Cheung MC, Albers JJ. Distribution of high densitylipoprotein particles with different apoproteincomposition: particles with A-I and A-II and particleswith A-I but no A-II. J Lipid Res 1982; 23:747–53

16. Cheung MC, Albers JJ. Characterization of lipopro-tein particles isolated by immunoaffinity chromatog-raphy. Particles containing A-I and A-II and particlescontaining A-I but no A-II. J Biol Chem 1984;259:12201–9

17. Asztalos BF, Sloop CH, Wong L, et al. Two-dimen-sional electrophoresis of plasma lipoproteins: recog-nition of new apo A-I-containing subpopulations.Biochim Biophys Acta 1993; 1169:291–300

18. Huang Y, von Eckardstein A, Wu S, et al. Effects of theapolipoprotein E polymorphism on uptake and trans-fer of cell-derived cholesterol in plasma. J Clin Invest1995; 96:2693–701

19. Kunitake ST, La Sala KJ, Kane JP. Apolipoprotein A-I-containing lipoproteins with pre-beta electrophoreticmobility. J Lipid Res 1985; 26:549–55

20. Sparks DL, Lund-Katz S, Phillips MC. The charge andstructural stability of apolipoprotein A-I in discoidaland spherical recombinant high density lipoproteinparticles. J Biol Chem 1992; 267:25839–47

21. Oram JF. HDL Apolipoproteins and ABCA1. Partners

References

ch08 14/7/05 4:46 pm Page 89

in the removal of excess cellular cholesterol.Arterioscler Thromb Vasc Biol 2003; 23:720–7

22. Jonas A, von Eckardstein A, Kezdy KE, et al. Structuraland functional properties of reconstituted highdensity lipoprotein discs prepared with sixapolipoprotein A-I variants. J Lipid Res 1991; 32:97–106

23. Liadaki KN, Liu T, Xu S, et al. Binding of high densitylipoprotein (HDL) and discoidal reconstituted HDLto the HDL receptor scavenger receptor class B typeI. Effect of lipid association and APOA-I mutations onreceptor binding. J Biol Chem 2000; 275:21262–71

24. Lawn RM, Wade DP, Garvin MR, et al. The Tangierdisease gene product ABC1 controls the cellularapolipoprotein-mediated lipid removal pathway. JClin Invest 1999; 104:R25–R31

25. Redgrave TG, Small DM. Quantitation of the transferof surface phospholipid of chylomicrons to the highdensity lipoprotein fraction during the catabolism ofchylomicrons in the rat. J Clin Invest 1979; 64:162–71

26. Tall AR, Blum CB, Forester GP, et al. Changes in thedistribution and composition of plasma high densitylipoproteins after ingestion of fat. J Biol Chem 1982;257:198–207

27. Clay MA, Pyle DH, Rye K-A, et al. Formation of spher-ical, reconstituted high density lipoproteins contain-ing both apolipoproteins A-I and A-II is mediated bylecithin:cholesterol acyltransferase. J Biol Chem 2000;275:9019–25

28. Liang H-Q, Rye K-A, Barter PJ., Remodelling of recon-stituted high density lipoproteins by lecithin: choles-terol acyltransferase. J Lipid Res 1996; 37:1962–70

29. Barter PJ, Hopkins GJ, Calvert GD. Pathways for theincorporation of esterified cholesterol into very lowdensity and low density lipoproteins in plasmaincubated in vitro. Biochim Biophys Acta 1982;713:136–48

30. Tall AR. Plasma cholesteryl ester transfer protein. JLipid Res 1993; 34:1255–74

31. Glass C, Pittman RC, Weinstein DB, et al. Dissociationof tissue uptake of cholesterol ester from that ofapoprotein A-I of rat plasma high density lipoprotein:selective delivery of cholesterol ester to liver, adrenal,and gonad. Proc Natl Acad Sci USA 1983; 80:5435–9

32. Liang HQ, Rye KA, Barter PJ. Dissociation of lipid-free apolipoprotein A-I from high density lipopro-teins. J Lipid Res 1994; 35:1187–99

33. Liang HQ, Rye KA, Barter PJ. Cycling of apolipopro-tein A-I between lipid-associated and lipid-free pools.Biochim Biophys Acta 1995; 1257:31–7

34. Rye K-A, Wee K, Curtiss L, et al. Apolipoproteins A-IIinhibits high density lipoprotein remodeling andlipid-poor apolipoproteins A-I formation. J Biol Chem2003; 278:22530–6

35. Horowitz BS, Goldberg IJ, Merab J, et al. Increasedplasma and renal clearance of an exchangeable poolof apolipoprotein A-I in subjects with low levels ofhigh density lipoprotein cholesterol. J Clin Invest1993; 91:1743–52

36. van Tol A. Phospholipid transfer protein. Curr OpinLipidol 2002; 13:135–9

37. Lusa S, Jauhiainen M, Metso J, et al. The mechanismof human plasma phospholipid transfer protein-induced enlargement of high-density lipoproteinparticles: evidence for particle fusion. Biochem J1996; 313:275–82

38. Rye KA, Jauhiainen M, Barter PJ, et al. Triglyceride-enrichment of high density lipoproteins enhancestheir remodelling by phospholipid transfer protein. JLipid Res 1998; 39:613–22

39. Rye KA, Clay MA, Barter PJ. Remodelling of highdensity lipoproteins by plasma factors. Atherosclerosis1999; 145:227–38

40. Clay MA, Barter PJ. Formation of new HDL particlesfrom lipid-free apolipoprotein A-I. J Lipid Res 1996;7:1722–32

41. Wang N, Tall AR. Regulation and mechanism of ATP-binding cassette transporter A-mediated cellularcholesterol efflux. Arterioscler Thromb Vasc Biol2003; 23:1178–84

42. Francis GA, Knopp RH, Oram JF. Defective removalof cellular cholesterol and phospholipids byapolipoprotein A-I in Tangier disease. J Clin Invest1995; 96:78–87

43. Oram JF, Lawn RM. ABCA1: the gatekeeper for elimi-nating excess tissue cholesterol. J Lipid Res 2001;42:1173–9

44. Wang N, Lan D, Chen W, et al. ATP-binding cassettetransporters G1 and G4 mediate cellular cholesterolefflux to high-density lipoproteins. Proc Natl Acad SciUSA 2004; 101:9774–9.

45. Nakamura K, Kennedy MA, Baldan A, et al.Expression and regulation of multiple murine ATP-binding cassette transporter G1 mRNAs/isoformsthat stimulate cellular cholesterol efflux to highdensity lipoprotein. J Biol Chem 2004; 279:45980–9.

46. Fielding PE, Nagao K, Hakamata H, et al. A two-stepmechanism for free cholesterol and phospholipidefflux from human vascular cells to apolipoprotein A-1. Biochemistry 2000; 39:14113–20

47. Acton S, Rigotti A, Landschulz KT, et al. Identi-fication of scavenger receptor SR-BI as a high densitylipoprotein receptor. Science 1996; 271:518–20

48. Ji Y, Jian B, Wang N, et al. Scavenger receptor BIpromotes high density lipoprotein-mediated cellularcholesterol efflux. J Biol Chem 1997; 272:20982–5

49. Gu X, Kozarsky K, Krieger M. Scavenger receptorclass B, type I-mediated [3H]cholesterol efflux tohigh and low density lipoproteins is dependent onlipoprotein binding to the receptor. J Biol Chem2000; 275:29993–30001

50. Yancey PG, Bortnick AE, Kellner-Weibel G, et al.Importance of different pathways of cellular choles-terol efflux. Arterioscler Thromb Vasc Biol 2003;23:712–19

51. Schwartz CC, Zech LA, VandenBroek JM, et al.Cholesterol kinetics in subjects with bile fistula.


ch08 14/7/05 4:46 pm Page 90

Positive relationship between size of the bile acidprecursor pool and bile acid synthetic rate. J ClinInvest 1993; 91:923–38

52. Clee SM, Kastelein JJ, van Dam M, et al. Age and resid-ual cholesterol efflux affect HDL cholesterol levelsand coronary artery disease in ABCA1 heterozygotes.J Clin Invest 2000; 106:1263–70

53. Attie AD, Kastelein JP, Hayden MR. Pivotal role ofABCA1 in reverse cholesterol transport influencingHDL levels and susceptibility to atherosclerosis. JLipid Res 2001; 42:1717–26

54. Aiello RJ, Brees, D, Bourassa P-A, et al. Increasedatherosclerosis in hyperlipidemic mice with inactiva-tion of ABCA1 in macrophages. Arterioscler ThrombVasc Biol 2002; 22:630–7

55. Steinberg D, Parthasarathy S, Carew TE, et al. Beyondcholesterol. Modifications of low-density lipoproteinthat increase its atherogenicity. N Engl J Med 1989;320:915–24

56. Navab M, Imes SS, Hama SY, et al. Monocyte transmi-gration induced by modification of low densitylipoprotein in cocultures of human aortic wall cells isdue to induction of monocyte chemotactic protein 1synthesis and is abolished by high density lipoprotein.J Clin Invest 1991; 88:2039–46

57. Hessler JR, Robertson AL Jr, Chisolm GM III. LDL-induced cytotoxicity and its inhibition by HDL inhuman vascular smooth muscle and endothelial cellsin culture. Atherosclerosis 1979; 32:213–29

58. Maier JA, Barenghi L, Bradamante S, et al.Modulators of oxidized LDL-induced hyperadhesive-ness in human endothelial cells. Biochem BiophysRes Commun 1994; 204:673–7

59. Decossin C, Tailleux A, Fruchart JC, et al. Preventionof in vitro low-density lipoprotein oxidation by analbumin-containing Lp A-I subfraction. BiochimBiophys Acta 1995; 1255:31–8

60. Kunitake ST, Jarvis MR, Hamilton RL, et al. Bindingof transition metals by apolipoprotein A-I-containingplasma lipoproteins: inhibition of oxidation of lowdensity lipoproteins. Proc Natl Acad Sci USA 1992;89:6993–7

61. Klimov AN, Gurevich VS, Nikiforova AA, et al.Antioxidative activity of high density lipoproteins invivo. Atherosclerosis 1993; 100:13–18

62. Mackness MI, Arrol S, Abbott C, et al. Protection oflow-density lipoprotein against oxidative modificationby high-density lipoprotein associated paraoxonase.Atherosclerosis 1993; 104:129–35

63. Mackness MI, Arrol S, Durrington PN. Paraoxonaseprevents accumulation of lipoperoxides in low-densitylipoprotein. FEBS Lett 1991; 286:152–4

64. Watson AD, Berliner JA, Hama SY, et al. Protectiveeffect of high density lipoprotein associated paraox-onase. Inhibition of the biological activity ofminimally oxidized low density lipoprotein. J ClinInvest 1995; 96:2882–91

65. Garner B, Waldeck AR, Witting PK, et al. Oxidation ofhigh density lipoproteins. II. Evidence for direct

reduction of lipid hydroperoxides by methionineresidues of apolipoproteins AI and AII. J Biol Chem1998; 273:6088–95

66. Cockerill GW, Rye KA, Gamble JR, et al. High-densitylipoproteins inhibit cytokine-induced expression ofendothelial cell adhesion molecules. ArteriosclerThromb Vasc Biol 1995; 15:1987–94

67. Xia P, Vadas MA, Rye KA, et al. High density lipopro-teins (HDL) interrupt the sphingosine kinase signal-ing pathway. A possible mechanism for protectionagainst atherosclerosis by HDL. J Biol Chem 1999;274:33143–7

68. Calabresi L, Gomaraschi M, Villa B, et al. Elevatedsoluble cellular adhesion molecules in subjects withlow HDL-cholesterol. Arterioscler Thromb Vasc Biol2002; 22:656–61

69. O’Connell BJ, Genest J Jr. High-density lipoproteinsand endothelial function. Circulation 2001; 104:1978–83

70. Yuhanna IS, Zhu Y, Cox BE, et al. High-densitylipoprotein binding to scavenger receptor-BI activatesendothelial nitric oxide synthase. Nat Med 2001;7:853–7

71. Spieker LE, Sudano I, Hurlimann D, et al. High-density lipoprotein restores endothelial function inhypercholesterolemic men. Circulation 2002;105:1399–402

72. Badimon JJ, Badimon L, Fuster V. Regression ofatherosclerotic lesions by high density lipoproteinplasma fraction in the cholesterol-fed rabbit. J ClinInvest 1990; 85:1234–41

73. Chiesa G, Monteggia E, Marchesi M, et al.Recombinant apolipoprotein A-I Milano infusion intorabbit carotid artery rapidly removes lipid from fattystreaks. Circ Res 2002; 90:974–80

74. Tangirala RK, Tsukamoto K, Chun SH, et al.Regression of atherosclerosis induced by liver-directed gene transfer of apolipoprotein A-I in mice.Circulation 1999; 100:1816–22

75. Liu AC, Lawn RM, Verstuyft JG, et al. Humanapolipoprotein A-I prevents atherosclerosis associatedwith apolipoprotein[a] in transgenic mice. J LipidRes 1994; 35:2263–7

76. Schultz JR, Verstuyft JG, Gong EL, et al. Protein compo-sition determines the anti-atherogenic properties ofHDL in transgenic mice. Nature 1993; 365:762–4

77. Tailleux A, Bouly M, Luc G, et al. Decreased suscepti-bility to diet-induced atherosclerosis in humanapolipoprotein A-II transgenic mice. ArteriosclerThromb Vasc Biol 2000; 20:2453–8

78. Warden CH, Hedrick CC, Qiao JH, et al.Atherosclerosis in transgenic mice overexpressingapolipoprotein A-II. Science 1993; 261:469–72

79. The Lipid Research Clinics Coronary PrimaryPrevention Trial results. I. Reduction in incidence ofcoronary heart disease. JAMA 1984; 251:351–64

80. Pedersen TR, Olsson AG, Faergeman O, et al.Lipoprotein changes and reduction in the incidenceof major coronary heart disease events in the


ch08 14/7/05 4:46 pm Page 91

Scandinavian Simvastatin Survival Study (4S).Circulation. 1998; 97:1453–60

81. Manninen V, Tenkanen L, Koskinen P, et al. Jointeffects of serum triglyceride and LDL cholesterol andHDL cholesterol concentrations on coronary heartdisease risk in the Helsinki Heart Study. Implicationsfor treatment. Circulation 1992; 85:37–45

82. Nissen SE, Tsunoda T, Tuzcu EM, et al. Effect of recom-binant ApoA-I Milano on coronary atherosclerosis inpatients with acute coronary syndromes: a randomizedcontrolled trial. JAMA 2003; 290:2292–300

83. Fruchart J-C, Staels B, Duriez P. The role of fibric

acids in atherosclerosis. Curr Athero Rep 2001;3:83–92


85. Martin G, Duez H, Blanquart C, et al. Statin-inducedinhibition of the Rho-signaling pathway activatesPPARalpha and induces HDL apoA-I. J Clin Invest2001; 107:1423–32

86. McPherson R. Comparative effects of simvastatin andcholestyramine on plasma lipoproteins and CETP inhumans. Can J Clin Pharmacol 1999; 6:85–90


ch08 14/7/05 4:46 pm Page 92

BACKGROUND: NUCLEARRECEPTORS AS REGULATORS OFMETABOLISM

Nuclear receptors comprise a superfamily ofligand-dependent transcription factors thatregulate genetic networks controlling cellgrowth, development and metabolism. Con-sisting of over 100 different proteins (48 in thehuman genome), the superfamily includes thewell-known receptors for steroids, thyroidhormones and vitamins.1 Members of thenuclear receptor superfamily are characterizedby a conserved structural and functionalorganization (Figure 1) consisting of a hetero-geneous amino terminal domain, a highlyconserved central DNA binding domain(DBD) and a functionally complex carboxyterminal ligand-binding domain (LBD). TheLBD mediates ligand binding, receptor homo-and heterodimerization, repression oftranscription in the absence of ligand andligand-dependent activation of transcriptionwhen agonist ligands are bound (for review,see reference 2).

Crystal structures of several LBDs supportmolecular and biochemical studies indicatingthat ligand binding promotes a conformationalchange in receptor structure. What appears tobe a relatively flexible conserved helix near thecarboxy terminus (helix 12) occupies uniquepositions when structures of unliganded,agonist-occupied and antagonist-occupiedLBDs are compared (for review, see references3, 4). Importantly, mutagenesis experimentsindicate that helix 12, referred to as activationfunction 2 (AF-2), is necessary for ligand-dependent transactivation by nuclear recep-tors. Recent work indicates the AF-2 helixcontributes an essential surface to the forma-tion of an agonist-dependent hydrophobicpocket that serves as a binding site for co-activators. The alternative positions occupiedby helix 12 in the unliganded or antagonist-occupied conformations preclude the forma-tion of this binding pocket (for review, seereferences 5, 6).

Numerous experiments have defined theeffects of glucocorticoids and thyroid hormoneon metabolic control and provided the founda-tion for the endocrine regulation of metabo-lism (for review, see references 7, 8). Morerecently, studies have identified the peroxi-some proliferator activated receptors (PPAR-α,PPAR-δ and PPAR-γ), liver X receptors (LXRαand LXRβ) and the farnesoid X receptor(FXR) as direct sensors of metabolic status.The PPARs, LXRs and FXR directly bind fattyacids and cholesterol derivatives and regulatetarget genes that control the transport and

LXR as a therapeutic target foratherosclerosisI.G. Schulman and R.A. Heyman

9

DBD LBD -CN-

Figure 1 Structural organization of nuclear receptors.The DNA-binding domain (DBD) and the ligand-binding domain (LBD) are highlighted

ch09 14/7/05 4:46 pm Page 93

ultimate metabolic fate of the cognateligand.9–14 Thus these receptors are poised tosense and respond to small changes in the fluxthrough the metabolic pathways that theycontrol. Importantly, these fatty-acid- andcholesterol-derived natural ligands bind toreceptors with dissociation constants close tothe physiological concentrations known toexist for these metabolites.11,12,15–18 The identifi-cation of the thiazolidinedione class of insulinsensitizers as synthetic ligands for PPAR-γ19,20

served to validate the idea that these metabolicsensors can be therapeutic targets for the treat-ment of human disease.

REGULATION OF HEPATIC LIPIDMETABOLISM BY LXR

The LXR subgroup of the nuclear receptorsuperfamily is comprised of two isotypes, LXRαand LXRβ, that are encoded by separate genes.The founding member of the subgroup LXRαwas originally cloned from a liver cDNA library,hence the name liver X receptor, using theDNA binding of the retinoic acid receptor as aprobe, and found to be highly expressed in theliver, kidney and intestine.21 In contrast, LXRβis more ubiquitously expressed.22 Both LXRsbind to DNA and regulate transcription asheterodimers with retinoid X receptors servingas the common heterodimeric partner.21,22 Thefirst link between LXR and lipid metabolismcame from the identification of cholesterolderivatives including 22(R)-hydroxycholesterol,24(S)-hydroxycholesterol and 24(S),25-epoxy-cholesterol as ligands that directly bind to bothLXR-α and LXR-β and increase their transcrip-tional activity.16–18) More recent studies havealso demonstrated that 27-hydroxycholesteroland cholestenoic acid are LXR ligands.23,24 Theidentification of hydroxycholesterols as naturalLXR ligands dovetailed nicely with the charac-terization of LXR-α knockout mice. Whileapparently normal under standard laboratoryconditions, when challenged with a diet rich incholesterol LXR-α–/– mice accumulate massiveamounts of cholesterol in the liver. Molecularanalysis uncovered aberrant regulation of

several genes involved in lipid and cholesterolmetabolism including Cyp7a, which encodescholesterol 7α hydroxylase, the rate-limitingenzyme in the conversion of cholesterol to bileacids.25 Subsequently, the ATP-binding cassettetransporters (ABC) ABCG5 and ABCG8, whichmove cholesterol out of the liver and into theintestine, were identified as LXR targetgenes.26,27 Thus an increase in cholesterol levelsis predicted to lead to an elevation in theconcentration of cholesterol-derived LXRligands resulting in the catabolism of choles-terol to bile acid and the excretion of choles-terol out of the liver (Figure 2). Importantly,Cyp7a, ABCG5 and ABCG8 all appear to bedirectly regulated by LXR18,26 although thebinding site for LXR present in the murineCyp7a gene is not conserved in the humangene.18

Along with effects on cholesterol metabo-lism, activation of LXR agonists also increasesexpression of genes involved in fatty acidmetabolism including the master transcrip-tional regulator of fatty acid synthesis, the


Bile acidsCyp7A

SREBP1c

LXR

LXR

ABCG5ABCG8

Cholesterol

HDL ABCA1

Triglycerides

Figure 2 Hepatic activity of LXR. Activation of LXR inthe liver results in upregulation of the genes encodingcholesterol 7 hydroxylase (Cyp7a), the rate-limitingenzyme in the catabolism of cholesterol to bile acids,the half ABC transporters ABCG5 and ABCG8 thatfunction together to excrete cholesterol from the liverinto the intestine and ABCA1, which promotes theefflux of cholesterol to HDL. LXR also regulatesexpression of SREBP1c, a master transcriptionalregulator of genes involved in fatty acid and triglyceridesynthesis. See text for details

ch09 14/7/05 4:46 pm Page 94

sterol response element binding protein 1c(SREBP1c)28,29 (Figure 2). Additionally, severalof the genes encoding the enzymes involved infatty acid metabolism, including fatty acidsynthase (FAS) and stearoyl CoA desaturase 1(SCD-1), are regulated directly or indirectly byLXR.30–32 The coordinated upregulation offatty acid synthesis with reverse cholesteroltransport is most likely to provide lipids for thetransport and storage of cholesterol. Incontrast to the agonist activity of cholesterolmetabolites, however, fatty acids act as antago-nists of LXR transcriptional activity, suggestingthe possibility of a negative feedback loopwhereby the metabolic end product inhibitsthe inducer.33

REGULATION OF REVERSECHOLESTEROL TRANSPORT BY LXR

Based on the defined role for LXR in hepaticcholesterol catabolism and excretion, onemight have expected that synthetic LXRagonists would lower plasma cholesterol levels.Quite surprisingly, however, treatment ofanimals with LXR agonists significantly elevatesHDL-cholesterol.29 Gene expression analysis inthe livers and intestines of LXR-agonist-treatedmice identified ABCA1 as a direct LXR geneand this discovery stimulated great interest inthe therapeutic potential of LXR agonists giventhe links between ABCA1, HDL metabolismand atherosclerosis.34–36 ABCA1 is required forthe process of reverse cholesterol transport,whereby cells efflux internal cholesterol toacceptor proteins on pre-β-HDL particles. Lossof functional ABCA1 results in Tangier disease,a condition in which patients have extremelylow levels of circulating HDL and an increasedrisk for developing atherosclerosis. Examina-tion of fibroblasts isolated from subjects withTangier disease reveals that ABCA1-defectivecells are unable to efflux cholesterol, suggest-ing that the low HDL levels and increased riskof atherosclerosis result from a loss of reversecholesterol transport. Historically, Tangierdisease patients present with large accumula-tions of cholesterol-laden macrophages in their

lymph tissues, highlighting the role of ABCA1and reverse cholesterol transport in macro-phage cholesterol homeostasis (for review, seereferences 37–39). Importantly, accumulationof oxidized LDL-cholesterol by macrophages inthe arterial wall is an initiating step in thedevelopment of atherosclerotic lesions (forreview, see reference 40) and recent studieswith mouse knockouts of ABCA1 furthersupport a link between reverse cholesteroltransport and atherosclerosis.41–43 In support ofthe role of LXR as a direct regulator of ABCA1expression and activity, treatment of primarymacrophages or cell lines with LXR agonistsresults in induction of the ABCA1 gene,increased levels of ABCA1 protein and anincrease in cholesterol efflux.34,36,44 A bindingsite for LXR–RXR heterodimers in the ABCA1promoter has also been described.34 Subse-quent studies identified other proteinsinvolved in the reverse cholesterol transportincluding ABCG1, ABCG4 and apolipoproteinE (apoE) as direct LXR target genes.44–47 Thusactivation of LXR results in the mobilization ofcholesterol in the periphery (Figure 3) and

LXR AS A THERAPEUTIC TARGET FOR ATHEROSCLEROSIS 95

HDL

ABCA1ABCG1ABCG4apoE

LXR

TNFαiNosCOX2MMP-9

Figure 3 Activity of LXR in macrophages. Activation ofLXR in macrophages results in the upregulation ofgenes encoding proteins that participate in reversecholesterol transport including ABCA1, ABCG1, ABCG4and apoE. LXR agonists can also induce reversecholesterol transport in other tissues such as skeletalmuscle.61 Additionally, LXR also mediates an anti-inflammatory effect by repressing the ability of NFκB toinduce proinflammatory genes. See text for details

ch09 14/7/05 4:46 pm Page 95

stimulates the catabolism and excretion ofcholesterol when it arrives in the liver (Figure2). Interesting genetic deletion of LXR activityin mice (LXRαβ–/–) results in the accumulationof cholesterol-laden macrophages and spleno-megaly similar to that observed in Tangierdisease patients.48,49

The accumulation of oxidized LDL-choles-terol by macrophages in blood vessel walls is anearly event in the pathogenesis of atherosclero-sis and it had long been suggested that revers-ing this process by pumping cholesterol out ofmacrophage foam cells would have aninhibitory effect on the progression of athero-sclerosis (for review, see reference 40). Theability of LXR to directly regulate reversecholesterol transport in macrophages allowedtwo experiments to be carried out to test thishypothesis. First, transplantation of lethallyirradiated apoE–/– and LDL receptor knockout(LDLR–/–) mice with bone marrow fromLXRαβ+/+ or LXRαβ–/– mice demonstrated thatgenetic deletion of LXR leads to an increase inatherosclerosis is these well-established mousemodels.49 Second, treatment of apoE–/– andLDLR–/– mice with synthetic LXR agonistsleads to a reduction in atherosclerosis.50,51

Together the combination of genetic analysisand pharmacology clearly demonstrated theanti-atherogenic activity of LXR. Not surpris-ingly, the mRNA levels for LXR target genesincluding ABCA1 and apoE are elevated in theatherosclerotic lesions of mice treated withLXR agonists.51 Subsequent studies in ourlaboratory combining bone marrow transplan-tation with the administration of synthetic LXRagonists have demonstrated that LXR activityin macrophages is necessary for the anti-atherogenic effect of LXR ligands.52

ANTI-INFLAMMATORY ACTIVITY OFLXR

It was certainly easy to assume that the ability ofLXR to regulate lipid metabolism and reversecholesterol transport provides the mechanisticbasis for the anti-atherogenic activity of

LXR.49–51 Recent studies in macrophages,however, also demonstrate that LXR agonistscan inhibit the expression of several pro-inflammatory genes including iNOS, COX-2,and MMP-9 and these compounds are effectivein a murine model of irritant contact dermati-tis.53,54 Additionally, studies in our laboratoryhave shown that LXR agonists can inhibit LPS-dependent induction of TNF-α levels in vivoand this effect of LXR agonists is absent inLXR-αβ–/– mice (I. Schulman, personal obser-vation). Molecular studies indicate that activa-tion of LXR decreases the transcriptionalactivity of NFκB,54 although the mechanisticbasis for this inhibition has not yet been deter-mined.

Since atherosclerosis is considered aninflammatory disease (for review, see reference40) the question remains whether LXRmediates its anti-atherogenic via control ofreverse cholesterol transport, by limiting theinflammatory response, or both. Future studiesthat combine genetically altered macrophages(i.e. ABCA1–/–) introduced by bone marrowtransplantation along with the administrationof LXR agonists can be used to define theindividual contributions of reverse cholesteroltransport and anti-inflammatory activity totherapeutic effects of LXR ligands.Additionally, studies with the glucocorticoidreceptor have shown that it is possible toidentify nuclear receptor ligands that repressinflammatory genes but do not activatepositively regulated glucocorticoid receptortarget genes.55,56 One expects that such dissoci-ated ligands will also be identified for LXR.

A number of studies have suggested a linkbetween viral or bacterial infections and ather-osclerosis (for review, see reference 57) andenhanced expression of Toll-like receptors(TLR), which mediate the rapid innateimmunity to invading pathogens via stimula-tion of pro-inflammatory pathways, has beendetected in human atherosclerotic lesions.58

Interestingly, recent studies by Castrillo et al59

indicate that activation of TLR4 inhibits thetranscriptional activity of LXR and the ability ofmacrophages to efflux cholesterol. The cross-talk between TLR4 signalling and LXR activity


ch09 14/7/05 4:46 pm Page 96

suggests one potential mechanistic basis for theimpact of infectious agents on cardiovasculardisease.

LXR AND DIABETES

Treatment of experimental animals with LXRagonists leads to increases in hepatic fatty acidsynthesis and plasma triglyceride levels.28,29

Since elevations in fatty acids have been linkedto insulin resistance and type II diabetes,several investigators have examined the cross-talk between LXR activity and glucose metabo-lism. Interestingly, along with upregulatingfatty acid synthesis, activation of LXR alsorepresses expression of the genes encoding theenzymes of gluconeogenesis includingphosphoenolpyruvate carboxy kinase (PEPCK)and glucose 6-phosphatase60,61 and inducesexpression of GLUT4 in adipose tissue.60 Thus,in many ways, activation of LXR mimics treat-ment with insulin. Perhaps not surprisingly inlight of this ‘insulin-like’ activity, LXR ligandsdecrease hepatic glucose output and lowerblood glucose levels in animal models of type 2diabetes.60,61 The observation, however, thatLXR ligands can behave as insulin sensitizers,even in the face of relatively large increases inplasma triglyceride levels, suggests the possibil-ity of a broader role for LXR in regulatingglucose homeostasis. Indeed, the observationthat LXRs are also active in skeletal muscle62

and that LXR-αβ–/– mice are resistant to diet-induced obesity (I. Schulman, personal obser-vation) supports a role for the LXRs asimportant co-ordinators of energy metabolism.

THERAPEUTIC POTENTIAL OF LXRLIGANDS

The anti-atherogenic, anti-inflammatory andanti-diabetic activities of LXR agonists in animalmodels have highlighted the therapeutic poten-tial of LXR-α and LXR-β. Nevertheless, the linkbetween LXR activity and triglyceride metabo-lism has clearly dampened the enthusiasmsurrounding this target class. Treatment of

mice and hamsters with synthetic LXR agonistsresults in a significant increase in plasma triglyc-eride levels,29 although the kinetics and magni-tude of the triglyceride elevation varydepending on the specific LXR ligand understudy (I. Schulman and R. Heyman, personalobservations). Treatment of patients withcardiovascular disease or metabolic syndromewith a drug that raises triglycerides, however, isnot a viable option and approaches to separatethe beneficial activities of LXR ligands fromunwanted side-effects need to be explored.Furthermore, studies in human cells haveshown that LXR agonists also increase expres-sion of the gene encoding the cholesterol estertransfer protein (CETP).63 CETP functions totransfer cholesterol esters from HDL toapolipoprotein-B-containing lipoprotein parti-cles and CETP activity has been shown toinversely correlate with atherosclerosis.64–66

Indeed CETP inhibitors are currently beingexplored for the treatment of atherosclerosis.65

Interestingly, defects in hepatic cholesterolmetabolism are detected in LXR-α–/– singleknockout mice, indicating that LXR-β is notfunctionally redundant with LXR-α.25 Incontrast, cholesterol and triglyceride levelsappear normal in LXR-β–/– mice, suggestingthat LXR-α mediates most, if not all, of theeffects of LXR ligands on triglyceride metabo-lism.67 The relatively low level of LXR-β in theliver most likely accounts for lack of functionalredundancy in this tissue. Nevertheless, inmacrophages either LXR-α or LXR-β aloneappears to be sufficient to mediate the effectsof LXR ligands on reverse cholesterol transportand inflammatory gene expression (I.Schulman, personal observation). Takentogether these observations have led severalinvestigators to suggest that LXR-β-selectiveligands may provide a mechanistic basis foridentification of LXR ligands with improvedtherapeutic profiles.68 The enthusiasm forLXR-β-selective ligands must be tempered withthe realization that the spectrum of activitiesmeasured in the complete absence of LXR-αactivity may differ when an isotype selectivesynthetic ligand is used. Additionally, theobservation that ligand-binding pockets of


ch09 14/7/05 4:46 pm Page 97

LXR-α and LXR-β defined by crystallogra-phy69–72 differ by only one amino acid suggeststhat identification of selective ligands may notbe simple.

While the therapeutic activity of LXR-β-selec-tive ligands is still an open question, it has beenpossible to identify ligands for other nuclearreceptors that exhibit a restricted set of activi-ties and therefore allow the separation ofbeneficial therapeutic activities from unwantedside-effects. Perhaps the best examples of suchcompounds are the selective oestrogen recep-tor modulators such as roloxifene that functionas oestrogen receptor agonists in some tissuesand oestrogen receptor antagonists in others(for review, see reference 73). More recently,synthetic ligands for PPAR-γ have been identi-fied that appear to separate the insulin-sensitiz-ing activity of PPAR-γ from unwanted effects on

weight gain.74,75 A common feature of all theseselective receptor modulators is that theyappear to function as partial or weak agonistswhen characterized in vitro. When bound toreceptors, selective modulators produceunique conformational changes that cannot beachieved by more typical agonists (for review,see references 3, 4). The outcome of theseunique conformations is an alteration in inter-actions between receptors and the down-stream co-regulator proteins that mediate thetranscriptional response leading to ligand-specific effects on gene expression.73,76 Sincethe LXRs function in multiple tissues tomediate effects on lipid metabolism, glucosehomoeostasis and inflammation we expect thatthe identification of selective LXR modulatorswill yield compounds with beneficial therapeu-tic activities.


1. Maglich JM, Sluder A, Guan X, et al. Comparison ofcomplete nuclear receptor sets from the human,Caenorhabditis elegans and Drosophila genomes.Genome Biol 2001; 2:RESEARCH0029:1–7

2. Mangelsdorf DJ, Thummel C, Beato M, et al. Thenuclear receptor superfamily: the second decade. Cell1995; 83:835–9

3. Steinmetz AC, Renaud JP, Moras D. Binding of ligandsand activation of transcription by nuclear receptors.Annu Rev Biophys Biomol Struct 2001; 30:329–59

4. Greschik H, Moras D. Structure–activity relationshipof nuclear receptor–ligand interactions. Curr TopMed Chem 2003; 3:1573–99

5. McKenna NJ, O’Malley BW. Combinatorial control ofgene expression by nuclear receptors and coregula-tors. Cell 2002; 108:465–74

6. Westin S, Rosenfeld MG, Glass CK. Nuclear receptorcoactivators. Adv Pharmacol 2000; 47:89–112

7. Yen PM. Physiological and molecular basis of thyroidhormone action. Physiol Rev 2001; 81:1097–142

8. Schacke H, Docke WD, Asadullah K. Mechanismsinvolved in the side effects of glucocorticoids.Pharmacol Ther 2002; 96:23–43

9. Chawla A, Repa JJ, Evans RM, et al. Nuclear receptorsand lipid physiology: opening the X-files. Science2001; 294:1866–70

10. Edwards PA, Kast HR, Anisfeld AM. BAREing it all.The adoption of lxr and fxr and their roles in lipidhomeostasis. J Lipid Res 2002; 43:2–12

11. Forman BM, Chen J, Evans RM. Hypolipidemic drugs,polyunsaturated fatty acids, and eicosanoids areligands for peroxisome proliferator-activated recep-tors alpha and delta. Proc Natl Acad Sci USA 1997;94:4312–7

12. Kliewer SA, Sundseth SS, Jones SA, et al. Fatty acidsand eicosanoids regulate gene expression throughdirect interactions with peroxisome proliferator-activated receptors alpha and gamma. Proc Natl AcadSci USA 1997; 94:4318–23

13. Lee CH, Olson P, Evans RM. Minireview: lipid metab-olism, metabolic diseases, and peroxisome prolifera-tor-activated receptors. Endocrinology 2003;144:2201–7

14. Repa JJ, Mangelsdorf DJ. Nuclear receptor regulationof cholesterol and bile acid metabolism. Curr OpinBiotechnol 1999; 10:557–63

15. Forman BM, Ruan B, Chen J, et al. The orphannuclear receptor LXRalpha is positively andnegatively regulated by distinct products of meval-onate metabolism. Proc Natl Acad Sci USA 1997;94:10588–93

16. Janowski BA, Grogan MJ, Jones SA, et al. Structuralrequirements of ligands for the oxysterol liver Xreceptors LXRalpha and LXRbeta. Proc Natl Acad SciUSA 1999; 96:266–71

17. Janowski BA, Willy PJ, Devi TR, et al. An oxysterolsignalling pathway mediated by the nuclear receptorLXRα. Nature 1996; 383:728–31

References

ch09 14/7/05 4:46 pm Page 98

18. Lehmann JM, Kliewer SA, Moore LB, et al. Activationof the nuclear receptor LXR by oxysterols defines anew hormone response pathway. J Biol Chem 1997;272:3137–40

19. Lehmann JM, Moore LB, Smith-Oliver TA, et al. Anantidiabetic thiazolidinedione is a high affinity ligandfor peroxisome proliferator-activated receptorgamma (PPAR gamma). J Biol Chem 1995; 270:12953–6

20. Forman BM, Tontonoz P, Chen J, et al. 15-Deoxy-delta 12, 14-prostaglandin J2 is a ligand for theadipocyte determination factor PPAR gamma. Cell1995; 83:803–12

21. Willy PJ, Umesono K, Ong ES, et al. LXR, a nuclearreceptor that defines a distinct retinoid responsepathway. Genes Devel 1995; 9:1033–45

22. Shinar DM, Endo N, Rutledge SJ, et al. NER, a newmember of the gene family encoding the humansteroid hormone nuclear receptor. Gene 1994;147:273–6

23. Song C, Liao S. Cholestenoic acid is a naturally occur-ring ligand for liver X receptor alpha. Endocrinology2000; 141:4180–4

24. Fu X, Menke JG, Chen Y, et al. 27-hydroxycholesterolis an endogenous ligand for liver X receptor incholesterol-loaded cells. J Biol Chem 2001; 276:38378–87

25. Peet DJ, Turley SD, Ma W, et al. Cholesterol and bileacid metabolism are impaired in mice lacking thenuclear oxysterol receptor LXRα. Cell 1998; 93:693–704

26. Repa JJ, Berge KE, Pomajzl C, et al. Regulation ofATP-binding cassette sterol transporters ABCG5 andABCG8 by the liver X receptors alpha and beta. J BiolChem 2002; 277:18793–800

27. Berge KE, Tian H, Graf GA, et al. Accumulation ofdietary cholesterol in sitosterolemia caused bymutations in adjacent ABC transporters. Science2000; 290:1771–5

28. Repa JJ, Liang G, Ou J, et al. Regulation of mousesterol regulatory element-binding protein-1c gene(SREBP-1c) by oxysterol receptors, LXRalpha andLXRbeta. Genes Dev 2000; 14:2819–30

29. Schultz JR, Tu H, Luk A, et al. Role of LXRs in controlof lipogenesis. Genes Dev 2000; 14:2831–8

30. Zhang Y, Repa JJ, Gauthier K, et al. Regulation oflipoprotein lipase by the oxysterol receptors,LXRalpha and LXRbeta. J Biol Chem 2001;276:43018–24

31. Wang Y, Kurdi-Haidar B, Oram JF. LXR-mediatedactivation of macrophage stearoyl-CoA desaturasegenerates unsaturated fatty acids that destabilizeABCA1. J Lipid Res 2004; 45:972–80

32. Joseph SB, Laffitte BA, Patel PH, et al. Direct andindirect mechanisms for regulation of fatty acidsynthase gene expression by LXRs. J Biol Chem 2002;277:11019–25

33. Ou J, Tu H, Shan B, et al. Unsaturated fatty acidsinhibit transcription of the sterol regulatory element-binding protein-1c (SREBP-1c) gene by antagonizing

ligand-dependent activation of the LXR. Proc NatlAcad Sci USA 2001; 98:6027–32

34. Costet P, Luo Y, Wang N, et al. Sterol-dependenttransactivation of the ABC1 promoter by the liver Xreceptor/retinoid X receptor. J Biol Chem 2000;275:28240–5

35. Repa JJ, Turley SD, Lobaccaro JA, et al. Regulation ofabsorption and ABC1–mediated efflux of cholesterolby RXR heterodimers. Science 2000; 289:1524–9

36. Venkateswaran A, Laffitte BA, Joseph SB, et al.Control of cellular cholesterol efflux by the nuclearoxysterol receptor LXR alpha. Proc Natl Acad SciUSA 2000; 97:12097–102

37. Hayden MR, Clee SM, Brooks-Wilson A, et al.Cholesterol efflux regulatory protein, Tangier diseaseand familial high-density lipoprotein deficiency. CurrOpin Lipidol 2000; 11:117–22

38. Hobbs HH, Rader DJ. ABC1: connecting yellowtonsils, neuropathy, and very low HDL. J Clin Invest1999; 104:1015–17

39. Oram JF, Lawn RM. ABCA1. The gatekeeper foreliminating excess tissue cholesterol. J Lipid Res2001; 42:1173–9

40. Glass CK, Witztum JL. Atherosclerosis. The roadahead. Cell 2001; 104:503–16

41. Aiello RJ, Brees D, Francone OL. ABCA1–deficientmice: insights into the role of monocyte lipid efflux inHDL formation and inflammation. ArteriosclerThromb Vasc Biol 2003; 23:972–80

42. Singaraja RR, Fievet C, Castro G, et al. IncreasedABCA1 activity protects against atherosclerosis. J ClinInvest 2002; 110:35–42

43. Joyce CW, Amar MJ, Lambert G, et al. The ATPbinding cassette transporter A1 (ABCA1) modulatesthe development of aortic atherosclerosis in C57BL/6and apoE-knockout mice. Proc Natl Acad Sci USA2002; 99:407–12

44. Wagner BL, Valledor AF, Shao G, et al. Promoter-specific roles for liver X receptor/corepressorcomplexes in the regulation of ABCA1 and SREBP1gene expression. Mol Cell Biol 2003; 23:5780–9

45. Kennedy MA, Venkateswaran A, Tarr PT, et al.Characterization of the human ABCG1 gene: liver Xreceptor activates an internal promoter that producesa novel transcript encoding an alternative form of theprotein. J Biol Chem 2001; 276:39438–47

46. Laffitte BA, Repa JJ, Joseph SB, et al. LXRs controllipid-inducible expression of the apolipoprotein Egene in macrophages and adipocytes. Proc Natl AcadSci USA 2001; 98:507–12

47. Wang N, Lan D, Chen W, et al. ATP-binding cassettetransporters G1 and G4 mediate cellular cholesterolefflux to high-density lipoproteins. Proc Natl Acad SciUSA 2004; 101:9774–9

48. Schuster GU, Parini P, Wang L, et al. Accumulation offoam cells in liver X receptor-deficient mice.Circulation 2002; 106:1147–53

49. Tangirala RK, Bischoff ED, Joseph SB, et al.Identification of macrophage liver X receptors as


ch09 14/7/05 4:46 pm Page 99

inhibitors of atherosclerosis. Proc Natl Acad Sci USA2002; 99:11896–901

50. Terasaka N, Hiroshima A, Koieyama T, et al. T-0901317, a synthetic liver X receptor ligand, inhibitsdevelopment of atherosclerosis in LDL receptor-deficient mice. FEBS Lett 2003; 536:6–11

51. Joseph SB, McKilligin E, Pei L, et al. Synthetic LXRligand inhibits the development of atherosclerosis inmice. Proc Natl Acad Sci USA 2002; 99:7604–9

52. Lenin N, Bischoff ED, Daige Dl et al. Macrophageliver X receptor is required for antiatherogenic activ-ity of LXR agonists. Arterioscler Thromb Vasc Biol2005; 25(1):135–42

53. Fowler AJ, Sheu MY, Schmuth M, et al. Liver X recep-tor activators display anti-inflammatory activity inirritant and allergic contact dermatitis models: liver-X-receptor-specific inhibition of inflammation andprimary cytokine production. J Invest Dermatol 2003;120:246–55

54. Joseph SB, Castrillo A, Laffitte BA, et al. Reciprocalregulation of inflammation and lipid metabolism byliver X receptors. Nat Med 2003; 9:213–9

55. Miner JN. Designer glucocorticoids. BiochemPharmacol 2002; 64:355–61

56. Coghlan MJ, Jacobson PB, Lane B, et al. A novel anti-inflammatory maintains glucocorticoid efficacy withreduced side effects. Mol Endocrinol 2003; 17:860–9

57. Buja LM. Does atherosclerosis have an infectiousetiology? Circulation 1996; 94:872–3

58. Edfeldt K, Swedenborg J, Hansson GK, et al.Expression of toll-like receptors in human atheroscle-rotic lesions: a possible pathway for plaque activation.Circulation 2002; 105:1158–61

59. Castrillo A, Joseph SB, Vaidya SA, et al. Crosstalkbetween LXR and toll-like receptor signalingmediates bacterial and viral antagonism of cholesterolmetabolism. Mol Cell 2003; 12:805–16

60. Laffitte BA, Chao LC, Li J, et al. Activation of liver Xreceptor improves glucose tolerance through coordi-nate regulation of glucose metabolism in liver andadipose tissue. Proc Natl Acad Sci USA 2003;100:5419–24

61. Cao G, Liang Y, Broderick CL, et al. Antidiabeticaction of a liver x receptor agonist mediated by inhibi-tion of hepatic gluconeogenesis. J Biol Chem 2003;278:1131–6

62. Muscat GE, Wagner BL, Hou J, et al. Regulation ofcholesterol homeostasis and lipid metabolism inskeletal muscle by liver X receptors. J Biol Chem2002; 277:40722–8

63. Luo Y, Tall AR. Sterol upregulation of human CETPexpression in vitro and in transgenic mice by an LXRelement. J Clin Invest 2000; 105:513–20

64. de Grooth GJ, Smilde TJ, Van Wissen S, et al. Therelationship between cholesteryl ester transferprotein levels and risk factor profile in patients withfamilial hypercholesterolemia. Atherosclerosis 2004;173:261–7

65. Brousseau ME, Schaefer EJ, Wolfe ML, et al. Effects ofan inhibitor of cholesteryl ester transfer protein onHDL cholesterol. N Engl J Med 2004; 350:1505–15

66. Klerkx AH, de Grooth GJ, Zwinderman AH, et al.Cholesteryl ester transfer protein concentration isassociated with progression of atherosclerosis andresponse to pravastatin in men with coronary arterydisease (REGRESS). Eur J Clin Invest 2004; 34:21–8

67. Alberti S, Schuster G, Parini P, et al. Hepatic choles-terol metabolism and resistance to dietary cholesterolin LXRbeta-deficient mice. J Clin Invest 2001;107:565–73

68. Lund EG, Menke JG, Sparrow CP. Liver X receptoragonists as potential therapeutic agents for dyslipi-demia and atherosclerosis. Arterioscler Thromb VascBiol 2003; 23:1169–77

69. Hoerer S, Schmid A, Heckel A, et al. Crystal structureof the human liver X receptor beta ligand-bindingdomain in complex with a synthetic agonist. J MolBiol 2003; 334:853–61

70. Farnegardh M, Bonn T, Sun S, et al. The three-dimen-sional structure of the liver X receptor beta reveals aflexible ligand-binding pocket that can accommodatefundamentally different ligands. J Biol Chem 2003;278:38821–8

71. Williams S, Bledsoe RK, Collins JL, et al. X-ray crystalstructure of the liver X receptor beta ligand bindingdomain: regulation by a histidine–tryptophan switch.J Biol Chem 2003; 278:27138–43

72. Svensson S, Ostberg T, Jacobsson M, et al. Crystalstructure of the heterodimeric complex of LXRalphaand RXRbeta ligand-binding domains in a fullyagonistic conformation. Embo J 2003; 22:4625–33

73. McDonnell DP, Connor CE, Wijayaratne A, et al.Definition of the molecular and cellular mechanismsunderlying the tissue-selective agonist/antagonistactivities of selective estrogen receptor modulators.Recent Prog Horm Res 2002; 57:295–316

74. Oberfield JL, Collins JL, Holmes CP, et al. A peroxi-some proliferator-activated receptor gamma ligandinhibits adipocyte differentiation. Proc Natl Acad SciUSA 1999; 96:6102–6

75. Rocchi S, Picard F, Vamecq J, et al. A uniquePPARgamma ligand with potent insulin-sensitizing yetweak adipogenic activity. Mol Cell 2002; 8:737–47

76. Schulman IG, Heyman RA. The flip side: identifyingsmall molecule regulators of nuclear receptors. ChemBiol 2004; 11:639–46


ch09 14/7/05 4:46 pm Page 100

INTRODUCTION

Since the initial report of an inverse relation-ship between high density lipoprotein (HDL)levels and cardiovascular risk,1 plasma HDLlevels have been a clinically important targetfor prevention of atherosclerosis. With thediscovery of the role of ABCA12–4 in the vascu-lar efflux of cholesterol to form nascent HDLparticles and SR-B15–7 in the hepatic uptake ofcholesterol from HDL came heightened inter-est in identifying novel genes in regulatingplasma HDL levels.

With its discovery in 1999,8,9 endotheliallipase joined lipoprotein lipase (LPL) andhepatic lipase (HL) as important plasma regula-tors of lipoprotein cholesterol levels. Althougheach of these enzymes is capable of hydrolysingtriglycerides and phospholipids, there aredifferences in their substrate preferences.While LPL and, to a lesser extent, HL arepredominantly triglyceride lipases, endotheliallipase (EL) is predominantly a phospholipase.10

EL is also unique in its expression by endothe-lial cells.8,9 EL, like LPL and HL, is secreted andthen binds to heparan sulphate proteoglycanson the endothelial surface. There it interactswith HDL, hydrolysing HDL phospholipids andgenerating free fatty acids at the local tissue site.The ability of EL to profoundly decrease HDLlevels in mice has stimulated significant interestin its relevance to human HDL metabolism.The focus of this chapter will be recent reportsof both in vitro and in vivo function of endothe-lial lipase.

STRUCTURE AND FUNCTION OFENDOTHELIAL LIPASE

Endothelial lipase is a 500-amino-acid proteinwith a predicted molecular weight of 55 kDa.Immunoblot analysis of EL expressed in 293HEK cells, using an anti-human EL polyclonalantibody, reveals bands of apparent molecularweight 68 kDa, 40 kDa and 28 kDa. Aminoterminal sequencing demonstrated that the40 kDa and 28 kDa forms were proteolyticcleavage products of the full-length protein.Based on a preliminary analysis with glycosi-dases, the 68 kDa protein is probably the fullyglycosylated monomer (unpublished data).

Two alternatively spliced products of the EL(LIPG) gene were recently reported by Ishidaet al.11 These isoforms, labelled EL2a andEL2b, were 480 amino acids and 346 aminoacids, respectively. EL2a and EL2b have exon Iwhile the original full-length form containsexon II, which encodes the aminoterminalsequence with the signal peptide. EL2b lacksexon VI, which encodes the catalytic asparticacid residue and a portion of the lid region.EL2a and EL2b were localized to the cytosoland lacked enzymatic activity.

By comparison of their primary amino acidsequences, endothelial lipase has 45% homol-ogy with LPL and 40% homology with HL.8 Asdisplayed in Figure 1, the location of the 10cysteine residues is conserved. The samecatalytic residues, 149serine, 173aspartic acid and254histidine, assign this enzyme to the triglyc-eride lipase family. The GXSXG lipase motif

Endothelial lipase and the regulation ofHDL metabolismK.O. Badellino and W. Jin

10

ch10 14/7/05 4:47 pm Page 101

surrounding the active site serine, commonamong other members of the lipase family, isalso found in endothelial lipase.

An examination of the EL sequence alsoshows five areas that may represent N-linkedglycosylation sites. Two potential lipid-bindingregions have been identified by comparison tothe crystal structure of pancreatic lipase:12,13

170GLDPAGP177 and 204RSFGLSIGIQM214. ELalso has four regions of positively chargedresidues that resemble heparin-binding sites:14KLHKPK19, 282RFKK285 and 292RKNR295,304KKMRNKR310 and 427RRIRVK432. Thesepotential heparin-binding, or proteoglycan-binding, sites are similar to regions found inLPL.14,15 Endothelial lipase also has a 19-amino-acid putative lid region with a significantly

different sequence from the same regions inLPL and HL. This lid region may be importantin determining the more predominantphospholipase activity of EL.

The phospholipid preference of EL wasnoted in both initial reports of the discovery ofEL.8,9 Jaye et al8 also reported that EL hydro-lysed a radiolabelled triolein substrate at a lowlevel. McCoy et al10 more definitively charac-terized EL lipolytic activity and compared it toHL and LPL. In the absence of serum as asource of apolipoprotein CII (apoC-II), ELhydrolysis of tributyrin, a 3-carbon acyl chainsubstrate, and triolein, an unsaturated octade-cenoic acid, was similar to that of HL and LPL.The addition of heat-inactivated serumcompletely inhibited the ability of EL to


P L E V T P S T E N KM

RW

DG R C C

TE

DP

E

NT

SISPGR

ELW

FR

KF T L K R Q T E G S K V R I

RR

IN

LEGRPNRPQSLYSRFEK

WL

NY

WSQSAGE

WTLQIKLLDGLDEETYVLFTNTANQEI

RE

V

I

E L P L T Q S D A N T G Y L T V Y F TP

E I E G M N K

Y S F V H I K M Q Y H Y V R F P M G A R T K L Y M K S N RK

NR

M

KKANYGIS

N

CC

R

N

KRSLCC

IG

K

KF

RN

S

D

IQFAFSPKDQNVLSD

VFLHVAR

EHE

G

CC

KV

VE

T

I

TGYA

IS

GL

V

D

N

L

GPQFDGGNPYID

I

H

G VP M Q I G I S L

G FS R T Y T H L V

D VF D A D D P S L

R KH I D A G E F M P G

A P D L G T I RG

V TG

KV

F

N

GAYG

AVHA

GL

SYGILH

V

NG

LS

F

DD

KEQLWD

LM

RA

ISHGVV

RT

NN

VA

DTYLQH

ALP

LW

DV

VV

VN

AD

K E R T H LA

S V LK H

LW

NE

F IG

SM

TW G H I I

FF T

KA T

MN F S

CD

EL

P

QS H G V L

S

Y

C G E H E P D K ST

R

LN

VKP

SVRF

ETQ

T

A

KP

KHLK

DE

LRG

EPGFP

V

PSGAAFC

YCLS

WF

CL

LP

VS

NS

M

Signal peptide 1

10

20 30

406070

8090

100110 120

130

140 149

160

170173180190200210

220 230 254

250

260 270 290

280300

310

320330340

350

240

Lidregion

360

370 380 390 400 410 420

430440460

470480= potential N-glycosylation site= potential heparin-binding site= potential lipid-binding site= catalytic residues

50

Figure 1 A model of endothelial lipase was developed, based on the crystal structure of pancreatic lipase and apublished molecular model of LPL. The locations of cysteine residues, heparin-binding sites and the catalytic triadresidues are conserved

ch10 14/7/05 4:47 pm Page 102

hydrolyse both triolein and phospholipids.Addition of the same amount of serum hadthe expected effect of potentiating LPLhydrolysis of the triolein substrate. Theaddition of purified apoC-II also potentiatedLPL. In contrast to the effect of serum, apoC-II had no effect on EL hydrolysis of eithertriolein or phospholipids.

McCoy et al also compared the triglyceridelipase to phospholipase activity ratio of EL toHL and LPL. Using lipases expressed byadenoviral gene transfer into COS cells, thetriglyceride lipase to phospholipase activityratio was determined to be 0.65 for EL, 24 forHL and 140 for LPL (plus serum) (Figure 2).These ratios demonstrate the preference of ELfor phospholipids, versus the preference ofboth HL and LPL for triolein substrates.

While synthetic substrates can be used tocharacterize the lipase activity of theseenzymes, their ability to hydrolyse naturallyoccurring lipoproteins more accuratelydisplays their in vivo activity. McCoy et alisolated lipoproteins from non-fasted hyperli-paemic subjects. Lipoprotein fractions contain-ing 1.25 mM phospholipids were incubatedwith either EL-, LPL- or HL-conditionedmedium. A comparison of the free fatty acidsreleased per reaction showed that, while ELhydrolyses triglyceride-containing lipoproteins,it is more active in hydrolysing HDL thaneither HL or LPL.

Duong et al16 explored the influence of thephospholipid composition of recombinantspherical HDL on the ability of HL and EL tohydrolyse both phospholipid and triglyceride.For a comparison of phospholipid substrates,they created recombinant HDL (rHDL) withpalmitic acid at sn-1 and that differed only intheir phospholipid composition at the sn-2position. They found that the Vmax was greatestfor EL hydrolysis of phospholipid with thelonger chain docosodecanoic acid in the sn-2position (PDPC), while HL preferred oleic acidat the sn-2 position (POPC). In contrast, forboth EL and HL, triglyceride hydrolysis wasmodulated by the type of phospholipidincluded in triolein-containing rHDL. WhileEL is an extremely poor triglyceride lipase, theVmax for triolein in rHDL containing onlyarachidonic acid was highest, 41.4 nmol freefatty acid ml–1h–1, slightly increased over therate for PDPC. HL was most active with linoleicacid in the sn-2 position, slightly increased overthe Vmax for POPC. The phospholipid contentof HDL varies with the source of dietary fat. Asnoted by Duong et al, a diet rich in fish wouldenrich HDL with PDPC, making them bettersubstrates for EL. Rye et al17 found that alonger, more unsaturated, phospholipidcontent of HDL promotes the formation of thelipid-poor apoA-I particles involved in reversecholesterol transport. Together, these findingssuggest that EL hydrolysis of phospholipidsmay interfere with reverse cholesterol trans-port and promote the development of athero-sclerosis.

ENDOTHELIAL LIPASE AND THE REGULATION OF HDL METABOLISM 103

Triglyceridase activity

200

100

0LPL HL EL

Act

ivity

(pm

ol/m

in-m

l)A

ctiv

ity (

pmol

/min

-ml)

Phospholipase activity

7.5

5.0

2.5

0.0LPL HL EL

Figure 2 McCoy et al10 compared the triglyceride lipaseand phospholipase activities of LPL, HL and endotheliallipase. Culture medium from COS cells infected withadenoviruses containing the respective cDNAs was used.Used with permission from Journal of Lipid Research2002; 43:921–9

ch10 14/7/05 4:47 pm Page 103

In contrast, Gauster et al18 studied the effectof EL modification of HDL on both SR-B1 andABCA1-mediated free cholesterol efflux. HDLisolated from human plasma was incubated withCOS7 cells infected with adenovirus encodinghuman EL. This HDL was found to be depletedin phosphatidylcholine and enriched in non-esterified fatty acids, with an increase in negativecharge. EL-modified HDL had impaired abilityto promote 3H-cholesterol efflux from Chinesehamster ovary cells overexpressing SR-B1. Therewas no impairment of cholesterol efflux fromABCA1-expressing cells, however. Thesefindings suggest that EL may promote reversecholesterol transport by hydrolysing HDLphospholipid and modifying its surface charge.

TISSUE EXPRESSION OFENDOTHELIAL LIPASE

The gene encoding endothelial lipase was firstcloned in 1999 by two different groups8,9 using

two different strategies. Jaye et al8 treated amonocyte cell line, THP-1, with phorbol 12-myristate 13 acetate, then exposed the cells tooxidized LDL-cholesterol. An RNA amplifica-tion product, found in the LDL-treated cellsbut not in untreated cells, was used to probe ahuman placental cDNA library. The cloneobtained had significant homology to HL,pancreatic lipase and LPL. By Northern blot,they found EL is expressed in placenta, lung,liver, kidney, testis and thyroid in both mouseand human tissues (Figure 3). By Western blot,they showed that EL is constitutively anduniquely expressed by human coronary arteryendothelial cells and human umbilical veinendothelial cells. Similarly, Hirata et al9 identi-fied EL as a gene upregulated in endothelialcells undergoing tube formation. They found asimilar pattern of tissue expression.

Yu et al19 examined mouse and rat tissues todiscern whether EL is expressed exclusively byendothelial cells or whether it is first expressedby other cells such as hepatocytes and then


140

120

100

80

60

40

20

0

Thy

roid

Pla

cent

a

Live

r

Kid

ney

Lung

Trac

hea

Test

is

Col

on

App

endi

x

Sm

inte

st

Pro

stat

e

Cer

ebel

lum

Fet

al k

idne

y

Fet

al li

ver

Fet

al lu

ng

Fet

al b

rain

Exp

ress

ion

leve

l

Figure 3 Relative expression of endothelial lipase mRNA in human tissues. Figure courtesy of M. Jaye

ch10 14/7/05 4:47 pm Page 104

translocates to the endothelial surface. Byimmunostaining, they found that EL isexpressed only by endothelial cells in liver,large vessels of the thyroid, the lung andadrenal glands.

The physiological role of EL in the placenta,thyroid, lung and adrenal gland is as yetunknown. Fetal HDL is apoE-rich, but has thesame phospholipid content as adult HDL.20 Itis interesting to note that Yu et al19 found thatEL expression was decreased in the liver ofapoE knockout mice, but increased in theiraorta. This suggests a potential regulatory roleof apoE in EL expression.

Thyroid hormone is known to decrease HLactivity.21 Its effect on EL expression and activ-ity has yet to be examined. It is tempting tospeculate that EL is involved in local vascularlipid metabolism, especially in certainendocrine tissues such as the thyroid andadrenal gland. This may be true in the lung aswell. In their initial report of EL, Hirata et al9

suggested that EL co-localizes with type IIepithelial cells and may be involved in surfac-tant metabolism.

REGULATION OF ENDOTHELIALLIPASE EXPRESSION

Both acute and chronic inflammatory diseasesare associated with decreased HDL levels.22,23

Two groups have shown that two early phasepro-inflammatory cytokines, tumour necroticfactor alpha (TNF-α) and interleukin-1 beta (IL-1β), both markedly increase the expression ofEL mRNA in several lines of primary endothelialcells (Figure 4),24,25 which is in sharp contradic-tion to their effects on HL and LPL.26,27 Thesecretion of EL protein was induced by IL-1βand TNF-α in a dose- and time-dependentmanner, and was the major source of extracel-lular lipase activity in endothelial cells aftercytokine treatment.25 In the presence of SN50 (aP50 inhibitor), the induction of EL by these twocytokines was partially blunted, indicating thatthe NF-κB pathway might be involved in theregulation of EL. In addition, two forms ofmechanical force involved in vascular diseases,fluid shear stress and cyclic stretch, also upregu-lated EL mRNA expression in endothelialcells.24 All these data suggest the possibility that


Figure 4 Effect of cytokine stimulationon endothelial lipase protein andmRNA expression in HUVEC. In panelA, protein expression and in panel B,mRNA expression in response to (1–9):0, 1, 3, 10, 30, 100, 300, 1000 and 3000pg/ml IL-1β. In panel C, proteinexpression and in panel D, mRNAexpression in response to (1–8): 0,10 pg, 30 pg, 100 pg, 300 pg, 1 ng, 3 ng,and 10 ng/ml TNFα. Fold change inEL protein is based on the 68 kDaform only. Reproduced with permissionfrom Jin W et al.25 Endothelial cellssecrete triglyceride lipase andphospholipase activities in response tocytokines as a result of endotheliallipase. Circulation Research 2003;92:644–50

A

B

C

D

ch10 14/7/05 4:47 pm Page 105

inflammation-induced upregulation of ELcould be a causal factor in reducing HDL levelsseen in inflammatory conditions.

When fed with a high-fat diet, apoE–/– miceare susceptible to development of atheroscle-rosis, which is increasingly considered aninflammatory state. Yu et al measured ELmRNA by real-time PCR and Western blotting.Aortic EL mRNA and protein expression werehigher than that in wild-type mice, whereas itshepatic expression was significantly de-creased.19 Cholesterol feeding alone, or withadded saturated fat, further reduced hepaticEL mRNA expression. However, aortic ELmRNA was not regulated by these differentdiets. Interestingly, addition of cholic acid,which may induce hepatic inflammation,significant increased EL mRNA expression.28

Together these data suggest a tissue-specificmanner of regulation of EL. In human non-atherosclerotic coronary arteries, Azumi et alfound that EL was expressed in endothelialcells and medial smooth muscle cells. However,in atherosclerotic coronary arteries, EL wasexpressed in macrophages within atheroma-tous plaques as well as endothelial and smoothmuscle cells.28 Thus, additional EL supplied byinfiltrating macrophages may furtherstrengthen its local effects on HDL at the siteof atherosclerotic lesions. Under acute inflam-matory conditions, upregulation of EL inendothelial cells at local sites of inflammationmay be part of innate defence responses.Generated free fatty acids from HDL may beutilized by local tissues, including possibly theendothelial cells themselves, as an energysource. While initially potentially beneficial,the change of HDL by EL may be proathero-genic if present for an extended period.

MECHANISMS BY WHICH ELAFFECTS HDL METABOLISM

The lipolytic activity of EL against phospholipidsof HDL is thought to play a major role in reduc-ing HDL-cholesterol (HDL-C) levels. Whenoverexpressed in HepG2 cells, EL effectivelyhydrolysed phospholipids from HDL and liber-

ated fatty acids, which were then incorporatedinto cellular phospholipids and triglycerides.This leads to an increase in the amount of cellu-lar lipids and suppressed the rate of fatty acidsynthesis. In the presence of CD36, more fattyacids were incorporated into cellular lipids.29

Maugeais et al demonstrated that, when over-expressed in mice, EL resulted in a dose-dependent increase in post-heparin plasmaphospholipase activity. Kinetic studies demon-strated a dose-dependent acceleration of HDL-apolipoprotein catabolism, which was highlycorrelated to post-heparin plasma phospholipaseactivity (r = 0.89, p <0.01). This effect was not thesole result of the reduction in HDL pool size.30

Conversely, inhibition of EL activity caused a21% slower fractional catabolic rate (FCR) ofphospholipids in HDL particles (0.42 ± 0.03 vs0.53 ± 0.02 pool per hour).31 Post-heparinplasma phospholipase activity was dramaticallyreduced in EL–/– mice.32 All these data suggestthat the phospholipase activity of EL may directlymodulate the phospholipid content of HDL andsubsequently affect its turnover.

Independent of its lipolytic activity, EL canserve as a ligand to enhance binding and cellu-lar processing of HDL particles via heparansulphate proteoglycan-mediated pathways.Fuki et al33 found, in in vitro studies usingcultured cells expressing EL, that 70% of thesurface-bound HDL was released back into themedium, and only 30% of HDL holoparticleswere processed intracellularly. In addition,Strauss et al showed that the selective uptake ofHDL-associated cholesterol esters wasmarkedly elevated in EL-expressing cells andindependent of SR-BI.34 Broedl et al showedthat, in vivo, a catalytic inactive EL mutant,ELS149A, led to an intermediate reduction ofHDL-C in HL–/– mice compared to catalyticactive EL, but not in wild-type and apoA-I trans-genic mice.35 Thus, both lipolytic and non-lipolytic effects of EL may contribute toaccelerate the catabolism of HDL-C. Indeed,Maugeais et al noted that the uptake of HDLparticles was increased up to 158% and 189%by liver and kidney, respectively, in EL-express-ing mice as assessed using [125I]-tyramine-cellobiose-labelled HDL.30 On the other hand,


ch10 14/7/05 4:47 pm Page 106

Ma et al reported that the FCR of cholesterylester-labelled HDL was delayed in their kineticstudy in EL–/– mice (2.22 ± 0.28 vs 3.04 ± 0.31pool per hour).36

EFFECTS OF EL ON HDLMETABOLISM IN MICE

Our understanding of the physiological role ofEL in HDL metabolism is greatly advancedfrom several studies performed in mice.Hepatic EL overexpression, using an adenovi-ral vector encoding a human EL cDNA, firstshed light on its effects on HDL metabolism inwild-type and human apoA-I transgenic mice.Three days after vector injection, overexpres-sion of EL resulted in barely detectable plasmaHDL-C and apoA-I levels. Impressively, theseextremely low levels of HDL-C and apoA-I weresustained during the 28-day experimentalcourse, although human EL protein isexpected to have diminished to an almostundetectable level at day 28 post-injection.8 Adecreased HDL size was seen in EL-expressingmice as shown by gel filtration and non-denaturing gradient gel electrophoresis.31 Inline with these results, analysis of BAC-humanEL-transgenic mice revealed an approximate20% decrease in the plasma concentration ofHDL-C, a 26% decrease in apoA-I but other-wise similar composition of their HDL parti-cles. Of note, these mice expressed human ELtranscript in brain, aorta, heart, lung, kidneyand spleen, but not liver.33

Loss of function studies of EL in micefurther established that EL is a major geneticmodulator of HDL levels. A bolus injection of aneutralizing IgG specific for murine ELresulted in a 25–60% increased HDL-C in wild-type, apoA-I transgenic and HL–/– mice, withthe peak HDL-C levels occurring at 48 hoursafter injection. In addition, acute inhibition ofEL in HL–/– mice generated larger HDL parti-cles.32 Examination of two EL–/– mouse linesgenerated by two different laboratoriesdemonstrated ~57% increase and ~25%increase of HDL-C levels in homozygous andheterozygous EL–/– mice, respectively.33,36 Ishita

et al33 also showed both male and female EL–/–

mice had similar changes of HDL levels and itscomposition was not altered, suggesting anincreased number of HDL particles. Ma et al36

reported that EL–/– mice displayed higherlevels of HDL even with high-cholesterol dietfeeding. NMR analysis showed that EL–/– micehave an abundance of large HDL particles.Furthermore, EL–/– mice intercrossed onto aSR-BI–/– background showed a stepwise changein HDL-C levels, suggesting that the effects ofEL on HDL levels are independent of SR-BI.All these data demonstrated that EL inverselyregulates plasma HDL particle sizes and levels.

GENETIC STUDIES IN HUMANS

The relevance of studies of EL in cultured cellsand mouse models to human cholesterolmetabolism was first examined by deLemos etal.37 They asked whether elevated HDL-C levelsmay be a result of functional polymorphismsand mutations in the EL gene. They sequenced1200 base pairs of the promoter and all exonsof the EL gene in 20 individuals with HDL-Clevels greater than the 90th percentile for age,sex and race. They identified and confirmed 17polymorphic sites in these individuals (Table1). Six of these were potentially functional.Four polymorphisms resulted in amino acidchanges: Gly26Ser, Thr111Ile, Thr298Ser andAsn396Ser. Two were in the promoter(–303A/C and –410C/G). The genotypicfrequencies of each variant were measured in176 black controls, 165 white controls and 123whites with high HDL-C. The Thr111Ilepolymorphism occurred in 32.6% of whitesubjects with high HDL-C, 31.2% of controlwhite subjects and 10.3% of control blacksubjects. There was no association betweenHDL-C levels and the presence of this polymor-phism. Three variants, Gly26Ser, Thr298Serand -303A/C, were found in the cohort of highHDL-C whites and in blacks but not in thecontrol white population.

Ma et al36 examined 372 members of theLipoprotein and Coronary AtherosclerosisStudy for the presence of the single nucleotide


ch10 14/7/05 4:47 pm Page 107

polymorphism (SNP) 584C/T, which producesthe Thr111Ile variant. They found a significantunivariate association between the 584C/Tallele and mean plasma levels of HDL-C. Theyfound an allele-dependent variation with therank order TT>CT>CC for both HDL-C andapoA-I/apoB ratios. They followed theseindividuals for 2.5 years and found no differ-ence between genotypes in progression orregression of disease.

These studies were extended to a populationof 340 Japanese children between the ages of 9and 15 years without known lipid abnormali-ties.38 Yamakawa-Kobayashi et al identified twonew polymorphisms, –384A/C and 2237G/A.Using multiple linear regression with sex, age

and body mass index as co-variates, they foundsignificant associations between the –384A/Cpolymorphism and serum HDL-C and apoA-Ilevels, and between the 2237G/A polymor-phism and serum HDL-C levels.

An additional 13 polymorphisms were identi-fied in genomic DNA from immortalized B-lymphocyte cell lines from 93 subjects ofvarying ethnic backgrounds in theGennaissance Index Repository.39 DNA fromtwo cohorts of individuals with high and lowHDL, derived from subjects enrolled in theACCESS study, was examined for the sevenpolymorphisms which were found at greaterthan 5% frequency. The T229G exon 1 andC53T exon3 polymorphisms, as well as four


Table 1 Single nucleotide polymorphisms identified in the endothelial lipase gene summary of results fromdeLemos et al,37 Yamakawa-Kobayashi et al38 (in italics) and Mank-Seymour et al39 (in bold)

Polymorphism Frequency

Amino acid Reference Black White (high Japanese Hispanic Asianchange HDL white)

–410 C/G 37 0 0.3–303 A/C 37 1.8 0 (0.4)328 G/A Gly26Ser 37 5.7 0 (0.8)1145C/G Thr298Ser 37 2.3 0 (0.4)1439A/G Asn396Ser 37 0.6 1.2 (2.4)584C/T Thr111Ile 37, 38, 39 10.3, 9 31.2, 30 (32.6) 23 34229T/G 5'UTR 37, 39 12 35 31 3551C/T Int1 3755C/G Int1 3798C/A Int4 37, 39 12 22 11 12264C/A Ser4Ser 37, 39 0 14 0 0315C/T Ser21Ser 371922T/C Int2 37, 39 7 18 21 2642C/T Int5 37, 39 16 48 58 65840A/G Int7 53 42 42 622725T/C Int8 372237A/G 3'UTR 37, 38 13.863C/T Arg54Cys 39–384A/C 38, 39 2.12037T/C 38 7.72842T/A 38 6.23082T/C 38 6.5

The following were found in reference 39 only:–50G 2742A/G 2884G/T 5781C/A 83G/A 38C/G 669T/G 33G/A

Int1 Int1 Int4 Int5 Ala277Ala Int6 Int8120A/T Int8 3600G/A 304C/T

Int9 3'UTR

ch10 14/7/05 4:47 pm Page 108

intronic polymorphisms, were found to signifi-cantly associate with HDL levels. Two polymor-phisms were found to be potentially protectivefor myocardial infraction, the C53T(Thr111Ile) and G98A Intron4. They foundthat the contribution of EL gene variation tooverall variation in HDL-C was only 1–3%.

INHIBITION OF EL AS APHARMACOLOGICAL MEANS OFELEVATING HDL LEVELS

Collectively, studies that have examined theinfluence of EL on HDL particles confirm thatEL is able to modulate the lipid contents ofHDL and its function. Studies using mousemodels of EL overexpression or, conversely, ELgene deletion or inhibition, strongly supportthe conclusion that EL is a significant enzymein controlling the catabolism of HDL. Hence,pharmacological inhibition of EL expressionor activity may provide a novel way to raiseHDL levels.

If cytokine-induced upregulation of EL is acausal factor in reducing HDL levels seen ininflammatory conditions, one potentialapproach to the management of low HDL-Clevels under these conditions might be inhibi-

tion of EL expression through the use of anti-inflammatory agents. These agents may also beuseful to elevate HDL levels in patients withatherosclerosis since it is increasingly beingrecognized as a chronic inflammatory process.

The findings of Broedl et al40 must be consid-ered when entertaining the use of agents thatwould target EL for inhibition. They examinedthe effect of EL overexpression in three mousemodels of elevated apoB-containing lipopro-teins: apoE–/–, LDL-receptor–/– and humanapoB transgenics. They found that hepaticexpression of EL resulted in markedlydecreased levels of VLDL/LDL cholesterol,phospholipid and apoB, with increased catabo-lism of apoB-containing lipoproteins. Thissuggests that inhibition of EL might result inelevated levels of atherogenic lipoproteins andVLDL and LDL-C levels as well as HDL-C.

Extensive study of the mass and activity of ELin human subjects will be necessary to deter-mine the applicability of these findings tohuman cholesterol metabolism. The associa-tions between HDL levels and certain polymor-phisms in the EL gene suggest that EL may bequite relevant to human HDL metabolism.Extending these studies to examine associa-tions between EL gene variants and LDL levelswill be essential to determining therisk/benefit balance of EL inhibition.


1. Gordon T, Castelli WP, Hjortland MC, et al. Highdensity lipoprotein as a protective factor againstcoronary heart disease. The Framingham Study. Am JMed 1977; 62:707–14

2. Brooks-Wilson A, Marcil M, Clee SM, et al. Mutationsin ABC1 in Tangier disease and familial high-densitylipoprotein deficiency. Nat Genet 1999; 22:336–45

3. Bodzioch M, Orso E, Klucken J, et al. The gene encod-ing ATP-binding cassette transporter 1 is mutated inTangier disease. Nat Genet 1999; 22:347–51

4. Rust S, Rosier M, Funke H, et al. Tangier disease iscaused by mutations in the gene encoding ATP-binding cassette transporter 1. Nat Genet 1999;22:352–5

5. Calvo D, Vega MA. Identification, primary structure,and distribution of CLA-1, a novel member of the

CD36/LIMPII gene family. J Biol Chem 1993;268:18929–35

6. Acton SL, Scherer PE, Lodish HF, et al. Expressioncloning of SR-BI, a CD36–related class B scavengerreceptor. J Biol Chem 1994; 269:21003–9

7. Acton S, Rigotti A, Landschulz KT, et al.Identification of scavenger receptor SR-BI as a highdensity lipoprotein receptor. Science 1996; 271:518–20

8. Jaye M, Lynch KJ, Krawiec J, et al. A novel endothelial-derived lipase that modulates HDL metabolism. NatGenet 1999; 21:424–8

9. Hirata K, Dichek HL, Cioffi JA, et al. Cloning of aunique lipase from endothelial cells extends thelipase gene family. J Biol Chem 1999; 274:14170–5

10. McCoy MG, Sun GS, Marchadier D, et al.

References

ch10 14/7/05 4:47 pm Page 109

Characterization of the lipolytic activity of endotheliallipase. J Lipid Res 2002; 43:921–9

11. Ishida T, Zheng Z, Dichek HL, et al. Molecularcloning of nonsecreted endothelial cell-derived lipaseisoforms. Genomics 2004; 83:24–33

12. Bourne Y, Martinez C, Kerfelec B, 2004. Horsepancreatic lipase. The crystal structure refined at 2.3A resolution. J Mol Biol 1994; 238:709–32

13. van Tilbeurgh H, Roussel A, Lalouel JM, et al.Lipoprotein lipase. Molecular model based on thepancreatic lipase x-ray structure: consequences forheparin binding and catalysis. J Biol Chem 1994;269:4626–33

14. Hata A, Ridinger DN, Sutherland S, et al. Binding oflipoprotein lipase to heparin. Identification of fivecritical residues in two distinct segments of the amino-terminal domain. J Biol Chem 1993; 268:8447–57

15. Lookene A, Nielsen MS, Gliemann J, et al. Contributionof the carboxy-terminal domain of lipoprotein lipase tointeraction with heparin and lipoproteins. BiochemBiophys Res Commun 2000; 271:15–21

16. Duong M, Psaltis M, Rader DJ, et al. Evidence thathepatic lipase and endothelial lipase have differentsubstrate specificities for high-density lipoproteinphospholipids. Biochemistry 2003; 42:13778–85

17. Rye KA, Duong M, Psaltis MK, et al. Evidence thatphospholipids play a key role in pre-beta apoA-Iformation and high-density lipoprotein remodeling.Biochemistry 2002; 41:12538–45

18. Gauster M, Oskolkova OV, Innerlohinger J, et al.Endothelial lipase-modified high-density lipoproteinexhibits diminished ability to mediate SR-BI(scavenger receptor B type I)-dependent free-choles-terol efflux. Biochem J 2004; 382:75–82

19. Yu KCWDC, Kadambi SJ, Stahl A, et al. Endotheliallipase is synthesized by hepatic and aorta endothelialcells and its expression is altered in apoE-deficientmice. J Lipid Res 2004; 45:1614–23

20. Nagasaka H, Chiba H, Kikuta H, et al. Unique charac-ter and metabolism of high density lipoprotein(HDL) in fetus. Atherosclerosis 2002; 161:215–23

21. Tan KC, Shiu SW, Kung AW. Effect of thyroid dysfunc-tion on high-density lipoprotein subfraction metabo-lism: roles of hepatic lipase and cholesteryl ester transferprotein. J Clin Endocrinol Metab 1998; 83:2921–4

22. Bausserman LL, Bernier DN, McAdam KP, et al.Serum amyloid A and high density lipoproteinsduring the acute phase response. Eur J Clin Invest1988; 18:619–26

23. Barter P. Effects of inflammation on high-densitylipoproteins. Arterioscler Thromb Vasc Biol 2002;22:1062–3

24. Hirata K, Ishida T, Matsushita H, et al. Regulatedexpression of endothelial cell-derived lipase.Biochem Biophys Res Commun 2000; 272:90–3

25. Jin W, Sun GS, Marchadier D, et al. Endothelial cellssecrete triglyceride lipase and phospholipase activitiesin response to cytokines as a result of endotheliallipase. Circ Res 2003; 92:644–50

26. Feingold KR, Marshall M, Gulli R, et al. Effect ofendotoxin and cytokines on lipoprotein lipase activityin mice. Arterioscler Thromb 1994; 14:1866–72

27. Feingold KR, Memon RA, Moser AH, et al. Endotoxinand interleukin-1 decrease hepatic lipase mRNAlevels. Atherosclerosis 1999; 142:379–87

28. Azumi H, Hirata K, Ishida T, et al.Immunohistochemical localization of endothelialcell-derived lipase in atherosclerotic human coronaryarteries. Cardiovasc Res 2003; 58:647–54

29. Strauss JG, Hayn M, Zechner R, et al. Fatty acids liber-ated from high-density lipoprotein phospholipids byendothelial-derived lipase are incorporated intolipids in HepG2 cells. Biochem J 2003; 371:981–8

30. Maugeais C, Tietge UJ, Broedl UC, et al. Dose-depen-dent acceleration of high-density lipoprotein catabo-lism by endothelial lipase. Circulation 2003; 108:2121–6

31. Jin W, Millar JS, Broedl U, et al. Inhibition ofendothelial lipase causes increased HDL cholesterollevels in vivo. J Clin Invest 2003; 111:357–62

32. Ishida T, Choi S, Kundu RK, et al. Endothelial lipaseis a major determinant of HDL level. J Clin Invest2003; 111:347–55

33. Fuki IV, Blanchard N, Jin W, et al. Endogenouslyproduced endothelial lipase enhances binding andcellular processing of plasma lipoproteins via heparansulfate proteoglycan-mediated pathway. J Biol Chem2003; 278:34331–8

34. Strauss JG, Zimmermann R, Hrzenjak A, et al.Endothelial cell-derived lipase mediates uptake andbinding of high-density lipoprotein (HDL) particlesand the selective uptake of HDL-associated choles-terol esters independent of its enzymic activity.Biochem J 2002; 368:69–79

35. Broedl UC, Maugeais C, Marchadier D, et al. Effectsof nonlipolytic ligand function of endothelial lipaseon high density lipoprotein metabolism in vivo. J BiolChem 2003; 278:40688–93

36. Ma K, Cilingiroglu M, Otvos JD, et al. Endotheliallipase is a major genetic determinant for high-densitylipoprotein concentration, structure, and metabo-lism. Proc Natl Acad Sci USA 2003; 100:2748–53

37. deLemos AS, Wolfe ML, Long CJ, et al. Identificationof genetic variants in endothelial lipase in personswith elevated high-density lipoprotein cholesterol.Circulation 2002; 106:1321–6

38. Yamakawa-Kobayashi K, Yanagi H, Endo K, et al.Relationship between serum HDL-C levels andcommon genetic variants of the endothelial lipasegene in Japanese school-aged children. Hum Genet2003; 113:311–15

39. Mank-Seymour AR, Durham KL, Thompson JF, et al.Association between single-nucleotide polymor-phisms in the endothelial lipase (LIPG) gene andhigh-density lipoprotein cholesterol levels. BiochimBiophys Acta 2004; 1636:40–6

40. Broedl UC, Maugeais C, Millar JS, et al. Endotheliallipase promotes the catabolism of ApoB-containinglipoproteins. Circ Res 2004; 94:1554–61


ch10 14/7/05 4:47 pm Page 110

INTRODUCTION

The risk of atherosclerotic cardiovasculardisease (ASCVD) is strongly associated withplasma levels of both atherogenic and anti-atherogenic lipoproteins. Low density lipopro-teins (LDLs) are the paradigm of theatherogenic lipoprotein. After modification,they are taken up by macrophages within thearterial intima and promote the generation offoam cells, the classic cell type of the athero-sclerotic lesion. Therapy to reduce LDL choles-terol levels is proven to reduce the risk ofASCVD clinical events by about one third over5 years.1 A low level of HDL cholesterol (HDL-C) on statin therapy is an independent predic-tor of future cardiovascular events.2 Therefore,HDL-C is a major target for the development ofnew therapies designed to raise HDL-C levels.3

The cholesteryl ester transfer protein (CETP)transfers cholesteryl esters from HDL to apoB-containing lipoproteins such as LDL.Pharmacological inhibition of CETP raisesHDL-C levels in animals and humans andreduces atherosclerosis in animals; it iscurrently being studied for its ability to reduceatherosclerosis in humans. Here we review thephysiology of CETP and the issues around itsinhibition as a novel therapeutic strategy toraise HDL-C.

THE PHYSIOLOGY OF CETP

Lipoproteins contain a core of hydrophobiclipids (triglycerides and cholesteryl esters)surrounded by phospholipids and apolipopro-teins. Very low density lipoproteins (VLDLs)have as their major protein apoB-100. They aresecreted by the liver, their core triglyceridesare hydrolysed by lipoprotein lipase in muscleand adipose and they are eventually convertedto LDL. LDL also contains the same apoB-100protein, which binds to the LDL receptor andmediates its uptake by the liver. HDL and itsmajor protein apoA-I are secreted by the intes-tine and the liver. Lipid-poor apoA-I acquiresphospholipids and unesterified cholesterolfrom tissues through efflux mediated by ATP-binding cassette transporter proteins such asABCA1. The enzyme lecithin:cholesterolacyltransferase (LCAT) converts unesterifiedcholesterol on HDL to cholesteryl ester (CE).

HDL cholesteryl ester has several possibleroutes of metabolism. One is direct selectiveuptake into steroidogenic tissues and the livermediated by the cell-surface receptor SR-BI.Another is transfer via CETP to apoB-contain-ing lipoproteins such as VLDL and LDL. CETPpromotes exchange of cholesteryl esters fromHDL to apoB-containing lipoproteins inexchange for triglyceride which moves from

Cholesteryl ester transfer protein (CETP)inhibition as a therapeutic strategy forraising HDL-cholesterol levels andreducing atherosclerosisD.J. Rader

11

ch11 14/7/05 4:47 pm Page 111

apoB-containing lipoproteins to HDL. Theaction of CETP therefore influences lipopro-tein metabolism in a variety of ways. Itincreases the CE content of apoB-containinglipoproteins and promotes the generation ofsmall, dense LDL. It results in depletion of CEand enrichment of TG in HDL, making HDL abetter substrate for lipases such as hepaticlipase (HL) and resulting in the generation ofsmaller, denser HDL particles that are morerapidly catabolized via the kidney.4 Rodentslack CETP and the SR-BI pathway for selectiveuptake of HDL CE into the liver is very active.In species that have CETP (such as rabbits andprimates), the CETP pathway appears to bethe major route of metabolism of HDL CE andthe importance of the SR-BI pathway isuncertain.

One of the major questions about the physi-ology of CETP in humans regards its role inreverse cholesterol transport. In rabbits theCETP pathway accounts for a major amountof the clearance of HDL CE.5 Schwartz andcolleagues have done most detailed work inhumans regarding the pathways of HDL-Cmetabolism. Studies in healthy human volun-teers in which HDL was labelled with choles-teryl ester ex vivo and reinjected, followed bymultiple sampling of plasma and continuouscollection of bile for 6 days, revealed that thevast majority of tracer that appeared in bilewas derived after transfer to apoB-containinglipoproteins,6 presumably by CETP. Thus,transfer of HDL CE to apoB-containinglipoproteins may be a major route of deliveryof HDL CE to the liver, suggesting thatinhibiting CETP could potentially slow thatpathway. However, about 30% of cholesterolin HDL is unesterified, and studies bySchwartz and colleagues demonstrated thatthe majority of HDL unesterified cholesterolgets to the liver and bile via direct transferfrom HDL.6,7 Thus, substantial RCT couldtheoretically occur via direct transport byHDL of unesterified cholesterol to the liver, apathway for which CETP is not required.Indeed, both CETP deficiency and CETPinhibition are associated with absoluteincreases in HDL free cholesterol.

GENETIC DEFICIENCY OF ANDVARIATION IN CETP IN HUMANS:INSIGHT INTO THE ROLE OF CETPIN HUMAN LIPOPROTEINMETABOLISM ANDATHEROSCLEROSIS

Japanese subjects with extremely elevated levelsof HDL-C were found to have markely reducedCETP activity in plasma and subsequentlyfound to have mutations in the CETP gene thatwere the cause of the low CETP activity.8,9 Theinitial CETP mutations included a splice sitemutation of intron 14 and an aspartic acid toglycine substitution at the 442 position of exon15.10 Homozygotes for the intron 14 mutationhave no measurable CETP mass or activity intheir plasma. In contrast, homozygosity for theD442G mutation results in partial CETPdeficiency. Other mutations in CETP causingdeficiency have subsequently been described.Heterozygotes for loss-of-function mutations inthe CETP gene usually have slightly more thanhalf of the CETP activity of normal controlsand generally have only a modest increase inHDL-C levels. The finding that CETPdeficiency causes very high HDL-C levels leddirectly to the concept that pharmacologicalinhibition of CETP could be a novel strategyfor raising HDL-C levels.

Kinetic studies in patients with homozygousCETP deficiency have demonstrated that theturnover of apoA-I is significantly reduced inCETP-deficient subjects, as is the turnover ofthe second most abundant HDL apolipopro-tein, apoA-II.11 Even though CETP-deficientLDL had delayed catabolism in normalsubjects (consistent with decreased LDL recep-tor affinity), there was increased catabolism ofLDL apoB in CETP-deficient subjects, consis-tent with endogenous upregulation of theLDL-receptor.12

There is no consensus about whetherhomozygous CETP-deficient subjects are atreduced risk of cardiovascular diseasecompared with the normal population; indeed,one group in Japan believes they may be atincreased risk. In the Omigari region of Japan,of 201 patients with HDL-C levels >100 mg/dl,


ch11 14/7/05 4:47 pm Page 112

29 were homozygotes or compound heterozy-gotes for CETP deficiency. One of thesesubjects had clinical coronary heart disease(CHD),13 but was also found to have substan-tially reduced hepatic lipase activity.14 Giventhe relatively small number of homozygousCETP-deficient subjects in Japan, the issue oftheir cardiovascular risk relative to the generalpopulation may not be easily resolved. Thereare considerably greater numbers of heterozy-gous CETP-deficient subjects for addressingthis question. In the same series of 201Japanese patients from the Omigari regionwith HDL-C levels >100 mg/dl describedabove, 52 were heterozygotes for CETPdeficiency and eight had clinical CHD.13

However, all had reduced hepatic lipase activ-ity, suggesting that markedly elevated HDL-Cin the setting of heterozygous CETP deficiencyand reduced hepatic lipase may not necessarilyafford protection from CHD. In a population-based study of 104 505 persons in the Omigariregion of Japan, there was found to be anincreased prevalence of ECG changes insubjects with very high HDL-C levels,15 but thenumber of documented heterozygous subjectswas low. In contrast, in another study amongJapanese subjects with HDL-C >80 mg/dl and alow risk of cardiovascular disease, heterozygousCETP gene mutations were identified in aboutone quarter of the subjects,16 suggesting apossible protective effect.

The Honolulu Heart Study (which includesmany subjects of Japanese origin) initiallyreported in a cross-sectional study thatheterozygous D442G carriers with normalHDL-C concentrations of 40–60 mg/dl had anapparent increase in CHD compared withcontrols in the same HDL range.17 However, arecent prospective analysis reported after a 7-year follow-up that, after adjustment for otherCV risk factors, heterozygotes with the CETPD442G mutation had no evidence of increasedCHD risk at any HDL-C strata and, in fact, hadthe lowest rates of CHD (although this was notstatistically significant).18 Therefore, the issueof whether heterozygous CETP deficiency isassociated with altered cardiovascular risk hasnot been resolved.

Insights into the relationship between CETPand cardiovascular disease may also beprovided by studies of common polymor-phisms in the human CETP gene. The Taq1BRFLP in intron 1 of the human CETP gene isthe most extensively studied CETP polymor-phism. Subjects from the FraminghamOffspring Study carrying the Taq1B B2 allelehad lower CETP levels and higher HDL-Clevels and a reduced risk of CHD (in men)compared with homozygotes for the B1 allele.19

The VA-HIT study in men with CHD and lowHDL showed a similar result.20 The commonSNP resulting in conversion of the isoleucine atpostion 405 to valine (I405V) has also beenstudied extensively. CHD risk was higher inthose with the VV genotype among men in theHonolulu Heart Study21 and carriers of I405VSNP has been associated with increased HDL-Cand increased risk of CHD in women in theCopenhagen City Heart Study.22 In a case-control study of Ashkenazi Jews with excep-tional longevity, probands had a nearly 3-foldincreased frequency of homozygosity for theCETP 405V allele compared with controls, andalso had larger HDL and LDL particle size.23 Aformal meta-analysis of the Taq1B and I405VSNPs suggested that the Taq1B B2 allele wassignificantly associated with reduced CETPmass and activity and increased HDL-C levelsand the B2B2 genotype with possibly reducedcardiovascular disease, but suggested that therewas no convincing evidence that the I405V wasassociated with altered risk of cardiovasculardisease.24 Two additional CETP-coding SNPs,A373P and R451Q, are in linkage disequilib-rium. In one study, carriers of the rarer373P/451Q haplotype had reduced levels ofHDL-C and a lower CHD risk.25 Commonpromoter SNPs in the CETP gene, including–629A/C and a variable number tandemrepeat (VNTR) 1946 upstream of the transcrip-tion initiation site, are associated with variationin CETP mass and HDL-C levels,26,27 but notwith CHD. The human CETP gene divides intotwo linkage disequilibrium groups, a 5' blockand a 3' block.27,28 The 5' haplotype block,which contains the promoter polymorphismsand the Taq1B SNP, has a significant

CETP INHIBITION FOR RAISING HDL-CHOLESTEROL AND REDUCING ATHEROSCLEROSIS 113

ch11 14/7/05 4:47 pm Page 113

association with CETP mass and HDL-C levels,whereas the 3' haplotype block does not.Neither haplotype block has been shown tohave an unequivocal association with cardio-vascular risk. Thus, the relationship betweenCETP SNPs and atherosclerosis is variable anddoes not provide major insight into therelationship of CETP and ASCVD.

OBSERVATIONAL STUDIES OFPLASMA CETP LEVELS ANDCARDIOVASCULAR RISK

It is surprising that there have been relativelyfew observational studies of the association ofplasma CETP levels (mass or activity) withcardiovascular outcomes. Patients with lowerCETP mass had more rapid progression ofcarotid intimal medial thickness29 and angio-graphic coronary disease.30 A small coronaryangiographic study in Japanese patients foundno correlation of CETP mass levels with extentof coronary atherosclerosis.31 Another smallcase-control study in Chinese patients foundthat subjects with MI or stroke had higherCETP mass and activity compared to controls.32

A small case-control study found no differencebetween CETP activity levels in CHD cases andcontrols.33 The largest observational study ofplasma CETP levels and cardiovascular risk wasthe EPIC-Norfolk cohort study, in which CETPmass was measured in 755 originally healthyindividuals who developed CAD during follow-up and 1400 matched controls who remainedfree of CAD.34 CETP levels were inverselyrelated to HDL-C levels and directly related toLDL-C levels. Mean CETP levels were notsignificantly different in cases and controls;however, subjects in the highest quintile ofCETP mass had a 1.5-fold increased risk ofCAD compared to those in the lowest quintileafter adjusting for several cardiovascular riskfactors but not lipids. Furthermore, a stratifiedanalysis in subjects above or below the median(non-fasting) triglyceride level showed a signif-icant positive association for those above theTG median but no association of CETP levelsand CAD in those below the TG median.

CETP INHIBITION IN ANIMALMODELS: EFFECTS ON LIPIDS ANDATHEROSCLEROSIS

Rodents

Rodents such as mice and rats are naturallydeficient in CETP, and thus they cannot beused as models for addressing the effects ofCETP inhibition on atherosclerosis. Studies inwhich the CETP gene was introduced trans-genically into these species have uniformlydemonstrated reduction in HDL-C levels, buthave provided mixed results with regard toeffects on atherosclerosis. Expression of CETPin C57BL/6 mice fed an atherogenic diet,35

apoE knockout mice fed a chow diet,36 LDL-receptor knockout mice fed an atherogenicdiet36 and Dahl salt-sensitive hypertensive ratsfed an atherogenic diet37 all resulted inincreased atherosclerosis, suggesting that CETPis pro-atherogenic in rodents. However, expres-sion of CETP in human apoC-III transgenicmice, in diabetic mice and in LPL-deficientmice reduced atherosclerosis38 and expressionof CETP in LCAT transgenic mice also reducedatherosclerosis.39 Overall, the reasons for thesediscrepant results are unclear, but may relate tothe effects of CETP on the cholesterol contentin atherogenic VLDL remnant particles. In anycase, it is difficult to extrapolate from thesestudies in rodents to the potential effects ofinhibiting CETP in humans with regard toeffects on atherogenesis.

Rabbits

Rabbits have high levels of CETP, approxi-mately four times those found in humanplasma. Several approaches to CETP inhibitionhave been utilized in rabbits. Reduction ofCETP expression by injection of anti-senseoligodeoxynucleotides in rabbits resulted in anincrease in HDL-C and a reduction in aorticcholesterol content as a marker of atheroscle-rosis.40 Immunization of rabbits with a peptidebased on the CETP sequence generated auto-antibodies against CETP, resulting in reducedplasma CETP activity, increased HDL-C andreduced aortic atherosclerosis.41 Treatment of


ch11 14/7/05 4:47 pm Page 114

rabbits for 6 months with the small-moleculeCETP inhibitor JTT-705 reduced CETP activity,increased HDL-C by 90% and reduced aorticatherosclerosis by 70%.42 However, the inter-pretation of this study was confounded by thefact that JTT-705 decreased non-HDL-C levelssubstantially, and another control groupadministered simvastatin had a similar reduc-tion in non-HDL-C without a change in HDL-Cand also had significant reduction in athero-sclerosis. In another study in which rabbitswere made even more severely hypercholes-terolaemic, 3 months of treatment with JTT-705 did not reduce non-HDL-C levels and,although it raised HDL-C levels, it failed toreduce atherosclerosis.43 Finally, a study inrabbits with the small molecule CETP inhibitortorcetrapib showed increased HDL-C withoutreduction in non-HDL-C and neverthelessdemonstrated significant reduction in athero-sclerosis.44 Thus, CETP inhibition in rabbitsreduces atherosclerosis. Although the rabbithas much higher levels of CETP than humans,these data are among the most encouragingthat CETP inhibition may reduce atherosclero-sis in humans.

CETP INHIBITION IN HUMANS:EFFECTS ON LIPIDS ANDLIPOPROTEIN METABOLISM

A phase I clinical trial of a CETP peptideimmunization approach in 36 subjects reportedthat one subject who received a single injectionof the CETP peptide at the highest dose devel-oped anti-CETP antibodies.45 In an extensionstudy of 15 subjects who received a secondinjection of the active vaccine, eight developedanti-CETP antibodies. There were no changesin HDL-C levels in this phase I study.

Two small molecule inhibitors of CETP, JTT-705 and torcetrapib, have been studied inhumans and some of the data have beenpublished. In a phase II study of three doses ofJTT-705 (300, 600 and 900 mg) in 198 healthyvolunteers with modest hyperlipidaemia, JTT-705 900 mg for 4 weeks reduced CETP activityby 37%, decreased LDL-C by 7% and increased

HDL-C by 34%.46 JTT-705 also increased apoA-I levels modestly but significantly comparedwith placebo at each dose.

In a phase I study in healthy volunteers, torce-trapib was administered at doses of 10, 30, 60 and120 mg daily and 120 mg twice daily for 14 days.47

CETP inhibition ranged from 12% at the 30 mgdose to 80% at the 120 mg bid dose. There wasalso a dose-dependent increase in HDL-C,ranging from 16% to 91%. There was also a dose-dependent reduction in non-HDL-C, with 8%reduction in LDL-C at the 60 mg dose and 42%reduction of LDL-C at the highest dose.

Finally, a small phase II study designed toaddress the effects of torcetrapib on lipopro-tein metabolism in subjects with low HDL-C(<40 mg/dl) has been published.48 Subjectswith LDL-C levels >160 mg/dl were initiallytreated with 20 mg of atorvastatin daily forreduction in LDL-C (nine subjects) and theremaining subjects were on no statin therapy(10 subjects). All subjects received placebo for4 weeks, then 120 mg of torcetrapib daily for 4weeks; a subgroup of the subjects not takingatorvastatin went on to receive 120 mg of torce-trapib twice daily for another 4 weeks. Plasmalipids and lipoprotein kinetic studies wereperformed at the end of each 4-week period.Torcetrapib resulted in a 46% increase inHDL-C when given at 120 mg daily withoutatorvastatin, a 61% increase in HDL-C whengiven at 120 mg with atorvastatin and a 106%increase in HDL-C when given at 120 mg twicedaily without atorvastatin. Correspondingincreases in apoA-I were 16%, 13% and 36%.Torcetrapib elevated all types of HDL particles,but increased the large, more buoyant HDL2

particles to a greater extent than the smallerand denser HDL3 particles. Torcetrapib alsoreduced LDL-C and apoB by about 14–17%when administered with atorvastatin. Finally,torcetrapib significantly altered the distribu-tion of cholesterol among HDL and LDLsubclasses, resulting in increases in the meanparticle size of HDL and LDL in each cohort.Kinetic studies of HDL apolipoprotein metab-olism using endogenous labelling with deuter-ated leucine demonstrated slower catabolismof apoA-I after torcetrapib administration.49


ch11 14/7/05 4:47 pm Page 115

SUMMARY

It is now clear that pharmacological inhibition ofCETP in humans substantially increases HDL-Clevels. The question of whether CETP inhibitionwill reduce atherosclerosis burden and clinicalcardiovascular events in humans is one of themost important and fascinating questions in thefields of translational lipidology and atheroscle-rosis. Arguments can be made on both sides. Onone hand, CETP may indeed play an importantrole in reverse cholesterol transport, humansdeficient in CETP are not obviously protectedfrom atherosclerotic cardiovascular disease andthe rodent studies are equivocal in their resultsregarding the effects of CETP expression on

atherosclerosis. On the other hand, much RCTmay occur as free cholesterol, HDL may inhibitatherosclerosis through a variety of propertiesunrelated to RCT, observational studies suggestelevated CETP levels are associated withincreased cardiovascular risk and the rabbitstudies of CETP inhibition indicate significantreduction in atherosclerosis. Large randomizedclinical trials designed to assess the effects ofCETP inhibition on atherosclerosis burdenusing imaging modalities as well as on hard clini-cal cardiovascular events are under way. It will bea fascinating chapter in the progress towardtargeting HDL metabolism as a therapeuticapproach against atherosclerotic cardiovasculardisease.50,51


1. Expert Panel ATP-III. Executive Summary of TheThird Report of The National Cholesterol EducationProgram (NCEP) Expert Panel on Detection,Evaluation, And Treatment of High BloodCholesterol In Adults (Adult Treatment Panel III).JAMA 2001; 285:2486–97

2. Gordon DJ, Rifkind BM. High-density lipoprotein—the clinical implications of recent studies. N Engl JMed 1989; 321:1311–16

3. Sacks FM. The role of high-density lipoprotein (HDL)cholesterol in the prevention and treatment ofcoronary heart disease: expert group recommenda-tions. Am J Cardiol 2002; 90:139–43

4. Rye K-A, Clay MA, Barter PJ. Remodelling of highdensity lipoproteins by plasma factors. Atherosclerosis1999; 145:227–38

5. Goldberg DI, Beltz WF, Pittman RC. Evaluation ofpathways for the cellular uptake of high densitylipoprotein cholesteryl esters in rabbits. J Clin Invest1991; 87:331–46

6. Schwartz CC, VandenBroek JM, Cooper PS.Lipoprotein cholesteryl ester production, transferand output in vivo in humans. J Lipid Res 2004;45:1594–607

7. Schwartz CC, Zech LA, VandenBroek JM, et al.Cholesterol kinetics in subjects with bile fistula.Positive relationship between size of the bile acidprecursor pool and bile acid synthetic rate. J ClinInvest 1993; 91:923–38

8. Brown ML, Inazu A, Hesler CB, et al. Molecular basisof lipid transfer protein deficiency in a family withincreased high-density lipoproteins. Nature 1989;342:448–51

9. Inazu A, Brown ML, Hesler CB, et al. Increased high-density lipoprotein levels caused by a common choles-teryl-ester transfer protein gene mutation. N Engl JMed 1990; 323:1234–8

10. Inazu A, Jiang XC, Haraki T, et al. Genetic cholesterylester transfer protein deficiency caused by two preva-lent mutations as a major determinant of increasedlevels of high density lipoprotein cholesterol. J ClinInvest 1994; 94:1872–82

11. Ikewaki K, Rader DJ, Sakamoto T, et al. Delayedcatabolism of high density lipoprotein apolipoproteinA-I and A-II in human cholesteryl ester transferprotein deficiency. J Clin Invest 1993; 92:1650–8

12. Ikewaki K, Nishiwaki M, Sakamoto T, et al. Increasedcatabolic rate of low density lipoproteins in humanswith cholesteryl ester transfer protein deficiency. JClin Invest 1995; 96:1573–81

13. Hirano K, Yamashita S, Kuga Y, et al. Atheroscleroticdisease in marked hyperalphalipoproteinemia:combined reduction of cholesteryl ester transferprotein and hepatic triglyceride lipase. ArteriosclerThromb Vasc Biol 1995; 15:1849–56

14. Sakai N, Yamashita S, Hirano K, et al. Frequency ofexon 15 missense mutation (D442: G) in cholesterylester transfer protein gene in hyperalphalipopro-teinemic Japanese subjects. Atherosclerosis 1995;114:139–46

15. Hirano K, Yamashita S, Nakajima N, et al. Geneticcholesteryl ester transfer protein deficiency isextremely frequent in the Omagari area of Japan.Marked hyperalphalipoproteinemia caused by CETPgene mutation is not associated with longevity.Arterioscler Thromb Vasc Biol 1997; 17:1053–9

References

ch11 14/7/05 4:47 pm Page 116

16. Moriyama Y, Okamura T, Inazu A, et al. A low preva-lence of coronary heart disease among subjects withincreased high density lipoprotein cholesterol levels,including those with plasma cholesteryl ester transferprotein deficiency. Prev Med 1998; 27:659–67

17. Zhong S, Sharp DS, Grove JS, et al. Increasedcoronary heart disease in Japanese-American menwith mutation in the cholesteryl ester transfer proteingene despite increased HDL levels. J Clin Invest 1996;97:2917–23

18. Curb JD, Abbott RD, Rodriguez BL, et al. A prospec-tive study of HDL-C and cholesteryl ester transferprotein gene mutations and the risk of coronary heartdisease in the elderly. J Lipid Res 2004; 45:948–53

19. Ordovas JM, Cupples LA, Corella D, et al. Associationof cholesteryl ester transfer protein-TaqIB polymor-phism with variations in lipoprotein subclasses andcoronary heart disease risk. Arterioscler Thromb VascBiol 2000; 20:1323–9

20. Brousseau ME, O’Connor JJ Jr, Ordovas JM, et al.Cholesteryl ester transfer protein TaqI B2B2genotype is associated with higher HDL cholesterollevels and lower risk of coronary heart disease endpoints in men with HDL deficiency: Veterans AffairsHDL Cholesterol Intervention Trial. ArteriosclerThromb Vasc Biol 2002; 22:1148–54

21. Bruce C, Sharp DS, Tall AR. Relationship of HDL andcoronary heart disease to a common amino acidpolymorphism in the cholesteryl ester transferprotein in men with and without hypertriglyc-eridemia. J Lipid Res 1998; 39:1071–8

22. Agerholm-Larsen B, Nordestgaard BG, Steffensen R,et al. Elevated HDL cholesterol is a risk factor forischemic heart disease in white women when causedby a common mutation in the cholesteryl ester trans-fer protein gene. Circulation 2000; 101:1907–12

23. Barzilai N, Atzmon G, Schechter C, et al. Uniquelipoprotein phenotype and genotype associated withexceptional longevity. JAMA 2003; 290:2030–40

24. Boekholdt SM, Thompson JF. Natural genetic varia-tion as a tool in understanding the role of CETP inlipid levels and disease. J Lipid Res 2003; 44:1080–93

25. Agerholm-Larsen B, Tybjaerg-Hansen A, Schnohr P,et al. Common cholesteryl ester transfer proteinmutations, decreased HDL cholesterol, and possibledecreased risk of ischemic heart disease. TheCopenhagen City Heart Study. Circulation 2000;102:2197–203

26. Dachet C, Poirier O, Cambien F, et al. New functionalpromoter polymorphism, CETP/-629, in cholesterylester transfer protein (CETO) gene related to CETPmass and high density lipoprotein levels: role ofSp1/Sp3 in transcriptional regulation. ArteriosclerThromb Vasc Biol 2000; 20:507–15

27. Thompson JF, Lira ME, Durham LK, et al.Polymorphisms in the CETP gene and associationwith CETP mass and HDL levels. Atherosclerosis2003; 167:195–204

28. Corbex M, Poirier O, Fumeron F, et al. Extensive

association analysis between the CETP gene andcoronary heart disease phenotypes reveals severalputative functional polymorphisms and gene–environment interaction. Genet Epidemiol 2000;19:64–80

29. de Grooth GJ, Smilde TJ, van Wissen S, et al. Therelationship between cholesteryl ester transferprotein levels and risk factor profile in patients withfamilial hypercholesterolemia. Atherosclerosis 2004;173:261–7

30. Klerkx AHEM, de Grooth GJ, Zwinderman AH, et al.Cholesteryl ester transfer protein concentration isassociated with progression of atherosclerosis andresponse to pravastatin in men with coronary arterydisease (REGRESS). Eur J Clin Invest 2004; 34:21–8

31. Goto A, Sasai K, Suzuki S, et al. Cholesteryl ester trans-fer protein and atherosclerosis in Japanese subjects: astudy based on coronary angiography. Atherosclerosis2001; 159:153–63

32. Zhuang Y, Wang J, Qiang H, et al. Cholesteryl estertransfer protein levels and gene deficiency in Chinesepatients with cardio-cerebrovascular diseases. ChinMed J 2002; 115:371–4

33. Blankenberg S, Rupprecht HJ, Bickel C, et al.Common genetic variation of the cholesteryl estertransfer protein gene strongly predicts future cardio-vascular death in patients with coronary arterydisease. J Am Coll Cardiol 2003; 41:1983–9

34. Boekholdt SM, Kuivenhoven JA, Wareham NJ, et al.Plasma levels of cholesteryl ester transfer protein andthe risk of future coronary artery disease in appar-ently healthy men and women; the prospective EPIC-Norfolk population study. Circulation 2004;110:1418–23

35. Marotti KR, Castle CK, Boyle TP, et al. Severe athero-sclerosis in transgenic mice expressing simian choles-teryl ester transfer protein. Nature 1993; 364:73–5

36. Plump AS, Masucci-Magoulas L, Bruce C, et al.Increased atherosclerosis in apoE and LDL receptorgene knock-out mice as a result of human cholesterylester transfer protein transgene expression.Arterioscler Thromb Vasc Biol 1999; 19:1105–10

37. Herrera VL, Makrides SC, Xie HX, et al. Spontaneouscombined hyperlipidemia, coronary heart diseaseand decreased survival in Dahl salt-sensitive hyperten-sive rats transgenic for human cholesteryl ester trans-fer protein. Nat Med 1999; 5:1383–9

38. Hayek T, Masucci-Magoulas L, Jiang X, et al.Decreased early atherosclerotic lesions in hyper-triglyceridemic mice expressing cholesteryl estertransfer protein transgene. J Clin Invest 1995;96:2071–4

39. Foger B, Chase M, Amar MJ, et al. Cholesteryl estertransfer protein corrects dysfunctional high densitylipoproteins and reduces aortic atherosclerosis inlecithin cholesterol acyltransferase transgenic mice. JBiol Chem 1999; 274:36912–20

40. Sugano M, Makino N, Sawada S, et al. Effect ofantisense oligonucleotides against cholesteryl ester


ch11 14/7/05 4:47 pm Page 117

transfer protein on the development of atherosclero-sis in cholesterol-fed rabbits. J Biol Chem 1998;273:5033–6

41. Rittershaus CW, Miller DP, Thomas LJ, et al. Vaccine-induced antibodies inhibit CETP activity in vivo andreduce aortic lesions in a rabbit model of atheroscle-rosis. Arterioscler Thromb Vasc Biol 2000;20:2106–12

42. Okamoto H, Yonemori F, Wakitani K, et al. A choles-teryl ester transfer protein inhibitor attenuates ather-osclerosis in rabbits. Nature 2000; 406:203–7

43. Huang Z, Inazu A, Nohara A, et al. Cholesteryl estertransfer protein inhibitor (JTT-705) and the develop-ment of atherosclerosis in rabbits with severe hyperc-holesterolemia. Clin Sci 2002; 103:587–94

44. Morehouse LA, Sugarman ED, Bourassa PA, et al.HDL elevation by the CETP-inhibitor torcetrapibprevents aortic atherosclerosis in rabbits. Circulation2004; 110:III-243 (abstr)

45. Davidson MH, Maki K, Umporowicz D, et al. Thesafety and immunogenicity of a CETP vaccine inhealthy adults. Atherosclerosis 2003; 169:113–20

46. de Grooth GJ, Kuivenhoven JA, Stalenhoef AF, et al.

Efficacy and safety of a novel cholesteryl ester transferprotein inhibitor, JTT-705, in humans: a randomizedphase II dose-response study. Circulation 2002;105:2159–65

47. Clark RW, Sutfin TA, Ruggeri RB, et al. Raising high-density lipoprotein in humans through inhibition ofcholesteryl ester transfer protein: an initial multidosestudy of torcetrapib. Arterioscler Thromb Vasc Biol2004; 24:490–7

48. Brousseau ME, Schaefer EJ, Wolfe ML, et al. Effects ofan inhibitor of cholesteryl ester transfer protein onHDL cholesterol. N Engl J Med 2004; 350:1505–15

49. Brousseau ME, Diffenderfer MR, Millar JS, et al.Effects of cholesteryl ester transfer protein inhibitionon high-density lipoprotein subspecies andapolipoprotein A-I metabolism. Circulation 2004;110:III-144 (abstr)

50. van der Steeg WA, Kuivenhoven JA, Klerkx AH, et al.Role of CETP inhibitors in the treatment of dyslipi-demia. Curr Opin Lipidol 2004; 15:631–6

51. Rader DJ. High-density lipoproteins as an emergingtherapeutic target for atherosclerosis. JAMA 2003;290:2322–4


ch11 14/7/05 4:47 pm Page 118

INTRODUCTION

The fundamental role of insulin is to facilitatecellular uptake of glucose in skeletal muscle; inaddition, insulin suppresses hepatic gluconeo-genesis, the other key determinant of steady-state plasma glucose levels. In the steady (orfasted) state the quantity of insulin required tomaintain a plasma glucose level depends onmuscle mass and hepatic glucose output.However, there is more than a 2-fold variationin the plasma insulin levels required tomaintain identical plasma glucose levels innormal subjects.1 This variation in insulinrequirement for glucose disposal has beentermed insulin resistance, whereby subjectsneeding higher amounts of insulin are ‘insulinresistant’ compared to those who need lesseramounts of insulin. Insulin response is a linearvariable across populations; insulin resistance(or insulin sensitivity) is a relative concept innormal glucose-tolerant subjects, and there areno absolute cut-off values.

In pathogenic terms, insulin resistance is aprincipal feature of type 2 diabetes andprecedes the clinical development of thedisease by 10 to 20 years.2 Insulin resistance iscaused by the decreased ability of peripheraltarget tissues (muscle and liver) to respondproperly to normal insulin levels. Initially,increasing pancreatic insulin secretion is ableto counteract insulin resistance and thusnormal glucose homoeostasis can bemaintained. However, pancreatic reserveeventually diminishes in the face of increasing

peripheral demands and glucose concentra-tions rise, heralding a diagnosis of type 2diabetes once plasma glucose concentrationsgo beyond universally agreed diagnostic cut-offs, whether fasting or post-glucose loading.

It would be important to mention at thispoint that there are eponymous syndromes ofextreme insulin resistance which are due tomolecular defects in either the insulinmolecule, its receptor or in post-receptorsignalling. Plasma insulin levels are extremelyelevated in such circumstances.3 These subjectsoften have morphological abnormalities andhave a shortened life expectancy due to severedysregulation of intermediary metabolism. Forthe purposes of this chapter, however, we shalluse the term insulin resistance with regard tomore subtle changes in plasma insulin levels, ashappens across normal populations andsubjects with impaired glucose tolerance, type2 diabetes.

Insulin resistance is relevant not only to thepathogenesis of type 2 diabetes but also toatherogenesis and vascular disease. Althoughsomewhat described earlier by others, therelationship between serum insulin levels andcardiovascular disease was perhaps bestconceptualized by Gerald Reaven in hisBanting lecture of 1988.4 Reaven explainedthat insulin resistance and hyperinsulinaemiawere associated not only with glucose intoler-ance but also with hypertriglyceridaemia, lowHDL-cholesterol (HDL-C) and hypertension(Figure 1). He thus argued that insulin resis-tance may be causally related to the risk of

Overview of insulin resistance and the metabolic syndromeN. Sattar and B. Mukhopadhyay

12

ch12 14/7/05 4:47 pm Page 119

coronary heart disease (CHD).4 Similarly,others have suggested that both vasculardisease and diabetes arise from a ‘commonsoil’5 and such hypotheses currently form thebasis of our understanding of the central rolethat insulin resistance plays in the pathogenesisof diabetes and macrovascular disease.

This chapter will review recent advances inour understanding of the pathogenesis ofinsulin resistance, discuss its relevance to CHDand describe recent attempts to encapsulateinsulin-resistance-associated vascular risk by thecreation of metabolic syndrome criteria.Finally, therapeutic opportunities for tacklinginsulin resistance and metabolic syndrome willbe briefly reviewed.

FACTORS INVOLVED IN THEPATHOGENESIS OF INSULINRESISTANCE

Obesity

Obesity has long been considered the majorrisk factor for insulin resistance and type 2diabetes. For example, in a recent 10-yearfollow-up of the Bruneck study,6 the relativerisk for new onset type 2 diabetes in non-diabetic individuals at baseline with BMI>30 kg/m2 was 9.9 (95% CI 4.5 to 21.4) ascompared to similar subjects with BMI<25 kg/m2. In overweight subjects the relativerisk for type 2 diabetes was 3.4-fold (95% CI 1.8

to 6.3). The rising prevalence of obesity world-wide is contributing to substantially increasingrates of type 2 diabetes. Indeed, the totalnumber of people with diabetes is projected torise from 171 million in 2000 to 366 million in2030, an increase contributed to by bothhigher obesity rates and an aging population.7

At the other end of the age spectrum, manyrecent cases of type 2 diabetes in children havebeen reported8 and caused alarm in the UKand elsewhere.

But how does obesity increase the risk fordiabetes? In a study of 42 obese and 36 non-obese subjects, Bonadonna and colleagues9

demonstrated that in the physiological rangeof plasma insulin concentrations, the increasein total body glucose uptake and suppressionof hepatic glucose output were both signifi-cantly impaired in the obese group. Theglucose uptake and hepatic glucose outputcould be normalized by supraphysiologicalinsulin levels in obese subjects. Moreover, bothin the basal state and hyperglycaemic state,obese subjects were hyperinsulinaemic, whichdirectly correlated with the degree of insulinresistance.9 More recently, such findings havebeen corroborated in a study of 356 childrenaged 11 to 14 years.10 In that study, body massindex directly correlated with fasting insulinand inversely with insulin sensitivity. There wasalso a clustering of conventional cardiovascularrisk factors in those within the highest quartileof insulin resistance.10


hyperinsulinaemiahypertension

↓ ↑ triglyceride

insulin resistance

Atheroma

IGT/IFG ortype 2

HDL-cholesterol

Figure 1 CHD risk factorslinked to insulin resistance asinitially proposed by Reaven4

ch12 14/7/05 4:47 pm Page 120

At this stage it would therefore be safe to saythat obese subjects are more insulin resistantthan non-obese subjects across all age groups.However, the curvilinear relationship betweenBMI and insulin sensitivity is not precise. Forexample, by examining Figure 2 it can be seenthat even within a normal glucose tolerancecohort, the variability of insulin sensitivity atany given BMI is substantial.11 In this respectthe relevance of fat distribution to insulin resis-tance beyond total adiposity (or BMI) deservesdiscussion.

Body fat location: visceral versussubcutaneous

Many, but not all, clinical investigators havenoted that visceral adiposity is more stronglycorrelated with insulin resistance as comparedto subcutaneous adiposity. Visceral adiposetissue is morphologically and functionallydifferent from subcutaneous adipose tissue.Visceral fat originates from brown adiposetissue, has higher rates of lipolysis and glycolysisthan subcutaneous fat and drains directly intothe liver via the portal circulation.12 It is thusargued to be more relevant to the developmentof insulin resistance than subcutaneous fat.

Wajchenberg et al13 studied clinical charac-teristics and resistance to free fatty acid suppres-

sion in lean and obese non-diabetic women;obese subjects were further subdivided into twogroups, based on normal or increased amountof visceral fat. As predicted, obese women (withnormal visceral fat) had higher BMI, waist:hipratio, blood pressure and resistance to free fattyacid (FFA) suppression during oral glucosetolerance test (OGTT) in comparison to thelean controls. However, when the two BMI-similar obese groups were compared, thosewith higher visceral fat were noted to havegreater insulin resistance and higher plasmatriglycerides. Subsequently correlation analysisshowed that visceral fat contributed more toinsulin resistance than subcutaneous fat.13

A recent elegant study provided furtherevidence for the relative importance of regionaladiposity. Klein et al14 studied 15 obese womenbefore and 10 to 12 weeks after abdominalliposuction. Although liposuction decreasedthe volume of subcutaneous abdominal adiposetissue by significant amounts in both womenwith normal glucose tolerance and in thosewith diabetes, it did not significantly alter theinsulin sensitivity of muscle, liver or adiposetissue. By contrast, in animal studies, the surgi-cal removal of visceral adipose tissue resulted inmarked and nearly immediate improvements ininsulin resistance.15 In line with these observa-tions, Weiss et al have shown a near doubling of

OVERVIEW OF INSULIN RESISTANCE AND THE METABOLIC SYNDROME 121

16 26 36 46

BMI (kg/m2)

80

70

60

50

40

30

20

10

0

Insu

lin s

ensi

tivity

Figure 2 Inverse, curvilinearrelationship between obesity(quantitated by BMI) and insulinsensitivity (in arbitrary units11)across different stages of glucosetolerance. Grey squares, NGT (n =483); white circles, IGT (n = 71);black diamonds, type 2 diabetes(n = 65). Reproduced withpermission from Stumvoll99 withkind permission of SpringerScience and Business Media.

ch12 14/7/05 4:47 pm Page 121

the visceral to subcutaneous fat ratio in obesechildren with impaired glucose tolerancecompared to BMI-matched children withnormal glucose tolerance.16

There are two main inferences that can bedrawn from the above observations. First,visceral fat accumulation is significantly corre-lated to the development of insulin resistance.Second, body mass index is not an idealobesity-related marker of insulin resistancesince it gives an estimate of overall adiposity,not central (visceral) adiposity. As a result, anumber of investigators have proposed waistcircumference as a better marker of visceral fatand related metabolic complications.17–19 Infact, the extent of evidence is sufficientlyrobust for waist circumference to be recom-mended as a simpler screening tool for healthrisks linked to obesity inclusive of cardiovascu-lar disease and type 2 diabetes by the NationalInstitute of Health and Scottish IntercollegiateGuidelines Network.

Free fatty acids

Of course, one of the major roles of adipocytesis to store fat as triacylglycerol. Lipolysis ofstored fat releases FFAs into the circulation inthe fasting state. Insulin suppresses lipolysisand promotes esterification. It is well acceptedthat elevations in FFAs can impair insulin-mediated glucose uptake by cells.20,21

Moreover, FFAs might also impair beta cellfunction and release of insulin. In vivo studieshave shown that acute rises in FFAs increaseinsulin secretion whereas prolonged elevationcauses a reduction in glucose-stimulatedinsulin secretion from beta cells.22 This effectmay be more pronounced in genetically predis-posed individuals, such as those with impairedglucose tolerance who are thus at risk of devel-oping type 2 diabetes.23 Elevated FFAs alsoimpair insulin-mediated suppression of hepaticgluconeogenesis and augment glycogenolysis,thereby increasing glucose efflux from theliver.24 Visceral body fat is positively associatedwith increased FFA levels and insulin resis-tance/type 2 diabetes.25 Interestingly, FFA fluxfrom visceral fat is relatively resistant to

suppression by insulin in obese type 2 diabetespatients compared to obese non-diabeticindividuals.26 Finally there is evidence to showthat insulin clearance by the liver is impairedby circulating levels of FFAs.27

Overall, therefore, adiposity in general, andvisceral adiposity in particular, can impartinsulin resistance by decreasing cellularglucose uptake, increasing hepatic glucoseoutput and possibly impairing insulin secre-tion; all these three effects mediated in part byexcess FFAs released from visceral fat.

Muscle fat accumulation

In parallel with the growing interest in visceraladiposity as a mediator of insulin resistance,there is growing interest in the potential rolesof hepatic and intramyocellular (IMCL) fataccumulation. Muscle is an important site ofinsulin action. The relationship between skele-tal muscle triglyceride content and insulinresistance was first proposed in animal studiesby Falholt.28 Interestingly, this relationship wasnoted to be independent of obesity. Subse-quently they demonstrated that insulin-resis-tant type 2 diabetes patients had highertriglycerides in rectus abdominis musclecompared to control subjects.29

From ensuing work by others, it is now clearthat the degree of IMCL is inversely related tothe extent of this organ’s insulin sensitivityand positively related to the degree of visceralfat accumulation.30 It has been speculated thatelevated FFA delivery and/or impaired fattyacid (FA) oxidation result in IMCL fataccumulation of triacylglycerol and FAmetabolites (diacylglycerol, long-chain acylCoA), which are likely to induce defects in theinsulin signalling cascade, causing insulinresistance.30 In support of the relevance ofIMCL fat, a recent elegant study fromSchulman’s group demonstrated a signifi-cantly higher IMCL fat content in parallel with60% lower insulin sensitivity in offspring ofdiabetic parents relative to age- and body massindex (BMI)-matched healthy glucose tolerantcontrols.31 Of additional interest, usingmagnetic resonance spectroscopy the authors


ch12 14/7/05 4:47 pm Page 122

were able to demonstrate that, whereas lower-ing of FFA post-glucose load was similar inboth groups, the offspring of diabetic subjectshad 30% lower mitochondrial oxidativephosphorylation, indicative of lower betaoxidative capacity. The authors speculatedthat offspring of diabetic subjects have a lowerratio of type 1 (mostly oxidative) to type II(mostly glycolytic) fibres, and thereby have aninherited reduction in mitochondrial contentin muscle. These intriguing data clearlyrequire confirmation in other cohorts. Itshould be noted that the ‘insulin-resistant’offspring were generally of acceptable weight(mean BMI 23). An obvious interpretation isthat individuals genetically prone to insulinresistance require lower increments in BMI tomanifest adverse metabolic control. Putanother way, such individuals should have farmore incentive to maintain a healthy weightand remain physically active.

Hepatic fat accumulation

The role of the liver, and hepatic fat accumula-tion in particular, in the pathogenesis of type 2diabetes has attracted recent interest. In a recentstudy,32 liver fat content was shown to correlatewith several features of insulin resistance innormal weight and moderately overweightsubjects independent of BMI and intra-abdomi-nal or overall obesity. Moreover, the surprisinglycommon condition of non-alcoholic fatty liverdisease (NAFLD), prevalent in 5–20% of thepopulation and up to 75% of obese individualsor those with type 2 diabetes, has been shown tobe linked to insulin resistance.33

Why then does the liver accumulate fat? Onepossibility is simply excess flux of fatty acids tothe liver from abdominal or visceral fatdepots.34 However, others suggest thatincreased liver fat content may relate better todietary fat intake.35 In this respect, a recentstudy in rats demonstrated that even short-termfat feeding can lead to ~3-fold increase in livertriglyceride and total fatty acyl-CoA contentwithout any significant increase in visceral orskeletal muscle fat content.36 Further possibili-

ties for liver fat accumulation include excessiveintravascular lipolysis of triglyceride-richlipoproteins or indeed impaired FFA clear-ance. Whether acquired or genetic defects inhepatic β-oxidation are involved in liver fataccumulation, as they appear to be for IMCLfat accumulation, requires direct examination.Whatever the exact mechanism for fat accumu-lation, recent animal data indicate that thisincrease in hepatic fat leads to decreasedinsulin activation of glycogen synthase andincreased gluconeogenesis by inducing defectsin the insulin-signalling cascade.36

Interestingly, indirect estimates of liver fat(e.g. liver enzymes, in particular alanineaminotransferase (ALT)), are linked to hepaticinsulin resistance and predict diabetesindependent of other risk factors. Indeed, werecently demonstrated a graded association ofhigher ALTs with risk for new onset diabetesover a 4.9-year follow-up in the West ofScotland Coronary Prevention Study(WOSCOPS, Figure 3),37 a finding extendingprior observations from other groups.38,39

Thus, hepatic and muscle fat accumulationis highly relevant to the pathogenesis of insulinresistance and type 2 diabetes. Their place inthe scheme of abnormalities contributing totype 2 diabetes and CHD is shown in Figure 4.Clearly, future studies aimed at improving ourunderstanding of the mechanism(s) for suchfat accumulation might lead to novel therapiesto target insulin resistance.


0

1

2

3

4

5

0 1 2 3 4 5

Years

≥29 U/I

22–28 U/I

17–21 U/I<17 U/I

Per

cent

age

deve

lopi

ng d

iabe

tes

Figure 3 Association of ALT quartiles (U/L) with newonset diabetes in WOSCOPS (modified from reference37)

ch12 14/7/05 4:47 pm Page 123

Inflammatory factors

Although considerable research has addressedthe potential role of inflammatory mediators(e.g. C-reactive protein (CRP), interleukin-6,TNF-α) in the pathogenesis of coronary heartdisease,40 the relevance of this pathway to thepathogenesis of insulin resistance and type 2diabetes has only recently attracted interest. Ithas now been shown in cross-sectional studiesthat circulating inflammatory marker levelscorrelate with obesity and insulin resistance andare elevated in groups at risk of type 2 diabetes,i.e. offspring of diabetics, women with polycysticovary syndrome and subjects of South Asianorigin.41–43 The link between obesity and inflam-mation is intriguing and predominantly stemsfrom the observation that adipocytes release IL-6and TNF-α together with a range of ‘adipokines’,such as leptin, resistin and adiponectin, all ofwhich have relevance to insulin action.44

Beyond cross-sectional findings, severalprospective studies now indicate that C-reactiveprotein and white cell count, together withother acute phase markers, predict incidentdiabetes independently of established predic-

tors.45–47 One of the best examples of such datacomes from WOSCOPS.45 In this study, we wereable not only to show a graded relationshipbetween baseline CRP and new onset diabetes(Figure 5) but also that such prediction wasindependent of several routinely measuredparameters associated with diabetes risk,including fasting triglyceride, blood pressure,BMI and fasting glucose.45

In light of the foregoing associations, thesearch for mechanisms linking inflammationto insulin resistance and diabetes pathogenesishas become an important research avenue andseveral potential mechanisms (reviewed inreference 48), both direct and indirect, aresummarized in Table 1. Many of these illustratethe intimate links between inflammatory andnutrient pathways, with dysregulation of oneoften leading to dysregulation of the other. Forexample, cytokines can accelerate adipocytelipolysis and enhance hepatic fatty acid andtriglyceride synthesis. Cytokines may thereforecontribute to hepatic fat accumulation andinsulin resistance by this mechanism.

It is important to note, however, that insulinhas anti-inflammatory properties so that, rather


Lifestylefactors

Genes+/–

programming

Higher relative fatmass

±Altered location of fat

(visceral vssubcutaneous)

±Greater tendency to

muscle and hepatic fataccumulation due in

part to defects infat oxidation

↑ Insulin resistance↑ Glucose↑ Triglyceride↓ HDL-cholesterol

↑ C-reactive protein↑ IL-6↑ PAI-1↓ Adiponectin

↑ Diabetes&

↑ CHD

Figure 4 Individuals at risk for developing type 2 diabetes are likely to preferentially store fat in visceral stores andutilize fat less efficiently in muscle and liver, thereby accumulating fat in these organs. The consequences are greaterinsulin resistance and associated metabolic consequences for equivalent and even modest levels of total adiposity. Ofcourse, if such individuals are able to stay lean by remaining active and adhering to good diet, they will protectthemselves from such metabolic consequences as increasing obesity is germane to the expression of an underlyinggenetic potential towards insulin resistance

ch12 14/7/05 4:47 pm Page 124

than inflammation causing insulin resistance,the reverse may also be true.49 Further studiesare required to disentangle causal pathwaysand the advent of specific cytokine blockersmay help in this respect.

Adiponectin

Adiponectin is perhaps the most relevant ofthe spectrum of adipokines to the pathogenesisof insulin resistance. It is a recently identifiedadipose-tissue-derived protein with importantmetabolic effects.50 It is a 244-amino-acidprotein, which, despite being solely derivedfrom adipose tissue, is paradoxically reduced inobesity.50 Adiponectin enhances hepatic andmuscle fatty acid oxidation and thereby lessensfat accumulation in these organs. It does so byenhancing AMP-activated protein kinase(AMPK) activity.51 In line with this and previ-ously discussed observations on the relevanceof IMCL and hepatic fat accumulation, highadiponectin concentrations correlate withgreater insulin sensitivity and independentlypredict a reduced risk of type 2 diabetes.52

Moreover, recombinant adiponectin increasesinsulin sensitivity and improves glucose toler-ance in various animal models.53 Therefore,


0

1

2

3

4

5

0 0.5 1 1.5 2 2.5 3 3.5 4 4.5 5

Q1 : <0.66 Q2: >0.66 to Q3: >1.28 toQ4: >2.27 to Q5: > 4.18

% diabetic

Years in study

Quintiles of CRP (mg/l)

Figure 5 Association of baseline C-reactive protein (CRP) with new onset diabetes in WOSCOPS (modified fromreference 45)

Table 1 Potential mechanisms linking inflammation todiabetes development

Levels of acute phase mediators may reflect parallelrelease of other adipocyte products that cause insulinresistance

Direct effects of cytokines on insulin-signalling cascade Pro-inflammatory cytokine-mediated lipolytic effects and

hepatic de novo fatty acid synthesisPro-inflammatory cytokine-mediated endothelial

dysfunctionImpaired anti-inflammatory cytokine (IL-10) release and

thus potentially impaired endothelial function andinsulin signalling

IL-6-mediated increased glucocorticoid receptor densityand responsiveness

ch12 14/7/05 4:47 pm Page 125

low adiponectin may contribute to thedecrease in whole-body insulin sensitivity thataccompanies obesity in humans and reversal oralleviation of hypoadiponectinaemia mayrepresent a target for development of drugsimproving insulin sensitivity and glucose toler-ance. It is also relevant that adiponectin isstrongly anti-inflammatory, acting through theNFκB pathway,54 and downregulates adhesionmolecule expression on endothelial cells.55

Recent data have linked low circulatingadiponectin to the subsequent occurrence ofvascular events, independently of traditionalmarkers of CHD, but additional studies arerequired to confirm this.56

Plasminogen activator inhibitor 1

Increased plasminogen activator inhibitor 1(PAI-1) has been linked not only to thrombosisand fibrosis, but also to obesity and insulinresistance. Recent data suggest increased PAI-1levels predict diabetes independently ofobesity, insulin levels and indeed inflammatoryproteins.57 Furthermore, recent animal studieshave reported enhanced insulin sensitivity inPAI-1–/– mice relative to wild type on a high-fatand high-carbohydrate diet.58 Although directcausal mechanisms can presently only bespeculated, the authors argued that inhibitionof PAI-1 might provide a novel anti-obesity andanti-insulin resistance treatment.58

Endothelial dysfunction

The endothelium plays an important role inregulating blood flow and thus glucose uptakein insulin-sensitive tissues.59 Although thephysiological importance of the latter mecha-nism remains controversial, circulating factorsthat impede endothelial-dependent vasodilata-tion (such as FFAs or cytokines) may reduceglucose uptake in response to insulin.59,60 Insupport of the relevance of this pathway is arecent nested case-control analysis of theNurses’ Health Study examining plasma levelsof biomarkers reflecting endothelial dysfunc-tion (adhesion molecules) as predictors ofnew-onset diabetes.61 Meigs and colleagues

noted that baseline levels of several adhesionmolecules were significantly higher amongcases than among controls. More importantly,elevated E-selectin and ICAM-1 levels predictedincident diabetes in logistic regression modelsadjusted for BMI, family history of diabetes,smoking, diet score, alcohol intake, activityindex and post-menopausal hormone use.Further adjustment for waist circumferenceinstead of BMI or for baseline levels of CRP,fasting insulin and haemoglobin A(1c) did notalter these associations.61

Relevance of insulin resistance tocoronary heart disease

From the foregoing discussion, it is patently clearthat insulin resistance must be relevant to thepathogenesis of CHD; the associated elevatedtriglyceride, blood pressure and lower HDL-Care themselves sufficient to implicate insulinresistance in CHD pathogenesis, as elegantlydescribed by Reaven.4 Additionally, it is nowclear that many other parameters correlated withinsulin resistance and potentially relevant to itspathogenesis might also play a critical role inatherogenesis (e.g. inflammatory mediators,adhesion molecules, PAI-1, adiponectin).

Several studies have directly associatedinsulin resistance with CHD events. Forexample, the Verona Diabetes study62 showedthat, along with sex, age, smoking, HDL/totalcholesterol ratio and hypertension, HOMA-IR(see below) was an independent predictor ofboth prevalent and incident CHD. A 1-unitincrease in (log)HOMA-IR value was associatedwith an odds ratio for incident CHD duringfollow-up of 1.56 (95% CI 1.14–2.12, p <0.001).Similarly, hyperinsulinaemia predicted CHDrisk in Helsinki policemen over the 22-yearfollow-up, and to a large extent independentlyof other CHD risk factors.63

Measurement of insulin resistance

For clinical and study purposes, measurementof insulin resistance needs a reliable and repro-ducible method. The choice of techniquevaries among investigators. The following are


ch12 14/7/05 4:47 pm Page 126

the commonly employed techniques in theexperimental situation:

• Euglycaemic hyperinsulinaemic clamp.This procedure measures glucose disposalunder hyperinsulinaemic conditions, insulinbeing infused at a constant rate and 20%glucose at a variable rate to give a constant(clamped) plasma glucose level. Thisremains the gold standard test for assess-ment of whole-body insulin sensitivity.64

• HOMA. Homeostatic model assessmentconsiders fasting glucose and insulin levels:easier to perform and gives an estimate ofmainly hepatic insulin resistance.65

• Intravenous glucose tolerance test. One ofthe earlier techniques, insulin resistance isestimated from the insulin response toglucose load and fractional glucoseremoval; it has been largely superseded bythe above methods.66

• Fasting insulin resistance index (FIRI).Similar to HOMA, insulin resistance isassessed from fasting insulin and glucosevalues; it correlates well with the eugly-caemic clamp technique.67

• Finally, the quantitative insulin sensitivitycheck index (QUICKI) is derived fromlogarithmic-transformed fasting plasmaglucose (FPG) and insulin levels68 and is auseful index of insulin resistance incomparison with clamp-IR.

Advent of metabolic syndrome criteria

Clearly, many of the techniques needed tomeasure insulin resistance are complex andmost require insulin measurements. Samplesfor fasting insulin measurement need to becentrifuged and plasma separated rapidly(recommended 30 minutes). Such require-ments limit the use of many of thesetechniques for clinical application andepidemiological research. Thus, researchershave attempted to define sets of criteria toapproximate the dysfunctional metabolismassociated with insulin resistance. Severaldifferent names have been applied to suchcriteria but ‘metabolic syndrome’ is now the

favoured term. While many more factors, suchas small, dense LDL, C-reactive protein,plasminogen activator inhibitor-1, adiponectin,etc., have now been added to the list of riskfactors linked to metabolic syndrome andinsulin resistance (Figure 6), the core clinicaldrivers to identify individuals with metabolicsyndrome continue to be obesity, dyslipi-daemia, hypertension and markers of glucosedysregulation.

The first formal attempt to define themetabolic syndrome was made by the WorldHealth Organization (WHO).69 However, thisdefinition focuses on patients with existingevidence of glucose dysregulation, at whichstage the risk of conversion to diabetes isalready high (Table 2). The modified WHOcriteria allow metabolic syndrome to bediagnosed in individuals with normal glucosetolerance but the criteria are more complexand prescriptive, requiring documentedevidence of insulin resistance or at least asurrogate measure such as fasting insulin. Asmentioned above, the difficulties involved inobtaining these data will preclude widespreadclinical use of these criteria.

In 2001 the Adult Treatment Panel (ATP) IIIof the National Cholesterol EducationProgram (NCEP)70 proposed a new definition


↑ CRPIL-6, IL-18,

TNFfibrinogen

↑ PAI-1

↑TriglycerideLow HDLC

↑ Apo-BSmall LDL

Micro-albuminuria

↑ Insulin↑ Glucose

Hyperuricaemia

Hypertension

Decreasedadiponectin

Endothelialdysfunction

ICAM-1/vWF

METABOLICSYNDROME

Accelerated atherogenesis

Figure 6 Expanded list of parameters correlated withinsulin resistance (by no means comprehensive)

ch12 14/7/05 4:47 pm Page 127

of the metabolic syndrome using thresholds forfive easily measured variables linked to insulinresistance: waist circumference, triglyceride,HDL-C, fasting plasma glucose concentrationand blood pressure (Table 2). It is triggeredwhen pre-defined limits of any three criteriaare exceeded and, therefore, many suchindividuals will have normal fasting glucoseconcentrations. This definition allows popula-tion data to be more easily gathered.

There are now numerous studies using thesecriteria. The first four published studiesexamined the prevalence of metabolicsyndrome in middle-aged men and womenusing the NCEP criteria as shown in Table 3.Three of these71–74 suggested that around aquarter of men and women aged 50–55 yearshave metabolic syndrome, and others havesince confirmed this incidence. When thesedata were combined with US census data, it wasestimated that 47 million adults had themetabolic syndrome.75

Of course, the most important criterion fordefinitions of metabolic syndrome is whetherthey have any clinical utility. Several studieshave demonstrated that the NCEP criteriapredict vascular events or mortality71–74 (Table3). Moreover, such prediction appears to beindependent of classical risk factors in a smallnumber of studies thus far examined. This is

probably because the classification incorpo-rates variables such as BMI, triglyceride,glucose and diastolic BP that are not includedin current risk factor stratification. In addition,where examined,72 the NCEP definition wasmore strikingly correlated with diabetes thanwith CHD risk (Table 3), a valuable benefitsince predictive charts for diabetes are gener-ally lacking. This is despite the fact that the vastmajority of men diagnosed with metabolicsyndrome had ‘normal’ glucose levels atbaseline. This point is particularly importantsince, by the time impaired glucose tolerancehas appeared, the conversion rate to frankdiabetes is very high and opportunity forsuccessful intervention is limited.

It should be noted that metabolic syndromeas defined by NCEP is only modestly correlatedwith directly measured insulin resistance.76

Thus, whether individual clinicians will incor-porate the NCEP definition of metabolicsyndrome into clinical practice remains uncer-tain. There is no doubt that the NCEP defini-tions could be useful for recruiting individualsinto intervention trials, and indeed are beingused for this purpose. Lifestyle measures havealready been shown to reduce diabetes risk by58% in subjects at high risk for diabetes77,78 andmetformin77 and acarbose79 also have beneficialeffects, although these studies were conducted


Table 2 Definitions of the metabolic syndrome derived from the NCEP and WHO criteria

NCEP* definition of metabolic syndrome in men WHO* definition of metabolic syndrome in men

Presence of at least three (≥3) of the following:• Fasting plasma glucose ≥110 mg/dl (6.1 mmol/l)• Serum triglyceride ≥150 mg/dl (1.7 mmol/l)• Serum HDL-cholesterol <40 mg/dl (1.04 mmol/l)• Blood pressure† ≥130 systolic or ≥85 diastolic

or on medication• Abdominal obesity waist girth >102 cm

Presence of at least one (≥1) of the following:• Fasting plasma glucose ≥110 mg/dl (6.1 mmol/l)• Hyperinsulinaemia (upper quartile of fasting insulin

in non-diabetic population)• Diabetes or impaired glucose toleranceandPresence of at least two (≥2) of the following:• Hyperlipidaemia: serum triglyceride ≥150 mg/dl (1.7

mmol/l) or serum HDL-C <35 mg/dl (0.9 mmol/l)• Blood pressure†: ≥140 systolic or ≥90 diastolic or on

medication• Obesity: waist/hip ratio >0.90 or BMI ≥30 kg/m2

*NCEP, National Cholesterol Education Programme; WHO, World Health Organization† Blood pressure is measured in mm/Hg

ch12 14/7/05 4:47 pm Page 128


Tab

le 3

Prev

alen

ce a

nd

pred

icti

ve a

bilit

y of

met

abol

ic s

yndr

ome

usin

g re

cen

t de

fin

itio

ns

Inve

stig

ator

sC

rite

ria

Popu

latio

n st

udie

dM

ean

age

Ave

rage

Pr

eval

ence

of

Not

able

fea

ture

sR

R (

95%

CI)

for

inc

iden

t at

bas

elin

efo

llow

-up

met

abol

ic

CH

D e

vent

s/m

orta

lity

or

peri

odsy

ndro

me

at

diab

etes

in

subj

ects

with

ba

selin

em

etab

olic

syn

drom

e

Ford

et

al74

NC

EP

8814

men

an

d w

omen

––

24%

men

Hig

her

in

cide

nce

in

His

pan

ic m

en–

20 y

ears

or

olde

r23

% w

omen

and

espe

cial

ly H

ispa

nic

wom

enH

igh

er i

nci

den

ce b

y ag

e

Rid

ker

et a

l71N

CE

P*14

719

hea

lth

y w

omen

54

yea

rs8

year

s24

.4%

C

RP

leve

ls h

igh

er i

n t

hos

e w

ith

C

HD

eve

nts

>45

year

sm

etab

olic

syn

drom

e2.

3 (1

.6–3

.3)

to 4

.0

(3.0

–5.4

) fo

r C

RP

abov

ean

d be

low

3m

g/l

Satt

ar e

t al

72N

CE

P*64

47 m

en w

ith

hig

h

55 y

ears

4.9

year

s26

.2%

90

% w

ith

FB

G <

6.1

mm

ol/l

CH

D e

ven

tsch

oles

tero

lC

RP

leve

ls h

igh

er i

n t

hos

e w

ith

1.

8 (1

.4–2

.2)

to 1

.3m

etab

olic

syn

drom

e(1

.0–1

.7)†

Dia

bete

s: 3

.5 (

3.5

to 4

.9)

Lak

ka e

t al

74W

HO

1209

hea

lth

y m

en

52 y

ears

11.6

yea

rsW

HO

14.

2%U

ncl

ear

wh

y in

cide

nce

low

er,

CH

D m

orta

lity

NC

EP

aged

42–

60 y

ears

NC

EP

8.8%

but

lipid

s le

vels

app

eare

d su

peri

or

WH

O:

2.9

(1.2

–6.8

)**

to o

ther

gro

ups

NC

EP:

4.2

(1.

6–10

.8)*

*

WH

O,

Wor

ld H

ealt

h O

rgan

izat

ion

; N

CE

P, N

atio

nal

Ch

oles

tero

l E

duca

tion

Pro

gram

; C

RP,

C-r

eact

ive

prot

ein

*Stu

dies

use

d a

BM

I cu

t-off

in

pla

ce o

f th

e w

aist

th

resh

old

†Adj

uste

d fo

r ag

e, S

BP,

ch

oles

tero

l to

HD

L r

atio

an

d sm

okin

g**

Adj

uste

d fo

r ag

e, e

xam

inat

ion

yea

r, L

DL

-ch

oles

tero

l, sm

okin

g, f

amily

his

tory

of

CH

D

ch12 14/7/05 4:47 pm Page 129

in obese subjects who already had impairedglucose tolerance, in whom much of the cardio-vascular risk associated with metabolic dysfunc-tion had already been accrued.

Clearly, there is considerable potential forfurther refinement of metabolic syndrome defin-itions. For example, two trials have noted higherCRP levels in men and women with themetabolic syndrome.71,72 Moreover, CRPretained its independence as a predictor of CHDevents in women with or without the metabolicsyndrome71,72 (Figure 7). Further data demon-strated slightly improved prediction of CHD anddiabetes following reduction of the glucose cut-off from 6.1 mmol/l to 5.5 mmol/l.72 These dataare clearly not comprehensive but demonstratethe potential for refining the NCEP definition toincrease the predictive value for CHD and/ordiabetes. In other words, current metabolicsyndrome criteria should be considered to be‘works in progress’.73

TREATMENTS FOR METABOLICSYNDROME

Downstream targets

Of course, current clinical practice is predomi-nantly geared towards treating the downstreamconsequences of insulin resistance by individu-ally targeting the dyslipidaemia (lipid-loweringagents), blood pressure (anti-hypertensives),

elevated glucose (oral hypoglycaemic agents)and pro-thrombotic (anti-platelet agents)aspects of metabolic syndrome (Figure 8).However, it makes sense to consider treatmentsaimed at the upstream cause of metabolicsyndrome, namely insulin resistance. There issome preliminary evidence that specific anti-hypertensive agents (ACE inhibitors andangiotensin receptor blockers) may also lessenthe risk for type 2 diabetes.80 Similarly, thereare plentiful data to suggest statins and ACEinhibitors exert anti-inflammatory properties,but the clinical relevance of these effectsremain to be fully elucidated.

Upstream targets

Lifestyle changes

Ideally, to treat metabolic syndrome, oneshould target the precursor causes. This wouldinvolve increased physical activity combinedwith dietary improvement both in terms ofreduced caloric intake and altered composi-tion of diet. Such measures would result inweight loss and there are plentiful data todemonstrate improvements in a range ofmetabolic parameters secondary to lifestylechanges81,82 (Table 4). Of note, increasingphysical activity levels alone, i.e. without neces-sarily altering diet or losing weight, canimprove several metabolic-syndrome-related


0

2

4

6

8

10

12

14

0 1 2 3 4 5 6

Years

Per

cent

age

with

eve

ntCRP>=3, MS

CRP <3, MS

CRP>=3, no MS

CRP<3, no MS

Figure 7 Risk of CHDevents in relation tometabolic syndrome (MS)and elevated C-reactiveprotein (CRP) in WOSCOPS(modified from reference72)

ch12 14/7/05 4:47 pm Page 130


HyperinsulinaemiaHypertension

Hyperglycaemia

Insulin resistance

Antihypertensives

AtherosclerosisStatins

Triglyceridaemiasmall, dense LDL

low HDLInflammation

Hyper-coagulability

Insulin sensitizingagents or lifestyle

measures

Figure 8 Upstream (insulin sensitizing) vs downstream (e.g. anti-hypertensive, lipid-lowering) targeting of metabolicsyndrome

Table 4 Summary of metabolic and clinical effects of insulin sensitizing modalities

Agent Primary mode of action Other metabolic benefits Clinical trial data Ongoing studies

Physical activityand weight loss

Metformin

PPAR-γ agonists

Improve insulinsensitivity


Lower weight


Weight unchangedor slightlyincreased

Improved lipid profileLower inflammatorylevels

Improved vascular andhaemostatic function

Lower blood pressure

Improved lipid profileImproved haemostaticfunction

Some suggestion ofdecreased hepaticsteatosis

Improved lipid profileLower inflammatorylevels

Improved vascular andhaemostatic function

Lower blood pressureDecreased hepaticsteatosis

Lower risk fordiabetes

Reduce risk ofvascular events post-infarct as comparedto CABG

Lower risks for type2 diabetes

Lower risk of CHDevents in type 2diabetes

Reduce stentrestenosis

Reduced carotidIMT

Improvedendothelialfunction

To determine levels ofexercise needed tolessen CHD risk

Best mechanisms topromote increasedphysical activity (e.g. useof pedometers, etc.)

Studies in subjects withmetabolic syndrome

Ongoing work inpolycystic ovarysyndrome, a condition inyoung women linked toinsulin resistance,elevated risk for type 2diabetes and CHD

Vascular endpoint studiesin type 2 diabetes and insubjects with metabolicsyndrome

Studies in inflammatoryconditions

ch12 14/7/05 4:47 pm Page 131

parameters.83 Moreover, other evidencesuggests physical activity may be more impor-tant to vascular risk than initially appreciated.As an example, a recent randomized study tocompare the effects of exercise training versusstandard percutaneous coronary interventionwith stenting on clinical symptoms and clinicalvascular endpoints demonstrated significantlysuperior event-free survival in those random-ized to exercise.84

Insulin sensitizing agents: metformin

Metformin is an effective anti-diabetic drugthat lowers blood glucose concentrations bydecreasing hepatic glucose production andincreasing glucose disposal in skeletal muscle;however, the molecular site of metforminaction has only recently received attention.Recent studies suggest that, similar to theendogenously synthesized adiponectin,metformin increases AMPK activity.85,86 AMPKhas been implicated in the stimulation ofglucose uptake into skeletal muscle and theinhibition of liver gluconeogenesis. Inaddition, via the increase in APMK, metforminmight also increase hepatic fatty acid oxidationand may thus lessen hepatic fat accumulationas reported in one recent study,87 although amore recent trial did not show an effect ofmetformin on liver fat content.88 Interestingly,exercise also increases AMPK activity.

In terms of metformin’s metabolic effectsoutwith glucose lowering, there are data tosuggest it improves lipids (variably reducestriglyceride, LDL-C, apo B, remnants concen-trations and elevates HDL-C), decreasescoagualibility/thrombosis (in particular lowersPAI-1 and t-PA levels), and preliminary datasuggest it reduces some inflammation-relatedmolecules (CAMs, CRP).89,90 More impressiveare metformin’s effects in reducing risk fordiabetes77 and emerging evidence for its CHDevents risk reduction in diabetes subjects ascompared to non-insulin-sensitizing thera-pies.91,92

Metformin does not provoke hyperinsuli-naemia and it has a good clinical track record.It is now recommended as first-line therapy in

overweight patients with diabetes by mostleading clinical associations (e.g. SIGN guide-lines, Diabetes UK).

Insulin sensitizing agents: peroxisomeproliferator activated receptors

Peroxisome proliferator activated receptors(PPARs) are members of a superfamily ofnuclear hormone receptors. PPARs are ligand-activated transcription factors and have asubfamily of three different isoforms: PPAR-α,PPAR-γ and PPAR-β/δ. All isoforms hetero-dimerize with the 9-cis-retinoic acid receptorRXR and play an important part in the regula-tion of several metabolic pathways, includingadipocyte metabolism, lipid biosynthesis andglucose metabolism.93 They are discussed ingreater detail in Chapter 13.

In terms of evidence presented previously inthis chapter, it is relevant that a major mecha-nism of action of PPAR-γ agonists is to involvethe alteration of the tissue distribution of FFAuptake and utilization. Indeed, PPAR-γ agonistsinduce the differentiation of pre-adipocytesinto adipocytes and stimulate triglyceridestorage in expanded ‘safe’ subcutaneous butnot visceral fat depots.94 Moreover, they lessenhepatic and IMCL fat accumulation by enhanc-ing β-oxidation, an action that may be due totheir ability to significantly enhance adipo-nectin concentrations. Such effects in turnpromote glucose utilization. They increaseHDL-C and are also potently anti-inflamma-tory, acting to lower cytokine synthesis inadipocyte and other tissues.95

Thiazolidinediones are activators of PPAR-γand are now used as hypoglycaemic, muscleinsulin-sensitizing agents in type 2 diabetes.They are far more potent insulin sensitizersthan metformin. They have also been shown toexert anti-proliferative effects, and to antago-nize angiotensin II actions in vivo and invitro.96 The latter actions may help to explainevidence of clinically relevant blood pressurelowering by the thiazolidinediones, rosiglita-zone and pioglitazone, whereas metformindoes not appear to lessen blood pressure signif-icantly. Given the excellent spectrum of


ch12 14/7/05 4:47 pm Page 132

favourable metabolic actions of thiazolidine-diones, they are prime agents to be tested inindividuals with metabolic syndrome.Moreover, their relative efficacy in reducingCHD risk factors and vascular events, incomparison with conventional anti-diabeticagents, is of significant interest and relevantstudies are ongoing.97,98

CONCLUSION

There is now a major interest in insulin resis-tance as a candidate pathway in the pathogen-esis of vascular disease. Part of this intereststems from the rapidly increasing rates ofobesity worldwide which fuel insulin resistance,particularly in susceptible ‘at risk’ individuals.Insulin resistance is associated with a plethoraof metabolic perturbances (as summarized

herein) and many of these can directly orindirectly accelerate the atherogenic process.There is thus great clinical interest in assessingthe degree of insulin resistance in subjects atrisk for vascular disease or type 2 diabetes.Direct measurements of insulin resistance aregenerally unsuitable for widespread clinicaluse. Rather simple criteria based on readilymeasured factors associated with it have beenproposed under the broader term of metabolicsyndrome. Abundant data to suggest suchmetabolic syndrome criteria predict CHDevents and more strongly risk for type 2diabetes now exist. However, ongoing work isrequired to refine the existing criteria topredict CHD events and scrutinize clinicalapplication. In parallel, ongoing clinical trialswill determine the role of lifestyle factors andinsulin-sensitizing agents in reducing risk ofdisease in such high-risk subjects.


1. Hollenbeck C, Reaven GM. Variations in insulin-stimulated glucose uptake in healthy individuals withnormal glucose tolerance. J Clin Endocrinol Metab1987; 64:1169–73

2. Warram JH, Martin BC, Krolewski AS, et al. Slowglucose removal rate and hyperinsulinemia precedethe development of type II diabetes in the offspring ofdiabetic parents. Ann Intern Med 1990; 113:909–15

3. Taylor SI. Lilly lecture: molecular mechanisms ofinsulin resistance: lessons from patients withmutations in the insulin-receptor gene. Diabetes1992; 41:1473–90

4. Reaven GM. Banting lecture 1988. Role of insulinresistance in human disease. Diabetes 1988;37:1595–607

5. Stern MP. Diabetes and cardiovascular disease. The‘common soil’ hypothesis. Diabetes 1995; 44:369–74

6. Bonora E, Formentini G, Calcaterra F, et al. HOMA-estimated insulin resistance is an independent predic-tor of cardiovascular disease in type 2 diabeticsubjects: prospective data from the Verona DiabetesComplications Study. Diabetes Care 2002; 25:1135–41

7. Wild S, Roglic G, Green A, et al. Global prevalence ofdiabetes: estimates for the year 2000 and projectionsfor 2030. Diabetes Care 2004; 27:1047–53

8. Ehtisham S, Hattersley AT, Dunger DB, et al. BritishSociety for Paediatric Endocrinology and DiabetesClinical Trials Group. First UK survey of paediatric

type 2 diabetes and MODY. Arch Dis Child 2004;89:526–9

9. Bonadonna RC, Groop L, Kraemer N, et al. Obesityand insulin resistance in humans: a dose-responsestudy. Metabolism 1990; 39:452–9

10. Sinaiko AR, Jacobs DR Jr, Steinberger J, et al. Insulinresistance syndrome in childhood: associations of theeuglycemic insulin clamp and fasting insulin withfatness and other risk factors. J Pediatr 2001; 139:700–7

11. Matsuda M, DeFronzo RA. Insulin sensitivity indicesobtained from oral glucose tolerance testing: compar-ison with the euglycemic insulin clamp. Diabetes Care1999; 22:1462–70

12. Montague CT, O’Rahilly S. The perils of portliness:causes and consequences of visceral adiposity.Diabetes 2000; 49:883–8

13. Wajchenberg BL, Giannella-Neto D, da Silva ME, etal. Depot-specific hormonal characteristics of subcu-taneous and visceral adipose tissue and their relationto the metabolic syndrome. Horm Metab Res 2002;34:616–21

14. Klein S, Fontana L, Young VL, et al. Absence of aneffect of liposuction on insulin action and risk factorsfor coronary heart disease. N Engl J Med 2004;350:2549–57

15. Gabriely I, Ma XH, Yang XM, et al. Removal ofvisceral fat prevents insulin resistance and glucose

References

ch12 14/7/05 4:47 pm Page 133

intolerance of aging: an adipokine-mediated process?Diabetes 2002; 51:2951–8

16. Weiss R, Dufour S, Taksali SE, et al. Prediabetes inobese youth: a syndrome of impaired glucose toler-ance, severe insulin resistance, and altered myocellu-lar and abdominal fat partitioning. Lancet 2003;362:951–7

17. Han TS, van Leer EM, Seidell JC, et al. Waist circum-ference action levels in the identification of cardio-vascular risk factors: prevalence study in a randomsample. BMJ 1995; 311:1401–5

18. Sattar N, Tan CE, Han TS, et al. Associations of indicesof adiposity with atherogenic lipoprotein subfractions.Int J Obes Relat Metab Disord 1998; 22:432–9

19. Rexrode KM, Carey VJ, Hennekens CH, et al.Abdominal adiposity and coronary heart disease inwomen. JAMA 1998; 280:1843–8

20. Boden G, Jadali F, White J, et al. Effects of fat oninsulin stimulated carbohydrate metabolism innormal men. J Clin Invest 1991; 88:960–6

21. Dresner A, Laurent D, Marcucci M, et al. Effects offree fatty acids on glucose transport and IRS-1-associ-ated phosphatidylinositol 3–kinase activity. J ClinInvest 1999; 103:253–9

22. Carpentier A, Mittelman SD, Lamarche B, et al. Acuteenhancement of insulin secretion by FFA in humansis lost with prolonged FFA elevation. Am J Physiol1999; 276:E1055–66

23. Storgaard H, Jensen CB, Vaag AA, et al. Insulin secre-tion after short- and long-term low-grade free fattyacid infusion in men with increased risk of developingtype 2 diabetes. Metab Clin Exp 2003; 52:885–94l

24. Chen X, Iqbal N, Boden G. The effects of free fattyacids on gluconeogenesis and glycogenolysis innormal subjects. J Clin Invest 1999; 103:365–72

25. Pouliot MC, Despres JP, Nadeau A, et al. Visceralobesity in men. Associations with glucose tolerance,plasma insulin, and lipoprotein levels. Diabetes 1992;41:826–34

26. Basu A, Basu R, Shah P, et al. Systemic and regionalfree fatty acid metabolism in type 2 diabetes. Am JPhysiol Endocrinol Metab 2001; 280:E1000–6

27. Wiesenthal S, Sandhu H, McCall R, et al. Free fattyacids impair hepatic insulin extraction in vivo.Diabetes 1999; 48:766–74

28. Falholt K, Cutfield R, Alejandro R, et al. The effects ofhyperinsulinaemia on arterial wall and peripheralmuscle metabolism in dogs. Metabolism 1985;34:1146–9

29. Falholt K, Jensen I, Lindkaer JS, et al. Carbohydrateand lipid metabolism of skeletal muscle in type 2diabetic patients. Diab Med 1988; 5:27–31

30. Petersen KF, Shulman GI. Pathogenesis of skeletalmuscle insulin resistance in type 2 diabetes mellitus.Am J Cardiol 2002; 90:11G-18G

31. Petersen KF, Dufour S, Befroy D, et al. Impairedmitochondrial activity in the insulin-resistantoffspring of patients with type 2 diabetes. N Engl JMed 2004; 350:664–71

32. Seppala-Lindroos A, Vehkavaara S, Hakkinen AM, etal. Fat accumulation in the liver is associated withdefects in insulin suppression of glucose productionand serum free fatty acids independent of obesity innormal men. J Clin Endocrinol Metab 2002;87:3023–8

33. Medina J, Fernandez-Salazar LI, Garcia-Buey L, et al.Approach to the pathogenesis and treatment ofnonalcoholic steatohepatitis. Diab Care 2004;27:2057–66

34. Kelley DE, McKolanis TM, Hegazi RA, et al. Fatty liverin type 2 diabetes mellitus: relation to regionaladiposity, fatty acids, and insulin resistance. Am JPhysiol Endocrinol Metab 2003; 285:E906–16

35. Tiikkainen M, Bergholm R, Vehkavaara S, et al.Effects of identical weight loss on body compositionand features of insulin resistance in obese womenwith high and low liver fat content. Diabetes 2003;52:701–7

36. Samuel VT, Liu ZX, Qu X, et al. Mechanism ofhepatic insulin resistance in non-alcoholic fatty liverdisease. J Biol Chem 2004; 279:32345–53

37. Sattar N, Scherbakova O, Ford I, et al. Elevatedalanine aminotransferase predicts new-onset type 2diabetes independently of classical risk factors,metabolic syndrome and C-reactive protein in theWest of Scotland Coronary Prevention Study.Diabetes 2004; 53:2855–60

38. Ohlson LO, Larsson B, Bjorntorp P, et al. Risk factorsfor type 2 (non-insulin-dependent) diabetes mellitus.Thirteen and one-half years of follow-up of the partic-ipants in a study of Swedish men born in 1913.Diabetologia 1988; 31:798–805

39. Vozarova B, Stefan N, Lindsay RS, et al. High alanineaminotransferase is associated with decreased hepaticinsulin sensitivity and predicts the development oftype 2 diabetes. Diabetes 2002; 51:1889–95

40. Willerson JT, Ridker PM. Inflammation as a cardio-vascular risk factor. Circulation 2004; 109(Suppl1):II2–10

41. Kelly CCJ, Lyall H, Petrie JR, et al. Low-grade chronicInflammation in women with PCOS. J ClinEndocrinol Metab 2001; 86:2453–5

42. Forouhi NG, Sattar N, McKeigue P. Relation of C-reactive protein to cardiovascular risk factors inEuropeans and South Asians. Int J Obesity 2001;25:1327–31

43. Chambers JC, Eda S, Bassett P, et al. C-reactiveprotein, insulin resistance, central obesity, andcoronary heart disease risk in Indian Asians from theUnited Kingdom compared with European whites.Circulation 2001; 104:145–50

44. Lehrke M, Lazar MA. Inflamed about obesity. NatMed 2004; 10:126–7

45. Freeman DJ, Norrie J, Caslake MJ, et al. C-reactiveprotein is an independent predictor of risk for thedevelopment of diabetes in the West of ScotlandCoronary Prevention Study. Diabetes 2002;51:1596–600


ch12 14/7/05 4:47 pm Page 134

46. Pradhan AD, Manson JE, Rifai N, et al. C-reactiveprotein, interleukin-6, and risk of developing type 2diabetes mellitus. JAMA 2001; 286:327–34

47. Barzilay JI, Abraham L, Heckbert SR, et al. Therelation of markers of inflammation to the develop-ment of glucose disorders in the elderly: theCardiovascular Health Study. Diabetes 2001;50:2384–9

48. Sattar N, Perry CG, Petrie JR. Type 2 diabetes as aninflammatory disorder. Br J Diabet Vasc Dis 2003;3:36–41

49. Dandona P, Aljada A, Dhindsa S, et al. Insulin as ananti-inflammatory and antiatherosclerotic hormone.Clin Cornerstone 2003; (Suppl 4):S13–20

50. Weyer C, Funahashi T, Tanaka S, et al.Hypoadiponectinemia in obesity and type 2 diabetes:close association with insulin resistance and hyperin-sulinemia. J Clin Endocrinol Metab 2001; 86:1930–5

51. Yamauchi T, Kamon J, Minokoshi Y, et al.Adiponectin stimulates glucose utilization and fatty-acid oxidation by activating AMP-activated proteinkinase. Nat Med 2002; 8:1288–95

52. Lindsay RS, Funahashi T, Matsuzawa Y, et al.Adiponectin protects against development of type 2diabetes in the Pima Indian population. Lancet 2002;360:57–8

53. Yamauchi T, Kamon J, Waki H, et al. The fat-derivedhormone adiponectin reverses insulin resistanceassociated with both lipoatrophy and obesity. NatMed 2001; 7:941–6

54. Ouchi N, Kihara S, Arita Y, et al. Adiponectin, anadipocyte-derived plasma protein, inhibits endothe-lial NF-kappaB signaling through a cAMP-dependentpathway. Circulation 2000; 102:1296–301

55. Ouchi N, Kihara S, Arita Y, et al. Novel modulator forendothelial adhesion molecules: adipocyte-derivedplasma protein adiponectin. Circulation 1999;100:2473–6

56. Pischon T, Girman CJ, Hotamisligil GS, et al. Plasmaadiponectin levels and risk of myocardial infarction inmen. JAMA 2004; 291:1730–7

57. Festa A, D’Agostino R Jr, Tracy RP, et al. InsulinResistance Atherosclerosis Study. Elevated levels ofacute-phase proteins and plasminogen activatorinhibitor-1 predict the development of type 2diabetes: the insulin resistance atherosclerosis study.Diabetes 2003; 51:1131–7

58. Ma LJ, Mao SL, Taylor KL, et al. Prevention of obesityand insulin resistance in mice lacking plasminogenactivator inhibitor 1. Diabetes 2004; 53:336–46

59. Steinberg HO, Baron AD, Steinberg H, et al. Vascularfunction, insulin resistance and fatty acids.Diabetologia 2002; 45:623–34

60. Sattar N. Inflammation and endothelial dysfunction:intimate companions in the pathogenesis of vasculardisease? Clin Sci (Lond) 2004; 106:443–5

61. Meigs JB, Hu FB, Rifai N, et al. Biomarkers ofendothelial dysfunction and risk of type 2 diabetesmellitus. JAMA 2004; 291:1978–86

62. Bonora E, Formentini G, Calcaterra F, et al. HOMA-estimated insulin resistance is an independent predic-tor of cardiovascular disease in type 2 diabeticsubjects: prospective data from the Verona DiabetesComplications Study. Diabetes Care 2002; 25:1135–41

63. Pyorala M, Miettinen H, Laakso M, et al.Hyperinsulinemia predicts coronary heart disease riskin healthy middle-aged men: the 22–year follow-upresults of the Helsinki Policemen Study. Circulation1998; 98:398–404

64. DeFronzo R, Tobin J, Andres R. Glucose clamptechnique: a method for quantifying insulin secretionand resistance. Am J Physiol 1979; 237:E214–23

65. Mathews D, Hosker J, Rudenski A, et al. Homeostasismodel assessment: insulin resistance and β-cellfunction from plasma fasting glucose and insulinconcentrations in man. Diabetologia 1985; 28:412–19

66. Reaven GM, Olefsky JM. Relationship between insulinresponse during the intravenous glucose tolerancetest, rate of fractional glucose removal and the degreeof insulin resistance in normal adults. Diabetes 1974;23:454–9

67. Cleland S, Petrie J, Morris A, et al. FIRI: a fair insulinresistance index? Lancet 1996; 347:770

68. Katz A, Nambi SS, Mather K, et al. Quantitativeinsulin sensitivity check index: a simple, accuratemethod for assessing insulin sensitivity in humans. JClin Endocrinol Metab 2000; 85:2402–10

69. Alberti KG, Zimmet PZ. Definition, diagnosis andclassification of diabetes mellitus and its complica-tions. Part 1: diagnosis and classification of diabetesmellitus provisional report of a WHO consultation.Diabet Med 1998; 15:539–53

70. Expert Panel on Detection, Evaluation, andTreatment of High Blood Cholesterol in Adults.Executive Summary of The Third Report of TheNational Cholesterol Education Program (NCEP)Expert Panel on Detection, Evaluation, AndTreatment of High Blood Cholesterol In Adults(Adult Treatment Panel III). JAMA 2001;285:2486–97

71. Ridker PM, Buring JE, Cook NR, et al. C-reactiveprotein, the metabolic syndrome, and risk of incidentcardiovascular events: an 8-year follow-up of 14 719initially healthy American women. Circulation 2003;107:391–7

72. Sattar N, Gaw A, Scherbakova O, et al. Metabolicsyndrome with and without CRP as a predictor ofCHD and diabetes in West of Scotland CoronaryPrevention Study. Circulation 2003; 108:414–9

73. Sattar N, Forouhi NG. Metabolic syndrome criteria:ready for clinical prime time or work in progress? EurHeart J 2005 Apr 13 [Epub ahead of print]

74. Lakka HM, Laaksonen DE, Lakka TA, et al. Themetabolic syndrome and total and cardiovasculardisease mortality in middle-aged men. JAMA 2002;288:2709–16

75. Ford ES, Giles WH, Dietz WH. Prevalence of themetabolic syndrome among US adults: findings from


ch12 14/7/05 4:47 pm Page 135

the third National Health and Nutrition ExaminationSurvey. JAMA 2002; 16:356–9

76. Cheal KL, Abbasi F, Lamendola C, et al. Relationshipto insulin resistance of the adult treatment panel IIIdiagnostic criteria for identification of the metabolicsyndrome. Diabetes 2004; 53:1195–200

77. Knowler WC, Barrett-Connor E, Fowler SE, et al.Reduction in the incidence of type 2 diabetes withlifestyle intervention or metformin. N Engl J Med2002; 346:393–403

78. Tuomilehto J, Lindström J, Erickson JG, et al, for theFinnish Diabetes Prevention Study Group. Preventionof type 2 diabetes mellitus by changes in lifestyleamong subjects with impaired glucose tolerance. NEngl J Med 2001; 344:1343–50

79. Chiasson JL, Josse RG, Gomis R, et al. STOP-NIDDMTrial Research Group. Acarbose for prevention oftype 2 diabetes mellitus: the STOP-NIDDMrandomised trial. Lancet 2002; 359:2072–7

80. Teo K, Yusuf S, Anderson C, et al. ONTARGET/TRANSCEND Investigators. Rationale, design, andbaseline characteristics of 2 large, simple, random-ized trials evaluating telmisartan, ramipril, and theircombination in high-risk patients: the OngoingTelmisartan Alone and in Combination with RamiprilGlobal Endpoint Trial/Telmisartan RandomizedAssessment Study in ACE Intolerant Subjects withCardiovascular Disease (ONTARGET/TRANSCEND)trials. Am Heart J 2004; 148:52–61

81. Anderson JW, Kendall CW, Jenkins DJ. Importance ofweight management in type 2 diabetes: review withmeta-analysis of clinical studies. J Am Coll Nutr 2003;22:331–9

82. Tuomilehto J. Reducing coronary heart diseaseassociated with type 2 diabetes: lifestyle interventionand treatment of dyslipidaemia. Diabetes Res ClinPract 2003; 61 (Suppl 1):S27–34

83. Wannamethee SG, Lowe GD, Whincup PH, et al.Physical activity and hemostatic and inflammatoryvariables in elderly men. Circulation 2002;105:1785–90

84. Hambrecht R, Walther C, Mobius-Winkler S, et al.Percutaneous coronary angioplasty compared withexercise training in patients with stable coronaryartery disease: a randomized trial. Circulation 2004;109:1371–8

85. Musi N, Hirshman MF, Nygren J, et al. Metforminincreases AMP-activated protein kinase activity inskeletal muscle of subjects with type 2 diabetes.Diabetes 2002; 51:2074–81

86. Zhou G, Myers R, Li Y, et al. Role of AMP-activatedprotein kinase in mechanism of metformin action. JClin Invest 2001; 108:1167–74

87. Marchesini G, Brizi M, Bianchi G, et al. Metformin innon-alcoholic steatohepatitis. Lancet 2001; 358:893–4

88. Tiikkainen M, Hakkinen AM, Korsheninnikova E, etal. Effects of rosiglitazone and metformin on liver fatcontent, hepatic insulin resistance, insulin clearance,and gene expression in adipose tissue in patients withtype 2 diabetes. Diabetes 2004; 53:2169–76

89. Shin JJ, Rothman J, Farag A, et al. Role of oral anti-diabetic agents in modifying cardiovascular riskfactors. Minerva Med 2003; 94:401–8

90. Morin-Papunen L, Rautio K, Ruokonen A, et al.Metformin reduces serum C-reactive protein levels inwomen with polycystic ovary syndrome. J ClinEndocrinol Metab 2003; 88:4649–54

91. UK Prospective Diabetes Study (UKPDS) Group. Effectof intensive blood-glucose control with metformin oncomplications in overweight patients with type 2diabetes (UKPDS 34). Lancet 1998; 352:854–65

92. Kao J, Tobis J, McClelland RL, et al. Investigators inthe Prevention of Restenosis With Tranilast and ItsOutcomes Trial. Relation of metformin treatment toclinical events in diabetic patients undergoing percu-taneous intervention. Am J Cardiol 2004; 93:1347–50

93. Evans RM, Barish GD, Wang YX. PPARs and thecomplex journey to obesity. Nat Med 2004; 10:355–61

94. Berthiaume M, Sell H, Lalonde J, et al. Actions ofPPARgamma agonism on adipose tissue remodeling,insulin sensitivity and lipemia in absence of glucocor-ticoids. Am J Physiol Regul Integr Comp Physiol 2004;287:R1116–23

95. Marx N, Duez H, Fruchart JC, et al. Peroxisomeproliferator-activated receptors and atherogenesis:regulators of gene expression in vascular cells. CircRes 2004; 94:1168–78

96. Schiffrin EL, Amiri F, Benkirane K, et al. Peroxisomeproliferator-activated receptors: vascular and cardiaceffects in hypertension. Hypertension 2003; 42:664–8

97. Charbonnel B, Dormandy J, Erdmann E, et al.PROactive Study Group. The prospective pioglita-zone clinical trial in macrovascular events(PROactive): can pioglitazone reduce cardiovascularevents in diabetes? Study design and baseline charac-teristics of 5238 patients. Diabetes Care 2004;27:1647–53

98. Viberti G, Kahn SE, Greene DA, et al. A diabetesoutcome progression trial (ADOPT): an internationalmulticenter study of the comparative efficacy ofrosiglitazone, glyburide, and metformin in recentlydiagnosed type 2 diabetes. Diabetes Care 2002;25:1737–43

99. Stumvoll M. Control of glycemia: from molecules tomen. Minowski Lecture 2003. Diabetologia 2004;47:770–81.


ch12 14/7/05 4:47 pm Page 136

INTRODUCTION

Carbohydrates and lipids constitute the mainsource of energy for eukaryote organisms.Optimal energy homeostasis depends on a finecontrol of lipid and carbohydrate metabolismand is influenced by a diversity of environmen-tal and physiological conditions. Such regula-tion occurs in a co-ordinated fashion at severallevels of metabolic pathways by differentfactors triggering an adapted modulation ofmetabolic parameters. Some factors may affectenzyme activities while others act at the level ofgene transcription of proteins with keyfunctions in anabolic or catabolic pathways.Several transcription factors play pivotalregulatory functions in lipid and carbohydratemetabolism. The sterol regulatory elementbinding proteins (SREBPs) regulate theexpression of genes involved in cholesterol andfatty acid (FA) metabolism.1 The nuclearreceptors are a family of ligand-activatedtranscription factors involved in translating theeffects of lipid-soluble signalling molecules,such as hormones, vitamins, FAs and variousdrugs, at the gene expression level. Forinstance, the liver X receptors (LXRs) areactivated by cholesterol derivatives and controlgenes involved in cholesterol homeostasis.2 Inthis chapter, we will focus on the role of perox-isome proliferator activated receptors (PPARs)in energy metabolism. PPARs are FA-activatednuclear receptors3 that control lipid andglucose homeostasis via the modulation ofgene expression of various proteins orchestrat-

ing lipid biosynthesis, degradation, uptake,extracellular and intracellular transport andstorage. Dysfunctional regulation of thesegenes gives rise to metabolic disorders such asdyslipidaemia, glucose intolerance, hyperin-sulinaemia and obesity that ultimately lead tothe development of cardiovascular diseases.PPAR activity may also be modulated bysynthetic ligands and as such this class oftranscription factors offers interesting opportu-nities for therapeutic intervention.

MOLECULAR CHARACTERISTICS

Three distinct PPARs, α (NR1C1), β(δ),(NR1C2) and γ (NR1C3), each encoded by aseparate gene and displaying different tissue4–7

and developmental4 expression patterns, havebeen identified. PPAR-α, the first identifiedPPAR family member, is principally expressedin tissues exhibiting high rates of fatty acidoxidation such as liver, kidney, heart andmuscle. PPAR-γ, on the other hand, isexpressed at high levels in brown and whiteadipose tissue, and the intestine. PPAR-β/δ isexpressed in a wide range of tissues includingheart, adipose tissue, brain, intestine, muscle,spleen, lung and adrenal glands.8 All PPARsare also expressed, to variable extents, in thedifferent cell types of the vascular wall, includ-ing endothelial cells, smooth muscle cells andmonocytes/macrophages. Whereas PPAR-αand PPAR-β/δ modulate the transcription ofgenes implicated in lipid and lipoprotein

Peroxisome proliferator activatedreceptors and energy metabolismP. Gervois, J-C. Fruchart and B. Staels

13

ch13 18/7/05 5:56 pm Page 137

metabolism in liver and muscle,9 PPAR-γcontrols cellular differentiation, mediatesadipogenesis and modulates insulin action.10

Following ligand binding, PPARs associatewith the retinoic X receptor (RXR) to form aPPAR/RXR heterodimeric complex, capableof binding to a specific response element,termed peroxisome proliferator responseelement (PPRE), located in the promoterregion of target genes (Figure 1). Although themajority of PPREs identified so far consist of adirect repeat (DR) of the canonical AGGTCAsequence spaced by one nucleotide (DR-1),3

DR-2 elements may also function as PPRE.11

PPARs regulate target genes through this DNA-binding-dependent mechanism termed trans-activation. In addition, PPARs may repressgene transcription in a DNA-binding-indepen-dent manner via a transrepression mechanism.As such, PPARs interfere with other transcrip-tion factor pathways such as NF-κB, STAT,

C/EBP and AP-1. The transrepression medi-ated by PPARs occurs at different levels ofthese signal transduction pathways includingprotein–protein interaction between PPARsand these transcription factors leading to theformation of inactive complexes,12–18 competi-tion for cofactors, induction of IκB, the majorinhibitor of the NF-κB pathway, and down-regulation of plasma membrane receptor genetranscription (Figure 1).11,19,20

PPAR activity may be regulated at thetranscriptional, post-transcriptional andprotein level. In rats, PPAR-α expression isregulated by hormones, such as glucocorti-coids, insulin and leptin, by physiologicalstimuli such as stress and fasting and follows adiurnal rhythm.21–26 Although little is knownabout factors regulating PPAR-α expression inhumans, the observation that PPAR-α expres-sion in liver varies significantly among individ-uals27,28 suggests that PPAR-α is also strongly


Ligand

Plasma membranereceptor

1

2

4

PPAR

3TF

Inhibitor

TF

TF

TF PPAR RXR

Ligand

Transrepression Transactivation

PPRETFRE

RXR

Figure 1 Mechanisms of transcriptional gene regulation by PPARs. Activated PPARs form a heterodimeric complexwith the retinoic X receptor (RXR), which subsequently binds to a PPRE located within the regulatory sequence oftarget genes to modulate gene transcription (transactivation). PPARs negatively regulate gene transcription(transrepression) through several mechanism including (1) competition for co-factors, (2) protein–proteininteraction between PPAR with the STATs, NF-κB, C/EBPs or Fos/Jun transcription factors, (3) induction of theexpression of the NF-κB inhibitor IκB and (4) repression of membrane receptor expression. PPRE, peroxisomeproliferator response element; TFRE, transcription factor response element

ch13 18/7/05 5:56 pm Page 138

regulated at the gene level in humans bygenetic and environmental factors. Thetranscriptional regulation of PPAR-γ occurs viaalternative promoters, which results both inrodents and humans29,30 in the production oftwo distinct proteins, PPAR-γ1 and PPAR-γ2,with distinct activation capacities.31 Inaddition, glucocorticoids are implicated in theregulation of PPAR-γ expression during differ-entiation of 3T3-L1 preadipocytes.32 PPARactivity is also modulated at the protein level.Protein phosphorylation constitutes acommon post-translational mechanism ofregulation. Both PPAR-α and PPAR-γ activitiesdepend on their phosphorylation status, thataffects their transactivation potential andmodulates their biological functions.33–38 PPARprotein levels are also subject to regulation.PPAR-γ activators induce PPAR-γ receptorubiquitination and subsequent degradation bythe proteasome,39 whereas PPAR-α activatorsinhibit PPAR-α ubiquitination and increasePPAR-α protein half-life.40,41 The productionof isoforms with repressive activity on the wild-type receptor represents another mechanismof regulation of PPAR signalling occurring atthe protein level. Such an isoform, resultingfrom alternative splicing and giving rise to atruncated protein, has been identified forPPAR-α.27,28 This variant acts in a ligand-independent manner and alters PPAR-α wild-type transcriptional capacity.

PPARs are activated by natural ligandsderived from fatty acids (FAs), such as 8(S)hydroxyeicosatetraenoic acid, 8(S) hydroxy-eicosapentaenoic acid and leukotriene B4 forPPAR-α42 and prostaglandin-J2 and compo-nents of oxidized low-density lipoprotein(oxLDL) for PPAR-γ.42,43 Hypolipidaemic drugsof the fibrate class are synthetic agonists forPPAR-α and anti-diabetic glitazones are high-affinity synthetic ligands for PPAR-γ.42 Naturaland synthetic ligands have also been identifiedfor PPAR-β/δ,44,45 but these are not yet used inclinical practice. PPAR-β/δ ligands includenatural polyunsaturated FAs and prosta-glandins and synthetic ligands such ascarbaprostacyclin and other potent, subtype-specific agonists.46

PPARS AND THE CONTROL OFENERGY METABOLISM

PPAR-αα and intracellular metabolism

PPAR-α plays a pivotal role in the regulation ofintracellular lipid metabolism (Figure 2). Ahigh level of PPAR-α expression is observed intissues with elevated FA catabolism. PPREs havebeen identified in the promoter region ofrodent genes coding for enzymes implicated inthe peroxisomal β-oxidation pathway such asacyl-CoA oxidase (ACO), multi-functional

PEROXISOME PROLIFERATOR ACTIVATED RECEPTORS AND ENERGY METABOLISM 139

Adipose tissue

Liver

PPAR-α

Lipolysis

RCT

FA oxidation

FA uptake

TG synthesis

FA secretion

VLDL production

Figure 2 By modulating gene expression, PPAR-αdirectly or indirectly affects several regulatory processesthat maintain lipid homeostasis: (1) Limitation ofhepatic TG synthesis and VLDL production due toincreased fatty acid (FA) uptake, enhanced FAcatabolism and reduced FA synthesis. PPAR-α activationenhances FA β-oxidation and therefore diminishes theFA pool to be incorporated into triglyceride (TG)-richlipoproteins. Consequently, PPAR-α stimulates lipid fluxby controlling the FA flux from peripheral tissues suchas adipose tissue to the liver. (2) Induction oflipoprotein lipolysis as a result of either an increase inintrinsic lipoprotein lipase activity or an increasedaccessibility of TG-rich lipoprotein particles for lipolysisdue to reduced TG-rich lipoprotein apoC-III contentand induction of ApoA-V. (3) Increase in HDLproduction and stimulation of reverse cholesteroltransport. PPAR-α activators increase the production ofapolipoprotein A-I and A-II in human liver, which leadsto increased plasma HDL concentrations and enhancedreverse cholesterol transport (RCT)

ch13 18/7/05 5:56 pm Page 139

enzyme and 3-ketoacyl-CoA thiolase,3 linkingPPAR-α to the regulation of FA catabolism.Interestingly, these enzymes are also involvedin the biosynthesis of endogenous PPAR-αligands.47,48 PPAR-α also modulates genesinvolved in FA uptake, activation to acyl-CoAesters, mitochondrial β-oxidation and ketonebody synthesis.3,49 Intracellular FA concentra-tions are controlled, in part, by the activity ofthe FA transport protein-1 (FATP-1), whichcontrols the entry of FAs through the cellmembrane, and by acyl-CoA synthetase (ACS),which traps FAs inside the cells by their conver-sion to ester derivatives. PPAR-α activatesFATP-1 expression in liver and intestine andACS expression in liver and kidney.50 Theimplication of PPAR-α in FA transport wasfurther demonstrated by the lack of inductionof FATP and FA translocase mRNA in liver byPPAR-α activators in PPAR-α-null mice.51 FAmetabolism is also tightly linked to the rate ofmitochondrial FA uptake. PPAR-α has beendemonstrated to affect FA import intomitochondria by upregulating the expressionof the muscle-52–54 and liver-type carnitinepalmitoyltransferase I genes.25 Interestingly,further inhibition of mitochondrial FA importin PPAR-α-null mice causes hepatic and cardiaclipid accumulation, hypoglycaemia and deathin 100% of males and 25% of females.55

Furthermore, PPAR-α-deficient mice fed ahigh-fat diet display a massive accumulation oflipids in liver,25,56 highlighting the crucial rolethat PPAR-α plays in lipid metabolism. Aconcordant phenotype was observed in PPAR-α-deficient mice fasted for 24 hours whodisplayed hypoglycaemia, hypoketonaemia andelevated plasma FA levels.25 Other studiesrevealed that PPAR-α might increase energyexpenditure by upregulating the expression ofuncoupling proteins (UCPs).57–59 These datastrongly argue in favour of a critical role forPPAR-α in lipid homeostasis. Through theireffects on the expression of FA transporter andFA oxidation genes, PPAR-α directs the FA fluxto the β-oxidation pathway and thereforediminishes the FA pool to be incorporated intriglyceride (TG)-rich lipoproteins. Conse-quently, PPAR-α maintains lipid homeostasis

by controlling the FA flux from peripheraltissues, such as adipose tissue, to the liver(Figure 2).

PPAR-αα and extracellular lipidmetabolism

PPAR-α is an important regulator of extracel-lular lipid metabolism. The role of PPARs inTG metabolism is convincingly established.3

The hypotriglyceridaemic effect of PPAR-αactivators is the result of increased lipoproteinlipolysis and enhanced FA oxidation. Theprocess of TG clearance is directly under thecontrol of lipoprotein lipase (LPL) and apoC-III. PPAR-α activators repress apolipoproteinC-III (apoC-III) expression60,61 and induce LPLin liver.62 Moreover, PPAR-α enhances FAcatabolic rate, which subsequently affects TGsynthesis and VLDL production.3 PPAR-α playsa central role in hepatic apoC-III gene regula-tion.63 ApoC-III repression by PPAR-α occursvia PPAR-α-dependent induction of humanand rat Rev-erbα, a nuclear orphan receptorthat is a strong repressor of transcription.11,64

The induction of Rev-erbα is mediated byPPAR-α, which interacts with a PPRE locatedwithin the human Rev-erbα promoter.11 Such amechanism is consistent with the observationthat Rev-erbα-deficient mice exhibit increasedplasma TG and apoC-III concentrations andelevated hepatic apoC-III mRNA levels.65

Therefore, by upregulating Rev-erbα, PPAR-αindirectly modulates apoC-III expression, aneffect that may contribute to the normoli-paemic action of PPAR-α activators. Recently, anew apolipoprotein, apoA-V, has been identi-fied to be an important determinant of plasmatriglycerides.66–68 Vu-Dac et al demonstratedthat fibrates induce apoA-V expression inhuman hepatocytes.69 These findings define anovel mechanism via which PPAR-α activatorscan influence triglyceride homeostasis.

Both apoA-I and HDL are inversely corre-lated with the incidence of coronary arterydisease.70 There is evidence supporting theprotective role of HDL through the removaland the recycling of cholesterol excess fromperipheral tissues to the liver. PPAR-α has


ch13 18/7/05 5:56 pm Page 140

important actions on HDL metabolism. PPAR-α activators affect HDL metabolism in anopposite manner in rodents and humans.Whereas fibrate treatment of rats lowersplasma HDL, an increase is generally observedin humans.71–73 Such an increase in plasmaHDL levels is related, at least in part, tochanges in apoA-I and apoA-II gene expressionin liver. The human apoA-I gene contains aPPRE that mediates upregulation by PPAR-αactivators.74,75 Studies on the regulation ofhepatic apoA-I gene expression, carried out inhuman apoA-I transgenic mice, demonstratedan induction of plasma HDL and human apoA-I concentrations by PPAR-α activators.74 PPAR-α also influences the expression of humanapoA-II, the other major protein component ofHDL.76 Administration of fenofibrate topatients with coronary artery disease results ina marked increase in plasma apoA-II concen-trations. This increase in plasma apoA-II is dueto a direct effect on hepatic apoA-II produc-tion, since fibrates induce apoA-II mRNA levelsboth in primary cultures of human hepatocytesand in human hepatoblastoma HepG2 cells.76

The induction of apoA-II mRNA levels isfollowed by an increase in apoA-II secretion inboth cell culture systems. PPAR-α binds withhigh affinity to a PPRE located in the humanapoA-II promoter, thereby activating apoA-IIgene transcription. These data demonstratethat PPAR-α activators increase human apoA-IIplasma levels by stimulating transcription of itsgene through the interaction of activatedPPAR-α with the apoA-II-PPRE.

The reverse cholesterol transport (RCT)pathway mediates the transport of cholesterolfrom peripheral tissues to the liver. Peripheralcell cholesterol is captured by HDL particles,which bring it back to the liver for direct elimi-nation into the bile or, after metabolism, in bileacids. PPARs influence the RCT pathways byregulating macrophage cholesterol efflux, HDLtransport in plasma and bile acid synthesis.Cholesterol efflux is the first step of the RCTthat occurs either via passive diffusion or viatransmembrane receptors such as the scavengerreceptor (SR) class B type 1 (SR-B1/CLA-1)and the ATP-binding cassette A1 (ABCA1)

proteins. Murine scavenger receptor BI (SR-BI)and its human homologue CLA-1 have beenidentified as HDL receptors, which bind HDLwith high affinity and mediate the selectiveuptake of cholesteryl esters in liver andsteroidogenic tissues.77,78 Interestingly, PPAR-αis expressed in human macrophages79 andinduces the expression of SR-B1/CLA-180 andABCA1.46,81,82 The induction of ABCA1 occursvia an indirect mechanism involving inductionof LXR-α expression,83 a major regulator ofABCA1. Increased ABCA1 expression results ina higher cholesterol efflux from macrophages.The major contribution of ABCA1 in thecontrol of plasma HDL-cholesterol levels washighlighted by the identification of mutationsin the ABCA1 gene in patients with familialHDL deficiency and Tangier disease.

In addition to influencing the transcription ofHDL apolipoproteins and receptors regulatingHDL metabolism, PPARs also affect the expres-sion of enzymes involved in HDL metabolism. Inrats, PPAR-α activators decrease the productionof HDL-remodelling enzymes, such as hepaticlipase and lecithin:cholesterol acyltransferase(LCAT).84,85 PPAR-α also induces the expressionand activity of phospholipid transfer protein(PLTP),86 an enzyme promoting the transfer ofphospholipids from VLDL/LDL to HDL inmice. However, gene regulation studies onPLTP and LCAT by PPARs have beenperformed so far only in rodent models andneed to be investigated in humans.

PPAR-αα adiposity and steatosis

Although the expression of PPAR-α is very lowin white adipose tissue, several arguments arein favour of a function of PPAR-α in thecontrol of adiposity. PPAR-α may also play arole in adipose tissue as a mediator of leptin-induced lipolysis.23 Indeed, hyperleptinaemiain rodents depletes adipocyte fat while upregu-lating FA oxidation enzymes, UCPs and PPAR-α, whose expression is normally low in whiteadipocytes. The essential role of PPAR-α inmediating the actions of leptin on lipid metab-olism in liver and adipose tissue was shown byLee et al.87 More directly, PPAR-α activators


ch13 18/7/05 5:56 pm Page 141

reduce adiposity in rodent models ofobesity.88–90 Despite similar degrees of hyper-leptinaemia and reduction in food intake,epididymal fat was strongly reduced in wild-type but only weakly in PPAR-α-deficient mice.Accumulation of FAs in the liver is controlledby an equilibrium between hepatic lipid secre-tion and intracellular lipid metabolism path-ways. Potentiation of FA oxidation preventsconversion into triglycerides and hepaticaccumulation of lipids. PPAR-α exerts anti-steatotic actions in mice. PPAR-α plays a criticalrole in the cellular response to fasting. Uponshort-term starvation, PPAR-α-deficient miceexhibit hepatic steatosis, myocardial lipidaccumulation and hypoglycaemia with aninadequate ketogenic response.56 Moreover,diet-induced liver triglyceride accumulation isdecreased by PPAR-α activation in wild-type butnot in PPAR-α-deficient mice that displaysevere steatohepatitis.91 In addition, Chou et alperformed similar investigations in a lipoat-rophic mouse model.92 Activation of PPAR-αled to reduced triglyceride levels in liver andmuscle, an effect attributed to an increasedexpression of FA oxidation enzymes ratherthan reduced lipogenesis or lipid uptake.Theseobservations suggest that PPAR-α maymodulate obesity-linked disorders such assteatosis, whether such effects also occur inhumans awaits further studies.

PPAR-ββ//δδ

As PPAR-β/δ is ubiquitously expressed, numer-ous and various functions have been attributedto this receptor. PPAR-β/δ was found tomodulate tumorigenic pathways in the APCmodel of intestinal adenomatous polyposis.93

PPAR-β/δ also influences blastocyst implanta-tion.94 PPAR-β/δ may also control cell prolifera-tion in skin.95 However, recent studies identifiedspecific functions for this nuclear receptor inenergy metabolism as well (Figure 3).

PPAR-ββ//δδ and adiposity

PPAR-β/δ plays an important role in adiposetissue. Transgenic overexpression of a VP16-

PPAR-β/δ chimeric protein in adipocytesresults in a reduction of body weight andadiposity.96 Upon high-fat-diet feeding, thesemice are resistant to weight gain, have anormal lipid profile, display no hypertrophy ofwhite and brown adipose tissue and develop nofatty liver. Crossing these transgenic mice intothe db/db background of a genetically predis-posed mouse model results in a reversal of theobesity phenotype. In contrast, PPAR-β/δ-deficient mice fed a high-fat diet displayreduced energy uncoupling and are prone toobesity.96 In addition, pharmacological activa-tion of PPAR-β/δ in db/db mice reduces thesize of lipid droplets in brown fat and preventslipid accumulation in liver due to enhanced FAoxidation. PPAR-β/δ treatment reversed thediet-induced obesity and insulin resistance inmice.97 PPAR-β/δ can thus be considered as aregulator of fat burning by coordinating FAoxidation and energy uncoupling. However,the relevance of these findings in rodentmodels, that have substantial amounts ofbrown adipose tissue, cannot be easily


PPAR-β

FA oxidation FA oxidation

Energy expenditure

Lipid accumulation

Liver

Muscle

Adipose tissue

RCT Fibre number

Figure 3 PPAR-β/δ functions in energy metabolism.PPAR-β/δ regulates the expression of genes implicated inβ-oxidation, energy uncoupling and cholesterol efflux.Activation of PPAR-β/δ in muscle results in an increasein fibre number displaying high oxidative metabolismand improves exercise performance. In adipose tissue,PPAR-β/δ acts as a regulator of fat burning by co-ordinating FA oxidation and energy uncoupling. PPAR-β/δ stimulates cholesterol efflux from the cells, uptakeby HDL and transport back to the liver. Although lessexpressed than PPAR-α, PPAR-β/δ may enhance β-oxidation in liver

ch13 18/7/05 5:56 pm Page 142

extended to the human situation and investi-gation is required in other animal models suchas the insulin-resistant monkeys.

PPAR-ββ//δδ and lipid metabolism

Recently, in vitro and in vivo genetic andpharmacological studies firmly established arole for PPAR-β/δ in lipid metabolism. PPAR-β/δ regulates acyl-CoA synthetase 2 expressionin rat brain cells8 and promotes reverse choles-terol transport.46 Moreover, incubation ofskeletal and cardiac muscle cells with thePPAR-β/δ agonist GW501516 led to themodulation of genes implicated in β-oxidation,cholesterol efflux and energy uncoupling.98,99

Since PPAR-α-deficient mice do not display acompromised β-oxidation in muscle duringexercise and fasting, Muoio et al proposed acompensatory regulation by PPAR-β/δ.100

Several studies corroborated this hypothesis.Indeed, overexpression of PPAR-β/δ in muscleresulted in an increase in fibre number display-ing high oxidative metabolism, upregulation ofCPT1b, UCP2 and m-FABP,98,100–102 and in areduction of fat. PPAR-β/δ acts as a mediator ofthe adaptative response to exercise in muscleand has a remodelling function through theincrease of muscle fibre101 that consequentlyimproves exercise performance.

PPAR-β/δ also plays a role in lipoproteinmetabolism. Treatment of obese rhesusmonkeys with the PPAR-β/δ agonist GW501516not only significantly increased HDL levels via amechanism involving the induction of apoA-Iand apoA-II, but also decreased plasma VLDLand the number of small dense LDL particles.These findings support a model in which PPAR-β/δ stimulates cholesterol efflux from the cells,capture by HDL and transport back to the liver.Evidence that PPAR-β/δ regulates cholesterolefflux was provided using PPAR-β/δ-selectiveagonists.46,103 PPAR-β/δ activation increasesABCA1 expression and apoA-I mediated choles-terol efflux. However, a controversial role ofPPAR-β/δ in macrophage lipid homeostasis isillustrated by the findings that PPAR-β/δ maypromote VLDL-triglyceride and cholesterolloading and storage in macrophages via the

induction of the scavenger receptors CD36 andSRA, as well as ADRP and a/FABP expres-sion.103,104 Thus, the overall effect of PPAR-β/δon macrophage cholesterol homeostasis needsfurther clarification.

Therefore, PPAR-β/δ constitutes a potentialpharmacological target of interest for themetabolic syndrome.46 Whether an improve-ment in the lipid profile is also observed inhumans will remain elusive until specificagonists are being tested in clinical trials.

PPAR-γγ

PPAR-γγ and adipogenesis

Adipose tissue is the main site for lipid storageand plays a crucial role in the modulation oflipid levels in response to hormonal stimula-tion. Pharmacological activation of PPAR-γimproves insulin sensitivity.105 PPAR-γmodulates lipid and glucose homeostasisthrough its function in adipose tissue, but alsoexerts indirect actions in liver and muscle(Figure 4). Several crucial genes for adipocytedifferentiation and lipid uptake and storagecontain a PPRE. Gene expression of CD36,106

aP2,107 phosphoenol pyruvate carboxykinase,108


PPAR-γ

Gluconeogenesis

FA oxidation

Glucose utilization

FA oxidation

Glucose utilization

FA uptake

TNF-αIL-6

Leptin

FA uptakeLipogenesis

Liver

Muscle

Adipose tissue

Adiponectin

Figure 4 Activation of adipose PPAR-γ stimulates fattyacid (FA) and triglyceride flux toward the adipocyte andlipogenesis. Modulation of adipokines leads to indirecteffects in other tissues. PPAR-γ action potentiatesglucose utilization by muscle and liver to the detrimentof fatty acid oxidation

ch13 18/7/05 5:56 pm Page 143

ACS,109,110 FAT50 and LPL62 are regulated byPPAR-γ. The induction of LPL promotes FAdelivery to adipocytes while induction of FAT,FATP-1 and ACS results in enhanced FA storageby the adipocyte. Moreover, activation of PPAR-γ induces expression of glycerol kinase111 thatconverts glycerol to glycerol-phosphate, allow-ing FA re-incorporation into triglycerides.Consequently, PPAR-γ activation prevents freeFA release from the adipocyte, thus resulting inlower levels of circulating FA. These effectsshould result indirectly in a decrease in FAuptake and oxidation by the liver and themuscle. Thus, PPAR-γ action promotes glucoseutilization in both tissues and consequentlydecreases gluconeogenesis in the liver andimproves glucose oxidation in muscle. PPAR-γalso interferes with tumour necrosis factor alpha(TNF-α) signalling. TNF-α inhibits insulinsignalling and promotes lipolysis in adiposetissue, enhancing the development of insulinresistance in muscle and liver. Inhibition ofTNF-α signalling by PPAR-γ preserves normalglucose and lipid homeostasis. PPAR-γ alsomodulates the expression of adipokines such asleptin and adiponectin.112–115 PPAR-γ inhibitsleptin-induced lipolysis via repression of leptingene expression. Adiponectin, stronglyexpressed in adipose tissue, is upregulated byPPAR-γ.116 Adiponectin stimulates FA oxidationin muscle and decreases production of glucoseby the liver. These actions lead to the loweringof circulating FA, triglycerides and glucose, andcontribute to improved insulin sensitivity.Altogether, the actions of PPAR-γ preserve amature, insulin-sensitive adipocyte phenotypeand appear to be crucial for lipid and glucosehomeostasis. As anti-diabetic agents, glitazonesare primarily studied for their effect on glucosehomeostasis. Nevertheless, these drugs exerthypotriglyceridaemic actions in rodents byincreasing lipolysis and clearance of TG-richlipoproteins.117–122 This effect may be attributed,at least in part, to the induction of adiposetissue LPL expression.62,123 However, whetherthese mechanisms are operative in humans isunclear.

PPAR-γ is expressed in human macro-phages.79 PPAR-γ affects the RCT in a manner

similar to PPAR-α. PPAR-γ induces the expres-sion of SR-B1/CLA-180 and ABCA1.46,81,82 PPAR-γ-induced expression of LXR-α83 results inABCA1 expression.

PPAR-- γγ and steatosis

Although PPAR-α is the main PPAR isotypeexpressed in liver under normal conditions,PPAR-γ is induced in fatty livers. The roles ofPPAR-γ and thiazolidinediones in liver havebeen investigated in various animal models.Herzig et al. showed that CREB-deficient micedevelop a fatty liver phenotype and displayelevated expression of PPAR-γ. The authorsdemonstrated that CREB inhibits hepaticPPAR-γ expression in the fasted state andproposed that PPAR-γ probably behaves as apro-steatotic factor by a direct action in theliver.124 Chronic treatment of obese (NZO ×NON)F1 mice125 or diabetic KKAy mice126 withanti-diabetic thiazolidinediones led to severehepatic steatosis, while normal mice were notaffected. The function of liver PPAR-γ was alsoinvestigated in lipoatrophic A-ZIP/F-1 (AZIP)mice. In AZIP mice, ablation of liver PPAR-γreduced hepatic steatosis and also abolishedthe hypoglycaemic and hypolipidaemic effectsof rosiglitazone, demonstrating that, in theabsence of adipose tissue, the liver is an impor-tant site of PPAR-γ action. Thus, in mice, liverPPAR-γ contributes to hepatic steatosis, butprevents triglyceride accumulation and insulinresistance in other tissues.127 Interestingly, micelacking PPAR-γ in adipose tissue displayed amarked increase in hepatic glucogenesis andinsulin resistance and were significantly moresusceptible to high-fat diet-induced steatosis.128

Therefore, in mice, fasting or pathologicalconditions, such as obesity and diabetes, stimu-late hepatic PPAR-γ expression and activity thatconsequently promotes hepatic steatosis.

In humans, by contrast, PPAR-γ activationappears to improve hepatic steatosis. Patho-physiological evaluation of patients withdominant-negative PPAR-γ mutations revealednot only all the features of the metabolicsyndrome but also the existence of non-alco-holic steatohepatitis.129 Moreover, treatment of


ch13 18/7/05 5:56 pm Page 144

type 2 diabetic patients with rosiglitazone orpioglitazone decreases liver fat content.130–132

PPAR-γ activation also decreases liver fatcontent in subjects with non-alcoholic steato-hepatitis.133,134 Therefore, PPAR-γ appears toplay an important opposite role in hepaticsteatosis in humans and rodents.

PPAR-γ in clinical practice

Several studies have tested the effects of glita-zone treatment on glucose homeostasis indiabetic and insulin-resistant patients.135,136

Glitazone therapy lowers fasting and post-prandial glucose levels and improves insulin-stimulated glucose disposal. However, theeffect of glitazones on the plasma lipid profilein humans depends on the molecule tested,and intervention trials assessing the influenceof these compounds on the incidence ofcardiovascular disease, are still lacking. Never-theless, glitazone treatment of patients withtype 2 diabetes suggested that these drugs mayinhibit early atherosclerotic lesion progres-sion.135,136 In patients with type 2 diabetes,pioglitazone and rosiglitazone appear to havedistinct effects on the plasma levels of triglyc-erides and LDL. Pioglitazone treatment lowersserum concentrations of LDL and triglycerides,whereas rosiglitazone treatment does notappear to lower triglycerides and increases thelevels of LDL-cholesterol.137–139 Although themechanistic basis for these differences isunclear, one possible explanation may be thatpioglitazone has, albeit limited, PPAR-α activ-ity.140 Despite their atheroprotective proper-ties, certain glitazones may present a risk forheart failure in advanced diabetic patients.Thus, their clinical use should be carefullymonitored.141 The risk of heart failure withglitazones, which is probably due to plasmavolume expansion, appears to be enhancedwhen they are used in combination therapy

with insulin.141 Troglitazone has beenwithdrawn worldwide because of its hepato-toxic effects. Such effects are not observed withother glitazones.

CONCLUSION

PPARs are activated by dietary fatty acids andeicosanoids, as well as by pharmacologicaldrugs, such as fibrates for PPAR-α and glita-zones for PPAR-γ. PPAR-α and PPAR-β/δ areconsidered major regulators of intra- andextracellular lipid metabolism. PPAR-γimproves glucose metabolism via induction ofkey genes of adipocyte differentiation andfunction. As such, PPARs control plasma lipidand glucose levels, elevated levels of which areassociated with diabetes and cardiovasculardiseases. Recent investigations suggest thatPPAR-α and PPAR-γ activation attenuatesatherosclerosis progression, not only bycorrecting metabolic disorders but alsothrough direct effects on the vascular wall. Theintimate and causative relationships betweenlipid metabolism and coronary heart diseasehave stimulated research on the physiologicalfunctions of the lipid-activated transcriptionfactors of the PPAR family. This research hasresulted in the development and understand-ing of the action mechanism of drugs useful inthe treatment of metabolic disorders predis-posing to atherosclerosis. It is now anticipatedthat future development of drugs exertingtheir activity via PPARs will focus not so muchon the search of more potent activators butrather on selective receptor modulators in anattempt to separate desirable from unwantedside-effects. Finally, ongoing clinical studieswith specific PPAR-α/PPAR-β/δ/PPAR-γ(co)activators should prove (or disprove) theiractivity in reducing coronary events and totalmortality.


ch13 18/7/05 5:56 pm Page 145

1. Brown MS, Goldstein JL. The SREBP pathway: regula-tion of cholesterol metabolism by proteolysis of amembrane-bound transcription factor. Cell 1997;89:331–40

2. Peet D, Janowski B, Mangelsdorf D. The LXRs: a newclass of oxysterol receptors. Curr Opin Genet Dev1998; 8:571–5

3. Schoonjans K, Staels B, Auwerx J. The peroxisomeproliferator activated receptors (PPARs) and theireffects on lipid metabolism and adipocyte differentia-tion. Biochim Biophys Acta 1996; 1302:93–109

4. Braissant O, Wahli W. Differential expression ofperoxisome proliferator-activated receptor-alpha, -beta, and -gamma during rat embryonic develop-ment. Endocrinology 1998; 139:2748–54

5. Auboeuf D, Rieusset J, Fajas L, et al. Tissue distribu-tion and quantification of the expression of mRNAsof peroxisome proliferator-activated receptors andliver X receptor-alpha in humans: no alteration inadipose tissue of obese and NIDDM patients. Diabetes1997; 46:1319–27

6. Cullingford TE, Bhakoo K, Peuchen S, et al.Distribution of mRNAs encoding the peroxisomeproliferator-activated receptor alpha, beta, andgamma and the retinoid X receptor alpha, beta, andgamma in rat central nervous system. J Neurochem1998; 70:1366–75

7. Dreyer C, Krey G, Keller H, et al. Control of theperoxisomal beta-oxidation pathway by a novel familyof nuclear hormone receptors. Cell 1992; 68:879–87

8. Basu-Modak S, Braisssant O, Escher P, et al.Peroxisome proliferator-activated receptor betaregulates acyl-CoA synthetase 2 in reaggregated ratbrain cell cultures. J Biol Chem 1999; 274:35881–8


10. Spiegelman BM. PPAR-gamma: adipogenic regulatorand thiazolidinedione receptor. Diabetes 1998;47:507–14

11. Gervois P, Chopin-Delannoy S, Fadel A, et al. Fibratesincrease human REV-ERBalpha expression in liver viaa novel peroxisome proliferator-activated receptorresponse element. Mol Endocrinol 1999; 13:400–9

12. Delerive P, De Bosscher K, Besnard S, et al. PPARαnegatively regulates the vascular wall inflammatorygene response by negative cross-talk with transcrip-tion factors NF-κB and AP-1. J Biol Chem 1999;274:32048–54

13. Delerive P, Martin F, Chinetti G, et al. PPAR activatorsinhibit thrombin-induced endothelin-1 production inhuman vascular endothelial cells by inhibiting the AP-1 signalling pathways. Circ Res 1999; 85:394–402

14. Jiang C, Ting AT, Seed B. PPAR-gamma agonistsinhibit production of monocyte inflammatorycytokines. Nature (Lond) 1998; 391:82–6

15. Ricote M, Li AC, Willson TM, et al. The peroxisomeproliferator-activated receptor-gamma is a negativeregulator of macrophage activation. Nature (Lond)1998; 391:79–82

16. Sakai M, Matsushima-Hibiya Y, Nishizawa M, et al.Suppression of rat gluthatione transferase P expres-sion by peroxisome proliferators:interaction betweenJun and peroxisome proliferator-activated receptor α.Cancer Res 1995; 53:5370–6

17. Staels B, Koenig W, Habib A, et al. Activation ofhuman aortic smooth-muscle cells is inhibited byPPARalpha but not by PPARgamma activators. Nature(Lond) 1998; 393:790–3

18. Zhou YC, Waxman DJ. Cross-talk between januskinase-signal transducer and activator of transcription(JAK-STAT) and peroxisome proliferator-activatedreceptor-alpha (PPARα) signalling pathways. Growthhormone inhibition of PPARα transcriptional activitymediated by STAT5b. J Biol Chem 1999; 274:2672–81

19. Delerive P, Gervois P, Fruchart JC, et al. Induction ofIkappaBalpha expression as a mechanism contribut-ing to the anti-inflammatory activities of peroxisomeproliferator-activated receptor-alpha activators. J BiolChem 2000; 275:36703–7

20. Gervois P, Kleemann R, Pilon A, et al. Global suppres-sion of IL-6–induced acute phase response geneexpression after in vivo chronic treatment with theperoxisome proliferator-activated receptor α activatorfenofibrate. J Biol Chem 2004; 279:16154–60

21. Lemberger T, Staels B, Saladin R, et al. Regulation ofthe peroxisome proliferator-activated receptor alphagene by glucocorticoids. J Biol Chem 1994;269:24527–30

22. Steineger HH, Sorensen HN, Tugwood JD, et al.Dexamethasone and insulin demonstrate marked andopposite regulation of the steady-state mRNA level ofthe peroxisomal proliferator-activated receptor(PPAR) in hepatic cells. Hormonal modulation offatty acid-induced transcription. Eur J Biochem 1994;225:967–74

23. Wang MY, Unger RH. Novel form of lipolysis inducedby leptin. J Biol Chem 1999; 274:17541–4

24. Zhou YT, Wang ZW, Higa M, et al. Reversingadipocyte differentiation: implications for treatmentof obesity. Proc Natl Acad Sci USA 1999; 96:2391–5

25. Kersten S, Seydoux J, Peters JM, et al. Peroxisomeproliferator-activated receptor alpha mediates theadaptive response to fasting. J Clin Invest 1999;103:1489–98

26. Lemberger T, Saladin R, Vazquez M, et al. Expressionof the peroxisome proliferator-activated receptoralpha gene is stimulated by stress and follows adiurnal rhythm. J Biol Chem 1996; 271:1764–9

27. Palmer CN, Hsu MH, Griffin KJ, et al. Peroxisomeproliferator activated receptor-alpha expression inhuman liver. Mol Pharmacol 1998; 53:14–22


References

ch13 18/7/05 5:56 pm Page 146

28. Gervois P, Pineda Torra I, Chinetti G, et al. Atruncated human PPARα splice variant with dominantnegative activity. Mol Endocrinol 1999; 13:1535–49

29. Zhu Y, Qi C, Korenberg JR, et al. Structural organizationof mouse peroxisome proliferator-activated receptorgamma (mPPAR gamma) gene: alternative promoteruse and different splicing yield two mPPAR gammaisoforms. Proc Natl Acad Sci USA 1995; 92:7921–5

30. Fajas L, Auboeuf D, Raspe E, et al. The organization,promoter analysis, and expression of the humanPPARgamma gene. J Biol Chem 1997; 272:18779–89

31. Werman A, Hollenberg A, Solanes G, et al. Ligand-independent activation domain in the N terminus ofperoxisome proliferator-activated receptor gamma(PPARgamma). Differential activity of PPARgamma1and -2 isoforms and influence of insulin. J Biol Chem1997; 272:20230–5

32. Wu Z, Bucher NL, Farmer SR. Induction of peroxi-some proliferator-activated receptor gamma duringthe conversion of 3T3 fibroblasts into adipocytes ismediated by C/EBPbeta, C/EBPdelta, and glucocor-ticoids. Mol Cell Biol 1996; 16:4128–36

33. Juge-Aubry CE, Hammar E, Siegrist-Kaiser C, et al.Regulation of the transcriptional activity of the perox-isome proliferator-activated receptor alpha byphosphorylation of a ligand-independent trans-activating domain. J Biol Chem 1999; 274:10505–10

34. Shao D, Rangwala SM, Bailey ST, et al. Interdomaincommunication regulating ligand binding by PPAR-gamma. Nature (Lond) 1998; 396:377–80

35. Shalev A, Siegrist-Kaiser CA, Yen PM, et al. The perox-isome proliferator-activated receptor alpha is aphosphoprotein: regulation by insulin. Endocrinol-ogy 1996; 137:4499–502

36. Adams M, Reginato MJ, Shao D, et al. Transcriptionalactivation by peroxisome proliferator-activated recep-tor gamma is inhibited by phosphorylation at aconsensus mitogen-activated protein kinase site. J BiolChem 1997; 272:5128–32

37. Hu E, Kim JB, Sarraf P, et al. Inhibition of adipogen-esis through MAP kinase-mediated phosphorylationof PPARgamma. Science 1996; 274:2100–3

38. Camp HS, Tafuri SR, Leff T. c-Jun N-terminal kinasephosphorylates peroxisome proliferator-activatedreceptor-γ1 and negatively regulates its transcrip-tional activity. Endocrinology 1999; 140:392–7

39. Hauser S, Adelmant G, Sarraf P, et al. Degradation ofthe peroxisome proliferator-activated receptorgamma is linked to ligand-dependent activation. JBiol Chem 2000; 275:18527–33

40. Blanquart C, Barbier O, Fruchart JC, et al.Peroxisome proliferator-activated receptor alpha(PPARalpha) turnover by the ubiquitin-proteasomesystem controls the ligand-induced expression level ofits target genes. J Biol Chem 2002; 277:37254–9

41. Hirotani M, Tsukamoto T, Bourdeaux J, et al.Stabilization of peroxisome proliferator-activatedreceptor alpha by the ligand. Biochem Biophys ResCommun 2001; 288:106–10

42. Willson TM, Wahli W. Peroxisome proliferator-activated receptor agonists. Curr Opin Chem Biol1997; 1:235–41

43. Nagy L, Tontonoz P, Alvarez JG, et al. Oxidized LDLregulates macrophage gene expression throughligand activation of PPARγ. Cell 1998; 93:229–40

44. Xu HE, Lambert MH, Montana VG, et al. Molecularrecognition of fatty acids by peroxisome proliferator-activated receptors. Mol Cell 1999; 3:397–403

45. Berger J, Leibowitz MD, Doebber TW, et al. Novelperoxisome proliferator-activated receptor (PPAR)gamma and PPARdelta ligands produce distinctbiological effects. J Biol Chem 1999; 274:6718–25

46. Oliver WR, Shenk JL, Snaith MR, et al. A selectiveperoxisome proliferator-activated receptor deltaagonist promotes reverse cholesterol transport. ProcNatl Acad Sci USA 2001; 98:5306–11

47. Hashimoto T, Fujita T, Usuda N, et al. Peroxisomaland mitochondrial fatty acid beta-oxidation in micenullizygous for both peroxisome proliferator-activated receptor alpha and peroxisomal fatty acyl-CoA oxidase. Genotype correlation with fatty liverphenotype. J Biol Chem 1999; 274:19228–36

48. Qi C, Zhu Y, Pan J, et al. Absence of spontaneousperoxisome proliferation in enoyl-CoA Hydratase/L-3–hydroxyacyl-CoA dehydrogenase-deficient mouseliver. Further support for the role of fatty acyl-CoAoxidase in PPAR alpha ligand metabolism. J BiolChem 1999; 274:15775–80

49. Aoyama T, Peters JM, Iritani N, et al. Altered consti-tutive expression of fatty acid-metabolizing enzymesin mice lacking the peroxisome proliferator-activatedreceptor alpha (PPARalpha). J Biol Chem 1997;273:5678–84

50. Martin G, Schoonjans K, Lefebvre AM, et al.Coordinate regulation of the expression of the fattyacid transport protein and acyl-CoA synthetase genesby PPARalpha and PPARgamma activators. J BiolChem 1997; 272:28210–17

51. Motojima K, Passilly P, Peters JM, et al. Expression ofputative fatty acid transporter genes are regulated byperoxisome proliferator-activated receptor alpha andgamma activators in a tissue- and inducer-specificmanner. J Biol Chem 1998; 273:16710–14

52. Mascaro C, Acosta E, Ortiz JA, et al. Control of humanmuscle-type carnitine palmitoyltransferase I genetranscription by peroxisome proliferator-activatedreceptor. J Biol Chem 1998; 273:8560–73

53. Brandt JM, Djouadi F, Kelly DP. Fatty acids activatetranscription of the muscle carnitine palmitoyltrans-ferase I gene in cardiac myocytes via the peroxisomeproliferator-activated receptor alpha. J Biol Chem1998; 273:23786–92

54. Yu GS, Lu YC, Gulick T. Co-regulation of tissue-specific alternative human carnitine palmitoyltrans-ferase Ibeta gene promoters by fatty acid enzymesubstrate. J Biol Chem 1998; 273:32901–9

55. Djouadi F, Weinheimer CJ, Saffitz JE, et al. A gender-related defect in lipid metabolism and glucose


ch13 18/7/05 5:56 pm Page 147

homeostasis in peroxisome proliferator-activatedreceptor alpha-deficient mice. J Clin Invest 1998;102:1083–91

56. Leone TC, Weinheimer CJ, Kelly DP. A critical rolefor the peroxisome proliferator-activated receptoralpha (PPARalpha) in the cellular fasting response:The PPARalpha-null mouse as a model of fatty acidoxidation disorders. Proc Natl Acad Sci USA 1999;96:7473–8

57. Kelly LJ, Vicario PP, Thompson GM, et al.Peroxisome proliferator-activated receptors gammaand alpha mediate in vivo regulation of uncouplingprotein (UCP-1, UCP-2, UCP-3) gene expression.Endocrinology 1998; 139:4920–7

58. Brun S, Carmona MC, Mampel T, et al. Activators ofperoxisome proliferator-activated receptor-alphainduce the expression of the uncoupling protein-3gene in skeletal muscle: a potential mechanism forthe lipid intake-dependent activation of uncouplingprotein-3 gene expression at birth. Diabetes 1999;48:1217–22

59. Tsuboyama-Kasaoka N, Takahashi M, Kim H, et al.Up-regulation of liver uncoupling protein-2 mRNA byeither fish oil feeding or fibrate administration inmice. Biochem Biophys Res Commun 1999; 257:879–85

60. Hertz R, Bishara-Shieban J, Bar-Tana J. Mode ofaction of peroxisome proliferators as hypolipidemicdrugs. Suppression of apolipoprotein C-III. J BiolChem 1995; 270:13470–5

61. Staels B, Vu-Dac N, Kosykh V, et al. Fibrates down-regulate apolipoprotein C-III expression indepen-dent of induction of peroxisomal acyl co-enzyme Aoxidase. J Clin Invest 1995; 95:705–12

62. Schoonjans K, Peinado-Onsurbe J, Lefebvre AM, et al.PPARalpha and PPARgamma activators direct adistinct tissue-specific transcriptional response via aPPRE in the lipoprotein lipase gene. EMBO J 1996;15:5336–48

63. Peters JM, Hennuyer N, Staels B, et al. Alterations inlipoprotein metabolism in peroxisome proliferator-activated receptor alpha-deficient mice. J Biol Chem1997; 272:27307–12

64. Vu-Dac N, Chopin-Delannoy S, Gervois P, et al. Thenuclear receptors peroxisome proliferator-activatedreceptorα and Rev-erbα mediate the species-specificregulation of apolipoprotein A-I expression byfibrates. J Biol Chem 1998; 273:25713–20

65. Raspe E, Duez H, Mansen A, et al. Identification ofRev-erbalpha as a physiological repressor of apoC-IIIgene transcription. J Lipid Res 2002; 43:2172–9

66. Packard CJ. Overview of fenofibrate. Eur Heart J1998; 19 (Suppl A):A62–5

67. Pennacchio LA, Olivier M, Hubacek JA, et al. Anapolipoprotein influencing triglycerides in humansand mice revealed by comparative sequencing.Science 2001; 294:169–73

68. Tugwood JD, Aldridge TC, Lambe KG, et al.Peroxisome proliferator-activated receptor-alpha and

the pleiotropic responses to peroxisome proliferators.Arch Toxicol Suppl 1998; 20:377–86

69. Vu-Dac N, Gervois P, Jakel H, et al. ApolipoproteinA5, a crucial determinant of plasma triglyceridelevels, is highly responsive to peroxisome proliferatoractivated receptor alpha. J Biol Chem 2003;278:17982–5

70. Miller GJ, Miller NE. Plasma-high-density-lipoproteinconcentration and development of ischaemic heart-disease. Lancet 1975; 1:16–19

71. Balfour JA, McTavish D, Heel RC. Fenofibrate. Areview of its pharmacodynamic and pharmacokineticproperties and therapeutic use in dyslipidaemia.Drugs 1990; 40:260–90

72. Bard JM, Parra HJ, Camare R, et al. A multicentercomparison of the effects of simvastatin and fenofi-brate therapy in severe primary hypercholes-terolemia, with particular emphasis on lipoproteinsdefined by their apolipoprotein composition.Metabolism 1992; 41:498–503

73. Malmendier CL, Delcroix C. Effects of fenofibrate onhigh and low density lipoprotein metabolism inheterozygous familial hypercholesterolemia.Arteriosclerosis 1985; 55:161–9

74. Berthou L, Duverger N, Emmanuel F, et al. Oppositeregulation of human versus mouse apolipoprotein A-I by fibrates in human apolipoprotein A-I transgenicmice. J Clin Invest 1996; 97:2408–16

75. Vu-Dac N, Schoonjans K, Laine B, et al. Negativeregulation of the human apolipoprotein A-Ipromoter by fibrates can be attenuated by the inter-action of the peroxisome proliferator-activated recep-tor with its response element. J Biol Chem 1994;269:31012–18

76. Vu-Dac N, Schoonjans K, Kosykh V, et al. Fibratesincrease human apolipoprotein A-II expressionthrough activation of the peroxisome proliferator-activated receptor. J Clin Invest 1995; 96:741–50

77. Acton S, Rigotti A, Landschulz KT, et al. Identi-fication of scavenger receptor SR-BI as a high densitylipoprotein receptor. Science 1996; 271:518–20

78. Calvo D, Gomez-Coronado D, Lasuncion MA, et al.CLA-1 is an 85–kD plasma membrane glycoproteinthat acts as a high-affinity receptor for both native(HDL, LDL, and VLDL) and modified (OxLDL andAcLDL) lipoproteins. Arterioscler Thromb Vasc Biol1997; 17:2341–9

79. Chinetti G, Griglio S, Antonucci M, et al. Activation ofproliferator-activated receptors alpha and gammainduces apoptosis of human monocyte-derivedmacrophages. J Biol Chem 1998; 273:25573–80

80. Chinetti G, Gbaguidi GF, Griglio S, et al. CLA-1/SR-BI is expressed in atherosclerotic lesion macrophagesand regulated by activators of peroxisome prolifera-tor-activated receptors. Circulation 2000; 101:2411–17

81. Chinetti G, Lestavel S, Bocher V, et al. PPAR-alphaand PPAR-gamma activators induce cholesterolremoval from human macrophage foam cells through


ch13 18/7/05 5:56 pm Page 148

stimulation of the ABCA1 pathway. Nature Med 2001;7:53–8

82. Chawla A, Boisvert WA, Lee CH, et al. A PPARgamma-LXR-ABCA1 pathway in macrophages isinvolved in cholesterol efflux and atherogenesis. MolCell 2001; 7:161–71

83. Tobin KA, Steineger HH, Alberti S, et al. Cross-talkbetween fatty acid and cholesterol metabolismmediated by liver X receptor-alpha. Mol Endocrinol2000; 14:741–52

84. Staels B, Peinado-Onsurbe J, Auwerx J. Down-regula-tion of hepatic lipase gene expression and activity byfenofibrate. Biochim Biophys Acta 1992; 1123:227–30

85. Staels B, van Tol A, Skretting G, et al. Lecithin:choles-terol acyltransferase gene expression is regulated in atissue-selective manner by fibrates. J Lipid Res 1992;33:727–35

86. Bouly M, Masson D, Gross B, et al. Induction of thephospholipid transfer protein gene accounts for thehigh density lipoprotein enlargement in mice treatedwith fenofibrate. J Biol Chem 2001; 276:25841–7

87. Lee Y, Yu X, Gonzales F, et al. PPAR alpha is necessaryfor the lipopenic action of hyperleptinemia on whiteadipose and liver tissue. Proc Natl Acad Sci USA 2002;99:11848–53

88. Guerre-Millo M, Gervois P, Raspe E, et al. Peroxisomeproliferator-activated receptor alpha activatorsimprove insulin sensitivity and reduce adiposity. J BiolChem 2000; 275:16638–42

89. Mancini FP, Lanni A, Sabatino L, et al. Fenofibrateprevents and reduces body weight gain and adiposityin diet-induced obese rats. FEBS Lett 2001; 491:154–8

90. Ye JM, Iglesias PJ, Watson DG, et al. Peroxisomeproliferator-activated receptor (PPAR)-alpha activa-tion lowers muscle lipids and improves insulin sensi-tivity in high fat-fed rats: comparison with PPAR-gamma activation. Diabetes 2001; 50:411–17

91. Ip E, Farrell GC, Robertson G, et al. Central role ofPPARalpha-dependent hepatic lipid turnover indietary steatohepatitis in mice. Hepatology 2003;38:123–32

92. Chou CJ, Haluzik M, Gregory C, et al. WY14,643, aperoxisome proliferator-activated receptor alpha(PPARalpha ) agonist, improves hepatic and musclesteatosis and reverses insulin resistance in lipoat-rophic A-ZIP/F-1 mice. J Biol Chem 2002; 277:24484–9

93. He TC, Chan TA, Vogelstein B, et al. PPARdelta is anAPC-regulated target of nonsteroidal anti-inflamma-tory drugs. Cell 1999; 99:335–45

94. Lim H, Dey SK. PPARdelta functions as a prostacyclinreceptor in blastocyst implantation. TrendsEndocrinol Metab 2000; 11:137–42

95. Tan NS, Michalik L, Noy N, et al. Critical roles ofPPAR beta/delta in keratinocyte response to inflam-mation. Genes Dev 2001; 15:3263–77

96. Wang YX, Lee CH, Tiep S, et al. Peroxisome prolifer-ator-activated receptor delta activates fat metabolismto prevent obesity. Cell 2003; 113:159–70

97. Tanaka T, Yamamoto J, Iwasaki S, et al. Activation ofperoxisome proliferator-activated receptor deltainduces fatty acid beta-oxidation in skeletal muscleand attenuates metabolic syndrome. Proc Natl AcadSci USA 2003; 100:15924–9

98. Dressel U, Allen TL, Pippal JB, et al. The peroxisomeproliferator-activated receptor beta/delta agonist,GW501516, regulates the expression of genesinvolved in lipid catabolism and energy uncoupling inskeletal muscle cells. Mol Endocrinol 2003;17:2477–93

99. Gilde AJ, van der Lee KA, Willemsen PH, et al.Peroxisome proliferator-activated receptor (PPAR)alpha and PPARbeta/delta, but not PPARgamma,modulate the expression of genes involved in cardiaclipid metabolism. Circ Res 2003; 92:518–24

100. Muoio DM, MacLean PS, Lang DB, et al. Fatty acidhomeostasis and induction of lipid regulatory genesin skeletal muscles of peroxisome proliferator-activated receptor (PPAR) alpha knock-out mice.Evidence for compensatory regulation by PPAR delta.J Biol Chem 2002; 277:26089–97

101. Luquet S, Lopez-Soriano J, Holst D, et al. Peroxisomeproliferator-activated receptor delta controls muscledevelopment and oxidative capability. FASEB J 2003;17:2299–301

102. Holst D, Luquet S, Kristiansen K, et al. Roles of perox-isome proliferator-activated receptors delta andgamma in myoblast transdifferentiation. Exp Cell Res2003; 288:168–76

103. Vosper H, Patel L, Graham TL, et al. The peroxisomeproliferator-activated receptor delta promotes lipidaccumulation in human macrophages. J Biol Chem2001; 276:44258–65

104. Chawla A, Lee CH, Barak Y, et al. PPARdelta is a verylow-density lipoprotein sensor in macrophages. ProcNatl Acad Sci USA 2003; 100:1268–73

105. Sood V, Colleran K, Burge MR. Thiazolidinediones: acomparative review of approved uses. DiabetesTechnol Ther 2000; 2:429–40

106. Sfeir Z, Ibrahimi A, Amri E, et al. Regulation ofFAT/CD36 gene expression: further evidence insupport of a role of the protein in fatty acidbinding/transport. Prost Leuk Ess Fatty Acids 1997;57:17–21

107. Tontonoz P, Hu E, Graves RA, et al. mPPAR gamma2: tissue-specific regulator of an adipocyte enhancer.Genes Dev 1994; 8:1224–34

108. Tontonoz P, Hu E, Devine J, et al. PPAR gamma 2regulates adipose expression of the phospho-enolpyruvate carboxykinase gene. Mol Cell Biol 1995;15:351–7

109. Schoonjans K, Watanabe M, Suzuki H, et al.Induction of the acyl-coenzyme A synthetase gene byfibrates and fatty acids is mediated by a peroxisomeproliferator response element in the C promoter. JBiol Chem 1995; 270:19269–76

110. Schoonjans K, Staels B, Grimaldi P, et al. Acyl-CoAsynthetase mRNA expression is controlled by fibric-


ch13 18/7/05 5:56 pm Page 149

acid derivatives, feeding and liver proliferation. Eur JBiochem 1993; 216:615–22

111. Guan Y, Zhang Y, Davis L, et al. Expression of peroxi-some proliferator-activated receptors in urinary tractof rabbits and humans. Am J Physiol 1997;273:F1013–22

112. De Vos P, Lefebvre AM, Miller SG, et al.Thiazolidinediones repress ob gene expression inrodents via activation of peroxisome proliferator-activated receptor gamma. J Clin Invest 1996;98:1004–9

113. Yamauchi T, Kamon J, Waki H, et al. The fat-derivedhormone adiponectin reverses insulin resistanceassociated with both lipoatrophy and obesity. NatureMed 2001; 7:941–6

114. Maeda N, Takahashi M, Funahashi T, et al.PPARgamma ligands increase expression and plasmaconcentrations of adiponectin, an adipose-derivedprotein. Diabetes 2001; 50:2094–9

115. Berg AH, Combs TP, Du X, et al. The adipocyte-secreted protein Acrp30 enhances hepatic insulinaction. Nature Med 2001; 7:947–53

116. Iwaki M, Matsuda M, Maeda N, et al. Induction ofadiponectin, a fat-derived antidiabetic andantiatherogenic factor, by nuclear receptors. Diabetes2003; 52:1655–63

117. Stevenson RW, Hutson NJ, Krupp MN, et al. Actionsof novel antidiabetic agent englitazone in hyper-glycemic hyperinsulinemic ob/ob mice. Diabetes1990; 39:1218–27

118. Sohda T, Mizuno K, Momose Y, et al. Studies onantidiabetic agents. 11. Novel thiazolidinedione deriv-atives as potent hypoglycemic and hypolipidemicagents. J Med Chem 1992; 35:2617–26

119. Kemnitz JW, Elson DF, Roecker EB, et al. Pioglitazoneincreases insulin sensitivity, reduces blood glucose,insulin, and lipid levels, and lowers blood pressure, inobese, insulin- resistant rhesus monkeys. Diabetes1994; 43:204–11

120. Young PW, Cawthorne MA, Coyle PJ, et al. Repeattreatment of obese mice with BRL 49653, a newpotent insulin sensitizer, enhances insulin action inwhite adipocytes. Association with increased insulinbinding and cell-surface GLUT4 as measured byphotoaffinity labeling. Diabetes 1995; 44:1087–92

121. Lohray BB, Bhushan V, Rao BP, et al. Noveleuglycemic and hypolipidemic agents. 1. J Med Chem1998; 41:1619–30

122. Reddy KA, Lohray BB, Bhushan V, et al. Novel anti-diabetic and hypolipidemic agents. 5. Hydroxyl versusbenzyloxy containing chroman derivatives. J MedChem 1999; 42:3265–78

123. Lefebvre AM, Peinado-Onsurbe J, Leitersdorf I, et al.Regulation of lipoprotein metabolism by thiazolidine-diones occurs through a distinct but complementarymechanism relative to fibrates. Arterioscler ThrombVasc Biol 1997; 17:1756–64

124. Herzig S, Hedrick S, Morantte I, et al. CREB controlshepatic lipid metabolism through nuclear hormone

receptor PPAR-gamma. Nature (Lond) 2003;426:190–3

125. Watkins SM, Reifsnyder PR, Pan HJ, et al. Lipidmetabolome-wide effects of the PPARgamma agonistrosiglitazone. J Lipid Res 2002; 43:1809–17

126. Bedoucha M, Atzpodien E, Boelsterli UA. DiabeticKKAy mice exhibit increased hepatic PPARgamma1gene expression and develop hepatic steatosis uponchronic treatment with antidiabetic thiazolidine-diones. J Hepatol 2001; 35:17–23

127. Gavrilova O, Haluzik M, Matsusue K, et al. Liverperoxisome proliferator-activated receptor gammacontributes to hepatic steatosis, triglyceride clear-ance, and regulation of body fat mass. J Biol Chem2003; 278:34268–76

128. He W, Barak Y, Hevener A, et al. Adipose-specificperoxisome proliferator-activated receptor gammaknockout causes insulin resistance in fat and liver butnot in muscle. Proc Natl Acad Sci USA 2003;100:15712–17

129. Savage DB, Tan GD, Acerini CL, et al. Humanmetabolic syndrome resulting from dominant-negative mutations in the nuclear receptor peroxi-some proliferator-activated receptor-gamma.Diabetes 2003; 52:910–17


131. Carey DG, Cowin GJ, Galloway GJ, et al. Effect ofrosiglitazone on insulin sensitivity and body composi-tion in type 2 diabetic patients. Obes Res 2002;10:1008–15

132. Bajaj M, Suraamornkul S, Pratipanawatr T, et al.Pioglitazone reduces hepatic fat content andaugments splanchnic glucose uptake in patients withtype 2 diabetes. Diabetes 2003; 52:1364–70

133. Neuschwander-Tetri BA, Brunt EM, Wehmeier KR, etal. Improved nonalcoholic steatohepatitis after 48weeks of treatment with the PPAR-gamma ligandrosiglitazone. Hepatology 2003; 38:1008–17

134. Neuschwander-Tetri BA, Brunt EM, Wehmeier KR, etal. Interim results of a pilot study demonstrating theearly effects of the PPAR-gamma ligand rosiglitazoneon insulin sensitivity, aminotransferases, hepaticsteatosis and body weight in patients with non-alcoholic steatohepatitis. J Hepatol 2003; 38:434–40

135. Kaplan F, Al-Majali K, Betteridge DJ. PPARS, insulinresistance and type 2 diabetes. J Cardiovasc Risk 2001;8:211–17

136. Minamikawa J, Tanaka S, Yamauchi M, et al. Potentinhibitory effect of troglitazone on carotid arterialwall thickness in type 2 diabetes. J Clin EndocrinolMetab 1998; 83:1818–20

137. Kipnes MS, Krosnick A, Rendell MS, et al.Pioglitazone hydrochloride in combination withsulfonylurea therapy improves glycemic control inpatients with type 2 diabetes mellitus: a randomized,


ch13 18/7/05 5:56 pm Page 150

placebo-controlled study. Am J Med 2001; 111:10–17138. Gegick CG, Altheimer MD. Comparison of effects of

thiazolidinediones on cardiovascular risk factors:observations from a clinical practice. Endocr Pract2001; 7:162–9

139. Sidhu JS, Kaposzta Z, Markus HS, et al. Effect ofrosiglitazone on common carotid intima-media thick-

ness progression in coronary artery disease patientswithout diabetes mellitus. Arterioscler Thromb VascBiol 2004; 24:930–4

140. Smith U. Pioglitazone: mechanism of action. InternalJ Clin Pract Suppl 2001; 13–18

141. Gale EA. Lessons from the glitazones: a story of drugdevelopment. Lancet 2001; 357:1870–5


ch13 18/7/05 5:56 pm Page 151

ch13 18/7/05 5:56 pm Page 152

BACKGROUND

History of adipokines

As the prevalence of obesity is at near-epidemicproportions in many societies scientific andmedical attention has focused on the fat cell.Traditionally regarded as a cell providingenergy storage and insulation, the adipocyte isnow recognized to be a complex and metaboli-cally active cell with a prominent endocrinefunction.1,2 The adipocyte is highly responsiveto central and local signals and nutritionalstatus, and produces a number of proteins andmolecules which have autocrine, paracrine andendocrine effects. These molecules are collec-tively known as adipokines.

The existence of adipokines has been postu-lated for decades,3 and then the discovery ofleptin in 19944 set in train a remarkable era ofresearch and discovery that has completelychanged our understanding of, and approachto, adipose cells and adipose tissue. Thisresearch effort shows no signs of abating,encouraged by both an intense scientific inter-est in a completely ‘new’ endocrine organ –adipose tissue – and by the need for treatmentsfor obesity and its related metabolic disorders.

For both the researcher in the field and forinterested onlookers, this era of discovery hasbeen exciting and impressive in its pace. Fromthe initial discovery of each adipokine to thestage of establishing a reasonable understand-ing of its role, regulation and effects hassometimes taken only months. Whilst we still

lack an overall understanding of the adipokinesystem and network, significant patterns arenow emerging regarding the secretory aspectsof adipose tissue.

Adipokines and the metabolic syndrome

A major reason for the intense research inter-est in adipokines is their likely role as media-tors of the metabolic syndrome.5–7 It is nowclear that dysregulation of adipokine produc-tion occurs in obesity, and that the alteredadipokine milieu underpins, at a biochemicallevel, the insulin resistance and cardiovascularand metabolic abnormalities that characterizethe syndrome. The question still remains as towhich occurs first – the adipokine dysregula-tion or the obesity? The observation thatabnormalities in adipokine production seen inobesity can be largely reversed with weight lossargues that obesity may be the primaryproblem. However recent data indicating rolesfor adipokines in appetite and metabolic rateregulation suggest that inherited ‘abnormali-ties’ in adipokine production (which may beevolutionally advantageous) may be the funda-mental defect, only expressed in settings ofpositive energy balance. It is remarkable thatmost identified adipokines appear to havesome relationship with the metabolicsyndrome. It is as yet unclear whether this isindeed the case, or whether it is a biasproduced by experimental approaches basedon metabolic syndrome or obesity models.

Adipocytes and their secretory productsJ.B. Prins

14

ch14 14/7/05 4:47 pm Page 153

Adipokine regulation: evolutionary andgenetic aspects

Whilst much adipokine research is centred onobesity models, from an evolutionary point ofview, many adipokines are starvationmolecules.8,9 Indeed, the levels of obesity seentoday (and represented by some animalmodels) have a very short history and areunlikely to have significantly impacted on thegenome. For example, leptin was touted as anobesity hormone immediately after its discov-ery. As further research was undertaken it hasbecome apparent that its major role is to signalthat sufficient energy stores exist to allow repro-duction.8,9 It is now no surprise to researchersthat the regulation and function of leptin areimpaired in obese states, and similar patterns ofregulation are recognized much more quicklyin the newly discovered adipokines.

The corollary to this is that the fundamentalfunction of many adipokines is to maximizeweight gain. For millennia this was an essentialsystem for survival and those with the most‘active’ adipokine system survived best in situa-tions of low food availability. In our currentsituation of excess food availability in manysocieties, this metabolic efficiency has driventhe obesity epidemic. The discovery ofadipokines has afforded some insight into themediators of this ‘thrifty genotype’.

We still understand little about the diversityof body weight, and why some individualsremain lean in situations of energy excess andvice versa. Very lean individuals remain littlestudied, but will reveal insights into energyhomeostasis and adipokine regulation. Twinand other studies demonstrate clearly a substan-tial inheritability of body weight and metabolicdysfunction,10–13 but in very few individuals canthis be explained on the basis of single-genedefects. The single-gene defects leading toobesity or the metabolic syndrome are often ofadipokine genes, and have provided superbinsight into adipokine physiology.14

Adipokines and medical therapy

The strong relationship between adipokinedysregulation and metabolic dysfunction

provides clearly identified targets and opportu-nities for therapeutic intervention. Theseopportunities have attracted scientists andpharmaceutical companies to the research areawith a resultant rapid advance in understand-ing. To date, however, strategies to block theaction or reduce expression of over-producedadipokines have had little impact on metabolicdysfunction in man,15 despite, in someinstances, promising results in murine models.Such results have re-affirmed the complexity ofmetabolic regulation, and have also led to theidentification of important differences betweenman and other species and recognition of theneed for good experimental models.

ADIPOKINES

Adipocyte secretory products include hor-mones, cytokines, growth factors and fats,leading to the adoption of the ‘umbrella’ term– adipokine. In this chapter, adipokines aregrouped based on their putative majorfunction, but it is apparent that most adipo-kines have multiple roles. In many instances,the major function is controversial or unclear,and may also change according to the bodyweight of the subject. For example, in starva-tion, leptin has a metabolic role, signalling tothe hypothalamus that energy stores are lowand thus stimulating appetite and suppressingovulation. In the setting of obesity, thesefunctions of leptin are superfluous, and promi-nent functions now include deleterious effectson the vasculature and glucose homeostasis.

‘Metabolic’ adipokines

Leptin

The discovery of leptin,4 first postulateddecades before, began the current era ofintense research into adipose tissue as anendocrine organ. It is still regarded as theprototypical adipokine whose secretion isnormal at normal body weight but is abnormalin situations of either decreased16 or increasedfat stores.1


ch14 14/7/05 4:47 pm Page 154

Leptin is secreted almost exclusively byadipocytes and acts via a family of plasmamembrane receptors.17,18 The functional longform of the receptor (Ob-Rb) is expressed inseveral sites in the hypothalamus and brain-stem19 and signals via the STAT-3 pathway toalter expression of a number of genes involvedin energy homeostasis.20,21 Leptin receptors arealso expressed in many peripheral tissues22–33

and leptin appears to have actions in adiposetissue, skeletal muscle and the pancreas.Regulation of energy balance appears to bethrough both central mechanisms34 (appetiteand metabolic rate) and peripheral mecha-nisms (activation of 5'-AMP-activated proteinkinase (AMPK) expression in muscle andliver).35 In addition to regulation of energybalance, leptin has important roles in fertility(outlined above)36–39 and immune function asthe immunosuppression seen in starvation ornutrient deprivation appears due to leptindeficiency.40,41

Circulating concentrations of leptin areproportional to fat mass, but the fact that thehigh leptin levels in obesity do not suppressappetite indicates that leptin resistance may bea fundamental pathology in obesity.32,33 Thisresistance is in part due to impaired transportof leptin into the cerebrospinal fluid, and inpart due to reduced activation at the targettissue level.

The importance of normal leptin levels iselegantly demonstrated by the rare individualswith leptin deficiency and their response toleptin replacement therapy, with virtualnormalization of extreme metabolic dysfunc-tion.14,42,43 In contrast, trials of leptin supple-mentation in obese humans have shownconsiderably less promise.44

Adiponectin

Adiponectin (AdipQ, Acrp30, apMI) issecreted exclusively by adipocytes and, incontrast to leptin, circulating levels areinversely proportional to fat mass.45,46 Serumlevels are very high (3–30 nM) and, like leptin,are 2–3-fold higher in females than in males.6

Adiponectin has roles in glucose and lipid

homeostasis5–7,47,48 and as an anti-atherogenicprotein.7

Adiponectin is a prototypical adipokineinvolved in the ‘cross-talk’ between insulintarget tissues.49 This cross-talk is demonstratedby the FIRKO mouse, which lacks adiposetissue insulin receptors, has low adiponectinlevels and is insulin resistant in muscle andliver.50 Adiponectin has a major role toenhance hepatic insulin sensitivity, and the fallin adiponectin levels with weight gain appearsresponsible for associated hepatic insulin resis-tance and non-suppression of hepatic glucoseoutput. With weight loss, the rise inadiponectin levels correlates strongly with thereturn of insulin sensitivity.6

Adiponectin expression is increased byPPAR-γ agonists such as thiazolidinediones andthis function may be central to their efficacy asinsulin sensitizers5,51,52 Additionally, the de-creased adiponectin seen in obesity maymediate the enhanced adipocyte TNF-αproduction, postulated to underpin the insulinresistance.5

In mice, recombinant adiponectin decreaseshepatic gluconeogenesis by reducing expres-sion of phosphoenolpyruvate carboxykinase(PEPCK). Adiponectin also reduces lipidaccumulation in non-adipose tissues via activa-tion of AMPK.6 Adiponectin actions aremediated by the receptors AdipoR1 andAdipoR2.53 These findings indicate substantialpromise for adiponectin as a therapeutic agentin patients with the metabolic syndrome.

Free fatty acids

Free, or non-esterified, fatty acids are a mecha-nism for energy transport and redistribution.54

They are derived from diet, hepatic synthesisand lipolysis, dependent on nutritional statusand acute energy balance. Because lipolysis is amajor source of free fatty acids (FFAs)(especially in obesity), these molecules areregarded as adipokines. The major regulatorsof FFA concentration are catecholamines(favouring lipolysis) and insulin (favouringlipogenesis, which is essentially triglyceride re-esterification). Obesity is a state of high

ADIPOCYTES AND THEIR SECRETORY PRODUCTS 155

ch14 14/7/05 4:47 pm Page 155

catecholamine levels and insulin resistance andthe net effect of these abnormalities, plusongoing positive energy balance, leads toincrease FFA concentrations in the circulation.

Increased FFA concentrations lead tofurther metabolic dysfunction through effectsto decrease insulin production55,56 and toimpair insulin action57 leading to reducedinsulin-stimulated glucose uptake. It is clearthat all FFA are not alike, with some providingmetabolic benefit. This benefit, particularlywith respect to insulin sensitivity, may be due toalterations in plasma membrane fluidity or maybe due to particular FFA (or metabolites)activating members of the PPAR family oftranscription factors.58,59

Steroid hormones

Adipocytes have an important role in steroidhormone metabolism, and there has beenmuch recent emphasis on this role in the devel-opment of the metabolic syndrome.

Adipocytes express the enzyme 11β-hydroxy-steroid dehydrogenase (11β-HSD) whichconverts circulating inactive cortisone to theactive hormone cortisol, thus regulatingadipose tissue (and perhaps systemic) cortico-steroid concentrations.60 Adipose tissue corti-costeroid produced by this reaction could thenpromote pre-adipocyte replication and differ-entiation, thus promoting obesity, andcontribute to insulin resistance. Observationsin support of this theory are that 11β-HSDactivity is greater in omental adipose tissue anddysregulated in obesity.60,61 Compellingsupporting data come from transgenic studiesin which mice overexpressing 11β-HSDdemonstrate features of the metabolicsyndrome62 whilst the 11β-HSD knockoutmouse is relatively protected from metabolicdysfunction.63 Further evidence defining theimportance of the 11β-HSD system will comefrom clinical trials of the 11β-HSD inhibitorsunder development by many groups.

Adipocytes also have an important role insex-steroid metabolism. Androgens areconverted to oestrogens in fat by a cytochromeP450-dependent aromatase. This reaction is

the major source of oestrogens in men andpost-menopausal women, and a significantcontributor in pre-menopausal women.Additionally, adipose tissue 17β-hydroxysteroidoxidoreductase converts oestrone to oestradioland androstenedione (from the adrenal) totestosterone. In obesity there is increased netsex steroid interconversion which maycontribute to the metabolic syndrome and toregulation of adipose tissue distribution.1

‘Vascular’ adipokines

The relationship between the microvasculaturewithin adipose tissue and adipose cell growthand activity has been an active area of researchfor some time. This relates to the observationthat (sometimes enormous) change in fat massmust be accompanied by a similar change invasculature.64 It is unclear which comes first. Insettings of positive energy balance, an obvioussequence of events is that the need for lipidstorage drives increase in size and number ofadipocytes, which in turn signals growth in thevasculature.65 This possibility is supported bythe identification of pro-vascularizationadipokines. The alternative possibility is thatneo-vascularization occurs first, with themicrovasculature then signalling adipogene-sis.66 This is supported by observations thatanti-angiogenic drugs can cause weight loss66

and that adipose-tissue-derived microvascularendothelial cells produce factors that promotehuman pre-adipocyte replication.67

Adipokines that promote vasculargrowth

Monobutyrin is a pro-angiogenic lipid that maypromote expansion of the microvasculatureduring adipose tissue growth.68 Angiopoietin-1is produced by adipocytes and endothelial cellsand expression is altered in settings of adiposetissue growth and regression.69 A number ofgroups have presented evidence that vascularendothelial growth factor (VEGF) is anadipokine involved in neo-vascularization, andthat the relationship may be reciprocal.70 Thisimportance of the relationship is confirmed by


ch14 14/7/05 4:47 pm Page 156

the demonstration that blocking VEGF recep-tors inhibits adipose tissue development invivo.71

Hormones of the renin–angiotensin–aldosterone system

Adipose tissue synthesizes all components ofthe renin–angiotensin–aldosterone system(RAAS) and expresses angiotensin II recep-tors.72 Adipose tissue angiotensinogen produc-tion is increased by overnutrition and insulin,suggesting that overactivity of the RAAS may bea link between obesity and hypertension.73

Angiotensin II increases lipogenesis74 inhuman adipocytes and promotes pre-adipocytedifferentiation,72 suggesting a role for theRAAS in regulation of body weight. The RAAShas also been implicated in the development ofmyocardial dysfunction and hence may have arole in the development of the cardiomyopathycommonly seen in obesity and diabetes.75

Plasminogin activator inhibitor-1

Plasminogin activator inhibitor-1 (PAI-1) is amajor anti-fibrinolytic protein produced byliver and fat. Circulating levels of PAI-1 areincreased in obesity and PAI-1 production isgreater in omental than subcutaneous adiposetissue,76 suggesting a link between adiposetissue PAI-1 production and the cardiovasculardisease component of the metabolicsyndrome.7 PAI-1 appears to be involved in thepathogenesis of cardiovascular disease as levelsare increased in association with myocardialinfarction and venous thrombosis.1 PAI-1production by adipocytes is increased byinsulin, glucose and corticosteroids anddecreased by thiazolidinediones.77–79 It is alsopossible that statin therapy lowers PAI-1.7

Increased PAI-1 levels also predict type 2diabetes and are a core feature of the insulinresistance syndrome.

‘Immune’ adipokines

Adipsin (complement factor D) was one of thefirst recognized80 adipokines and is now but one

of a large group of ‘immune’ or ‘inflammatory’compounds recognized to be produced by fat.81

Many of these adipokines have purportedmetabolic function in addition to their immunerole as most cause insulin resistance in vitro or,when overexpressed, in vivo. Debate continuesas to their true role in adipose tissue, and thisdebate has been accelerated by the identifica-tion of C-reactive protein (CRP) as a risk factorfor cardiovascular disease.6 It has been recentlyproposed that ‘immune’ adipokines maycontribute to the elevated CRP seen in obesityand the metabolic syndrome. This family ofadipokines has increased production in obesity,with return toward normal levels with weightloss.6 In health, it appears that adipose tissue isa significant contributor to circulating levels ofmost of these compounds. All are alsoproduced by immune tissues and levels seen ininflammation far exceed those seen in (evenmorbid) obesity.

Tumour necrosis factor alpha

Tumour necrosis factor alpha (TNF-α) isproduced by lymphoid cells, skeletal muscleand fat cells, and is a postulated ‘link’ betweenobesity and type 2 diabetes.82–86 This postulateis based on observations that TNF-α knockoutmice are relatively protected from obesity-related diabetes,87 and anti-TNF-α strategieshave been shown to ameliorate diabetes inobese animals.84

Adipose tissue and circulating TNF-α levelsare increased in obese animals and humans,and these levels tend to normalize with weightloss.88 In several studies, change in insulinsensitivity parallels the change in TNF-α level.TNF-α disrupts insulin signalling at a numberof levels, including the insulin receptor, IRS-1and GLUT 4.85,89–95 Despite these findings, fewdata have been published showing TNF-α-induced inhibition of glucose uptake ininsulin-responsive tissues. Of importance,effective anti-TNF-α strategies have no effecton glucose homeostasis in humans with type 2diabetes.96,97

TNF-α has also been proposed to have a rolein the regulation of adipose tissue mass. It


ch14 14/7/05 4:47 pm Page 157

impairs human pre-adipocyte differentiation invitro,98 and induces apoptosis of human pre-adipocytes and adipocytes in vitro.99 It alsoinduces lipolysis,100,101 so TNF-α may have a netrole in reducing adipose tissue mass. Thiswould certainly be logical in an evolutionarysense, as it would serve to mobilize energy insettings of inflammation.

Overall, despite the wealth of data aboutTNF-α as a key adipokine, currently availableevidence does not strongly support a role forTNF-α as underpinning the metabolic syn-drome in man.

Interleukin 6

Interleukin-6 (IL-6) is a potent anti-inflamma-tory cytokine which also has endocrineactions.2,102 These include postulated roles inglucose and lipid metabolism.5 Like TNF-α,circulating IL-6 levels and production byadipose tissue are increased in obesity and fallwith weight loss.103 In health, adipose tissue is amajor source of circulating IL-6. IL-6 produc-tion is greater in omental than subcutaneousfat, and this may be relevant to the metabolicsyndrome.104 IL-6 has a postulated role in thedevelopment of vascular disease in obesity.7 IL-6 knockout mice have altered appetite andenergy expenditure and IL-6 is produced inthe hypothalamus.105 These findings suggest abroader role for IL-6 than TNF-α in energymetabolism.

Complement pathway components

Complement factors B, C3 and D (adipsin) areproduced by adipose tissue.81 Adipsin was thefirst adipokine identified80 and a potentialmetabolic role was suggested by the observa-tion that adipsin levels were decreased inmurine obese models.106 Subsequent humanstudies demonstrated the opposite – thatadipsin expression is increased in obesity andwith feeding and reduced in cachexia, lipodys-trophy and fasting.107

In recent years it has been postulated thatthe metabolic role for factors B, C3 and adipsinas adipokines is through the formation of acyla-

tion-stimulating protein (ASP).108 ASP levelincreases post-prandially and is involved in thestimulus and storage of triglycerides.109 ASPknockout animals are lean and have high post-prandial FFA levels and increased metabolicrate,110 but humans with adipsin deficiencyhave no obvious metabolic derangement andare normal weight.111

RELATIONSHIP OF ADIPOKINES TOTHE METABOLIC SYNDROME

Overall, the relationship of adipokines to themetabolic syndrome remains one of correla-tion, rather than cause–effect. It is becomingclear that murine models of obesity do nottruly reflect the obese human state. Forexample, the ob/ob mouse allowed facilitatedidentification of leptin and provided enormousamounts of data and impetus to obesityresearch. However, the leptin-deficient mousediffers significantly from the leptin-deficienthumans described to date, the former display-ing decreased metabolic rate and skeletal andmuscle abnormalities not prominent in thehumans.42 Also, transgenic studies providevaluable clues and suggestions as to the roles ofadipokines, but do not necessarily provide keydata of relevance to the ‘average’ obese man orwoman. This is exemplified by the TNF-αstudies, which demonstrate efficacy of anti-TNF-α strategies in mice with the metabolicsyndrome, whilst identical strategies in man areineffective.

The fundamental issue is that prolongedoverfeeding induces obesity which, in evolu-tionary terms, is an abnormal state. Strategiesutilized by adipose tissue in settings of starva-tion are ill developed in obesity and hence aredysregulated. This concept is exemplified byleptin and adiponectin. In starvation, leptinlevels are low, signalling to the hypothalamusto both increase feeding behavior and suppressovulation – strategies of clear survival advan-tage. Similarly, adiponectin levels are high,maximizing insulin sensitivity to promoteenergy storage and utilization. In obesity, theseactions of leptin and adiponectin are not


ch14 14/7/05 4:47 pm Page 158

needed for survival, so appropriate regulatorymechanisms are not in place.

It seems likely that obesity drives abnormaladipokine production and this, in turn,contributes to the many facets of the metabolicsyndrome. It is unclear whether inheritedperturbations in adipokine production mightunderlie the metabolic syndrome. In someways it is surprising that numerous large-scalegenetics studies have failed to show clearevidence for adipokine sequence variations aspotential causes of the metabolic syndrome.On the other hand, this could be expected inview of the fact that the metabolic syndrome isso common, and is quite a subtle disorder,taking many decades to develop in mostinstances. It seems certain that the metabolicsyndrome is a multi-gene disorder.112 Futurelarge-scale genetic studies with extensivephenotyping and genotyping may indeedunravel and identify contributory mutations.

A further layer of complexity is the inter-relationship between many of the adipokines. Ifalteration in the production or action of oneadipokine was the primary event (which seemsunlikely), this could lead to abnormalities ofexpression or function in many of the others. Itmay therefore be unrealistic to expect thattherapeutic modulation of a single adipokinewould be efficacious in the metabolic syndrome.

REGULATION OF KEY ADIPOKINEPRODUCTION

A number of regulators of adipokine produc-tion have been identified. This section willconcentrate on acute regulation, as opposed todefining the relationship between fat mass andadipokine production, which is detailed in thesections above. Adipokine production isregulated acutely by both nutritional status anda number of drugs, as well as by otheradipokines.

Leptin

Feeding has an obvious effect to reduce appetiteand this may in part be due to stimulation of

leptin expression.1 In contrast, food deprivationin the mouse and man is associated with an acutefall in leptin expression.5,113,114 These observa-tions are consistent with the appetite-suppressiveeffects of the adipokine. Leptin expression isreduced by thiazolinedione (TZDs) and β-agonists and increased by insulin and glucocorti-coids.5,115 These observations are consistent withthe appetite-suppressive effects of the adipokine.Weight loss induced by lifestyle interventioninduces a fall in leptin,116 but this may simplyreflect the reduction in fat mass. Insulin is apotent acute regulator of leptin expression andmay mediate the post-prandial changes seen.21 Itis not clear whether the hyperinsulinaemia seenin obesity or the metabolic syndrome underliesthe relationship between leptin levels and fatmass.

Adiponectin

Adiponectin expression is reduced in fasting andrestored by refeeding.5,114 This may be due to aneffect of the stomach-derived peptide, ghrelin, tosuppress adiponectin expression.117 TZDs havebeen shown by many groups to increaseadiponectin expression and this may be a signif-icant contributor to their efficacy as insulin sensi-tizers.51,52 Interestingly, this effect may bedepot-specific, with a greater adiponectinresponse in omental, compared to subcuta-neous, adipose tissue.52 Of further therapeuticinterest, adiponectin expression is increased byblockade of the RAS with either angiotensin-converting enzyme inhibitor or angiotensin IIreceptor blockade.118 Of interest, exercise doesnot alter adiponectin expression despite signifi-cant promotion of insulin sensitivity.116,119 Incontrast to leptin, adiponectin expression isdecreased by insulin and glucocorticoids,5

perhaps explaining the decrease in adiponectinlevels seen in obese, insulin-resistant subjects.

POTENTIAL FOR ADIPOKINE-BASEDMEDICAL THERAPIES

As outlined above, the relationship betweenmetabolic dysfunction and altered adipokine


ch14 14/7/05 4:47 pm Page 159

levels is now well established. The opportunityis thus provided for therapeutic interventionsaimed at normalizing adipokine levels andhence improving metabolic dysfunction. At asuperficial level, strategies to increase leptin inobesity with a view to suppressing appetite orincreasing adiponectin levels to improveinsulin sensitivity would seem likely to besuccessful. However, the complexities of theregulatory systems mean that such simplisticstrategies are not always efficacious. On a morepositive note, such therapeutic attempts haveincreased our understanding of the metabolicroles of adipokines.

Another important consideration is that theunderlying reason for adipokine dysregulationmay in itself make therapy ineffective. Anexample is seen with insulin therapy in type 2diabetes, where progressively larger doses ofinsulin are needed to overcome the insulinresistance – the basic initial abnormality.

Leptin therapy has been trialled in obesitywith disappointing results. The leptin ‘resis-tance’ of obesity necessitates high dosage, andthis in turn leads to increased insulin resistancewith little reduction in appetite or weight.Current strategies are aimed at developingreceptor-specific leptin analogues in an effortto improve the efficacy/adverse effect ratio.

Adiponectin has obvious therapeutic poten-tial with the theoretical attraction of being a‘replacement’ therapy rather than a supple-mentation to supra-physiological concentra-tions. If effective, adiponectin therapy couldinduce insulin sensitivity, improve lipids anddecrease vascular inflammation. Such strate-gies are effective in rodent models with evident

significant side-effects. A potential concern isthat significant improvement in insulin sensi-tivity may induce weight gain, but this is notapparent in animal studies published to date.

Strategies to reduce circulating levels ofimmune adipokines are also theoreticallyattractive. Anti-TNF-α therapeutics are effec-tive in rheumatoid arthritis and Crohn’sdisease in man, and in murine models haveefficacy in type 2 diabetes. In human trials,however, similar agents do not alter insulinsensitivity or weight. The reason for this speciesspecificity of effect is not clear. Strategies toreduce levels of IL-6 or PAI-1 may also be ofbenefit in reducing insulin resistance andvascular disease. The potential problem withsuch interventions is the risk of compromisingthe immune response, with consequentincreased risk of infection, autoimmunity andmalignancy.

As outlined above, corticosteroids are impli-cated in many aspects of the metabolicsyndrome. Many companies have inhibitors of11β-HSD in development with the promise ofreducing adipose tissue and possibly systemiccorticosteroid levels. The advantage of thisstrategy over, for example, corticosteroidreceptor blockers is that the stress response isunlikely to be significantly compromised.

The development of effective therapeuticinterventions based on the adipokine system iscompletely reliant on a sound and detailedunderstanding of the (patho)-physiology of thesystem in sickness and in health. At the currentrate of research progress, such understandingwill be rapidly obtained, and it is hoped thateffective therapies will result in the near future.


1. Ahima RS, Flier J. Adipose tissue as an endocrineorgan. Trends Endocrinol Metab 2000; 11:327–9

2. Mohamed-Ali V, Pinkney J, Coppack AW. Adiposetissue as an endocrine and paracrine organ. Int JObesity 1998; 22:1145–58

3. Coleman DL. Effects of parabiosis of obese withdiabetes and normal mice. Diabetologia 1973;9:294–8

4. Zhang Y, Proenca R, Maffei M, et al. Positionalcloning of the mouse obese gene and its humanhomologue. Nature 1994; 372:425–32

5. Fasshauer M, Paschke R. Regulation of adipocy-tokines and insulin resistance. Diabetologia 2003

6. Rajala M, Scherer P. Minireview: The adipocyte – atthe crossroads of energy homeostasis, inflammation,and athersclerosis. Endocrinology 2003; 144:3765–73

References

ch14 14/7/05 4:47 pm Page 160

7. Lyon C, Law R, Hsueh WA. Minireview: Adiposity,inflammation and atherogenesis. Endocrinology2003; 144:2195–200

8. Ahima RS, Prabakaran D, Mantzoros C, et al. Role ofleptin in the neuroendocrine response to fasting.Nature 1996; 382:250–2

9. Flier JS. What’s in a name? In search of leptin’s physi-ological role. J Clin Endocrinol Metab 1998;83:1407–13

10. Borjesson M. The aetiology of obesety in children. Astudy of 101 twin pairs. Acta Paediatr Scand 1976;65:279–87

11. Bouchard C, Tremblay A, Depr‚s J-P, et al.Overfeeding in identical twins: 5–tear postoverfeed-ing results. Metabolism 1996; 45:1042–50

12. Carey D, Campbell L, Chisholm D, et al. Geneticinfluences on central abdominal fat: a twin study. IntJ Obesity (Suppl) 1994; 18 (Supp 2):116

13. Mauriege P, Depres J-P, Marcotte M, et al. Adiposetissue lipolysis after long-term overfeeding in identi-cal twins. Int J Obesity 1992; 16:219–25

14. O’Rahilly S, Farooqi IS, Yeo GSH, et al. Minireview:human obesity—lessons from monogenic disorders.Endocrinology 2003; 144:3757–64

15. Lee DW, Leinung MC, Rozhavskaya-Arena M, et al.Leptin and the treatment of obesity: its current status.Eur J Pharmacol 2002; 440:129–39

16. Oral E, Simha V, Ruiz E, et al. Leptin-replacementtherapy for lipodystrophy. N Engl J Med 2002;346:570–2

17. Tartaglia LA. The leptin receptor. J Biol Chem 1997;272:6093–6

18. Tartaglia LA, Dembski M, Weng X, et al.Identification and expression cloning of a leptinreceptor, OB-R. Cell 1995; 83:1263–71

19. Lynn RB, Cao G-Y, Considine RV, et al. Autoradio-graphic localization of leptin binding in the choroidplexus of {Iob/ob} and {Idb/db} mice. BiochemBiophys Res Commun 1996; 219:884–9

20. White DW, Kuropatwinski KK, Devos R, et al. Leptinreceptor (OB-R) signalling. J Biol Chem 1997;272:4065–71

21. Zigman JM, Elmquist JK. Minireview: From anorexiato obesity—the yin and yang of body weight control.Endocrinology 2003; 144:3749–56

22. Wang Y, Kuropatwinski KKK, White DW, et al. Leptinreceptor action in hepatic cells. J Biol Chem 1997;272:16216–23

23. Serradeil-Le Gal C, Raufaste D, Brossard G, et al.Characterization and localization of leptin receptorsin the kidney. FEBS Lett 1997; 404:185–91

24. Hoggard N, Mercer JG, Rayner DV, et al. Localizationof leptin receptor mRNA splice variants in murineperipheral tissues by RT-PCR and {in situ} hybrid-ization. Biochem Biophys Res Commun 1997;232:383–7

25. Cao G-Y, Considine RV, Lynn RB. Leptin receptors inthe adrenal medulla of the rat. Am J Physiol 1997;273:E448–52

26. Walder K, Filippis A, Clark S, et al. Leptin inhibitsinsulin binding in isolated rat adipocytes. JEndocrinol 1997; 155:R5–7

27. Karlsson C, Lindell K, Svensson E, et al. Expression offunctional leptin receptors in the human ovary. J ClinEndocrinol Metab 1997; 82:4144–8

28. Lostao MP, Urdaneta E, Martinez-Ans¢ E, et al.Presence of leptin receptors in rat small intestine andleptin effect on sugar absorption. FEBS Lett 1998;423:302–6

29. Wang M-Y, Zhou YT, Newgard CB, et al. A novelleptin receptor isoform in rat. FEBS Lett 1996;392:87–90

30. Kieffer TJ, Heller RS, Habener JF. Leptin receptorsexpressed on pancreatic á-cells. Biochem Biophys ResCommun 1996; 224:522–7

31. Emilsson V, Liu Y-L, Cawthorne MS, et al. Expressionof the functional leptin receptor mRNA in pancreaticislets and direct inhibitory action of leptin on insulinsecretion. Diabetes 1997; 46:313–16

32. El-Haschimi K, Pierroz D, Hileman S, et al. Twodefects contribute to hypothalamic leptin resistancein mice with diet-induced obesity. J Clin Invest 2000;105:1827–32

33. Bjírb’k C, Elmquist JK, Frantz JD, et al. Identificationof SOCS-3 as a potential mediator of central leptinresistance. Mol Cell 1998; 1:619–25

34. White DW, Tartaglia LA. Leptin and OB-R: bodyweight regulation by a cytokine receptor. CytokineGrowth Fact Rev 1997; 7:303–9

35. Minokoshi Y, Kim YB, Peroni OD, et al. Leptin stimu-lates fatty-acid oxidation by activating AMP-activatedprotein kinase. Nature 2002; 415:268–9

36. Barash IA, Cheung CC, Weigle DS, et al. Leptin is ametabolic signal to the reproductive system.Endocrinology 1996; 137:3144–7

37. Mounzih K, Lu R, Chehab FF. Leptin treatmentrescues the sterility of genetically obese {1ob/ob}males. Endocrinology 1997; 138:1190–3

38. Hamilton BS. A new role for a fat actor. Nature Med1996; 2:272–3

39. Chehab FF, Lim ME, Lu R. Correction of the sterilitydefect in homozygous obese female mice by treat-ment with the human recombinant leptin. NatureGenet 1996; 12:318–20

40. Farooqi IS, Matarese G, Lord GM, et al. Beneficialeffects of leptin on obesity, T cell hyporesponsiveness,and neuroendocrine/metabolic dysfunction ofhuman congenital leptin deficiency. J Clin Invest2002; 110:1093–103

41. Lord GM, Matarese G, Howard JK, et al. Leptinmodulates the T-cell immune response and reversesstarvation-induced immunosuppression. Nature 1998;394:897–901

42. Montague CT, Farooqi IS, Whitehead JP, et al.Congenital leptin deficiency is associated with severeearly-onset obesity in humans. Nature 1997;387:903–8

43. Farooqi I, Jebb SA, Langmack G, et al. Effects of


ch14 14/7/05 4:47 pm Page 161

recombinant leptin therapy in a child with congenitalleptin deficiency. N Engl J Med 1999; 341:879–84

44. Heymsfield S, Greenberg A, Fujioka K, et al.Recombinant leptin for weight loss in obese and leanadults: a randomized, controlled, dose-escalationtrial. JAMA 2000; 283:1567–8

45. Hu E, Liang P, Spiegelman BM. AdipoQ is a noveladipose-specific gene dysregulated in obesity. J BiolChem 1996; 271:10697–703

46. Arita Y, Kihara S, Ouchi N, et al. Paradoxical decreaseon an adipose-specific protein, adiponectin, inobesity. Biochem Biophys Res Commun 1999;257:79–83

47. Combs T, Berg A, Obici S, et al. Endogenous glucoseproduction is inhibited by the adipose-derivedprotein Acrp30. J Clin Invest 2001; 108:1875–81

48. Berg A, Combs T, Scherer P. ACRP30/adiponectin:an adipokine regulating glucose and lipid metabo-lism. Trends Endocrinol Metab 2002; 13:84–9

49. Berg A, Combs T, Du X, et al. The adipocyte-secretedprotein Acrp30 enhances hepatic insulin action.Nature Med 2001; 7:947–53

50. Bluher M, Michael M, Peroni O, et al. Adipose tissueselective insulin receptor knockout protects againstobesity and obesity-related glucose intolerance.Develop Cell 2002; 3:25–38

51. Hirose H, Kawai T, Yamamoto Y, et al. Effects ofpioglitazone on metabolic parameters, body fat distri-bution, and serum adiponectin levels in Japanesemale patients with type 2 diabetes*1. Metabolism2002; 51:314–17

52. Motoshima H, Wu X, Sinha MK, et al. Differentialregulation of adiponectin secretion from culturedhuman omental and subcutaneous adipocytes: effectsof insulin and rosiglitazone. J Clin Endocrinol Metab2002; 87:5662–7

53. Yamauchi T, Kamon J, Ito Y, et al. Cloning ofadiponectin receptors that mediate antidiabeticmetabolic effects. Nature 2003; 423:762–9

54. Kraegan E, Cooney GJ. The role of free fatty acids inmuscle insulin resistance. In: Marshall S, Home P,Rizza R, eds. The Diabetes Annual 12. Amsterdam:Elsevier, 1999: 141–60

55. Bollheimer LC, Skelly RH, Chester MW, et al.Chronic exposure to free fatty acid reduces pancre-atic β cell insulin content by increasing basal insulinsecretion that is not compensated for by a corre-sponding increase in proinsulin biosynthesis transla-tion. J Clin Invest 1998; 101:1094–101

56. Dobbins RL, Chester MW, Stevenson BE, et al. A fattyacid-dependent step is critically important for bothglucose- and non-glucose stimulated insulin secre-tion. J Clin Invest 1998; 101:2370–6

57. Storlein LH, Kriketos AD, Calvert GD, et al. Fattyacids, triglycerides and syndromes of insulin resis-tance. Prost Leuk Ess Fatty Acids 1997; 57:379–85

58. Kliewer SA, Sundseth SS, Jones SA, et al. Fatty acidsand eicosanoids regulate gene expression throughdirect interactions with peroxisome proliferator-

activated receptors α and γ. Proc Natl Acad Sci USA1997; 94:4318–23

59. Forman BM, Chen J, Evans RM. Hypolipidemic drugs,polyunsaturated fatty acids, and eicosanoids areligands for peroxisome proliferator-activated recep-tors α and γ. Proc Natl Acad Sci USA 1997;94:4312–17

60. Stewart P, Boulton A, Kumar S, et al. Cortisol metab-olism in human obesity: impaired cortisone to cortisolconversion in subjects with central adiposity. J ClinEndocrinol Metab 1999; 84:1022–7

61. Bujalska I, Kumar S, Stewart PM. Does central obesityreflect ‘Cushing’s disease of the omentum’? Lancet1997; 349:1210–13

62. Masuzaki H, Paterson J, Shinyama H, et al. A trans-genic model of visceral obesity and the metabolicsyndrome. Science 2001; 294:2166–70

63. Kotelevtsev Y, Holmes MC, Burchell A, et al. 11beta-Hydroxysteroid dehydrogenase type 1 knockout miceshow attenuated glucocorticoid-inducible responsesand resist hyperglycemia on obesity or stress. ProcNatl Acad Sci USA 1997; 94:14924–9

64. Ailhaud G, Grimaldi P, Negrel J. Cellular and molec-ular aspects of adipose tissue development. Ann RevNutr 1992; 12:207–33

65. Prins JB, O’Rahilly S. Regulation of adipose cellnumber in man. Clin Sci 1997; 92:3–11

66. Rupnick MA, Panigrahy D, Zhang C-Y, et al. From thecover: adipose tissue mass can be regulated through thevasculature. Proc Natl Acad Sci USA 2002; 99:10730–5

67. Hutley L, Herington A, Shurety W, et al. Humanadipose tissue endothelial cells promote preadipocyteproliferation. Am J Physiol Endocrinol Metab 2001;281:E1037–44

68. Wilkison W, Spiegelman B. Biosynthesis of the vasoac-tive lipid monobutyrin. Central role of diacylglycerol.J Biol Chem 1993; 268:2844–9

69. Dallabrida SM, Zurakowski D, Shih S-C, et al. Adiposetissue growth and regression are regulated byangiopoietin-1. Biochem Biophys Res Commun 2003;311:563–71

70. Fukumura D, Ushiyama A, Duda DG, et al. Paracrineregulation of angiogenesis and adipocyte differentia-tion during in vivo adipogenesis. Circ Res 2003;93:88–97

71. Shibuya M. VEGF-receptor inhibitors for anti-angio-genesis. Folia Pharmacol Jap 2003; 122:498–503

72. Karlsson C, Lindell K, Ottosson M, et al. Humanadipose tissue expresses angiotensinogen andenzymes required for its conversion to angiotensin II.J Clin Endocrinol Metab 1998; 83:3925–9

73. Frederich RC, Kahn B, Peach M, et al. Tissue-specificnutritional regulation of angiotensinogen in adiposetissue. Hypertension 1992; 19:339–44

74. Jones BH, Standridge MK, Moustaid N. AngiotensinII increases lipogenesis in 3T3-L1 and human adiposecells. Endocrinology 1997; 138:1512–19

75. Malik F, Lavie C, Mehra M, et al. Renin–angiotensinsystem: genes to bedside. Am Heart J 1998; 136:562–3


ch14 14/7/05 4:47 pm Page 162

76. Shimomura I, Funahashi T, Takahashi M, et al.Enhanced expression of PAI-1 in visceral fat: possiblecontributor to vascular disease in obesity. Nature Med1994; 2:800–3

77. Morange P-E, Aubert J, Pieretti F, et al.Glucocorticoids and insulin promote plasminogenactivator inhibitor 1 production by human adiposetissue. Diabetes 1999; 48:890–5

78. Marx N, Bourcier T, Sukhova G, et al. PPARγ activa-tion in human endothelial cells increases plasmino-gen activator inhibitor type-1 expression. ThrombVasc Biol 1999; 19:546–51

79. He G, Bruun JM, Lihn AS, et al. Stimulation of PAI-1and adipokines by glucose in human adipose tissue invitro. Biochem Biophys Res Commun 2003;310:878–83

80. Flier JS, Cook KS, Usher P, et al. Severely impairedadipsin expression in genetic and acquired obesity.Science 1987; 237:405–8

81. Choy LN, Rosen BS, Spiegelman BM. Adipsin and anendogenous pathway of complement from adiposecells. J Biol Chem 1992; 267:12736–41

82. Hotamisligil GS, Arner P, Caro JF, et al. Increasedadipose tissue expression of tumor necrosis factor-αin human obesity and insulin resistance. J Clin Invest1995; 95:2409–15

83. Hotamisligil GS, Peraldi P, Spiegelman BM. Themolecular link between obesity and diabetes. CurrOpin Endocrinol Diabetes 1996; 3:16–23

84. Hotamisligil GS, Shargill NS, Spiegelman BM.Adipose expression of tumor necrosis factor-α: directrole in obesity-linked insulin resistance. Science 1993;259:87–91

85. Hotamisligil GS, Spiegelman BM. Tumor necrosisfactor alpha: a key component of the obesity-diabeteslink. Diabetes 1994; 43:1271–8

86. Jequier E. A metabolic perspective on the interactionbetween obesity and diabetes. Curr Opin EndocrinolDiabetes 1996; 3:10–15

87. Uysal KT, Wiesbrock SM, Marino MW, et al.Protection from obesity-induced insulin resistance inmice lacking TNF-α function. Nature 1997;389:610–14

88. Ledgerwood EC, Prins JB. TNF alpha. In: MarshallSM, Home PD, Rizza R, eds. Diabetes Annual 12, 1stedn. Oxford: Elsevier, 1998

89. Hotamisligil GS. Molecular mechanisms of insulinresistance and the role of the adipocyte. Int J Obesity2000; 24(Suppl 4):S23–7

90. Liu LS, Spelleken M, R’hrig K, et al. Tumor necrosisfactor α acutely inhibits insulin signaling in humanadipocytes. Diabetes 1998; 47:515–22

91. Peraldi P, Hotamisligil GS, Buurman WA, et al.Tumor necrosis factor (TNF)-α inhibits insulin signal-ing through stimulation of the p55 TNF receptor andactivation of sphingomyelinase. J Biol Chem 1996;271:13018–22

92. Hube F, Hauner H. The two tumor necrosis factorreceptors mediate opposite effects on differentiation

and glucose metabolism in human adipocytes inprimary culture. Endocrinology 2000; 141:2582–8

93. Stephens JM, Lee J, Pilch PF. Tumor necrosis factor-α-induced insulin resistance in 3T3–L1 adipocytes isaccompanied by a loss of insulin receptor substrate-1and GLUT4 expression without a loss of insulin-receptor-mediated signal transduction. J Biol Chem1997; 272:971–6

94. Kanety H, Feinstein R, Papa MZ, et al. Tumor necro-sis factor α-induced phosphorylation of insulin recep-tor substrate-1 (IRS-1). J Biol Chem 1995;270:23780–4

95. Hotamisligil GS, Peraldi P, Budavari A, et al. IRS-1–mediated inhibition of insulin receptor tyrosinekinase activity in TNF-α- and obesity-induced insulinresistance. Science 1996; 271:665–8

96. Scheen AJ, Castillo MJ, Paquot N, et al. Neutralizationof TNFα in obese insulin-resistant human subjects: noeffect on insulin sensitivity. Diabetes (Suppl) 1996;45:286A

97. Ofei F, Hurel S, Newkirk J, et al. Effects of engineeredhuman anti-TNFα antibody (CDP571) on insulinsensitivity and glycemic control in patients withNIDDM. Diabetes 1996; 45:881–5

98. Petruschke TH, Hauner H. Tumor necrosis factor-αprevents the differentiation of human adipocyteprecurser cells and causes delipidation of newly devel-oped fat cells. J Clin Endocrinol Metab 1993;76:742–7

99. Prins JB, Niesler CU, Winterford CM, et al. Tumornecrosis factor-α induces apoptosis of human adiposecells. Diabetes 1997; 46:1939–44

100. Hauner H, Petruschke T, Russ M, et al. Effects oftumour necrosis factor alpha (TNFα) on glucosetransport and lipid metabolism of newly-differenti-ated human fat cells in culture. Diabetologia 1995;38:764–71

101. Van der Poll T, Romijn JA, Endert E, et al. Tumornecrosis factor mimics the metabolic response toacute infection in healthy humans. Am J Physiol 1991;261:E457–65

102. Jones TH. Interleukin-6 an endocrine cytokine. ClinEndocrinol 1994; 40:703–13

103. Bastard J-P, Jardel C, Bruckert E, et al. Elevated levelsof interleukin 6 are reduced in serum and subcuta-neous adipose tissue of obese women after weightloss. J Clin Endocrinol Metab 2000; 85:3338–42

104. Fried SK, Bunkin DA, Greenberg AS. Omental andsubcutaneous adipose tissue depots of obese subjectsrelease interleukin-6: depot difference and regulationby glucocorticoid. J Clin Endocrinol Metab 1998;83:847–50

105. Wallenius V, Wallenius K, Ahrén B, et al. Interleukin-6–deficient mice develop mature-onset obesity.Nature Med 2002; 8:75–9

106. Rosen BS, Cook KS, Yaglom J, et al. Adipsin andcomplement factor D activity: an immune-relateddefect in obesity. Science 1989; 244:1483–7

107. Napolitano A, Lowell BB, Damm D, et al.


ch14 14/7/05 4:48 pm Page 163

Concentration of adipsin in blood and rates ofadipsin secretion by adipose tissue in humans withnormal, elevated and diminished adipose tissue mass.Int J Obesity 1994; 18:213–18

108. Sniderman AD, Cianflone K. The adipsin-ASPpathway and regulation of adipocyte function. AnnMed 1994; 26:389–93

109. Baldo A, Sniderman AD, St-Luce S, et al. The adipsin-acylation stimulating protein system and regulation ofintracellular triglyceride synthesis. J Clin Invest 1993;92:1543–7

110. Murray I, Havel P, Sniderman AD, et al. Reducedbody weight, adipose tissue, and leptin levels despiteincreased energy intake in female mice lacking acyla-tion-stimulating protein. Endocrinology 2000;141:1041–9

111. Biesma D, Hannema A, van Velzen-Blad H, et al. Afamily with complement factor D deficiency. J ClinInvest 2001; 108:233–40

112. Tang W, Miller MB, Rich SS, et al. Linkage analysis ofa composite factor for the multiple metabolicsyndrome: The National Heart, Lung, and BloodInstitute Family Heart Study. Diabetes 2003;52:2840–7

113. Zhang Y, Matheny M, Zolotukhin S, et al. Regulationof adiponectin and leptin gene expression in whiteand brown adipose tissues: influence of

[beta]3–adrenergic agonists, retinoic acid, leptin andfasting. Biochim Biophys Acta (BBA) Mol Cell BiolLipids 2002; 1584:115–22

114. Gui Y, Silha JV, Mishra S, et al. Changes in adipokineexpression during food deprivation in the mouse andthe relationship to fasting-induced insulin resistance.Can J Physiol Pharmacol 2003; 81:979–85

115. Zhang B, Graziano MP, Doebber TW, et al. Down-regulation of the expression of the {1obese} gene byand antidiabetic thiazolidinedione in Zucker diabeticfatty rats and {1db/db} mice. J Biol Chem 1996;271:9455–9

116. Monzillo LU, Hamdy O, Horton ES, et al. Effect oflifestyle modification on adipokine levels in obesesubjects with insulin resistance. Obes Res 2003;11:1048–54

117. Ott V, Fasshauer M, Dalski A, et al. Direct peripheraleffects of ghrelin include suppression of adiponectinexpression. Horm Metab Res 2002; 34:640–5

118. Furuhashi M, Ura N, Higashiura K, et al. Blockade ofthe renin–angiotensin system increases adiponectinconcentrations in patients with essential hyperten-sion. Hypertension 2003; 42:76–81

119. Hulver MW, Zheng D, Tanner CJ, et al. Adiponectinis not altered with exercise training despite enhancedinsulin action. Am J Physiol Endocrinol Metab 2002;283:E861–5


ch14 14/7/05 4:48 pm Page 164

ADIPOCYTE TRIGLYCERIDESTORAGE AND CARDIOVASCULARRISK FACTORS

A major function of adipose tissue is to storeenergy for later use. In this respect, adiposetissue is the most specialized and efficientorgan for energy storage. In fact, adipose tissueis the ultimate repository of dietary fat. Inobesity, i.e. a condition of long-term energyexcess, adipose tissue might store triglyceridesworth months of whole-body energy expendi-ture. Obesity is a high-risk syndrome for cardio-vascular disease and some of the underlyingrisk factors can be traced back to inefficientstorage of fatty acids in adipose tissue.1 It is alsorecognized that differentiated adipocytes fromvarious adipose tissue stores have functionaldifferences in terms of fatty acid handling. Thischapter will describe the principal pathways forstorage and release of fatty acids in adipocytesand how they are regulated. The clinicalrelevance for the respective pathways and howthis might relate to cardiovascular risk factorswill also be highlighted.

BACKGROUND TO ADIPOCYTE FATSTORAGE AND FAT RELEASE

Adipogenesis is regulated by two majortranscriptional factors: peroxisomal prolifera-tor activated receptor gamma (PPAR-γ) andthe sterol regulatory element binding protein 1(SREBP-1).2–4 PPAR-γ appears to regulate a setof genes for triglyceride storage, whereas

SREBP-1 is often seen as the ultimate transcrip-tional signal for the lipogenic pathway inducedby insulin signalling. The CCAAT/enhancer-binding protein (C/EBP) also plays a role inthe early differentiation of adipocytes.2–4

Triglycerides are stored in a large lipid dropletwithin the adipocyte. Adipocyte triglyceridesynthesis is under acute and long-term transcrip-tional regulation by insulin. The acute regulationmay involve cAMP activated dependent proteinkinase regulation of glycerol-3 phosphate acyl-transferase (GPAT). Elevated intracellularconcentrations of cAMP inactivate GPAT. Theacute regulation of diacylglycerol acyl-transferase(DGAT) is, however, largely unknown.5

Mobilization of fatty acids from theadipocyte triglyceride stores is mediated byhormone-sensitive lipase (HSL). A recentlydiscovered additional triglyceride hydrolasemight also be of importance in the mobiliza-tion of intracellular triglycerides.6 Recentcharacterization of the processes of lipidmobilization has shown that the abundantadipocyte protein perilipin A covers the lipiddroplet, which on the one hand appears tofunction as a barrier and on the other as a facil-itating protein for HSL action. Regulation oflipolysis by HSL has recently been outlined byHolm et al.7 HSL is regulated by the reversiblephosphorylation of serine residues. This eventmarks the end of several lipolytic signals culmi-nating in the modulation of the intracellularcAMP concentration. Most lipolytic signalsderive from a range of G-protein-coupledreceptors, which are outlined later in thischapter. The dominant lipolytic stimulus is

Regulation of adipocyte triglyceride storageF. Karpe and G.D. Tan

15

ch15 14/7/05 4:48 pm Page 165

likely to be the signalling of catecholaminesthrough β-adrenoreceptors. Upon stimulation,adenylyl cyclase activity increases and thisgenerates cAMP from ATP. Subsequentphosphorylation of the HSL sites Ser659 andSer660 by cAMP-dependent protein kinase Aappears to activate HSL. However, HSL activa-tion alone is not enough for effective lipolysisto occur; HSL needs to be in physical contactwith its substrate, the triglyceride droplet.Crucially, phosphorylated HSL translocatesfrom the cytoplasm to the lipid droplet.8 Here,the protein envelope of the lipid droplet,perilipin A, plays a role. Like HSL, perilipin Ais phosphorylated by the lipolytic signal, and itis thought that perilipin A phosphorylationfacilitates HSL’s access to its substrate. Incontrast, the unphosphorylated perilipin Aprotects the lipid droplet from lipolysis byHSL. The physiological role of perilipin A hasbeen investigated by the targeted disruption ofthe perilipin A gene (PLIN) in mice. Thesemice were healthy and exhibited resistance todiet-induced obesity. They were leaner thancontrol mice and had reduced adipocytetriglyceride stores with small and dispersedintracellular lipid droplets. Physiologically,they showed constitutively activated lipolysis.9,10

These data clearly highlight the important roleof perilipin A in lipid droplet formation andthe control of lipolysis.

Insulin provides the powerful inhibitorycontrol of fat mobilization from adipocytes.This is mediated through a signal chain involv-ing the insulin receptor, via PI3-kinase formingPIP3, which in turn activates PKB/Akt. PKBthen phosphorylates and activates cAMP-phosphodiesterase 3B (PDE3B). PDE3Bhydrolyses cAMP to AMP, thereby reducingcAMP concentrations.7 The cellular cAMPconcentration is therefore an integrator of theregulation of fat mobilization.

STORAGE

Transport of fatty acids

Most of the fat stored in adipocytes derivesfrom triglyceride-rich lipoproteins such as

chylomicrons and very low density lipoproteins(VLDLs). The triglycerides contained in theselipoproteins are hydrolysed at the vascularendothelium by lipoprotein lipase. The fattyacids produced are transported along theprevailing concentration gradient. In the fedstate, when the generation of fatty acids byadipocytes is low, fatty acids flow through theendothelium to the adipocyte. In contrast, inthe fasted state, the gradient is in the oppositedirection and a majority of the fatty acidsproduced by the hydrolysis of VLDL areunlikely to enter the adipocyte, and insteadenter the systemic circulation.

The mechanism by which fatty acids aretransported across the adipocyte plasmamembrane is controversial. There are twodistinct pathways, which may co-exist, althoughit is unclear which is dominant. Fatty acids aretranslocated across membranes either by trans-porter proteins or by non-protein-mediatedmechanisms. A number of fatty acid transportproteins have been identified. In adipocytes,FATP-1 and FAT/CD36 are the two majorproteins with this function. The activation ofFATP-1 by insulin has recently been describedin adipocytes.11 Insulin mediates translocationof FATP-1 to the plasma membrane withinminutes of insulin exposure, to facilitate theuptake of long-chain fatty acids. Alternatively,there is a distinct possibility that a majorproportion of fatty acid uptake is non-protein-mediated.12,13 This movement can be describedby a simple flip-flop mechanism by which fattyacids bind with high affinity to the phospho-lipids in the plasma membrane, traverse themembrane and are esterified in the cell.

Both these mechanisms are passive anddependent on the concentration gradient offatty acids from the endothelium to theadipocytes. Interference with these mecha-nisms is therefore unlikely to provide promis-ing targets for the alteration of adipocyte fattyacid uptake.

Lipogenesis and triglyceride synthesis

Adipocyte lipogenesis can be divided intotriglyceride synthesis and de novo lipogenesis.


ch15 14/7/05 4:48 pm Page 166

Data in humans indicate that de novo lipogen-esis is a minor pathway in adipose tissue. Undermost metabolic conditions, it is assumed thatde novo lipogenesis is minimal. However,under extreme conditions such as hypercaloriccarbohydrate feeding, it may occur. De novolipogenesis may also take place in the liver, butthe quantitative contribution is likely to besmall. It has therefore been assumed that othertissues such as adipose tissue might contribute.The very last step in de novo lipogenesis,stearoyl-CoA desaturase, is enhanced by thePPAR-γ agonist rosiglitazone in humans.14

Although lipogenesis might not be particularlyactive, some of the enzymes and regulatoryproperties of the pathways involved appear tohave significant effects on adipocyte function.Essentially, all steps of lipogenesis and trigly-ceride synthesis are positively regulated bySREBP-1c. SREBP-1c is localized to theendoplasmic reticulum (ER). The proteinundergoes a sequence of protein cleavages torelease a final peptide fragment that trans-locates to the nucleus to act as a transcriptionalactivator. A cofactor of SREBP-1c, themembrane-spanning SREBP cleavage activat-ing peptide (SCAP) is sensitive to themembrane lipid/cholesterol content. TheSCAP molecule has seven membrane-spanningdomains, which convey sterol sensing. Whenactivated, presumably by a certain cholesteroldensity in the membrane or a low abundanceof polyunsaturated acyl chains in the phospho-lipid structures, SCAP promotes the activity ofa site-1 protease. This allows, in turn, for a site-2 protease that cleaves off the final signallingpeptide derived from the SREBP-1c protein.The cleaved peptide leaves the ER to enter thenucleus and binds to a sterol-regulatedelement (SRE). A large number of genesinvolved in fatty acid synthesis are induced byactivation of SREBP-1c. Fatty acid synthase(FAS), acetyl CoA carboxylase (ACC), stearoylCoA desaturase (SCD-1) and GPAT have SREsand they will co-ordinately promote synthesisof triglycerides.

The SCD-1 step is normally seen as the finalstep in de novo lipogenesis. It converts palmiticacid to monounsaturated palmitoleic acid. The

conversion appears to have major significancein biological terms. Hypothetically, this can beexplained by the biological effect of thebiophysical state of the fatty acid: palmitic acidis crystalline, whereas palmitoleic acid is liquidat 37°C. Therefore, overincorporation ofpalmitic acid into phospholipids in a biologicalmembrane might substantially alter function.The activation of SCD-1 by SREBP-1c mighttherefore be seen as an important feedbackloop to control membrane function. In thiscontext it has also been proposed that theactual cholesterol content of the adipocytemight serve as an internal signal for triglyc-eride storage.15

In mice with targeted disruption of the SCD-1 gene, the storage of triglycerides in adiposetissue is reduced. The mice are lean and do notaccumulate adipose tissue when fed a high-fatdiet. Fat oxidation is increased and the animalsare insulin sensitive, suggesting that the SCD-1gene disruption alters the fuel utilization.16

The final committed step in triglyceridesynthesis in the adipocyte is DGAT. This path-way is dealt with in detail in Chapter 3.

Insulin signalling co-ordinates adipocytelipogenesis. A review of the molecular events ofinsulin signalling has recently beenpublished.17 It has been proposed that aprolonged or intensified signal would enhancethe overall action of insulin18 and in turnadipocyte lipogenesis. In general terms, thiscan be achieved by insulin sensitizers such asthiazolidinediones, but their molecular mecha-nism of action is still unclear. Another option isto interfere with the insulin signalling pathwaydirectly, i.e. with one of the sequential stepsregulating the intracellular phosphorylationstatus. For this reason protein phosphataseinhibitors can be seen as insulin sensitizers andpotentially promote adipocyte fat storage. Theparticular step that has been targeted so far isthe protein tyrosine phosphatase 1B (PTP1B),for which there is abundant in vitro evidence offunctional effects by using PTP1B inhibitors.18

However, the action of this system in terms ofadipocyte triglyceride storage is unclear. As aproof of concept, ablation of PTP1B activitycreates a lean phenotype in mice,19 which

REGULATION OF ADIPOCYTE TRIGLYCERIDE STORAGE 167

ch15 14/7/05 4:48 pm Page 167

would be incompatible with enhancedadipocyte triglyceride storage due to amaintained high state of insulin signalling.These mice have increased insulin sensitivity,which is particularly enhanced in skeletalmuscle. This paradox is explained by anincreased energy expenditure, whose relation-ship to the PTP1B has not been fully explored.

Thiazolidinediones are known to be insulinsensitizers and ligands to PPAR-γ, which is atranscription factor enhancing severallipogenic enzymes. PPAR-γ is highly expressedin adipocytes, but it is unclear whether theeffects on lipogenesis or fatty acid metabolismin adipose tissue are related to the insulinsensitization. Thiazolidinediones appear toreduce the concentration of non-esterifiedfatty acids in rodents,20,21 but this effect is muchless marked in humans.22–25 It was recentlyproposed that thiazolidinediones induceglycerol kinase activity in adipocytes and thatthis would be the mechanistic background tolowering of plasma NEFA concentrations aswell as retention of triglycerides withinadipocytes.26 However, this does not seem tohold true for humans, in whom induction ofglycerol kinase is absent in response to rosigli-tazone.27

METABOLISM OF FATTY ACIDS INHUMAN ADIPOCYTES

Human adipose tissue and adipocytes have avery low oxidative capacity. This is a key featureof white adipose tissue in humans: it stores fatbut does not metabolize it. In contrast, brownadipose tissue has high oxidative capacity andthe uncoupling process gives rise to non-shiver-ing thermogenesis. Neonates may have somebrown adipose tissue, but the feature of non-shivering thermogenesis is lost early duringdevelopment.

The activation of brown adipose tissue hasbeen studied in detail in hibernating animalsand in rodents. There is a strong centralregulation through sympathetic activationsignalling through β-adrenoreceptors. Theatypical β3-adrenoreceptor has attracted partic-

ular attention as it seems specific for this tissuein mammals. It has therefore been suggestedthat agonists for this receptor might play a rolein enhancing uncoupling in brown adiposetissue.28 The role of the receptor in humanadipose tissue is still controversial. The recep-tor is also found in human white adipose tissueand it has therefore been speculated that suchagonists might stimulate non-shivering thermo-genesis in humans. Alternatively, it mightsimply be involved in the regulation of lipolysis,in particular the differential regulationbetween subcutaneous and visceral adiposetissue.29 One line of evidence to suggest afunction of the β3-adrenoreceptor in humanshas been the phenotype of a common geneticvariant, which sometimes shows an associationwith obesity; but the results from a largenumber of association studies are conflict-ing,30,31 One of the problems in employing β3-adrenorecoptor agonists has been the lack ofspecific compounds; activation of β2- and β1-adrenoreceptors might be an underlying causefor unwanted cardiovascular side-effects.

The expression of uncoupling protein-1(UCP-1) is crucial for the mitochondrialuncoupling process; it cannot be replaced byother uncoupling proteins such as UCP-2 orUCP-3.32 As more and more factors that deter-mine adipocyte differentiation become known,it has been speculated that white adipose tissueadipocyte characteristics might not be static;could it be possible to interfere with theprocess and actually turn white adipocytes intobrown adipocytes that burn fat?33,34

Theoretically, this would be a very attractiveapproach to dispose of excess energy andtherefore serve as a potential anti-obesity treat-ment. The expression of PPAR-γ cofactor-1α(PGC-1α) in brown adipose tissue is linked toUCP-1 expression. Tiraby and colleaguesrecently tested whether transient overexpres-sion of PGC-1α in adipocytes affected UCP-1expression.35 Indeed, the mRNA content ofUCP-1 increased, suggesting that a transcrip-tional program linked to energy dissipationhad been turned on. Similar effects were seenin mice after transgenic induction of the PGC-1α like product, PGC-1β.36 The latter observa-


ch15 14/7/05 4:48 pm Page 168

tion is, however, made in a species that alreadyharbours brown adipocytes and is thereforeless conclusive. There are two other examplesof UCP-1 induction in mouse adipose tissue byenhancing upstream regulatory factors. First,adipose tissue-specific induction of PPAR-δappears to induce UCP-1 and physiologicalstudies also demonstrated increased uncou-pling.37 These mice gained less weight whenprovoked with a high-fat diet compared withthe wild type. Second, the translationalinhibitor 4E-BP1, which is a protein that ishighly phosphorylated in response to insulinsignalling and thereby releases a polypeptide(eIF4E) that is involved in translation initia-tion, appears to constitutively depress PGC-1expression and thereby UCP-1.38 Mice with adisruption of the 4E-BP1 gene show a lean andfat-burning phenotype.

In summary, these data demonstrate thedistinct possibility of converting the character-istics of adipocytes from white adipose tissue tothose from brown adipose tissue. If this couldbe therapeutically exploited it might open newways for pharmacological treatment of obesityand its related cardiovascular complications.PPAR-δ agonists are promising agents in thisrespect.

LIPOLYSIS

Adrenoreceptor function in adipocytes

Catecholamine-induced lipolysis is probablythe major trigger for fat mobilization. Inhuman white adipose tissue adipocytes, theadrenergic response is balanced by the lipolyticresponse from β2-adrenoreceptors and theanti-lipolytic response of α2-adrenoreceptors.The relative role of β3-adrenoreceptors iscontroversial.39 The receptor distributionvaries depending on adipose tissue locationand this is likely to be an important underlyingfactor determining the characteristics of thedifferent adiopose tissue depots.39 Buttock fatappears to be resistant to lipolytic stimulus,whereas adipocytes from visceral fat stores arevery sensitive. The upper body subcutaneousadipose tissue is more responsive than the

lower body store. The endogenous ligands forthe adrenoreceptors (adrenaline (epineph-rine) and noradrenaline (norepinephrine))have different relative agonist actions to the β2-and α2-adrenoreceptors, respectively. Thesecatecholamines reach the tissue by differentroutes. There is evidence for innervation ofadipose tissue,40 which would invoke noradren-aline release upon sympathetic activation. Incontrast adrenaline is released from theadrenal medulla and reaches the tissue by theblood stream. Noradrenaline has higheragonist specificity for β2-adrenoreceptors thanadrenaline, but is a weaker α2-adrenoreceptoragonist. In isolated cellular systems, noradren-aline therefore gives rise to a higher lipolyticstimulus than adrenaline.41

The effects resulting from sympathetic activa-tion of adipose tissue are complex. The deliveryof the products of lipolysis, non-esterified fattyacids, from adipose tissue is enhanced by asimultaneous increase in blood flow. Foodintake is followed by insulin secretion thatinhibits HSL action, whilst the post-meal sympa-thetic activation enhances the blood flow.42

Obviously, the inhibitory lipolytic signal ofinsulin is stronger than that of the insulinaemia-induced lipolytic stimulus by sympathetic activa-tion. However, there might be tissue-specificdifferences whereby certain adipose tissuedepots, perhaps in particular visceral fat stores,are particularly sensitive to sympathetic activa-tion and insensitive to the anti-lipolytic action ofinsulin. It has been speculated that the readinessto respond to sympathetic activation ismaintained by pulsatile signals to the tissue inthe resting state,43–45 which has recently beenconfirmed in humans.

In exercise there is a need for lipid mobiliza-tion. In physical exercise, lipolysis is partlypromoted by β2-adrenoreceptor activation inadipocytes.

There is substantial genetic heterogeneity inthe β2-adrenoreceptor gene and this has impli-cations for responsiveness to the lipolyticsignals. Three common haplotypes have beendescribed for the β2-adrenoreceptor gene.46

Recent in vitro data show that there is a 500-fold difference in terbutaline sensitivity


ch15 14/7/05 4:48 pm Page 169

between the extreme variants.47 This finding islikely to have substantial pharmacogenomicimpact, either when the dyslipidaemic effectsof β-adrenoreceptor blockers are evaluated orwhen novel therapeutic approaches are testedin this system.

Atrial natriuretic peptide

Natriuretic peptides have emerged as new andpotent lipolytic factors.48–50 Atrial natriureticpeptide (ANP) is released from the atriumwhereas the brain natriuretic peptide (BNP) isprimarily released from the cardiac ventricle,although it was first identified in the brain. Thestimulus for release is stretching of thecardiomyocytes. ANP and BNP bind to eitherof two G-protein-coupled receptors onadipocytes, the A-type primarily expressed byadipocytes and the more ubiquitouslyexpressed B-type, both of which activate guany-late cyclase giving rise to increased concentra-tions of cGMP. A third receptor, the C-type,lacks the intracellular domain of the receptorstructure and has therefore been suggested totake part in catabolizing ANP and BNP. Almostall other lipolytic responses in adipocytes signalthrough cAMP, and thus the signalling of ANPby cGMP is unique in this respect. It gives riseto HSL and perilipin phosphorylation throughthe cGMP-dependent protein kinase I.51

The plasma concentrations of these peptidesare in the range which is found to stimulatelipolysis from adipocytes in vitro. Physical activ-ity leads to stimulation of sympathetic activa-tion and in turn lipolysis. However, there is aparadoxically low proportion of this lipolyticstimulus that can be blocked by β-adrenore-ceptor blockade. It is likely that ANP releasedas a result of the increased performance of theheart plays a role. Systemic administration of β-adrenoreceptor blockers increases the plasmaconcentrations of ANP. This is likely to dependon the negative chronotropic effect of thesedrugs and the increased end-diastolic fillingpressure of the atrium. It is tempting to specu-late that this might lead to increased lipolysisand partly explain the dysmetabolic effects ofβ-adrenoreceptor blockers. It is therapeutically

interesting to interfere with the ANP systemfrom a metabolic point of view, but there is stilla lack of potent non-peptide antagonists.

Adenosine receptors

Adenosine produced by the adipocyte exerts anegative feedback in response to lipolyticstimuli.52,53 It was later discovered that adeno-sine acts via an inhibitory G-protein-coupledreceptor, reducing the adenylate cyclase activ-ity and, in turn, lowering cAMP concentra-tions.54 Adipocytes express two types ofadenosine receptors, A1 and A2. The A1 recep-tor is found in mature adipocytes and conveysthe classic inhibition of cAMP, whereas A2 isfound in pre-adipocytes and is stimulatory.55

Overexpression of A1 in mice protects themagainst obesity-related insulin resistance.56 Thedevelopment of A1 agonists is, however,hampered by the wide tissue distribution of thereceptor, leading to potential side-effects onheart rate or kidney function. In adipocytes A1appears to be tonically activated at physiologi-cal nanomolar concentrations. There is asubstantial receptor reserve57 and it has beenpostulated that this receptor reserve can beused therapeutically.58 The rapid downregula-tion of the A1 receptor is another caveat in thedevelopment of long-term anti-lipolytic drugs.

Nicotinic acid receptor

Specific binding of nicotinic acid to adiposetissue was demonstrated 40 years ago.59 Thebinding of nicotinic acid to a putative G-protein-coupled receptor has been demon-strated in membranes prepared fromadipocytes.60 More recently, Tunaru andcolleagues61 and Wise and colleagues62 usedgenomic approaches to screen for genesencoding orphan G-protein-coupled receptorswith a tissue distribution resembling that ofnicotinic acid binding, i.e. adipose tissue,spleen and macrophages. Both groups recog-nized that the new receptor had already beencloned as PUMA-G in macrophages (proteinupregulated in macrophages by interferon-γ)63

and as HM7464 in monocytes, respectively. In


ch15 14/7/05 4:48 pm Page 170

addition, Wise et al cloned a gene with a 96%predicted amino acid sequence similarity toHM74 (HM74A),62 whose gene productappears to be the high-affinity receptor fornicotinic acid. Receptor activation by nicotinicacid showed binding to the G-protein-coupledreceptor subunit GTPγS with subsequent cellu-lar response.62 Mice made deficient of thereceptor were viable without any obviouslyabnormal phenotype, but failed to suppresslipolysis after nicotinic acid administration.61

Although nicotinic acid has proved to be avery useful lipid-modifying agent, its use hasalways been hampered by side-effects thatpatients have found difficult to tolerate. Thenicotinic-acid-specific flush is uncomfortableand if new molecules signalling through thisreceptor pathway could be developed thatwould be seen as a major step forward in themanagement of hypertriglyceridaemias. Theendogenous ligand of the nicotinic acid recep-tor is unknown, but the cloning of this receptoris likely to pave the way for anti-lipolytic drugsthat can be used either as hypolipidaemicdrugs or possibly as anti-diabetic drugs.

ADIPOCYTE CORTISOLMETABOLISM

Hypercortisolism has profound effects on

adipose tissue distribution and adipocytefunction. Systemic exposure to cortisolproduces a specific pattern of adipose tissuedepletion in limbs, whilst trunk and visceraladipose tissue expands. The clinical similarityof the adipose tissue distribution in Cushing’ssyndrome and in the male-pattern fat distribu-tion is enigmatic, in the sense that the hypothal-amo–pituitary axis is not obviously disturbed inthe latter condition. The paradox could beexplained by local adipose tissue generation ofcortisol which increases with increasing obesity.Cortisol generated locally would be functionallyantagonistic to insulin. The local cortisol gener-ation affects the adipocytes but has little effecton systemic cortisol concentrations. Theenzyme responsible for the conversion ofbiologically inert cortisone to cortisol inadipose tissue is the 11β-hydroxysteroiddehydrogenase type 1 (11β-HSD1). The role of11β-HSD has recently been reviewed.65,66

Inhibitors of 11β-HSD1 might be a particularlyinteresting target to alter adipocyte function inthe metabolic syndrome.67

ACKNOWLEDGEMENTS

F Karpe is a Senior Wellcome Trust ClinicalResearch Fellow and G Tan is a MedicalResearch Council Training Fellow.


1. Sniderman AD, Cianflone K, Arner P, et al. Theadipocyte, fatty acid trapping and atherogenesis.Arterioscler Thromb Vasc Biol 1998; 18:147–51

2. Rangwala SM, Lazar MA. Transcriptional control ofadipogenesis. Annu Rev Nutr 2000; 20:535–59

3. Tong Q, Hotamisligil GS. Molecular mechanisms ofadipocyte differentiation. Rev Endocr Metab Disord2001; 2:349–55

4. Camp HS, Ren D, Leff T. Adipogenesis and fat-cellfunction in obesity and diabetes. Trends Mol Med2002; 8:442–7

5. Coleman RA, Lewin TM, Muoio DM. Physiologicaland nutritional regulation of enzymes of triacylglyc-erol synthesis. Annu Rev Nutr 2000; 20:77–103

6. Zimmerman R, Strauss JG, Haemmerle G, et al. Fat

mobilization in adipose tissue is promoted by adiposetriglyceride lipase. Science 2004; 306:1383–6

7. Holm C, Osterlund T, Laurell H, et al. Molecularmechanisms regulating hormone-sensitive lipase andlipolysis. Annu Rev Nutr 2000; 20:365–93

8. Clifford GM, Londos C, Kraemer FB, et al.Translocation of hormone-sensitive lipase andperilipin upon lipolytic stimulation of rat adipocytes.J Biol Chem 2000; 275:5011–15

9. Martinez-Botas J, Anderson JB, Tessier D, et al.Absence of perilipin results in leanness and reversesobesity in Lepr(db/db) mice. Nat Genet 2000;26:474–9

10. Tansey JT, Sztalryd C, Gruia-Gray J, et al. Perilipinablation results in a lean mouse with aberrant

References

ch15 14/7/05 4:48 pm Page 171

adipocyte lipolysis, enhanced leptin production, andresistance to diet-induced obesity. Proc Natl Acad SciUSA 2001; 98:6494–9

11. Stahl A, Evans JG, Pattel S, et al. Insulin causes fattyacid transport protein translocation and enhancedfatty acid uptake in adipocytes. Dev Cell 2002;2:477–88

12. Kamp F, Guo W, Souto R, et al. Rapid flip-flop of oleicacid across the plasma membrane of adipocytes. J BiolChem 2003; 278:7988–95

13. Scow RO, Blanchette-Mackie EJ. Why fatty acids flowin cell membranes. Prog Lipid Res 1985; 24:197–241

14. Riserus U, Tan GD, Fielding BA, et al. Rosiglitazoneincreases indexes of stearoyl-CoA desaturase activityin humans: link to insulin sensitization and the role ofdominant-negative mutation in Peroxisome Pro-liferator-Activated Receptor-γ. Diabetes 2005;54:1379–84

15. Dugail I, Le Lay S, Varret M, et al. New insights intohow adipocytes sense their triglyceride stores. Ischolesterol a signal? Horm Metab Res 2003; 35:204–10

16. Ntambi JM, Miyazaki M, Stoehr JP, et al. Loss ofstearoyl-CoA desaturase-1 function protects miceagainst adiposity. Proc Natl Acad Sci USA 2002;99:11482–6

17. Saltiel AR, Pessin JE. Insulin signaling pathways intime and space. Trends Cell Biol 2002; 12:65–71

18. Xie L, Lee SY, Andersen JN, et al. Cellular effects ofsmall molecule PTP1B inhibitors on insulin signaling.Biochemistry 2003; 42:12792–804

19. Klaman LD, Boss O, Peroni OD, et al. Increasedenergy expenditure, decreased adiposity, and tissue-specific insulin sensitivity in protein-tyrosine phos-phatase 1B-deficient mice. Mol Cell Biol 2000;20:5479–89

20. Zhou YT, Grayburn P, Karim A, et al. Lipotoxic heartdisease in obese rats: implications for human obesity.Proc Natl Acad Sci USA 2000; 97:1784–9

21. Finegood DT, McArthur MD, Kojwang D, et al. Beta-cell mass dynamics in Zucker diabetic fatty rats.Rosiglitazone prevents the rise in net cell death.Diabetes 2001; 50:1021–9

22. Maggs DG, Buchanan TA, Burant CF, et al. Metaboliceffects of troglitazone monotherapy in type 2 diabetesmellitus. A randomized, double-blind, placebo-controlled trial. Ann Intern Med 1998; 128:176–85


24. Hällsten K, Virtanen KA, Lönnqvist F, et al.Rosiglitazone but not metformin enhances insulin-and exercise-stimulated skeletal muscle glucoseuptake in patients with newly diagnosed type 2diabetes. Diabetes 2002; 51:3479–85

25. Tan GD, Fielding BA, Currie JM, et al. Nitric oxideand beta-adrenergic stimulation are majhor regula-tors of preprandial and postprandial subcutaneous

adipose tissue blood flow in humans. Circulation2004; 109:47–52

26. Guan HP, Li Y, Jensen MV, et al. A futile metaboliccycle activated in adipocytes by antidiabetic agents.Nat Med 2002; 8:1122–8

27. Tan GD, Debard C, Tiraby C, et al. A ‘futile cycle’induced by thiazolidinediones in human adiposetissue? Nat Med 2003; 9:811–12

28. Arch JR. beta(3)-Adrenoceptor agonists: potential,pitfalls and progress. Eur J Pharmacol 2002;440:99–107

29. Lönnqvist F, Thome A, Nilsell K, et al. A pathogenicrole of visceral fat beta 3-adrenoceptors in obesity. JClin Invest 1995; 95:1109–16

30. Fujisawa T, Ikegami H, Kawaguchi Y, et al. Meta-analysis of the association of Trp64Arg polymorphismof beta 3-adrenergic receptor gene with body massindex. J Clin Endocrinol Metab 1998; 83:2441–4

31. Allison DB, Heo M, Faith MS, et al. Meta-analysis ofthe association of the Trp64Arg polymorphism in thebeta3 adrenergic receptor with body mass index. Int JObes Relat Metab Disord 1998; 22:559–66

32. Nedergaard J, Golozoubova V, Matthias A, et al.UCP1: the only protein able to mediate adaptive non-shivering thermogenesis and metabolic inefficiency.Biochim Biophys Acta 2001; 1504:82–106

33. Chen HC, Farese RV, Jr. Turning WAT into BAT getsrid of fat. Nat Med 2001; 7:1102–3

34. Walczak R, Tontonoz P. Setting fat on fire. Nat Med2003; 9:1348–9

35. Tiraby C, Tavernier G, Lefort C, et al. Acquirement ofbrown fat cell features by human white adipocytes. JBiol Chem 2003; 278:33370–6

36. Kamei Y, Ohizumi H, Fujitani Y, et al. PPARgammacoactivator 1beta/ERR ligand 1 is an ERR proteinligand, whose expression induces a high-energyexpenditure and antagonizes obesity. Proc Natl AcadSci USA 2003; 100:12378–83

37. Wang YX, Lee CH, Tiep S, et al. Peroxisome-prolifer-ator-activated receptor delta activates fat metabolismto prevent obesity. Cell 2003; 113:159–70

38. Tsukiyama-Kohara K, Poulin F, Kohara M, et al.Adipose tissue reduction in mice lacking the transla-tional inhibitor 4E-BP1. Nat Med 2001; 7:1128–32

39. Lafontan M, Berlan M. Do regional differences inadipocyte biology provide new pathophysiologicalinsights? Trends Pharmacol Sci 2003; 24:276–83

40. Dodt C, Lönnroth P, Wellhoner JP, et al. Sympatheticcontrol of white adipose tissue in lean and obesehumans. Acta Physiol Scand 2003; 177:351–7

41. Carpene C, Bousquet-Melou A, Galitzky J, et al.Lipolytic effects of beta 1-, beta 2-, and beta 3-adren-ergic agonists in white adipose tissue of mammals.Ann NY Acad Sci 1998; 839:186–9

42. Ardilouze JL, Fielding BA, Currie JM, et al. Nitricoxide and beta-adrenergic stimulation are majorregulators of preprandial and postprandial subcuta-neous adipose tissue blood flow in humans.Circulation 2004; 109:47–52


ch15 14/7/05 4:48 pm Page 172

43. Hucking K, Hamilton-Wessler M, Ellmerer M, et al.Burst-like control of lipolysis by the sympatheticnervous system in vivo. J Clin Invest 2003; 111:257–64

44. Getty L, Panteleon AE, Mittelman SD, et al. Rapidoscillations in omental lipolysis are independent ofchanging insulin levels in vivo. J Clin Invest 2000;106:421–30

45. Karpe F, Fielding BA, Coppack SW, et al. Oscillationsof fatty acid and glycerol release from human subcu-taneous adipose tissue in vivo. Diabetes 2005;54:1297–303

46. Drysdale CM, McGraw DW, Stack CB, et al. Complexpromoter and coding region beta 2-adrenergic recep-tor haplotypes alter receptor expression and predictin vivo responsiveness. Proc Natl Acad Sci USA 2000;97:10483–8

47. Eriksson P, Dahlman I, Ryden M, et al. Relationshipbetween beta-2 adrenoceptor gene haplotypes andadipocyte lipolysis in women. Int J Obes Relat MetabDisord 2004; 28:185–90

48. Sengenes C, Berlan M, De Glisezinski I, et al.Natriuretic peptides: a new lipolytic pathway inhuman adipocytes. FASEB J 2000; 14:1345–51

49. Moro C, Galitzky J, Sengenes C, et al. Functional andpharmacological characterization of the natriureticpeptide-dependent lipolytic pathway in human fatcells. J Pharmacol Exp Ther 2004; 308:984–92

50. Dessi-Fulgheri P, Sarzani R, Rappelli A. Role of thenatriuretic peptide system in lipogenesis/lipolysis.Nutr Metab Cardiovasc Dis 2003; 13:244–9

51. Sengenes C, Bouloumie A, Hauner H, et al.Involvement of a cGMP-dependent pathway in thenatriuretic peptide-mediated hormone-sensitivelipase phosphorylation in human adipocytes. J BiolChem 2003; 278:48617–26

52. Ho R, Russell TR, Asakawa T, et al. Cellular levels offeedback regulator of adenylate cyclase and the effectof epinephrine and insulin. Proc Natl Acad Sci USA1975; 72:4739–43

53. Ho RJ, Sutherland EW. Action of feedback regulatoron adenylate cyclase. Proc Natl Acad Sci USA 1975;72:1773–7

54. Londos C, Cooper DM, Schlegel W, et al. Adenosineanalogs inhibit adipocyte adenylate cyclase by a GTP-dependent process: basis for actions of adenosine andmethylxanthines on cyclic AMP production and lipol-ysis. Proc Natl Acad Sci USA 1978; 75:5362–6

55. Borglum JD, Vassaux G, Richelsen B, et al. Changes inadenosine A1- and A2-receptor expression duringadipose cell differentiation. Mol Cell Endocrinol1996; 117:17–25

56. Dong Q, Ginsberg HN, Erlanger BF. Overexpressionof the A1 adenosine receptor in adipose tissueprotects mice from obesity-related insulin resistance.Diabetes Obes Metab 2001; 3:360–6

57. Liang HX, Belardinelli L, Ozeck MJ, et al. Tonic activ-ity of the rat adipocyte A1-adenosine receptor. Br JPharmacol 2002; 135:1457–66

58. Dhalla AK, Shryock JC, Shreeniwas R, et al.Pharmacology and therapeutic applications of A1adenosine receptor ligands. Curr Top Med Chem2003; 3:369–85

59. Carlson LA, Hanngren Å. Initial distribution in miceof 3H-labelled nicotinic acid studied with autoradiog-raphy. Life Sci 1964; 3:867–71

60. Lorenzen A, Stannek C, Lang H, et al. Character-ization of a G protein-coupled receptor for nicotinicacid. Mol Pharmacol 2001; 59:349–57

61. Tunaru S, Kero J, Schaub A, et al. PUMA-G and HM74are receptors for nicotinic acid and mediate its anti-lipolytic effect. Nat Med 2003; 9:352–5

62. Wise A, Foord SM, Fraser NJ, et al. Molecular identi-fication of high and low affinity receptors fornicotinic acid. J Biol Chem 2003; 278:9869–74

63. Schaub A, Futterer A, Pfeffer K. PUMA-G, an IFN-gamma-inducible gene in macrophages is a novelmember of the seven transmembrane spanning recep-tor superfamily. Eur J Immunol 2001; 31:3714–25

64. Nomura H, Nielsen BW, Matsushima K. Molecularcloning of cDNAs encoding a LD78 receptor andputative leukocyte chemotactic peptide receptors. IntImmunol 1993; 5:1239–49

65. Seckl JR, Walker BR. Minireview: 11beta-hydroxy-steroid dehydrogenase type 1 – a tissue-specific ampli-fier of glucocorticoid action. Endocrinology 2001;142:1371–6

66. Stulnig TM, Waldhäusl W. 11beta-Hydroxysteroiddehydrogenase type 1 in obesity and type 2 diabetes.Diabetologia 2004; 47:1–11

67. Walker BR, Seckl JR. 11beta-Hydroxysteroid dehydro-genase type 1 as a novel therapeutic target inmetabolic and neurodegenerative disease. ExpertOpin Ther Targets 2003; 6:771–83


ch15 14/7/05 4:48 pm Page 173

ch15 14/7/05 4:48 pm Page 174

Humans live in close association with vastnumbers of microorganisms that are presenton the external and internal surfaces of ourbody. The ability to mount a prominent inflam-matory response to pathogens confers a contin-uous advantage in our fight against pathogens.All metazoan organisms have evolved compleximmune defence systems, used to repelinvasive microbes that would parasitize or killthem. Innate immunity is the most universaland the most rapidly acting and most organ-isms survive through the action of this mecha-nism alone. After any trauma or infection,organisms mount a homeostatic response toinjury which is called an acute-phase response,a highly complex process. In the acute phase,the response is protective because it counter-acts the effects of injury and improves survival.Long-term exposure to stressful stimuli(mucositis, aging, increased fat intake, peri-odontitis, etc.), however, may result in disease(atherosclerosis) rather than repair.

There exist two arms of innate immunity: thesensing arm (those mechanisms involved in thecontinuous sensing and perception of infec-tion) and the effector arm, the sophisticatedprocesses aimed at eradication of infection andtissue repair. Each of these arms may be subdi-vided into humoral and cellular processeswhich are tightly co-ordinated in the inflam-matory process.1

In this chapter we will evaluate each part ofthe innate immune defence in relation to theatherosclerotic process.

AFFERENT ARM OF INNATEIMMUNITY AND ATHEROSCLEROSIS(Figure 1)

Cellular sensing

The chronic inflammatory process of athero-sclerosis is triggered and sustained by unknownfactors. Among the candidate triggers areoxidized or enzymatically modified low-densitylipoproteins, heat shock proteins and infec-tious pathogens. Interestingly, in recent years ithas become clear that all these triggers and

Role of innate immunity inatherosclerosis: immune recognition,immune activation and theatherosclerostic processJ-M. Fernández-Real

16

Figure 1 Afferent arm of innate immunity andatherosclerosis. Cellular components simultaneouslyinvolved in innate immune sensing and theatherosclerotic process

TLR4

CD14

ch16 14/7/05 4:48 pm Page 175

ligands could be recognized and sensed by thesame cell and the same receptor.

Macrophages play a primary role in hostdefence against infection, utilizing a range ofreceptors to recognize microbes by opsonic aswell as direct interactions. The term ‘pathogen-associated molecular patterns’ (PAMPs) wascoined to describe those microbial principlesthat triggered an innate immune response.2

PAMPs were said to act via ‘pattern recognitionreceptors’ (PRR),2 i.e. those sensors that couldrecognize a pattern on a microbe. Binding oftargets via PRR results in phagocytosis andkilling. Macrophages express a broad reper-toire of PRR (e.g. scavenger and lectin-like). Inthis sense, the amount of lipid retained inmacrophages during the atheroscleroticprocess depends on the unregulated uptake ofoxidized lipoproteins by scavenger receptors,counterbalanced by degradation and choles-terol efflux. This scavenger receptor also playsa major role in microbial uptake in the absenceof opsonins.

The principal signalling receptors of theinnate immune system – through which thegreater part of the host awareness of infectionis processed – are the toll-like receptor (TLR)family of transmembrane molecules. The bestunderstood TLR, both in terms of ligandbinding and signal transduction, is thelipopolysaccharide (LPS) receptor, TLR4.

Toll-like receptor-4

An involvement of toll-like receptor-4 (TLR4)in the pathogenesis of atherosclerosis has beenrecently emphasized. Expression of TLR4 isclearly detected by immunohistochemistry inmacrophages of lipid-rich, human atheroscle-rotic plaques and in lesions of atherosclerosis-susceptible mice deficient for apolipoproteinE.3 This expression (evaluated by immunohis-tochemistry and semiquantitative polymerasechain reaction) was paralleled with inflamma-tory activation in the lesions as assessed bynuclear translocation of NF-κB.4 Fibrousplaques and normal arteries show almost noTLR4 expression.3 Basal TLR4 messenger RNAexpression of human monocyte-derived

macrophages is markedly upregulated upon in-vitro loading by oxidized LDL.

TLR ligation transmits transmembranesignals that activate NF-κB and mitogen-activated protein kinase (MAPK) pathways,inducing expression of a wide variety of genessuch as those encoding proteins involved inleukocyte recruitment, production of reactiveoxygen species and phagocytosis. Activation ofTLR will also elicit the production of cytokinesthat potentiate local inflammation.

In the last few years, retrospective andprospective studies of markers of inflammationin humans have shown that a range of differentinfectious agents may be associated with accel-erated atherosclerosis. Very recently, a directlink between microbial infection and lipidaccumulation in macrophages has also beendemonstrated.5 A family of transcriptionalregulators in macrophages (named liver Xreceptors, LXR) can modulate LPS-inducedTLR4 pathways, and selected TLRs can inhibitLXR function by cross-talk.5,6 Given that LXRpromotes synthesis of ABCA1 and other trans-porters, which reduce intracellular levels ofcholesterol, these observations could aid in theprevention of atherosclerosis.

A polymorphism in the TLR4 gene affects theinflammatory response to lipopolysaccharide(LPS), with impact on the risk of Gram-negativeinfections and septic shock.7 Two co-segregat-ing mutations in the gene region coding for theextracellular domain of TLR4, characterized bya substitution at amino acid position 299(glycine for aspartate) and another at position399 (isoleucine for threonine), are relativelycommon. A study reported an association ofthis TLR4 polymorphism with atherogenesis.8

These variants, Asp299Gly and Thr399Ile, leadto a blunted immunological response toinhaled LPS9 and to lower levels of pro-inflam-matory cytokines, acute-phase reactants,fibrinogen, systemic interleukin-6 and solubleadhesion molecules.8 Last, and most strikingly,they appear to be associated with reducedextent and progression of carotid atherosclero-sis as quantified by B-mode ultrasound.8

The TLR4 gene polymorphism has also beenevaluated in a study in 655 men with


ch16 14/7/05 4:48 pm Page 176

angiographically documented coronary athero-sclerosis, who participated in a prospectivecholesterol-lowering trial. These patients wererandomly assigned to either pravastatin orplacebo for 2 years, for evaluation of the effecton coronary artery disease. There were nosignificant differences between geneticallydefined sub-groups with respect to baseline riskfactors, treatment or in-trial changes of lipid,lipoprotein or angiographic measurements.Genotype was not associated with progressionof atherosclerosis. In the pravastatin group,299Gly carriers had a lower risk of cardiovascu-lar events during follow-up than non-carriers(2.0% versus 11.5%, p = 0.045). Among non-carriers, pravastatin reduced the risk of cardio-vascular events from 18.1% to 11.5% (p = 0.03),whereas among 299Gly carriers this risk wasstrikingly reduced from 29.6% to 2.0% (p =0.0002, p = 0.025 for interaction).10 Theseresults suggested that the TLR4 Asp299Glypolymorphism was associated with the risk ofcardiovascular events and also modified theefficacy of statins in preventing cardiovascularevents.10

In another study, however, no association ofthe TLR4 polymorphism with individualparameters of sub-clinical inflammation orwith parameters of the metabolic syndrome (asituation with a well-described increased pro-inflammatory response) was found.11 Subjectswith one or two alleles causing the 299GlyTLR4 did not differ from carriers of TLR4 Asphomozygotes with regard to hypertension,obesity, waist circumference or HDL-choles-terol levels. Differences were also not observedfor systemic levels of IL-6, IL-6 receptor, Creactive protein or fibrinogen. Of all parame-ters analysed, only the prevalence of hyperten-sion showed a trend (p = 0.07), leaving thepossibility of a mild protective effect of theGly299 TLR4 allele.11

Heat shock protein 60 (hsp60) is a putativeendogenous ligand of the TLR4 complex.12

Since hsp60 is an important player in chronicinflammatory conditions, the TLR4 polymor-phism may regulate through hsp60 the subclin-ical inflammation underlying the pathogenesisof arteriosclerosis. In fact, in a femoral artery

cuff model in the atherosclerotic ApoE3(Leiden) transgenic mouse, TLR4 activation byLPS stimulated plaque formation and subse-quent outward arterial remodelling. With theuse of the same model in wild-type mice, neoin-tima formation and outward remodellingoccurred. In TLR4-deficient mice, however, nooutward arterial remodelling was observedindependent of neointima formation. Carotidartery ligation in wild-type mice resulted inoutward remodelling without neointimaformation in the contralateral artery. This wasassociated with an increase in TLR4 expressionand Hsp60 mRNA levels. In contrast, outwardremodelling was not observed after carotidligation in TLR4-deficient mice. These findingsprovided genetic evidence that TLR4 isinvolved in outward arterial remodelling.13

TLR4-deficient mice also sustained smallerinfarctions and exhibit less inflammation aftermyocardial ischaemia-reperfusion injury.14

CD14 receptor

The CD14 receptor, as TLR4, is in the cross-roads between infection and the immunesystem. Several molecules bind lipopolysaccha-ride (LPS) and subsequently activate theresting monocyte/macrophages, playing animportant role in the internalization anddetoxification of LPS. However, CD14 is themain LPS receptor that can activate monocytesin conjunction with serum LPS-bindingprotein and TLR4 at low (<10 ng/ml) clinicallysignificant concentrations of LPS. CD14 inter-acts with different components of Gram-negative and -positive bacteria and fungi,defining CD14 as a central PRR in innateimmunity. Other agonists, notably the heatshock protein HSP70, also activate monocytesby binding to CD14.15

CD14 is a membrane glycoprotein (with aglycosyl-phosphatidyl inositol (GPI) anchor)which is present on the surface of differentmyeloid cells and at very low numbers on B-lymphocytes, basophils and gingival fibroblasts.The CD14 protein has no transmembrane orcytoplasmic domain. CD14 is a constituent of amulti-ligand PRR complex, in which TLRs

ROLE OF INNATE IMMUNITY IN ATHEROSCLEROSIS 177

ch16 14/7/05 4:48 pm Page 177

show a major impact on CD14 signalling inmacrophages.

Two European groups reported a relation-ship between promoter polymorphism in theCD14 gene and increased risk of atherosclero-sis. One group found that T>C at position –260was associated with increased risk of myocar-dial infarction, and the other reported C>T wasmore frequent in myocardial infarctionsurvivors than controls.16,17

Two Japanese groups confirmed the occur-rence of the C (–260) nucleotide change andan associated predisposition to increased riskof coronary artery disease.18,19 The polymor-phism was apparently associated with anenhanced risk for myocardial infarction (MI),particularly in patients who did not other-wise have any significant risk profile for athero-sclerosis.

More recently, this functional polymorphismin the promoter region of the CD14 gene wasstudied to determine its impact on commoncarotid artery (CCA) intima-media thickness(IMT) and any interactions with environmen-tal inflammatory stimuli. The CC genotype wasassociated with increased CCA IMT. The age-and sex-adjusted odds ratio for IMT above the75th percentile was 1.63 (95% CI, 1.19 to 2.24;p = 0.002) and 1.70 (95% CI, 1.18 to 2.44; p =0.004) after additional adjustment for conven-tional risk factors. This gene effect was foundonly in current smokers and ex-smokers. Multi-variate analysis in this group (n = 503)increased the odds ratio to 2.02 (95% CI, 1.23to 3.34; p = 0.006). No significant interactionswere found in non-smokers. This studysuggested that CD14 may modulate the inflam-matory effects of smoking in atherogenesis.20

These observations are interesting becausebacterial LPS, a potent mediator of inflamma-tion, has been identified as an active compo-nent of cigarette smoke,21 and smokers haveelevated plasma levels of LPS.22 Circulatinglevels of LPS, in turn, have been shown toindependently predict incident atherosclerosismeasured by carotid ultrasound,22 but theability of LPS to promote atherogenesisappears to be dependent on the degree ofinflammatory response it provokes.23

Because the polymorphism is associated withan upregulation of CD14 receptors onmonocytes, these observations corroborate thegrowing evidence that chronic infections of,e.g. Chlamydia pneumonia, Helicobacter pylori,Epstein Barr virus, etc., or other inflammatorytriggers (smoking) may be important riskfactors for the development of atherosclerosisand consequently of MI.

Humoral sensing (Figure 2)

Pentraxins

The pentraxin family, named, for its electronmicrographic appearance, from the Greek penta(five) and ragos (berries),24 comprises C-reactiveprotein (CRP) and serum amyloid P (SAP)component in man, and is highly conserved inevolution, with homologous proteins throughoutthe vertebrates and even in the phylogeneticallydistant arthropod, Limulus polyphemus, and thehorseshoe crab.25 SAP, named for its universalpresence in amyloid deposits, is a constitutive,non-acute-phase plasma glycoprotein in man andall other species studied, except the mouse, inwhich it is the major acute-phase protein.

Human CRP is a calcium-dependent ligand-binding protein, which binds with highestaffinity to phosphocholine residues. Extrinsicligands include many glycan, phospholipid andother components of micro-organisms, such ascapsular and somatic components of bacteria,fungi and parasites, as well as plant products.Autologous ligands include native andmodified plasma lipoproteins, damaged cellmembranes, a number of different phospho-lipids and related compounds, small nuclearribonucleoprotein particles and apoptotic cells(reviewed in references 26 and 27).

When human CRP is ligand-bound, it isrecognized by C1q and potently activates theclassical complement pathway, engaging C3,the main adhesion molecule of the comple-ment system, and the terminal membraneattack complex, C5–C9. Bound CRP may alsoprovide secondary binding sites for factor H,and thereby regulate alternative pathwayamplification and C5 convertases.26


ch16 14/7/05 4:48 pm Page 178

The first suggestion of a possible relation-ship of CRP to atherosclerosis came when itwas demonstrated that aggregated, but notnative, non-aggregated, CRP selectively boundjust LDL and some VLDL from whole serum.Native CRP also binds to oxidized LDL. Thereis robust immunohistochemical evidence ofCRP deposition within all acute myocardialinfarcts, co-localized with activated comple-ment components.28 Thus, CRP couldcontribute to complement activation andinflammation in the plaques.27

Endothelial dysfunction, a marker of athero-sclerosis related to coronary events, is associatedin epidemiological studies with markers ofsystemic inflammation, including CRP produc-tion. It is well demonstrated that circulatingCRP concentration in apparently healthypeople predicts furure cardiovascular events.29–32

Finally, there exists an epidemiological associa-tion between higher peak CRP values and poorprognosis of ischaemic heart disease.27

In recent years, a number of new pentraxinshave been discovered, including pentraxin 3(PTX3), neuronal pentraxin 1 and neuronalpentraxin 2. These molecules are known aslong pentraxins and are approximately twicethe size of the prototypic pentraxins CRP and

SAP. PTX3 was the first long pentraxin to bediscovered and its expression is induced inresponse to inflammatory stimuli, includingtumour necrosis factor-alpha (TNF-α), inter-leukin (IL)-1β and LPS. PTX3 is produced by arange of cell types, including monocytes/macrophages, endothelial cells and fibroblasts,but is not produced by hepatocytes, which area major source of CRP. Like CRP, PTX3 is ableto bind the C1q complement component andit has been proposed that PTX3 may play thesame function in the periphery as CRP does inthe circulation.

Strong PTX3 staining has been found inmacrophages and endothelial cells in advancedatherosclerotic lesions and in smooth musclecells.33 In contrast, sections from non-athero-sclerotic internal mammary arteries did notexpress PTX3. Increased serum PTX3 has beenalso detected in the blood of patients withacute MI.34

Soluble CD14

CD14 also exists as a soluble form (sCD14)found in normal human serum. sCD14 is appar-ently derived both from secretion of CD14 andfrom enzymatically cleaved GPI-anchored tissue


Pentraxins Soluble CD14

Mannose-binding lectin

Complement

Adiponectin

Figure 2 Afferentarm of innateimmunity andatherosclerosis.Humoral componentssimultaneouslyinvolved in innateimmune sensing andthe atheroscleroticprocess

ch16 14/7/05 4:48 pm Page 179

CD14. Soluble CD14 has been shown toenhance the endotoxin-neutralization capacityof plasma.35–37 In fact, plasma lipoproteinspromote the release of bacterial LPS from themonocyte cell surface and sCD14 is involved inthis process. Neutralization of LPS by reconsti-tuted lipoprotein particles is accelerated morethan 30-fold by addition of sCD14.38,39

Circulating sCD14 concentration has beenfound to be associated with several cardiovascu-lar risk factors such as waist diameter, bloodpressure, insulin resistance, plasma triglyc-erides and serum uric acid concentration.40

This observation is important in the sense thatLPS, one of the most potent biological responsemodifiers currently recognized, circulates innormal humans attached to triglyceride-richlipoproteins.41,42 LPS is extraordinarily ubiqui-tous in nature, being present in food and water,and in normal indoor environments as aconstituent of house dust.43 Endogenous LPS iscontinually produced within the gut by thedeath of Gram-negative bacteria and isabsorbed into intestinal capillaries. Low-gradeportal venous LPS has been claimed to be thestatus quo in humans.44 Decreased efficiency inneutralizing LPS-induced responses washypothesized to lead to a chronic pro-inflam-matory response and insulin resistance.40 Thiswas further supported by the finding ofnegative correlations betweeen serum sCD14and circulating concentrations of soluble TNF-α receptors in healthy subjects.40 sCD14 concen-tration has also been described to be linked toaortic stiffness.45

Mannose-binding lectin

Mannose-binding lectin (MBL) is a circulatingimmune factor responsible for opsonization ofpathogens by binding mannose moieties ontheir surface and directly activating comple-ment via the lectin pathway before antibodyformation. Common variations in the MBLgene are responsible for an opsonic deficiencythat affects 5% to 7% of Caucasian subjects.Deficiencies in MBL can be caused by threesingle nucleotide polymorphisms within exon 1of the MBL gene on chromosome 10: allele B at

codon 54 (G54D), allele C at codon 57 (G57E)and allele D at codon 52 (R52C), with the mostcommon codon at these loci designated alleleA. This effect is substantially modulated by atleast four promoter polymorphisms, includingthe H/L and X/Y systems, which show reduc-tions of MBL of up to 85% among individualshomozygous for the LX (‘low’) promoters.46

The structural variations have typically beenlabelled ‘O’ alleles, in contrast to the mostcommon ‘A’ allele. The presence of a heterozy-gous genotype (AO) results in an approximate8-fold reduction in MBL levels, but there isconsiderable overlap in the distribution of MBLlevels in those with AA and AO genotypes.These structural variations leading to decreasedcirculating MBL concentration were associatedwith increased risk of certain infectious condi-tions47,48 and, interestingly, were also predictiveof coronary artery disease (CAD) in a recentstudy in American Indians of different ethnicityliving in three different locations.49 This wasparticularly remarkable in this population giventhe marked presence of other CAD risk factorssuch as type 2 diabetes, hypertension andalbuminuria. A significant association betweenpersistent infection with Chlamydia pneumoniaeand CAD has also been reported, but only inthe context of OO or AO structural MBLgenotypes.50 An early report of an associationbetween MBL genotypes and CAD from Norwayindicated that the prevalence of homozygousstructural (but not heterozygous) genotypespredicting low levels of MBL was increasedamong those with prior coronary artery bypassprocedures compared with normal blooddonors.51 MBL variants were also associatedwith a slightly higher mean area of plaquedetected in the carotid artery in whites at highrisk for CAD.52 In another study of US physi-cians, no relationship was noted between MBLlevels and self-reported peripheral arterialdisease.53

Complement

Complement is a term referring to a collection ofplasma proteins, specific cellular receptors andcell surface regulatory molecules. Complement


ch16 14/7/05 4:48 pm Page 180

represents an important innate immune defencesystem to discriminate ‘self’ from ‘non-self’.Invading pathogens are normally attacked byalternative and MBL pathways (triggered by thesurface composition of the invader) and also bythe classical pathway (triggered by specificantibodies targeted towards the intruder ordirectly as, for example, in the case of severalviruses and bacteria). Chemotaxis of phagocyticcells, opsonization and lysis of the microbe thenmostly lead to limitation of the attack and controlof the infection. This type of humoral innate hostdefence plays a crucial role and is executed onviruses, bacteria, fungi and parasites.54

Activation of the complement system tocompletion results in the formation of C5b-9terminal complexes. These complexes havebeen observed in human atherosclerotic lesionsby immunohistochemistry. Endothelial celldamage leads to complement activation andendothelial cells overlying atheroscleroticlesions have been observed to contain C3 andC5b-9 antigens. Cardiac myocytes stain forcomplement proteins (C3, C4 and C5b-9)following MI. Infarct size and extent of inflam-matory cell infiltrates are diminished by decom-plementation prior to experimentally inducedmyocardial ischaemia. Following MI and ulcer-ation of atherosclerotic lesions in humanpatients there is an increase in circulatingcomplement activation products and a decreasein the level of native C1 through C4 proteins.55

Macrophage complement receptors, C3breceptor (CR1) and C3bi receptor, were alsoexpressed in the atherosclerotic lesions whenthe complement system was activated.56

Thus, complement plays an important rolein immune recognition, immune response, inatherogenesis and its sequelae.

Adiponectin

Adiponectin (also called Acrp30 or adipoQ inmice) is a 244-amino-acid protein synthesizedand secreted exclusively by the adiposetissue.57,58 It is a close homologue of thecomplement protein C1q, which is involved inthe recognition of microbial surfaces.

In vitro, adiponectin inhibited monocyteadhesion to endothelial cells, decreased lipidaccumulation in human monocyte-derivedmacrophages and diminished scavenger recep-tor expression in these cells.59,60 In culturedhuman endothelial cells, adiponectin down-regulated expression of intracellular adhesionmolecules.60 It shows anti-inflammatory proper-ties as suggested by the suppressive effect ofadiponectin on phagocytic activity and lipo-polysaccharide-induced TNF-α production incultured macrophages.61 Adiponectin has beenshown to inhibit TNF-α induction of nuclearfactor κB through activation of the cAMP–protein kinase A pathway.61 Adiponectin alsoseems to stimulate the production of nitricoxide in vascular endothelial cells in in vitrostudies.62

Decreased adiponectin was originallydescribed in patients with coronary arterydisease.63,64 The incidence of cardiovasculardeath was found to be higher in renal failurepatients with low plasma adiponectincompared with those with higher plasmaadiponectin levels.65 Adiponectin has alsobeen found to be associated with vascularfunction in addition to its well-known positiveeffects on insulin action. Shimabukuro et alevaluated forearm blood flow (FBF) usingplethysmography in 76 Japanese subjectswithout a history of cardiovascular disease ordiabetes mellitus.66 They found positive associ-ations between FBF and adiponectin. Ouchi etal also reported a significant and positiveassociation between adiponectin and endothe-lium-dependent vasodilation among hyperten-sive patients.67

Increased nitroglycerin-induced vasodilationof forearm conduit vessels has also beenobserved in those apparently healthy subjectswith the highest circulating adiponectinconcentration.68 However, when establishedcardiovascular risk factors were present, suchas impaired fasting glucose, glucose intoler-ance or type 2 diabetes, no significant associa-tions between endothelial or vasculardysfunction and adiponectin were found. Itcould be that, once hypoadiponectinaemiadevelops, homeostatic mechanisms are lost.68


ch16 14/7/05 4:48 pm Page 181

Adiponectin null-mice formed 2-fold moreneointima in response to external vascular cuffinjury than wild-type mice.69 In fact, in wild-type mice, adiponectin infiltrated rapidly intothe subendothelial space of the vascular wallwhen the endothelial barrier of the arterialwall was injured by balloon angioplasty.70

Adenovirus-mediated supplement of adipo-nectin improved the intimal thickening inadiponectin null-mice to the wild-type level.71

In these studies, the protective effect of adipo-nectin seemed to be a direct consequence ofadiponectin action on the vascular wall and/ormacrophages rather than an indirect conse-quence of alteration of conventional athero-sclerotic risk factors in vivo.69,71

EFFERENT ARM OF INNATEIMMUNITY AND ATHEROSCLEROSIS

From a historical point of view, the cellulareffector arm was the first to be characterizedand the best understood in this context.

Cellular response (Figure 3)

It is not enough for the host to sense microbes.It must kill microbes as well. In vertebrates,innate immunity is largely dependent uponmyeloid cells: professional immunocytes thatengulf and destroy pathogens.1 Myeloid cellsinclude mononuclear phagocytes and polymor-phonuclear phagocytes. The mononuclearphagocytes are the macrophages, derived fromblood monocytes. A higher peripheral whiteblood cell count has been associated withinsulin resistance and with atherosclerosis.72

Peripheral white blood cell count correlatedsignificantly with insulin-mediated glucosedisposal during a euglycaemic clamp.72 In subse-quent studies, it was demonstrated thatneutrophil and lymphocyte count correlatedpositively with several components of the insulinresistance syndrome, and that plasma insulinconcentration was specifically associated withthe number of lymphocytes and monocytes.73

Macrophages are distributed throughout thebody of the host, in some cases (e.g. heart,


Lipases,glycosidases,proteases

Cell adhesion molecules H2O2, hydroxyl radical,oxygen hialides,singlet oxygen

Nitric oxide,peroxynitrite

Macrophage

Figure 3 Efferent armof innate immunity andatherosclerosis. Cellularcomponentssimultaneously involvedin innate immuneaction and theatherosclerotic process

ch16 14/7/05 4:48 pm Page 182

brain, lung and liver) lying within the paren-chyma of major organs. Macrophages are not auniform population of cells; rather they aremorphologically diverse, encompassing thespindle-shaped tissue histiocyte, the flattenedKüpffer cell of the hepatic sinusoids and thestellate microglial cell of the central nervoussystem.1

Should an infectious inoculum be intro-duced by any route, a macrophage will rarelybe far away from the invasive organism. Thisfunction seems directly involved in develop-ment of atherosclerosis. The accumulation ofcholesterol-rich lipoproteins in the arterial wallresults in recruitment of blood monocytes andtheir differentiation into lipid-laden foam cells,which drive the disease process of atheroscle-rosis. The most convincing evidence that ather-osclerosis is an inflammatory process, and notmerely a process of depositions of lipids in thearterial wall, is the continous presence andaccumulation of monocyte-derived macro-phages and T lymphocytes in fatty streaks andadvanced atherosclerotic lesions.

Monocyte migration is integral to the devel-opment of atherosclerosis. Early in the processof atherosclerosis, circulating monocytesadhere to the endothelial layer of the vesselwall, migrate into the vascular interstitium andphagocytize oxidized low density lipoproteincholesterol. This process results in the forma-tion of lipid-laden foam cells, which accumu-late within the arterial wall to form fatty streaks.Ultimately, these early lesions evolve intoadvanced atherosclerotic plaques that containnecrotic lipid cores surrounded by a proteogly-can matrix and covered by a fibrous cap andthickened intima. This structure defines anorganized atherosclerotic plaque.74

According to a recent review, a number ofapproaches have been used to cripple macro-phage activity in genetically prone mousemodels of atherosclerosis, all of which attenu-ated the atherosclerotic process.74 These includemouse models deficient for expression of:

• Macrophage chemoattractive protein-1(MCP-1), which stimulates macrophagemovement into the vessel wall;

• Chemokine receptor-2, a macrophagereceptor that binds MCP-1;

• Macrophage colony stimulating factor,which enhances conversion of monocytesto macrophages;

• Macrophage osteopontin, which mayprevent macrophage apoptosis, similar toits effects on endothelial cell survival.

These observations underscore the prominentrole of the macrophage in the pathogenesis ofatherosclerosis.

The cellular response results in the produc-tion of lipases, glycosidases, proteases, anti-microbial peptides, cell adhesion molecules,H2O2, hydroxyl radicals, oxygen hialides, singletoxygen, nitric oxide (NO), peroxynitrite andothers. These substances are simultaneouslyinvolved in the fight against infection and inthe atherosclerotic process. For instance, theNO system appears to play a major role.75 NO isan important messenger molecule that plays acritical role in a wide variety of physiologicalfunctions, including immune modulation,vascular relaxation, neuronal transmission andcytotoxicity. There are at least three isotypes ofNO synthase (NOS): endothelial cell NOS(eNOS), the neuronal type NOS (nNOS) andthe so-called inducible NOS (iNOS). iNOS isimplicated in host defence and is synthesizedde novo in response to a variety of inflammatorystimuli. Studies in humans indicate that NOproduction is decreased during hypertension.76

A polymorphism within the promoter of theiNOS candidate gene, NOS2A, revealed bothincreased allele sharing among sibpairs andpositive association of NOS2A to essentialhypertension.77 These facts are linked toanother major cardiovascular risk factor,insulin resistance, because insulin stimulationof glucose uptake in skeletal muscles andadipose tissues in vivo seem to be NO depen-dent.78 Moreover, iNOS has recently beenshown to be crucial for the development ofinsulin resistance.79 NO also antagonizes theeffects of angiotensin II on vascular tone andgrowth and also downregulates the synthesis ofangiotensin-converting enzyme and angio-tensin II type 1 (AT-1) receptors.80 Angiotensin


ch16 14/7/05 4:48 pm Page 183

II is important in stimulating the production ofreactive oxygen species and the activation ofancient inflammatory mechanisms through itsAT1 receptor.81

Humoral response (Figure 4)

The humoral response of the innate immunesystem results in the production of differentcytokines (the major cytokines are TNF-α andIL-6), acute phase reactants, lysozyme, lactofer-rin and myeloperoxidase, among others, aimedat tissue repair.

Tumour necrosis factor-alpha andinterleukin-6

The available information on the effects ofTNF-α and IL-6 in experimental models andthe transversal and prospective observations inhumans suggest that they are involved in the

pathophysiology of hypertension, abdominalobesity, dyslipidaemia and disorders of glucosemetabolism.82

The liver is the target of these systemicinflammatory mediators. Among the mostimportant aspects of this response is the repri-oritization of hepatic protein synthesis with theincreased production of a number of plasmaproteins (positive acute-phase proteins) andreduced production of a number of normalexport proteins (negative acute-phase pro-teins). Although the concentrations of multi-ple components of the acute-phase responseincrease together, not all of them increaseuniformly in all patients. These variationsindicate that the components of the acute-phase response are individually regulated, andthis may be caused in part by differences in thepattern of production of specific cytokines.These facts would explain increased suscepti-bility to increased inflammatory activity amonghealthy volunteers with genetically increased


TNF-α Interleukin-6

Acute-phase proteins

Myeloperoxidase

Macrophage

Figure 4 Efferent arm of innate immunity and atherosclerosis. Humoral components simultaneously involved ininnate immune action and the atherosclerotic process

ch16 14/7/05 4:48 pm Page 184

rates of some cytokines (reviewed in reference82).

Given that abnormalities in immune systemfunction and inflammatory mediators havebeen found to be associated with several classi-cal cardiovascular risk factors such as hyperten-sion, dyslipidaemia, obesity, insulin resistanceand others like endothelial dysfunction andclotting activation, cardiovascular diseaseseems to be the endpoint, common tometabolic and inflammatory pathways.

Human atherosclerotic lesions have beenfound to contain TNF-α mRNA. The accumu-lation of cholesteryl esters in macrophagesexposed to LDL is associated with increasedsynthesis and release of TNF-α. The associa-tions of TNF-α gene polymorphism and MIhave been investigated in several studies.83–87

The TNFA2 allele, linked to increasedtranscription rate of the cytokine, was associ-ated with parental history of MI.83 The oddsratio for myocardial infarction tended to behigher (albeit not significantly) in TNF2homozygotes in Brazil.84 These authors alsodescribed a tendency toward increased risk ofMI conferred by obesity. In another study, theTNF2 allele was associated with increasedplasma homocysteine levels, a known potentia-tor of lipid-related oxidation.85 TNF2 homozy-gotes (n = 10) also tended to have more fibrouslesions and calcification in their coronary arter-ies in an autopsy series.86 An associationbetween –308 TNF-α gene polymorphism andischaemic heart disease has been recentlydescribed, mainly attributed to women withtype 2 diabetes.87

IL-6 has also been speculated to play a keyrole in the development of coronary arterydisease through a number of metabolic,endothelial and procoagulant mechanisms(reviewed in references 88 and 89). Damage tothe vessel wall causes endothelial cell disrup-tion, resulting in exposure of the underlyingvascular smooth muscle cells. Endothelial andsmooth muscle cells produce IL-6 and IL-6gene transcripts are expressed in human ather-osclerotic lesions. Prospective studies of appar-ently healthy and high-risk individuals indicatethat increased IL-6 levels predict cardiovascular

events.82 IL-6 has been demonstrated to be astrong independent marker of increasedmortality in unstable CAD and identifiespatients who benefit most from a strategy ofearly invasive management.

Myeloperoxidase

Myeloperoxidase (MPO), an enzyme princi-pally associated with host defence mechanisms,has also been associated with CRP levels andcardiovascular risk.90 MPO seems to modulatevascular signalling and vasodilatory function ofnitric oxide, linking fight against infection withmetabolic events.91 Importantly, circulatingmyeloperoxidase levels, in contrast to troponinT, creatine kinase MB isoform and CRP levels,identified patients at risk for cardiac events inthe absence of myocardial necrosis, highlight-ing its potential usefulness for risk stratificationamong patients who present with chest pain.92

CONCLUSIONS

The ability to mount a prominent inflamma-tory response to pathogens confers an advan-tage in innate immune defence. Several factorsseem to be implicated in the recognition ofmicrobial surfaces, external and endogenousligands and in their elimination. Thesepathways are simultaneously involved in ather-osclerosis and in metabolic pathways.82 Thus,different metabolic pathways might haveevolved in parallel with several mechanismsinvolved in our fight against infection.

All the associations described between innateimmune components and atherosclerosis mightbe interpretated as a body’s response to chronictissue injury (the response-to-injury hypothe-sis), cardiovascular disease being a byproduct ofthe inflammatory cascade triggered by physical,environmental and infectious agents. Althoughit seems unlikely that one specific agent causesatherosclerosis, the infectious burden andenvironmental exposure related to bacterialproducts (smoking, LPS) are increasinglyclaimed to be involved in the triggering anddevelopment of atherosclerosis.93


ch16 14/7/05 4:48 pm Page 185

Research on factors involved in innateimmunity and the inflammatory cascade willprobably help to characterize individuals

prone to cardiovascular disease, and to developnew therapeutic targets aimed at preventingthis important cause of death.


1. Beutler B. Innate immunity: an overview. MolImmunol 2004; 40:845–59

2. Janeway CA Jr, Medzhitov R. Innate immune recogni-tion. Ann Rev Immunol 2002; 20:197–216

3. Xu XH, Shah PK, Faure E, et al. Toll-like receptor-4 isexpressed by macrophages in murine and humanlipid-rich atherosclerotic plaques and upregulated byoxidized LDL. Circulation 2001;104:3103–8

4. Edfeldt K, Swedenborg J, Hansson GK, et al.Expression of toll-like receptors in human atheroscle-rotic lesions: a possible pathway for plaque activation.Circulation 2002; 105:1158–61

5. Castrillo A, Joseph SB, Vaidya SA, et al. Crosstalkbetween LXR and Toll-like receptor signalingmediates bacterial and viral antagonism of cholesterolmetabolism. Mol Cell 2003; 12:805–16

6. Joseph SB, Castrillo A, Laffitte BA, et al. Reciprocalregulation of inflammation and lipid metabolism byliver X receptors. Nat Med 2003; 9:213–19

7. Agnese DM, Calvano JE, Hahm SJ, et al. Human toll-like receptor 4 mutations but not CD14 polymor-phisms are associated with an increased risk ofgram-negative infections. J Infect Dis 2002;186:1522–5

8. Kiechl S, Lorenz E, Reindl M, et al. Toll-like receptor4 polymorphisms and atherogenesis. N Engl J Med2002; 347:185–92

9. Arbour NC, Lorenz E, Schutte BC, et al. TLR4mutations are associated with endotoxin hyporespon-siveness in humans. Nat Genet 2000; 25:187–91

10. Boekholdt SM, Agema WR, Peters RJ, et al. Variants oftoll-like receptor 4 modify the efficacy of statintherapy and the risk of cardiovascular events.Circulation 2003; 107:2416–21

11. Illig T, Bongardt F, Schopfer A, et al. The endotoxinreceptor TLR4 polymorphism is not associated withdiabetes or components of the metabolic syndrome.Diabetes 2003; 52:2861–4

12. Ohashi K, Burkart V, Flohe S, et al. Cutting edge: heatshock protein 60 is a putative endogenous ligand ofthe toll-like receptor-4 complex. J Immunol 2000;164:558–61

13. Hollestelle SC, De Vries MR, Van Keulen JK, et al.Toll-like receptor 4 is involved in outward arterialremodeling. Circulation 2004; 109:393–8

14. Oyama J, Blais C Jr, Liu X, et al. Reduced myocardialischemia-reperfusion injury in toll-like receptor 4-deficient mice. Circulation 2004;109:784–9

15. Asea A, Kraeft SK, Kurt-Jones EA, et al. HSP70 stimu-lates cytokine production through a CD14-dependentpathway, demonstrating its dual role as a chaperoneand cytokine. Nat Med 2000; 6:435–42

16. Unkelbach K, Gardemann A, Kostrzewa M, et al. Anew promoter polymorphism in the gene oflipopolysaccharide receptor CD14 is associated withexpired myocardial infarction in patients with lowatherosclerotic risk profile. Arterioscler Thromb VascBiol 1999; 19:932–8

17. Hubacek JA, Rothe G, Pit’ha J, et al. C (-260)- Tpolymorphism in the promoter of the CD14monocyte receptor gene as a risk factor for myocar-dial infarction. Circulation 1999; 99:3218–20

18. Ito D, Murata M, Tanahashi N, et al. Polymorphism inthe promoter of lipopolysaccharide receptor CD14and ischemic cerebrovascular disease. Stroke 2000;31:2661–4

19. Shimada K, Watanabe Y, Mokuno H, et al. Commonpolymorphism in the promoter of the CD14monocyte receptor gene is associated with acutemyocardial infarction in Japanese men. Am J Cardiol2000; 86:682–4

20. Risley P, Jerrard-Dunne P, Sitzer M, et al. CarotidAtherosclerosis Progression Study. Promoterpolymorphism in the endotoxin receptor (CD14) isassociated with increased carotid atherosclerosis onlyin smokers: the Carotid Atherosclerosis ProgressionStudy (CAPS). Stroke 2003; 34:600–4

21. Hasday JD, Bascom R, Costa JJ, et al. Bacterialendotoxin is an active component of cigarette smoke.Chest 1999; 115:829–35

22. Wiedermann CJ, Kiechl S, Dunzendorfer S, et al.Association of endotoxemia with carotid atheroscle-rosis and cardiovascular disease: prospective resultsfrom the Bruneck study. J Am Coll Cardiol 1999;34:1975–81

23. Wiedermann CJ, Kiechl S, Schratzberger P, et al. Therole of immune activation in endotoxin-inducedatherogenesis. J Endotoxin Res 2001; 7:322–6

24. Osmand AP, Friedenson B, Gewurz H, et al.Characterisation of C-reactive protein and thecomplement subcomponent Clt as homologousproteins displaying cyclic pentameric symmetry(pentraxins). Proc Natl Acad Sci USA 1977;74:739–43

25. Robey FA, Liu T-Y. Limulin: a C-reactive protein fromLimulus polyphemus. J Biol Chem 1981; 256:969–75

References

ch16 14/7/05 4:48 pm Page 186

26. Ablij H, Meinders A. C-reactive protein: history andrevival. Eur J Intern Med 2002; 13:412–22

27. Pepys MB, Hirschfield GM. C-reactive protein: a criti-cal update. J Clin Invest 2003; 111:1805–12

28. Lagrand WK, Niessen HW, Wolbink GJ, et al. C-reactive protein colocalizes with complement inhuman hearts during acute myocardial infarction.Circulation 1997; 95:97–103

29. Ridker PM, Buring JE, Shih J, et al. Prospective studyof C-reactive protein and the risk of future cardiovas-cular events among apparently healthy women.Circulation 1998; 98:731–3

30. Kuller LH, Tracy RP, Shaten J, et al for the MRFITResearch Group. Relationship of C-reactive proteinand coronary heart disease in the MRFIT nested case-control study. Am J Epidemiol 1996; 144:537–47

31. Harris TB, Ferrucci L, Tracy RP, et al. Associations ofelevated interleukin-6 and C-reactive protein levelswith mortality in the elderly. Am J Med 1999;106:506–12

32. Ridker PM, Rifai N, Stampfer MJ, et al. Plasmaconcentration of interleukin-6 and the risk of futuremyocardial infarction among apparently healthymen. Circulation 2000; 101:1767–72

33. Rolph MS, Zimmer S, Bottazzi B, et al. Production ofthe long pentraxin PTX3 in advanced atheroscleroticplaques. Arterioscler Thromb Vasc Biol 2002; 22:e10-4

34. Peri G, Introna M, Corradi D, et al. PTX3, a proto-typical long pentraxin, is an early indicator of acutemyocardial infarction in humans. Circulation 2000;102:636–41

35. Hiki N, Berger D, Dentener MA, et al. Changes inendotoxin-binding proteins during major electivesurgery: important role for soluble CD14 in regula-tion of biological activity of systemic endotoxin. ClinDiagn Lab Immunol 1999; 6:844–50

36. Haziot A, Rong GW, Lin XY, et al. Recombinantsoluble CD14 prevents mortality in mice treated withendotoxin (lipopolysaccharide). J Immunol 1995;154:6529–32

37. Haziot A, Rong GW, Bazil V, et al. Recombinantsoluble CD14 inhibits LPS-induced tumor necrosisfactor-alpha production by cells in whole blood. JImmunol 1994; 152:5868–76

38. Wurfel MM, Hailman E, Wright SD. Soluble CD14 actsas a shuttle in the neutralization of lipopolysaccharide(LPS) by LPS-binding protein and reconstituted highdensity lipoprotein. J Exp Med 1995; 181:1743–54

39. Kitchens RL, Wolfbauer G, Albers JJ, et al. Plasmalipoproteins promote the release of bacteriallipopolysaccharide from the monocyte cell surface. JBiol Chem 1999; 274:34116–22

40. Fernandez-Real JM, Broch M, Richart C, et al. CD14monocyte receptor, involved in the inflammatorycascade, and insulin sensitivity. J Clin EndocrinolMetab 2003; 88:1780–4

41. Harris HW, Grunfeld C, Feingold KR, et al. Humanvery low density lipoproteins and chylomicrons can

protect against endotoxin-induced death in mice. JClin Invest 1990; 86:696–702

42. Eggesbo JB, Lyberg T, Aspelin T, et al. Differentbinding of 125I-LPS to plasma proteins from personswith high or low HDL. Scand J Clin Lab Invest 1996;56:533–43

43. Michel OJ, Kips J, Duchateau J, et al. Severity ofasthma is related to endotoxin in house dust. Am JRespir Crit Care Med 1996; 154:1641–6

44. Jacob AI, Goldberg PK, Bloom N, et al. Endotoxinand bacteria in portal blood. Gastroenterology 1977;72:1268–70

45. Amar J, Ruidavets JB, Bal Dit Sollier C, et al. SolubleCD14 and aortic stiffness in a population-based study.J Hypertens 2003; 21:1869–77

46. Madsen HO, Garred P, Thiel S, et al. Interplaybetween promoter and structural gene variantscontrol basal serum level of mannan-binding protein.J Immunol 1995; 155:3013–20

47. Summerfield JA, Sumiya M, Levin M, et al.Association of mutations in mannose binding proteingene with childhood infection in consecutive hospitalseries. BMJ 1997; 314:1229–32

48. Koch A, Melbye M, Sorensen P, et al. Acute respira-tory tract infections and mannose-binding lectininsufficiency during early childhood. JAMA 2001;285:1316–21

49. Best LG, Davidson M, North KE, et al. Prospectiveanalysis of mannose-binding lectin genotypes andcoronary artery disease in American Indians: theStrong Heart Study. Circulation 2004; 109:471–5

50. Rugonfalvi-Kiss S, Endresz V, Madsen HO, et al.Association of Chlamydia pneumoniae with coronaryartery disease and its progression is dependent on themodifying effect of mannose-binding lectin.Circulation 2002; 106:1071–6

51. Madsen HO, Videm V, Svejgaard A, et al. Associationof mannose-binding-lectin deficiency with severeatherosclerosis. Lancet 1998; 352:959–60

52. Hegele RA, Ban MR, Anderson CM, et al. Infection-susceptibility alleles of mannose-binding lectin areassociated with increased carotid plaque area. JInvestig Med 2000; 48:198–202

53. Albert MA, Rifai N, Ridker PM. Plasma levels ofcystatin-C and mannose binding protein are notassociated with risk of developing systemic atheroscle-rosis. Vasc Med 2001; 6:145–9

54. Joiner KA. Complement evasion by bacteria andparasites. Annu Rev Microbiol 1988; 42:201–30

55. Seifert PS, Kazatchkine MD. The complement systemin atherosclerosis. Atherosclerosis 1988; 73:91–104

56. Saito E, Fujioka T, Kanno H, et al. Complementreceptors in atherosclerotic lesions. Artery 1992;19:47–62

57. Maeda K, Okubo K, Shimomura I, et al. cDNAcloning and expression of a novel adipose specificcollagen-like factor, apM1 (AdiPose Most abundantGene transcript 1). Biochem Biophys Res Commun1996; 221:286–9


ch16 14/7/05 4:48 pm Page 187

58. Scherer PE, Williams S, Fogliano M, et al. A novelserum protein similar to C1q, produced exclusively inadipocytes. J Biol Chem 1995; 270:26746–9

59. Ouchi N, Kihara S, Arita Y, et al. Adipocyte-derivedplasma protein, adiponectin, suppresses lipidaccumulation and class A scavenger receptor expres-sion in human monocyte-derived macrophages.Circulation 2001; 103:1057–63

60. Yokota T, Oritani K, Takahashi I, et al. Adiponectin, anew member of the family of soluble defense colla-gens, negatively regulates the growth of myelomono-cytic progenitors and the functions of macrophages.Blood 2000; 96:1723–32

61. Ouchi N, Kihara S, Arita Y, et al. Adiponectin, anadipocyte-derived plasma protein, inhibits endothe-lial NF-kappaB signaling through a cAMP-dependentpathway. Circulation 2000; 102:1296–301

62. Chen H, Montagnani M, Funahashi T, et al.Adiponectin stimulates production of nitric oxide invascular endothelial cells. J Biol Chem 2003;278:45021–6

63. Hotta K, Funahashi T, Arita Y, et al. Plasma concen-trations of a novel, adipose-specific protein,adiponectin, in type 2 diabetic patients. ArteriosclerThromb Vasc Biol 2000; 20:1595–9

64. Ouchi N, Kihara S, Arita Y, et al. Novel modulator forendothelial adhesion molecules: adipocyte-derivedplasma protein adiponectin. Circulation 1999;100:2473–6

65. Zoccali C, Mallamaci F, Tripepi G, et al. Adiponectin,metabolic risk factors, and cardiovascular eventsamong patients with end-stage renal disease. J Am SocNephrol 2002; 13:134–41

66. Shimabukuro M, Higa N, Asahi T, et al. Hypo-adiponectinemia is closely linked to endothelialdysfunction in man. J Clin Endocrinol Metab 2003;88:3236–40

67. Ouchi N, Ohishi M, Kihara S, et al. Association ofhypoadiponectinemia with impaired vasoreactivity.Hypertension 2003; 42:231–4

68. Fernandez-Real JM, Castro A, Vazquez G, et al.Adiponectin is associated with vascular functionindependent of insulin sensitivity. Diabetes Care2004; 27:739–45

69. Kubota N, Terauchi Y, Yamauchi T, et al. Disruptionof adiponectin causes insulin resistance and neointi-mal formation. J Biol Chem 2002; 277:25863–6

70. Okamoto Y, Arita Y, Nishida M, et al. An adipocyte-derived plasma protein, adiponectin, adheres toinjured vascular walls. Horm Metab Res 2000; 32:47–50

71. Matsuda M, Shimomura I, Sata M, et al. Role ofadiponectin in preventing vascular stenosis. Themissing link of adipo-vascular axis. J Biol Chem 2002;277:37487–91

72. Facchini F, Hollenbeck CB, Chen YN, et al.Demonstration of a relationship between white bloodcell count, insulin resistance, and several risk factorsfor coronary heart disease in women. J Intern Med1992; 232:267–72

73. Targher G, Seidell JC, Tonoli M, et al. The whiteblood cell count: its relationship to plasma insulinand other cardiovascular risk factors in healthy maleindividuals. J Intern Med 1996; 239:435–41

74. Hansson GK, Peter Libby P, Schönbeck U, et al.Innate and adaptive immunity in the pathogenesis ofatherosclerosis. Circ Res 2002; 91:281–91

75. Manning RD Jr, Hu L, Tan DY, et al. Role of abnor-mal nitric oxide systems in salt-sensitive hypertension.Am J Hypertens 2001; 14(6 Pt 2):68S–73S

76. Leclercq B, Jaimes EA, Raij L. Nitric oxide synthaseand hypertension. Curr Opin Nephrol Hypertens2002; 11:185–9

77. Rutherford S, Johnson MP, Curtain RP, et al.Chromosome 17 and the inducible nitric oxidesynthase gene in human essential hypertension. HumGenet 2001; 109:408–15

78. Roy D, Perreault M, Marette A. Insulin stimulation ofglucose uptake in skeletal muscles and adipose tissuesin vivo is NO dependent. Am J Physiol 1998; 274(4 Pt1):E692–9

79. Perreault M, Marette A. Targeted disruption ofinducible nitric oxide synthase protects againstobesity-linked insulin resistance in muscle. Nat Med2001; 7:1138–43

80. Bataineh A, Raij L. Angiotensin II, nitric oxide, andend-organ damage in hypertension. Kidney Int(Suppl) 1998; 68:S14–19

81. Luft FC. Angiotensin, inflammation, hypertension,and cardiovascular disease. Curr Hypertens Rep 2001;3:61–7

82. Fernández-Real JM, Ricart W. Insulin resistance andchronic cardiovascular inflammatory syndrome.Endocr Rev 2003; 24:278–301

83. Herrmann SM, Ricard S, Nicaud V, et al.Polymorphisms of the tumour necrosis factor-α gene,coronary heart disease and obesity. Eur J Clin Invest1998; 28:59–66

84. Padovani JC, Pazin-Filho A, Simoes MV, et al. Genepolymorphisms in the TNF locus and the risk ofmyocardial infarction. Thromb Res 2000; 100:263–9

85. Wang XL, Oosterhof J. Tumour necrosis factor αG–308→ A polymorphism and risk for coronary arterydisease. Clin Sci 2000; 98:435–7

86. Keso T, Perola M, Laippala P, et al. Polymorphismswithin the tumor necrosis factor locus and prevalenceof coronary artery disease in middle-aged men.Atherosclerosis 2001; 154:691–7

87. Vendrell J, Fernandez-Real JM, Gutierrez C, et al. Apolymorphism in the promoter of the tumor necrosisfactor-alpha gene (–308) is associated with coronaryheart disease in type 2 diabetic patients. Athero-sclerosis 2003; 167:257–64

88. Yudkin JS, Kumari M, Humphries SE, et al.Inflammation, obesity, stress and coronary heartdisease: is interleukin-6 the link? Atherosclerosis2000; 148:209–14

89. Woods A, Brull DJ, Humprhies SE, et al. Genetics ofinflammation and risk of coronary artery disease: the


ch16 14/7/05 4:48 pm Page 188

central role of interleukin 6. Eur Heart J 2000;21:1574–83

90. Zhang R, Brennan ML, Fu X, et al. Associationbetween myeloperoxidase levels and risk of coronaryartery disease. JAMA 2001; 286:2136–42

91. Eiserich JP, Baldus S, Brennan ML, et al.Myeloperoxidase, a leukocyte-derived vascular NOoxidase. Science 2002; 296:2391–4

92. Brennan ML, Penn MS, Van Lente F, et al. Prognosticvalue of myeloperoxidase in patients with chest pain.N Engl J Med 2003; 349:1595–604

93. Espinola-Klein C, Rupprecht HJ, Blankenberg S, et al.Impact of infectious burden on extent and long-termprognosis of atherosclerosis. Circulation 2002;105:15–21


ch16 14/7/05 4:48 pm Page 189

ch16 14/7/05 4:48 pm Page 190

INTRODUCTION

In spite of the significant progress being madein understanding the cellular and molecularbasis of atherosclerosis, cardiovascular diseaseremains the leading cause of morbidity andmortality in the United States. Multiple riskfactors for development of atherosclerosis havebeen identified, including elevated levels ofLDL-cholesterol and reduced levels of HDL. Inaddition to the contribution of lipoproteinmetabolism, the inflammatory response in thevessel wall has proven to be a major componentof the development and progression of athero-sclerotic lesions, as well as an important factorin the risk of acute thrombotic events.

Atherosclerotic lesions develop at anatomi-cal sites characterized by variations in haemo-dynamic and mechanical forces that inducethe local expression of chemokines, cytokinesand endothelial cell adhesion molecules.1 Theinfiltrating inflammatory cells secrete solublemediators that perpetuate the inflammatoryresponse underlying the progression of athero-sclerotic lesions. Vascular cells also play a criti-cal part in disease progression. Activatedendothelial cells (ECs), in addition to mediat-ing inflammatory cell recruitment, regulate thefunction of vascular smooth muscle cells(VSMCs). In normal vessels, VSMCs residepredominantly in the media in a quiescent,contractile state, maintained in part byproducts of ECs such as prostacyclin, nitricoxide and TGF-β. Atherosclerotic lesions arecharacterized by dedifferentiation of VSMCs to

a synthetic state characterized by proliferationand migration into the neointima.2 Inaddition, VSMC dedifferentiation is associatedwith downregulation of contractile proteinssuch as α-smooth muscle actin, and upregula-tion of adhesion molecules,3 as well as matrixcomponents and inflammatory mediators.2 Inaddition to mediating leukocyte recruitment,cell adhesion molecules (CAMs) mediatesignal transduction and thereby regulate thefunction of leukocytes as well as vascular cells,thus providing potential targets for novel thera-peutics designed to selectively regulate theinflammatory response in atherosclerosis.

CD44 is a widely expressed CAM that servesas a principal receptor for the extracellularmatrix glycosaminoglycan hyaluronan (HA).CD44 can mediate atherogenic processesincluding inflammatory cell recruitment andcellular activation.4–6 Importantly, the pro-inflammatory low molecular weight forms ofHA (LMW-HA) accumulate in atheroscleroticlesions.7–9 This evidence led to investigation ofthe potential for CD44 to modulate atherogen-esis. Using a genetic approach it was demon-strated that CD44-null mice had markedlyreduced atherosclerosis compared with CD44heterozygote and wild-type littermates.10

Furthermore, it was shown that CD44 promotesatherosclerosis by mediating inflammatory cellrecruitment and leukocyte and vascular cellactivation.10 In addition, CD44-deficient micewere found to be protected against ischaemicbrain injury.11 Thus, CD44 appears to be animportant factor, and potential therapeutic

Role of CD44 in atherogenesis and itspotential role as a therapeutic targetE. Puré

17

ch17 14/7/05 4:48 pm Page 191

target, in the two most devastating outcomes ofcardiovascular disease, myocardial infarctionand stroke. In this chapter, the mechanisms bywhich CD44 can contribute to atherogenesis atmultiple levels and their potential as targets fortherapeutic intervention in cardiovasculardisease are discussed.

STRUCTURE, LIGANDS ANDFUNCTION OF CD44 (Figure 1)

CD44 is a family of type I transmembraneglycoprotein products of a single gene gener-ated by alternative RNA splicing. The gene forCD44 maps to chromosome 11 in humans and2 in mice.12 The CD44 gene consists of 20exons. The most abundant isoform of CD44

(referred to as CD44s for ‘standard’) isencoded by 12 non-variant exons that encodefor the extracellular domain, a transmembranedomain and an intracellular domain. The alter-natively spliced isoforms are generated byinclusion of various numbers and combina-tions of one to 10 or 11 (depending on thespecies) variant exons that are inserted at asingle site in the membrane-proximal region ofthe extracellular domain.13,14 The predomi-nant, and highly conserved, 72-amino-acidcytoplasmic domain can also be replaced by analternatively spliced exon encoding a trun-cated form of the receptor.

CD44 is also subject to myriad post-transla-tional modifications, including variations inthe N- and O-linked carbohydrate structuresand covalent modification with glycosamino-


Ser 323

Ser 325

Cytoplasm

Membraneproximalregion

N-glycosylation

O-glycosylation

Chondroitin sulphate

BX7B HA binding motif

223

SS

S S

S S

Variant exonsV2–v10

v2

v3

v4

v5

v6

v7

v8

v9

v10

GAG

GAG�

�P

P

Figure 1 Schematic representation of CD44

ch17 14/7/05 4:48 pm Page 192

glycans, in a cell-type-specific fashion that alsovaries with activation and differentiation insome cell lineages. These modificationscontribute to the large variation in the appar-ent molecular weight, pI and ligand bindingfunction of CD44.

The first ~100 residues of the amino termi-nus of the extracellular domain constitute astructural ‘link module’ that contains one ofthe three potential HA binding motifs,B(X7)B,15 and defines the receptor as amember of the link module superfamily.16 TheN-terminus of the extracellular domain,including six cysteines that participate inintramoleclar disulphide bridging, and thetransmembrane and intracellular domains arehighly conserved (80–90% among species),while the membrane proximal region of theextracellular domain is less well conserved(~50%).

A soluble form of CD44 is found in plasmaand other body fluids.17,18 Although found insignificant levels in normal plasma, immuneactivation and inflammation are often associ-ated with increased plasma levels of solubleCD44. These findings suggest that release ofCD44 correlates with enhanced local prote-olytic activity and matrix remodelling and havegenerated interest in CD44 as a potentialbiomarker for inflammation. Recent evidencesuggests that, in addition to the transmem-brane and soluble forms, CD44 also exists as anintegral component of extracellular matrixthat may be an important reservoir of thesoluble CD44 released into the fluid phase ininflammation.19,20 Furthermore, cleavage of theextracellular domain leads to sequential prote-olytic processing of the transmembranedomain and the release of the intracellulardomain (by a presenilin-dependent gammasecretase), that translocates to the nucleuswhere it may regulate gene transcription, andthe extracellular release of a so-called CD44β-peptide.21,22 The function of the latter isunknown, but is reminiscent of the presenilin-dependent processing of β-amyloid precursorprotein leading to the release of amyloid βpeptide(s) implicated in Alzheimer’s disease.Cleavage of CD44 may also be ligand induced

as cross-linking the receptor with anti-CD44antibodies leads to release of soluble CD44.23,24

Endogenous membrane metalloproteinases,including MT1-MMP and MT3-MMP, anddisintegrins and metalloproteases (ADAM)25

have been implicated in the shedding of CD44based on pharmacological evidence.24,26

Shedding of CD44 appears to be controlled atleast in part by Ras and Rho GTPases (Cdc42and Rac1), possibly through regulation of theactin cytoskeleton.24,27 Finally, an alternativelyspliced variant may generate a soluble form ofCD44 by de novo synthesis.28

An important aspect of CD44 is the regula-tion of its affinity for its principal ligand, HA.CD44 expressed on the vast majority of primarycells exhibits low affinity for HA while cellularactivation is associated with increased affinityof CD44 for HA, reportedly due to one or moremechanisms including increased expression,receptor oligomerization, expression of alter-natively spliced variants and post-translationalmodifications, including changes in glycosyla-tion, sulphation and phosphorylation.29–38

The principal ligand implicated in many ofthe defined functions of CD44 is HA.12,30 Oneimportant function of CD44 is in fact its role inthe assembly and turnover of HA-containingmatrices.39,40 However, it is important to notethat CD44 also binds other matrix compo-nents, including collagen types I and VI andfibrinogen and fibronectin, and osteopon-tin12,41–43 that may also be important in thefunction of CD44 in atherogenesis. In addition,it has recently been demonstrated that dockingof matrix metalloproteinases to CD44 may bean important mechanism for the local activa-tion of TGF-β1 and the regulation of inflam-mation/fibrosis.44

CD44 PROMOTESATHEROSCLEROSIS IN A MURINEMODEL (Figure 2)

Using a genetic approach it was demonstratedthat CD44-null mice had markedly reducedatherosclerosis compared with CD44 heterozy-gote and wild-type littermates.10

ROLE OF CD44 IN ATHEROGENESIS AND ITS POTENTIAL ROLE AS A THERAPEUTIC TARGET 193

ch17 14/7/05 4:48 pm Page 193

The extent of atherosclerotic lesions inapoE–/–.CD44+/+ and apoE–/–.CD44–/– mice wascompared. The extent of atherosclerosis wasmarkedly reduced in CD44–/–.apoE–/– andCD44+/–.apoE–/– mice compared toCD44+/+.apoE–/– mice maintained on a chowdiet without having any effect on plasmacholesterol levels. Further investigationindicated that CD44 can potentiate atherogen-esis through several mechanisms involving cellsof both the haematopoietic and non-haematopoietic compartments:

1. CD44 mediates the recruitment of mono-cytes/macrophages to atheroscleroticlesions.10

2. The pro-inflammatory low molecularweight form of the CD44 ligand HA(LMW-HA) that accumulates in athero-

sclerotic lesions can induce the produc-tion of inflammatory mediators known topromote atherogenesis including thecytokine interleukin-12 (IL-12).5 IL-12 iscritical for the generation of a TH1 typeinflammatory response, in part throughthe induction of interferon-gamma (IFN-γ) production. The IL-12/IFN-γ pathway isan important pro-atherogenic componentof the inflammatory response. In studyingthe signalling pathways that regulateCD44/HA-induced IL-12 production, a novel 12/15-lipoxygenase-dependentpathway that mediates the expression ofIL-12 in macrophages and in atheroscle-rotic lesions was discovered.45

3. CD44 is required for the phenotypic de-differentiation of VSMCs in atheroscleroticlesions to the ‘synthetic’ state.10


Figure 2 CD44 has the potential to regulate atherogenesis through multiple mechanisms

ch17 14/7/05 4:48 pm Page 194

4. The low molecular weight pro-inflamma-tory forms of HA that accumulate in ather-osclerotic lesions stimulate VCAM-1expression and proliferation of culturedprimary aortic VSMCs, whereas highmolecular weight forms of HA (HMW-HA)inhibit VSMC proliferation.

CD44 MEDIATES RECRUITMENT OFMONOCYTES/MACROPHAGES AND TCELLS (Figure 3)

CD44 is expressed on both inflammatory cellsand activated endothelial cells and canmediate the adhesion of inflammatory cells toECs.4,46 Using a genetic approach it wasdemonstrated that expression of CD44 onmonocytes/macrophages was critical for therecruitment of macrophages to atheroscleroticlesions.10 Thus, trafficking of monocytes/macrophages from apoE–/–.CD44–/– mice toatherosclerotic lesions when adoptively trans-ferred into apoE–/–.CD44+/+ mice was markedlyreduced (by 70–90%) when compared to the

trafficking of apoE–/–.CD44+/+ monocytes/macrophages to atherosclerotic lesions inapoE–/–, CD44+/+ mice. These results indicatethat expression of CD44 on monocytes/macrophages is required for their recruitmentto lesions. In addition, it has been reportedthat recruitment of activated T cells to sites ofinflammation is mediated by CD44. The neteffect of T cells also appears to be to promoteatherogenesis. Therefore it will be of consider-able interest to extend the studies ofmonocyte/macrophage recruitment to athero-sclerotic lesions to determine whether CD44-mediated interactions are also important forthe recruitment of T cells to atheroscleroticlesions. Activated ECs also upregulate CD44 aswell as their production of HA that binds tothe EC surface via CD44.46,47 In studies of T cellhoming, it was shown that, in addition to CD44expressed on T cells, the upregulation ofendothelial cell HA production and itsdocking to EC CD44 also play a role in therecruitment of T cells to sites of inflamma-tion.48 Therefore, it will also be of interest infuture studies to determine whether EC CD44


Figure 3 CD44 mediates leukocyte recruitment to atherosclerotic lesions. CD44 mediates the initial binding ofcirculating monocytes and activated T cells to activated endothelium. CD44 on leukocytes binds to HA produced byactivated endothelium and docked to the lumenal surface via CD44 on the endothelial surface. Subsequentengagement of leukocyte integrins such as VLA-4 by endothelial cell adhesion molecules such as VCAM-1 results infirm adhesion leading to transendothelial migration

ch17 14/7/05 4:48 pm Page 195

and HA mediate monocyte/macrophage andT cell recruitment to atherosclerotic lesions, ashas been described in the case of other inflam-matory lesions.

CD44 REGULATES THEPRODUCTION OF INFLAMMATORYMEDIATORS BY MACROPHAGES ANDT CELLS (Figure 4)

CD44 stimulates the production of inflamma-tory mediators by macrophages (Figure 4) suchas MCP-1, TNF and iNOS5,6,49 that have beenimplicated in the pathogenesis of atherosclero-sis. In addition, the engagement of CD44 byLMW-HA was found to be one of the mostpotent inducers of IL-12 production in

macrophages. Release of IL-12 leads to thegeneration of a TH1 type inflammatoryresponse, in part through the induction of IFN-γ. The pro-atherogenic effects of IL-12 andIFN-γ are now well established.50–53

A potentially very interesting recent findingis that CD44/HA, as well as lipopolysaccha-rides, can activate a novel 12/15-lipoxygenase-dependent pathway leading to IL-12 geneexpression.45 Importantly, the IL-12 produc-tion by macrophages in atherosclerotic lesionswas found to be predominantly dependent onthis novel 12/15-lipoxygenase-mediatedmechanism since the levels of IL-12 and IFN-γdetected in the atherosclerotic lesions ofapobec–/–.LDLR–/–.12/15-LO–/– mice weremarkedly reduced compared to the levelsdetected in apobec–/–.LDLR–/–.12/15-LO+/+


Figure 4 CD44 induces the interleukin-12/interferon-γ (IL-12/IFN-γ) pro-inflammatory axis in atheroscleroticlesions. LMW-HA binding to CD44 induces macrophage release of pro-inflammatory mediators, including IL-12 thatpromotes a TH1 type inflammatory response. An important function of IL-12 is to induce the production of IFN-γ byseveral cell types including TH1 T cells and natural killer cells, thus establishing a positive amplification loop by, inturn, inducing enhanced production of IL-12 by macrophages

ch17 14/7/05 4:48 pm Page 196

mice.45 In contrast, in the same 12/15-lipoxy-genase-deficient mice, the IL-12 and IFN-γresponse to infection with Toxoplasma gondiiwas comparable to that seen in 12/15-LO+/+

littermate controls. These results indicate thatthe signalling pathways involved in the produc-tion of IL-12 may vary in a cell type and/orstimulus-dependent manner. If this indeedproves to be the case it will open up the possi-bility of pharmacologically inhibiting the pro-atherogenic IL-12/IFN-γ axis in atheroscleroticlesions without necessarily disrupting theinduction of TH1 responses required, forexample, to resolve a number of infections.Further understanding of the signallingpathways and transcriptional regulation of IL-12 may therefore provide novel targets for therational design of therapeutic interventions tomodulate the inflammatory response in ather-osclerotic lesions.

CD44 HAS THE POTENTIAL TOREGULATE VASCULARREMODELLING THROUGHMULTIPLE MECHANISMS (Figure 5)

Development of atherosclerotic lesions is associ-ated with vascular remodelling. The prolifer-ation, de-differentiation, activation andmigration of VSMC and matrix reorganizationare all important aspects of vascular remodel-ling. The role of CD44 in leukocyte homing toatherosclerotic lesions and leukocyte activationnotwithstanding, data indicate that CD44 alsoplays a critical role in vascular homeostasis andin regulating the function of VSMCs in athero-sclerotic lesions. A characteristic feature ofatherosclerotic lesions is transition of VSMCsfrom a contractile to a synthetic phenotype. Thesynthetic phenotype is characterized by changesin gene expression (increased expression of


Figure 5 CD44 may play a role in vascular remodelling by regulating vascular smooth muscle cell de-differentiation,proliferation and migration and matrix assembly and organization

ch17 14/7/05 4:48 pm Page 197

CAMs and decreased α-smooth muscle actin),3as well as proliferation and migration of medialVSMCs into the neointima.2,54,55 It appears thatCD44 may play an important role in regulatingthe phenotypic de-differentiation, growth andmigration of VSMC. Together with the evidencethat both cell-surface and matrix-associatedCD44 play important roles in the assembly andorganization of extracellular matrix, these datasuggest that CD44 may be an important factor invascular remodelling.

CD44/HA regulates vascular smoothmuscle cell growth and de-differentiation

Recent evidence suggests that CD44 mediatessignals that maintain quiescence of VSMC innormal vessel, but mediates pro-atherogenicsignals in atherosclerotic lesions in hypercho-lesterolaemic mice.10 The suggestion that CD44plays a role in homeostasis is in part based ondata that the high molecular weight form of HA(HMW-HA) that predominates in extracellularmatrix under homeostatic conditions inhibitsVSMC growth. In contrast, interaction of CD44with the intermediate to low molecular weightforms of HA (LMW-HA) that accumulate inatherosclerotic lesions stimulates the growthand phenotypic de-differentiation of VSMC.CD44 was shown to promote human VSMCproliferation.7 Furthermore, the expression ofhigh levels of CD44 on cultured murine VSMCsis maintained by the production of endogenousendothelin and an endothelin-convertingenzyme inhibitor was shown to block HA-induced VSMC proliferation.56 LMW-HA alsoinduced increased expression of VCAM-1 onmurine VSMCs in vitro, at least in part throughthe activation of NF-κB.10 It was also demon-strated that CD44 is required for the upregula-tion of VCAM-1 on VSMCs in murineatherosclerotic lesions. Similar to the findingsin human VSMCs, LMW-HA induced the prolif-eration of murine VSMCs through a CD44-dependent mechanism.10 Thus, the ability ofHA to stimulate VSMC activation is dependentupon the form of HA. The intermediate to lowmolecular weight forms of HA that accumulatein atherosclerotic lesions as the result of degra-

dation of HMW-HA, either by hyaluonidases orby oxygen/nitrogen radicals,57 stimulate thegrowth and activation of human as well asmurine VSMC and do so largely via CD44. Inmarked contrast, HMW-HA, which is thepredominant form found in normal vessels,inhibits growth factor-induced proliferation ofVSMC in vitro, also through interaction withCD44. Based on these results it was hypothe-sized that, rather than a default pathway, themaintenance of VSMCs in a quiescent state mayrequire receptor-mediated interactions withcomponents of extracellular matrix. In addi-tion, ECs in normal vessels produce severalfactors that contribute to maintaining VSMCsin a quiescent state, including prostacyclin,nitric oxide and TGF-β. Taken together, thesefindings suggest the HMW-HA may contributeto maintenance of the contractile state ofVCSMs, while LMW-HA may promote de-differ-entiation to the synthetic state.

The role of CD44 in vascular smoothmuscle cell adhesion and migration

Cellular adhesion, motility and migration arepartially dependent on the integrity ofintegrin–adhesion complexes and organizationof the actin cytoskeleton.58,59 Cell movement isdetermined in part by the balance betweenadhesion and de-adhesion of integrin–adhesion complexes.60,61 Data are accumulat-ing that implicate CD44 in the regulation ofcell adhesion and motility. CD44 plays a role inthe homing and migration of leukocytes, andits association with increased invasiveness ofcertain tumours suggests that it may participatein adhesion and cell motility.23,62 Its affinity forHA contributes to the potential for CD44 toparticipate in cellular migration, particularly inHA-rich matrices.63 Interaction of CD44 withcytoskeletal components provides the basis bywhich CD44 can regulate adhesion complexesand play a role in tissue remodelling. CD44interactions with ERM proteins, and thecoupling of CD44 to signal transductionpathways involving Rho family GTPases, PI3-kinases and tyrosine kinases, suggest that CD44may participate in cytoskeletal reorganization


ch17 14/7/05 4:48 pm Page 198

and signalling events important for establish-ing cell polarity.

Actin polymerization is dependent on avariety of factors: formation of integrin–adhesion complexes with stable focal adhesioncontacts, activation of src kinases and RhoAGTPases and downstream effectors such as focaladhesion kinase (FAK) and paxillin, and growthfactors such as TGF-β.64–67 Studies have impli-cated CD44 as a co-receptor for α4β1 integrin inadhesion,68 and as having a role in α4β1-mediated cell trafficking.69 However, directevidence supporting a role for CD44 in integrin-dependent adhesion or signalling is still lacking.A more convincing link exists between CD44and the cytoskeleton. The cytoplasmic domainof CD44 undergoes post-translational modifica-tions, which impacts on its binding affinity forcytoskeletal-associated proteins such as ankyrinand the ezrin/radixin/moesin (ERM) pro-teins.70–72 Furthermore, CD44 can mediate theactivation of src kinases and may associate withsmall GTPases. CD44 may also regulate GTPaseactivity through its association with Rho GDP-dissociation inhibitor.73,74 Based on these obser-vations, the potential for CD44 to participate inintegrin-dependent adhesion and its impact onleukocyte and vascular cell function in lesionswarrants further investigation.

Rho-family GTPases have been implicated incell polarization, with Cdc42 and Rac contribut-ing to a balanced formation of lamellapodiaand filopodia.75,76 The integrated formation oflamellapodia and filopodia leads to both stableprotrusions at the leading edge and detach-ments in the rear, providing a basis for direc-tional migration.60,77,78 These observationssuggest that regulation of the actin cytoskeletonand adhesion is a critical feature contributingto directional migration. Differentially activatedintegrins located at the leading edge translateto specific intracellular signalling events,leading to stable lamellapodia formation.

The ERM proteins serve as importantmembrane cytoskeletal linkers, contributing tothe formation of integrin complexes and actinreorganization.79 Phosphatidylinositol-4,5-bis-phosphate (PIP2) has been shown to stabilizebinding between CD44 and the ERM proteins

and phosphatidylinositol 3,4,5-triphosphatekinase (PI3K) and integrin activation participatein maintaining polarization.80,81 CD44 localizesto the leading edge following activation of keyGTPases, indicating it may play a role in deter-mining cell polarity through stabilization of F-actin.81–83 Thus CD44 may participate indirectional migration of VSMCs through itsinteraction with intracellular proteins impli-cated in cell polarity. Furthermore, these obser-vations provide a potential mechanism by whichCD44 may participate in cytoskeletal re-organi-zation and signalling events, establishing the cellpolarity critical for directional migration andthereby playing an important role in vascularremodelling. Further studies to determine themechanisms by which CD44 regulates VSMClocomotion, polarity and migration are likely toprovide additional potential targets for thera-peutic intervention in atherosclerosis.

Matrix composition

CD44 is critical to both the formation andturnover HA-rich peri-cellular matrix that influ-ences adhesion and migration of cells.23,63 Also,specific proteoglycan forms of CD44 have beenshown to exhibit affinity for ECM proteins suchas fibronectin, collagen and osteopontin. ECMproteins such as fibronectin contain RGDmotifs and, upon binding to integrins, induceactivation of RhoA, formation of stress fibresand establishment of focal adhesion complexesthat provide a link between the extracellularenvironment and the cytoskeleton.58,84 Certainproteoglycans, such as syndecan-4, serve asimilar function by exposing integrin-bindingsites on fibronectin, facilitating the formationof adhesion complexes and playing a role in theorganization of extracellular matrix.85 Thisprovides a mechanism by which CD44 mayimpact the composition and organization ofECM and thus affect vascular remodelling.

CD44 may regulate atherogenesis byregulating the activation of TGF-ββ

VSMCs reside in the media of normal vessels ina quiescent, contractile state. ECs contribute to


ch17 14/7/05 4:48 pm Page 199

maintaining VSMC quiescence by the produc-tion of prostacyclin, nitric oxide and TGF-β, allof which are downregulated at sites prone tothe development of atherosclerotic lesions inresponse to variations in haemodynamic andmechanical forces. In contrast to the effects ofTGF-β in normal vessels, TGF-β serves as achemotactic, mitogenic and differentiationstimulus for parenchymal cells in areas of tissueinjury. However, the role of TGF-β is complexin that it can contribute to inflammation atsome stages, but can also play a role in thesuppression of inflammation. Cell surfaceadhesion molecules including αvβ6 andproteases such as MMP-9 play a role in activa-tion of latent TGF-β, thereby localizing theeffects of active TGF-β in an autocrine/paracrine fashion.44,86

Recent studies in a murine model ofbleomycin-induced lung fibrosis87 demon-strated a pivotal role for CD44 in the resolutionof lung inflammation and the transition to areparative fibrotic response. This studyrevealed that CD44-deficient mice had normalto enhanced levels of latent TGF-β, butreduced levels of active TGF-β in bronchialalveolar lavage fluid, suggesting that CD44 maybe critical for regulating TGF-β activity.

The effects of TGF-β may be many, includingmodification of the integrin repertoire, alter-ation of ECM components and induction ofcore protein expression along with associatedglycosaminoglycans. Furthermore, there aredata to indicate that an intact cytoskeleton isrequired for αvβ6 integrin-mediated activationof latent TGF-β, suggesting that without astable cytoskeleton, localized activation of TGF-β may be altered.86 Also intriguing is the notionthat CD44 may regulate activation of TGF-β byserving as a docking site for MMP-9 at the cellsurface, thereby locally activating latent TGF-β.44 Previous studies have also suggested thatCD44 may interact with the TGF-β receptor 1.88

TGF-β is an important factor in determiningthe balance between inflammation and fibro-sis. As the balance between the fibrotic andinflammatory responses may be critical indetermining the susceptibility of atheroscle-rotic plaques to rupture, it is important to gain

a better understanding of the effects of TGF-βand the role of CD44 in mediating the activa-tion of TGF-β in the vessel wall.

Other potential roles for CD44 inatherogenesis

Atherosclerotic lesions are characterized byneovascularization. CD44 has been implicatedin angiogenesis44,89,90 although the mechanismsinvolved are not well defined. Regulation ofboth EC and VSMC growth and migration iscritical to angiogenesis and, as discussed above,can be regulated by CD44. Thus, the impact ofCD44 on neovascularization of atheroscleroticlesions is another area worthy of further inves-tigation.

Finally, considerable experimental evidencesuggests that CD44 promotes cell growth7,91

and, in some cell types, prevents cell deaththrough apoptosis.92–94 The influence of CD44on both expansion and death of macrophagesand VSMCs may have profound effects onlesion size and the morphology of atheroscle-rotic lesions and therefore warrants furtherinvestigation.

Potential role of soluble CD44 inatherosclerosis

Soluble CD44 (sCD44) is present at substantiallevels in normal mouse and human plasma andfurther increases are associated with immuneactivation, while sCD44 levels are reduced instates of immunodeficiency.17 Although thephysiological and pathophysiological signifi-cance of soluble CD44 is not yet known, it isevident that sCD44 has affinity for HA.95

Protease-mediated release of CD44 also leadsto further presenilin-dependent gamma-secre-tase-mediated proteolytic processing of thevestiges of the receptor.20 This leads to therelease of a fragment of the intracellulardomain that can translocate to the nucleus,where it regulates gene expression.21 In view ofthe need for novel biomarkers of inflammationin cardiovascular disease, the results ofongoing studies to assess whether sCD44 levelsare significantly increased in the plasma of


ch17 14/7/05 4:48 pm Page 200

patients with confirmed CVD compared to age-and gender-matched controls will be of inter-est. In addition to its potential as a biomarker,sCD44 is known to have ligand binding capac-ity and therefore to have the potential toimpact on disease progression. Thus, futurestudies to investigate the impact of sCD44 onatherogenesis in mice may provide useful infor-mation regarding targeting of this adhesionreceptor with novel therapeutic approaches toCVD.

CONCLUSIONS AND IMPLICATIONSWITH REGARD TO DEVELOPINGNOVEL THERAPEUTICS

In spite of the rapid advances in identifyingrisk factors of CVD and in understanding thecellular and molecular mechanisms underlyingatherogenesis, CVD continues to be theleading cause of death in the US. Appreciationof the role of inflammation in atherogenesissuggests new opportunities for developingnovel therapeutics that, in combination withlipid-lowering regimens, may enhance pharma-cological approaches to attenuate the develop-ment and progression of atheroscleroticlesions and/or possibly stabilize or regress pre-existing lesions.

It is worth considering the potential oftargeting CD44 in CVD and other inflamma-tory diseases for several reasons. One impor-tant feature of CD44 that makes it aparticularly attractive target is that, althoughCD44 appears to play an important role ininflammation, all evidence to date suggests thatthis receptor is not critical to any essentialhomeostatic functions, at least in mice. Inaddition, CD44 may be particularly amenableto pharmacological intervention in view of themultiple levels at which this receptor isregulated in its function and the multiplemechanisms by which the receptor may

regulate the response of both inflammatorycells and vascular cells in atheroscleroticlesions (Table 1). Specifically, CD44 is regu-lated at the level of its expression and also in itsaffinity for its ligand. Furthermore, proteolyticprocessing of the receptor can modify cell–celland cell–matrix interactions and generatebiologically active fragments of CD44. Addi-tional studies are required to fully delineatethe mechanisms by which CD44 regulates geneexpression, cell growth and cell migration.Once understood, the underlying mechanismsare likely to provide opportunities for modulat-ing the structure and function of CD44 and/orthe pathways downstream of CD44 to selec-tively interfere with the pro-atherogenic activi-ties of this receptor in the vessel wall.

ACKNOWLEDGEMENTS

The author expresses her sincere gratitude toher many colleagues including Dr DanielRader and Dr Richard Assoian, who continu-ously fuel her interest in the role of inflamma-tion in atherosclerosis. The author also greatlyappreciates the opportunity to engage in thisresearch in collaboration with past and presentmembers of her laboratory, especially DrCarolyn Cuff, Dr Liang Zhao and MelissaMiddleton. Ms Adrienne Whitmore providedinvaluable assistance in the preparation of thismanuscript.

Studies in the author’s laboratory have beensupported by the PHS through grants from theNHBLI.


Table 1 Levels of regulation of CD44

• Expression• Adhesion function• Signal transduction• Proteolytic processing

ch17 14/7/05 4:48 pm Page 201

1. Topper JN, Gimbrone MA Jr. Blood flow and vasculargene expression: fluid shear stress as a modulator ofendothelial phenotype. Mol Med Today 1999; 5:40–6

2. Braun M, Pietsch P, Schror K, et al. Cellular adhesionmolecules on vascular smooth muscle cells.Cardiovasc Res 1999; 41:395–401

3. Rolfe BE, Muddiman JD, Smith NJ, et al. ICAM-1expression by vascular smooth muscle is phenotype-dependent. Atherosclerosis 2000; 149:99–110

4. DeGrendele HC, Estess P, Siegelman MH.Requirement for CD44 in activated T cell extravasa-tion into an inflammatory site. Science 1997;278:672–5

5. Hodge-Dufour J, Noble PW, Horton MR, et al.Induction of IL-12 and chemokines by hyaluronanrequires adhesion dependent priming of resident butnot elicited macrophages. J Immunol 1997; 159:2492–500

6. McKee CM, Penno MB, Cowman M, et al. Hyaluronan(HA) fragments induce chemokine gene expressionin alveolar macrophages. The role of HA size andCD44. J Clin Invest 1996; 98:2403–13

7. Jain M, He Q, Lee W-S, et al. Role of CD44 in thereaction of vascular smooth muscle cells to arterialwall injury. J Clin Invest 1996; 97:596–603

8. Evanko SP, Raines EW, Ross R, et al. Proteoglycandistribution in lesions of atherosclerosis depends onlesion severity, structural characteristics, and theproximity of platelet-derived growth factor and trans-forming growth factor-β. Am J Pathol 1998;152:533–46

9. Riessen R, Wight TN, Pastore C, et al. Distribution ofhyaluronan during extracellular matrix remodelingin human restenotic arteries and balloon-injured ratcarotid arteries. Circulation 1996; 93:1141–7

10. Cuff CA, Kothapalli D, Azonobi I, et al. The adhesionreceptor CD44 promotes atherosclerosis by mediatinginflammatory cell recruitment and vascular cell activa-tion. J Clin Invest 2001; 108:1031–40

11. Wang X, Xu L, Wang H, et al. CD44 deficiency inmice protects brain from cerebral ischemia injury. JNeurochem 2002; 83:1172–9

12. Lesley J, Hyman R, Kincade PW. CD44 and its inter-action with extracellular matrix. Adv Immunol 1993;54:271–335

13. Naor D, Sionov RV, Ish-Shalom D. CD44: structure,function, and association with the malignant process.Adv Cancer Res 1997; 71:241–319

14. Ponta H, Sherman L, Herrlich PA. CD44: fromadhesion molecules to signalling regulators. Nat RevMol Cell Biol 2003; 4:33–45

15. Yang B, Yang BL, Savani RC, et al. Identification of acommon hyaluronan binding motif in the hyaluro-nan binding proteins RHAMM, CD44 and linkprotein. EMBO J 1994; 13:286–96

16. Kohda D, Morton CJ, Parkar AA, et al. Solution struc-ture of the link module: a hyaluronan-bindingdomain involved in extracellular matrix stability andcell migration. Cell 1996; 86:767–75

17. Katoh S, McCarthy JB, Kincade PW. Characterizationof soluble CD44 in the circulation of mice. Levels areaffected by immune activity and tumor growth. JImmunol 1994; 153:3440–9

18. Katoh S, Taniguchi H, Matsubara Y, et al.Overexpression of CD44 on alveolar eosinophils withhigh concentrations of soluble CD44 in bronchoalve-olar lavage fluid in patients with eosinophilicpneumonia. Allergy 1999; 54:1286–92

19. Cichy J, Bals R, Potempa J, et al. Proteinase-mediatedrelease of epithelial cell-associated CD44. Extra-cellular CD44 complexes with components of cellularmatrices. J Biol Chem 2002; 277:44440–7

20. Cichy J, Puré E. The liberation of CD44. J Cell Biol2003; 161:839–43

21. Lammich S, Okochi M, Takeda M, et al. Presenilindependent intramembrane proteolysis of CD44 leadsto the liberation of its intracellular domain and thesecretion of an Abeta-like peptide. J Biol Chem 2002;277:44754–9

22. Okamoto I, Kawano Y, Murakami D, et al. Proteolyticrelease of CD44 intracellular domain and its role inthe CD44 signaling pathway. J Cell Biol 2001;255:755–62

23. Camp RL, Scheynius A, Johansson C, et al. CD44 isnecessary for optimal contact allergic responses but isnot required for normal leukocyte extravasation. JExp Med 1993; 178:497–508

24. Shi M, Dennis K, Peschon JJ, et al. Antibody-inducedshedding of CD44 from adherent cells is linked to theassembly of the cytoskeleton. J Immunol 2001;167:123–31

25. Nakamura H, Suenaga N, Taniwaki K, et al.Constitutive and induced CD44 shedding by ADAM-like proteases and membrane-type 1 matrix metallo-proteinase. Cancer Res 2004; 64:876–82

26. Kajita M, Itoh Y, Chiba T, et al. Membrane-type 1matrix metalloproteinase cleaves CD44 and promotescell migration. J Cell Biol 2001; 153:893–904

27. Kawano Y, Okamoto I, Murakami D, et al. Rasoncoprotein induces CD44 cleavage throughphosphoinositide 3-OH kinase and the Rho family ofsmall G proteins. J Biol Chem 2000; 275:29628–35

28. Yu Q, Toole BP. A new alternatively spliced exonbetween v9 and v10 provides a molecular basis forsynthesis of soluble CD44. J Biol Chem 1996;271:20603–7

29. Puré E, Camp RL, Peritt D, et al. Defective phospho-rylation and hyaluronate binding of CD44 with pointmutations in the cytoplasmic domain. J Exp Med1995; 181:55–62


References

ch17 14/7/05 4:48 pm Page 202

30. Lazaar AL, Puré E. CD44: a model for regulatedadhesion function. Immunologist 1995; 3:19–25

31. Perschl A, Lesley J, English N, et al. Role of CD44cytoplasmic domain in hyaluronan binding. Eur JImmunol 1995; 25:495–501

32. Katoh S, Zheng Z, Oritani K, et al. Glycosylation ofCD44 negatively regulates its recognition of hyaluro-nan. J Exp Med 1995; 182:419–29

33. Lesley J, English N, Perschl A, et al. Variant cell linesselected for alterations in the function of the hyaluro-nan receptor CD44 show differences in glycosylation.J Exp Med 1995; 182:431–7

34. Maiti A, Maki G, Johnson P. TNF-alpha induction ofCD44-mediated leukocyte adhesion by sulfation.Science 1998; 282:941–3

35. Cichy J, Puré E. Oncostatin M and transforminggrowth factor-β1 induce post-translational modifica-tion and hyaluronan binding to CD44 in lung-derivedepithelial tumor cells. J Biol Chem 2000; 275:18061–9

36. Delcommenne M, Kannagi R, Johnson P. TNF-alphaincreases the carbohydrate sulfation of CD44: induc-tion of 6-sulfo N-acetyl lactosamine on N- and O-linked glycans. Glycobiology 2002; 12:613–22

37. Brown KL, Maiti A, Johnson P. Role of sulfation inCD44-mediated hyaluronan binding induced byinflammatory mediators in human CD14(+) periph-eral blood monocytes. J Immunol 2001; 167:5367–74

38. Johnson P, Maiti A, Brown KL, et al. A role for the celladhesion molecule CD44 and sulfation in leukocyte–endothelial cell adhesion during an inflammatoryresponse? Biochem Pharmacol 2000; 59:455–65

39. Knudson W, Bartnik E, Knudson CB. Assembly ofpericellular matrices by COS-7 cells transfected withCD44 lymphocyte-homing receptor genes. Proc NatlAcad Sci USA 1993; 90:4003–7

40. Underhill CB, Nguyen HA, Shizari M, et al. CD44positive macrophages take up hyaluronan duringlung development. Devel Biol 1993; 155:324–36

41. Henke CA, Roongta U, Mickelson DJ, et al. CD44-related chondroitin sulfate proteoglycan, a cellsurface receptor implicated with tumor cell invasion,mediates endothelial cell migration on fibrinogenand invasion into a fibrin matrix. J Clin Invest 1996;97:2541–52

42. Jalkanen S, Jalkanen M. Lymphocyte CD44 binds theCOOH-terminal heparin-binding domain offibronectin. J Cell Biol 1992; 116:817–25

43. Weber GF, Ashkar S, Glimcher MJ, et al. Receptor–ligand interaction between CD44 and osteopontin(Eta-1). Science 1996; 271:509–12

44. Yu Q, Stamenkovic I. Cell surface-localized matrixmetalloproteinase-9 proteolytically activates TGF-βand promotes tumor invasion and angiogenesis.Genes Devel 2000; 14:163–76

45. Zhao L, Cuff CA, Moss E, et al. Selective interleukin-12 synthesis defect in 12/15-lipoxygenase-deficientmacrophages associated with reduced atherosclerosisin a mouse model of familial hypercholesterolemia. JBiol Chem 2002; 277:35350–6

46. DeGrendele HC, Estess P, Picker LJ, et al. CD44 andits ligand hyaluronate mediate rolling under physio-logic flow: a novel lymphocyte–endothelial cellprimary adhesion pathway. J Exp Med 1996;183:1119–30

47. Nandi A, Estess P, Siegelman MH. Hyaluronananchoring and regulation on the surface of vascularendothelial cells is mediated through the functionallyactive form of CD44. J Biol Chem 2000; 275:14939–48

48. Mohamadzadeh M, DeGrendele H, Arizpe H, et al.Proinflammatory stimuli regulate endothelialhyaluronan expression and CD44/HA-dependentprimary adhesion. J Clin Invest 1998; 101:97–108

49. Horton MR, McKee CM, Bao C, et al. Hyaluronanfragments synergize with interferon-γ to induce the C-X-C chemokines mig and interferon-inducibleprotein-10 in mouse macrophages. J Biol Chem 1998;273:35088–94

50. Lee T-S, Yen H-C, Pan C-C, et al. The role of inter-leukin 12 in the development of atherosclerosis inApoE-deficient mice. Arterioscler Thromb Vasc Biol1999; 19:734–42

51. Gupta S, Pablo AM, Jiang X, et al. IFN-gamma poten-tiates atherosclerosis in ApoE knock-out mice. J ClinInvest 1997; 99:2752–61

52. Davenport P, Tipping PG. The role of interleukin-4and interleukin-12 in the progression of atherosclero-sis in apolipoprotein E-deficient mice. Am J Pathol2003; 163:1117–25

53. Benagiano M, Azzurri A, Ciervo A, et al. T helper type1 lymphocytes drive inflammation in human athero-sclerotic lesions. Proc Natl Acad Sci USA 2003; 100:6658–63

54. Cybulsky MI, Iiyama K, Li H, et al. A major role forVCAM-1, but not ICAM-1, in early atherosclerosis. JClin Invest 2001; 107:1255–62

55. Li H, Cybulsky MI, Gimbrone MA, et al. Inducibleexpression of vascular cell adhesion molecule-1 byvascular smooth muscle cells in vitro and within rabbitatheroma. Am J Pathol 1993; 143:1551–9

56. Tanaka Y, Makiyama Y, Mitsui Y. Endothelin-1 isinvolved in the growth promotion of vascular smoothcells by hyaluronic acid. Int J Cardiol 2000; 76:39–47

57. Li M, Rosenfeld L, Vilar RE, et al. Degradation ofhyaluronan by peroxynitrite. Arch BiochemBiophysics 1997; 341:245–50

58. Hynes RO. Integrins: versatility, modulation andsignaling in cell adhesion. Cell 1992; 69:11–25

59. Burridge K, Fath K. Focal contacts: transmembranelinks between the extracellular matrix and thecytoskeleton. BioEss 1989; 10:104–8

60. Lauffenburger DA, Horwitz AF. Cell migration: aphysically integrated molecular process. Cell 1996;84:359–69

61. Huttenlocher A, Ginsberg MH, Horwitz AF.Modulation of cell migration by integrin-mediatedcytoskeletal linkages and ligand-binding affinity. JCell Biol 1996; 134:1551–62

62. Bourguignon LY, Gunja-Smith Z, Iida N, et al.


ch17 14/7/05 4:48 pm Page 203

CD44v(3,8-10) is involved in cytoskeleton-mediatedtumor cell migration and matrix metalloproteinase(MMP-9) association in metastatic breast cancer cells.J Cell Physiol 1998; 176:206–15

63. Knudson CB. Hyaluronan receptor-directed assemblyof chondrocyte pericellular matrix. J Cell Biol 1993;120:825–34

64. Bhowmick NA, Ghiassi M, Aakre M, et al. TGF-beta-induced RhoA and p160ROCK activation is involvedin the inhibition of Cdc25A with resultant cell-cyclearrest. Proc Natl Acad Sci USA 2003; 100:15548–53

65. Ridley AJ, Hall A. The small GTP-binding protein rhoregulates the assembly of focal adhesions and actinstress fibers in response to growth factors. Cell 1992;70:389–99

66. Nobes CD, Hall A. Rho, rac, and cdc42 GTPasesregulate the assembly of multimolecular focalcomplexes associated with actin stress fibers, lamel-lipodia, and filopodia. Cell 1995; 81:53–62

67. Parsons JT, Martin KH, Slack JK, et al. Focal adhesionkinase: a regulator of focal adhesion dynamics andcell movement. Oncogene 2000; 19:5606–13

68. Verfaillie CM, Benis A, Iida J, et al. Adhesion ofcommitted human hematopoietic progenitors tosynthetic peptides from the C-terminal heparin-binding domain of fibronectin: cooperation betweenthe integrin alpha 4 beta 1 and the CD44 adhesionreceptor. Blood 1994; 84:1802–11

69. Kawakami N, Nishizawa F, Sakane N, et al. Roles ofintegrins and CD44 on the adhesion and migration offetal liver cells to the fetal thymus. J Immunol 1999;163:3211–16

70. Tsukita S, Oishi K, Sato N, et al. ERM family membersas molecular linkers between the cell surface glyco-protein CD44 and actin-based cytoskeletons. J CellBiol 1994; 126:391–401

71. Legg JW, Lewis CA, Parsons M, et al. A novel PKC-regulated mechanism controls CD44 ezrin associationand directional cell motility. Nat Cell Biol 2002;4:399–407

72. Zhu D, Bourguignon LY. Interaction between CD44and the repeat domain of ankyrin promoteshyaluronic acid-mediated ovarian tumor cell migra-tion. J Cell Physiol 2000; 183:182–95

73. Bourguignon LY, Zhu H, Shao L, et al. Rho-kinase(ROK) promotes CD44v(3,8-10)-ankyrin interactionand tumor cell migration in metastatic breast cancercells. Cell Motil Cytoskel 1999; 43:269–87

74. Takahashi K, Sasaki T, Mammoto A, et al. Direct inter-action of the Rho GDP dissociation inhibitor withezrin/radixin/moesin initiates the activation of theRho small G protein. J Biol Chem 1997; 272:23371–5

75. Hall A. Rho GTPases and the actin cytoskeleton.Science 1998; 279:509–14

76. Hall A, Nobes CD. Rho GTPases: molecular switchesthat control the organization and dynamics of theactin cytoskeleton. Phil Trans R Soc Lond B Biol Sci2000; 355:965–70

77. Etienne-Manneville S, Hall A. Integrin-mediatedactivation of Cdc42 controls cell polarity in migratingastrocytes through PKCzeta. Cell 2001; 106:489–98

78. Webb DJ, Parsons JT, Horwitz AF. Adhesion assembly,disassembly and turnover in migrating cells – overand over and over again. Nat Cell Biol 2002;4:E97–E100

79. Mackay DJ, Esch F, Furthmayr H, et al. Rho- and Rac-dependent assembly of focal adhesion complexes andactin filaments in permeabilized fibroblasts: an essen-tial role for ezrin/radixin/moesin proteins. J CellBiol 1997; 138:927–38

80. Hirao M, Sato N, Kondo T, et al. Regulation mecha-nism of ERM (Ezrin/Radixin/Moesin) protein/plasma membrane association: possible involvementof phosphatidylinositol turnover and Rho-dependentsignaling pathway. J Cell Biol 1996; 135:37–51

81. Katagiri K, Shimonaka M, Kinashi T. Rap1-mediatedlymphocyte function-associated antigen-1 activationby the T cell antigen receptor is dependent onphospholipase C-gamma1. J Biol Chem 2004; 279:11875–81

82. Seveau S, Eddy RJ, Pierini MFR, et al. Cytoskeleton-dependent membrane domain segregation duringneutrophil polarization. Mol Biol Cell 2001;12:3550–62

83. del Pozo MA, Vicente-Manzanares M, Tejedor R, et al.Rho GTPases control migration and polarization ofadhesion molecules and cytoskeletal ERM compo-nents in T lymphocytes. Eur J Immunol 1999;29:3609–20

84. Clark EA, Brugge JS. Integrins and signal transduc-tion pathways: the road taken. Science 1995;268:233–9

85. Saoncella S, Echtermeyer F, Denhez F, et al.Syndecan-4 signals cooperatively with integrins in aRho-dependent manner in the assembly of focaladhesions and actin stress fibers. Proc Natl Acad SciUSA 1999; 96:2805–10

86. Munger JS, Huang X, Kawakatsu H, et al. Theintegrin alpha v beta 6 binds and activates latent TGFbeta 1: a mechanism for regulating pulmonary inflam-mation and fibrosis. Cell 1996; 96:319–28

87. Teder P, Vandivier RW, Jiang D, et al. Resolution oflung inflammation by CD44. Science 2002; 296:155–8

88. Bourguignon LYW, Singleton PA, Zhu H, et al.Hyaluronan promotes signaling interaction betweenCD44 and transforming growth factor β receptor I inmetastatic breast tumor cells. J Biol Chem 2002;277:39703–12

89. Trochon V, Mabilat C, Bertrand P, et al. Evidence ofinvolvement of CD44 in endothelial cell proliferation,migration and angiogenesis in vitro. Int J Cancer1996; 66:664–8

90. Savani RC, Cao G, Pooler PM, et al. Differentialinvolvement of the hyaluronan (HA) receptors CD44and receptor for HA mediated motility in endothelialcell function and angiogenesis. J Biol Chem 2001;276:36770–8


ch17 14/7/05 4:48 pm Page 204

91. Khaldoyanidi S, Denzel A, Zoller M. Requirementfor CD44 in proliferation and homing of hematopoi-etic precursor cells. J Leukocyte Biol 1996;60:579–92

92. Ayroldi E, Cannarile L, Migliorati G, et al. CD44 (Pgp-1) inhibits CD3 and dexamethasone-induced apopto-sis. Blood 1995; 86:2672–8

93. Gunthert AR, Strater J, Von Reyher U, et al. Earlydetachment of colon carcinoma cells during CD95(APO-1/Fas)-mediated apoptosis. I. De-adhesion

from hyaluronate by shedding of CD44. J Cell Biol1996; 134:1089–96

94. Yu Q, Toole BP, Stamenkovic I. Induction of apopto-sis of metastatic mammary carcinoma cells in vivo bydisruption of tumor cell surface CD44 function. J ExpMed 1997; 186:1985–96

95. Skelton TP, Zeng C, Nocks A, et al. Glycosylationprovides both stimulatory and inhibitory effects oncell surface and soluble CD44 binding to hyaluronan.Cell Biol 1998; 140:431–46


ch17 14/7/05 4:48 pm Page 205

ch17 14/7/05 4:48 pm Page 206

LIPOXYGENASES: BACKGROUNDINFORMATION

The lipoxygenases (LOs) constitute a family ofnon-haem iron dioxygenase enzymes thatstereospecifically insert molecular oxygen intopolyunsaturated fatty acids.1–4 Based on thespecific position where arachidonic acid (AA) isoxidized, they are divided into 5-, 8-, 12-, 15- or12/15-LO. There are seven functional lipoxyge-nases in mice and six in humans and the firstgenetic defects in human LO pathways werereported recently.5 Both plants and animals

express lipoxygenases and, in the case ofmammals, the preferred substrate is often AAand the product a hydro(pero)xy-eicosa-tetraenoic acid (H(P)ETE). In this chapter, wewill consider two members, 12/15-LO and 5-LO.

We use the collective term 12/15-LO tocover the ‘leukocyte type’ 12-lipoxygenase (L-12LO) of mice (also present in pigs and ratsbut not present in the human genome6,7) and15-LO-1 of humans (also present in rabbits, buta distinct 15-LO-1 gene does not appear to bein the published mouse genome8).9 12/15-LOfrom different species produces variable ratiosof 12-HETE to 15-HETE. However, suchspecies differences are not observed with 5-LOsince there is conversion of AA strictly to 5-HPETE and then to leukotriene (LT)A4

10–14

(Figure 1).

BASIC SCIENCE OF 12/15-LO AND 5-LO: MOLECULAR STRUCTURE ANDEXPRESSION

The mouse/human 12/15-LO genes span7.8/10.7 kb including 14 exons and 13 intronson mouse chromosome 11/human chromo-some 17. The primary sequences share 74%identity between species. The three-dimen-sional crystal structure of the rabbit enzyme hasbeen elucidated. The 75 kDa protein possessesan N-terminal β barrel domain15 related to theC-terminal domain in lipases and is probablyinvolved in membrane binding for proper

Lipoxygenases: potential therapeutictarget in atherosclerosisL. Zhao and C.D. Funk

18

12/15-LO12-HPETE

12/15-LO

15-HPETE

5-LO and FLAP

5-HPETE

5-LO

LTA4

LTB4

LTD4

LTE4

LTC4LTC4 synthase

LTA4 hydrolase

γ-Glutamyl transpeptidase

γ-Glutamyl leukotrienase

Arachidonic acid

Dipeptidase

Figure 1 12/15-LO and 5-LO pathways

ch18 14/7/05 4:48 pm Page 207

acquisition of substrate. The C-terminalcatalytic domain contains the iron atominvolved in catalysis that is bound by threestrictly conserved histidines and the carboxylmoiety of the C-terminal isoleucine residue.

Mouse 12/15-LO is distributed in severaltissues with the highest expression inperitoneal macrophages, and lesser expressionin adipose tissue, pancreatic islet cells and avariety of brain regions (cortex, hippocampus,striatum, brainstem, cerebellum, pituitary andpineal glands).16,17 12/15-LO resides in thecytosol of resting macrophages. When macro-phages are incubated with apoptotic cells, thisenzyme translocates from cytosol to the plasmamembrane and is more extensively concen-trated at sites where macrophages bindapoptotic cells, co-localizing with polymerizedactin of emerging filopodia.18 The enzymeappears to translocate to the plasma mem-brane in a cultured macrophage cell line(J774.1), overexpressing 12/15-LO when incu-bated with LDL in a LRP-dependent fashion,19

but peritoneal macrophage 12/15-LO does nottranslocate in the presence of oxidized LDL.20

12/15-LO–/– mice, in which expression of12/15-LO is abolished and subsequent forma-tion of 12-HETE/15-HETE is absent, have beencreated.21 This mouse model has become anessential tool in unravelling biological andpathophysiological functions of 12/15-LO. Sofar, growing evidence has suggested roles of12/15-LO in various pathological processes,including atherosclerosis,22–26 drug addictionresponses,17 prostate carcinoma27 andischaemic preconditioning-induced cardiopro-tection.28 In the current chapter, our discussionwill focus on the importance of this enzyme inatherosclerosis and related mechanisms.

The human 5-LO gene contains 14 exonsspread out over 72 kb on human chromosome10q11.2 in a CpG-rich DNA segment. Itencodes a 673-amino-acid protein with a calcu-lated molecular mass of 78 kDa.29,30 The mouse5-LO gene maps to the central region ofchromosome 6 with a similar exon/intronformat.31 The crystal structure of 5-LO has notbeen elucidated. However, a number of impor-tant features have been inferred from homol-

ogy modelling and verified by biochemicaldata.32–38 Like 12/15-LO, 5-LO is predicted tocontain two domains. The N-terminal domainappears to resemble a C2 domain, a Ca++-dependent membrane targeting module,found in several signalling molecules likeprotein kinase C and cytosolic phospholipaseA2. Two bound Ca++ ions are present in thisdomain and, together with conserved trypto-phan residues, they govern subcellular localiza-tion to phosphatidylcholine-rich membraneslike the nuclear envelope upon cellular activa-tion.32–34 The catalytic domain with the essen-tial iron atom may resemble the elucidated15-LO structure, but in addition contains sitesfor phosphorylation not present in 15-LO, aswell as an SH3-binding domain and nuclearlocalization sequences.35–37

In resting bone-marrow-derived mast celland some macrophage populations, 5-LO isfound within the nucleus. Recent data suggestthat the positioning of 5-LO within the nucleusof resting cells is a powerful determinant of thecapacity to generate LTB4 upon subsequentactivation.38 Studies using 5-LO–/– mice haveprovided extensive evidence for the impor-tance of 5-LO in multiple disease models suchas inflammation,39 arthritis,40 pulmonary fibro-sis,41 autoimmune disease,42 pancreatitis,43

asthma,44 microbial infection,45,46 atherosclero-sis47 and renal transplant rejection.48 Currentunderstanding of the action of 5-LO in athero-sclerosis is discussed in this chapter.

MODEL SYSTEMS: ROLE OF 12/15-LOIN ATHEROSCLEROSIS

Atherosclerosis, a chronic inflammatory disor-der of the vascular wall, is characterized by theprogressive formation of fatty streak lesions,stable plaques and unstable or rupturedplaques which triggers acute clinical complica-tions such as infarction and stroke. Althoughconsiderable experimental and clinical studieshave provided evidence in understanding thepathogenesis of this disease, the mechanisms ofatherogenesis still remain to be investigatedfurther.


ch18 14/7/05 4:48 pm Page 208

Mouse models of atherosclerosis havebecome the key research tools to explore thepathogenesis of atherosclerosis.49–52 Studies of12/15-LO deficiency and human 15-LO-1overexpression in mouse models of atheroscle-rosis have provided considerable evidencerevealing the functional importance of 12/15-LO in atherogenesis. The apolipoprotein E(apoE)-deficient mouse is the hallmark modelsince it develops typical lesions on a normalchow diet that faithfully mimic human diseaseprogression, including monocyte adhesion,foamy macrophages, fatty streaks and advancedfibrosis.49,50 Disruption of 12/15-LO in apoE–/–

mice reduced atherosclerotic lesion progres-sion in aortic vasculature throughout the lifes-pan of the mice, despite no differences incholesterol, triglyceride levels and lipoproteinlevels.22,23 Recent data demonstrate thatdeficiency of 12/15-LO in bone marrow-derived cells protects apoE–/– mice fed aWestern diet from atherosclerosis to the sameextent as complete absence of 12/15-LO.Consistent with this observation, reconstitutionof apoE–/–, 12/15-LO–/– mice with bonemarrow of apoE–/– mice fully restores lesionburden to the levels of apoE–/– mice, suggestingthat 12/15-LO in bone-marrow-derived cellsaccounts for the pro-atherogenic activity of thisenzyme.52a These data suggest an essential roleof 12/15-LO in atherosclerotic lesion develop-ment in the apoE–/– mouse model.

The low-density lipoprotein receptor (LDL-R)-deficient mouse model is also well estab-lished, but requires fat feeding to induce lesiondevelopment and elevated LDL cholesterollevels.51 Collective studies by two independentgroups have demonstrated significantlyreduced atherosclerotic lesions at both theaortic root and the entire aorta in LDL-R–/–,12/15-LO–/– mice on a high-fat, high-choles-terol diet for 3, 9, 12 and 18 weeks, in theabsence of alterations in plasma lipids.24 In aseparate study, overexpression of human 15-LO-1 using the preproendothelin-1 promoterled to enhanced lesion development in LDL-R–/– mice on an atherogenic diet for 3 and 6weeks.25 Thus, the pro-atherogenic role ofmouse 12/15-LO or human 15-LO-1 was

verified in a second mouse model of athero-sclerosis.

The apobec-1/LDL-R double knockoutmodel is perhaps one of the best models ofatherosclerosis that mimics human familialhypercholesterolaemia.52 In order to furtherconfirm the role of 12/15-LO in atherogenesis,we crossed the 12/15-LO–/– mice to this thirdmouse model of atherosclerosis. Significantlyreduced (~50%) lesion size throughout theaorta was observed in both male and femaleapobec-1/LDL-R, 12/15-LO triple knockout 8-month-old mice compared to apobec-1, LDL-Rdouble knockout mice, despite no differencesin plasma total cholesterol levels.26 Thus,studies in mice consistently support a role of12/15-LO in promoting atherogenesis in threedifferent mouse models of atherosclerosis.

In hypercholesterolaemic rabbit models,modulation of 15-LO activity appears to play animportant role in atherogenesis. Two studieswith a putatively selective 15-LO inhibitor (thecompound PD146176) lacking significant anti-oxidant properties have shown a significantreduction in diet-induced atherosclerosis andmonocyte/macrophage accumulation inlesions.53,54 However, conflicting data withoverexpression of 15-LO from the lysozymepromoter or anaemia-induced 15-LO expres-sion leading to a paradoxical reduction oflesion formation have been reported.55,56 Thepro- and anti-atherogenic possibilities of 15-LOinvolvement in atherogenesis in rabbits remainto be further determined.

The role of 12/15-LO in restenosis has beenexamined in a rat model. One study demon-strated increased 12/15-LO mRNA and proteinexpression in the neointima of balloon-injuredrat carotid arteries.57 Specific 12/15-LO inacti-vation using a ribozyme significantly reducedthe intima-to-media thickness ratio in the leftcommon carotid arteries of rats 12 days afterballoon catheter injury.58 Another line ofevidence supporting the pro-atherogenic roleof 12/15-LO arises from a pig model.Accelerated atherosclerosis in hyperlipaemicgroups was associated with enhanced expres-sion of 12-LO, both at the mRNA and proteinlevels.59

LIPOXYGENASES: POTENTIAL THERAPEUTIC TARGET IN ATHEROSCLEROSIS 209

ch18 14/7/05 4:48 pm Page 209

MECHANISMS INVOLVED IN THEPRO-ATHEROGENIC ROLE OF 12/15-LO

LDL oxidation

Considerable evidence has accumulated overthe years supporting the hypothesis that LDLoxidation plays an important role in atheroge-nesis.60–68 12/15-LO catalyses the oxygenationof AA or linoleic acid in free form or whenesterified to cholesterol or phospholipids togenerate predominantly the 12/15-H(P)ETEor 13-H(P)ODE forms (free or esterified),respectively.69,70 Recent data suggest that low-density lipoprotein receptor-related protein(LRP)-mediated membrane translocation of12/15-LO is required for oxidation of LDL inmacrophages.19,71 12/15-LO represents one ofthe many pathways to generate oxidized LDL(oxLDL) in vivo,67,72–75 although some in vitrostudies challenge lipoxygenase involve-ment.76,77 15-LO-1 co-localizes with epitopes ofoxLDL in human and rabbit atheroscleroticlesions, and stereospecific products of thelipoxygenase reaction have been demonstratedin atherosclerotic lesions of rabbits andhumans.78–80 Retroviral transfer of the 15-LOgene into rabbit iliac arteries of rabbits fed ahigh-fat diet led to accumulation of oxLDL-likeepitopes.81 In the apoE–/– mouse model,disruption of 12/15-LO resulted in decreasedurinary and plasma levels of oxidant stressmarkers (isoprostane 8,12–iso-iPF2α-VI and IgGantibodies directed against malondialdehyde-modified LDL epitopes) in parallel withdecreased atherosclerosis.22,23 We also providedevidence that anti-oxidant vitamin E reducedlesion size and urinary 8,12–iso-iPF2α-VI levelsin apoE–/– mice, but had no further effect inapoE–/–,12/15-LO–/– mice, which verified thatthe pro-atherogenic effect of 12/15-LO couldbe attributed to its action on lipid peroxidation(unpublished data). Decreased isoprostanelevels associated with reduced atherosclerosisprogression were further confirmed in theapobec-1/LDL-R–/–,12/15-LO–/– mousemodel.26 In the apoE–/– bone marrow trans-plantation (BMT) model, reduced atheroscle-rotic lesion size in apoE–/– mice receiving

apoE–/–, 12/15-LO–/– bone marrow cells corre-lated with decreased plasma level of autoanti-bodies to oxLDL.52a Enhanced atheroscleroticlesion development was associated with greatersusceptibility of LDL oxidation in LDL-R–/–, 15-LO-1 transgenic mice than in LDL-R–/– mice.25

These combined data are strongly suggestive ofa 12/15-LO component to LDL oxidation andatherogenesis.

Inflammatory response

The importance of inflammation and theunderlying molecular and cellular mechanismscontributing to atherogenesis is clearly recog-nized.82,83 Hypercholesterolaemia and inflam-mation should be considered different aspectsof a single, shared pathogenetic pathway inatherosclerosis.84 Monocytes/macrophages,endothelial cells, lymphocytes and smoothmuscle cells are essential cellular componentsof the inflammatory response in differentstages of atherosclerosis progression, fromdevelopment of the fatty streak to processesthat ultimately contribute to plaque ruptureand myocardial infarction.85–88 The initialadhesion of monocytes to the endothelium ismediated by several adhesion molecules, suchas vascular cell adhesion molecule-1 (VCAM-1), P-selectin and integrin α4β1 (VLA4).89–96

Recent data demonstrated that deficient12/15-LO activity in mouse peritoneal macro-phages led to reduced endothelial activation(in terms of VCAM-1 expression) in thepresence of LDL.52a In contrast, transgenic12/15-LO overexpressing mice have enhancedmonocyte/endothelial cell interactionsthrough molecular regulation of endothelialadhesion molecules.97 Inhibition of 12/15-LO(with either an adenovirus expressing aribozyme to 12/15-LO or the 12/15-LOinhibitor cinnamyl-3,4-dihydroxy-alpha-cyano-cinnamate) also reduced monocyte/endo-thelial interactions in db/db mice, possiblythrough interactions of α4β1 integrins onmonocytes with endothelial VCAM-1 andconnecting segment 1 fibronectin and interac-tions of β2 integrins with endothelial intercel-lular adhesion molecule 1.98


ch18 14/7/05 4:48 pm Page 210

Lesional T cells appear to be activated andproduce Th1 (such as interleukin-12 (IL-12)and interferon-γ (IFN-γ)) and Th2 cytokines(such as IL-4 and IL-13). IL-12 is a key factor inthe induction of T-cell dependent activation ofmacrophages and plays an active role inregulating the immune response of atheroscle-rosis in apoE-deficient mice.99,100 Peritonealmacrophages from 12/15-LO–/– mice exhibit aprominent defect of IL-12p40 productioncompared to wild-type mice. Significantdecreases in IL-12p40 and IFN-γ mRNA as wellas IL-12p40 protein expression in apobec-1/LDL-R, 12/15-LO triple knockout miceaortas were observed. These data provide ameans whereby 12/15-LO influences atheroge-nesis via altered synthesis of Th1 cytokines.26

13-HPODE, a product of 12/15-LO, activatesnuclear factor-κB (NF-κB), as well as Ras,mitogen-activated protein kinases (MAPK1/2),p38 and c-Jun amino-terminal kinase inporcine VSMCs.101 In mouse vascular smoothmuscle cells, 12/15-LO deficiency resulted insignificantly reduced growth-factor-inducedcell migration, proliferation, AP-1-, p38- andcAMP-response element binding proteinactivation, as well as diminished superoxideand fibronectin production.102 These resultssuggest that 12/15-LO is responsible formodulation of multiple key inflammatoryresponses in atherosclerosis.

MODEL SYSTEMS: ROLE OF 5-LO INATHEROSCLEROSIS

Leukotrienes, products of the 5-LO pathway,are known to exert pro-inflammatory effects invivo.103,104 LTB4, an inflammatory cell (especiallyneutrophils) chemoattractant, and LTE4 havebeen detected in human and rabbit atheroscle-rotic lesions.105,106 However, the role of 5-LO inanimal models of atherosclerosis has not beenappreciated until recently, when Mehrabian etal47 reported that 5-LO expression is detected inmouse atherosclerotic lesions. A recent studydemonstrated that deficiency of B-LT1, thehigh-affinity chemoattractant LTB4 receptor,significantly reduced atherosclerotic lesion size

in apoE–/– mice on Western diet for 4 weeks.107

Aiello et al108 provided evidence that an LTB4

antagonist (CP-105,696), via pharmacologicalblockade of its receptor B-LT1, contributes toreduced monocyte infiltration in developingatherosclerotic lesions in apoE–/– mice. Furtherstudy in LDL-R–/– mice revealed that LTB4 mayexert its pro-atherogenic effect through anMCP-1 pathway.108

Genetic studies47,109,110 with the atherosclero-sis-susceptible strain C57BL/6 (B6) and eitherthe resistant strain CAST/Ei or MOLF/Ei ledto the conclusion that one or more genes onmouse chromosome 6 centred over the 63 cM(40–89 cM) interval is/are responsible foratherosclerosis susceptibility. The mouse 5-LOgene resides within this locus (53 cM, 117 Mbfrom telomere31). Interestingly, reduced levelsof 5-LO expression and LT production werefound in a congenic strain (CON6) in whichthe CAST/Ei chromosome 6 segment wasintegrated onto the B6 background,47 suggest-ing that the 5-LO locus in this strain might beresponsible for atherosclerosis resistance.Sequence analysis of the CAST/Ei 5-LOsequence revealed two amino acid differenceswith the B6 strain (Val645/Ile646 vs Ile645/Val646, respectively) and these substitutions,when introduced into human 5-LO (alsoIle645/Val646), led to impaired 5-LO activ-ity.111 It should be noted, however, that theoriginal determined sequence in a hybridB6x129 strain of mice has the sequenceIle645/Ile64631 and displays completelynormal 5-LO activity.

Striking results indicated that the 5-LO locuswas responsible for atherosclerosis susceptibil-ity on an LDL-R knockout background, reveal-ing a profound effect much greater than anyother genes to date.47 Two important caveatswith these studies include limited numbers ofanimals examined in the atherosclerotic lesionanalysis and the effect of 5-LO being observedat the heterozygous level; i.e. not with LDL-R,5-LO double knockouts.47,112 Based on theseconcerns, we examined the role of 5-LO in theapoE–/– mouse model. Our data do not supporta role of 5-LO in spontaneous lesion develop-ment in apoE–/– mice, however, on a pro-


ch18 14/7/05 4:48 pm Page 211

atherogenic, pro-inflammatory diet effects onand aneurysm formation become apparent.113

Recently, 5-LO expression has beendocumented in advanced human atheroscle-rotic lesions, localized to macrophages,dendritic cells, foam cells, mast cells andneutrophilic granulocytes.114 The number of 5-LO-expressing cells markedly increased inadvanced lesions, suggesting an associationbetween 5-LO and lesion development. Apreliminary study claims a role of 5-LO inhuman plaque instability.115 Thus, enhanced 5-LO expression was observed in plaques charac-terized as ‘unstable’ vs ‘stable’ and wasassociated with increased LTB4 production,

increased matrix metalloproteinase-2 (MMP-2)and MMP-9 activity and decreased collagencontent in the former samples.115 A geneticstudy in humans has attempted to link 5-LO tocardiovascular disease (CVD) susceptibility.116

Carotid intima-media thickness (IMT) wasdetermined bilaterally with B-mode ultra-sound, evaluating the risk of atherosclerosisand CVD, and a polymorphism in thetranscription-factor-binding region of the 5-LOpromoter was genotyped in 470 samples. Thefindings suggest that carriers of two variantalleles of the repeat GC-box/Sp1 region (< or>5 copies) in the 5-LO gene place about 6% ofthe population at markedly increased risk for


Table 1 Modulation of 12/15-LO (or 15-LO-1) and 5-LO pathways in animal models of atherosclerosis

Pathway Species Animal model Major observation Reference

12/15-LO15-LO-1

5-LO

Mouse

Rabbit

Mouse

apoE–/–

apoE–/–

LDL-R–/–

LDL-R–/–

apobec-1/LDL-R–/–

Diet-inducedhypercholesterolaemicmodel

Diet-inducedhypercholesterolaemicmodel

apoE–/–

apoE–/–

apoE–/–

LDL-R–/–

22, 23

Unpublisheddata

24

25

26

53, 54

55, 56

107

108

113

47

Significant reduction of aortic lesion size in 12/15-LO–/– mice at 10 w, 15 w, 8 m, 12 m and 15 m onchow diet

Significant reduction of aortic lesion size in apoE–/–

mice receiving apoE–/–, 12/15-LO–/– mice bonemarrow cells on Western diet

Significant reduction of aortic lesion size in12/15–LO–/– mice at 3, 9, 12 and 18 weeks on high-fat, high-cholesterol diet

15–LO-1 overexpression in endothelial cells enhanceslesion development on high-fat, high-cholesteroldiet for 3 and 6 weeks

Significant decrease of aortic lesion percentage in12/15-LO–/– mice on chow diet at 8 months

PD146176 significantly reduces atherosclerotic lesionsize and monocyte/macrophage accumulation

15-LO-1 overexpression or anaemia-induced 15-LO-1expression significantly decreases atheroscleroticlesion size

B-LT1 deficiency significantly reduces atheroscleroticlesion size on Western diet for 4 weeks

LTB4 antagonist, CP-105,696, significantly reducesatherosclerotic lesion size and monocye infiltration

Significant reduction of aneurysm formation in 5-LO–/–

mice fed high-fat, high-cholesterol diet for 8 weeks

Atherosclerotic lesion development significantlyreduced in 5-LO+/– mice

ch18 14/7/05 4:48 pm Page 212

atherosclerosis and CVD. This 5-LO poly-morphism-dependent atherogenic effect isenhanced by dietary intake of arachidonicacid, blunted by n-3 fatty acids and associatedwith increased plasma level of C-reactiveprotein.116 However, it must be stressed herethat no biochemical parameters of the 5-LOpathway correlating the variant alleles withdisease propensity have been measured, so thedata remain highly speculative.

POTENTIAL MECHANISMSINVOLVED IN THE ROLE OF 5-LO INATHEROSCLEROSIS

The studies mentioned above are beginning toreveal potential mechanisms for 5-LO involve-ment in multiple stages of atherosclerosisprogression, including adhesion of monocytesto endothelial cells, chemotaxis and migrationof monocytes/macrophages and plaque insta-bility. The participation of 5-LO in theseprocesses could be mediated through biosyn-thesis of pro-inflammatory leukotrienes, 5-oxo-6,8,11,14-eicosatetraenoic acid (5-oxo-ETE)and/or 5-HETE. Pre-treatment of humanendothelial cells with a 5-LO inhibitor blockedIL-1β-induced VCAM-1 expression.117 A recentstudy by Friedrich et al118 demonstrated thatLTB4, a major product of the 5-LO pathway, isan agonist of monocyte adhesion, possiblythrough triggering β1- and β2-integrin-depen-dent adhesion in vascular models. Monocytechemoattractant protein-1 (MCP-1), a proto-type of CC chemokines, through its receptorCCR2, plays an important role in the patho-genesis of atherosclerosis.119,120 A specific LTB4

receptor antagonist (CP-105,696) inhibitedcaecal ligation and puncture-induced recruit-ment of both neutrophils and macrophages,which was accompanied by a reduced level ofMCP-1 in a murine model of septic peritoni-tis.121 A similar pattern of cross-talk betweenLTB4 and MCP-1 may also occur in the arterywall.122 In fact, evidence of LTB4/MCP-1 inter-action has been reported, whereby LTB4

promotes atherogenesis in LDL-R–/– mice viaMCP-1.108 In addition, 5-oxo-ETE and 5-HETE

induce directional migration and actin poly-merization of human monocytes in vitro.123

Matrix metalloproteinases (MMPs) are keymodulators of plaque stability in atherosclero-sis.124,125 5-LO activity has been shown tomediate CD147-induced generation of pro-MMP-2 from fibroblasts.126 LTB4 induces MMP-2, MMP-3 and MMP-9 secretion in culturedTsup-1 cells (T lymphoblastoma cell line).127

These data suggest a potential role of 5-LO inplaque vulnerability.

THERAPEUTIC POTENTIAL OF 15-LOAND 5-LO INHIBITORS INATHEROGENESIS

15-LO has been recognized as a potentialtarget in human atherogenesis since at least199078 and therapeutic strategies have beenproposed.128 Evidence from several experimen-tal animal models (mice and rabbits)presented above supports the rationale forsuch an approach. Plausible mechanismsinvolved in 15-LO inhibition include reductionof lipid peroxidation, diminished inflamma-tory responses and prevention of monocyte–endothelial cell interactions. Although several15-LO inhibitors have been developed,53,54,129,130

none of them has been applied beyond animalstudies.

A considerable controversy erupted in thepast year with respect to human lipoxygenaseexpression patterns in atherosclerotic tissue.The dogma in the field since 1990 has beenthat 15-LO-1 expression is high in macro-phages of atherosclerotic lesions and that 5-LOexpression is absent or negligible.78 A study bySpanbroek et al114 has come to the oppositeconclusions: 5-LO/leukotriene pathway bio-synthetic and signalling components areabundantly expressed, whereas 15-LO-1 expres-sion is negligible. Who is right? Are thereexplanations for the divergent results? Do thesolid mouse 12/15-LO atherosclerosis datawith atherosclerotic models provide accuratepredictors for human disease intervention?Perhaps the development of novel 15-LOspecific inhibitors and the evaluation of their


ch18 14/7/05 4:48 pm Page 213

efficacy in clinical trials will have to wait until afirm consensus is reached on human lipoxyge-nase expression patterns in atheroscleroticdisease samples and resolving the answers tothese questions.

5-LO inhibition in preventing inflammatorydisease has been established for two decades.103

The premise that this pathway is relevant to theinflammatory component of atherosclerosisonly gained momentum in the last year or two,based on the recent studies in mice andhumans.47,108,114 Some of the reasons for thislong delay since inception of 5-LO inhibitorsinclude:

• The long-known fact that LTB4 is predom-inantly a neutrophil chemoattractant (aminor player in atherosclerosis), withmuch lesser effects on monocytes/macro-phages (key cell type in atherogenesis);

• The relatively recent recognition of ather-osclerosis as an inflammatory diseaserather than a vascular injury disorder;

• The apparent absence of 5-LO expressionin lesional macrophages.78

Numerous 5-LO pathway inhibitors(leukotriene modifiers) have been developedover the past two decades.131–134 Zileuton is theonly 5-LO inhibitor currently approved for usein humans.133,134 While it has demonstratedtherapeutic benefits in asthma symptommanagement, its application in the clinics hasbeen limited due to requirements for liverfunction testing and poor pharmacokinetics.135

The recent exciting data in humans and micein coronary disease samples and atheroscle-rotic models should encourage the pharma-ceutical industry to renew efforts at developingbetter specific 5-LO inhibitors with excellentpharmacodynamics for testing in people at riskfor cardiovascular disease.

ACKNOWLEDGEMENTS

This work was supported by NIH grantHL53558 and CIHR grant MOP-67146 to CDFunk, and American Heart AssociationPostdoctoral Fellowship 0225369U to L Zhao.


1. Funk CD. Prostaglandins and leukotrienes: advancesin eicosanoid biology. Science 2001; 294:1871–5

2. Brash AR. Lipoxygenases: occurrence, functions,catalysis, and acquisition of substrate. J Biol Chem1999; 274:23679–82

3. Funk CD. The molecular biology of mammalianlipoxygenases and the quest for eicosanoid functionsusing lipoxygenase-deficient mice. Biochim BiophysActa 1996; 1304:65–84

4. Yamamoto S. ‘Enzymatic’ lipid peroxidation:reactions of mammalian lipoxygenases. Free RadicBiol Med 1991; 10:149–59

5. Jobard F, Lefevre C, Karaduman A, et al. Lipoxy-genase-3 (ALOXE3) and 12(R)-lipoxygenase(ALOX12B) are mutated in non-bullous congenitalichthyosiform erythroderma (NCIE) linked tochromosome 17p13.1. Hum Mol Genet 2002;11:107–13

6. Lander ES, Linton LM, Birren B, et al. Initialsequencing and analysis of the human genome.Nature 2001; 409:860–921

7. Venter JC, Adams MD, Myers EW, et al. The sequence

of the human genome. Science 2001; 291:1304–518. Waterston RH, Lindblad-Toh K, Birney E, et al. Initial

sequencing and comparative analysis of the mousegenome. Nature 2002; 420:520–62

9. Kuhn H, Walther M, Kuban RJ. Mammalian arachi-donate 15-lipoxygenases: structure, function, andbiological implications. Prost Other Lipid Mediat2002; 68/69:263–90

10. Shimizu T, Radmark O, Samuelsson B. Enzyme withdual lipoxygenase activities catalyzes leukotriene A4synthesis from arachidonic acid. Proc Natl Acad SciUSA 1984; 81:689–93

11. Rouzer CA, Matsumoto T, Samuelsson B. Singleprotein from human leukocytes possesses 5-lipoxyge-nase and leukotriene A4 synthase activities. Proc NatlAcad Sci USA 1986; 83:857–61

12. Ueda N, Kaneko S, Yoshimoto T, et al. Purification ofarachidonate 5-lipoxygenase from porcine leukocytesand its reactivity with hydroperoxyeicosatetraenoicacids. J Biol Chem 1986; 261: 7982–8

13. Shimizu T, Izumi T, Seyama Y, et al. Characterizationof leukotriene A4 synthase from murine mast cells:

References

ch18 14/7/05 4:48 pm Page 214

evidence for its identity to arachidonate 5-lipoxyge-nase. Proc Natl Acad Sci USA 1986; 83:4175–9

14. Hogaboom GK, Cook M, Newton JF et al.Purification, characterization and structural proper-ties of a single protein from rat basophilic leukemia(RBL-1) cells possessing 5-lipoxygenase andleukotriene A4 synthetase activities. Mol Pharmacol1986; 30:510–19

15. Gillmor SA, Villasenor A, Fletterick R, et al. The struc-ture of mammalian 15-lipoxygenase reveals similarityto the lipases and the determinants of substrate speci-ficity. Nat Struct Biol 1997; 4:1003–9

16. Chen XS, Kurre U, Jenkins NA, et al. cDNA cloning,expression, mutagenesis of C-terminal isoleucine,genomic structure, and chromosomal localizations ofmurine 12-lipoxygenases. J Biol Chem 1994;269:13979–87

17. Walters CL, Wang BC, Godfrey M, et al. Augmentedresponses to morphine and cocaine in mice with a 12-lipoxygenase gene disruption. Psychopharmacology(Berl) 2003; 170:124–31

18. Miller YI, Chang MK, Funk CD, et al. 12/15-lipoxy-genase translocation enhances site-specific actinpolymerization in macrophages phagocytosingapoptotic cells. J Biol Chem 2001; 276:19431–9

19. Zhu H, Takahashi Y, Xu W, et al. Low density lipopro-tein receptor-related protein-mediated membranetranslocation of 12/15-lipoxygenase is required foroxidation of low density lipoprotein by macrophages.J Biol Chem 2003; 278:13350–5

20. Miller YI, Worrall DS, Funk CD, et al. Actin polymer-ization in macrophages in response to oxidized LDLand apoptotic cells: role of 12/15-lipoxygenase andphosphoinositide 3-kinase. Mol Biol Cell 2003;14:4196–206.

21. Sun D, Funk CD. Disruption of 12/15-lipoxygenaseexpression in peritoneal macrophages. Enhancedutilization of the 5-lipoxygenase pathway and dimin-ished oxidation of low density lipoprotein. J BiolChem 1996; 271:24055–62

22. Cyrus T, Witztum JL, Rader DJ, et al. Disruption of12/15-lipoxygenase results in inhibition of athero-sclerotic lesion development in mice lackingapolipoprotein E. J Clin Invest 1999; 103:1597–604

23. Cyrus T, Pratico D, Zhao L, et al. Absence of 12/15-lipoxygenase expression decreases lipid peroxidationand atherogenesis in apolipoprotein E-deficient mice.Circulation 2001; 103:2277–82

24. George J, Afek A, Shaish A, et al. 12/15-Lipoxygenasegene disruption attenuates atherogenesis in LDL-receptor deficient mice. Circulation 2001;104:1646–50

25. Harats D, Shaish A, George J, et al. Overexpression of15-lipoxygenase in vascular endothelium acceleratesearly atherosclerosis in LDL receptor-deficient mice.Arterioscler Thromb Vasc Biol 2000; 20:2100–5

26. Zhao L, Cuff CA, Moss E, et al. Selective interleukin-12 synthesis defect in 12/15-lipoxygenase deficientmacrophages associated with reduced atherosclerosis

in mouse model of familial hypercholesterolemia. JBiol Chem 2002; 277:35350–6

27. Shappell SB, Olson SJ, Hannah SE, et al. Elevatedexpression of 12/15-lipoxygenase and cyclooxyge-nase-2 in a transgenic mouse model of prostate carci-noma. Cancer Res 2003; 63:2256–67.

28. Gabel SA, London RE, Funk CD, et al. Leukocyte-type12-lipoxygenase-deficient mice show impairedischemic preconditioning-induced cardioprotection.Am J Physiol Heart Circ Physiol 2001; 280:H1963–9

29. Funk CD, Matsumoto T, Hoshiko S, et al.Characterization of the human 5-lipoxygenase gene.Proc Natl Acad Sci USA 1989; 86:2587–91

30. Matsumoto T, Funk CD, Radmark O, et al. Molecularcloning and amino acid sequence of human 5-lipoxy-genase. Proc Natl Acad Sci USA 1988; 85:26–30

31. Chen XS, Naumann T, Kurre U, et al. cDNA cloning,expression, mutagenesis, intracellular localizationand gene chromosomal assignment of mouse 5-lipoxygenase. J Biol Chem 1995; 270:17993–9

32. Hammarberg T, Provost P, Persson B, et al. The N-terminal domain of 5-lipoxygenase binds calcium andmediates calcium stimulation of enzyme activity. JBiol Chem 2000; 275:38787–93.

33. Kulkarni S, Das S, Funk CD, et al. Molecular basis ofthe specific subcellular localization of the C2-likedomain of 5-lipoxygenase. J Biol Chem 2002;277:13167–74

34. Chen XS, Funk CD. The N-terminal ‘beta-barrel’domain of 5-lipoxygenase is essential for nuclearmembrane translocation. J Biol Chem 2001; 276:811–18

35. Werz O, Szellas D, Steinhilber D, et al. Arachidonicacid promotes phosphorylation of 5-lipoxygenase atSer-271 by MAPK-activated protein kinase 2 (MK2). JBiol Chem 2002; 277:14793–800.

36. Lepley RA, Fitzpatrick FA. 5-lipoxygenase contains afunctional Src homology 3-binding motif that inter-acts with the Src homology 3 domain of Grb2 andcytoskeletal proteins. J Biol Chem 1994; 269:24163–8.

37. Jones SM, Luo M, Peters-Golden M, et al.Identification of two novel nuclear import sequenceson the 5-lipoxygenase protein. J Biol Chem 2003;278:10257–63.

38. Luo M, Jones SM, Peters-Golden M, et al. Nuclearlocalization of 5-lipoxygenase as a determinant ofleukotriene B4 synthetic capacity. Proc Natl Acad SciUSA 2003; 100:12165–70

39. Chen XS, Sheller JR, Johnson EN, et al. Role ofleukotrienes revealed by targeted disruption of the 5-lipoxygenase gene. Nature 1994; 372:179–82

40. Griffiths RJ, Smith MA, Roach ML, et al. Collagen-induced arthritis is reduced in 5-lipoxygenase-activat-ing protein-deficient mice. J Exp Med 1997;185:1123–9

41. Peters-Golden M, Bailie M, Marshall T, et al.Protection from pulmonary fibrosis in leukotriene-deficient mice. Am J Respir Crit Care Med 2002;165:229–35


ch18 14/7/05 4:48 pm Page 215

42. Goulet JL, Griffiths RC, Ruiz P, et al. Deficiency of 5-lipoxygenase abolishes sex-related survival differencesin MRL-lpr/lpr mice. J Immunol 1999; 163:359–66

43. Cuzzocrea S, Rossi A, Serraino I, et al. 5-lipoxygenaseknockout mice exhibit a resistance to acute pancre-atitis induced by cerulein. Immunology 2003;110:120–30

44. Coffey M, Peters-Golden M. Extending the under-standing of leukotrienes in asthma. Curr Opin AllergyClin Immunol 2003; 3:57–63

45. Aliberti J, Serhan C, Sher A. Parasite-induced lipoxinA4 is an endogenous regulator of IL-12 productionand immunopathology in Toxoplasma gondii infection.J Exp Med 2002; 196:1253–62

46. Bailie MB, Standiford TJ, Laichalk LL, et al. Leuko-triene-deficient mice manifest enhanced lethalityfrom Klebsiella pneumonia in association withdecreased alveolar macrophage phagocytic and bacte-ricidal activities. J Immunol 1996; 157:5221–4.

47. Mehrabian M, Allayee H, Wong J, et al. Identificationof 5-lipoxygenase as a major gene contributing toatherosclerosis susceptibility in mice. Circ Res 2002;91:120–6

48. Goulet JL, Griffiths RC, Ruiz P, et al. Deficiency of 5-lipoxygenase accelerates renal allograft rejection inmice. J Immunol 2001; 167:6631–6

49. Reddick RL, Zhang SH, Maeda N. Atherosclerosis inmice lacking apo E. Evaluation of lesional develop-ment and progression. Arterioscler Thromb 1994;14:141–7

50. Nakashima Y, Plump AS, Raines EW, et al. ApoE-deficient mice develop lesions of all phases of athero-sclerosis throughout the arterial tree. ArteriosclerThromb 1994; 14:133–40

51. Ishibashi S, Goldstein JL, Brown MS, et al. Massivexanthomatosis and atherosclerosis in cholesterol-fedlow density lipoprotein receptor-negative mice. J ClinInvest 1994; 93:1885–93

52. Powell-Braxton L, Veniant M, Latvala RD, et al. Amouse model of human familial hypercholes-terolemia: markedly elevated low density lipoproteincholesterol levels and severe atherosclerosis on a low-fat chow diet. Nat Med 1998; 4:934–8

52a. Huo Y, Zhao L, Hyman MC, et al. Critical role ofmacrophage 12/15-lipoxygenase for atherosclerosisin apolipoprotein E-deficient mice. Circulation 2004;110:2024–31

53. Bocan TM, Rosebury WS, Mueller SB, et al. A specific15-lipoxygenase inhibitor limits the progression andmonocyte-macrophage enrichment of hypercholes-terolemia-induced atherosclerosis in the rabbit.Atherosclerosis 1998; 136:203–16

54. Sendobry SM, Cornicelli JA, Welch K, et al.Attenuation of diet-induced atherosclerosis in rabbitswith a highly selective 15-lipoxygenase inhibitorlacking significant antioxidant properties. Br J Pharm1997; 120:1199–206.

55. Shen J, Herderick E, Cornhill JF, et al. Macrophage-mediated 15-lipoxygenase expression protects against

atherosclerosis development. J Clin Invest 1996;98:2201–8

56. Trebus F, Heydeck D, Schimke I, et al. Transientexperimental anemia in cholesterol-fed rabbitsinduces systemic overexpression of the reticulocyte-type 15-lipoxygenase and protects from aortic lipiddeposition. Prost Leuk Ess Fatty Acids 2002; 67:419–28

57. Natarajan R, Pei H, Gu JL, et al. Evidence of 12-lipoxy-genase expression in balloon injured rat carotid arter-ies. Cardiovasc Res 1999; 41:481–99

58. Gu JL, Pei H, Thomas L, et al. Ribozyme-mediatedinhibition of rat leukocyte-type 12-lipoxygenaseprevents intimal hyperplasia in balloon-injured ratcarotid arteries. Circulation 2001; 103:1446–52.

59. Natarajan R, Gerrity RG, Gu JL, et al. Role of 12-lipoxygenase and oxidant stress in hyperglycaemia-induced acceleration of atherosclerosis in a diabeticpig model. Diabetologia 2002; 45:125–33

60. Funk CD, Cyrus T. 12/15-Lipoxygenase, oxidativemodification of LDL and atherogenesis. TrendsCardiovasc Med 2001; 11:116–24

61. Cathcart MK, Folcik VA. Lipoxygenases and athero-sclerosis: protection versus pathogenesis. Free RadicBiol Med 2000; 28:1726–34

62. Kuhn H, Chan L. The role of 15-lipoxygenase inatherogenesis: pro- and antiatherogenic actions. CurrOpin Lipidol 1997; 8:111–17

63. Steinberg D, Parthasarathy S, Carew TE, et al. Beyondcholesterol. Modifications of low-density lipoproteinthat increase its atherogenicity. New Engl J Med 1989;320:915–24

64. Yla-Herttuala S, Palinski W, Rosenfeld ME, et al.Evidence for the presence of oxidatively modified lowdensity lipoprotein in atherosclerotic lesions of rabbitand man. J Clin Invest 1989; 84:1086–95

65. Parthasarathy S, Rankin SM. Role of oxidized lowdensity lipoprotein in atherogenesis. Prog Lipid Res1992; 31:127–43

66. Berliner JA, Navab M, Fogelman AM, et al. Athero-sclerosis: basic mechanisms. Oxidation, inflamma-tion, and genetics. Circulation 1995; 91:2488–96

67. Berliner JA, Heinecke JW. The role of oxidizedlipoproteins in atherogenesis. Free Rad Biol Med1996; 20:707–27

68. Steinberg D. Oxidative modification of LDL andatherogenesis. Circulation 1997; 95:1062–71

69. Belkner J, Wiesner R, Rathman J, et al. Oxygenationof lipoproteins by mammalian lipoxygenases. Eur JBiochem 1993; 213:251–61

70. Belkner J, Stender H, Kuhn H. The rabbit 15-lipoxy-genase preferentially oxygenates LDL cholesterolesters, and this reaction does not require vitamin E. JBiol Chem 1998; 273:23225–32

71. Xu W, Takahashi Y, Sakashita T, et al. Low densitylipoprotein receptor-related protein is required formacrophage-mediated oxidation of low densitylipoprotein by 12/15-lipoxygenase. J Biol Chem 2001;276:36454–9

72. Kuhn H, Belkner J, Zaiss S, et al. Involvement of 15-


ch18 14/7/05 4:48 pm Page 216

lipoxygenase in early stages of atherogenesis. J ExpMed 1994; 179:1903–11

73. Rankin SM, Parthasarathy S, Steinberg D. Evidencefor a dominant role of lipoxygenase(s) in the oxida-tion of LDL by mouse peritoneal macrophages. JLipid Res 1991; 32:449–56

74. Parthasarathy S, Wieland E, Steinberg D. A role forendothelial cell lipoxygenase in the oxidative modifi-cation of low density lipoprotein. Proc Natl Acad SciUSA 1989; 86:1046–50

75. Sigari F, Lee C, Witztum JL, et al. Fibroblasts thatoverexpress 15-lipoxygenase generate bioactive andminimally modified LDL. Arterioscler Thromb VascBiol 1997; 17:3639–45

76. Sparrow CP, Olszewski J. Cellular oxidative modifica-tion of low density lipoprotein does not require lipoxy-genases. Proc Natl Acad Sci USA 1992; 89:128–31

77. Jessup W, Darley-Usmar V, O’Leary V. et al. 5-lipoxy-genase is not essential in macrophage-mediatedoxidation of low-density lipoprotein. Biochem J 1991;278 (Pt 1): 163–9

78. Yla-Herttuala S, Rosenfeld ME, Parthasarathy S, et al.Colocalization of 15-lipoxygenase mRNA and proteinwith epitopes of oxidized low density lipoprotein inmacrophage-rich areas of atherosclerotic lesions.Proc Natl Acad Sci USA 1990; 87:6959–63

79. Folcik VA, Nivar-Aristy RA, Krajewski LP, et al.Lipoxygenase contributes to the oxidation of lipids inhuman atherosclerotic plaques. J Clin Invest 1995;96:504–10

80. Kuhn H, Heydeck D, Hugou I, et al. In vivo action of15-lipoxygenase in early stages of human atherogene-sis. J Clin Invest 1997; 99:888–93

81. Yla-Herttuala S, Luoma J, Viita H, et al. Transfer of 15-lipoxygenase gene into rabbit iliac arteries results inthe appearance of oxidation-specific lipid–proteinadducts characteristic of oxidized low density lipopro-tein. J Clin Invest 1995; 95: 2692–8

82. Libby P. Inflammation in atherosclerosis. Nature2002; 420:868–74

83. Libby P, Ridker PM, Maseri A. Inflammation andatherosclerosis. Circulation 2002; 105:1135–43

84. Steinberg D. Atherogenesis in perspective: hyper-cholesterolemia and inflammation as partners incrime. Nat Med 2002; 8:1211–17

85. Li AC, Glass CK. The macrophage foam cell as atarget for therapeutic intervention. Nat Med 2002;8:1235–42

86. Hansson GK. Regulation of immune mechanisms inatherosclerosis. Ann NY Acad Sci 2001; 947:157–65;discussion 165–6

87. Hansson GK. Inflammation and immune response inatherosclerosis. Curr Atheroscler Rep 1999; 1:150–5

88. Gouni-Berthold I, Sachinidis A. Does the coronaryrisk factor low density lipoprotein alter growth andsignaling in vascular smooth muscle cells? FASEB J2002; 16:1477–87

89. Huo Y, Hafezi-Moghadam A, Ley K. Role of vascularcell adhesion molecule-1 and fibronectin connecting

segment-1 in monocyte rolling and adhesion on earlyatherosclerotic lesions. Circ Res 2000; 87:153–9

90. Nageh MF, Sandberg ET, Marotti KR. Deficiency ofinflammatory cell adhesion molecules protectsagainst atherosclerosis in mice. Arterioscler ThrombVasc Biol 1997; 17:1517–20

91. Collins RG, Velji R, Guevara NV, et al. P-selectin orICAM-1 deficiency substantially protects against ather-osclerosis in apo E deficient mice. J Exp Med 2000;191:189–94

92. Dong ZM, Chapman SM, Brown AA, et al. Thecombined role of P- and E-selectins in atherosclerosis.J Clin Invest 1998; 102:145–52

93. Dong ZM, Brown AA, Wagner DD. Prominent role ofP-selectin in the development of advanced athero-sclerosis in ApoE-deficient mice. Circulation 2000;101:2290–5

94. Shih PT, Brennan ML, Vora DK. Blocking very lateantigen-4 integrin decreases leukocyte entry and fattystreak formation in mice fed an atherogenic diet. CircRes 1999; 84:345–51

95. Shih PT, Elices MJ, Fang ZT. Minimally modified low-density lipoprotein induces monocyte adhesion toendothelial connecting segment-1 by activating beta 1integrin. J Clin Invest 1999; 103:613–25

96. Huo Y, Ley K. Adhesion molecules and atherogenesis.Acta Physiol Scand 2001; 173:35–43

97. Reilly KB, Srinivasan S, Hatley ME, et al. 12/15 lipoxy-genase activity mediates inflammatory monocyte:endothelial interactions and atherosclerosis in vivo. JBiol Chem 2004; 279:9440–50

98. Hatley ME, Srinivasan S, Reilly KB, et al. Increasedproduction of 12/15 lipoxygenase eicosanoids accel-erates monocyte/endothelial interactions in diabeticdb/db mice. J Biol Chem 2003; 278:25369–75

99. Lee TS, Yenm HC, Pan CC, et al. The role of inter-leukin 12 in the development of atherosclerosis inapoE-deficient mice. Arterioscler Thromb Vasc Biol1999; 9: 734–42

100. Uyemura K, Demer LL, Castle SC, et al. Cross-regula-tory roles of interleukin (IL)-12 and IL-10 in athero-sclerosis. J Clin Invest 1996; 97:2130–8

101. Natarajan R, Reddy MA, Malik KU, et al. Signalingmechanism of nuclear factor-kappa B-mediated activa-tion of inflammatory genes by 13-hydroperoxy-octadecadienoic acid in cultured vascular smoothmuscle cells. Arterioscler Thromb Vasc Biol 2001;21:1408–13

102. Reddy MA, Kim YS, Lanting L, et al. Reduced growthfactor responses in vascular smooth muscle cellsderived from 12/15-lipoxygenase-deficient mice.Hypertension 2003; 41:1294–300

103. Samuelsson B. Leukotrienes: mediators of immediatehypersensitivity actions and inflammation. Science1983; 220:568–75

104. Lewis RA, Austen KF, Soberan RJ. Leukotrienes andother products of 5-lipoxygenase pathway. Bio-chemistry and relation to pathobiology in humandiseases. N Engl J Med 1990; 323:645–55


ch18 14/7/05 4:48 pm Page 217

105. De Caterina R, Mazzone A, Giannessi D, et al.Leukotriene B4 production in human atheroscleroticplaques. Biomed Biochim Acta 1988; 47:S182–5

106. Patrignani P, Daffonchio L, Hernandez A, et al.Release of contracting autacoids by aortae of normaland atherosclerotic rabbits. J Cardiovasc Pharmacol1992; 20(Suppl 12):S208–10

107. Subbarao K, Jala VR, Mathis S, et al. Role ofleukotriene B4 receptor in the development of ather-osclerosis: potential mechanisms. ArteriosclerThromb Vasc Biol 2004; 24:369–75

108. Aiello RJ, Bourassa PA, Lindsey S, et al. Leukotriene B4

receptor antagonism reduces monocytic foam cells inmice. Arterioscler Thromb Vasc Biol 2002; 22:443–9

109. Mehrabian M, Wong J, Wang X, et al. Genetic locus inmice that blocks development of atherosclerosisdespite extreme hyperlipidemia. Circ Res 2001;89:125–30

110. Welch CL, Bretschger S, Latib N, et al. Localization ofatherosclerosis susceptibility loci to chromosomes 4and 6 using the Ldlr knockout mouse model. ProcNatl Acad Sci USA 2001; 98:7946–51

111. Kuhn H, Anton M, Gerth C, et al. Amino acid differ-ences in the deduced 5-lipoxygenase sequence ofCAST atherosclerosis-resistance mice confer impairedactivity when introduced into the human ortholog.Arterioscler Thromb Vasc Biol 2003; 23:1072–6

112. Mehrabian M, Allayee H. 5-lipoxygenase and athero-sclerosis. Curr Opin Lipidol 2003; 14:447–57

113. Zhao L, Moos MP, Gräbner R, et al. The 5-lipoxyge-nase pathway promotes pathogenesis of hyperlipi-demia-dependent aortic aneurysm. Nat Med 2004; 10:966–73

114. Spanbroek R, Gräbner R, Lotzer K, et al. Expandingexpression of the 5-lipoxygenase pathway within thearterial wall during human atherogenesis. Proc NatlAcad Sci USA 2003; 100:1238–43

115. Cipollone F, Mezzetti A, Fazia M, et al. Identificationof 5-lipoxygenase as a major gene contributing toatherosclerotic plaque instability in humans.Circulation 2003; 108:IV–223

116. Dwyer JH, Allayee H, Dwyer KM, et al. Arachidonate5-lipoxygenase promoter genotype, dietary arachi-donic acid, and atherosclerosis. N Engl J Med 2004;350:29–37

117. Lee S, Felts KA, Parry GC, et al. Inhibition of 5-lipoxy-genase blocks IL-1 beta-induced vascular adhesionmolecule-1 gene expression in human endothelialcells. J Immunol 1997; 158:3401–7

118. Friedrich EB, Tager AM, Liu E, et al. Mechanisms ofleukotriene B4-triggered monocyte adhesion.Arterioscler Thromb Vasc Biol 2003; 23:1761–7

119. Gu L, Okada Y, Clinton SK, et al. Absence ofmonocyte chemoattractant protein-1 reduces athero-sclerosis in low density lipoprotein receptor-deficientmice. Mol Cell 1998; 2:275–81

120. Boring L, Gosling J, Cleary M, et al. Decreased lesionformation in CCR2–/– mice reveals a role forchemokines in the initiation of atherosclerosis.

Nature 1998; 394:894–7121. Matsukawa A, Hogaboam CM, Lukacs NW, et al.

Endogenous monocyte chemoattractant protein-1(MCP-1) protects mice in a model of acute septicperitonitis: cross-talk between MCP-1 and leukotrieneB4. J Immunol 1999; 163:6148–54

122. Rosenfeld ME. Leukocyte recruitment into developingatherosclerotic lesions: the complex interactionbetween multiple molecules keeps getting morecomplex. Arterioscler Thromb Vasc Biol 2002; 22:361–3

123. Sozzani S, Zhou D, Locati M, et al. Stimulatingproperties of 5-oxo-eicosanoids for human mono-cytes: synergism with monocyte chemotactic protein-1and -3. J Immunol 1996; 157:4664–71

124. Fabunmi RP, Sukhova GK, Sugiyama S, et al.Expression of tissue inhibitor of metalloproteinases-3in human atheroma and regulation in lesion-associ-ated cells: a potential protective mechanism in plaquestability. Circ Res 1998; 83:270–8

125. Rajagopalan S, Meng XP, Ramasamy S, et al. Reactiveoxygen species produced by macrophage-derivedfoam cells regulate the activity of vascular matrixmetalloproteinases in vitro. Implications for athero-sclerotic plaque stability. J Clin Invest 1996; 98:2572–9

126. Taylor PM, Woodfield RJ, Hodgkin MN, et al. Breastcancer cell-derived EMMPRIN stimulates fibroblastMMP2 release through a phospholipase A(2) and 5-lipoxygenase catalyzed pathway. Oncogene 2002;21:5765–72

127. Leppert D, Hauser SL, Kishiyama JL, et al.Stimulation of matrix metalloproteinase-dependentmigration of T cells by eicosanoids. FASEB J 1995;9:1473–81

128. Cornicelli JA, Trivedi BK. 15-Lipoxygenase and itsinhibition: a novel therapeutic target for vasculardisease. Curr Pharm Des 1999; 5:11–20

129. Schewe T, Sadik C, Klotz LO, et al. Polyphenols ofcocoa: inhibition of mammalian 15-lipoxygenase. BiolChem 2001; 382:1687–96

130. Sadik CD, Sies H, Schewe T. Inhibition of 15-lipoxy-genases by flavonoids: structure–activity relations andmode of action. Biochem Pharmacol 2003; 65:773–81

131. Koshihara Y, Neichi T, Murota S, et al. Caffeic acid isa selective inhibitor for leukotriene biosynthesis.Biochim Biophys Acta 1984; 792: 92–7

132. Evans JF, Leville C, Mancini JA, et al. 5-Lipoxygenase-activating protein is the target of a quinoline class ofleukotriene synthesis inhibitors. Mol Pharmacol 1991;40:22–7

133. Dupont R, Goossens JF, Cotelle N, et al. New bis-catechols 5-lipoxygenase inhibitors. Bioorg MedChem 2001; 9:229–35

134. Israel E, Rubin P, Kemp JP, et al. The effect of inhibi-tion of 5-lipoxygenase by zileuton in mild-to-moder-ate asthma. Ann Intern Med 1993; 119:1059–66

135. Busse WW, McGill KA, Horwitz RJ. Leukotrienepathway inhibitors in asthma and chronic obstructivepulmonary disease. Clin Exp Allergy 1999; 29(Suppl2):110–15


ch18 14/7/05 4:48 pm Page 218

INTRODUCTION

Atherosclerosis, the major cause of coronaryheart disease (CHD), has been proposed tobegin during early childhood and to progressin a non-linear fashion throughout adulthood.The aetiology of atherogenesis can be bestdescribed as a chronic inflammatory disease ofthe arterial intima, dominated by a leukocyteinfiltrate comprising predominately T-lympho-cytes and monocyte-derived macrophages. Thischapter will focus on the putative role ofsecreted phospholipase A2 (sPLA2) isozymes1,2

in atherosclerosis, with a particular emphasison lipoprotein-associated phospholipase A2

(Lp-PLA2), also known as plasma platelet-activating factor acetylhydrolase (PAF-AH) ortype VIIA PLA2.

BACKGROUND

Although both secreted and cytosolic enzymescomprise the growing PLA2 superfamily,1,2 onlythe sPLA2 sub-group will be considered in thischapter because an acknowledged criticalprocess in the pathophysiology of atherosclero-sis is the extracellular metabolism of lipopro-teins. In order to evaluate fully the potentialrole of the sPLA2 family in atherogenesis it isfirst necessary to summarize the currentconcepts concerning the pathogenesis of ather-osclerosis as well as list the cell types involved.

Epidemiological, clinical and genetic studiesindicate that elevated plasma levels of lowdensity lipoprotein (LDL) greatly increase the

risk for atherosclerosis.3 Furthermore, it isgenerally accepted that a key pathogenic eventis the retention4 and subsequent oxidation5,6 ofapoB-rich lipoproteins (in particular, LDL)within the arterial wall. Whilst proteoglycans7

appear to play an important role in promotinglipoprotein retention, the precise mechanismsby which LDL becomes oxidatively modifiedremain rather unclear. Evidence, however, isaccumulating that supports a major role formyeloperoxidase in promoting lipid oxidationin human, but not mouse, atheroscleroticlesion development.8,9 Suffice it to say thatvascular cells, including all constituents of theatherosclerotic plaque, can produce and usereactive oxygen species (ROS)10,11 whichultimately can oxidize LDL. Furthermore, thereis mounting epidemiological evidence thatlesion development, as well as lesion stability,may be significantly influenced by infection.12

Induction of inflammation is an importantcomponent in the defence against micro-organ-isms and bacterial lipopolysaccharides are wellknown stimuli for enhanced ROS generation,as is any type of systemic inflammation13 regard-less of cause. Thus, the precise processes whichmodulate lipid peroxidation of LDL may varyaccording to the stage of lesion development aswell as the presence or absence of an underly-ing infection or systemic inflammation.

Once modified, oxidized LDL (oxLDL) canpromote a plethora of pro-inflammatoryatherogenic effects, influencing all cell typesthat comprise the plaque, i.e. monocyte-derived macrophages, T lymphocytes, endothe-lial and smooth muscle cells.5,6,14–16 Oxidized

Role of secretory phospholipase A2

isozymes (Lp-PLA2)C.H. Macphee

19

ch19 14/7/05 4:49 pm Page 219

LDL, for example, has been demonstrated toinduce diverse effects including endothelialdysfunction17 and cell death,18,19 two clinicallyrelevant activities that could help develop apro-thrombotic state and promote plaqueinstability, respectively. One of the earliestevents following LDL oxidation is the PLA2-dependent hydrolysis of the oxidized phospho-lipids generating lysophosphatidylcholine(lyso-PC) and oxidized fatty acids,20,21 both ofwhich have been shown to be causal agents ininflammation. Indeed, it has been demon-strated that the concentration of lyso-PC issignificantly elevated in atherosclerotic arter-ies.22,23 With this albeit brief and somewhatsimplified overview of atherogenesis we cannow assess which, if any, of the various isozymesof sPLA2 could be causative factors in influenc-ing human disease progression and giving riseto major adverse coronary events. Specificattention, therefore, will be focused on whichsPLA2s fit the following biological profile:participate in lipoprotein metabolism(especially oxLDL), are upregulated in athero-sclerotic lesions, contribute to inflamma-tion/host defence and are positively associatedwith cardiovascular disease and/or events.

sPLA2 ISOZYMES ANDATHEROGENESIS

PLA2s are an ubiquitous class of enzymes thathydrolyse the sn-2 acyl bond of phospholipids ofcell membranes and lipoproteins to yield freefatty acids (FFAs) and lysophospholipids.1,2 Bothof these products are themselves pro-inflamma-tory but are also precursors of other pro-inflam-matory lipid mediators such as leukotrienes,eicosanoids, prostaglandins and platelet-activat-ing factor.24 In general, the mammalian sPLA2s(groups IB, IIA,C–F, II, V, X, XII) have lowmolecular masses (13–19 kDa), require calciumfor activity and lack specificity for arachidonate-containing phospholipids. Whilst the biologicalfunctions of these sPLA2s are still under investi-gation, it appears that a primary action is tomediate the second larger wave of arachidonicacid release for subsequent eicosanoid forma-

tion, which is somewhat surprising consideringthat they are not arachidonyl-sensitiveenzymes.24 These phospholipases contain aHis/Asp dyad active site and act optimally onaggregated substrates such as micelles ormembranes, a phenomenon termed ‘interfacialactivation.’ Lp-PLA2 (i.e. type VIIA), a 45 kDacalcium-independent sPLA2, on the otherhand, is not an interfacial enzyme but utilizes adistinct Ser/His/Asp triad active site thataccesses its substrates from the aqueous phase.25

This property of Lp-PLA2 enables a broadsubstrate specificity that is governed primarilyby aqueous phase solubility.26 Although Lp-PLA2 was discovered due to its ability to hydrol-yse platelet-activating factor,27 its mechanism ofaction raises the distinct possibility that it canhydrolyse a wide variety of physiologicalsubstrates, including phospholipids containingnon-truncated oxidatively modified (i.e. polar)polyunsaturated fatty acids.

Table 1 summarizes key functions of the 11known mammalian extracellular sPLA2s as theypertain to atherogenesis, namely involvementin LDL oxidation, expression in disease andwhether evidence exists for any to be a positiverisk factor for coronary heart disease inhumans. Although data for some of the sPLA2sare incomplete, Table 1 clearly shows that twoof these extracellular enzymes, Lp-PLA2 andtype IIA, qualify as potential causative factors inthe promotion of atherosclerosis. Thus, the restof this chapter will focus entirely on discussingand comparing the evidence for these twoenzymes. It must be noted that, although sPLA2-V and -X are not yet sufficiently charac-terized to include in a full analysis, both holdsome potential to be linked with atherosclerosisdue to their leukocyte expression and ability tohydrolyse anionic phospholipids.28

Lp-PLA2 (PAF-AH) AND sPLA2-IIA

Plaque expression and oxidation of LDL

Lp-PLA2

Leukocytes,29 and in particular monocytes,macrophages and T-cells,30–32 appear to be the


ch19 14/7/05 4:49 pm Page 220

only source of circulating Lp-PLA2. Given thatthese cells represent the exact same leukocytepopulation that is intimately involved inatherogenesis,6 it comes as no surprise to notethat the enzyme is greatly upregulated in ather-osclerotic lesions.32 The observation thatlipopolysaccharide (LPS) represents the mostpowerful natural stimulus for enhancedexpression and secretion implies a major roleof Lp-PLA2 for this enzyme in host defence andthe acute phase response.31,33 Tumour necrosisfactor and interleukin-1 are key cytokines inbacterial infection and, interestingly, both havebeen reported to increase plasma Lp-PLA2

activity, but only modestly when compared withLPS, suggesting that they may partly mediatethe effect of LPS.33 Thus, Lp-PLA2 can beviewed as one of the acute-phase proteins.

Lp-PLA2 is associated predominantly withLDL in human plasma, with the remainder(~20%) distributed across high density lipopro-tein (HDL) and very low density lipoprotein(VLDL).34 It has been reported that Lp-PLA2

binds preferentially to the highly atherogenicsmall dense LDL35 and is found enriched inpro-inflammatory electronegative LDL parti-cles36 as well as lipoprotein(a).37 The criticaldeterminant in lipoprotein binding appears tobe a specific interaction between two domainson human Lp-PLA2 and the carboxy terminusof apolipoprotein B-100 on LDL.38 Moreover, ithas been suggested that the catalytic properties

of Lp-PLA2 are influenced by its lipoproteinenvironment.39,40 Since Lp-PLA2 accesses itssubstrate from the aqueous phase,26 caution isadvised when only activity measurements areutilized to determine the contribution of Lp-PLA2 across various lipoprotein fractions. Thisis because the catalytic rate will vary withchanges in the concentrations of components,such as lipids and proteins, that can bindsubstrate and alter its aqueous phase concen-tration. Thus, both Lp-PLA2 activity and massmeasurements are required before a definitiveconclusion can be made on the contributionsof this enzyme across different lipoproteinfractions.

The association of the majority of plasma Lp-PLA2 with apoB-containing lipoproteins inhuman plasma is highly significant because Lp-PLA2 remains latent until LDL undergoesoxidative modification. Lp-PLA2 is solelyresponsible for a well-known consequence ofLDL oxidation, the rapid hydrolysis of oxidizedphosphatidylcholine moieties within modifiedLDL.41–43 This generates biologically relevantquantities of lyso-PC and oxidized free fattyacids (ox-FFAs). Both of these lipid productshave been demonstrated to be causal agents ininflammation and atherosclerosis.43 This isespecially true for lyso-PC, whose concentra-tion is significantly elevated in plaques.23 Forexample, in vitro studies of lyso-PC demon-strate that it has the following pro-atherogenic

ROLE OF SECRETORY PHOSPHOLIPASE A2 ISOZYMES (Lp-PLA2) 221

Table 1 Secreted mammalian PLA2 enzymes

sPLA2 type Other name Catalytic site Involved during Upregulated in human Levels associated LDL oxidation atherosclerotic plaque with risk for CHD

IB Pancreatic His/Asp dyad No No NoIIA Synovial His/Asp dyad Yes Yes YesIIC His/Asp dyad No No NoIID His/Asp dyad No No NoIIE His/Asp dyad No No NoIIF His/Asp dyad No No NoIII His/Asp dyad No No NoV His/Asp dyad ? ? ?VIIA Lp-PLA2 Ser/His/Asp triad Yes Yes (macrophages and T-cells) Yes

PAF-AHX His/Asp dyad ? Yes (macrophages) ?XII His/Asp dyad No No No

ch19 14/7/05 4:49 pm Page 221

activities: chemoattractant for monocytes andT-cells, induces the elaboration of chemokinesand adhesion molecules, upregulates therelease of MPO, stimulates superoxide produc-tion, impairs endothelial-dependent relax-ation, inhibits endothelial migration afterinjury and can cause cell death.43–49 It thereforefollows that inhibition of Lp-PLA2 couldprovide a novel approach in the treatment ofatherosclerosis through prevention of thegeneration of two bioactive lipid mediators.

Support for this concept comes from theobservation that inhibition of Lp-PLA2 not onlyabolished the increase in lyso-PC during oxida-tion of LDL, but prevented the generation ofoxidized non-esterified fatty acid moities thatwere very effective human monocyte chemo-attractants.43 In addition, related experimentshave demonstrated that inhibition of Lp-PLA2

can significantly diminish the cytotoxic andapoptosis-inducing effects of oxidized LDL onhuman macrophages.50 This finding may berelevant to the death of macrophage foam cellsand smooth muscle cells known to occur inatherosclerotic lesions and associated withprogression to a state vulnerable to rupture.51

Although these findings support a pro-inflammatory role for Lp-PLA2 in athero-genesis, a body of evidence suggests theopposite.52–54 This view arose primarily becauseof the ability of Lp-PLA2 (i.e. PAF-AH) todegrade and inactivate exogenously addedPAF, a molecule attributed with pro-inflamma-tory and pro-thrombotic properties. While therole of PAF in inflammation has been inten-sively studied by numerous researchers formany years, the demonstration of its role inpathological conditions remains elusive. Forinstance, although PAF was thought to have apotential significant role in the pathogenesis ofasthma, this has been questioned due to lack ofclinical efficacy of a variety of structurallydiverse PAF receptor antagonists.55

Furthermore, it remains to be determinedwhether Lp-PLA2 actually plays a significantrole in the metabolism of endogenously gener-ated PAF.

This proposed anti-inflammatory role for Lp-PLA2 was subsequently used to explain, at least

in part, the anti-atherogenic properties ofHDL, even though the majority (~80%) of theenzyme in human plasma is located on anotherlipoprotein, LDL.34 The situation is very differ-ent in rodents and rabbits, where the vastmajority of plasma Lp-PLA2 is found associatedwith HDL.52 The hypothesis in this instance isbased upon the premise that the oxidizedphospholipids represent the inflammatorymediators and not the two aqueous-phasesoluble products, lyso-PC and oxidized fattyacids.52,56 This concept raises several paradoxes,even excluding the fact that the enzyme isfound primarily on LDL in humans. First,although it is generally accepted that Lp-PLA2

is solely responsible for the generation of lyso-PC during the oxidation of LDL, the sameenzyme appears to play only a minor role in thegeneration of lyso-PC that occurs during theoxidation of human HDL.43 Second, whereasother sPLA2s are consistently viewed as pro-inflammatory due to their ability to generatelyso-PC and FFAs, this link is somehow ignoredor forgotten for Lp-PLA2. Third, recent studieshave demonstrated that oxidized phosphatidyl-cholines can inhibit ligand activation of toll-like receptors57,58 and upregulate theexpression of haem oxygenase-1,59 both ofwhich provide protection against inflamma-tion. The authors propose that these lipidoxidation products can function as a feedbackmechanism to limit inflammation and associ-ated tissue damage, especially when inconjunction with oxidative stress. Interestingly,a polymorphism in the toll-like receptor 4known to attenuate receptor signalling hasbeen associated with a decreased risk of ather-osclerosis.60,61

Support for an anti-inflammatory role forLp-PLA2 does exist from various pre-clinicalstudies which showed that recombinanthuman Lp-PLA2 could prevent or attenuatepathological inflammation in a number ofanimal models.52 Similar efficacy, however,could not be replicated in the clinic, the laststudy being in individuals with sepsis.62 Theconsistent lack of clinical efficacy has resultedin the termination of this particularapproach.


ch19 14/7/05 4:49 pm Page 222

sPLA2- I IA

The plasma concentration of group IIA PLA2

can increase many fold during the acute phaseso, like Lp-PLA2, it can be viewed as anothermember of the acute-phase proteins. It follows,therefore, that large amounts of sPLA2-IIA aredetected in the exudating fluids and plasma ofpatients with various systemic and local inflam-matory diseases.2 This particular sPLA2 has abroad cellular distribution and, unlike Lp-PLA2, its expression is not limited to leuko-cytes. Inflammatory effector cells do, however,store sPLA2-IIA in their secretory granules andrelease it promptly following cell activation.2,63

In normal arteries sPLA2-IIA is mainly foundwith smooth muscle cells of the media, whereasin human atherosclerotic plaques it has beendetected in some, but not all, macrophage-derived foam cells.63,64 A similar finding wasobserved in intimal lesions from WHHLrabbits, where sPLA2-IIA was expressed in asmall population of foam cells.65

It has been demonstrated that the phospho-lipids of native LDL as well as high densitylipoprotein (HDL) can be hydrolysed byhuman sPLA2-IIA.63,66,67 Thus, unlike Lp-PLA2,sPLA2-IIA is not dependent upon LDL becom-ing oxidized to enable cleavage of phospho-lipids. Moreover, sPLA2-IIA-modified LDLshows increased affinity for proteoglycans, anincreased propensity to aggregate and anenhanced ability to deliver cholesterol tocells.66 Interestingly, the modification of HDLby sPLA2-IIA treatment also resulted inenhanced lipid deposition following incuba-tion with macrophages, endothelial andhepatoma cells. The experimental dataindicate that all sPLA2-IIA-treated lipoproteinspossess an enhanced ability to transfer choles-terol to both vascular and non-vascular tissues.Thus, sPLA2-IIA clearly has the potential toalter lipoprotein metabolism in a broad sense,but its ability to influence the hydrolysis ofapoB-containing lipoproteins within the arter-ial intima is highly questionable. This is due toits apparent inability to act on phosphatidyl-choline,67,68 the major phospholipid of bothlipoproteins and cell membranes. Recall that

the concentration of lyso-PC is greatly elevatedin atherosclerotic plaques.23

Thus, it would appear that Lp-PLA2 is betterplaced to contribute to the hydrolysis ofoxidized LDL within the developing athero-sclerotic lesion, whilst sPLA2-IIA may still play apart by enhancing the intimal retention ofapoB-containing lipoproteins. As noted earlier,the group V and X secretory PLA2s mayultimately evolve into stronger candidates sinceboth are efficient in binding and hydrolysingphosphatidylcholines.28 They, like sPLA2-IIA,are not dependent upon oxidative modifica-tion for substrate generation, a characteristicthat remains unique to Lp-PLA2. Both,however, are very efficient in the hydrolysis ofboth LDL and HDL.65,67 Recent findings, forexample, suggest that the lipolytic modifica-tion of HDL by both group V and X sPLA2s canlead to a reduction of its anti-atherogenicfunctions, including an impaired ability tomediate cellular cholesterol efflux.65

Association with coronary heart disease

Lp-PLA2

This section will be split into two parts, the firstdealing with plasma levels of Lp-PLA2 and thesecond concentrating on known geneticpolymorphisms. Since Lp-PLA2 is responsiblefor approximately 95% of the hydrolysis ofexogenously added PAF (when used at around50 µM) in human plasma, this method repre-sents a straightforward means of comparingcirculating levels in controls versus patients.The first studies to use this method during thelate 1980s consistently demonstrated a positiveassociation with cardiovascular disease.Documented elevations in plasma Lp-PLA2

levels were noted in individuals with athero-sclerosis,69–71 diabetes,72 stroke73,74 and evenessential hypertension.75 More recently, theseobservations have been confirmed andextended using both Lp-PLA2 activity and mass(determined via a specific immunoassay)measurements.

Blankenberg et al showed increasing plasmaLp-PLA2 activity levels in both male and female


ch19 14/7/05 4:49 pm Page 223

patients with stable and unstable coronaryartery disease as compared with those withoutangiographically confirmed coronary arterydisease (CAD).76 Interestingly, no correlationcould be demonstrated between Lp-PLA2 activ-ity and inflammatory markers, including otheracute-phase reactants such as C-reactiveprotein (CRP). This latter observationconfirmed the findings of Packard et al, whoearlier used the Lp-PLA2 immunoassay for thefirst time to show a strong positive associationwith the risk of CAD (defined as myocardialinfarction, revascularization or death) in anested case-control study involving individualswho participated in the West of ScotlandCoronary Prevention Study (WOSCOPS).77 Lp-PLA2 mass was found to be an independent riskmarker, with a 2-fold increased risk associatedwith Lp-PLA2 levels in the highest quintilecompared with the lowest quintile. Thisrelationship was not influenced by markers ofinflammation such as CRP and white cellcount, or other traditional risk factors such asfibrinogen and lipid parameters includingLDL cholesterol. An earlier study by the samegroup had demonstrated significantly higherLp-PLA2 levels in men with angiographicallyproven CAD compared with age-matchedcontrols, independent of LDL- and HDL-cholesterol, smoking and systolic bloodpressure.78 The independence of Lp-PLA2

when compared with CRP is of note becauseboth are acute-phase proteins, but each appar-ently represents different inflammatorypathways. Based on the weight of scientificevidence, the American Heart Associationrecently stated that measurement of CRP couldbe used to help further evaluation and therapyin the primary prevention of cardiovasculardisease in patients at intermediate risk.79

The idea that Lp-PLA2 and CRP could becomplementary in identifying individuals athigh CAD risk found further support from avery recent case-control analysis of men andwomen who participated in the AtherosclerosisRisk in Communities (ARIC) study. Forindividuals in ARIC with LDL-cholesterolbelow 130 mg/dl, Lp-PLA2 and CRP were bothsignificantly and independently associated with

CAD in fully adjusted models.80 As inWOSCOPS, Lp-PLA2 mass was not correlatedwith CRP levels. However, although Lp-PLA2

levels were significantly higher in cases thannon-cases in the complete cohort, in a modeladjusted for traditional risk factors, and inparticular LDL-cholesterol, the association wasattenuated. A similar attenuation was alsodemonstrated in another, albeit much smaller,prospective study of apparently healthy middle-aged women.81 Thus, although Lp-PLA2 levelsare consistently elevated in patients at risk ofCHD, many more studies are needed toconfirm that Lp-PLA2 levels represent a strongindependent predictor of future cardiovascularrisk. The studies do acknowledge that there is asignificant positive correlation of Lp-PLA2 withLDL cholesterol (its main carrier in humanplasma), a significant inverse correlation withHDL-cholesterol and a lack of any correlationbetween Lp-PLA2 and CRP.

Whilst there is growing evidence that plasmalevels of Lp-PLA2 may represent an indepen-dent predictor of coronary heart disease, aquite different picture emerges from studies ofa loss of function mutation in Lp-PLA2, whichexists primarily in the Japanese population.52,82

This single nucleotide polymorphism (V279F)is found in around 30% of the Japanesepopulation (4% homozygous, 27% heterozy-gous), with plasma from homozygotes demon-strating a complete inability to hydrolyseexogenously added PAF. The inactivatingmissense mutation was first reported in 1988 byMiwa,83 who stated that there were no signifi-cant differences between the Lp-PLA2 enzymeactivities of patients with and without asthmaticattacks. The next group to comment on themutation concluded that none of the subjectswith the enzyme deficiency had a history ofallergy, circulatory shock or chronic inflamma-tory diseases.84 A similar observation was madeby others who showed that the mutant alleledid not correlate with asthma prevalence, typeor severity.85 Another manuscript published inthe same year, however, concluded that themutation was a modulating locus for the sever-ity of asthma.86 A similarly confusing pictureemerges when considering the association of


ch19 14/7/05 4:49 pm Page 224

the mutation with CAD risk, because thefindings tend to be completely the opposite towhat the growing epidemiology around plasmaLp-PLA2 levels is indicating.

The first CAD association study, and largestto-date, reported a positive association with theLp-PLA2 missense mutation in male, but notfemale, patients with diagnosed myocardialinfarction.87,88 A later study by the same group,however, failed to confirm the previouslyobserved association between genotype andCAD.89 Another publication showed a statisti-cally significant link with the Lp-PLA2 mutationand stroke patient.90 Intriguingly, the samegroup had previously demonstrated thatplasma Lp-PLA2 activities were significantlyelevated in stroke.73,74 However, there is someconsistency to this apparent paradox because,in the genetic association paper, it was shownthat in normal and heterozygous subjects, Lp-PLA2 activity was higher in stroke cases ascompared to controls. In three other studies,with various cardiovascular phenotypes, thisparadoxical pattern was also observed.91–93

Moreover, in the three publications by Yamadaand colleagues which showed the frequency ofthe 279F allele to be significantly higher inCAD cases as compared to controls, the Lp-PLA2 activity in non- or heterozygous carrierswas comparable.88,94,95

If low or absent Lp-PLA2 levels are trulypredisposing for an increased risk for cardio-vascular disease, then subjects with disease whodo not carry the 279F allele should also havesignificantly lower Lp-PLA2 activity ascompared to controls. Furthermore, it isnoteworthy that the generation of lyso-PC andwater-soluble products (i.e. oxidized fattyacids) from oxidized phospholipids was similarin plasma samples from both normals andindividuals heterozygous for the nullmutation.96 Homozygote subjects, on the otherhand, demonstrated a significant reduction inthese Lp-PLA2 products. These findingsindicate that functional differences can only beobserved between homozygotes and normals,which is not consistent with the publisheddisease associations in heterozygotes.

Thus, the data generated in Japanesesubjects are conflicting and more comprehen-sive studies are needed to resolve these discrep-ancies. This need was further highlighted by arecent investigation of another Lp-PLA2

polymorphism (Val379) that is also associatedwith reduced enzyme activity, which demon-strated a reduced risk of myocardial infarctionin a Caucasian population.97

sPLA2- I IA

Only one study has investigated the associationof sPLA2-IIA and CAD risk and found that itwas a significant risk marker for the presenceof CAD and was able to predict clinicalcoronary events independently of other riskfactors.98 Unlike Lp-PLA2, sPLA2-IIA levels weresignificantly correlated with CRP levels,suggesting they are common components ofthe same inflammatory pathway. In a multi-variate analysis, sPLA2-IIA, but not CRP,remained a significant predictor of futurecoronary events, although this requires confir-mation in many additional studies coveringdiverse population groups. Currently no dataexist linking an sPLA2-IIA polymorphism withCAD risk.

SUMMARY

The weight of biochemical evidence indicatesthat Lp-PLA2, rather than sPLA2-IIA, has thegreater potential to be a causal player inatherogenesis. Although further research isobviously required, especially to addressqueries around genetic associations, the realtest will be to demonstrate clinical benefitthrough the use of a small-molecule inhibitorof the enzyme. To this end, a highly potent andselective inhibitor of Lp-PLA2

99 is currently inphase II clinical studies. If effective, this thera-peutic approach of directly modifying plaqueinflammatory processes will complement LDLand HDL interventions for reduction of cardio-vascular risk associated with atheroscleroticvascular disease.


ch19 14/7/05 4:49 pm Page 225

1. Six DA, Dennis EA. The expanding superfamily ofphospholipase A2 enzymes, classification and charac-terisation. Biochim Biophys Acta 2000; 1488:1–19

2. Kudo I, Murakami M. Phospholipase A2 enzymes.Prost Other Lipid Mediat 2002; 68–69:3–58

3. Brown MS, Goldstein JL. Koch’s postulates for choles-terol. Cell 1992; 71:187–8

4. Williams KJ, Tabas I. The response-to-retentionhypothesis of early atherogenesis. Arteriol ThrombVasc Biol 1995; 15:551–61

5. Berliner JA, Heinecke JW. The role of oxidisedlipoproteins in atherogenesis. Free Radic Biol Med1996; 20:707–27

6. Ross R. Atherosclerosis – an inflammatory disease. NEngl J Med 1999; 340:115–26

7. Skalen K, Gustafsson M, Rydberg EK, et al.Subendothelial retention of atherogenic lipoproteinsin early atherosclerosis. Nature 2002; 417:750–4

8. Heinecke JW. Oxidised amino acids: culprits inhuman atherosclerosis and indicators of oxidativestress. Free Rad Biol Med 2002; 32:1090–1101

9. Brennan ML, Penn MS, Van Lente F, et al. Prognosticvalue of myeloperoxidase in patients with chest pain.N Engl J Med 2003; 349:1595–604.

10. Harrison D, Griendling KK, Landmesser U, et al Roleof oxidative stress in atherosclerosis. Am J Cardiol2003; 91:7A–11A

11. Griendling KK, FitzGerald GA. Oxidative stress andcardiovascular injury: animal and human studies.Circulation 2003; 108:2034–40

12. Kiechl S, Egger G, Mayr M, et al. Chronic infectionsand the risk of carotid atherosclerosis. Circulation2001; 103:1064–70

13. Buffon A, Biasucci LM, Liuzzo G, et al. Widespreadcoronary inflammation in unstable angina. N Engl JMed 2002; 347:5–12

14. Holvoet P. Oxidation of low density lipoproteins inatherosclerosis and thrombosis: inhibitory effects ofantioxoidants, estrogen and high density lipopro-teins. Vasc Dis 1997; 2:499–509

15. Dart AM, Chin-Dusting JPF. Lipids and the endothe-lium. Cardiovasc Res 1999; 43:308–22

16. Leitinger N. Oxidised phospholipids as modulators ofinflammation in atherosclerosis. Curr Opin Lipidol2003; 14:421–30

17. Cai H, Harrison DG. Endothelial dysfunction incardiovascular diseases: the role of oxidant stress. CircRes 2000; 87:840–4

18. Martinet W, Kockx MM. Apoptosis in atherosclerosis:focus on oxidised lipids and inflammation. Curr OpinLipidol 2001; 12: 535–41

19. Littlewood TD, Bennett MR. Apoptotic cell death inatherosclerosis. Curr Opin Lipidol 2003; 14:469–75

20. Parthasarathy S, Steinbrecher UP, Barnett J, et al.Essential role of a phospholipase A2 activity in

endothelial cell-induced modification of low densitylipoprotein. Proc Natl Acad Sci USA 1985; 82:3000–4

21. Quinn MT, Parthasarathy S, Steinberg D.Lysophosphatidylcholine: a chemotactic factor forhuman monocytes and its potential role in athero-genesis. Proc Natl Acad Sci USA 1988; 85:2805–9

22. Keaney JF, Xu A, Cunningham D, et al. Dietaryprobucol preserves endothelial function in choles-terol-fed rabbits by limiting vascular oxidative stressand superoxide generation. J Clin Invest 1995;95:2520–9

23. Thukkani AK, McHowat J, Hsu FF, et al. Identificationof α-chloro fatty aldehydes and unsaturated lysophos-phatidylcholine molecular species in human athero-sclerotic lesions. Circulation 2003; 108:3128–33

24. Balsinde J, Winstead MV, Dennis EA. PhospholipaseA2 regulation of arachidonic acid mobilisation. FEBSLett 2002; 531:2–6

25. Min JH, Jain MK, Wilder C, et al. Membrane-boundplasma platelet activating factor acetylhydrolase actson substrate in the aqueous phase. Biochemistry1999; 38:12935–42

26. Min JH, Wilder C, Aoki J, et al. Platelet-activatingfactor acetylhydrolase: broad substrate specificity andlipoprotein binding does not modulate the catalyticproperties of the plasma enzyme. Biochemistry 2001;40:4539–49

27. Farr RS, Cox CP, Wardlow ML, et al. Preliminarystudies of an acid-labile factor in human sera thatinactivates platelet-activating factor. Clin ImmunolImmunopathol 1980; 15:318–30

28. Murakami M, Kudo I. New phospholipase A2 isozymeswith a potential role in atherosclerosis. Curr OpinLipidol 2003; 14: 431–6

29. Asano K, Okamoto S, Fukunaga K, et al. Cellularsource(s) of platelet-activating factor acetylhydrolaseactivity in plasma. Biochem Biophys Res Commun1999; 261:511–14

30. Elstad MR, Stafforini DM, McIntyre TM, et al. Platelet-activating factor acetylhydrolase increases duringmacrophage differentiation. J Biol Chem 1989; 264:8467–70

31. Howard KM, Olson MS. The expression and localiza-tion of plasma platelet-activating factor acetylhydro-lase in endotoxemic rats. J Biol Chem 2000; 275:19891–6

32. Hakkinen T, Luoma J, Macphee CH, et al. Lipo-protein-associated phospholipase A2 is expressed bymacrophages in human and rabbit atheroscleroticlesions. Arterioscler Thromb Vasc Biol 1999; 19:2909–17

33. Memon RA, Fuller J, Moser AH, et al. In vivo regula-tion of plasma platelet-activating factor acetylhydro-lase during the acute phase response. Am J Physiol1999; 277:R94–R103


References

ch19 14/7/05 4:49 pm Page 226

34. Caslake MJ, Packard CJ, Suckling KE, et al.Lipoprotein-associated phospholipase A2, platelet-activating factor acetylhydrolase: a potential new riskfactor for coronary artery disease. Atherosclerosis2000; 150:413–19

35. Tselepis AD, Dentan C, Karabina S-AP, et al. PAF-degrading acetylhydrolase is preferentially associatedwith dense LDL and VHDL-1 in human plasma.Catalytic characteristics and relation to the monocyte-derived enzyme. Arterioscler Thromb Vasc Biol 1995;15:1764–73

36. Benitez S, Sanchez-Quesada JL, Ribas V, et al. Platelet-activating factor acetylhydrolase is mainly associatedwith electronegative low-density lipoprotein subfrac-tion. Circulation 2003; 108:92–6

37. Blencowe C, Hermetter A, Kostner GM, et al.Enhanced association of platelet-activating factoracetylhydrolase with lipoprotein (a) in comparisonwith low density lipoprotein. J Biol Chem 1995;270:31151–7

38. Stafforini DM, Tjoelker LW, McCormick SPA, et al.Molecular basis of the interaction between platelet-activating factor acetylhydrolase and low densitylipoprotein. J Biol Chem 1999; 274:7018–24

39. Stafforini DM, McIntyre TM, Carter ME, et al. Humanplasma platelet-activating factor acetylhydrolase:association with lipoprotein particles and role in thedegradation of platelet-activating factor. J Biol Chem1987; 262:4215–22

40. McCall MR, La Belle M, Forte TM, et al. Dissociableand nondissociable forms of platelet-activating factoracetylhydrolase in human plasma LDL: implicationsfor LDL oxidative susceptibility. Biochim BiophysActa 1999; 1437:23–36

41. Steinbrecher UP, Pritchard PH. Hydrolysis ofphosphatidylcholine during LDL oxidation ismediated by platelet-activating factor acetylhydrolase.J Lipid Res 1989; 30:305–15

42. Tew DG, Southan C, Rice SQJ, et al. Purification,properties, sequencing and cloning of a lipoprotein-associated, serine-dependent phospholipase involvedin the oxidative modification of low density lipopro-teins. Arterioscler Thromb Vasc Biol 1996; 16:591–9

43. Macphee CH, Moores KE, Boyd HF, et al.Lipoprotein-associated phospholipase A2, platelet-activating factor acetylhydrolase, generates two bio-active products during the oxidation of low densitylipoprotein: use of a novel inhibitor. Biochem J 1999;338:479–87

44. Kabarowski JHS, Xu Y, Witte ON. Lysophosphatidyl-choline as a ligand for immunoregulation. BiochemPharmacol 2002; 64:161–7

45. Murugesan G, Rani MRS, Gerber CE, et al. Lyso-phosphatidylcholine regulates human microvascularendothelial cell expression of chemokines. J Mol CellCardiol 2003; 35:1375–84

46. Rong JX, Berman JW, Taubman MB, et al.Lysophosphatidylcholine stimulates monocytechemoattractant protein-1 gene expression in rat

aortic smooth muscle cells. Arterioscler Thromb VascBiol 2002; 22:1617–23

47. Silliman CC, Elzi DJ, Ambruso DR, et al.Lysophosphatidylcholines prime the NADPH oxidaseand stimulate multiple neutrophil functions throughchanges in cytosolic calcium. J Leukocyte Biol 2003;73:511–24

48. Lum H, Qiao J, Walter RJ, et al. Inflammatory stressincreases receptor for lysophosphatidylcholine inhuman microvascular endothelial cells. Am J PhysiolCirc Physiol 2003; 285:H1786–9

49. Macphee CH. Lipoprotein-associated phospholipaseA2: a potential new risk factor for coronary arterydisease and a therapeutic target. Curr OpinPharmacol 2001; 1:121–5

50. Carpenter KL, Dennis IF, Challis IR, et al. Inhibition oflipoprotein-associated phospholipase A2 diminishesthe death-inducing effects of oxidised LDL on humanmonocyte-macrophages. FEBS Lett 2001; 505:357–63

51. Okura Y, Brink M, Itabe H, et al. Oxidized low-densitylipoprotein is associated with apoptosis of vascularsmooth muscle cells in human atheroscleroticplaques. Circulation 2000; 102:2680–6

52. Tjoelker LW, Stafforini DM. Platelet-activating factoracetylhydrolase in health and disease. BiochimBiophys Acta 2000; 1488:102–23

53. Tselepsis AD, Chapman MJ. Inflammation, bioactivelipids and atherosclerosis: potential roles of lipopro-tein-associated phospholipase A2, platelet activatingfactor-acetylhydrolase. Atheroscler Suppl 2002;3:57–68

54. Caslake MJ, Packard CJ. Lipoprotein-associatedphospholipase A2 (platelet activating factor-acetyl-hydrolase) and cardiovascular disease. Curr OpinLipidol 2003; 14:347–52

55. Gomez FP, Rodriguez-Roisin R. Platelet-activatingfactor antagonists: current status in asthma. BioDrugs2000; 14:21–30

56. Berliner JA, Subbanagounder G, Leitinger N, et al.Evidence for a role of phospholipid oxidationproducts in atherogenesis. Trends Cardiovasc Med2001; 11:142–7

57. Bochkov VN, Kadl A, Huber J, et al. Protective role ofphospholipid oxidation products in endotoxin-induced tissue damage. Nature 2002; 419:77–81

58. Walton KA, Cole AL, Yeh M, et al. Specific phospho-lipid oxidation products inhibit ligand activation oftoll-like receptors 4 and 2. Arterioscler Thromb VascBiol 2003; 23:1197–203

59. Kronke G, Bochkov VN, Huber J, et al. Oxidizedphospholipids induce expression of human hemeoxygenase-1 involving activation of cAMP-responsiveelement-binding protein. J Biol Chem 2003; 278:51006–14

60. Kiechl S, Lorenz E, Reindl M, et al. Toll-like receptor4 polymorphisms and atherogenesis. N Engl J Med2002; 347:185–92

61. Ameziane N, Beillat T, Verpillat P, et al. Association ofthe Toll-like receptor 4 gene Asp299Gly polymor-


ch19 14/7/05 4:49 pm Page 227

phism with acute coronary events. ArteriosclerThromb Vasc Biol 2003; 23:61–4

62. Opal S, Laterre P-F, Abraham E, et al. Recombinanthuman platelet-activating factor acetylhydrolase fortreatment of severe sepsis: results of a phase III, multi-center, randomized, double-blind, placebo-controlled, clinical trial. Crit Care Med 2004;32:332–41

63. Hurt-Camejo E, Camejo G, Peilot H, et al.Phospholipase A2 in vascular disease. Circ Res 2001;89:298–304

64. Menschikowski M, Kasper M, Lattke P, et al. Secretorygroup II phospholipase A2 in human atheroscleroticplaques. Atherosclerosis 1995; 118:173–81

65. Ishimoto Y, Yamada K, Yamamoto S, et al. Group Vand X secretory phospholipase A2s-induced modifica-tion of high-density lipoprotein linked to the reduc-tion of its antiatherogenic functions. BiochimBiophys Acta 2003; 1642:129–38

66. Jaross W, Eckey R, Menschikowski M. Biologicaleffects of secretory phospholipase A2 group IIA onlipoproteins and in atherogenesis. Eur J Clin Invest2002; 32:383–93

67. Gesquiere L, Cho W, Subbaiah PV. Role of group IIaand group V secretory phospholipase A2 in the metab-olism of lipoproteins. Substrate specificities of theenzymes and the regulation of their activities bysphingomyelin. Biochemistry 2002; 41:4911–20

68. Kim KP, Han SK, Hong M, et al. The molecular basisof phosphatidylcholine preference of human group-Vphospholipase A2. Biochem J 2000; 348:643–7

69. Ostermann G, Ruhling K, Zabel-Langhennig R, et al.Plasma from atherosclerotic patients exerts anincreased degradation of platelet-activating factor.Thromb Res 1987; 47:279–85

70. Ostermann G, Lang A, Holtz H, et al. The degrada-tion of platelet-activating factor in serum and itsdiscriminative value in atherosclerotic patients.Thromb Res 1988; 52:529–40

71. Graham RM, Stephens CJ, Sturm MJ, et al. Plasmaplatelet-activating factor degradation in patients withsevere coronary artery disease. Clin Sci 1992; 82:535–41

72. Hofmann B, Ruhling K, Spangenberg P, et al.Enhanced degradation of platelet-activating factor inserum from diabetic patients. Haemostasis 1989;19:180–4

73. Satoh K, Imaizumi T, Kawamura Y, et al. Activity ofplatelet-activating factor (PAF) acetylhydrolase inplasma from patients with ischemic cerebrovasculardisease. Prostaglandins 1988; 35:685–98

74. Satoh K, Yoshida H, Imaizumi T, et al. Platelet-activat-ing factor acetylhydrolase in plasma lipoproteinsfrom patients with ischemic stroke. Stroke 1992;23:1090–2

75. Satoh K, Imaizumi T, Kawamura Y, et al. Increasedactivity of the platelet-activating factor acetylhydrolasein plasma low density lipoprotein from patients withessential hypertension. Prostaglandins 1989;37:673–82.

76. Blankenberg S, Stengel D, Rupprecht HJ, et al.Plasma PAF-acetylhydrolase in patients with coronaryartery disease: results of a cross-sectional analysis. JLipid Res 2003; 44:1381–6

77. Packard CJ, O’Reilly DS, Caslake MJ, et al.Lipoprotein-associated phospholipase A2 as anindependent predictor of coronary heart disease.West of Scotland Coronary Prevention Study Group.N Engl J Med 2000; 343:1148–55

78. Caslake MJ, Packard CJ, Suckling KE, et al.Lipoprotein-associated phospholipase A2, platelet-activating factor acetylhydrolase: a potential new riskfactor for coronary artery disease. Atherosclerosis2000; 150:413–19

79. Pearson TA, Mensah GA, Alexander RW, et al.Markers of inflammation and cardiovascular disease:application to clinical and public health practice.Circulation 2003; 107:499–511

80. Ballantyne CM, Hoogeveen RC, Bang H, et al.Lipoprotein-associated phospholipase A2, high-sensi-tivity C-reactive protein, and risk for incidentcoronary heart disease in middle-aged men andwomen in the Atherosclerosis Risk in Communities(ARIC) study. Circulation 2004; 109:837–42.

81. Blake GJ, Dada N, Fox JC, et al. A prospective evalua-tion of lipoprotein-associated phospholipase A2 levelsand the risk of future cardiovascular events in women.J Am Coll Cardiol 2001; 38:1302–6.

82. Karasawa K, Harada A, Satoh N, et al. Plasma platelet-activating factor acetylhydrolase. Progress Lipid Res2003; 42:93–114

83. Miwa M, Miyake T, Yamanaka T, et al.Characterization of serum platelet-activating factor(PAF) acetylhydrolase. Correlation betweendeficiency of serum PAF acetylhydrolase and respira-tory symptoms in asthmatic children. J Clin Invest1988; 82:1983–91.

84. Yoshida H, Satoh K, Koyama M, et al. Deficiency ofplasma platelet-activating factor acetylhydrolase: rolesof blood cells. Am J Hematol 1996; 53:158–64

85. Satoh N, Asano K, Naoki K, et al. Plasma platelet-activating factor acetylhydrolase deficiency inJapanese patients with asthma. Am J Respir Crit CareMed 1999; 159:974–9

86. Stafforini DM, Numao T, Tsodikov A, et al.Deficiency of platelet-activating factor acetylhydrolaseis a severity factor for asthma. J Clin Invest 1999;103:989–97.

87. Yamada Y, Ichihara S, Fujimura T, et al. Identificationof the G994→ T missense in exon 9 of the plasmaplatelet-activating factor acetylhydrolase gene as anindependent risk factor for coronary artery disease inJapanese men. Metabolism 1998; 47:177–81

88. Yamada Y, Yoshida H, Ichihara S, et al. Correlationsbetween plasma platelet-activating factor acetylhydro-lase (PAF-AH) activity and PAF-AH genotype, age,and atherosclerosis in a Japanese population.Atherosclerosis 2000; 150:209–16

89. Yamada Y, Izawa H, Ichihara S, et al. Prediction of the


ch19 14/7/05 4:49 pm Page 228

risk of myocardial infarction from polymorphisms incandidate genes. N Engl J Med 2002; 347:1916–23

90. Hiramoto M, Yoshida H, Imaizumi T, et al. Amutation in plasma platelet-activating factor acetyl-hydrolase (Val279→Phe) is a genetic risk factor forstroke. Stroke 1997; 28:2417–20

91. Yoshida H, Imaizumi T, Fujimoto K, et al. A mutationin plasma platelet-activating factor acetylhydrolase(Val279Phe) is a genetic risk factor for cerebralhemorrhage but not for hypertension. ThrombHaemost 1998; 80:372–5

92. Unno N, Nakamura T, Kaneko H, et al. Plasmaplatelet-activating factor acetylhydrolase deficiency isassociated with atherosclerotic occlusive disease inJapan. J Vasc Surg 2000; 32:263–7

93. Unno N, Nakamura T, Mitsuoka H, et al. Singlenucleotide polymorphism (G994→T) in the plasmaplatelet-activating factor-acetylhydrolase gene isassociated with graft patency of femoropoplitealbypass. Surgery 2002; 132:66–71

94. Ichihara S, Yamada Y, Yokota M. Association of aG994→T missense mutation in the plasma platelet-activating factor acetylhydrolase gene with geneticsusceptibility to nonfamilial dilated cardiomyopathyin Japanese. Circulation 1998; 98:1881–5

95. Yamada Y, Ichihara S, Izawa H, et al. Association of aG994 → T (Val279 → Phe) polymorphism of theplasma platelet-activating factor acetylhydrolase genewith myocardial damage in Japanese patients withnonfamilial hypertrophic cardiomyopathy. J HumGenet 2001; 46:436–41

96. Subramanian VS, Goyal J, Miwa M, et al. Role oflecithin-cholesterol acyltransferase in the metabolismof oxidized phospholipids in plasma: studies withplatelet-activating factor-acetyl hydrolase-deficientplasma. Biochim Biophys Acta 1999; 1439:95–109

97. Abuzeid AM, Hawe E, Humphries SE, et al.Association between the Ala379Val variant of thelipoprotein associated phospholipase A2 and risk ofmyocardial infarction in the north and south ofEurope. Atherosclerosis 2003; 168:283–8

98. Kugiyama K, Ota Y, Takazoe K, et al. Circulating levelsof secretory type II phospholipase A2 predict coronaryevents in patients with coronary artery disease.Circulation 1999; 100:1280–4

99. Blackie JA, Bloomer JC, Brown MJ, et al. The identifi-cation of clinical candidate SB-480848: a potentinhibitor of lipoprotein-associated phospholipase A2.Bioorg Med Chem Lett 2003; 13:1067–70.


ch19 14/7/05 4:49 pm Page 229

ch19 14/7/05 4:49 pm Page 230

ABC transporters see ABCA1; ABCG1;ABCG5/ABCG8

ABCA1cholesterol absorption 58–9HDLs 88–9regulation, reverse cholesterol transport 95–6reverse cholesterol transport 85–6

ABCG1cholesterol absorption 58–9regulation, reverse cholesterol transport 95–6reverse cholesterol transport 85–6

ABCG5/ABCG8cholesterol absorption 58–9genetics 65role in cardiovascular risk 66therapeutic targets 66

absorptionbile acids 73cholesterol 46–7, 57–61dietary fats 57

ACAT see acyl coenzyme A:cholesterolacyltransferase

Acyl CoA, DGATs 31–9acyl coenzyme A:cholesterol acyltransferase

(ACAT)atherogenic properties 47–8atherosclerosis 41–54, 48–50biochemical roles 41–54cholesterol absorption 46–7cholesterol homeostasis 41–54chromosomal location 42–3enzyme activity regulation 44functional differences, ACAT1/ACAT2; 44–5gene family identification 41–2gene knockouts 49–50gene structure 42–3hepatic lipoprotein secretion 47–8inhibition, therapeutic potential 47introduction 41macrophages 48–9physiological roles 41–54regulation 44, 45–6structural studies 43–4transcriptional regulation 45–6

adenosine receptors, lipolysis 170adenosine-triphosphate-binding cassette (ABC)

transporters see ABC transportersadipocytes

adrenoreceptor function 169–70ANP 170cortisol metabolism 171fatty acids metabolism 168–9fatty acids transport 166lipogenesis 165–8secretory products see adipokinesTGs storage 165–73UCP 168–9

adipogenesis 165–8DGATs 167–8PPAR-γ 143–4, 165–8SREBP 165–8thiazolidinediones 168

adipokines 153–64adinopectin 155, 159, 160complement pathway components 158evolutionary aspects, regulation 154free fatty acids 155–6functions 154genetic aspects, evolution 154history 153IL-6 158‘immune’ 157–8leptin 154–5, 159, 160medical therapy 154‘metabolic’ 154–6metabolic syndrome 153, 158–9obesity 153–4PAI-1; 157production 159RAAS 157regulation 154, 159steroid hormones 156terminology 154therapeutic potential 159–60‘thrifty genotype’ 154TNF-α 157–8, 158–9‘vascular’ 156–7VEGF 156–7

Index

Index 14/7/05 4:49 pm Page 231

adiponectin, 155, 159, 160immunity 181–2insulin resistance 125–6

adiposityPPAR-α 141–2PPAR-β/δ 142–3

adrenoreceptor function, lipolysis 169–70afferent arm of innate immunity 175–82ANP see atrial natriuretic peptideanti-inflammatory activity, LXR 96–7apoA-I/apoA-II

cycling 83HDLs 82–3lipid-free/lipid-poor 84

apoB-100biosynthesis 3–9co-translational degradation 9–10co-translational loading 17–19co-translational partial lipidation 4degradation 9–10MTP 17–19post-translational degradation 9–10VLDL 3–9

apoB-48MTP inhibition 19–21VLDL 4–6

ARF 1activation 1–3functions 1–3VLDL 1–3, 8–9

atherogenesisCD44 191–205Lp-PLA2 219–29non-cholesterol sterols 65–6sPLA2- I I A 223, 225sPLA2 isozymes 219–29

atherogenic properties, ACAT 47–8atherosclerosis

ACAT 48–50CD44; 193–5immunity 175–89macrophages 48–9murine model 193–5

atrial natriuretic peptide (ANP), lipolysis 170

bile acidsabsorption 73biosynthesis 71–2detoxification 73–4FXR 69–80transport, enterohepatic circulation 72–3

biosynthesisapoB-100; 3–9

bile acids 71–2body fat location, insulin resistance 121–2body weight diversity 154

cardiovascular riskadipocytes TGs storage 165CETP 114

CD14 receptor, immunity 177–8CD14, soluble, immunity 179–80CD44

atherogenesis 191–205atherosclerosis 193–5ERM proteins 199functions 192–3HA 193, 194–5ligands 192–3matrix composition 199mechanisms 193–201monocytes/macrophages 195–6multiple mechanisms 197–201murine model 193–5potential roles 200–1regulation levels 201rho-family GTPases 199schematic representation 192soluble 193, 200–1structure 192–3T cells 195–7TGF-β 199–200therapeutic potential 200–1therapeutic target 191–205vascular smooth muscle 197–9

cellular response, immunity 182–4cellular sensing, immunity 175–8CETP see cholesteryl ester transfer proteinCHD see coronary heart diseasechemokine receptor-2, immunity 183cholesterol

absorption 57–61absorption, ACAT 46–7absorption inhibition 58dietary 64enterocytes transfer 58–9esterification 59ezetimibe 58hydrolysis 57inhibition, absorption 58micelle formation 57–8plant sterols/stanols 59–61transport, MTP 15–16transport, reverse, HDLs 84–92

cholesteryl ester transfer protein (CETP)animal models 114–15

232 INDEX

Index 14/7/05 4:49 pm Page 232

atherosclerosis effects 114–15cardiovascular risk 114deficiency 112–14genetic deficiency 112–14HDL-cholesterol levels 111–18HDLs 88–9human studies 115–16inhibition, therapeutic potential 111–18lipids effect 114–16lipoprotein metabolism 112–14, 115–16LXR 97–8physiology 111–12plasma levels 114role, LP metabolism/atherosclerosis 112–14variation 112–14VLDL 111–12

chromosomal location, ACAT 42–3complement, immunity 180–1complement pathway components, adipokines

158coronary heart disease (CHD)

insulin resistance 126Lp-PLA2 223–5risk factors, insulin resistance 119–20risk factors, metabolic syndrome 130

cortisol metabolism, adipocytes 171cytokine stimulation, endothelial lipase 105

DBD see DNA binding domainDGATs see diacylglycerol acyltransferasesdiabetes

inflammatory factors 124–5LXR 97see also insulin resistance

diacylglycerol acyltransferases (DGATs)adipogenesis 167–8gene knockouts 34–5inhibition, therapeutic potential 36–7inhibitors 34–5polymorphisms 34–5properties 32–4structure 32–4TG cytotoxic effects 36–7therapeutic targets 31–9therapy, current 35–6

dietary fat metabolism 55–7absorption 57cholesterol 64emulsification 56lipolysis 56–7stomach 56–7see also intestinal lipid metabolism

dietary sterols, hydrolysis 59–61

diglycerides (DGs), intestinal lipid metabolism56–7

DNA binding domain (DBD), nuclear receptors93–4

drug targets, VLDL 10–11

efferent arm of innate immunity 182–5emulsification, dietary fats 56endothelial dysfunction, insulin resistance 126endothelial lipase

comparisons, LPL/HL 101–4cytokine stimulation 105elevating HDL levels 109expression regulation 105–6function 101–4genetic studies 107–9HDLs metabolism 100–10human studies 107–9inhibition 109mechanisms 106–7mice studies 107polymorphisms 107–9regulation 100–10single nucleotide polymorphisms 107–9structure 101–4tissue expression 104–5

energy metabolismPPAR-α 138–42PPAR-β/δ 142–3PPARs 137–51

enterocytes transfer, cholesterol 58–9enterohepatic circulation

bile acids 72–3FXR 72–3

esterification, cholesterol 59euglycaemic hyperinsulinaemic clamp, insulin

resistance 127extracellular lipid metabolism, PPAR-α 140–1ezetimibe, cholesterol absorption 58

farnesoid X receptor (FXR) 69–80bile acids 69–80biology 70–1enterohepatic circulation 72–3guggulsterone 76HDLs 74, 75lipid metabolism 69–80, 74–6modulation 76triglycerides (TG) 74–6

fats, dietary see dietary fat metabolismfatty acids

metabolism 168–9transport 166

INDEX 233

Index 14/7/05 4:49 pm Page 233

free fatty acidsadipokines 155–6insulin resistance 122

FXR see farnesoid X receptor

gene expression studies, MTP 24gene knockouts

ACAT 49–50DGATs 34–5

gene structure, ACAT 42–3genetic deficiency, CETP 112–14GTPases, VLDL assembly 8–9guggulsterone, FXR 76

HA see hyaluronanhepatic fat accumulation, insulin resistance 123–4hepatic lipid metabolism, LXR, regulation 94–5hepatic lipoprotein secretion, ACAT 47–8high density lipoproteins (HDLs) 84

ABCA1; 88–9anti-oxidant properties 87apoA-I/apoA-II 82–3catabolism 83–4CETP 88–9composition 81elevating HDL levels 109endothelial lipase 100–10experimental animals 87–8formation 82–3FXR 74, 75HDL-cholesterol levels, CETP 111–18HDL-cholesterol levels, raising 111–18HDL-mediated cell cholesterol efflux 87HDL-mediated inhibition of adhesion molecule

expression 87heterogeneity 81–2human studies 88intervention studies 87–8LCAT 82–3, 88–9metabolism 82–4metabolism, regulation 100–10overview 81–92plasma 84potential new therapies 88–9PPAR-α agonists 88–9protection mechanisms 87remodelling 84reverse cholesterol transport 84–92size 84stimulation of endothelial NO production 87structure 81subpopulations 81–2therapeutic targets 88–9

HL, endothelial lipase comparison 101–4HOMA see homeostatic model assessmenthomeostasis

cholesterol, ACAT 41–54sterols 41–54, 63–4

homeostatic model assessment (HOMA), insulinresistance 127

HSL, adipocytes TGs storage 165–6human genetic studies, MTP 24humoral response, immunity 184–5humoral sensing, immunity 178–82hyaluronan (HA), CD44; 193, 194–5hydrolysis

dietary cholesterol 57dietary sterols 59–61

IL-6 see interleukin-6‘immune’ adipokines 157–8immunity

adiponectin 181–2afferent arm of innate immunity 175–82atherosclerosis 175–89CD14 receptor 177–8CD14, soluble 179–80cellular response 182–4cellular sensing 175–8chemokine receptor-2; 183complement 180–1efferent arm of innate immunity 182–5humoral response 184–5humoral sensing 178–82IL-6 184–5innate 175–89MCP-1; 183macrophage colony stimulating factor 183macrophage osteopontin 183MBL 180MPO 185pentraxins 178–9TLR4 176–7TNF-α 184–5

inflammationanti-inflammatory activity, LXR 96–7inflammatory factors, diabetes 124–5inflammatory factors, insulin resistance 124–5inflammatory response, LOs 210–11

inhibition, therapeutic potentialACAT 47CETP 111–18DGATs 36–7

inhibitorsDGATs 34–5MTP 17–19

234 INDEX

Index 14/7/05 4:49 pm Page 234

insulin resistanceadiponectin 125–6body fat location 121–2CHD 126CHD risk factors 119–20endothelial dysfunction 126euglycaemic hyperinsulinaemic clamp 127free fatty acids 122hepatic fat accumulation 123–4HOMA 127inflammatory factors 124–5measurement 126–7metabolic syndrome 119–36MTP inhibition 19–21muscle fat accumulation 122–3, 124obesity 120–1PAI-1; 126parameters list 127pathogenic factors 120–30subcutaneous fat 121–2visceral fat 121–2, 124

insulin sensitizing modalitiesclinical effects 130–3metabolic effects 130–3metformin 131–2PPAR-γ agonists 131–3PPARs 131–3

interleukin-6 (IL-6)adipokines 158immunity 184–5

intestinal lipid metabolism see lipid metabolismintracellular metabolism, PPAR-α 139–40

knockouts see gene knockouts

LBD see ligand-binding domainLCAT see lecithin:cholesterol acyltransferaseLDL see low density lipoproteinlecithin:cholesterol acyltransferase (LCAT), HDLs

82–3, 88–9leptin 154–5, 159, 160lifestyle changes, metabolic syndrome 130–2ligand-binding domain (LBD), nuclear receptors

93–4lipid metabolism

FXR 69–80, 74–6intestinal, dietary fat metabolism 55–7intestinal, overview 55–62PPAR-β/δ 142–3

lipogenesisSREBP 165–8TGs synthesis 166–8

lipolysis 169–71

adenosine receptors 170adrenoreceptor function 169–70ANP 170dietary fats 56–7nicotinic acid receptor 170–1

lipoprotein lipase (LPL)endothelial lipase comparison 101–4TGs 75–6

lipoprotein metabolism, CETP 112–14, 115–16lipoprotein secretion, ACAT 47–8lipoxygenases (LOs) 207–18

background 207expression 207–8inflammatory response 210–11LDL oxygenation 207–18mechanisms 210–11model systems 211–13molecular structure 207–8role, atherosclerosis 208–9, 211–14therapeutic potential 207–18, 213–14

liver X receptors (LXR)anti-inflammatory activity 96–7CETP 97–8diabetes 97ligands, therapeutic potential 97–8LXRα, LXRβ 65–6macrophages 95–6regulation, hepatic lipid metabolism 94–5regulation, reverse cholesterol transport 95–6therapeutic targets 93–100

low density lipoprotein (LDL)oxygenation 207–18, 220–3plaque expression 220–3

Lp-PLA2 219–29atherogenesis 219–29background 219–20CHD 223–5role 219–29secreted enzymes 221sPLA2- I I A 223, 225sPLA2 isozymes 219–29

LPL see lipoprotein lipaseLXR see liver X receptors

macrophage chemoattractive protein-1 (MCP-1),immunity 183

macrophage colony stimulating factor, immunity183

macrophage osteopontin, immunity 183macrophages

ACAT 48–9atherosclerosis 48–9LXR 95–6

INDEX 235

Index 14/7/05 4:49 pm Page 235

mannose-binding lectin (MBL), immunity 180‘metabolic’ adipokines 154–6metabolic syndrome

adipokines 153, 158–9CHD risk factors 130criteria 127–30definitions 128downstream targets 130insulin resistance 119–36lifestyle changes 130–2predictive ability 128–9prevalence 128–9treatments 130–3upstream targets 130–3

metabolism, regulation, HDLs 100–10metabolism regulators, nuclear receptors 93–4metformin, insulin sensitizing modalities 131–2micelle formation

cholesterol absorption 57–8plant sterols 60–1

microsomal triglyceride transfer protein (MTP)apoB-100; 17–19cardiovascular implications 15–29cholesterol transport 15–16expression reduction 19gene expression studies 24hepatic regulation 21–4human genetic studies 24inhibitors 17–19insulin resistance 19–21intestinal regulation 19–21MTP–PDI complex 17–19partial, tissue-specific silencing 15–29regulation 19–24triglyceride transport 15–16VLDL 4, 5, 11

modulation, non-cholesterol sterols 63–8monocytes/macrophages, CD44; 195–6MPO see myeloperoxidaseMTP see microsomal triglyceride transfer proteinmurine model

atherosclerosis 193–5CD44; 193–5

muscle fat accumulation, insulin resistance 122–3,124

myeloperoxidase (MPO), immunity 185

nicotinic acid receptor, lipolysis 170–1non-cholesterol sterols

atherogenesis 65–6modulation 63–8role in atherosclerosis 63–8sitosterolaemia 63–8

NR1H4 see FXRnuclear receptors

DBD 93–4LBD 93–4metabolism regulators 93–4

obesityadipokines 153–4insulin resistance 120–1oxysterols 65–6

PAI-1 see plasminogen activator inhibitor 1passive diffusion, reverse cholesterol transport 85–6pentraxins, immunity 178–9peroxisome proliferator-activated receptor α see

PPAR-αphospholipase D (PLD), VLDL 6–8, 9phytosterolaemia see sitosterolaemiasPLA2- I I A, atherogenesis 223, 225sPLA2- I I, Lp-PLA2 A 223sPLA2 isozymes, atherogenesis 219–29plant sterols/stanols

cholesterol absorption 59–61intestinal metabolism 59–61micelle formation 60–1

plasminogen activator inhibitor 1 (PAI-1)adipokines 157insulin resistance 126

PLD see phospholipase Dpolymorphisms

DGATs 34–5endothelial lipase 107–9

potential therapies see therapeutic potentialPPAR-α 75

adiposity 141–2energy metabolism 138–42extracellular lipid metabolism 140–1intracellular metabolism 139–40steatosis 141–2

PPAR-α agonistsadinopectin 155HDLs 88–9VLDL 10

PPAR-β/αδadiposity 142–3energy metabolism 142–3lipid metabolism 142–3

PPAR-γadipogenesis 143–4, 165–8clinical practice 145steatosis 144–5

PPAR-γ agonists, insulin sensitizing modalities131–3

236 INDEX

Index 14/7/05 4:49 pm Page 236

PPARsenergy metabolism 137–51insulin sensitizing modalities 131–3molecular characteristics 137–9

RAAS see renin–angiotensin–aldosterone systemregulation

ACAT 44, 45–6adipocytes TGs storage 165adipokines 154, 159CD44 201endothelial lipase 100–10endothelial lipase expression 105–6HDLs metabolism 100–10hepatic lipid metabolism, LXR 94–5MTP 19–24nuclear receptors metabolism 93–4reverse cholesterol transport 95–6

renin–angiotensin–aldosterone system (RAAS),adipokines 157

reverse cholesterol transportABCA1; 85–6ABCG1; 85–6delivery to liver 86–7HDLs 84–92passive diffusion 85–6regulation 95–6SR-B1; 85–6

rho-family GTPases, CD44; 199

SAR 1, VLDL 1–3scavenger receptor class B1 (SR-B1), reverse

cholesterol transport 85–6secretory phospholipase A2 isozymes see Lp-PLA2sitosterolaemia, non-cholesterol sterols 63–8small intestine, lipolysis 56–7SOAT see ACATsPLA2- I I A, atherogenesis 223, 225sPLA2- I I, Lp-PLA2 A 223sPLA2 isozymes, atherogenesis 219–29SR-B1 see scavenger receptor class B1SREBP see sterol response element binding

proteinstanols/sterols, cholesterol absorption 59–61statin therapy 63–4steatosis

PPAR-α 141–2PPAR-γ 144–5

steroid hormones, adipokines 156sterol response element binding protein (SREBP),

adipogenesis regulation 165–8sterols

homeostasis 41–54, 63–4

non-cholesterol 63–8sterols/stanols, cholesterol absorption 59–61stomach, dietary fat metabolism 56–7storage, adipocyte triglycerides 165–73structural studies, ACAT 43–4structure

ACAT 42–3DGATs 32–4

subcutaneous fat, insulin resistance 121–2syndecan 1; 76

T cells, CD44; 195–7targets, therapeutic see therapeutic targetsTGF-β, CD44 199–200TGs see triglyceridestherapeutic potential

ACAT 47adipokines 159–60CD44 200–1CETP 111–18DGATs 36–7HDLs 88–9LOs 207–18LXR ligands 97–8

therapeutic targetsABCG5/ABCG8 66CD44 191–205DGATs 31–9HDLs 88–9LXR 93–100

thiazolidinediones, adipogenesis 168‘thrifty genotype’, adipokines 154TLR4 see toll-like receptor-4TNF-α see tumour necrotic factor alphatoll-like receptor-4 (TLR4), immunity 176–7transport

cholesterol, MTP 15–16fatty acids 166reverse, cholesterol, HDLs 84–92

triglycerides (TGs)cytotoxic effects 36–7DGATs 36–7FXR 74–6LPL 75–6storage, adipocyte 165–73synthesis, lipogenesis 166–8transport, MTP 15–16

tumour necrotic factor alpha (TNF-α)adipokines 157–8, 158–9immunity 184–5

uncoupling proteins (UCP), fatty acidsmetabolism 168–9

INDEX 237

Index 14/7/05 4:49 pm Page 237

‘vascular’ adipokines 156–7VEGF 156–7

vascular endothelial growth factor (VEGF),adipokines 156–7

vascular smooth muscleadhesion/migration 198–9CD44; 197–9cell growth, de-differentiation, CD44 197, 198

VEGF see vascular endothelial growth factorvery low density lipoproteins (VLDL)

apoB-100; 3–9apoB-48; 4–6ARF 1; 1–3, 8–9

assembly 1–14assembly, overview 3–9CETP 111–12drug targets 10–11GTPases 8–9lipid droplets formation 6–8MTP 4, 5, 11PLD 6–8, 9PPAR-α agonists 10SAR 1; 1–3secretory pathway 1–3

visceral fat, insulin resistance 121–2, 124VLDL see very low density lipoproteins

238 INDEX

Index 14/7/05 4:49 pm Page 238

023 1842142291 lipid and asteroclorosis

Documents