Characterisation of the P19CL6 cardiovascular cell-line at the expression level; effects of long-chain fatty acids

A thesis submitted in accordance with the requirements of the University of Surrey for the degree of Doctor of Philosophy

by Iraj Khodadadi Kohlan

September 2005

Centre of Nutrition and Food Safety School of Biomedical and Molecular Sciences University of Surrey Guildford Surrey GU2 7XH ProQuest Number: 27610157

All rights reserved

INFORMATION TO ALL USERS The quality of this reproduction is dependent upon the quality of the copy submitted.

In the unlikely event that the author did not send a com plete manuscript and there are missing pages, these will be noted. Also, if material had to be removed, a note will indicate the deletion. uest

ProQuest 27610157

Published by ProQuest LLO (2019). Copyright of the Dissertation is held by the Author.

All rights reserved. This work is protected against unauthorized copying under Title 17, United States C ode Microform Edition © ProQuest LLO.

ProQuest LLO. 789 East Eisenhower Parkway P.Q. Box 1346 Ann Arbor, Ml 48106- 1346 Dedication

This thesis is dedicated to my parents and my wife for their enormous support, and to Alisa and Parita

11 Acknowledgment

I would like to thank my two supervisors: Dr Bruce Griffin and Dr Alfred Thumser for helping me to conduct this research in such an interesting and rapidly expanding field of science. I have been fortunate to use an excellent selection of techniques in this study. I am honoured for having worked with Dr Bruce Griffin, who has been an encouraging supervisor and great support during my study. I am also extremely grateful to Dr Alfred Thumser, for his invaluable guidance and help throughout my research at the University of Surrey. He was a great support at the times when I needed him most, and I would consider him a good fiiend as well as an excellent supervisor. I am also very grateful to Dr Nick Plant, who frequently helped during this project, to Dr George Kass for his kind help in fluorescent microscopy, and to Dr David Lovell for his advice in statistical analysis. Much gratitude is also due to Dr Vassilios Mersinias, whose help and advice were invaluable throughout the microarray experiments. Many thanks go to Prof. Colin Smith and Dr Giselda Bucca for their help in conducting the microarray experiments, and to Mr Ben Routley for his help in performing microarray data analysis. I am also thankful to Miss Victoria Leggett for DNA sequencing, and to Mr Peter Kentish for mouse tissue dissection. Finally, a big thank you to my sponsors: the Iranian Ministry of Health and Medical Sciences and the Hamadan University of Medical Sciences.

Ill Abstract

Dietary fats have been implicated in the progression of coronary heart disease and clear differences have been shown for the effects of saturated, monounsaturated and polyunsaturated fatty acids on the different stages of disease. In part, these effects can be explained by the regulation of gene transcription. In this study, a novel cardiac cell-line was targeted for the investigation of fatty acid effects on gene transcription. The aims of this study were to (i) characterise the P19CL6 cell-line at the transcriptional level, (ii) investigate the effects of long-chain fatty acids and clofibrate on mRNA levels of specific lipid metabolism-related , such as heart-type fatty acid-binding and peroxisome proliferator-activated receptor-a, -|3, and -y in the P19CL6 cell-line, and (iii) determine the effects of long chain fatty acids and clofibrate on global transcriptome levels in P19CL6 cells, using cDNA microarray analysis. P19CL6 cells displayed characteristics indicative of a cardiomyocyte phenotype including: the expression of cardiac-specific markers (a-MHC, P-MHC and CaCh-h), an increase in pulse rate with increasing adrenaline concentration and the display of mononuclear cardiomyocyte morphology. However, spontaneous contraction was inconsistent between cultures of P19CL6 cells and pulse rate decreased significantly with increasing passage number until no “beating” was eventually recorded. These observations led us to conclude that P19CL6 cells may display a heterogeneous phenotype, which was investigated further. Analysis of microarray data indicated that global transcriptome profiles observed in the P19CL6 cell-line are not indicative of a pure adult cardiac or skeletal muscle phenotype, and more closely related to mouse embryonic heart tissue. However, the H9C2(2-1) cell-line was more closely related to adult and embryonic heart tissue than the P19CL6 cell-line. Therefore it was concluded that the H9C2(2-1) cell-line represents a better cardiomyocyte model system than the P19CL6 cell-line.

The results of this study also showed that in P19CL6 cells long-chain fatty acids (but not clofibrate) significantly increased the abundance of peroxisome proliferator-activated receptor-a and -y, whereas no changes were observed in the expression levels of heart-type fatty acid-binding protein and peroxisome proliferator-activated receptor-P, under the conditions used. The non-response of the latter genes can probably be explained by the relatively short time of exposure of cells in culture. Microarray analysis

IV showed that linoleic and a-linolenic acids and clofibrate had similar effects, and that these differed fi’om those of palmitic and oleic acids. These results are in agreement with other studies showing that cellular responses to polyunsaturated fatty acids differ fi*om those observed with saturated and monounsaturated fatty acids.

V Abbreviations:

AA arachidonic acid ABCAl ATP-binding cassette transporter 1 ACBP acyl-Coenzyme A binding protein ACS acyl-Coenzyme A synthetase Ad-heart adult heart ADRP adipocyte differentiation related protein AF-1,2 activation factor-1, 2 A-I, II apolipoprotein A-I, II ALA a-Iinolenic acid A-LBP adipocyte lipid-binding protein ANF atrial natriuretic factor B-FABP brain fatty acid-binding protein BSA bovine serum albumin CaCh-h calcium channel (heart isoform) CaCh-m calcium channel (skeletal muscle isoforr CETP cholesterol ester transfer protein CHD coronary heart disease C-II, III apolipoprotein C-II, III Clo clofibrate CM chylomicron COX cyclooxygenase CPT carnitine palmitoyltransferase CRABP cellular retinoic acid binding-protein CRBP cellular retinol binding-protein DART diet and re-infarction trial DBD DNA binding domain DHA docosahexaenoic acid DNA deoxyribonucleic acid EFA essential fatty acid E-FABP epidermal fatty acid binding-protein

VI Em-heart embryonic heart EPA eicosapentaenoic acid FABP fatty acid-binding protein FABPpm plasmalemmal fatty acid-binding protein FAR fatty acid receptor FAT fatty acid translocase FATP fatty acid transport protein HAT histone acetyltransferase HDL high-density lipoprotein HDLC HDL-Cholesterol H-FABP heart fatty acid-binding protein HGMP genome mapping project HNF-4a hepatic nuclear factor-4a IDL intermediate-density lipoprotein I-FABP intestinal fatty acid-binding protein IL interleukin iLBP intracellular lipid-binding protein Kd dissociation constant LA linoleic acid LBD ligand binding domain LCAT lecithin-cholesterol acyltransferase LCFA long-chain fatty acid LDL low-density lipoprotein LDL-C LDL-Cholesterol 1-FABP ileal fatty acid binding-protein L-FABP liver fatty acid-binding protein LLAMA live linear analysis of mocroarray LOX lipooxygenase LPL lipoprotein lipase LT leukotriene LXR liver X receptor MHC myosin heavy chain M-LBP myelin-lipid binding protein; M-FABP

vu MLC2 myosin light chain-2 MUFA monounsaturated fatty acid

NFkB nuclear factor k B OA oleic acid PA palmitic acid PCA principal components analysis PG prostaglandins PLTP phospholipid transfer protein PPAR peroxisome proliferator-activated receptor PPRE peroxisome proliferator regulatory (responsive) element PUFA polyunsaturated fatty acid RXR retinoic X receptor SDS sodium dodecyl sulphate SFA saturated fatty acid SR-Bl scavenger receptor SREBP sterol regulatory element binding protein SSC sodium chloride and sodium citrate TAG triacylglycerol TFA trans fatty acid T-FABP testis fatty acid binding-protein TNFa tumour necrosis factor-a TX thromboxane TZD thiazollidinedione VCAM-1 vascular cell adhesion molecule-1 VLDL very low-density lipoprotein WHO world health organization

V lll Content

Chapter 1

Introduction page

1.1 Introduction 2 1.2 Coronary Heart Disease 3 1.2.1 Definition 3 1.2.2 Risk Factors 4 1.2.3 Atherosclerosis 5 1.3 Properties of fatty acids 5 1.3.1 Fatty acids function as an energy source in the cardiomyocytes 7 1.3.2 p-oxidation of fatty acids 7 1.3.3 Lipoprotein metabolism 8 1.3.4 Effect of fatty acids on PUFA metabolism 10 1.3.5 The link between dietary lipids and CHD 12 1.3.6 Cardioprotective effect of PUFA 13 1.3.7 Delivery of fatty acids to the cardiomyocytes 15 1.4 Membrane-associated fatty acid transporters 15 1.5 Intracellular fatty acid-binding (FABPs) 19 1.5.1 Classification of Intracellular FABPs 19 1.5.2 Ligand-binding characteristics of FABPs 21 1.6 Nuclear Receptors 25 1.6.1 Classification of nuclear receptors 25 1.6.2 Structure of nuclear receptors 27 1.7 Peroxisome proliferator-activated receptors (PPARs) 28 1.7.1 Distribution and PPAR isoforms 28 1.7.2 Transcriptional activity of PPARs 29 1.7.3 PPARs ligands 30 1.7.4 Synthetic ligands of PPARs 33 1.7.5 Complementary roles of PPARs in lipid metabolism 34 1.7.6 PPARs and atherosclerosis 36 1.8 Hypothesis, aim and objectives 38

IX Chapter 2 Materials and Methods page

2.1 Materials 41 2.2 Methods 42 2.2.1 Cell culture 42 2.2.1.1 Culture of P19CL6 cells under adherent conditions 42 2.2.1.2 Culture of P19CL6 cells under non-adherent conditions 43 2.2.1.3 Culture of H9C2(2-1 ) cells under adherent conditions 44 2.2.1.4 Culture of THP-1 cells under adherent conditions 44 2.2.1.5 Determination of PI 9CL6 cells pulse rate cultured with different serum samples 45 2.2.1.6 Culturing P19CL6 cells with concentration on adrenaline 45 2.2.1.7 Determination of PI 9CL6 cell morphology by microscopic analysis under non-adherent culture conditions 46 2.2.1.8 Determination of PI 9CL6 cells morphology using nuclei staining under adherent culture conditions 46 2.2.1.9 Incubation of PI 9CL6 cells with different LCF As 47 2.2.2 Total RNA isolation and purification 48 2.2.2.1 Isolation of total RNA from cell culture samples 49 2.22.1 Isolation of total RNA from mouse tissue samples 50 2.2.2.3 DNajg digestion of total RNA samples 51 2.2.2.4 Purification of total RNA from nucleotides 51 2.2.2.5 Determining the concentration and purity of the RNA 51 2.2.3 Gel electrophoresis 52 2.2.3.1 Formaldehyde-agarose RNA gel electrophoresis 52 2.2.3.2 DNA agarose gel electrophoresis 53 2.2.4 Determination of mRNA levels 54 2.2.4.1 Designing RT-PCR and nested PCR primers 54 2.2.4.2 Reverse transcription polymerase chain reaction (RT-PCR) 57 2.2.4.3 Nested PCR 58 2.2.5 Quantitation of PCR products from digitised gel images 60

X 2.2.6 Purification of PCR products from agarose gels 61 2.2.7 DNA sequencing of PCR products and sequence comparison 62 2.2.8 Microarray experiments , 64 2.2.8.1 Microarray procedure (HGMP protocol) 67 2.2.8.1.1 cDNA synthesis and target labelling 67 2.2.8.1.2 RNA dénaturation 68 2.2.8.1.3 Removal of unincorporated dye 69 2.2.8.1.4 Determining Cy3/Cy5 dye concentration and degree of incorporation into cDNA 69 2.2.8.1.5 Hybridisation of labelled-cDNA to microarrays and post-hybridisation wash up 70 2.2.8.2 Microarray procedure (ProntoPlus kit) 72 2.2.8.2.1 cDNA synthesis and target labelling 72 2.2.8.2.2 RNA dénaturation and removal of unincorporated dye 73 2.2.8.2.3 Pre-soak and pre-hybridisation wash 73 2.2.8.2.4 Hybridisation of labelled-cDNA to microarrays and post-hybridisation wash up 74 2.2.8.3 Microarray scanning and data capture 75 2.2.8.4 Image analysing 75 2.2.8.5 Data analysing 78 2.2.8.5.1 Normalisation and filtering of microarray data 78 2.2.8.5.2 Linear modelling and cluster analysis of microarray data 81

Chapter 3 Characterisation of the P19CL6 cell-line page

3.1 Introduction 83 3.2 Comparison of gene expression profiles of cardiac- and skeletal muscle- specific genes in P19CL6 and H9C2(2-1) cell-lines 86 3.2.1 Rationale 86 3.2.2 Results 86 3.3 Determining gene expression profiles of a-MHC, p-MHC, MyoD and

XI Myogenin in mouse heart and skeletal muscle tissues 90 3.3.1 Rationale 90 3.3.2 Results 91 3.4 Determining gene expression profiles of ANF, MLC2, CaCh-h, and CaCh-m in P19CL6 cells, mouse heart and skeletal muscle tissues 92 3.4.1 Rationale 92 3.4.2 Results 93 3.5 DNA sequencing of PCR products and sequence comparison 95 3.5.1 Rationale 95 3.5.2 Results 95 3.6 Investigating the effect of FBS, passage number and adrenaline on the “beating” phenotype of PI 9CL6 cells 96 3.6.1 Rationale 96 3.6.2 Results 97 3.7 Microscopic examination of P19CL6 cells under adherent culture conditions for nuclei staining 100 3.7.1 Rationale 100 3.7.2 Results 101 3.8 Discussion 103

Chapter 4 Characterisation of the P19CL6 cell-line by microarray analysis page

4.1 Introduction 112 4.2 Validation of microarray procedures 114 4.2.1 Rationale 114 4.2.2 Results 114 4.3 Comparison of global transcriptome levels in the P19CL6 and H9C2(2-1 ) Cell-lines, as well as mouse cardiac and skeletal muscle tissues using microarray analysis 116 4.3.1 Rationale 116

Xll 4.3.2 Results 116 4.4 Microarray analysis of global transcriptome levels in the P19CL6 cell-line, and mouse embryo and heart tissue 123 4.4.1 Rationale 123 4.4.2 Results 123 4.5 Microarray analysis of global transcriptome levels in the P19CL6 and H9C2(2-1) cell-lines, mouse skeletal muscle and mouse embryonic and adult heart tissue 128 4.5.1 Rationale 128 4.5.2 Results 128 4.6 Discussion 136

Chapter 5 Investigation of the effect of LCFAs and fibrates on gene transcription in P19CL6 cell-line page

5.1 Introduction 140 5.2 Determination of LCFAs and clofibrate effects on the expression of H-FABP, PPARa, PPARP and PPARy in P19CL6 cells 142 5.2.1 Rationale 142 5.2.2 Results 142 5.3 Determination of LCFAs and clofibrate effects on global transcriptome levels in P19CL6 cells, using microarray analysis 147 5.3.1 Rationale 147 5.3.2 Results 147 5.4 Discussion 151

Chapter 6

Conclusion page

6.1 Conclusion 155

X lll 6.2 Future works 158

Bibliography 60 i

XIV Figures

Chapter 1 p a g e

Figure 1.1: Diagrammatic representation different long-chain fatty acids 6 Figure 1.2: Diagrammatic representation of lipoprotein metabolism 9 Figure 1.3: Outline of the biosynthesis and metabolism of PUFA 11 Figure 1.4: Molecular mechanism of uptake and utilization of long-chain fatty acids 17 Figure 1.5: Structure of I-FABP showing the configuration of bound palmitic acid 21 Figure 1.6: Intracellular transports of fatty acids 22 Figure 1.7: Classification of nuclear receptors superfamily 26 Figure 1.8: Schematic representation of the structure of nuclear receptors 27 Figure 1.9: Diagrammatic representation of the PPARiRXR heterodimer bound to a PPAR-response element (PPRE) 30 Figure 1.10: Chemical structure of fibrates 33 Figure 1.11 : Complementary roles of PPAR isoforms in lipid metabolism 36

Chapter 2 page

Figure 2.1: Overview of total RNA isolation methods 48 Figure 2.2 : Example of the output from NanoDrop ND-1000 spectrophotometer for determining total RNA concentration 52 Figure 2.3: An illustration of the primer-encoding region of the Mouse musculus MyoD gene 55 Figure 2.4: Diagrammatic representation of RT-PCR products amplification by nested PCR. 59 Figure 2.5: An example of the process used to digitise gel images and semi- quantitate PCR products using Scion Image software 61 Figure 2.6: A typical elution profile and DNA sequence of a PCR product, e.g. MyoD, obtained by DNA sequencing 63

XV Figure 2.7: Representation of dual-hybridisation optimal interwoven loop designs for (A) three, (B) four, (C) six, and (D) seven conditional microarray experiments 65 Figure 2.8: A schematic representation of the steps required for microarray analysis 66 Figure 2.9: Schematic representation of the direct target-labelling procedure 68 Figure 2.10: Screen features of the NanoDrop ND-1000 spectrophotometer used to determine cDNA and Cy3/Cy5-dye concentrations 70 Figure 2.11: Representation of (A) design and (B) description of loop comparison for P19CL6 cells, mouse embryo and mouse heart total RNA 72 Figure 2.12: Layout of Mouse Known Gene SGC Oligo Set Array version 2 76 Figure 2.13: An example of an overlaid image generated after microarray scanning and analysis using the BlueFuse Software 77 Figure 2.14: An example of a scatter plot generated after microarray scanning and analysis using the BlueFuse Software 78 Figure 2.15: Illustration of the scatter plots for the self-against-self hybridisation of PI 9CL6 cells before (A) and after (B) data normalisation and filtering 79 Figure 2.16: An example of scatter plot generated by GeneSpring-7 software to represent the gene expression levels based on Cy3:Cy5 signal

ratio for P19CL6 v5H9C2(2-1) 80

Chapter 3 page

Figure 3.1: Expression of a-MHC, P-MHC, MyoD and Myogenin in (A) P19CL6 and (B) H9C2(2-1) cells cultured under adherent conditions on exposure to 1% DMSO, as detected by RT-PCR 87 Figure 3.2: Expression of a-MHC, p-MHC, MyoD and Myogenin in (A) P19CL6 and (B) H9C2(2-1) cells, cultured under adherent conditions on exposure to 1% DMSO, as detected by nested PCR 87 Figure 3.3: Expression of a-MHC, p-MHC, MyoD and Myogenin in P19CL6 cells cultured under non-adherent conditions, as determined by

XVI nested PCR 88 Figure 3.4: Expression of a-MHC, p-MHC, MyoD, and Myogenin in P19CL6 cells cultured under non-adherent conditions, as determined by nested PCR 89 Figure 3.5: Time course effect of incubation period on the expression of a-MHC, P-MHC, MyoD, and Myogenin in PI9CL6 cells cultured under non-adherent conditions, as determined by nested PCR 90 Figure 3.6: Detection of a-MHC, P-MHC, MyoD and Myogenin in mouse heart and skeletal muscle tissue, as determined by RT-PCR 91 Figure 3.7: Expression of a-MHC, P-MHC, MyoD and Myogenin in mouse heart and skeletal muscle tissue, as determined by nested PCR 92 Figure 3.8: Expression of a-MHC, P-MHC, ANF, MLC2 and CaCh-h in mouse heart, skeletal muscle tissue and P19CL6 cells, as detected by RT-PCR 93 Figure 3.9: Expression of cardiac and skeletal muscle isoforms of voltage- dependent calcium channel (CaCh-h and CaCh-m, respectively) in mouse tissues and P19CL6 cells, as detected by RT-PCR 94 Figure 3.10: An illustration of a sequence alignment for a PCR product and the targeted gene sequences 95 Figure 3.11: Determination of the effect of different FBS batches on the “beating” characteristic in P19CL6 cells cultured under adherent conditions 98 Figure 3.12: Determination of the effect of passage number on “beating” characteristic in P19CL6 cells cultured under adherent conditions 99 Figure 3.13: Determining the effect of different concentrations of adrenaline on “beating” in PI9CL6 cells cultured under adherent conditions 100 Figure 3.14: Phase-contrast microscopy of P19CL6 and H9C2(2-1) cells 101 Figure 3.15: Nuclear staining of P19CL6 cells cultured under adherent Conditions 102 Figure 3.16: Tissue-specific gene expression patterns obtained from the GeneCards database information for the expression of (A) a-MHC and (B) p-MHC in different human tissues 108 Figure 3.17: Tissue-specific gene expression patterns obtained from the

xvii GeneCards database information for the expression of (A) MyoD and (B) Myogenin in different human tissues 109

Chapter 4 page

Figure 4.1: Scatter plots for the self-hybridisation of P19CL6 cell-line (A) before, and (B) after data normalisation and filtering process 115 Figure 4.2: Experimental loop design to compare global transcriptome levels in P19CL6 and H9C2(2-1) cells, as well as mouse cardiac and skeletal muscle tissues 117 Figure 4.3: Scatter plots of samples: (A) P19CL6 cell-line V5 H9C2(2-1) cell-line, (B) H9C2(2-1) cell-line vs skeletal muscle tissue, (C) skeletal muscle vs Heart, and (D) Heart vj PI9CL6 cell-line 118 Figure 4.4: Hierarchical cluster analysis of mRNA expression levels in microarrays comparing the P19CL6 and H9C2(2-1 ) cell-lines, as well as mouse heart and skeletal muscle tissue 120 Figure 4.5: Principle component analysis (PCA) of global transcriptome levels in PI9CL6 and H9C2(2-1 ) cells, and mouse heart and skeletal muscle tissue 121 Figure 4.6: Experimental loop design to compare global transcriptome levels in PI9CL6, as well as mouse embryo and heart tissues 124 Figure 4.7: Scatter plots of samples: (A) P19CL6 cell-line V5 mouse embryo, (B) mouse embryo vs heart tissue, and (C) heart tissue vy P19CL6 cell-line 125 Figure 4.8: Hierarchical cluster analysis of mRNA expression levels in microarrays comparing the P19CL6 cell-line and mouse embryo and heart tissue - 126 Figure 4.9: Principal component analysis (PCA) of global transcriptome levels in the P19CL6 cell-line, mouse embryo and heart tissue 127 Figure 4.10: Representation of dual-hybridisation optimal interwoven loop design for microarray experiment 129 Figure 4.11: Scatter plots of samples: (A) P19CL6 cell-line adult heart, (B) reference vs adult heart, (C) embryonic heart vs H9C2(2-1) cell-line.

XVlll (D) embryonic heart vj reference, (E) P19CL6 cell-line vs skeletal muscle, and (F) reference V5 PI9CL6 cell-line 130 Figure 4.12: Scatter plots of samples: (A) reference V5 skeletal muscle, (B) skeletal muscle vs adult heart, (C) adult heart vs H9C2(2-1) cell-line, (D) P19CL6 cell-line vs reference, (E) P19CL6 cell-line vy embryonic heart, and (F) skeletal muscle vs embryonic heart 131 Figure 4.13: Scatter plots of samples: (A) adult heart vs P19CL6 cell-line, (B) P19CL6 cell-line V5 H9C2(2-1) cell-line, (C) H9C2(2-1) cell-line P19CL6 cell-line, (D) P19CL6 cell-line vs skeletal muscle, (E) P19CL6 cell-line vs embryonic heart, and (F) H9C2(2-1) cell-line V5 reference 132 Figure 4.14: Representation of theoretical loops 1 and 2 derived from optimal interwoven loop design for microarray experiment 133 Figure 4.15: Principal components analysis (PCA) of global transcriptome levels in the P19CL6 and H9C2(2-1) cell-lines, mouse skeletal muscle and embryonic and adult heart tissue obtained from theoretical loops 1 and 2 134

Chapter 5 page

Figure 5.1: Expression of H-FABP, PPARa, PPARp and PPARy in P19CL6 cells cultured under adherent conditions on exposure to 1% DMSO, as detected by RT-PCR 143 Figure 5.2: Expression of PPARa and PPARy in THP-1 cells, as detected by RT-PCR 144 Figure 5.3: Expression of PPARa and PPARy in THP-1 cells, as detected by RT-PCR and nested PCR 144 Figure 5.4: Expression of H-FABP, PPARa, PPARp, and PPARy in PI 9CL6 cells cultured under adherent conditions on exposure to 1% DMSO, as detected by nested PCR 145 Figure 5.5: Experimental loop design to compare global transcriptome levels in PI9CL6 cells cultured with LCFAs or clofibrate 148 Figure 5.6: Scatter plots of samples from P19CL6 cells treated with: (A) BSA

XIX V5 palmitic acid, (B) palmitic acid oleic acid, (C) oleic acid V 5 linoleic acid, (D) linoleic acid vs a-linolenic acid, (E) a-linolenic acid V5-clofibrate, and (F) clofibrate BSA 149 Figure 5.7: Hierarchical cluster analysis of mRNA expression levels in microarrays comparing the P19CL6 cells cultured with LCFAs and clofibrate 150 Figure 5.8: Principal component analysis (PCA) of global transcriptome levels in P19CL6 cultured with LCFAs or clofibrate 150

XX Tables

Chapter 1 page

Table 1.1: Major risk factors for CHD 4 Table 1.2: Membrane-associated proteins involved in fatty acid uptake and transport 16 Table 1.3: The FABP protein family. 20 Table 1.4: Examples of fatty acid-regulated genes involved in lipid metabolism and their corresponding function 24 Table 1.5: Natural and synthetic PPAR ligands 32

Chapter 2 page

Table 2.1: Sequences of oligonucleotide primers used for RT-PCR 56 Table 2.2: Sequences of oligonucleotide primers used for nested PCR 57 Table 2.3: RT-PCR cycling profile 58 Table 2.4: Components in nested PCR tubes 59 Table 2.5: Nested PCR cycling profile 60

Chapter 3 page

Table 3.1: Identity of PCR product sequences to those of target genes 96 Table 3.2: Comparison of pulse rate observed in P19CL6 cells cultured with different FBS samples 97 Table 3.3: Effect of adrenaline on the beating phenotype of P19CL6 cells 99

Chapter 4 page

Table 4.1 : Correlation matrix obtained by principal component analysis (PCA) of mRNA expression levels in the P19CL6 and H9C2(2-1) cell-lines and mouse heart and skeletal muscle tissue 122

XXI Table 4.2: Correlation matrix comparing mRNA expression levels in the P19CL6 and H9C2(2-1) cell-lines and mouse heart and skeletal muscle tissue 123 Table 4.3: Correlation matrix obtained by principal component analysis (PCA) of mRNA expression levels in the P19CL6 cell-line and mouse embryo and heart tissue 127 Table 4.4: Samples and Cy-dye labelling used on different microarray slides, based on the loop design 129 Table 4.5: A summary of correlation matrix obtained by principal components analysis (PCA) of mRNA expression levels in the P19CL6 and H9C2(2-1) cell-lines, mouse skeletal muscle and mouse embryonic and adult and heart tissue, obtained from the theoretical loop 1 135 Table 4.6: Correlation indexes of P19CL6 and H9C2(2-1 ) cell-lines to the mouse embryonic and adult heart tissues obtained for principle component analysis 136

Chapter 5 page

Table 5.1: Comparison of H-FABP, PPARa, PPARp, and PPARy gene expression relative to GAPDH observed in P19CL6 cells cultured with different LCFAs 146

xxii Chapter 1

Introduction Chapter 1:

1.1 Introduction

Coronary heart disease (CHD) is the leading cause of death in developed countries (Mata and Ortega, 2003; Talmud and Stephens, 2004; Chilton, 2004) and the prevention and treatment of CHD continues to be a significant challenge (Bartlett and Eaton, 2004). The link between dietary fat, blood lipid profile and heart disease has been clearly established (Szapary and Rader, 2004; Calder, 2004b; Wijendran and Hayes, 2004), with saturated fatty acids and cholesterol being considered as atherogenic fat, whereas polyunsaturated fatty acids are cardioprotective (Ascherio, 2002; Mori and Beilin, 2001; Wijendran and Hayes, 2004; Dyerberg et al., 2004; Calder, 2004a; Mesa et al., 2004; Calder, 2004b). Many of the potentially beneficial effects of dietary polyunsaturated fatty acids in preventing or inhibiting the progression of CHD are mediated through the control of gene transcription (Calder, 2004a; Calder, 2004b; Wijendran and Hayes, 2004). These fatty acids elicit their effects by co-ordinately suppressing lipid synthesis in the liver and up-regulating fatty acid up-take and oxidation in the liver, heart, and skeletal muscle, therefore improving the plasma lipid profile (Jump, 2004; Mesa et ah, 2004; Clarke, 2000; Pegorier et al., 2004).

Investigating the effect of different dietary fatty acids on the cardiovascular system has increased over the past two decades (Ascherio, 2002; Mori and Beilin, 2001; Wijendran and Hayes, 2004; Dyerberg et al., 2004; Calder, 2004a; Mesa et al., 2004; Calder, 2004b). However, there are a limited number of cardiovascular model systems available for the in vfrm investigation of fatty acid effects on cardiomyocytes (White et al., 2004; Mitcheson et al., 1998). Isolated embryonic and neonatal rat primary cardiomyocytes have been the most widely used models to study cardiac biology in vitro, but their use is somewhat limited because they lack many adult cardiomyocyte characteristics and become overgrown by non-myocytes after a few days in culture (White et al., 2004; Claycomb et al., 1998; Mitcheson et al., 1998). Different types of stem cells can differentiate into cardiac muscle cells in vitro, and the P19 mouse embryonal carcinoma (EC) cell-line was one of the first examples described and has been the most extensively characterized (van der Heyden and Defize, 2003). P19 embryonal carcinoma cells are multipotent and can differentiate into different cell types such as cardiac, skeletal muscle and neural cells (van der Heyden et al., 2003; van der Heyden and Defize, 2003; McBumey et al., 1982; Skeijanc, 1999; McBumey, 1993). Application of dimethyl sulphoxide (DMSO) to P19 cells, triggers signaling events that results in the differentiation into a cardiac-muscle phenotype (Skeijanc, 1999; McBumey, 1993). Therefore, as result of their capacity to form cardiomyocytes in vitro^ P19 cells have been widely used to study cardiac cell physiology (van der Heyden et al., 2003; Anisimov et al., 2002; van der Heyden and Defize, 2003).

The P19C16 cell-line is a suh-clone of PI 9 cells that has been reported to efficiently differentiate into cardiomyocytes upon exposure to DMSO under adherent culture conditions (Habara-Ohkubo, 1996). The P19CL6 cell-line has been widely used as an in vitro model of cardiovascular cells (Young et al., 2004; van der Heyden and Defize, 2003; van der Heyden et al., 2003; Peng et al., 2002; Anisimov et al., 2002; Paquin et al., 2002; Ridgeway et al., 2000; Monzen et al., 2001) and was used in this study to investigate the effect of long-chain fatty acids (LCFA) and fibrates on the expression of genes related to lipid metabolism.

1.2 Coronary Heart Disease (CHD) 1.2.1 Definition Coronary heart disease (CHD) has been the single most important cause of premature death in modem and industrialised countries over the last two decades (Lindsay and Gaw, 2001; Mata and Ortega, 2003; Talmud and Stephens, 2004; Chilton, 2004; Renaud and Lanzmann-Petithory, 2001). The World Health Organisation (WHO) has predicted that cardiovascular diseases will represent the major cause of morbidity and mortality worldwide before 2010 (Chapman, 2003). This will be accompanied by a marked increase in the prevalence of Diabetes Mellitus, obesity and the Metabolic Syndrome (Chapman, 2003). . 1.2.2 Risk factors

The importance of underlying risk factors in the diagnosis, treatment, and management of CHD is well known, with almost 300 variables being statistically associated with CHD (Poulter, 1999). However, evidence suggests that the vast majority of coronary events depend on genetic constitution, gene-environment interactions, lifestyle and factors such as age and gender (Assmann et al., 1999; Talmud, 2004; Mathers, 2004). Major CHD risk factors routinely assessed in clinical practice and epidemiological studies are classified in Table 1.1 into two main groups, namely those that are modifiable and unmodifiable (Lindsay and Gaw, 2001; Assmann et al., 1999; Fonarow and Horwich, 2003; Poulter, 2003):

Table 1.1: Major risk factors for CHD (Lindsay and Gaw, 2001)

CHD risk factors Modifiable Unmodifiable .Smoking Family history of CHD Elevated plasma cholesterol Personal history of CHD Obesity Diabetes Mellitus Physical inactivity Age Excess alcohol Gender Stress

Epidemiological and clinical studies indicate that three main CHD risk factors should be considered in monitoring and management processes, namely abnormal plasma lipid concentrations or dyslipidaemia, hypertension and smoking (Talmud and Stephens, 2004; Poulter, 2003; Ascherio, 2002). High plasma lipid levels, particularly cholesterol, play an important role in the development of CHD and atherosclerosis (Chilton, 2004) and lowering cholesterol levels, particularly low-density lipoprotein-cholesterol (LDL-C), has been the focus of the prevention of CHD for the last 25 years (Poulter, 1999; Chilton, 2004; Spratt, 2004). Recently, the importance of postprandial lipaemia and moderately raised serum triacylglycerol levels in the development of coronary heart disease has been recognised to generate pro-atherogenic changes in serum lipoproteins, including the formation of small dense LDL and the reduction of cardioprotective high-density lipoproteins (HDL) (Griffin and Fielding, 2001; Tulenko and Sumner, 2002; Wang and Briggs, 2004; Bemeis and Rizzo, 2004).

1.2.3 Atherosclerosis

Atherosclerosis is a pathological process that ultimately leads to a localized obstruction of an artery due to the progressive build-up of lipids in the arterial wall (Plutzky, 2003; Stocker and Keaney, 2004; Ross, 1999; Cullen, 2003). Several pathological processes participate in atherosclerotic plaque formation including: (1) Lipoprotein particles, and especially LDL, enter the arterial wall and undergo oxidation. (2) Oxidised LDL (ox-LDL) particles can incite a cascade of cellular responses that includes production of inflammatory mediators. (3) Vascular adhesion molecules (VCAM-1) are induced on the endothelial cell surface. (4) Monocytes attach to VCAM-1 and enter the arterial wall. (5) Migrated monocytes take-up the ox-LDL and differentiate into foam cells. (6) Vascular smooth muscle cells respond to these stimuli and migrate into the intima, where they proliferate and form early atherosclerotic lesions, leading to the formation of a fibrous cap and complicated atherosclerotic plaque in the later stages (Plutzky, 2003; Ross, 1993; Barish and Evans, 2004; Li and Glass, 2004; Cullen, 2003).

1.3 Properties of fatty acids

Dietary fat is an important macronutrient for the growth and development of all organisms (Wijendran and Hayes, 2004; Manco et al., 2004; Calder, 2004a). Changes in the quantity or composition of ingested dietary fat influences cellular processes and physiological systems (Jump, 2004). Many of these effects can be linked to changes in membrane lipid composition affecting membrane fluidity or eicosanoid signalling (Jump, 2002b; Clarke, 2000; Jump, 2004). The properties of dietary fats are primarily determined by the composition of their fatty acids, which may be saturated (no double bonds), monounsaturated (one double bond) or polyunsaturated (with greater than one double bond) (Calder and Grimble, 2002). Unsaturated fatty acids can adopt two distinct geometric configurations, denoted as the cis and trans configurations (Fig. 1.1), where cis fatty acids have a bend in the chain, and trans fatty acids adopt a straight configuration and function like saturated fatty acids (Yaqoob, 2003). There are four families of unsaturated fatty acids, denoted as n-9, n-7, n-6 and n-3 series (Calder and Grimble, 2002). This nomenclature indicates the position of the first double bond fi*om the methyl terminus of the hydrocarbon chain (Fig. 1.1) (Calder and Grimble, 2002). are unable to synthesise linoleic acid (LA; 18:2 n-6) and a-linolenic acid (ALA; 18:3 n-3), and therefore these fatty acids are classified as essential fatty acids (EFA), and must be supplied in the diet (Yaqooh, 2003; Wijendran and Hayes, 2004). It is estimated that a minimum human requirement for LA and ALA is 1% and 2% of daily energy intake, respectively (Calder, 2004b). These two fatty acids can be converted into other polyunsaturated fatty acids (PUFA) of the n-3 and n-6 series hy further elongation and desaturation (Calder, 2003; Jump, 2002b; Jump, 2004; Nakamura and Nara, 2004).

Saturated fatty acids, e.g. Palmitic acid (C16:0)

16 14 12 10 8 6 4 2

Monousaturated fatty acids (c/s), e.g. Oleic acid (18:1 n-9) 8 6 4 2

18 16 14 ?2 10 9

Monousaturated fatty acids (trans), e.g. Elaidic acid (18:1 n-9) 18 16 14 12 10 8 6 4 2

w V Y, X:OOM

Polyusaturated fatty acids, e.g. Linoleic acid (18:2 n-6) 18 16 14 11 8 6 4 2

13 12 10 9 COOH

Polyusaturated fatty acids, e.g. a- Linolenic acid (18:3 n-3) 8 6 4 2

16 15 13 12 1 COOH

Figure 1.1: Diagrammatic representation different long-chain fatty acids (Calder and Grimble, 2002). 1.3.1 Fatty acids function as an energy source in the cardiomyocytes

The heart is able to use a number of different substrates for aerobic energy production including fatty acids, glucose, pyruvate, lactate, acetate, ketone bodies and, to a minor extent, amino acids (Belke et ah, 1999; Soltys et al., 2002; van der Vusse et al., 2000). In the foetal heart, lactate oxidation and glycolysis are the preferred route of energy production, but following birth fatty acid oxidation becomes the predominant energy source providing over 70% of the energy needs for cardiac function under normal physiological conditions (Hauton et al., 2001; van Bilsen et ah, 2004; Lewin and Coleman, 2003; Makinde et al., 1998).

The heart relies heavily on the continuous supply of fatty acids, as endogenous synthesis is limited (Brinkmann et al., 2002; Stavinoha et al., 2004; Lewin and Coleman, 2003). However, the cardiac muscle may shift continuously from one source of energy to another, according to nutrient availability, which is controlled by diet, exercise, or pathophysiological state (Grynberg and Demaison, 1996; Ventura-Clapier et al., 2004; van Bilsen et al., 2004; van der Vusse et al., 2000). For example, under high glucose and low fatty acid availability, carbohydrates (and lactate) become the major energy substrate for the heart (Grynberg and Demaison, 1996; Belke et al., 1999), and the ability to metabolise FA by p-oxidation is reduced (Belke et al., 2002). In contrast, under fasting conditions fatty acids become the major source of energy and are preferentially oxidised in the heart, and carbohydrate utilization is decreased (Grynberg and Demaison, 1996; Frayn, 2003).

1.3.2 P-Oxidation of fatty acids

The oxidation of fatty acids is a central metabolic process for the provision of energy in the cell (Ghisla, 2004; Kim and Battaile, 2002). Fatty acid oxidation is achieved in a series of reactions that occur in mitochondria, peroxisomes and endoplasmic reticulum (Reddy and Hashimoto, 2001; Bartlett and Eaton, 2004; Ghisla, 2004; Kim and Battaile, 2002). The p-oxidation of fatty acids is confined to mitochondria and peroxisomes, whereas co-oxidation is a specialised pathway and occurs in the microsomes (Reddy and Hashimoto, 2001; Bartlett and Eaton, 2004). Mitochondria catalyse the complete p-oxidation of the bulk of dietary short- (

In contrast to mitochondria, peroxisomes are involved in the p-oxidation and chain shortening of long-chain, very-long-chain and branched-chain fatty acids, and prostaglandins (Reddy and Hashimoto, 2001; Kim and Battaile, 2002). Although long-chain fatty acids are predominantly p-oxidised in mitochondria, the P-oxidation of very-long-chain fatty acids (>C 2o) almost exclusively occurs in peroxisomes (Reddy and Hashimoto, 2001; Kim and Battaile, 2002). Unlike the mitochondrial pathway, peroxisomal p-oxidation system is carnitine-independent, and normally stops after a few cycles with the products exported to the mitochondria for complete oxidation (Reddy and Hashimoto, 2001; Kim and Battaile, 2002).

1.3.3 Lipoprotein metabolism

Chylomicrons (CM) containing apolipoprotein (apo) B-48 and large amounts of triacylglycerol (TAG) are synthesised in the intestine from dietary fats (Fig. 1.2) (Tulenko and Sumner, 2002; Wang and Briggs, 2004; Williams et al., 2004). In the circulation, CMs acquire apo-E and apo C-II and are reduced in size by the action of lipoprotein lipase (LPL), which catalyses removal of fatty acids from TAG pool. After its activation by apo C-II on CM, LPL hydrolyses most of the TAG to free fatty acids, the latter then bind to albumin and are deposited in tissues, mainly adipocytes and muscle (Wang and Briggs, 2004; Packard, 2003). VLDL, the other TAG-rich particle containing apo E, C-II and B-lOO, is synthesised in hepatocytes and secreted into the circulation for the delivery of lipids to the periphery (Tulenko and Sumner, 2002; Packard, 2003). In the circulation, LPL, is again activated by apo C-II, and hydrolyses and reduces VLDL triacylglycerol content, leaving the particle smaller, denser and cholesterol-enriched, termed intermediate-density lipoprotein (IDL) that with further hydrolysis can generate LDL (Fig. 1.2) (Tulenko and Sumner, 2002; Packard, 2003).

Cholesterol homeostasis involves the reverse transport of cholesterol from tissues back to the liver (Rader, 2003; Fredenrieh and Bayer, 2003; Linsel-Nitschke and Tall, 2005). Nascent HDL (pre-pHDL) containing apo A-I, which is synthesised in the liver and intestine binds to peripheral cell surface receptors, e.g. ATP-binding cassette transporter 1 (ABCAl) that facilitates the efflux of free cholesterol from the cells onto the surface of pre-pHDL (Duval et al., 2004; Tulenko and Sumner, 2002; Wang and Briggs, 2004; Fredenrieh and Bayer, 2003; Linsel-Nitschke and Tall, 2005). Absorbed

Adipose tissue

B48 FFA LPL FFA CM B100 Thoracic VLDL LPL duct lymph B48 CM ( remnant B100

In testin e

B100

LDL

Cholesterol LDL receptor delivery D iet B ile CETP/

HDL receptor Peripheral L iv er tissues

Cholesterol efflux HDL

Figure 1.2: Diagrammatic representation of lipoprotein metabolism.CM is synthesised by intestine from dietary fat, whereas VLDL and HDL are synthesised by liver from the endogenous sources. LPL in the circulation hydrolyses TAG core of CM and VLDL particles and releases free fatty acids. CETP: cholesterol ester transport protein; CM: chylomicron; FFA: free fatty acid; LDL: low-density lipoprotein; HDL: high-density lipoprotein; VLDL: very low-density lipoprotein; A, B48, BlOO, C and E are the major apolipoproteins associated with the particles (Tulenko and Sumner, 2002; Wang and Briggs, 2004). cholesterol is then esterified by lecithin-cholesterol acyltransferase (LCAT), which is activated by apo A-I, and internalised into the hydrophobic core of the pre-pHDL particle, the process that promotes further cholesterol efflux, mainly facilitated by scavenger receptor B1 (SR-Bl) and gradually matures pre-pHDL to HDL] and HDL] (Tulenko and Sumner, 2002; Wang and Briggs, 2004; Linsel-Nitschke and Tall, 2005; Fredenrieh and Bayer, 2003; Spady, 1999). HDL] and HDL] exchange cholesterol, TAG and phospholipid with VLDL/LDL particles, mediated by cholesterol ester transfer protein (CETP), TAG transfer protein and phospholipid transfer protein (PLTP), respectively or are taken-up by the liver for degradation and disposal of cholesterol in the form of bile acids (Fig. 1.2) (Duval et al., 2004; Tulenko and Sumner, 2002; Wang and Briggs, 2004; Linsel-Nitschke and Tall, 2005).

1.3.4 Effect of fatty acids on PUFA metabolism

As discussed earlier, dietary LA and ALA are essential fatty acids and must be supplied in the diet (Yaqoob, 2003), and are then converted into other n-3 and n-6 polyunsaturated fatty acids (Calder, 2003; Jump, 2002b). The major end-product of LA (18:2 n-6) metabolism is arachidonic acid (AA; 20:4 n-6), whereas ALA (18:3 n-3) is mainly metabolised to eicosapentaenoic acid (EPA; 20:5 n-3) and, to a much lesser extent, docosahexaenoic acid (DHA; 22:6 n-3) (Calder, 2003). The metabolism of n-3 and n-6 fatty acids is competitive and both pathways use the same set of desaturation and elongation enzymes, most importantly A6-desaturase (Fig. 1.3) (Belury, 2002; Calder, 2004b; Nakamura and Nara, 2004; Jump, 2004). Although ALA is the preferred substrate for the latter enzyme, LA is by far the most abundant substrate (Calder, 2004b). Due to this competitive effect, the composition of cellular lipid pools reflects, to some degree, alterations in dietary intakes of those two essential fatty acids (Yaqoob, 2003). Since LA is much more prevalent than ALA in human diets, the metabolism of n-6 fatty acids is quantitively the more important (Calder, 2004b), however dietary supplementation with LC n-3 PUFA is able to competitively suppress the n-6 pathway and vice versa (Calder and Grimble, 2002).

Both n-3 and n-6 polyunsaturated fatty acids are important components of membrane phospholipids, regulating membrane fluidity (Yaqoob, 2003). Although several

10 20-carbon PUF As are able to serve as eicosanoid precursors, arachidonic acid (n-6 PUFA) is usually the principal substrate for their synthesis (Calder, 2003). Arachidonic acid in cell membranes can be mobilized by phospholipase A%

(PLA2) (Calder, 2001) and subsequently acts as a substrate for cyclooxygenase (COX) and lipooxygenase (LOX) enzymes that give rise to the production of the inflammatory mediators such as prostaglandins (PG), thromboxanes (TX), and leukotrienes (LT) (Fig. 1.3) (Calder, 2001; Calder and Grimble, 2002; Calder and Miles, 2000; Calder, 2003; Yaqoob, 2003; Calder, 2004b). On the other hand, n-3 fatty acids, especially eicosapentaenoic and docosahexaenoic acids, are known to exert anti -inflammatory effects and act as arachidonic acid antagonists by decreasing the availability of AA for eicosanoid production, reducing oxygenation of AA by COX and producing derivatives that are considered to promote a less thrombotice inflammatory environment (Fig. 1.3) (Calder and Grimble, 2002; Calder, 2004a; Calder, 2001; Plutzky, 2003;

Linoleic acid a-Linolenic acid LA(18:2n-6) ALA(18:2n-3)

Ï A6-desaturase ; V '------— ^ V

Elongase V V

A5-desaturase COX'

Arachidonic acid Eicosapentaenoic acid AA (20:4n-6) EPA(20:5n-3) LO: LT, COX LOX

LT Docosahexaenoic acid DHA (22:6n-3)

Figure 1.3: Outline of the biosynthesis and metabolism of PUFA. Linoleic (n-6) and a-Linolenic (n-3) acids are competitively converted into their highly unsaturated metabolites by the action of desaturase and elongase enzymes (Calder and Grimble, 2002; Calder, 2004a). COX: cyclooxygenase; LOX: lipooxygenase; PG: prostaglandin; LT: Leukotriene; TX: thromboxanes. Calder, 2004b; Calder, 2003). The effects of n-3 and n-6 PUF As on inflammation and immunity has attracted much attention since atherosclerosis has, in part, an inflammatory basis (Ross, 1993; Chilton, 2004; Plutzky, 2003; Ross, 1999). Therefore, the ratio of n-6:n-3 fatty acids in both membranes and the diet may have important physiological implications (Wijendran and Hayes, 2004).

1.3.5 The link between dietary lipids and CHD

In the early 1950s it was known that populations with a high intake of saturated fat showed increased levels of atherosclerosis and CHD (Renaud and Lanzmann-Petithory, 2001). However, it was the Seven Countries Study (Keys, 1997), which clearly demonstrated that dietary saturated fatty acids (SFA) were a major environmental risk factor for CHD. Follow-up reports of the Seven Countries Study indicate that CHD mortality is closely and directly related to the intake of SFA and cholesterol (Panagiotakos et al., 2003; Ascherio, 2002; Kromhout et al., 1995). Over the last 50 years several large epidemiological studies have shown a strong relationship between high dietary saturated fat intake, increased plasma cholesterol, and CHD that has become known as diet-heart hypothesis (Chahoud et al., 2004; Panagiotakos et al., 2003; Ascherio, 2002; Kromhout et al., 1995; Sacks and Katan, 2002; Burr et al., 1989; GISSI-Trial, 1999; Cagÿula and Mustad, 1997; Kato et al., 1973; Keys, 1997). Since the establishment of the diet-heart hypothesis, many studies have been conducted to test the effects of different diets on CHD mortality (Chahoud et al., 2004; Panagiotakos et al., 2003; Ascherio, 2002; Kromhout et al., 1995; Sacks and Katan, 2002; Burr et al., 1989; GISSI-Trial, 1999).

Dietary interventional trials have confirmed that the type of fatty acid consumed is more important than the amount of total fat in determining the rate of CHD events (Chahoud et al., 2004). Dietary fat, and in particular saturated fatty acids, increase blood cholesterol and lipid levels (Ascherio, 2002; Wolfram, 2003; Stender and Dyerberg, 2004; Eckel et al., 2001b; Eckel et al., 2001a; Sacks and Katan, 2002; Chahoud et al., 2004). Unlike long-chain SFA and cholesterol, which markedly increase LDL-C production and concentration in plasma (Eckel et al., 2001a), the effect of shorter chain fatty acids (Cô-Cn) on plasma LDL-C levels is poorly documented (Kromhout et al., 1995; Tholstrup et al., 2004). Medium-chain fatty acids are absorbed into the portal vein and are rapidly oxidised by the liver, since the mitochondrial membrane is permeable to short- and medium-chain fatty acids (Olpin, 2004; Ramirez et al., 2001; St-Onge and Jones, 2002; Kromhout, 1999), whereas dietary LCFAs are packed into chylomicrons and bypass the liver via the lymphatic system, favouring uptake of fatty acids into adipose and muscle tissues (St-Onge and Jones, 2002; Leonhardt and Langhans, 2004).

In contrast to the atherogenic effect of saturated fatty acids, PUFAs display cardioprotective effects and reduce plasma total cholesterol and LDL-C (Ascherio, 2002; Mori and Beilin, 2001; Kris-Etherton et al., 2003), thereby correcting blood lipid profiles and reduce the risk of CHD mortality (Calder, 2004b; Wijendran and Hayes, 2004; Sacks and Katan, 2002; Chahoud et al., 2004; Calder, 2004b). Replacement of saturated fat with monounsaturated fat in the diet has been shown to yield a 30% reduction in CHD (Sacks and Katan, 2002). Similarly, the consumption of polyunsaturated fatty acids (PUFA) has been associated with a reduced risk of CHD and lower mortality (Oomen et al., 2000; Angerer and von Schacky, 2000; Hu et al., 2002). The beneficial effects of dietary PUFAs on cardiovascular diseases are now well established (Hu et al., 2002; Das, 2000; Wijendran and Hayes, 2004; Jump, 2004) and are discussed below in more detail.

Unlike cis unsaturated fatty acids, dietarytrans fatty acids {trans-VA) are regarded as atherogenic fat (2-to 10-fold greater than saturated fatty acids) (Stender and Dyerberg, 2004; Wolfi-am, 2003) since they lower serum high-density lipoprotein-cholesterol (HDL-C), increase LDL-C, decrease LDL particle size, and impair vascular endothelial function (De Roos et al., 2001; Mauger et al., 2003). Thus, high trans-VA intake is associated with increase risk of CHD and mortality (Dyerberg et al., 2004), but up to 1% of daily dietary energy intake as trans-¥A has been proposed to have no obvious damaging effects (De Roos et al., 2001).

1.3.6 Cardioprotective effects of PUFA

Linoleic acid (n-6 PUFA) is probably the most potent dietary PUFA in lowering serum cholesterol (Renaud and Lanzmann-Petithory, 2001; Calder, 2004a; Calder, 2004b; Wijendran and Hayes, 2004). By contrast, long-chain n-3 PUFA (a major component of fish oils) supplementation reduces circulating triacylglycerol (TAG) levels and influences

13 the susceptibility of LDL to oxidation (Leigh-Firbank et al., 2002). It has been proposed that eicosapentaenoic and docosahexaenoic n-3 fatty acids may be responsible for the low rates of CHD in populations with high fish consumption, such as the Greenland Inuits and Japanese (Das, 2000), as confirmed by the DART and GISSI dietary intervention trials and this is often referred to as the “fish oil effect” (GISSI-Trial, 1999; Burr et al., 1989). It has been suggested that three mechanisms contribute to the strong protective effect of long-chain n-3 fatty acids (Calder, 2004a; Calder, 2004b), particularly EPA and fish oil, towards cardiovascular events:

(i) An anti-thrombotic effect mediated through eicosanoid generation. When fish oil is consumed, EPA is incorporated into cell membrane phospholipids leading to a decrease in the amounts of AA in the cell membrane and therefore eicosanoid - synthesis (Calder, 2001; Calder, 2004a; Calder, 2004b). Furthermore, EPA, is released by phospholipase A] hydrolysis of phospholipids competitively inhibits the oxygenation of AA by cyclooxygenases (Calder and Grimble, 2002), reducing the likelihood of a thrombotic effect (Calder and Grimble, 2002; Calder, 2004b).

(ii) An anti-arrhythmic action by decreasing cardiomyocyte membrane electrical excitability. The presence of n-3 PUFAs in cardiomyocyte membranes decreases electrical excitability and modulates the activity of ion channels. This promotes electrical stability and prevents often fatal cardiac arrhythmias (Calder, 2004b).

(iii) An anti-inflammatory effect on the inflammatory progression of atherosclerosis. Long-chain n-3 PUFAs stabilise atherosclerotic plaques by decreasing infiltration of inflammatory and immune cells into the plaques and/or the function of those cells once in the plaque (Calder, 2004a; Calder, 2004b).

PUFAs are not only potentially strong cardioprotectants (Calder, 2004a), but also play a key role in the prevention of obesity, diabetes, and cancer (De Lorgeril et al., 1998; Ntambi and Bene, 2001; Wolfram, 2003; Norman et al., 2004). Though both n-6 and n-3 PUFAs have distinct biological effects that contribute to their cardioprotective action, the absolute levels of intake (g/day) and n-6:n-3 ratio required to achieve optimal CHD health are somewhat controversial (Wijendran and Hayes, 2004; Renaud and Lanzmann- Petithory, 2001). The International Society for the Study of Fatty Acids and Lipids (ISSFAL) statement on the recommended dietary supplementation for n-6 and n-3 fatty acids (based on a daily dietary energy intake of 2000 kcal) proposed 2%-3% from LA,

14 1% from ALA, and 0.3% from EPA+DHA, providing a n-6:n-3 ratio of 2:1 (Simopoulos et al., 2000). While some investigators propose that dietary supplementation of n-6 and n-3 fatty acids should not be higher than 5% of energy intake with a ratio of n-6:n-3 PUFA in the range of 3-5:1 (Renaud and Lanzmann-Petithory, 2001), the latest recommendation is an intake of about 6% LA, 0.75% ALA, and 0.5% from EPA+DHA corresponding to a n-6:n-3 ratio of ~6:1 (Wijendran and Hayes, 2004).

1.3.7 Delivery of fatty acids to the cardiomyocytes

Fatty acids are delivered to cardiac myocytes in three ways (Augustus et al., 2003; Glatz et al., 2003; Glatz and Storch, 2001). First, fatty acids are generated in the local capillary by lipoprotein lipase-mediated hydrolysis of TAG in circulating chylomicrons (CM) and very low-density lipoprotein (VLDL) (Hauton et al., 2001; Augustus et al., 2003; Mardy et al., 2001; Niu et al., 2004). Second, fatty acids are produced from intracellular hydrolysis of TAG in the core of internalised lipoproteins. Finally, fatty acids are obtained from circulating plasma albumin, and this represents the primary source of FA for the heart (Augustus et al., 2003; Niu et al., 2004). Cardiac fatty acid uptake and transport are mediated by a family of specific proteins that includes membrane-associated and cytoplasmic fatty acid-binding proteins (Glatz and Storch, 2001; Glatz et al., 2003).

1.4 Membrane-associated fatty acid transporters

It was originally believed that the uptake of long-chain fatty acids into cardiac myocytes was purely mediated by diffiision across the membrane (DeGrella and Light, 1985), as demonstrated by studies using model membrane systems (Kamp et al., 1993; Hamilton and Kamp, 1999; Kamp et al., 1995). However, this diffusional mechanism has been challenged by the identification of a number of membrane-associated fatty acid transporters (Abumrad et al., 1999) and it is now thought that membrane fatty acid transport is predominantly protein-mediated (Frohnert and Bemlohr, 2000; Pohl et al., 2004a; Pohl et al., 2004c; Bonen et al., 2004). These findings suggest that the mechanisms of passive and facilitated diffiision of fatty acids coexist, and their relative

15 contribution to overall fatty acid transport may depend on the cell type, hormonal states and the availability of fatty acids (Glatz and Storch, 2001). The tissue content of fatty acid transporters is responsive to various dietary, hormonal and pharmacological manipulations, as well as pathological states (Abumrad et al., 1999).

Cellular studies demonstrated a high-affinity protein-facilitated component in the transport of fatty acids across the cellular membrane (Gimeno et al., 2003; Glatz et al., 2003). This led to the identification of three major membrane fatty acid transporting proteins (Brinkmann et al., 2002): fatty acid translocase (FAT/CD36), plasma membrane fatty acid-binding protein (FABPpm), and fatty acid transport protein (FATP). All three proteins are present in the heart and are found to be associated with the cardiomyocyte plasmalemmal (van Nieuwenhoven et al., 1999). Several other membrane-bound proteins, such as fatty acid receptor (FAR) (Fujii et al., 1987), adipocyte differentiation related protein (ADRP) (Gao and Serrero, 1999) and caveolin (Trigatti et al., 1999), have also been implicated in fatty acid transport (Table 1.2).

Table 1.2: Membrane-associated proteins involved in fatty acid uptake and transport (Abumrad et al., 1999; Gimeno et al., 2003; Hajri and Abumrad, 2002).

Protein MW^ (kDa) Tissue locaiisation

liver, heart, muscle, brain, epidermis, kidney, intestine, testis, FAT/CD36 88 adipose tissue, spleen

FABPpm 40 liver, heart, muscle, intestine, adipose tissue

FATP1 63 heart, muscle, brain, kidney, epidermis, lung, adipose tissue

FATP2 62 liver, kidney, intestine

FATP3 64 lung, liver, testis

FATP4 65 intestine, heart, muscle, lung, brain, testis, kidney, spleen

FATP5 63 liver

F ATP6 70 heart

Caveolin 22 lung, adipose tissue, endothelium

ADRP 53 adipose tissue, liver

FAR 56-60 heart + MW: molecular weight

16 Following dissociation of fatty acids from plasma albumin, the transmembrane translocation of fatty acids most likely takes place either by passive diffusion through the lipid bilayer, and/or active transport facilitated by membrane-associated proteins (Fig. 1.4). The chain length of fatty acids and degree of unsaturation influenee their rate of uptake by the cell (van der Vusse et al., 2000). Fatty aeids that enter the cell by the membrane-associated transporters FABPpm and FAT/CD36, are delivered to intracellular acyl-eoenzyme A synthetase (ACS) for activation and estérification to acyl-CoA (Fig. 1.4) (Hertzel and Bemlohr, 2000). Non-esterified fatty acids bound to intracellular fatty acid-binding proteins (FABP), such as heart-type fatty acid-binding protein (H-FABP) that shuttle LCFAs to metabolic pathways. In addition to FABPpm and FAT/CD36 mediated transport, fatty acids also are taken-up by the cells through interaction with FATP, and are immediately activated to their acyl-CoA esters by sarcolemmal acyl-CoA synthetase (Glatz et al., 2003; Pohl et al., 2004c; Hertzel and Bemlohr, 2000). Aeyl-coenzyme A then binds to aeyl-CoA binding protein (ACBP), and translocates to different sites in intracellular metabolic pathways (Fig. 1.4) (Glatz and Storch, 2001; Glatz et al., 2003; Brinkmann et al., 2002).

interstftmi space Aibumm

FATP

acyl-CoA(ACBp) FAlH-FABP

f Signal Acylation Oxidation Estérification transduction

Figure 1.4: Molecular mechanism of uptake and utilization of long-chain fatty acids (Glatz et al., 2003). Fatty acids enter the cells through simple diffusion across the cell membrane or facilitated by membrane proteins such as FATP, FAT/CD36, and FABPpm. FATP functions as both fatty acid transporter and acyl-CoA synthetase. Fatty acid or acyl-CoA bound to a binding protein is transported to intracellular metabolic pathways. ACS: acyl-CoA synthetase; ACBP: acyl-CoA binding protein; FA: fatty acid; FAT: fatty acid translocase (FAT/CD36); FATP: fatty acid transport protein; FABPpm: plasmalemmal fatty acid-binding protein; H-FABP: heart-type fatty acid-binding protein. The first putative membrane-bound fatty acid transporter to be identified was a 40 kDa protein termed plasmalemmal fatty acid-binding protein (FABPpm) (Stremmel et al., 1985). Glatz et al. (1994) reported that FABPpm is an abundant component of the plasma membrane and it is estimated to constitute 2% of plasmalemmal protein and may co-operate with fatty acid translocase (FAT/CD36) to translocate fatty acids across the membrane (van der Vusse et al., 2000; Roepstorff et al., 2004).

Harmon and Abumrad (1993) identified an 88 kDa membrane protein involved in the rat adipose uptake and transport of fatty acids, that they referred to as FAT (fatty acid translocase), but this protein is homologous to human CD36 (Tandon et al., 1989). FAT/CD36 recognizes a broad range of ligands including oxidized LDL, anionic phospholipids, apoptotic cells, collagen, and fatty acids and is responsible for a large fi-action of fatty acid uptake by the heart (Hajri and Abumrad, 2002). FAT/CD36 is located in the plasma membrane, as well as intracellular pools of heart and skeletal muscle cells (Campbell et al., 2004; Bonen et al., 2004; Luiken et al., 2004). Translocation of FAT/CD36 fi*om the intracellular pools to the plasma membrane is induced by insulin and contraction (Chabowski et al., 2004; Luiken et al., 2004; Bonen et al., 2004). It is possible that FAT/CD36 acts as an acceptor or docking protein, either handing fatty acids to the other proteins or simply assisting their movement through hydrophobic environments (Campbell et al., 2004; Abumrad et al., 1999). It has recently been shown that FAT/CD36 is associated with lipid rafl;s in the plasma membrane that are used as platforms to bind or transport LCFA (Pohl et al., 2004a; Pohl et al., 2004b; Pohl et al., 2004c). Since disruption of lipid rafts decreases binding of LCFAs to FAT/CD36 it is likely that plasma membrane lipid rafts regulate the cell surface availability of FAT/CD36 and therefore control LCFA uptake and transport (Pohl et al., 2004a; Pohl et al., 2004b; Pohl et al., 2004c; Pohl et al., 2005).

Schaffer and Lodish (1994) identified a 63 kDa membrane-associated protein, named fatty acid transport protein (FATP), which increases the uptake of LCFA. Recently, a large family of FATPs has been reported (Table 1.2) (Gimeno et al., 2003). FATPs have sequence similarity to very-long-chain fatty-acyl CoA synthetase and possibly function both as a LCFA transporter and as an acyl-CoA synthetase (Frohnert and Bemlohr, 2000; Abumrad et al., 1998; Pohl et al., 2004a). An AMP-binding sequence that mediates acyl-CoA synthetase activity is found in all known FATPs (Coe et al., 1999; Pohl et al., 2004a; Pohl et al., 2004c). The different members of the FATP family show

18 characteristic expression patterns in different fatty acid metabolising tissues (Gimeno et al., 2003; Hajri and Abumrad, 2002). FATP4 is mainly expressed in the intestine. Likewise, FATPS and FATP6 are predominantly expressed in the liver and heart tissue, respectively, whereas FATPl is the major FATP in adipose tissue, but is also found in heart and skeletal muscle cells (Gimeno et al., 2003; Pohl et al., 2004c).

1.5 Intracellular Fatty Acid-Binding Proteins (FABPs)

1.5.1 Classification of intracellular FABPs

Intracellular fatty acid-binding proteins (FABP) are members of a conserved multi-gene family of the intracellular lipid-binding proteins (iLBPs) encoding 14-15 kDa proteins (Table 1.3) (Hertzel and Bemlohr, 2000; Haunerland and Spener, 2004; Veerkamp and Maatman, 1995). FABPs were discovered in the early 1970s as abundant cytoplasmic proteins that bind long-chain fatty acids in a non-covalent and reversible manner (Coe and Bemlohr, 1998; Ockner et al., 1972), and show a characteristic pattem of tissue distribution (Hertzel and Bemlohr, 2000; Zimmerman et al., 2001). At least 16 members of this protein family have been identified with 22%-73% amino acid identity (Table 1.3) (van der Vusse et al., 2000; Adida and Spener, 2002; Haunerland and Spener, 2004). Some FABPs, such as I-FABP and T-FABP, are found exclusively in one tissue, whereas others, such as L-FABP and H-FABP, are expressed in more than one tissue (Table 1.3), suggesting that these proteins have evolved separately in order to fulfil different physiological functions (Massolini and Calleri, 2003). FABPs are classified into four subfamilies, namely subfamily I, which comprises proteins specific for vitamin A derivatives; subfamily II, containing proteins with larger binding sites that allow binding of bulkier ligands, such as bile acids and heme; subfamily III with only one member (I-FABP), which is only expressed in intestine; and subfamily IV that includes the remaining FABPs (Haunerland and Spener, 2004).

Several functions have been proposed for the FABPs, and these include: (i) modulation of enzyme activity related to lipid metabolism; (ii) maintaining the fatty acid composition of cell membranes; and (iii) regulation of gene expression by interacting with LCFA-activated nuclear receptors, such as the peroxisome proliferators-activated

19 receptors (PPAR) (Storch and Thumser, 2000). Consistent with these general concepts, the FABP content of the most cells is generally proportional to the rate of fatty acid metabolism, the latter being highest in hepatocytes, cardiomyocytes, and adipocytes (Haunerland arid Spener, 2004).

In summary, membrane-associated and cytoplasmic LCFA transporters could co-operate to facilitate efficient LCFA uptake and utilization (Stahl et al., 2001). However, the fate of cellular LCFA depends on the cellular energy demand, pathophysiological states and the physiochemical properties of the FA (Brinkmann et al., 2002; Ferdinandusse et al., 2003).

Table 1.3: The FABP protein family.

ILBP-type Gene Expression Ligands

Subfamily 1 CRBP l-IV RBP1-4 Ubiquitous mainly in liver, kidney, lung Retinol CRABP 1, II RBP5, CRABP2 Ubiquitous mainly in epidermis and brain Retinoic acid

Subfamily II L-FABP FABP1 Liver, intestine, kidney, lung, and pancreas LCFA, acyl-CoA, and heme I-FABP FABP6 Ileum bile acids

Subfamily III I-FABP FABP2 Intestine, liver LCFA

Subfamily IV H-FABP FABPS Heart, skeletal muscle, brain, kidney, lung, LCFA mammary, placenta, testis, ovary, and stomach A-FABP FABP4 Adipose tissue, macrophages LCFA E-FABP FABPS Skin, adipose tissue, lung, brain, heart, LCFA skeletal muscle, testis, retina, and kidney B-FABP FABP7 Brain, retina, and Central Nervous System LCFA and DHA M-FABP FABPS Brain and Peripheral Nervous System LCFA T-FABP FABP9 Testis LCFA Table was compiled from data in reviews (Hertzel and Bemlohr, 2000; Haunerland and Spener, 2004; Storch and Thumser, 2000) as well as utilizing the SwissPort Database (http://ca.expasy.org/sprot/sprot-top.html).

20 1.5.2 Ligand-binding characteristics of FABPs

All FABPs contain 127-132 amino acids (Massolini and Calleri, 2003) and are composed of two a-helices and 10 anti-parallel p-strands organised into two p-sheets (Fig. 1.5) (Storch and Thumser, 2000). The p-sheets create an internal cavity lined with a large number of hydrophobic amino acid side chains, although the cavity also contains several polar amino acids (Massolini and Calleri, 2003; Zanotti, 1999). The fatty acid ligand binds in the hydrophobic cavity, with the carboxylate group oriented inwards, coordinated by tyrosine and arginine residues within the binding pocket (Zanotti, 1999; Hertzel and Bemlohr, 2000). The binding pocket of L-FABP (subfamily II) is considerably larger than that of other FABPs, allowing the binding of two fatty acid molecules (Haunerland and Spener, 2004; Pohl et al., 2005). Despite the overall similarity of FABPs, the ligand-binding configuration is specific for each FABP, though only slightly dependent on the type of fatty acid itself (Zanotti, 1999). For instance, I-FABP binds fatty acids in an almost linear conformation (Adida and Spener, 2002), whereas the members of subfamily IV bind fatty acids in a highly bent U-shaped conformation (Zanotti, 1999; Adida and Spener, 2002; Haunerland and Spener, 2004). %

Figure 1.5: Structure of I-FABP showing the configuration of bound palmitic acid.The a-helical portal domain (a-I and a-II) is shown in red, whereas the ten anti-parallel P-strands are labelled A through J and represented in blue. Palmitic acid is shown bound to the interior of the P-sheet domain demonstrating linear conformation with the carboxylate end oriented inwards and the acyl chain extending towards the a-helical portal domain (Storch and Thumser, 2000).

21 FABPs bind long-chain fatty acids (C 16-C20) with high affinity (Kd <500 nM) and a molar stoichiometry of 1:1, with the exception of L-FABP, which can bind two LCFA simultaneously (Thompson et ah, 1997; Storch and Thumser, 2000; Richieri et al., 2000; Adida and Spener, 2002; Massolini and Calleri, 2003). Ligand affinity is dependent on FABP-type, the length of acyl chain, and degree of fatty acid unsaturation (Massolini and Calleri, 2003). In general, affinities decrease with decreasing chain length, i.e. reduction in hydrophobicity, and increasing saturation (Richieri et al., 2000; Massolini and Calleri, 2003). LCFA binding affinity varies according to FABP type, with brain = myelin = heart > liver >intestine >adipocyte (Richieri et al., 2000; Zimmennan et al., 2001; Massolini and Calleri, 2003).

FABPs shuttle LCFA to metabolic pathways and LCFA-activated nuclear receptors in the nucleus, acting as a “cytosolic gateway”. Fatty acids and several of their metabolites function as messengers in cell signalling pathways and cell differentiation (Fig. 1.6) (Hertzel and Bemlohr, 1998; Calder, 2001; Jones-Villeneuve et al., 1983). They may also affect the abundance of transcription factors and regulate the activity of

Mitochondria FA Peroxisomes N u c le u s Lipid droplets

Transcription factor complex

Membrane composition mRNA metabolism Transcription factor activity/abundance P ro te in ^ activity/abundance

/^Signal Biological response transductiorv Metabolism Cell growth Differentiation

Figure 1.6: Intracellular transport of fatty acids.Fatty acids entered the cell may influence on the membrane composition and intracellular transduction or may oxidised in the mitochondria. FAs also affect the abundance of transcription factors and regulate the activity of nuclear receptors that control gene expression. FA: fatty acid (Jump, 2004; Storch and Thumser, 2000). nuclear receptors that control gene expression related to cell growth, differentiation, and metabolism (Jump, 2004; Khan and Vanden Heuvel, 2003; Clarke, 2001; Jump, 2002a; Jump, 2002b; Ntambi and Bene, 2001). The activation of nuclear receptors allows them to exert their biological function on gene expression (Yaqoob, 2004; Tan et al., 2002). LCFAs modulate gene transcription by interacting with transcription factors such as peroxisome proliferator-activated receptors (PPARs), liver X receptor (LXR), hepatic nuclear factor (HNF4-a) and sterol regulatory element-binding protein (SREBP) (Sampath and Ntambi, 2004). In cardiac and skeletal muscle cells, the expression of lipid-related genes is mainly under control of PPARs (Gilde and van Bilsen, 2003; Guan et al., 2002).

Thus, fatty acids and other lipids present in our diets are not only nutritionally important but also serve as ligands, or precursors for ligands, that bind to receptors in the nucleus (Chawla et al., 2001) and positively or negatively regulate the expression of lipid-related genes and the synthesis of proteins involved in lipid metabolism (Jump, 2002a; Jump, 2004; Jump and Clarke, 1999). Many of these genes encode for proteins that play a role in fatty acid transport or metabolism (Table 1.4) (Duplus et al., 2000; Jump, 2004; Clarke, 2000; Pegorier et al., 2004; Lee et al., 2003b; Ntambi and Bene, 2001). Changes in the amounts of these proteins is an adaptive response by cells to the variation in fatty acid concentration and availability (Duplus and Forest, 2002). Table 1.4: Examples of fatty acid-regulated genes involved in lipid metabolism and their corresponding function.

Regulation Gene Function

Glucose-6-phosphatase (G6Pase) Glycogenolysis Glucokinase (GK) Glycolysis Liver-pyruvate kinase (L-PK) Glycolysis Giucose-6-phosphate dehydrogenase (GePD) Glycolysis ATP citrate-lyase (ACL) Lipogenesis Fatty acid synthase (FAS) Lipogenesis Down-regulation Acetyl-CoA carboxylase (ACC) Lipogenesis 814 protein (814) Lipogenesis 8tearoyl-CoA desaturase-1 (8CD-1 ) Desaturation of FA A5-desaturase A5-desaturation A6-desaturase A6-desaturation Apolipoprotein-CIII (apo-CIII) Lipoprotein metabolism Fatty acid translocase (FAT/CD36) Membrane FA transport Liver-fatty acid-binding protein (L-FABP) Intracellular FA transport Heart-fatty acid-binding protein (H-FABP) Intracellular FA transport Adipocyte-fatty acid-binding protein (A-FABP) Intracellular FA transport Lipoprotein lipase (LPL) Lipoprotein metabolism Apolipoprotein-AI (apo-AI) Lipoprotein metabolism Apolipoprotein-AII (apo-AI 1) Lipoprotein metabolism Up-regulation Acyl-CoA synthetase (ACS) Fatty acid activation Acyl-CoA oxidase (AOX) Peroxisomal FA p-oxidation Cytochrom P450A2 (CYP4À2) Microsomal FA m-oxidation Carnitine palmitoyltransferase-1 (CPT-1 ) Acyl-chain transporter Uncoupling protein 2 and 3 (UCP2, UCP3) Controlling energy production Phosphoenolpyruvate carboxykinase (PEPCK) Glyceroneogenesis Glucose transporter (GLUT4) Glucose transport FA: fatty acid. Table was compiled from data in reviews (Jump and Clarke, 1999; Duplus et al., 2000; Duplus and Forest, 2002; Jump et al., 1994; Fruchart et al., 2001; Pegorier et al., 2004)

24 1.6 Nuclear Receptors

1.6.1 Classification of nuclear receptors

Nuclear receptors function as ligand-activated transcription factors that regulate the expression of target genes to affect processes as diverse as reproduction, development, differentiation and general metabolism (Chawla et al., 2001; Francis et al., 2003; Schulman and Heyman, 2004). To date, more than 70 gene products have been identified as members of the steroid receptor superfamily, primarily based on their similarity to known hormone receptors. However, the latest estimate of the number in the sequence, based on sequence alignment, is 48 nuclear receptor genes (Fig 1.7) (Zhang et al., 2004; Sladek and Giguere, 2000; Francis et ah, 2003; Maglich et al., 2001; Robinson-Rechavi et al., 2001). Nuclear receptors were first recognized as mediators of steroid hormone signalling (Nagy and Schwabe, 2004) and now include a large superfamily of ligand-modulated transcription factors (Nagy and Schwabe, 2004; Benoit et al., 2004). Nuclear receptors can be categorized into distinct functional classes according to their ligand binding properties (Benoit et al., 2004), including the hormone receptors, nutritional sensors, receptors with structural ligands, and a large number of so-called orphan nuclear receptors, whose ligands, target genes, and physiological functions were initially unknown (Fig 1.7) (Sladek and Giguere, 2000; Giguere, 1999; Escher and Wahli, 2000). The majority of nuclear receptors are constitutively localised in nuclei, although most steroid hormone receptors shuttle between the cytoplasm and the nucleus (Benoit et ah, 2004; Handschin and Meyer, 2005). H o rm o n e Nutritional Receptors with O rp h a n r e c e p to r s s e n s o r s structural r e c e p to r s lig a n d s

Steroid-hormone ROR, Empty LBP ; ' receptors PPAR a. p, T HNF-4 a. Y LRH-1 LXR ERcvp «.P ROR „ .. PR FXR ERRa, p. y AR PXR/SXR GR CAR No LBP MR NGFI- Nurr1 Non-steroid Norl hormone receptors Rev-Erb'a, P RAR.'a. P Y TR,V ,p VDR Unknown COUP-TF^ a. P. Y SF-1 GCNF DAX-1 SHP ILL ’ PNR TR2,4

Figure 1.7: Classification of nuclear receptors superfamily. 48 distinct members of the nuclear receptor superfamily were identified in the human genome. Nuclear receptors are categorised into subgroups, depending on type of their ligand and/or ligand-binding properties (Francis et al., 2003; Zhang et al., 2004; Maglich et al., 2001; Robinson-Rechavi et al., 2001; Chawla et al., 2001; Rangwala and Lazar, 2004; Benoit et al., 2004). AR: androgen receptor; CAR; constitutive androstane receptor; COUP-TF: chicken ovalbumin upstream promoter transcription factor; DAX: dosage-sensitive sex receptor; ER: estrogen receptor; ERR: ER-related receptor; FXR: famesoid X receptor; GCNF: germ cell nuclear factor; GR: glucocorticoid receptor; HNF4: hepatic nuclear factor-4; LRH-1: liver receptor homolog-1; LXR: liver X receptor; MR mineralocorticoid receptor; NGFI-B: nerve growth factor induced B; Nor 1: neuron-derived orphan receptor; Nurr 1: Nur-related factor 1; PNR: photoreceptor-specific nuclear receptor; PPAR: peroxisome proliferator-activated receptors; PR: progesterone receptor; PXR: pregnane X receptor; RAR: retinoic acid receptor; ROR: RAR-related orphan receptor; Rev-Erb: reverse-Erb; RXR: retinoid X receptor; SF-1: steroidogenic factor-1; SHP: short heterodimer partner; TEL: tailless; TR: thyroid hormone receptor; TR2,4: testicular receptor 2 and 4; VDR: vitamin D hormone receptor; LBP: ligand-binding pocket.

1.6.2 Structure of nuclear receptors

The structural organization of nuclear receptors is similar despite a wide variation in ligand sensitivity (Aranda and Pascual, 2001; Xu et al., 2001) and is composed of five different domains (Fig. 1.8) (Sladek and Giguere, 2000; Blanquart et al., 2003). The

NH2-terminal region, termed domain A/B, varies in both sequence and length and harbours a ligand-independent transcriptional activation function (AF-1). The

26 DNA-binding domain (DBD), or C domain, contains two highly conserved zinc-fmger motifs that target the receptor to DNA. The hinge region (domain D) pennits protein flexibility and allows simultaneous protein dimerization and binding to DNA. All nuclear receptors also have a large COOH-terminal region (Domain E) that encompasses the ligand-binding domain (LBD), dimerization interface, and ligand-dependent activation factor-2 (AF-2) (Khan and Vanden Heuvel, 2003; Gilde and van Bilsen, 2003; Willson et al., 2001). The function of domain F has not been identified (Blanquart et al., 2003).

A/BC D E F

AF-1 6 ^ 4 . j

•Hypervariable •Highly •Hypervariable •Moderately conserved •T ransactivation conserved •Co-regulator •Ligand binding •Influence DNA •DNA binding interaction •Dimerization binding activity •Co-factor •Influence DNA •Transactivation interaction binding activity •Co-regulator interaction

Figure 1.8: Schematic representation of the structure of nuclear receptors. The domain A/B contains the ligand-independent Activation Function-1. Domain C is the DNA-binding domain, the hinge region (domain D) allows the receptor to change conformation, and domain E contains the Ligand-Dependent Activation Function-2. The function of domain F is unknown (Gilde and van Bilsen, 2003; Blanquart et al., 2003; Sladek and Giguere, 2000).

Nuclear receptors are switched between active and inactive states by binding hydrophobic ligands (Nagy and Schwabe, 2004). The ligand-binding domain of nuclear receptors is formed by 10-13 a-helices, 2-5 P-strands and connecting loops (Schulman and Heyman, 2004; Nagy and Schwabe, 2004), creating a deep ligand-binding pocket (LBP) that accommodates the lipophilic ligands (Benoit et al., 2004). The size, shape, and flexibility of the LBP, together with its composition of hydrophobic amino acid residues, detenuines the specificity and efficiency of ligand binding (Benoit et al., 2004; Nagy and Schwabe, 2004). For example, steroid and thyroid hormones bind to their respective nuclear receptors at low nanomolar concentrations, whereas PPARs and LXRs are activated by dietary fatty acids, cholesterol, and their metabolites at micromolar levels (Nagy and Schwabe, 2004; Benoit et al., 2004; Schulman and Heyman, 2004; Handschin and Meyer, 2005). In addition, the LBP of some nuclear receptors, e.g. PXR, is flexible and can expand when the nuclear receptor interacts with a large ligand (Benoit et al., 2004). Some nuclear receptors, e.g. LHR-1 and ERRy, are in an active conformation even in the absence of exogenously added ligands (Benoit et al., 2004).

1.7 Peroxisome Proliferator-Activated Receptors (PPARs)

Issemann and Green (1990) reported the discovery of a ligand-activated transcription factor that specifically interacted with drugs that were known to cause peroxisomal proliferation in the liver. Accordingly, this transcription factor was referred to as the Peroxisome Proliferator-Activated Receptor (PPAR) (Gilde and van Bilsen, 2003; Smith, 2002). Since the discovery of this PPAR, latter named PPARa, two other PPAR isoforms have been cloned (Lemberger et al., 1996). Although the PPARs were originally cloned as orphan receptors (Xu et al., 2001) this subfamily of nuclear receptors can be activated by both fatty acids and their metabolic derivatives, and thus serve as lipid sensors (Evans et al., 2004). Numerous studies performed during the last decade have revealed that PPARs are implicated in several physiological processes, such as the regulation of lipoprotein and lipid metabolism, inflammatory responses, glucose homeostasis and cellular differentiation (Khan and Vanden Heuvel, 2003; Smith, 2002; Blanquart et al., 2003; Wahli, 2002; Manco et al., 2004).

1.7.1 Distribution of PPAR isoforms

To date, three isoforms of PPARs have been identified and these are PPARa, PPARp (also referred to as PPARS), and PPARy (Khan and Vanden Heuvel, 2003; Mandard et al., 2004). The tissue expression pattern of these PPAR isoforms (a, P, y) varies (Gilde and van Bilsen, 2003). PPARa is relatively abundant in tissues with a high capacity for fatty acid oxidation, such as liver, kidney, skeletal muscle, heart, brown adipose tissue and intestine (Escher and Wahli, 2000; Finck and Kelly, 2002; Kersten, 2002; Ferre, 2004; Mandard et al., 2004). PPARp is expressed ubiquitously and plays key functions in the regulation of lipid metabolism and energy homeostasis (Khan and Vanden Heuvel, 2003; Ferre, 2004; Evans et al., 2004; Dressel et al., 2003; Grimaldi, 2005). PPARy on the other hand, displays a tissue-selective pattern of expression (Smith, 2002). Three different PPARy transcripts, termed PPARyl, PPARy2 and PPARy3, are derived from the PPARy gene (Willson et al., 2001; Francis et al., 2003). However, PPARyl and PPARy3 mRNA give rise to the same protein product having 28 amino acids less than PPARy2 in NH 2-terminal (Willson et al., 2001; Francis et al., 2003; Ferre, 2004). PPARyl is broadly expressed, albeit at low levels (Willson et al., 2001), predominantly in adipocytes (Smith, 2002) whereas both PPARy2 and PPARy3 are highly expressed in adipose tissue (Willson et al., 2001; Smith, 2002). However, the functional significance of the multiple isoforms of PPARy is currently unclear (Willson et al., 2000).

1.7.2 Transcriptional activity of PPARs

All members of the nuclear receptor superfamily regulate transcription by binding to specific DNA sequences, known as hormone response elements, located in the 5'-flanking region of the target gene (Aranda and Pascual, 2001). In contrast to steroid nuclear receptors, which usually bind to these DNA response elements as homodimers (Khan and Vanden Heuvel, 2003), PPARs bind as a heterodimer with retinoid X receptor (RXR) (Fig. 1.9), the latter being activated by 9-cis retinoic acid (Escher and Wahli, 2000; Willson et al., 2001). The PPARiRXR heterodimer complex binds to a PPAR response element (PPRE), which consists of two direct repeat sequences of six nucleotides (consensus sequence: AGGTCA) separated by a single nucleotide spacer (Fig. 1.9) (Khan and Vanden Heuvel, 2003; Wahli, 2002; Willson et al., 2001 ; Frohnert et al., 1999).

PPARiRXR heterodimer-mediated transcription depends critically on the balance between co-activator and co-repressor complexes (Benko et al., 2003), which is controlled by availability of ligand (Fig. 1.9) (Escher and Wahli, 2000; Benko et al., 2003). In the absence of ligand, the nuclear receptor heterodimer complex, when bound to the relevant hormone response element, is associated with co-repressor proteins (and histone deacetylase), which maintain the chromatin in a compact state, and therefore transcription is repressed (Renaud and Moras, 2000). However, ligand binding triggers confonnational changes in PPAR and RXR that are accompanied by dissociation of the co-repressor complex and recruitment of a co-activator complex with histone acetyltransferase (HAT) activity. This results in histone acétylation, local unwrapping of

29 chromatin and activation of gene transcription (Fig. 1.9) (Jepsen and Rosenfeld, 2002; Escher and Wahli, 2000; Aranda and Pascual, 2001).

# ■ Ligands m

/ PPRE ^ / \ / \ General AAjTCTAGGTCAAAGGTCA Transcription Factors

Figure 1.9: Diagrammatic representation of the PFAR:RXR heterodimer bound to a PPAR-response element (PPRE). Ligand binding promotes the binding of a co-activator complex that initiates acétylation of histone proteins leading to the localised unwrapping of the chromatin and transcription by RNA polymerase. HAT: histone acetyl-transferase (Escher and Wahli, 2000; Renaud and Moras, 2000).

1.7.3 PPAR ligands

PPARs act as lipid sensors and bind ligands derived from dietary lipids or their metabolie products (Barish and Evans, 2004; Berger and Moller, 2002; Xu et al., 2001). PPARs can be activated by a wide variety of saturated and unsaturated fatty acids (Table 1.5). Long-chain fatty aeids, eicosanoids and inflammatory mediators such as leukotrienes and prostaglandins are natural activators of PPARs (Mandard et al., 2004; Willson et al., 2000; Willson et al., 2001). In addition to endogenous fatty acids and their metabolites, PPARs are also molecular targets of pharmaeeutieal compounds (Table 1.5) (Willson et al., 2000; Evans et al., 2004; Willson et al., 2001). For example, fibrates are PPARa agonists and widely used to treat hypertriglyceridemia (Willson et al., 2000; Evans et al., 2004), while PPARy is the only receptor for the thiazollidinedione

30 (TZD) class of antidiabetic drugs (Xu et al., 2001). Currently there is no specific pharmaceutical ligand for PPARp, although several are in clinical development (Willson et al., 2000; Marx et al., 2004). -

The ligand-binding domains of all PPAR isoforms bave sufficiently divergent amino acid sequences to allow some ligand specificity (Escher and Wahli, 2000; Xu et al., 2001). Unlike other nuclear receptors, PPARs contain a large ligand-binding pocket allowing them to bind bulky ligands, such as phospholipids and synthetic fibrates, as well as the relatively simple fatty acids (Li et al., 2003). Besides the size and shape, the hydrophobic/hydrophilic balance of the ligand-binding pocket plays an important role in ligand-binding specificity (Li et al., 2003). The capacity of fatty acids, differing in chain length and/or degree of unsaturation, as ligands for the three PPAR isoforms has received widespread attention (Gilde and van Bilsen, 2003). Studies demonstrate that both long-chain saturated, mono- and polyunsaturated fatty acids bind to PPARa and PPARp with more or less comparable affinity (Gilde and van Bilsen, 2003). In contrast, PPARy has very low affinity for saturated fatty acids and is preferably activated by PUP As over monounsaturated fatty acids, although a prostaglandin J 2 derivative, 15-deoxy-A^^’^'^-PGJ2 has been widely postulated as the main natural ligand for PPARy receptor (Khan and Vanden Heuvel, 2003; Gilde and van Bilsen, 2003; Forman et al., 1995).

31 Table 1.5: Natural and synthetic ligands for PPARs.

PPA Rs Ligand a P y a-ilnoienic acid (18:3, ALA) + + +

n-3 PU FA ElcGsapentaenoic acid (20:5, ERA) + + + Docosahexaenoic acid (22:6, DHA) + + + Linoleic acid (18:2, LA) + + +

n-6 PU FA y-linolenic acid (18:3, GLA) + + + Dihomo-y-linolenic acid (20:3, DGLA) + + + Arachidonic acid (20:4, AA) + + + Palmitoleic acid (16:1) + + ± n-9 PU FA Natural Oleic acid (18:1) + + ± lig an ds Capric acid (10:0) - --

Saturated FA Palmitic acid (16:0) + + - Stearic acid (18:0) + + - 8-hydroxyeicosapentaneoic (8-HEPE) + - -

8S-hydroxylate epoxyeicosatrienoic (8S-HETE) + - -

Eicosanoids Leukotriene 84 (LTB4) + - - 9-hydroxyoctadenoic (9-HODE) - - + 1 3 -hydroxyoctadenoic (13-HODE) - - +

15-deoxy-A^^’^''-prostagIandin J2 (15D-A^^'^^-PGJ2) -- + Clofibrate + - +

Hypolipidemic Ciprofibrate + - + d ru g s Bezafibrate + + + Pirinixic acid (Wy-14643) + - - S ynthetic Troglitazone (CS045); (TZD) -- + Hypoglycaemic ligands Rosiglitazone (TZD) - - + d ru g s L-796449; non-thiazollidinedione -- + Indomethacin -- + Anti-inflammatory Fenoprofen - - + dru g s Ibuprofen -- + +: binding affinity of ligand to PPAR no affinity has been reported due to these ligands ±: no consensus in the literature of the effects due to these ligands Table was compiled from data in reviews (Khan and Vanden Heuvel, 2003; Gilde and van Bilsen, 2003; Escher and Wahli, 2000; Desvergne and Wahli, 1999; Willson et al., 2000; Kota et al., 2005; Corton et al., 2000). 1.7.4 Synthetic ligands of PPARs

The fibrates encompass a class of clinically utilised hypolipidaemic agents that were first synthesized in the 1950s and came into clinical use in the 1960s (Thorp and Waring, 1962). Clofibrate (ethyl y?-chlorophenoxy isobutyrate) was the first of this class of drugs, which was followed by the introduction of other fibrates, such as bezafibrate, gemfibrozil, ciprofibrate, and fenofibrate (Fig. 1.10) (Bhatnagar, 1998). Gemfibrozil and fenofibrate are currently approved for use in the United States, while bezafibrate and ciprofibrate are mainly prescribed in Europe (Fazio and Linton, 2004). Fibrates bind mainly to PPARa, which is primarily expressed in liver, heart, skeletal muscle, and kidneys and regulate expression of the genes involved in lipid metabolism (Bremer, 2001; Chapman, 2003; Fruchart et al., 2001; Escher and Wahli, 2000). Many studies have been conducted over the past two decades to test the effects of fibrates on the plasma lipid profile and mortality from CHD (Fazio and Linton, 2004; Elkeles et al., 1998; Vakkilainen et al., 2003; Asztalos and Schaefer, 2003; Biijmohun et al., 2005). The wide spectrum of lipid-modulating actions of fibrates is mediated by their capacity to mimic the structure and biological functions of fatty acids (Chapman, 2003; Fazio and Linton, 2004).

© © % // Ci

clofibrcif® P —Eli "m fc^fiCiftbrate

Figure 1.10: Chemical structure of fibrates. Fibrates bind to PPARa and mimic the structure and biological functions of fatty acids (Chapman, 2003). In contrast to fibrates that mainly target PPARa, members of the thiazollidinedione family (TZD) of antidiabetic compounds are specific PPARy ligands (Ferre, 2004; Zhang and Young, 2002). Ciglitazone, pioglitazone and rosiglitazone are presently used in the treatment of Type-2 Diabetes and lower hyperglycaemia, hyperinsulinaemia and hypertriglyceridemia (Berger and Moller, 2002; Willson et al., 2000; Evans et al., 2004; Kota et al., 2005). TZDs act by enhancing the sensitivity of tissues to insulin, especially in skeletal muscle and adipocytes, resulting in increased insulin-dependent glucose utilisation in these tissues and thereby decreasing plasma glucose levels (Willson et al., 2001; Guan et al., 2002; Camejo, 2003).

1.7.5 Complementary roles of PPARs in lipid metabolism

PPARs activate genes that regulate several aspects of lipid metabolism involved with uptake, intracellular transport and oxidation of fatty acids, lipoprotein assembly and metabolism (Guan et al., 2002; Mandard et al., 2004). PPARa is a critical regulator of lipid oxidation, i.e. mitochondrial.and peroxisomal p-oxidation and co-oxidation in the endoplasmic reticulum, and lipid transport (Guan et al., 2002). Ligand-activation of PPARa increases fatty acid uptake and activation into fatty acyl-CoA by inducing FAT/CD36 and F ATP gene expression (Duval et al., 2004; Guan et al., 2002). PPARa activation also favours the up-regulation of FABPs that serves as a “cytoplasmic gateway” for transport of LCFAs to different metabolic pathways and the nucleus (Mandard et al., 2004; Guan et al., 2002; Tan et al., 2002).

In lipoprotein metabolic pathways (see Section 1.3.3 for more details), ligand-activation of PPARa, e.g. mediated by fibrates, increases the expression and activity of LPL, thereby increasing the clearance of the TAG-rich lipoproteins such as CM and VLDL (Duval et al., 2004; Fruchart, 2001; Camejo, 2003; Castrillo and Tontonoz, 2004; Chapman, 2003) and suppresses transcription of apo C-III, a potent inhibitor of LPL (Mandard et al., 2004; Guan et al., 2002; Fazio and Linton, 2004). PPARa activation enhances transcription of the major HDL apolipoproteins, apo A-I and apo A-II, thereby increasing HDL synthesis and reverse cholesterol transport (Guan et al., 2002; Mandard et al., 2004; Fruchart, 2001; Fazio and Linton, 2004). Activation of PPARa induces transcription of PLTP, a protein modulating the conversion of pre-PHDL

34 into mature HDL by transferring phospholipids from VLDL/LDL to HDL (Duval et aL, 2004; Fazio and Linton, 2004). PPARa activation also up-regulates the expression of HDL-receptors, /.e. ABCAl and SR-Bl, in vascular cells and macrophages, thereby allowing the removal of intracellular free cholesterol leading to a reduction in atherogenesis (Fruchart, 2001; Mandard et al., 2004; Duval et al., 2004; Fruchart et al., 1999), and induces cholesterol conversion to bile acids by increasing expression of LXR (Duval et al., 2004; Mandard et al., 2004). Moreover, ligand-activation of PPARa induces a reduction in the number of the atherogenic dense-LDL particles, and increases resistance of LDL to oxidation (Guo et al., 2001; Griffin and Fielding, 2001). These changes in lipoproteins are related to alterations in their apolipoprotein content, such as decreased apo B, apo C-III and apo E levels in TAG-rich lipoproteins and increased apo AI and apo All in HDL (Fruchart et al., 1999; Guo et al., 2001; Fruchart, 2001; Fazio and Linton, 2004).

PPARp is linked to cell renewal and differentiation (Escher and Wahli, 2000), and regulation of lipid metabolism and energy homeostasis (Evans et al., 2004; Dressel et al., 2003). In skeletal muscle, PPARP up-regulates fatty acid oxidation and energy expenditure to a greater extent than PPARa (Evans et al., 2004; Dressel et al., 2003; Wang et al., 2003). PPARp also improves the blood lipid profile by increasing HDL-C and lowering LDL and TAG levels (Guan et aL, 2002). In addition, PPARp promotes reverse cholesterol transport by up-regulating the expression of the ATP-binding cassette A1 (ABCAl), thereby increasing the efflux of free cholesterol from peripheral cells (Guan et al., 2002; Chen et al., 2003).

PPARy is a “master switch” for the differentiation of adipocytes, resulting in the formation of “fat” cells and lipid storage (Evans et al., 2004; Francis et al., 2003). Activation of PPARy stimulates the storage of fatty acids in mature adipocytes by acting at several steps: (1) activation of LPL to release FA from lipoproteins, (2) activation of the cellular FA transporters, and (3) activation of intracellular FA estérification and TAG synthesis (Guan et al., 2002; Ferre, 2004; Willson et aL, 2001; Escher and Wahli, 2000). Activation of PPARy also lowers hyperglycemia and hyperinsulinemia by stimulating glucose uptake into adipocytes and muscle tissue (Ferre, 2004; Marx et al., 2004). In addition, PPARy activation further increases insulin sensitivity by down-regulating the secretion of adipocyte hormones, such as adiponectin, TNFa and resistin, that inhibit

35 insulin-stimulated glucose transport and utilisation (Guan et aL, 2002; Ferre, 2004). Thus, PPARy induces TAG synthesis and storage in adipocytes and influences glucose metabolism and insulin sensitivity in muscle and liver tissues (Fig. 1.11) (Willson et al., 2001; Ferre, 2004). In summary PPARa, PPARp and PPARy co-operate to regulate blood lipid profile and glucose homeostasis. PPARa is abundant in the liver, whereas skeletal muscle tissue is the major source of PPARp. Ligand-activation of PPARa and PPARp increases fatty acid uptake and P-oxidation in liver, muscle, and heart, whereas PPARy ligands enhance lipid storage in adipocytes (Fig. 1.11).

'Adipogenests •Lipogenesis and lipid s t o ^ • Lipogenesis and «Adipokinc production lipid storage • Major target ot TZDs ■ Fatty acid oxidation ■ Fatty acid oxidation ' Fasting response PPAR- y ■Energy uncoupling PPAR-a PPAR-Ô • Regulation of whole - body insulin sensitivity Fatty acid oxidation ■ Fatty acid oxidation Energy uncoupling

Figure 1.11: Complementary roles of PPAR isoforms in lipid metabolism. Three PPAR isoforms regulate lipid and glucose homeostasis in liver, adipose tissue and skeletal muscle. Whereas PPARa and PPARP increase fatty acid oxidation, PPARy enhances fat storage in adipocytes and tissue insulin sensitivity (Fvans et al., 2004).

1.7.6 PPARs and atherosclerosis

PPARa plays a critical role in the inhibition of atherogenesis at several stages (Chen et al., 2003; Neve et al., 2001). PPARa down-regulates the expression of VCAM-1, thereby decrease migration of monocytes into the arterial wall (Willson et aL, 2000). PPARa also suppresses transcription of inflammatory mediators, such as interleukins and tumour necrosis factor-a that are required for progression of the atherosclerotic plaque (Chen et aL, 2003; Blanquart et aL, 2003; Cabrero et aL, 2002). Furthermore, ligand-activation of PPARa modulates expression of nuclear factor k B (NFk B), a potent

36 inflammatory mediator of the vascular wall cells, and inhibits progression of the atherosclerosis (Fruchart et aL, 1999).

PPARy reduces cytokine (TNFa, IL-ip, and IL-6) production, an anti-inflammatory effect of PPARy that could be beneficial in the inhibition of atherosclerotic progression (Tham et aL, 2003; Verges, 2004). In addition, PPARy reduces the expression of metalloproteinases that are implicated in plaque destabilization, leading to plaque rupture and myocardial infarction (Desvergne and Wahli, 1999; Ricote et aL, 1999; Verges, 2004). PPARy also plays critical roles in the macrophages/foam cells (Blanquart et aL, 2003; Chen et aL, 2003; Plutzky, 2003). Monocytes express PPARy when exposed to oxidised-LDL. Subsequently, ligand activation of PPARy induces monocyte differentiation and promotes uptake of oxidised-LDL through the induction of scavenger receptors, such as FAT/CD36 (Tham et aL, 2003; Plutzky, 2003; Guan et aL, 2002; Desvergne and Wahli, 1999; Zhang and Chawla, 2004). In foam cells ox-LDL activates PPARy and LXR (Barish and Evans, 2004), leading to an increase in the expression of FAT/CD36 and further uptake of ox-LDL itself (Marx et aL, 2004; Barish and Evans, 2004; Tham et aL, 2003; Zhang and Chawla, 2004). PPARy and LXR activation also enhances transcription of ABCA-1 protein and therefore promotes efflux of intracellular cholesterol (Barish and Evans, 2004; Castrillo and Tontonoz, 2004; Puddu et aL, 2003; Tham et aL, 2003). Thus, the net function of PPARy-LXR co-operation in macrophages is to reduce LDL-C by taking-up LDL/ox-LDL, and promoting cholesterol removal by enhancing reverse cholesterol transport, leading to cholesterol disposal in the liver (Barish and Evans, 2004).

In contrast to the inhibitory effects of PPARa and PPARy on the progression of atherosclerosis (Fruchart et aL, 1999; Tham et aL, 2003; Verges, 2004), PPARp plays a role in the pathology of diseases associated with vascular smooth muscle cells (VSMC) proliferation and atherosclerosis (Zhang et aL, 2002). PPARP is upregulated in VSMC during vascular lesion formation and modulates VSMC proliferation (Puddu et aL, 2003). In addition, deletion of PPARp from foam cells increases the availability of inflammatory suppressors and reduces atherosclerotic lesion formation. Therefore, PPARp controls the inflammatory status of the macrophage, promotes macrophage lipid accumulation and likely contributes to the development of atherosclerosis (Lee et aL, 2003a; Puddu et aL, 2003).

37 1.8 Hypothesis, aims and objectives

The research project described in this thesis originates from a desire to understand the effects of different LCFAs and fibrates on gene expression levels in cardiomyocytes. Studies on the physiological role of fatty acids have revealed that they influence lipid metabolism by functioning as regulators of gene expression (Jump, 2004; Pegorier et al., 2004; Clarke, 2000). The regulatory function of LCFAs on gene expression is achieved by controlling the abundance and activity of transcription factors (Sampath and Ntambi, 2004; Khan and Vanden Heuvel, 2003; Clarke, 2001; Jump, 2002a; Guan et al., 2002; Mandard et al., 2004; Jump, 2002a). Previous studies have shown that LCFA characteristics, such as chain length and degree of unsaturation, differentially affect gene expression levels in cardiomyocytes (Biagi et al., 1999; Bordoni et al., 1996; Dyerberg et al., 2004; Williams et al., 2004; Cameron-Smith et al., 2003; Belury, 2002; van der Lee et al., 2000a; Hickson-Bick et al., 2000; Clavel et al., 2002). In this study, the P19CL6 cell-line was used as a cardiovascular model system (Habara-Ohkubo, 1996) and it was hypothesised that LCFAs differentially affect gene expression levels in this cell-line. Thus, the research focused on both the characterisation of the cell-line and investigating the effects of different LCFAs (palmitic, oleic, linoleic, and a-linolenic acids) and clofibrate on gene expression levels and the aims and objectives of this study were as follows:

Aim-1: Characterisation of the P19CL6 cell line at the transcriptional level.

Objectives: • Comparison of the expression profiles of selected cardiac- (a-MHC and P-MHC) and skeletal muscle-specific genes (MyoD and Myogenin) in P19CL6 cells with those of the H9C2(2-1) skeletal muscle cell-line, as well as mouse heart and skeletal muscle tissue.

Evaluation of optimal culture conditions where the P19CL6 cells display a cardiomyocyte phenotype.

38 Investigation of the effect of different media and micronutrients on “beating” phenotype of the PI 9CL6 cell-line.

Microscopic analysis to define the cellular morphology and phenotype of P19CL6 cells.

Comparison of global transcriptome levels in P19CL6 and H9C2(2-1) cells, as well as mouse cardiac, skeletal muscle, and embryonic heart tissue, by microarray analysis to characterise the P19CL6 cells.

Aim-2: To determine the effects of LCFAs and clofibrate on gene expression levels in P19CL6 cells.

Objectives: • Ascertain the expression, at the mRNA level, of H-FABP, PPARa, PPARp, and PPARy in P19CL6 cells cultured in response to LCFAs and clofibrate.

• Compare global transcriptome levels in P19CL6 cells cultured with different LCFAs and clofibrate, using cDNA microarrays.

39 Chapter 2

Materials and Methods

40 Chapter 2:

Materials and Methods

2.1 Materials

Dulbecco’s Modified Eagle Medium (DMEM), DMEM-F12, RPMI-1640, a-Minimum Essential Medium (a-MEM), foetal bovine serum (FBS), bovine serum albumin (BSA), Trizol reagent, phosphate-buffered saline tablets (PBS), ultra-pure 20x SSC buffer (sodium chloride-sodium citrate), ultra-pure 10% SDS (sodium dodecyl

sulphate), Oligo dT(i 2-i8 ) primer. Mouse COT-1 DNA, poly A, Denhardt’s solution, and Superscript II Reverse Transcriptase kit were purchased from Invitrogen Corporation (Paisley, UK). Sigma-Aldrich Company Ltd. (Poole, UK) supplied lOOx penicillin-streptomycin (10,000 Units/ml and 10 mg/ml, respectively), bacteriological agar, long-chain fatty acids including palmitic (16:0 saturated FA), oleic (18:ln-9 monounsaturated FA), linoleic (18:2n-6 PUFA), and a-linolenic (18:3n-3 PUFA) acids, clofibrate, fatty acid-ffee bovine serum albumin (BSA), sodium acetate (DNfl5e/RNfl5e-free), sodium pyrophosphate, phorbol myristate acetate (PMA), chloroform, isopropyl alcohol, and dimethyl sulphoxide (DMSO). SV Total-RNA Isolation, SV Gel and PCR Clean-Up, and nested PCR kits, nucleotide mixtures, Taq DNA polymerase enzyme, 100 bp DNA ladder, Blue-Orange Loading dye, formamide and agarose for gel electrophoresis were purchased from Promega Corporation (Southampton, UK) and MWG (Milton Keynes, UK) supplied oligonucleotide primers. Ready-To-Go RT-PCR beads, AutoSeq G-50 columns. Microarray CodeLink Activated Slides (Amine-Binding Slides), HybriSlip microarray coverslips (24x60 mm), Cy3-dCTP and Cy5-dCTP fluorescent dyes were obtained from Amersham Biosciences (Chalfont St. Giles, UK) and Coming Life Sciences (Coming B.V. Life Sciences, Schiphol-Rijk, Netherlands) supplied ProntoPlus microarray kit. Ambion Ltd. (Huntingdon, UK) supplied ultra-pure DEPC-treated nuclease-ffee water, mouse heart and embryonic total-RNA, and RNA stabilization solution (RNA/a/er), whereas mouse embryonic heart and skeletal muscle total-RNA were purchased from Zyagen (Zyagen, San Diego, USA) and Panomics (Panomics Inc., Redwood City, USA), respectively. RNga^y MinElute

41 Cleanup kit, RN^^e-free DNase, and MinElute PCR purification kit were obtained from Qiagen Ltd. (Crawley, UK). Mixed-bed AG 501-X8 (D) molecular biology grade resin was purchased from Bio-Rad (Hemel Hempstead, UK), and Fisher Scientific supplied chloroform-resistant Oak Ridge centrifuge tubes (Loughborough, UK). Tissue culture flasks, petri dishes and 6-well plates were provided by Nunclon Products (Nunclon, Roskilde, Denmark) and 150 cm^ peel-off flasks were obtained from Helena BioSciences Europe Ltd. (Sunderland, UK). Specific batches of foetal bovine serum (FBS) were purchased from Bio west Ltd. (Ringmer, UK).

2.2 Methods

2.2.1 Cell culture

The P19CL6 cell-line is a sub-clone of the P19 cell-line (Habara-Ohkubo, 1996), the former having been specifically selected on their ability to differentiate into cardiomyocytes when cultured in the presence of DMSO (Habara-Ohkubo, 1996; Hosoda et al., 2001; Skeijanc, 1999). The literature on the culture and differentiation of PI 9 cells into cardiomyocytes describes the utilisation of two processes, namely culturing cells under standard conditions where cells are adhered to cell culture plastic (Habara-Ohkubo, 1996), or non-adherent conditions where cells grow on agar plates to produce cell aggregates prior to re-plating under adherent conditions (van der Heyden et al., 2003; Morley and Whitfield, 1993; van der Heyden and Defize, 2003; Skeijanc, 1999; McBumey, 1993; Smith et al., 1987). In this study both culture conditions and their effects on gene expression levels were investigated, as detailed below.

2.2.1.1 Culture of P19CL6 cell under adherent conditions

P19CL6 cells were purchased from Riken Cell Bank (Ibaraki, Japan) (Habara- Ohkubo, 1996) in growing flasks and cultured in normal culture media (NCM) containing a-MEM supplemented with 10% FBS and 1% penicillin-streptomycin. To split the cells, media was aspirated and washed with 5 ml of autoclaved phosphate-buffered saline (PBS; 10 mM phosphate buffer, 2.7 mM potassium chloride, 137 mM sodium chloride, pH 7.4).

42 Cells were then incubated with 2 ml of a 0.25% trypsin-EDTA solution at 37°C until cells detached from the flask (5 minutes). Trypsin was neutralised by adding fresh media (1:20; trypsin:media) and the cell suspension divided into two 175 cm^ flasks.

Differentiation of P19CL6 cells into cardiomyocytes was induced with a DMSO-containing differentiation media (Habara-Ohkubo, 1996). Cells were plated in 37 mm diameter wells (6-well culture plates) at a density of 2x10^ cells per well in 5 ml of differentiation media (DM; a-MEM/10% FBS/1% penicillin-streptomycin and 1% DMSO). Cells were maintained in DM for 15 days and the media refreshed every second day. Finally, cells were harvested by scraping into 1 ml PBS at the end of the incubation period and stored at -80°C until RNA extraction.

To prepare frozen stocks, cells were first treated with trypsin, as described above, and then centrifuged for 5 minutes at lOOOxg (2,000 rpm; Sigma-6K10 centrifuge; Sigma Ltd., Germany). After aspirating the supernatant, the cell pellet was re-suspended in freezing media containing 91% FBS and 9% DMSO at a density of 2x10^ cells/ml. Cells were stored under liquid nitrogen (-200°C).

2.2.1.2 Culture of P19CL6 cells under non-adherent conditions

P19CL6 cells were plated in 100 mm diameter bacterial petri dishes (1x10^ cells per dish) covered with a thin layer of 0.5% agar (0.5 g agar dissolved in 100 ml nuclease-free water and autoclaved) using DMEM-F12 media for 4 days, with or without 1% DMSO, as indicated, to form cell aggregates (Skeijanc^ 1999). Cell aggregates were collected by centrifugation at lOOOxg (2,000 rpm) for 5 minutes and re-suspending in 15 ml fresh media before re-plating into 75 cm^ culture flasks for 15 days, either in the presence or absence of 1% DMSO. Cells were observed daily by phase-contrast microscopy and collected at the end of the incubation period by scraping into 1 ml PBS and stored at -80°C. 2.2.1.3 Culture of H9C2(2-1) cells under adherent conditions

The H9C2(2-1) cell line was obtained from the European Collection of Cell Cultures (ECACC; Salisbury, UK) (Kimes and Brandt, 1976) in a frozen ampoule. The ampoule was thawed on ice and the cells transferred to a 25 cm^ tissue culture flask containing 5 ml of a-MEM media/10% FBS/1% penicillin-streptomycin. The culture media was refreshed after 8 hours once the cells had adhered to the flask (Kimes and Brandt, 1976). To plate new passages the cells were divided into new flasks, as described above for P19CL6 cells. Frozen stocks were prepared by re-suspending the cells in freezing media (91% FBS and 9% DMSO) at a density of 2-4x10^ cells/ml, as described for P19CL6 cells, and stored under liquid nitrogen (-200°C).

For experimental purposes the cells were plated in 37 mm diameter wells (6-well culture plates) at a density of 2x10^ cells/well in 5 ml of differentiation media (DM; a-MEM/10% FBS/1% penicillin-streptomycin and 1 % DMSO) (Kimes and Brandt, 1976). Cells were maintained in DM and the media refreshed every second day. Finally, cells were harvested by scraping into 1 ml PBS at the end of incubation period and stored at -80°C until RNA extraction.

2.2.1.4 Culture of THP-1 ceils under adherent conditions

THP-1 cells purchased from American Type Culture Collection (ATCC; (LGC Promochem, Teddington, UK) (Tsuchiya et al., 1980; Whatling et al., 2004) have suspension growth properties and are able to differentiate into adherent macrophages upon exposure to phorbol myristate acetate (PMA). Cells were plated in 25 cm^ tissue culture flasks in the presence of 5 ml RPMI-1640 media containing 10% FBS and 1% penicillin-streptomycin. To change the cell culture media the cells were collected by centrifugation for 5 min at lOOOxg (2,000 rpm; Sigma-6K10 centrifuge) and re-suspended in fresh media. Cells were split into new passages by dividing re-suspended cells at a ratio of 1:4 (cell suspension: fresh media).

To differentiate THP-1 cells into macrophages, the cells were centrifuged and the cell pellet re-suspended in RPMI-1640 media containing 10% FBS, 1% penicillin-streptomycin, and 100 nM PMA (100 pM stock solution in DMSO). Cells

44 were then cultured with PMA in 6-well plates at a density of 5x10 cells per well for 24, 48 and 72 hours, as indicated.

To prepare frozen stocks, cells were centrifuged at lOOOxg for 5 min. and the cell pellet re-suspended in freezing media (91% FBS and 9% DMSO) to achieve a density of 2x10^ cells/ml, and stored under liquid nitrogen (-200°C).

2.2.1.5 Determination of P19CL6 cells pulse rate cultured with different serum samples

Foetal bovine serum (FBS) samples from four different batches were obtained from Biowest Ltd. (Ringmer, UK). P19CL6 cells from different passage numbers were incubated with differentiation media (DM; a-MEM/10% FBS/1% DMSO and 1% penicillin-streptomycin) containing these specific FBS samples for 15 days with media refreshed every second day. Cells were observed daily by phase-contrast microscopy and changes in cell shape, morphology, and contractile activity (pulses/min) recorded. Finally, cells were collected at the end of day 15 by scraping into 1 ml PBS and stored at-80°C.

2.2.1.6 Culturing P19CL6 cells with different concentration of adrenaline

P19CL6 cells were cultured in differentiation media (6-well plates; 2x10^ cells/well) containing 0, 5, 10, 15, and 20 pM adrenaline for 15 days and the media refreshed every second day. The cell were examined daily by phase-contrast microscopy to assess the “beating” of cardiomyocytes and changes in contractile activity (pulses/min). Cells were collected at the end of the incubation period by scraping into 1 ml PBS and stored at -80°C.

45 2.2.1 7 Determination of P19CL6 cell morphology by microscopic analysis under non-adherent culture conditions

Aggregated cells were plated in 6-well plates containing autoclaved coverslips coated with polylysine (autoclaved coverslips were immersed in an aqueous solution of 0.01% polylysine for 30 min. at 37°C; then washed 3x with PBS before use). Coverslips were collected on different time points by removing the media and washing with PBS. Cells were fixed by immersing in 3% para-formaldehyde (pH 7.2) for 2 hours and mounted in 50% glycerol/0.1% azide/PBS on microscope slides (Rudnicki et al., 1990; Wilcox, 1993) and were subsequently photographed under 40x magnification using a Leica DMIL microscope (Leica GmbH, Wetzlar, Germany) with an attached Colour Video Sony 3CCD digital Camera (Sony Corporation, Tokyo, Japan).

2.2.1.8 Determination of P19CL6 cell morphology using nuclei staining under adherent culture conditions

Cells were cultured on coverslips under adherent culture conditions. Coverslips were collected at different time points and cells were fixed by immersing in para-formaldehyde, as described above (Rudnicki et al., 1990; Wilcox, 1993). The fixative was removed by washing 3x with PBS. Coverslips were then immersed in Hoechst dye (2 pg/ml in PBS) for 5 minutes at room temperature (protected from light) to stain nuclei. Finally, coverslips were washed once with PBS and mounted on microscope slides in mounting media (50% glycerol/0.1% azide/PBS). Microscopic examination was performed at 400x magnification using a Zeiss Axiovert 135 TV fluorescence microscope (Carl Zeiss AG, Oberkochen, Germany) setting the excitation and emission filter at 365 and 420 nm, respectively. Slides were then photographed using a Contax SLR-NX camera (Kyocera Yashica Ltd., Reading, UK) and images digitised using a Nikon LS4000 slide scanner (Nikon UK Ltd., Kingston upon Thames, UK).

46 2.2.1.9 Incubation of P19CL6 cells with different long-chain fatty acids

P19CL6 cells were cultured in 150 cm^ peel-off flasks at a density of 2x10^ cells per flask using 30 ml of DMSO-containing media for 14 days to differentiate into cardiomyocytes, as indicated. At the end of day-14, media was removed and cells washed twice with 15 ml PBS and incubated in 25 ml serum-free media (98% a-MEM/1% DMSO/1% penicillin-streptomycin) for 24 hours (day 15). Cells were then treated with media either containing bovine serum albumin (BSA), as carrier (98% a-MEM/1% penicillin-streptomycin 71% DMSO/0.15 mM BSA), BSA-fatty acid (98% a-MEM/1% penicillin-streptomycin /1% DMSO/0.15 mM BSA with 0.4 mM fatty acid), or BSA-clofibrate complex (98% a-MEM/1% penicillin-streptomycin /1% DMSO/0.15 mM BSA with 0.4 mM clofibrate) for 24 hours (Habara-Ohkubo, 1996; Bolon et al., 1997). Finally, media was aspirated off and cells washed with 15 ml PBS. Cell lysate was obtained by adding 15 ml of Trizol reagent (1 ml per 10 cm^ culture area) and by passing the cells through a pipette to breakdown genomic DNA, according to the manufacturer’s instructions (Life Technologies Ltd., Paisley, UK). The cell lysates were collected in nuclease-free centrifuge tubes and stored at -80°C until RNA extraction.

Preparation of BSA/FA and BSA/Fibrate complexes:

Serum-free media was prepared by the addition of DMSO (5 ml) and penicillin-streptomycin (5 ml) into 500 ml a-MEM to a final concentration of 98% a-MEM/1% DMSO/1% penicillin-streptomycin. An appropriate amount of bovine serum albumin was dissolved in serum-free media in a beaker to obtain a concentration of 0.15 mM BSA. It was stirred gently for 2 hours, filter-sterilized and stored at 4°C. In the next step, an appropriate amount of fatty acid or clofibrate powder was dissolved in 100 mM NaOH to prepare a 100 mM stock solution by stirring at 37°C for 30 minutes (Bolon et al., 1997). Finally, fatty acid (or clofibrate) stock solutions (100 mM) were diluted at the ratio of 1:250 with serum-free media containing 0.15 mM BSA to achieve a final concentration of 0.4 mM FA (or clofibrate). Media was then filter-sterilized and stored at 4°C.

47 2.2.2 Total-RNA isolation and purification

Two different RNA isolation methods were used to extract total-RNA from P19CL6, H9C2(2-1) and THP-1 cultured cells, mouse heart and skeletal muscle tissues. In the first method, the RNA was extracted and purified using a SV Total RNA Isolation kit (mainly used in PCR experiments) as illustrated in Fig. 2.1, whereas in the second method, the required RNA for microarray experiments was isolated by Trizol reagent and further purified, as recommended by the Human Genome Mapping Project (HGMP) protocol.

Cultured cells a n d Tissue samples

Total RNA SV Total RNA Trizol method isolation ■ - Isolation kit

DNase-1 digestion and PCR experiments column purification

Microarray experiments

Figure 2.1: Overview of total-RNA isolation methods. Two different RNA isolation methods (SV Total RNA Isolation and Trizol) were used for extraction of total-RNA from cultured cells or tissue samples. The RNA obtained using the SV Total RNA Isolation kit was used directly in PCR experiments, while the RNA obtained using the Trizol method was further purified before being labelled with Cy-dye and used in microarray experiments.

48 2.2.2.1 Isolation of total-RNA from cell culture samples

(/) SV Total RNA Isolation method

Frozen cells were thawed on ice and cell lysate prepared by adding lysis buffer [4M guanidine isothiocyanate, 0.01 M Tris (pH 7.5), and 0.97% p-Mercaptoethanol] according to the manufacturer’s instructions (Promega Corporation, Southampton, UK). The lysate mixture was passed through a 20-gauge needle to shear genomic DNA. The RNA was precipitated by adding 95% ethanol. Subsequently, the RNA pellet was washed by addition of RNA wash solution [60 mM potassium acetate, 10 mM Tris-HCl (pH 7.5), and 60% ethanol] and by utilising purification columns provided in the SV Total RNA Isolation kit. To erase DNA contaminant, DNA fragments were digested upon exposure to DNfl.ye-1 enzyme for 15 minutes at room temperature and the RNA was purified by centrifugation for 2 min at 14,000xg (12,300 rpm, Eppendorf microcentrifuge, Hamburg, Germany). Total-RNA was then cleaned-up by adding RNA wash solution and centrifugation for 2 minutes at 14,000xg (12,300 rpm). Finally, the RNA template was eluted into 100 pi nuclease-free water and stored at -80°C.

(//) Trizol method

In the Trizol method, cell lysates were obtained by the direct addition of Trizol reagent to the cells (1 ml per 10 cm^ of cell culture area) and passing the cells trough a pipette to breakdown genomic DNA. Then, 0.2 ml of chloroform per 1 ml of Trizol reagent was added and samples were incubated for 3 minutes at room temperature. To separate the aqueous phase containing total-RNA from phenol-chloroform phase, samples were centrifuged at 12,000xg for 15 minutes at 4°C (11,000 rpm using a J2-21 Beckman centrifuge and JA-17 Fixed-Angle rotor, USA) using chloroform-resistant nuclease-free centrifuge tubes.

The aqueous phase was transferred into an autoclaved chloroform-resistant centrifuge tube and 0.5 ml of isopropyl alcohol was added per 1 ml of Trizol. Samples were incubated for 10 minutes at room temperature and RNA precipitated by centrifugation at 12,000xg for 10 minutes at 4°C (11,000 rpm). The supernatant was removed and the RNA pellet washed with >1:1 ratio of 75% ethanokTrizol reagent (v/v).

49 Samples were re-centrifliged at 7,500xg (7,000 rpm) for 5 minutes at 4°C and the RNA pellet air-dried and dissolved in nuclease-free water and stored at -80°C.

2,22,2 Isolation of total-RNA from mouse tissue samples

(f) SV Total RNA Isolation method

Whole heart and skeletal muscle tissues (upper leg muscle) were dissected from 10 week old male GDI wild-type mice (+/+) (a kind gift from Miss Lisa Godfrey, School of Biomedical and Molecular Sciences, University of Surrey) and placed in sterile microcentrifuge tubes (4°C). Tissue samples were homogenised using an Ultra-Turrax T8 homogeniser (IKA Labortechnik, Staufen, Germany) until no visible tissue fragments remained, and the cell lysate was prepared by the addition of lysis buffer at the ratio of 1 ml of lysis buffer per 171 mg tissue mass, according to the manufacturer’s instructions (Promega Corporation, Southampton, UK). The RNA was precipitated, purified and eluted into nuclease-free water, as indicated previously for cell culture samples.

(ii) Trizol method

In the Trizol method, whole heart and skeletal muscle tissues (upper leg muscle) were immersed in sterile tubes containing 2.5 ml of KNAlater, an RNA stabilization solution, as indicated in the manufacturer’s instruction (Life Technologies Ltd., Paisley, UK). Samples were incubated overnight in KNAlater at 4°C and the reagent was removed before storage at -80°C.

To prepare tissue lysate, tissue samples were homogenised in Trizol reagent (1 ml per 50-100 mg tissue mass) using an Ultra-Turrax T8 homogeniser until no visible tissue fragments remained. Then 0.2 ml of chloroform per 1 ml of Trizol reagent was added and samples incubated for 3 minutes at room temperature. Total-RNA was isolated as described for cell culture samples. RNA pellet was precipitated at 12,000xg (11,400 rpm) for 10 minutes at 4°C using an Eppendorf microcentrifuge. The RNA pellet was washed with >1:1 ratio of 75% ethanokTrizol reagent (v/v), dissolved in nuclease-free water, and stored at -80°C.

50 2.2,23 DNase digestion of total-RNA samples

For microarray experiments the total-RNA samples were treated with DNa^e-I enzyme to remove potential genomic DNA contamination. Samples (< 87.5 pi total-RNA solution) were incubated with 2.5 units of RNfl^e-free DNa^e-I enzyme for 10 minutes at room temperature. Nucleotides and small oligonucleotides were removed by RNA purification process using RNeasy MinElute Spin columns, as described below.

2.2.2.4 Purification of total-RNA from nucleotides

To remove nucleotides and small oligonucleotides generated during DNa^e digestion, the RNA samples were treated with 350 pi of RLT buffer provided with the RNefl^y MinElute kit, and 250 pi of 96% ethanol. The RNA sample was then applied to an RNea^y MinElute Spin Column and centrifuged for 15 seconds at 14,000xg^ (12,300 rpm) using an Eppendorf microcentrifuge. The RNA, bound to the membrane, was washed with 500 pi of RPE buffer provided with the RNea^y MinElute kit, and centrifuged (14,000xg, 15 seconds). The washing was repeated by addition of 500 pi of 80% ethanol to the membrane and centrifugation for 2 min at 14,000xg (12,300 rpm) to dry the membrane. The RNA was eluted with 14 pi RNa^e-fi-ee water and collected by centrifugation for 1 min at 14,000xg (12,300 rpm) and stored (-80°C).

2.2.2.S Determining the concentration and purity of the RNA

A NanoDrop ND-1000 spectrophotometer (NanoDrop Technologies, Montchanin, USA) was used to determine total-RNA concentration by measuring the absorbance of sample at 220-350 nm. The spectrophotometer was zeroed using DEPC-treated nuclease-ffee water and the RNA sample (1 pi) was scanned at 220 nm to 350 nm, as shown in Figure 2.2. The concentration of RNA was determined based on the Beer-Lambert equation (c=A.s.b”^) where b is path length in cm and s is the manipulated wavelength-dependent extinction coefficient in ng.cm.pU^ to give the concentration (c) in ng/pl. The generally accepted extinction coefficients for nucleic acids are 40 for RNA and 50 for double-strand DNA samples. The ratio of absorbance at 260 and 280 nm

51 (A260/A280) was used to assess purity of RNA samples with a ratio of 1.8-2.0 generally being accepted as pure RNA (Yu et al., 2002; Brown, 2001). The purity of RNA samples was further confirmed by the ratio of absorbance at 260 and 230 nm (A260/A230) with a ratio of 1.8-2.2 being accepted as high quality, pure RNA (Fig. 2.2).

File Edit Show Context Help

Re-blank Print Screen Start Report I Measurement complete 2/4/2005 12:51 PM Exit Measure Blank Print Report Show Report User Iraj

85.00 Sample RNA-4Ü 80.00- Type ------

Total RNA 60.00- Sample ID PI 9CL6

u 50.00- Batch Index] 0

% 40.00H X ^ 230 Abs. I 38.754

6 30.00 A-260 10 mm path | 82.126

20.0 0 - A-280 10 mm path j 43.650

10.0 0 - 260/280) 1.88

260/230) 2.12

•9.06- 220 230 240 250 260 270 280 290 300 310 320 330 340 35050 ri9/uL I 3285.0 3.0.0 B839 Wavelength nm

Figure 2.2: Example of the output from NanoDrop ND-1000 spectrophotometer for determining total RNA concentration. The absorbance of RNA (1 pi) was measured between 220 nm and 350 nm. The absorbance at 260 nm was converted to a concentration (ng/pl) using an Fxtinction Coefficient (F) of 40 for RNA samples. The A260/A280 ratio gives an indication of RNA purity, with a ratio of 1.8-2.0 generally accepted as pure RNA (Yu et al., 2002).

2.2.3 Gel electrophoresis

2.2.3.1 Formaldehyde-agarose RNA gel electrophoresis

Formaldehyde-agarose RNA gel electrophoresis was used to eheek total-RNA purity. The 1% agarose gel was prepared by melting an appropriate amount of agarose in sterile water, and cooling to 60°C. Then, gel running buffer (100 mM MOPS, 40 mM sodium acetate, and 5 mM EDTA, pH 8.0) and 37% formaldehyde were added to a final coneentration of 20 mM MOPS, 8 mM sodium acetate, 1 mM EDTA, and 6.6% formaldehyde. The gel was cast under a fume hood and allowed to set for at least 30 minutes at room temperature.

52 Samples were prepared by mixing 4.5 pi total-RNA sample (up to 30 pg RNA) with 15.5 pi gel-denaturing buffer (12.8 mM MOPS, 5.1 mM sodium acetate, 0.64 mM EDTA, 8.3% formaldehyde, 64% formamide and ethidium bromide) in a sterile microcentrifuge tube and incubated at 65®C for 15 minutes, chilled on ice, and were then centrifuged for 5 seconds. Finally, 2 pi of gel-loading buffer (50% aqueous glycerol containing 0.5 mM EDTA and 0.25% bromophenol blue/xylene cyanol FF, pH 8.0) was added to the samples before loading onto the gel.

A pre-run electrophoresis step was applied for 5 minutes at 5 V/cm (measured as the distance between the electrodes) before the samples were loaded. The gel was submerged in gel running buffer (20 mM MOPS, 8 mM sodium acetate, and 1 mM EDTA, pH 8.0) and electrophoresed at 3-4 V/cm. Buffer from each reservoir was collected (after 2 hours), mixed and returned to the gel apparatus to prevent polarization of the buffer. Electrophoresis was stopped when the bromophenol blue had migrated three quarters of the length of the gel and the gel was photographed using an ImageMaster YDS UV-light camera (Amersham Pharmacia Biotech, San Francisco, USA) with LISCAP-Vl Capture Application Software (Amersham Pharmacia Biotech, San Francisco, USA).

2,23,2 DNA agarose gel electrophoresis

Agarose gels (2%) containing ethidium bromide (lxlO”^%w/v) were used to identify PGR products. The gels were prepared by dissolving and heating 2 g of analytical grade agarose per 100 ml TBE buffer (90 mM Tris, 90 mM Boric acid, and 4 mM EDTA, pH 8.0). Samples were prepared by adding 2 pi of Blue-Orange Loading dye [0.03% bromophenol blue, 0.03% xylene cyanol FF, 0.4% orange G, 10 mM Tris-HCl (pH 7.5), and 50 mM EDTA] to 10 pi of PGR product and loaded onto the gel. A 100 bp DNA ladder was loaded on the outer lanes of the gel (5 pi DNA plus 1 pi loading dye). The PGR products were separated by electrophoresis at 100 V and the gels photographed using an ImageMaster YDS UY-light camera with LISGAP-Yl Capture Application Software.

b3 2.2.4 Determination of mRNA levels

To determine gene expression levels, total-RNA was isolated either from cultured cells or mouse tissues and the levels of specific mRNAs determined by reverse transcriptase polymerase chain reaction (RT-PCR) using relevant gene-specific forward and reverse primers. In some experiments, RT-PCR products were re-amplified by nested PCR to achieve higher cDNA levels, as indicated (Newton and Graham, 1997; McPherson et al., 1995). PCR products were separated by electrophoresis on 2% agarose gels containing ethidium bromide (lxlO“^% w/v). Electrophoresis was performed at 100 V and gels photographed with an ImageMaster YDS UY-light camera and LISCAP-Yl Capture Application Software.

2.2.4.1 Designing RT-PCR and nested PCR primers

Primers for both RT-PCR and nested PCR techniques were designed using the web-based primer designing software, Primer-3 (http://frodo.wi.mit.edu/cgi- bin/primer3/primer3_www.cgi), and gene cDNA sequences (NCBI GenBank; http://www.ncbi.nlm.nih.gov/). The parameters were set at a product length of 300-900 base pairs (bp) for RT-PCR and 200-600 bp for nested PCR, a primer length of 20-25 bp, and %GC content at 45-60%. The theoretical product sizes were based on the cDNA length between the forward and reverse primers, as shown on Fig. 2.3 for MyoD gene. The oligonucleotide primers were obtained as lyophilised samples (MWG, Milton Keynes, UK) and re-suspended in nuclease-free water, aeeording to the manufacturer’s data sheet, to a concentration of 100 pmol/pl and stored at -20°C. Primers used in this study are shown in Table 2.1 (RT-PCR primers) and Table 2.2 (nested PCR primers).

54 GCT6ATCGCCGCAAGGCCGCCACCATGCGCGAGCGCCGCCGCCTGAGCAAAGTGAATGAG 540 AGTGAATGAG Forward primer GCCTTCGAGACGCTCAAGCGCTGCACGTCCAGCAACCCGAACCAGCGGCTACCCAAGGTG 600 GCCTTCGAGAC GAGATCCTGCGCAACGCCATCCGCTACATCGAAGGTCTGCAGGCTCTGCTGCGCGACCAG 660

GACGCCGCGCCCCCTGGCGCCGCTGCCTTCTACGCACCTGGACCGCTGCCCCCAGGCCGT 720

GGCAGCGAGCACTACAGTGGCGACTCAGACGCGTCCAGCCCGCGCTCCAACTGCTCTGAT 780

GGCATGATGGATTACAGCGGCCCCCCAAGCGGCCCCCGGCGGCAGAATGGCTACGACACC 840

GCCTACTACAGTGAGGCGGTGCGCGAGTCCAGGCCAGGGAAGAGTGCGGCTGTGTCGAGC 900

CTCGACTGCCTGTCCAGCATAGTGGAGCGCATCTCCACAGACAGCCCCGCTGCGCCTGCG 960

CTGCTTTTGGCAGATGCACCACCAGAGTCGCCTCCGGGTCCGCCAGAGGGGGCATCCCTA 1020

AGCGACACAGAACAGGGAACCCAGACCCCGTCTCCCGACGCCGCCCCTCAGTGTCCTGCA 1080 ACACAGAACAGGGAACCCAG Reverse primer

Figure 2.3: An illustration of the primer-encoding region of the Mouse musculus MyoD gene. The published cDNA sequences for Mouse musculus MyoD in the NCBI GenBank database was imported into the Primer-3 programme and forward and reverse RT-PCR primers designed. In this example, the PCR product is 513 bp (shown bold), starting and finishing at nucleotides 531 and 1044, respectively, in the shown sequences. The numbers indicated on the right-hand side relate to the nucleotide number relative to the start of the gene sequences.

00 Table 2.1: Sequences of oligonucleotide primers used for RT-PCR

Product Oiigo name Sequences %GC Tm°C size

GAPDH-F159^ ATT CAA CGG CAC AGT CAA GGC (21 bp) 52 59.8 556 bp GAPDH-R760 CAC ATT GGG GGT AGG AAC AC (20 bp) 55 59.4

a-MHC-F5464 ATG GAG GAG ACC ATC AAG GAC (21 bp) 52 59.8 491 bp a-MHC-R5955 TTG TGT ATT GGC CAC AGC GAG (21 bp) 52 59.8

P-MHC-F1457 CTT CAA CCA CCA CAT GTT CGT G (22 bp) 50 60.3 587 bp P-MHC-R2043 TCA CCC CTG GAG ACT TTG TCT (21 bp) 52 59.8 MyoD-F372 AGT GAA TGA GGC CTT CGA GAC (21 bp) 52 59.8 513 bp MyoD-R885 CTG GGT TCC CTG TTC TGT GT (20 bp) 55 59.4 MyoG-F53 AGC TGT ATG AGA CAT CCC CCT (21 bp) 52 59.8 608 bp MyoG-R661 ACG ATG GAC GTA AGG GAG TG (20 bp) 55 59.4 a-MHC-F5628* GAC GAG GAA TAA CCT CTC CAG CAG (24 bp) 54 64.4 302 bp a-MHC-R5931* GAG GAA GAG TGA GCG GCG CAT C (22 bp) 64 65.8

P-MHC-F5806* CCC TCA GAC CTG GAG CCT TTG C (22 bp) 64 65.8 205 bp P-MHC-R6007* GCC AAC ACC AAC CTG TCC AAG TTC (24 bp) 54 64.4 ANF-F206* TGA TAG ATG AAG GCA GGA AGC CGC (24 bp) 54 64.4 203 bp ANF-R429* AGG ATT GGA GCC CAG AGT GGA CTA (24 bp) 54 64.4 MLC2-F381* CCA CAG CCT CAG GTG ACC GAC (22 bp) 67 65.7 189 bp MLC2-R569* GAA GGC TGA CTA TGT CCG GGA GAT (24 bp) 54 64.4 CaCh-hF3771* GAA AAG GCA GTT CCC ATG CCG GAA (24 bp) 54 64.4 183 bp CaCh-hR3953* CGT CCA GCA CAC CTC CTT CAG G (22 bp) 64 65.8 CaCh-mF1232* GGG CGA GGT CAT GGA CGT GG (20 bp) 70 65.5 204 bp CaCh-mR1436* GGT CTT CTA TTG GCT GGT GAT CCT G (25 bp) 52 64.6 HFABP-F55 GAC TAC ATG AAG TCA CTC GGT GT (23 bp) 48 60.6 326 bp HFABP-R381 CCG AGT GCT CAC CAC ACT G (19 bp) 63 61.0

PPARa-F367 AAG GGC TTC TTT CGG CGA AC (20 bp) 55 59.4 880 bp PPARa-R1247 TGG TTG CTC TGC AGG TGG A (19 bp) 58 58.8 PPARP-F784 TTC GCC AAG AAC ATC CCC A (19 bp) 53 56.7 360 bp PPARP-R1143 TTC TAG AGC CCG CAG AAT GGT (21 bp) 52 59.8

PPARy-F658 AAG CTG TTG GCG GAG ATC TC (20 bp) 55 59.4 856 bp PPARY-R1514 TAC AAG TCC TTG TAG ATC TCC TG (23 bp) 44 58.9 tNumbers indicate the start site of the forward (F) and reverse (R) primers. *Primers have been designed based on previously used sequences in the literature (Fassler et al., 1996). Tn,: annealing temperature (°C) bp: .

56 Table 2.2: Sequences of oligonucleotide primers used for nested PCR

P ro d u ct Ollgo name Sequences %GC Tm°C size

GAPDH-F188^ GAC TCC ACT CAC GGC AAA TTC (21 bp) 52 59.8 413 bp GAPDH-R601 AGT CTT CTG GGT GGC AGT GAT (21 bp) 52 59.8

a-MHC-F5632 AAG AGT GAG CGG CGC ATC AA (21 bp) 55 59.4 302 bp a-MHC-R5931 CTG CTG GAG AGG TTA TTC CTC (21 bp) 52 59.8

P-MHC-F47 GAG GGC ATT GAG TGG ACA TTC (21 bp) 52 59.8 412 bp P-MHC-R458 CTT TCT TTG CCT TGC CTT TGC C (21 bp) 50 60.3

MyoD"F590 TAC CCA AGG TGG AGA TCC TG (21 bp) 55 59.4 392 bp MyoD-R982 GTG GTG CAT CTG CCA AAA GC (21 bp) 55 59.4

MyoG-F32 ACC AGG AGC CCC ACT TCT AT (20 bp) 55 59.4 337 bp MyoG-R369 GCG CAG GAT CTC CAC TTT AG (20 bp) 55 59.4 HFABP-F97 TCG GTA GCA TGA CCA AGC CTA (21 bp) 52 59.8 242 bp HFABP-R339 TTT CCC GTC AAC TAG CTC CCT (21 bp) 52 59.8

PPARa-F644 TGA AGA ACT TCA ACA TGA ACA AGG TCA A (28 bp) 36 60.7 350 bp PPARa-R993 CAG CAT CCC GTC TTT GTT CAT CA (23 bp) 48 60.6

PPARP-F847 TAT GGC GTG CAC GAG GCC (18 bp) 67 60.5 350 bp PPARP-R1094 TTC ATG AGG CCT GGC CGG T (19 bp) 63 61.0

PPARy-F751 ATA AAG TCC TTC CCG CTG ACC AA (23 bp) 48 60.6 525 bp PPARy-R1275 GCG GTC TCC ACT GAG AAT AAT GA (23 bp) 48 60.6 tNumbers indicate the start site of the forward (F) and reverse (R) primers. Tm: annealing temperature (°C) bp: base pair.

2.2 A.2 Reverse Transcription Polymerase Chain Reaction (RT-PCR)

Gene expression (mRNA) levels were determined by the reverse transcription polymerase chain reaction (RT-PCR) (Newton and Graham, 1997). The “Ready-To-Go” RT-PCR kit (Amersham Biosciences, Chalfont St. Giles, UK) was used that contains all the required constituents for RT-PCR (buffer, nucleotides, enzymes) in a solid bead, with the exception of RNA, primers and nuclease-ffee water. Beads were dissolved in 45 pi of

57 nuclease-ffee water and then 1 |il of the relevant gene-specific forward and reverse primers (100 pmol/pl) were added to the reaction mixture (4°C). Finally, 3 pi of the RNA template was added and the PCR samples transferred to a Thermo Hybaid thermal cycler (Thermo BioSciences GmbH, Ulm, Germany). The control reaction, supplied with the Ready-To-Go kit, was used as a positive control. The one-step protocol for RT-PCR was performed according to the scheme summarised in Table 2.3 as recommended in the manufacturer’s instructions. The annealing temperature was set at 5°C less than the lower Tm calculated by the supplier (MWG, Milton Keynes, UK) for a primer pair, and the extension temperature was adjusted to an optimal temperature for Taq DNA polymerase (Newton and Graham, 1997; McPherson et al., 1995).

Table 2.3: RT-PCR cycling profile.

42°C for 30 min Reverse Transcription First Strand DNA Synthesis 1 cycle 95°C for 5 min RNA-DNA Primer Dénaturation

95°C for 1 min Dénaturation Second Strand DNA Synthesis *T°C for 1 min Annealing 30 cycles and PCR Amplification

72°C for 2 min Extension

Final Extension 72°C for 10 min Final Extension 1 cycle

Holding Temperature 25°C - *T°C is usually 5°C less than the lower annealing temperature (Tm) calculated by the supplier for a primer pair

2.2.4.3 Nested PCR

In some samples, no PCR products were visible after an initial RT-PCR (Section 2.2.4.2). In these cases it was assumed that the initial mRNA levels, and therefore also the amplified PCR product levels, were very low and therefore the PCR products were ftirther amplified in a second round of PCR, using “nested”

58 primers (Newton and Graham, 1997; McPherson et al., 1995). Nested PCR primers were designed to bind to specific gene sequences in the hypothetical RT-PCR product (Fig. 2.4). Reagents were added to autoclaved 0.5 ml PCR tubes, as shown in Table 2.4 and the nested PCR protocol was performed according to the scheme summarised in Table 2.5.

m R N A

RT-PCR product

Nested PCR product

Figure 2.4: Diagrammatic representation of RT-PCR product amplification by nested PCR. A first set of gene-specific PCR primers (FI and Rl) was used in the RT-PCR reaction. In the second amplification, i.e. nested PCR, a separate set of primers (F2 and R2) was used to re-amplify the RT-PCR product. F: forward; R: reverse.

Table 2.4: Components in nested PCR tubes.

R e a g e n t Volume Final concentration

DEPC-treated water 35.5 1^1

MgCb (25 mM, 25 nmol/^l) 4 III 2 pM

DNA polymerase buffer (lOx) 5 pi Ix

dNTPs (10 mM each dNTP) 2 pi 0.4 mM each dNTP

Primers [(forward and reverse); 100 pmol/pl] 0.5 pi 1 pmol/pl

Taq DNA Polymerase (5 Unit/pl) 0.5 pi 0.05 u/pl

DNA template (RT-PCR product) 2 pi

Total volume 50 pi -

59 Table 2.5: Nested PCR cycling profile.

Dénaturation , 95°C for 1 min Dénaturation of double-strand DNA 1 cycle

95°C for 45 sec Dénaturation DNA Strands Synthesis 30 cycles and PCR Amplification *T°C for 45 sec Annealing

72°C for 2 min Elongation

Final Extension 72°Cfor 10 min Final Extension 1 cycle

Holding Temperature 25°C -

*T°C is usually5°C less than the lower annealing temperature (T^) for a primer pair

2.2.5 Quantitation of PCR products from digitised gel images

In order to get semi-quantitative values of gene expression (mRNA) levels from the RT-PCR and nested PCR experiments the photographic image of gels were digitised using the Scion Image programme (http://www.scioncorp.com/). In order to obtain digital values {i.e. gene expression levels relative to the expression of housekeeping gene) a saved image in “TIFF” format was used. First, a lane was marked, as shown for GAPDH (Fig. 2.5A). Next, the illumination of the lane was plotted for the marked area (Fig. 2.5B). The peak area was delineated by drawing tangents to remove the background ; brightness (Fig. 2.5B, tangent 1). Finally, the “effective pixels” within the peak area were taken to represent gene expression level (Fig. 2.5B, tangent 2).

60 GAPDH

0

Figure 2.5: An example of the process used to digitise gel images and semi-quantitate PCR products using Scion Image software.PCR products were separated on a 2% agarose gel containing ethidium bromide and the gel photographed. A saved image in “ lib” format was used for image digitisation. First, a lane was marked as shown for GAPDH (A). The illumination of the marked area was then plotted (B) and peak area was delineated by drawing tangents to remove the background brightness (tangent 1). Finally, the “effective pixels” within the peak area were taken to represent gene expression level (tangent 2).

2.2.6 Purification of PCR products from agarose gels

PCR products were separated by agarose gel electrophoresis and visualised, as described (Section 2.2.3.2). The PCR products were then excised from the gel and the gel slices weighed. The DNA was extracted from the agarose using a Wizard SV Gel and PCR Clean-Up kit (Promega Corporation, Southampton, UK). The gel slices were dissolved in a membrane-binding solution [4.5 M guanidine isothioeyanate, 0.5 M potassium acetate (pH 5.0)] at 65°C and the solution transferred to purification columns containing a DNA-binding membrane. The membranes, were washed with a membrane-washing solution [10 mM potassium acetate (pH 5.0), 80% ethanol, and 16.7 pM EDTA (pH 8.0)] and the DNA eluted with nuclease-ffee water. The concentration of PCR products was determined by absorbanee spectroscopy using a NanoDrop ND-1000 spectrophotometer, as discussed earlier (Section 2.2.2.5), using an extinction coefficient (s) of 50 ng.cm.pU’ for double-strand DNA samples. DNA purity was assessed by using the absorbance ratio at A260/A280, with a ratio of 1.8-2 regarded as aceeptable (Yu et ah, 2002; Brown, 2001).

61 2.2.7 DNA sequencing of PCR products and sequence comparison

PCR products were sequenced in the Functional Genomic and Microarray Centre (School of Biomedical and Molecular Sciences, University of Surrey) using a CEQ 2000XL DNA Analysis System (Beckman Coulter, Fullerton, USA), and relevant gene-specific primers (Fig. 2.6). In order to confirm identity, the DNA sequences of the PCR products were compared to the gene sequenees, as obtained from NCBI GeneBank Database (http://www.nebi.nlm.nih.gov/), taking into aecount the positions of the PCR primers utilised, using the ClustalW sequence analysis programme (http://www.ebi.ac.uk/elustalw/). Identity was further eonfirmed by comparison of the PCR product sequences to the Mouse musculus Genome database, using the Blast algorithm (http://www.nebi.nlm.nih.gov/genome/seq/MmBlast.html/).

62 s 3 CL

I - a I I I

aM) 'S g ÛÛ I 2 CL (Dû I .£0 1 ■ § 00S' I & B Î (0 (0 i Û Û II i ? sw O (0 > > (Q (Ü C c nCm < — ! <

IX) “IQ u I X Q o o e o a

, 1 1 § I l 11 ïtï 1a, 1 (U

9 10o 10 9 10 o » o 9 ■0 o 10 o 10 9 10 o

{% •3U*3S«JOn|J 93Ua380JOn|J

c l 2.2.8 Microarray experiments

P19C16 and H9C2(2-1) cells were cultured in 150 cm^ peel-off flasks at a density of 2x10^ cells per flask in the presence of DMSO for 15 days, a time point at which spontaneous contraction, i.e. a cardiac-type phenotype, of the P19C16 cells has been observed (Habara-Ohkubo, 1996). Total-RNA was isolated from the cultured cells, and mouse cardiac and skeletal muscle tissue, by the Trizol method. During the microarray experiments cDNA was synthesised, labelled and hybridised on arrays using two different methods. In the first method, HGMP protocols (Human Genome Mapping Project; http://www.hgmp.mrc.ac.uk/) were followed since microarray slides had been provided by the HGMP group, whereas in the second method labelling and hybridisation procedures were carried out using a ProntoPlus kit (Coming B.V. Life Sciences, Schiphol-Rijk, Netherlands). The ProntoPlus method requires less total-RNA template (5 pg) and fluorescent Cy-dye (1 pi) in comparison with the HGMP method (40 pg total-RNA and 3 pi dye), and therefore is more economical.

In order to obtain estimates of biologically important parameters, microarray experiments were designed as dual-hybridisation optimal interwoven loops, as shown in Fig. 2.7, using an online microarray experimental design programme available on the MARIE:ExGen Consortium website (http://exgen.ma.umist.ac.uk/) (Simon et al., 2002; Dobbin and Simon, 2002; Churchill, 2002; Vinciotti et al., 2005; Wit and McClure, 2004). The loop experiments were designed based on the number of conditions (total-RNA samples) and the number of comparisons (slides). Microarray experiments were carried out in a step-by-step procedure, as summarised in Fig. 2.8, including cDNA synthesis and labelling by the direct labelling method (Yu et al., 2002), hybridisation of labelled-cDNA on arrays, and image acquisition by scanning of arrays using an Affymetrix428 scanner (Affymetrix, Santa Clara, USA). BlueFuse (BlueGenome Ltd., Great Shelford, UK) software was used for image analysis and the data further analysed using Gene Spring-7 (Silicon Genetics, Redwood City, USA) and LLAMA (Live Linear Analysis of Micro Array, http://exgen.ma.umist.ac.uk/) software.

The variance between two comparisons in a loop design experiment is dependent on the number of paths connecting them (Wit and McClure, 2004; Dobbin and Simon, 2002). For example, the comparison of samples 1 and 2 shown in Fig. 2.7A, can be estimated via direct (1 to 2) or indirect (1 to 3 and 3 to 2) paths (Wit and McClure, 2004). Therefore,

64 the linear modelling approach (LLAMA) was performed to estimate comparisons between the samples for each gene via different paths, and to compute path-independent values with greater precision (Wit and McClure, 2004; Vinciotti et ah, 2005).

B

C D

Figure 2.7: Representation of dual-hybridisation optimal interwoven loop designs for (A) three, (B) four, (C) six, and (D) seven conditional microarray experiments. Experiments were efficiently designed in dual-hybridisation optimal interwoven loops using an online microarray experimental design software available on the MARIEiExGen Consortium website (http://exgen.ma.umist.ac.uk/) (Simon et al., 2002; Dobbin and Simon, 2002; Churchill, 2002; Vinciotti et al., 2005; Wit and McClure, 2004). The Cy3- and Cy5-labelled probes were mixed (40 pmol of each) in pairs, according to the experimental design, and hybridised on 60x24 mm microarray slides. Arrows represent Cy3 to Cy5 direction and show comparisons between the samples throughout the loop. Numbers 1-7 indicate total-RNA source depending on experiment.

65 Cell culture Mouse heart and samples skeletal muscle tissues

Total RNA isolation (Trizol method) Determining RNA concentration

DNase digestion and RNA purification

Determining RNA concentration cDNA direct labelling (Cy3 or Cy5 dye)

RNA denaturing and removing unincorporated dye Determining Cy3/Cy5 dye incorporation degree Hybridisation of labelled cDNA on array

Microarray post-hybridisation wash-up

Scanning and data capture

Image and Data analysing

Figure 2.8: A schematic representation of the steps required for microarray analysis. Details of each step are provided below.

66 2.2.8.1 Microarray procedure (HGMP protocol)

In order to perform microarray experiments based on the HGMP method, all required reagents were freshly prepared in an RNase-free environment and experiments were performed, as discussed below.

2.2.8.1.1 cDNA synthesis and target labelling

Total-RNA was initially isolated using the Trizol reagent according to the one-step procedure, as described in section 2.2.2. The RNA was further purified with the RNeasy MinElute kit to remove potential genomic DNA contamination. Finally, purified RNA (40 pg total-RNA) was used for cDNA synthesis and labelling with Cy3 or Cy5 fluorescent dyes according to the “direct incorporation” procedure (Yu et al., 2002).

The synthesis of cDNA and target labelling was achieved as follows: the total-RNA sample was placed in a microcentrifuge tube and volume adjusted to 12.9 pi with DEPC-treated nuclease-free water. Then, Oligo dT^MN (19 pM) was added (total reaction volume 15.4 pi), the solution mixed thoroughly and incubated on a heating block at 70°C for 10 min. At the end of the incubation, the tube was chilled immediately on ice for 10 min. To initiate cDNA synthesis, the first-strand synthesis buffer [50 mM Tris-HCl (pH 8.3), 75 mM KCl, 3 mM M gCy was added followed by addition of DTT (10 mM), nucleotide mixture (0.5 mM dATP, dOTP, dTTP, and 0.2 mM dCTP), Cy3/Cy5-dCTP dye (0.1 mM), and SuperScript-11 Reverse Transcriptase (14 units/pi), as shown in Fig. 2.9. Reagents were mixed (total volume 30 pi), and incubated at 42°C for 2 hours on a heating block, protected from light. Finally, residual RNA was degraded and unincorporated dye was removed before the labelled cDNA was hybridised to the microarrays, as discussed below.

67 V%/V%/VVVVVV^AAAAAAAAA mRNA

O lig o dT^^ 2-i8 )

V\ATU^A/\A/VV^AAAAAAAAA dATP, dGTP, dTTP TTTTTTTTTT Cy3/Cy5-dCTP Reverse transcriptase

VVXA>^yvVVVV^AAAAAAAAA XrvA/VVVVVVXr^AAAAAAAAA V\AA/VV\AA/V^TTTTTTTTTT V V A /\yy\A A /\/y^ TTTTTTTTTT -Ô:- -O'- -Ô:- -;ô> -:o:- -;o;- Cy3-labelled Cy5-labelled

Degrade mRNA

VXAA/X jTLTVVVV^TTTTTTTTT -:q > -:q > ■o:- -:o;- -:o? CyS-labelled cDNA CyS-labelled cDNA

Figure 2.9: Schematic representation of the direct target-labelling procedure.High quality RNA samples were used for synthesis and labelling of the cDNA in the presence of Oligo dT(i 2-i8 ) primer, buffer, nucleotide mixture, Cy3/Cy5-dCTP, and SuperScript-II reverse transcriptase (Yu et ah, 2002).

2.2.8.1.2 RNA dénaturation

Since hybridisation on microarray slide requires pure labelled-cDNA, but not RNA (mRNA and rRNA), the latter needed to be removed from the reaction mixture. The RNA was denatured by the addition of NaOH (0.17 M) and incubation at 37°C for 15 minutes followed by the addition of 0.5 M HEPES buffer (N-[2-Hydroxyethyl] piperazine-N’-[2-ethanesulfonic acid] sodium salt). 2.2.8.1.3 Removal of unincorporated dye

The fluorescent cDNA probes were purified to maximise the hybridisation and minimise non-specific background signal. In order to remove excess unincorporated Cy3/Cy5 dye and nucleotide fragments produced during the RNA-denaturing step, a MinElute PCR purification column was used (Qiagen Ltd., Crawley, UK). The labelling reaction mixture was diluted with 5 volumes of PB buffer provided in the MinElute PCR purification kit and applied to the MinElute purification column, allowing labelled-cDNA to bind to the membrane. The membrane was washed by the addition of PE buffer, as per the manufacturer’s instructions, and the column centrifuged at 14,000xg (12,300 rpm) for 1 min using an Eppendorf microcentrifuge. Finally, labelled-cDNA was eluted with 20 pi Elution buffer by centrifugation for 1 min at 14,000xg (12,300 rpm).

2.2.8.1.4 Determination of Cy3/Cy5 dye concentration and degree of incorporation into cDNA

Since an equal amount of Cy3- and Cy5-labelled cDNA should be added to the array, it was important to determine the concentration of labelled-cDNA in the samples before hybridisation. The Cy3 and Cy5 dyes display maximum absorbance at 550 nm and 650 nm, respectively, as indicated in the product datasheet (Amersham Biosciences, Chalfont St. Giles, UK). ' A NanoDrop ND-1000 spectrophotometer was used to determine the Cy3/Cy5 dye concentration by measuring the absorbance of samples between 250-750 nm. The spectrophotometer was zeroed using DEPC-treated nuclease-fi*ee water and the absorbance of the labelled-cDNA samples (1 pi) between 250-750 nm determined (Fig. 2.10). The concentration of labelled-cDNA was calculated in pmol/pl based on the Beer-Lambert equation c=A/(s.b), using an extinction coefficient (s) of 0.15 pmol.cm.pl”^ (150,000 M"\cm"^) and 0.25 pmol.cm.pr^ (250,000 M"\cm"') for Cy3 and Cy5 dye, respectively. The total cDNA yield (ng) was obtained using a modified Beer-Lambert equation (c=A.s.b”^) where A is absorbance at 260 nm, b is path length (cm) and s is a modified wavelength-dependent extinction coefficient (37 ng.cm.pl"^ for single strand DNA), taking into account the total volume of labelled-cDNA sample (pi).

69 The data obtained from NanoDrop spectroscopy, was imported into an online calculator (http ://www. prontosystems.com/technical_support/calculatore/index.asp) to determine the degree of Cy-dye incorporation and calculate the volume (pi) of labelled cDNA required for hybridisation on a 60x24 mm microarray slide.

Re-blank Print Screen Start Report I Measutemenl complete 12/19/2004 6:47 PM Exit M easure Blank Print Report Show Report U ser ilraj

^ 0 .1 0 Max A bsorbance Sample Type Other

Constant 37 Sam ple i PI 9CL6 BSA-Cy5 ID ! Batch Index 0 " 0.06 X 260 Im m A bs ; 0.338 S 0.05 Dye 1 Cy3

< 0 04 Dye 1 Abs 0 005

>- 0 03 pmol/ul Dye 1 ; 0.4 Dye 2 ______CyS______v j Dye 2 Abs C o 050

pmol/ul Dye 2 :i 2.0 260/280 I 1.99 220 250 300 350 400 450 500 550 600 650 700 750 ng/ul ; 121.5 3.0.0 B839 Wavelength nm

Figure 2.10: Screen features of the NanoDrop ND-1000 spectrophotometer used to determine cDNA and Cy3/Cy5-dye concentrations. Samples were scanned between 220-750 nm and the absorbance for Cy3 and Cy5 dye determined at 550 and 650 nm, respectively.

2.2.8.1.5 Hybridisation of labelled-cDNA to microarrays and post-hybridisation wash

Hybridisation of the fluorescent cDNA onto microarrays was performed in three steps, as discussed below.

(i) Preparing the hybridisation buffer: in the first step, formamide was deionised by mixing with AG501-X8 (D) Mixed-Bed resin (5% w/v) and stirring for 2 hours. The supernatant was removed using Whatman #1 filter paper, and filter-sterilized (2 pm). Hybridisation buffer was prepared according to the HGMP protocol (http://www.hgmp.mrc.ac.uk/). The reagents were mixed in a sterile tube to final concentrations as follows: deionised formamide (40% v/v), Denhardfs

70 solution (0.1 % Ficoll-400, 0.1 % polyvinylpyrrolidone, and 0.1 % BSA), SSC (750 mM NaCl and 75 mM sodium citrate), SDS (sodium dodecyl sulphate, 0.1% w/v), sodium pyrophosphate (1 mM), and Tris (50 mM, pH 7.4). The buffer was filter-sterilized (2 pm) and stored in small aliquots at -20°C.

(ii) Hybridisation of targets on microarrays: in the hybridisation step, purified Cy3- and Cy5-labelled cDNA samples (40 pmol each) were mixed in pairs, according to the experimental plan, as shown for the P19CL6-embryo-heart comparison loop design (Fig. 2.11). The sample volume was reduced until targets were dried, using a Speed-Vac Uniscience Gyrovap centrifuge (VA Howe Ltd., Banbury, UK). Pellets were dissolved in 1 pi nuclease-free water to accelerate the dissolution of targets in hybridisation buffer (39 pi). Then poly A oligonucleotide (0.115 pg/pl) and mouse Cot-1 DNA (0.17 pg/pl) were added to suppress cross-hybridisation of cDNA to mouse repetitive DNA and tubes were incubated at 100°C for 5 minutes on a heating block, protected from light.

Tubes were cooled at room temperature for 10 min (protected from light) and hybridisation mixtures were then applied to the centre of arrays that had been placed in hybridisation chambers. Coverslips were placed over the arrays, taking care not to introduce air bubbles. A humidified atmosphere was provided by applying 30 pi of SSC (750 mM NaCl and 75 mM sodium citrate) into the wells on two sides of the chambers. Chambers were sealed and incubated at 50°C for 20 hours (protected from light) using a pre-warmed water-bath.

(iii) Post-hybridisation wash of microarray slides: for the post-hybridisation washing step of microarrays the slides were immersed in SSC/SDS solution [150 mM NaCl, 15 mM sodium citrate, and 0.2% (w/v) sodium dodecyl sulphate] for 2 min at room temperature to facilitate removal of the coverslips. Slides were then washed sequentially (2x5 min each step) using the following wash solutions: (a) SSC solution (150 mM NaCl and 15 mM sodium citrate), (b) SSC/SDS solution (15 mM NaCl, 1.5 mM sodium citrate, and 0.1% (w/v) sodium dodecyl sulphate), and (c) SSC solution (15 mM NaCl, 1.5 mM sodium citrate). Finally, arrays were dried by centrifugation at 70xg (800 rpm) for 5 min using an ICE Centra CL3 centrifuge (International Equipment Company, Needham Heights, USA) and stored in the dark until scanning.

71 P19CL6-Cy3 M ouse M ouse Embryo-Cy3 Heart-Cy3

M ouse M ouse Embryo-CyS Heart-CyS P19CL6-Cy5

Figure 2.11: Representation of (A) design and (B) description of loop comparison for P19CL6 ceils, mouse embryo and mouse heart total RNA. The RNA samples from P19CL6 cultured cells and mouse embryo and heart were used for cDNA synthesis and labelling. The Cy3- and Cy5-labelled probes were mixed (40 pmol of each) in pairs, according to the experimental design (Simon et al., 2002; Dobbin and Simon, 2002), and hybridised on 60x24 mm mieroarray slides. A dye-swap technique was performed to minimise the effect of different Cy3/Cy5 dye intensities. Arrows represent Cy3 to Cy5 direction and show comparisons between the samples throughout the loop. Samples 1 -3 are P19CL6, embryo and mouse heart total RNA.

2.2.S.2 Microarray procedure (ProntoPlus kit)

The ProntoPlus kit (Corning B.V. Life Sciences) provided all required solutions for cDNA synthesis, labelling and purification, and hybridisation.

2.2.8.2.1 cDNA synthesis and target labelling

Total-RNA was isolated using the Trizol reagent and used for cDNA synthesis and labelling according to the “direct incorporation” procedure (Yu et al., 2002). The synthesis of cDNA and target labelling was achieved as follows: The total-RNA sample (5 pg total-RNA) was placed in a microcentrifuge tube (on ice) and the volume adjusted to 20 pi with nuclease-ffee water after addition of random (0.15 pg/pl) and Oligo dT (0.1 pg/pl) primers. This RNA/primer solution was mixed thoroughly and incubated on a heating block at 70°C for 10 min. At the end of the incubation, the tube was chilled immediately on ice.

While the RNA/primer solution was incubated at 70°C, the labelling mixture was prepared separately in microcentrifuge tubes for the Cy3 and Cy5 dyes. ChipShot reverse transcriptase 5x reaction buffer (8 pi), MgCh (4.8 pi), dNTP mix (2 pi for Cy3 and 3 pi

72 for Cy5 reaction tube), Cy-dye (1 pi into relevant reaction tube) and ChipShot reverse transcriptase (3.2 pi), all provided by the manufacturer, were added into tubes and final volume was adjusted at 20 pi using nuclease-free water. To initiate cDNA synthesis, the entire labelling mixture was added to the RNA/primer solution, mixed, centrifuged briefly and incubated at room temperature for 10 min, protected from light. The reaction mixture (total volume 40 pi) was then incubated at 42°C for 2 hours on a heating block, protected from light. Finally, residual RNA was degraded and unincorporated dye was removed before the labelled cDNA was hybridised.

2.2.S.2.2 RNA dénaturation and removal of unincorporated dye

The RNA was denatured by the addition of RNa^g H (1 pi) and RNaje solution (0.35 pi) to the cDNA synthesis reaction. Reagents were mixed gently and incubated at 37°C for 15 min. Labelled cDNA was then purified as follows: binding particles (18 pi), sodium acetate (4 pi, pH 5.2) and binding solution (310 pi), all provided by the manufacturer, were added to the labelled-cDNA sample, mixed gently for 10 sec and incubated at room temperature for 1 min. The entire content of tube was applied to a Spin-X column and centrifuged at 10,000xg (11,000 rpm) for 1 min using an Eppendorf microcentrifuge. The flowthrough was discarded and the column washed with 500 pi of 80% ethanol by centrifugation at 10,000xg (11,000 rpm) for 1 min. This wash step was repeated three times. Finally, labelled cDNA was eluted from the column using 60 pi of elution buffer, provided by the manufacturer, and centrifuged at 10,000xg (11,000 rpm) for 1 min after incubation of column at room temperature (1 min). The eluted cDNA was stored in a light-proof container at -80°C. Cy-dye incorporation was assessed with a NanoDrop ND-1000 spectrophotometer, as discussed above.

2.2.8.2.3 Pre-soak and pre-hybridisation wash

Dust and weakly bound oligonucleotides were removed from the microarray slides by washing with Pre-Soak and Pre-Hybridisation solutions, prepared as the manufacturer’s instructions. Microarray slides were immersed in pre-warmed (42°C for 30 min) Pre-Soak Solution and incubated at 42°C for 20 min. Arrays were transferred

73 into Wash Solution 2 (5% v/v Universal Wash Reagent A in deionised water) and washed twice by incubation at room temperature for 30 sec. Arrays were then transferred into pre-heated (42°C for 30 min) Universal Pre-Hybridisation Solution for 15 min. Arrays were incubated twice in Wash Solution 3 (1% v/v Universal Wash Reagent A in deionised water) at room temperature for 30 sec. Finally, arrays were rinsed with nuclease-ffee water at room temperature and dried by centrifugation at 70xg (800 rpm) for 2 min using an ICE Centra CL3 centrifuge. Slides were used in hybridisation or stored in a dust-free slide holder.

2.2.8.2.4 Hybridisation of labelled-cDNA to microarrays and post-hybridisation wash

In the hybridisation step, purified Cy3- and Cy5-labelled cDNA samples (40 pmol each) were mixed in pairs, according to the experimental plan, and the reaction volume minimised using a Speed-Vac Uniscience Gyrovap centrifuge (VA Howe Ltd., Banbury, UK) until targets were nearly dried. Pellets were dissolved in 40 pi of Universal Hybridisation Solution (1 pi for each pmol of labelled-cDNA) and tubes were incubated at 95°C for 5 minutes on a heating block, protected from light. Condensate was collected by centrifugation at 13,500xg (12,000 rpm) for 2 min and applied on the centre of the arrays that had been placed in hybridisation chambers, as discussed above for the HGMP method. Chambers were sealed and incubated at 42°C for 20 hours (protected from light) using a pre-warmed water-bath.

At the end of the incubation period microarray slides were immersed in pre-warmed (42°C for 30 min) Wash Solution 1 (10% v/v Universal Wash Reagent A and 0.5% v/v Universal Wash Reagent B in deionised water) for 2 min to remove coverslips. Slides were then transferred into fresh pre-warmed (42°C for 30 min) Wash Solution 1 for 5 min. Arrays were incubated in Wash Solution 2 (5% v/v Universal Wash Reagent A in deionised water) for 10 min at room temperature and were then incubated with Wash Solution 3 (1% v/v Universal Wash Reagent A in deionised water) at room temperature for 2 min. The latter wash step was repeated and slides were dried by centrifugation at 70xg (800 rpm) for 2 min using an ICE Centra CL3 centrifuge and stored in the dark until scanning. 2.2.S.3 Microarray scanning and data capture

Dust was removed from hybridised slides using dust-pro spray. Slides were then placed in an Afrymetrix 428 scanner (Affymetrix; Santa Clara, USA) that excited the Cy3/Cy5-dye labels with a laser and measured the emitted light at 532 and 635nm (Kawasaki, 2004). The laser power and photomultiplier tube settings (gain) were adjusted to achieve good signal-to-noise ratios while avoiding signal saturation. At each microarray coordinate, Cy3 or Cy5 fluorescence intensity is directly proportional to the amount of labelled-cDNA present. The pixel intensities were converted to a 16-bit “TIFF” image format for visualisation and saved for each channel separately, Cy3 and Cy5, until further image and data analysis.

2.2.8.4 Image analysis

Images scanned for Cy3 and Cy5 fluorescence were overlaid using BlueFuse software (BlueGenome Ltd., Great Shelford, UK) and analysed in two steps: placing of the grids, and quantification. Whereas placing the grids is a manual procedure, the BlueFuse software is able to quantitate the signals for each spot automatically based on the spot coordinates and the grid map. The software utilizes auto-segmentation and edge detection to calculate spot intensities and background noise. The overall dimensions of the microarrays are approximately 53x18 mm including 48 sub-grids, arranged in 12 rows and 4 columns, with 18 rows and 20 columns in each sub-grid (Fig. 2.12). Average spot diameter was 160 pm and the average distance from spot-to-spot centre was 220 pm. The array consists of 7,524 oligonucleotides, which represent 7,445 sequences, plus 79 controls. The array description file (Mm_SGC_AV2.gal) obtained from Human Genome Mapping Project website (HGMP; http://www.hgmp.mrc.ac.uk/), describes the coordinates of each spot on the array. Once scanned images for Cy3 and Cy5 were overlaid, coordinates of positive and negative “control spots” were defined and a theoretical sub-grid map adjusted on the spots manually.

75 MCI MC2 MCS MC4 MR1 MR1 MR1 MR1

MCI MC2 MCS MC4 MR2 MR2 MR2 MR2

MCI MC2 MCS MC4 MRS MRS MRS MRS

MCI MC2 MCS MC4 00000000000000000000 MR4 MR4 MR4 MR4

MC1 MC2 MCS MC4 MRS MRS MRS MRS

+++++++++++++++++++ MC1 MC2 MCS MC4 +++++++++++++++++ MR6 MR6 MR6 MR6 îîîîîîîîîîîîîîîîîîîî MCI MC2 MCS MC4 MR7 MR7 MR7 MR7

MC1 MC2 MCS MC4 MR8 MR8 MR8 MR8

MCI MC2 MCS MC4 MR9 MR9 MR9 MR9

Theoretical map of MCI MC2 MCS MC4 spots coordinates MR10 MR10 MR10 MR10 MCI MC2 MCS MC4 MR11 MR11 MR11 MR11 • Mouse known oligos MCI MC2 MCS MC4 MR12 MR12 MR12 MR12 ° Negative and positive control spots, bacterial probes, tissue-specific genes, and landing lights

Figure 2.12: Layout of Mouse Known Gene SGC Oligo Set Array, version 2 (~7.5k oligos). Array includes 48 sub-grids, arranged in 12 rows and 4 columns, with 18 rows and 20 columns in each sub-grid. There are -7,500 mouse known genes on the microarray slide (shown in black). The array map file identifies the spots based on theoretical coordinates of positive and negative control spots (shown in grey) and quantifies the pixel intensity for all spots. MC and MR of each sub-grid indicate the position of sub-grid on the microarray slide.

Once grid adjustment was completed, the pixel intensities for spots was determined. The BlueFuse software identified each spot in the overlaid image files fi"om a scanned array (Fig. 2.13) and quantified the pixel intensity of the spot and its (local) background. This software also identified and excluded regions of the spot that behaved significantly differently from the norm (such as specks of dust) and assigned flags to indicate the quality of the data. However, images were also checked manually and artefact areas were flagged. The spots were assigned confidence values ranging 0-1 automatically, with a value of one corresponding to the highest spot quality. Since the array consists of

76 duplicate spots for each gene, the measured intensities were therefore fused to obtain an average signal for each gene and an Excel-type spreadsheet was generated for each slide to use in further analyses. Scatter plots were generated (Fig. 2.14) based on the signal ratio of channel-2 to channel-1 (Ch2:Chl; Cy5 and Cy3 signals, respectively).

Figure 2.13: An example of an overlaid image generated after microarray scanning and analysis using the BlueFuse Software. Scanned images for Cy3 and Cy5 were imported into BlueFuse programme. Background and artefact signals were exeluded and signals for each microarray spot quantified based on the spot coordinates.

77 100000

10000

g 1000 0 1 Ol < 100

10

10 100 1000 toooo 100000 Amplitude Chl Figure 2.14: An example of a scatter plot generated after microarray scanning and analysis using the BlueFuse Software. The labelled cDNA from P19CL6 cells and mouse embryo were hybridised onto a microarray, which was then scanned. Cy3 and Cy5 signals (Chl and Ch2, respectively) were quantified for each microarray spot and spots were given a colour flag, ranging from red, orange, blue, light blue, and green representing confidence values from 0-1. Finally, the distribution of spots was plotted based on the signal ratio of P19CL6-Cy3 and mouse embryo-Cy5.

2.2.8.S Data analysis

2.2.8.5.1 Normalisation and filtering of microarray data

A number of statistical algorithms have been developed to normalise microarray data and to control for experimental variations (Goldsmith and Dhanasekaran, 2004). The procedure was optimised by self-against-self hybridisation of P19CL6 cDNA labelled with both Cy3 and Cy5, according to the HGMP protocols. In order to analyse microarray data, the Excel-type output data generated by BlueFuse software was imported into the Gene Spring-7 programme (Silicon Genetics; Redwood City, USA). For the all arrays, Cy3 signal (test signal) was compared with Cy5 signal (control signal), therefore, the gene expression levels were expressed as the Cy3:Cy5 signal ratio. Microarray data were then analysed in several steps. First, the intensity-dependent per spot and per chip normalisation, often called non-linear or LOWESS normalisation, was performed on each

78 array. Normalisation eliminated dye-related artefacts and corrected artefacts caused by non-linear rates of dye incorporation as well as inconsistencies in the relative fluorescence intensity between Cy3 and Cy5 dyes. Since such artefacts often result in a curve in the graph of raw (test signal) versus control signal, normalisation fits a straight line through the data and adjusts the control value for each spot. In addition, a dye-swap technique, used in the experimental loop design minimised the differences in Cy3 and Cy5 incorporation (Goldsmith and Dhanasekaran, 2004).

After normalisation, the data set was filtered to eliminate the spots with low quality. Filtering was performed based on “filter on flags” and “filter on data file” methods, the latter was applied in two steps including “filter on uniformity” and “filter on circularity”. For all experiments, spot uniformity and circularity were restricted at being greater than 0.5 (out of 1.0) in at least 50% of the arrays. Data normalisation and filtering were first performed on the self-against-self hybridisation experiment (Fig. 2.15 A) in order to optimise the procedures. Normalisation and filtering eliminated the spots with low quality and generated a scatter plot representing acceptable data (Fig. 2.15B).

Self hybridisation Self hybridisation ofP19CL6 y /.f' ofP19CL6 Before normalisation & filtering , *v ' After normalisation & filtering

P19CL6-Cy5 P19CL6-Cy5

Figure 2.15: Illustration of the scatter plots for the self-against-self hybridisation of P19CL6 cells before (A) and after (B) data normalisation and filtering. RNA template was isolated from P19CL6 cells and divided into two aliquots that were labelled with Cy3 and Cy5 dye, respectively. Targets were mixed and hybridised on a single array. The array was scanned and the image analysed. Data was normalised based on the intensity-dependent per spot and per chip normalisation method and filtered according to the flag statues, spot uniformity and spot circularity. The scatter plots of self-hybridisation experiment were generated (A) before and (B) after data normalisation and filtering processes.

79 Once normalisation and filtering was completed, a scatter plot was generated to demonstrate the relative distribution pattern of gene expression levels, where X- and Y-axes represent Cy5 (control) and Cy3 (test) channels, respectively (Fig. 2.16). Spots were expressed in different colours, based on Cy3:Cy5 signal ratio. Generally, the red spots represent the genes over-expressed in the sample labelled with Cy3, whereas the green spots indicate the genes under-expressed in the same sample, i.e. over-expressed in the sample that was labelled with Cy5 (Fig. 2.16). However, the majority of the genes, with Cy3:Cy5 signal ratio close to 1 (yellow spots), showed no differential expression. In order to demonstrate the genes with distinct changes in their expression levels, the cut-off point was defined at ±2-fold expression, therefore, the spots of the outer sides of ±2-fold gene expression lines were regarded as over- or under-expressed depending on the samples tested (Fig. 2.16).

P19CL6 vsH9C2(2-1) w 4A

. : # r >■«

H9C2(2-1)-Cy5

Figure 2.16: An example of scatter plot generated by GeneSpring-7 software to represent the gene expression levels based on Cy3:Cy5 signal ratio for P19CL6vs H9C2(2-1). Samples were labelled with Cy-dyes and hybridised onto a microarray. The array was scanned and data normalised and filtered. A scatter plot was generated to demonstrate the relative distribution pattern of gene expression levels based on the Cy3:Cy5 signal ratio represented in different colours (arrow). Red and green colours represent up- and down-regulation of the genes in the test sample (Cy3-labelled), respectively. The cut-off point of significant changes in gene expression level was defined at ±2-fold changes.

8 0 2.2.8.S.2 Linear modelling and cluster analysis of microarray data

The dual-hybridisation microarray data was analysed with a per-gene linear model, using the Live Linear Analysis of Micro Array programme (LLAMA) available on the MARIE:ExGen Consortium website (http://exgen.rna.umist.ac.uk/), to generate differential gene expression estimates between contrast pairs , i.e. Cy3:Cy5 ratios, in a log 2 scale. The data produced by linear modelling was further analysed by a web-based Complete Linkage Clustering software (http://rana.lbl.gov/) (Eisen et al., 1998) to generate a pair-wise linkage clusters. Analysis formed a hierarchical cluster, which was visualised with TreeView software (http://rana.lbl.gov/) (Eisen et al., 1998). Microarray data was also used to perform principal component analysis (PCA) and to generate correlation matrix in order to estimate correlation of samples using SPSS for Windows (SPSS Inc., Chicago, USA).

81 3

Characterisation of the P19CL6 cell-line

82 Chapter 3:

Characterisation of the P19CL6 cell-line

3.1 Introduction

Fatty acids and fibrates are able to regulate lipid-related genes expression in various cells (Mesa et al., 2004; Jump, 2004; Guo et al., 2001; Chapman, 2003). To determine the effect of fatty acids and fibrates on gene expression levels in cardiovascular cells the P19CL6 cell line, a clonal derivative of P19 mouse embryonal carcinoma cells (Habara- Ohkubo, 1996), was used as a cellular model for the cardiovascular system. During the course of various incubations of P19CL6 cells in the presence of DMSO, required for the differentiation of these cells into beating cardiomyocytes (Habara-Ohkubo, 1996; Skeijanc, 1999), it was noted that a “beating” phenotype was observed in some cultures, but not others. As one of the initial aims of this thesis was to investigate the effects of fatty acids and fibrates in cardiomyocytes it was imperative that the cell-type being used was both correctly characterised and cultured under the correct conditions to facilitate beating.

There are several potential reasons for not observing the “beating” phenotype, namely that; (a) the cells display a “mixed” phenotype, (b) have lost their cardiomyocyte properties, (c) cell density is too low or (d) the cell culture media affects the cells. These phenomena have been observed in other cell culture systems. For example, the H9C2(2-1) cell line was derived from H9C2 embryonic rat heart tissue, and was initially described to have skeletal muscle cells properties (Kimes and Brandt, 1976). However, subsequent investigations showed that these cells also exhibited a cardiomyocyte-specific subtype of calcium channel (Hescheler et al., 1991; van Nieuwenhoven et al., 1998). Furthermore, the properties of the Caco-2 enterocyte model system have been shown to be passage-dependent (Yu et al., 1997). In terms of the cell culture media it is known that adrenaline affects the “beating” of P19CL6 cells (van der Heyden and Defize, 2003; McBumey et al., 1982; Faquin et al., 2002; Takahashi et al., 2003) and therefore the serum used to make the culture media could affect cell characteristics.

83 In view of this finding, the aim of these studies was to characterise the P19CL6 cells in more depth, with the specific objectives and rationales as follows:

Objective 1: To measure the gene expression profiles of cardiac- and skeletal-muscle specific proteins, comparing PI9CL6 and H9C2(2-1) cells, mouse heart and skeletal muscle tissues.

Contractile muscle cells consist of thick and thin filament proteins, which are responsible for the striated appearance of cardiac and skeletal muscle myotubes (Gulick et al., 1991). Myosin heavy chain (MHO) is a member of a multigene family and a major protein in the thick muscle filaments (Robbins et al., 1986; Metzger et al., 1995). The distribution of MHC isoforms in different tissues is cell type-specific (Gulick et al., 1991), with cardiac isoforms (a-MHC and p-MHC) predominantly expressed in the heart (Gulick et al., 1991; Metzger et al., 1995). Likewise, myosin light chain-2 (MLC2), atrial natriuretic factor (ANF), and cardiac isoform of voltage-dependent a-calcium channel (CaCh-h) are known as cardiac-specific genes (Fassler et al., 1996). On the other hand, the skeletal muscle-specific isoform of the a-calcium channel (CaCh-m) and the genes encoding myogenic regulatory factors MyoD and Myogenin, are expressed only in skeletal muscle cells and are required for skeletal muscle formation and myogenesis (Sabourin and Rudnicki, 2000; Buckingham, 2001; Fassler et al., 1996). Thus, in this study the mRNA expression levels of cardiac- (a-MHC, P-MHC, MLC2, ANF, and CaCh-h) and skeletal muscle-specific genes (MyoD, Myogenin and CaCh-m) was used to characterise the P19CL6 cells, comparing them to H9C2(2-1) cells, and also mouse heart and skeletal muscle tissue.

Objective 2: Determine whether the “beating” characteristic of P19CL6 cells is dependent on passage number.

The morphogenic appearance, growth rate, gene expression profile, and endocrine function of some cells e.g. Caco-2 enterocyte, SaOS-2 osteoblast-like, and FRTL-5 thyroid cells is known to be passage-dependent (Green et al., 2001; Arkell and Jackson, 2003; Gao et al., 2002; Yu et al., 1997). Since P19CL6 cells are relatively poorly characterised, it was important to investigate the effect of passage number on the “beating” characteristic of PI 9CL6 cells.

84 Objective 3: Ascertain the effect of foetal bovine serum (FBS) and adrenaline on the “beating” characteristic of PI 9CL6 cells.

Foetal bovine serum contains a variety of nutrients required for cell growth and development, and differences in the composition or concentration of these nutrients may affect cellular differentiation and gene expression profiles (van der Heyden and Defize, 2003; Skeijanc, 1999; Takahashi et ak, 2003; Nag et al., 1985; Wilton and Skeijanc, 1999). In addition, vitamins and hormonal content of culture media such as ascorbic acid, adrenaline, oxytocin, and thyroid hormones (T3), have been reported to act as potent inducers of cell differentiation and beating in cardiomyocytes (Habara-Ohkubo, 1996; van der Heyden and Defize, 2003; Paquin et al., 2002; Takahashi et al., 2003; Rodriguez et al., 1994; Pennec et al., 2002; Wobus et al., 1994). Therefore, the effect of different FBS samples was investigated to determine whether different batches affected the growth characteristics, e.g. “beating” phenomenon of P19CL6 cells. In addition, the effect of adrenaline on the “beating” phenotype of PI 9CL6 cells was investigated in more detail.

Objective 4: To use microscopic cell biology techniques to characterise the P19CL6 cells.

P19-derived cardiomyocytes are mono-nucleated cells (Skeijanc, 1999), while skeletal muscle cells propagate as mono-nucleated myoblasts and, with time, begin to form multi-nucleated tubular structures (Kimes and Brandt, 1976; Skeijanc, 1999). Microscopic analysis, including nuclear staining, was used to differentiate between the potentially cardiac-like or skeletal muscle-like phenotype of P19CL6 cells.

85 3.2 Comparison of gene expression profiles of cardiac- and skeletal muscle-specific genes in P19CL6 and H9C2(2-1) cell-Iines

3.2.1 Rationale

A review of the literature on cell culture conditions for P19CL6 and PI9 cells showed that essentially two conditions were used in the culture of these cells (Habara- Ohkubo, 1996; Morley and Whitfield, 1993; van der Heyden et ak, 2003; van der Heyden and Defize, 2003; Skeijanc, 1999). The efficiency and rate at which P19 embryonic cells can be induced to differentiate into beating cardiomyocytes has been shown to vary according to the differentiation protocol (Anisimov et ak, 2002). The original paper describing the isolation of PI9CL6 cells and characterising them as cardiomyocytes used adherent culture conditions (Habara-Ohkubo, 1996). However, a substantial number of papers, especially those related to investigation with P19 cells (van der Heyden et ak, 2003; van der Heyden and Defize, 2003; Skeijanc, 1999; McBumey, 1993) used non-adherent conditions {i.e. growth on agar plates to produce cell aggregates) before re-plating under adherent conditions (Smith et ak, 1987; van der Heyden et ak, 2003; Skeijanc, 1999). In this investigation, both adherent and non-adherent culture conditions were studied to establish their effects on gene expression levels.

3.2.2 Results

Separation of RT-PCR products by agarose gel electrophoresis showed that GAPDH, a house-keeping gene (Ullmannova and Haskovec, 2003), was detected in both P19CL6 and H9C2(2-1) cells cultured under adherent conditions, thus confirming the presence of mRNA in the samples tested (Fig. 3.1). Under adherent conditions, P19CL6 cells expressed the cardiac-specific gene a-MHC, but not p-MHC, whereas H9C2(2-1) cells did not express either MyoD or Myogenin, which are specific to skeletal muscle cells, under the same conditions (Fig. 3.1). Similar results were produced when the experiment was repeated (data not shown).

86 Likewise, the detection of nested PCR products showed expression of GAPDH in both P19CL6 and H9C2(2-1) cells, although, multiple PCR products for GAPDH were detected (Fig. 3.2), indicating poor specificity of the GAPDH primers. Therefore new primers were designed for detecting GAPDH expression levels in further experiments. Under adherent conditions, P19CL6 cells expressed the cardiac-specific genes a-MHC

GAPDH a-MHC p-MHC MyoD Myogenin

500 bp

400 bp

300 bp ' f "

200bp ' ' t

Figure 3.1: Expression of a-MHC, P-MHC, MyoD and Myogenin in (A) P19CL6 and (B) H9C2(2-1) cells, cultured under adherent conditions on exposure to 1% DMSO, as detected by RT-PCR. Cells were cultured under adherent conditions for 15 days in the presence of 1% DMSO. Total RNA was isolated and mRNA levels determined by RT-PCR using gene-specific primers (Table 2.1). Theoretical sizes of gene-specific PCR products were: GAPDH: 362 bp; a-MHC: 302 bp; p-MHC: 513 bp; MyoD: 392 bp; Myogenin: 337 bp, as shown by arrows.

GAPDH a-MHC |3-MHC MyoD Myogenin A B A B A B A B A B

*

500 bp 400 bp ... . ; ■"j, mm

300 bp t t 200 bp

Figure 3.2: Expression of a-MHC, P-MHC, MyoD and Myogenin in (A) P19CL6 and (B) H9C2(2-1) cells, cultured under adherent conditions on exposure to 1% DMSO, as detected by nested PCR. Cells were cultured under adherent conditions for 15 days in the presence of 1% DMSO. Total RNA was isolated and mRNA levels determined by RT-PCR followed by nested PCR using gene-specific primers (Table 2.2). Theoretical sizes of gene-specific PCR products were: GAPDH: 275 bp; a-MHC: 302 bp; P-MHC: 410 bp; MyoD: 392 bp; Myogenin: 337 bp, as shown by arrows.

87 and p-MHC, as well as the skeletal muscle-specific markers MyoD and Myogenin, whereas H9C2(2-1) cells expressed only MyoD and Myogenin under the same conditions (Fig. 3.2).

In the P19CL6 cells cultured for 15 days under non-adherent conditions, expression of cardiac-specific a-MHC and p-MHC as well as skeletal muscle-specific genes (MyoD and Myogenin) was observed under all conditions, independent of the presence or absence of DMSO (Fig. 3.3). A time-dependent gene expression profile of P19CL6 cells cultured under non-adherent conditions showed expression of GAPDH occurred at all time points throughout the incubation period, confirming the presence of mRNA in the samples (Fig. 3.4). Under these conditions, P19CL6 cells expressed cardiac-specific genes (a-MHC and P-MHC) and skeletal muscle markers (MyoD, and Myogenin) at all time points.

PD/agar (NCM-NCM) PD/agar (DM-NCM) PD/agar (DM-DM) 1 ! i 111 H I l i Ï Ï 11

500 bp U U 1 400 bp mm ^ I J i . I

300 bp B Figure 3.3: Expression of a-MHC, P-MHC, MyoD and Myogenin in P19CL6 cells cultured under non-adherent conditions, as determined by nested PCR. P19CL6 cells were cultured on petri dishes in the presence of either differentiation (with DMSO) or normal culture media, i.e. DM and NCM, respectively. After four days, the aggregated cells were transferred into standard tissue culture flasks and incubated for 15 days with either DM or NCM. RNA was isolated from the cells at the end of the 15 days incubation period and mRNA levels determined by RT-PCR followed by nested PCR, using gene-specific primers (Table 2.2). Theoretical sizes of gene-specific PCR products were; GAPDH: 413 bp; a-MHC: 302 bp; P-MHC: 410 bp; MyoD: 392 bp; Myogenin: 337 bp, as shown by arrows.

88 A semi-quantitative estimation of gene expression levels, relative to GAPDH expression, showed that the expression of a-MHC, P-MHC, MyoD and Myogenin increased up to day-9 and then decreased (Fig. 3.5). The lowest relative gene expression levels were observed on day-3, whereas the highest values were displayed on day-9. In addition, relative expression of a-MHC and p-MHC was higher than MyoD and Myogenin at all time points. Moreover, unlike differences in the expression levels of a-MHC and p-MHC, the skeletal-specific genes (MyoD and Myogenin) were expressed at similar levels between day 7 and day 15 (Fig. 3.5).

Day-3 Day-7 Day-9 Day-12 Day-15 .£ “ .£ .£ . .£ X OI i à<à-S2 I t I OI i S I 5Î 5I CDI Ï 6 (à.i 2t 5I CDI i s<à22CDè I i II i

500 bp

400 bp

Figure 3.4: Expression of a-MHC, p-MHC, MyoD, and Myogenin in P19CL6 cells cultured under non-adherent conditions, as determined by nested PCR. P19CL6 cells were cultured on petri dishes in the presence of 1% DMSO. After four days, the aggregated cells were transferred into standard tissue culture flasks and incubated for 15 days with differentiation media (DM). Cells were collected by scraping into 1 ml phosphate buffered saline (PBS) on time points of day 3, 7, 9, 12, and 15. RNA was isolated from the cells and mRNA levels determined by RT-PCR followed by nested PCR, using gene-specific primers (Table 2.2). Theoretical sizes of gene-specific PCR products were: GAPDH: 413 bp; a-MHC: 302 bp; P-MHC: 410 bp; MyoD: 392 bp; Myogenin: 337 bp, as shown by arrows.

89 Day-12 Day-15

Figure 3.5: Time course effect of incubation period on the expression of a-MHC, P-MHC, MyoD, and Myogenin in P19CL6 cells cultured under non-adherent conditions, as determined by nested PCR. P19CL6 cells were cultured on petri dishes in the presence of 1 % DMSO. After four days, the aggregated cells were transferred into standard tissue culture flasks and incubated for 15 days with differentiation media (DM). Cells were collected by scraping into 1 ml phosphate buffered saline (PBS) on time points of day 3, 7, 9, 12, and 15. Total RNA was isolated from the cells and mRNA levels deterrnined by RT-PCR followed by nested PCR, using gene-specific primers (Table 2.2). PCR products were separated on 2% agarose gel and the gel photographed. The image was digitised using Scion Image programme to obtain a semi-quantitative estimation of gene expression levels, relative to GAPDH expression. Figure shows data from a single experiment. HI a-MHC; S P-MHC; C MyoD; D Myogenin

3.3 Determining gene expression profiles of a-MHC, P-MHC, MyoD and Myogenin in mouse heart and skeletal muscle tissues

3.3.1 Rationale

Previous work has shown that the expression of mouse cardiac myosin heavy chain isoforms (a-MHC and P-MHC) is restricted to contracting cardiac myocytes (Metzger et al., 1995; Morkin, 2000; Skeijanc, 1999), whereas regulation of skeletal muscle differentiation centers mainly on a family of transcription factors including MyoD and Myogenin (Edmondson and Olson, 1989; Yun and Wold, 1996; Buckingham, 2001; Black et al., 1995). Since a-MHC, p-MHC, MyoD and Myogenin mRNA was detected in P19CL6 cells, it was important to check the stringency of primers in mouse heart and skeletal muscle tissues. 3.3.2 Results

The expression of cardiac-specific genes, a-MHC and P-MHC, was detected in mouse heart tissue by RT-PCR (Fig. 3.6). In contrast, the skeletal muscle tissue did not express skeletal muscle-specific genes, MyoD and Myogenin, under the PCR conditions used (Fig. 3.6). Similar results were produced when the experiment was repeated (data not shown). These results could be explained by very low levels of mRNA for these genes, and therefore nested PCR was used to further amplify these signals. The detection of nested PCR products confirmed the expression of a-MHC and p-MHC in the heart tissue, while, under these conditions the expression of both MyoD and Myogenin was also apparent (Fig. 3.7). Similarly, skeletal muscle tissue displayed expression of MyoD and Myogenin (skeletal muscle-specific genes) as well as a-MHC and p-MHC (cardiac-specific genes).

Mouse Heart Mouse Skeletal Muscle

600 bp 500 bp 400 bp

Figure 3.6: Detection of a-MHC, P-MHC, MyoD and Myogenin in mouse heart and skeletal muscle tissue, as determined by RT-PCR. Heart and skeletal muscle tissues were dissected from 10 weeks old GDI wild type (+/+) male mice. Total RNA was extracted and gene expression levels determined by RT-PCR using gene-specific primers (Table 2.1). RT-PCR products were separated on 2% agarose gel containing ethidium bromide and gel photographed. Theoretical sizes of gene-specific PCR products were: GAPDH: 556 bp; a-MHC: 491 bp; p-MHC: 587 bp; MyoD: 513 bp; Myogenin: 608 bp, as shown by arrows.

91 Mouse Heart Mouse Skeletal Muscle I I i i I I I i Î I

500 bp

400 bp JL rnÊ rnm m 300 bp

Figure 3.7: Expression of a-MHC, P-MHC, MyoD and Myogenin in mouse heart and skeletal muscle tissue, as determined by nested PCR. Heart and skeletal muscle tissues were dissected from 10 weeks old GDI wild type (+/+) male mice. Total RNA was extracted and gene expression levels determined by RT-PCR followed by nested PCR using gene-specific primers (Table 2.2). PCR products were separated on 2% agarose gel containing ethidium bromide and gel photographed. Theoretical sizes of gene-specific PCR products were: GAPDH: 413 bp; a-MHC: 302 bp; p-MHC: 410 bp; MyoD: 392 bp; Myogenin: 337 bp, as shown by arrows.

3.4 Determining gene expression profiles of ANF, MLC2, CaCh-h, and CaCh-m in P19CL6 cells, mouse heart and skeletal muscle tissues

3.4.1 Rationale

Since the outcome from the investigation of a-MHC, P-MHC, MyoD, and Myogenin genes was inconclusive, the expression of other genes was examined. These were atrial natriuretic factor (ANF) and myosin light chain-2 polypeptide (MLC2) as well as cardiac (CaCh-h) and skeletal muscle (CaCh-m) isofonus of voltage-dependent L-type calcium channel (Fassler et ak, 1996).

92 3.4.2 Results

RT-PCR analysis confirmed that mouse heart tissue expressed the cardiac-specific genes including a-MHC, p-MHC, ANF, MLC2, and CaCh-h (Fig. 3.8). Under these conditions, the expression of a-MHC and P-MHC was not detected in skeletal muscle tissue and several faint bands were observed in a-MHC and p-MHC lanes that were different from the theoretical target sizes of the respective genes (Fig. 3.8), although expression of these genes was previously confirmed in skeletal muscle tissue by nested PCR (Fig. 3.7). Unlike a-MHC and P-MHC, the expression of ANF, MLC2, and CaCh-h was observed in the skeletal muscle tissue under conditions examined (Fig. 3.8).

Mouse heart tissue Mouse skeletal muscle tissue P19CL6 cells

X Ü Ü Ü 0 UL O s X 1 i l Î I 5 i I Î li ii

600 bp

500 bp

400 bp

300 bp I I 200 bp

Figure 3.8: Expression of a-MHC, p-MHC, ANF, MLC2 and CaCh-h in mouse heart, skeletal muscle tissue and P19CL6 cells, as detected by RT-PCR. Heart and skeletal muscle tissues were dissected from 10 weeks old CDl wild type (+/+) male mice and total-RNA was extracted from tissues as well as cultured P19CL6 cells, as indicated in Materials and Methods. Gene expression levels were determined by RT-PCR, using gene-specific primers (Table 2.1). PCR products were separated on 2% agarose gel containing ethidium bromide and gel photographed. Theoretical sizes of gene-specific PCR products were: GAPDH: 556 bp; a-MHC: 302 bp; P-MHC: 205 bp; ANF: 203 bp; MLC2: 189 bp; and CaCh-h: 183 bp, as shown by arrows.

93 In keeping with mouse heart and skeletal muscle tissues, P19CL6 cells expressed the CaCh-h gene, but again, the expression of a-MHC and p-MHC and other cardiac-specific genes (ANF and MLC2) was undetectable in the P19CL6 cells by RT-PCR (Fig. 3.8), although expression of a-MHC and p-MHC was previously confirmed in P19CL6 cells by nested PCR (Fig. 3.2, Fig. 3.3 and Fig. 3.4). Similar results were observed when the experiment was repeated (data not shown), confirming the expression of all cardiac-specific genes in the heart tissue, ANF, MLC2, and CaCh-h in skeletal muscle, and CaCh-h in P19CL6 cells. Examination of the cardiac and skeletal muscle isoforms of L-type calcium channel showed that mouse heart and P19CL6 cells expressed only the cardiac isoform (CaCh-h), whereas skeletal muscle tissue expressed both cardiac and skeletal muscle isoforms (Fig. 3.8 and Fig. 3.9).

m ouse skeletal Mouse heart muscle P19CL6 cells

_

500 bp ***^

300 bpjllljj illlHIHIIBli : 200 b p # # : I # B K # Figure 3.9: Expression of cardiac and skeletal muscle isoforms of voltage-dependent calcium channel (CaCh-h and CaCh-m, respectively) in mouse tissues and P19CL6 cells, as detected by RT-PCR. Heart and skeletal muscle tissues were dissected from 10 weeks old CDl wild type male mice (+/+) and RNA was isolated from tissues as well as cultured P19CL6 cells. Gene expression was determined by RT-PCR and PCR products separated on 2% agarose gels containing ethidium bromide. The primers used are described in Table 2.1 and theoretical sizes of gene-specific PCR products were: GAPDH: 556 bp; CaCh-h: 183 bp; and CaCh-m: 204 bp, as shown by arrows. 3.5 DNA sequencing of PCR products and sequence comparison

3.5.1 Rationale

RT-PCR and nested PCR analysis showed that certain cardiac-specific and skeletal muscle-specific genes, e.g. a-MHC, p-MHC, MyoD, and Myogenin, were expressed in P19CL6 cells. Since this finding had not previously been reported the validity of the results was checked by confirming the sequence identity of the PCR products.

3.5.2 Results

PCR product sequences were compared to the known gene sequences published in NCBI GenBank database, as shown for MyoD (Fig. 3.10), to determine the identity of PCR products in relation to known sequences. Identity was fiirther confirmed by comparison of the PCR product sequences to the Mouse musculus Genome database and

MyoD TCTGCAGGCTCTGCTGCGCGACCAGGACGCCGCGCCCCCTGGCGCCGCTGCCTTCTACGC 536 Sam ple — TGCAGGCTCTGCTGCGCGACCAGGACGCCGCGCCCCCTGGCGCCGCTGCCTTCTACGC 91 ********************************************************** MyoD ACCTGGACCGCTGCCCCCAGGCCGTGGCAGCGAGCACTACAGTGGCGACTCAGACGCGTC 596 Sam ple ACCTGGACCGCTGCCCCCAGGCCGTGGCAGCGAGCACTACAGTGGCGACTCAGACGCGTC 151 ************************************************************ MyoD CAGCCCGCGCTCCAACTGCTCTGATGGCATGATGGATTACAGCGGCCCCCCAAGCGGCCC 655 Sam ple CAGCCCGCGCTCCAACTGCTCTGATGGCATGATGGATTACAGCGGCCCCCCAAGCGGCCC 211 ************************************************************ MyoD CCGGCGGCAGAATGGCTACGACACCGCCTACTACAGTGAGGCGGTGCGCGAGTCCAGGCC 714 Sam ple CCGGCGGCAGAATGGCTACGACACCGCCTACTACAGTGAGGCGGTGCGCGAGTCCAGGCC 271 ************************************************************ MyoD AGGGAAGAGTGCGGCTGTGTCGAGCCTCGACTGCCTGTCCAGCATAGTGGAGCGCATCTC 774 Sam ple AGGGAAGAGTGCGGCTGTGTCGAGCCTCGACTGCCTGTCCAGCATAGTGGAGCGCATCTC 331 ************************************************************ MyoD - CACAGACAGCCCCGCTGCGCCTGCGCTGCTTTTGGCAGATGCACCACCAGAGTCGCCT 834 Sam ple CACAGACAGCCCCGCTGCGCCTGCGCTGCTTTTGGCAGATGCACCACCA ------382 *************************************************

Figure 3.10; An illustration of a sequence alignment for a PCR product and the targeted gene sequences. The DNA sequence of the nested PCR product was compared to the MyoD sequence, as obtained from NCBI, taking into account the positions of the PCR primers utilised, using the ClustalW sequence alignment software. The figure shows that the 382 bp PCR product (sample) is 100% identical to the Mouse musculus MyoD mRNA in the region starting and finishing at nucleotides 477 bp and 834 bp, respectively, based on the positions of the PCR primers utilised. The numbers indicated on the right-hand side relate to the nucleotide number relative to the start of the gene.

95 ranged from 93% to 100% (Table 3.1). This was considered to be acceptable in terms of identity between the published sequences and those obtained from the PCR experiments. Therefore, the PCR products were clearly related to the target genes used in the design of primers.

Table 3.1: Identity of PCR product sequences to those of target genes.

S o u r c e P19CL6 ceiis

G e n e GAPDH a-MHC p-MHC MyoD Myogenin

Id e n tity 99% 99% 98% 100% 98%

S o u r c e H9C2(2-1) cells

G e n e GAPDH a-MHC P-MHC MyoD Myogenin

Id e n tity 93% ND 96% 96%

S o u r c e Mouse Heart Tissue

G e n e GAPDH a-MHC P-MHC MyoD Myogenin

. Id e n tity 99% 99% 99% 98% 98%

S o u r c e Mouse Skeletal Muscle Tissue

G e n e GAPDH a-MHC P-MHC MyoD Myogenin

id e n tity 99% 99% 98% 98% 98%

ND: No primer products detected by agarose gel electrophoresis.

3.6 Investigating the effect of FBS, passage number and adrenaline on the ‘^beating” phenotype of P19CL6 cells

3.6.1 Rationale

The effect of culture media on heart cell growth and proliferation has been reported previously (Nag et al., 1985; Habara-Ohkubo, 1996; van der Heyden and Defize, 2003; Paquin et al., 2002). There was therefore evidence to suggest that FBS could influence cellular differentiation and gene expression in P19CL6 cells. In addition, the growth rate, gene expression, and function of some cells is passage-dependent (Green et al., 2001; Arkell and Jackson, 2003; Gao et al., 2002; Yu et al., 1997) and therefore passage number

96 could affect the phenotype of the P19CL6 cells. Furthermore, vitamins and hormones may induee cell differentiation in cardiomyocytes and in particular, the effect of adrenaline on P19CL6 cells has been established (Habara-Ohkubo, 1996; Nag et al., 1985; Pennec et al., 2002). Therefore, the effect of FBS, passage number and adrenaline on the “beating” characteristic of PI 9CL6 eells was investigated.

3.6.2 Results

Beating was observed mainly in localised areas with a high eell density (“islets”) independent of FBS type but only in passages 12-16 (Table 3.2). Beating was first observed on day-12, and pulse rates (beats/min) were measured on day-15. Pulse rates ranging from 17-41 beats per minute were recorded (Table 3.2). Statistical analysis showed that the “beating” phenotype was independent of the FBS used and was likely to be dependent on passage number (Table 3.2 and Fig. 3.11).

Table 3.2: Comparison of pulse rate observed in P19CL6 cells cultured with different FBS samples.

Pulse Rate (Mean ± SD) Media type P12 P13 P14 P15 P16 P17-22

DM1 30.6+6.6 (n=3) 33 (n=1) 26, 29 (n=2) 27+5.3 (n=3) 23+2.6 (n=3) ND

DM2 29,41 (n=2) 30.3+1.5 (n=3) 26±5.1 (n=4) 24 (n=1) 23, 21 (n=2) ND

DM3 30, 32 (n=2) 28.3±4 (n=3) 37, 23 (n=2) 23±2.6 (n=3) 1 8 ,18(n=2) ND

DM4 33.7+3.5 (n=4) 27, 29 (n=2) 23 (n=1) 25.3+4 (n=3) 23.315.6 (n=3) ND

FBS from four different batches (DM1-DM4) was used to prepare differentiation media. Pulse rate was recorded as beats per min. Data are shown as Mean ± Standard Deviation. PI2: passage number; ND: not detected;n: number of localised beating areas in the wells.

97 M ed ia

Figure 3.11: Determination of the effect of different FBS batches on the “beating” characteristic in P19CL6 cells cultured under adherent conditions. P19CL6 cells from different passage numbers (12-16) were cultured in the presence of 1% DMSO under adherent conditions using different FBS samples DM1-DM4 referring to media made up with four different FBS samples. Pulse rate (beats/min) was recorded on day-15 of incubation. Two-Way between-groups ANOVA analysis showed no significant difference between the FBS samples used.

Beating was observed only in passages 12-16 (Table 3.2). The lowest pulse rate (17 beats per min) was observed in passage 16, while passage 12 displayed the highest pulse rate (41 beats per min) and beating was never observed in passage 17 or above. Statistical analysis confirmed that the passage number of the P19CL6 cells had a significant effect on the “beating” phenotype (Fig. 3.12). Regression analysis showed an inverse correlation of pulse rate with passage number and a significant reduction of pulse rate with increasing passage number in P19CL6 cells (P<0.001) (Fig. 3.12). In addition, a Two-Way ANOVA revealed significant differences (P<0.01) between the passages examined. However, there was no significant effect of FBS on passage number, i.e. there was no significant effect of FBS on pulse rate in P19CL6 cells from different passage numbers.

98 Q. 20

12 13 14 15 16

Passage Number

Figure 3.12: Determination of the effect of passage number on “beating” characteristic in P19CL6 cells cultured under adherent conditions. The data shown in Table 3.2 has been plotted. A linear regression analysis of correlation between pulse rate and passage number showed a significant effect of passages number on pulse rate (y= -2.58x +63.3, R^=0.486, P<0.001). Two-Way ANOVA between-groups analysis confirmed significance difference between the different passage numbers of P19CL6 cells (P<0.01). a, b, and c represent passages with significant difference in the pulse rate of P19CL6 cells.

Beating was observed in P19CL6 cells cultured under adherent conditions with adrenaline concentrations ranging from 0-20 pM and pulse rates (pulses/min) were measured on day-15 (Table 3.3). Synchronous beating was observed on day 15 in a large culture area in the cells cultured with 20 pM adrenaline. Statistical analysis showed that pulse rate in P19CL6 cells correlated with the adrenaline concentration, and increased significantly (P<0.001) with increasing concentration (Fig. 3.13). In addition, One-Way ANOVA between-group analysis confirmed significant differences in the pulse rate of P19CL6 cells cultured with different adrenaline concentrations (Fig. 3.13).

Table 3.3: Effect of adrenaline on the beating phenotype of P19CL6 cells

Pulse Rate (Mean ± SD)

0 pM 5 pM 10 pM 15 pM 20 p M t

23+ 7 (a7=9) 26+5.9 (a7=9) 31+8 (A7=10) 40+ 8.7 (a7=8) 44+4.6 (n=6)

P19CL6 cells from passage 12 were plated in 6-well plates in three biologieal repeats. Pulse rate was recorded as beats/min and data are shown as Mean ± Standard Deviation, n: number of localised beating areas in the wells; f: synchronous beating (cells from a large culture area were beating together).

99 w 40

S 30

« 2 0 -

Adrenaline concentration (^M)

Figure 3.13: Determining the effect of different concentrations of adrenaline on “beating” in P19CL6 cells cultured under adherent conditions. P19CL6 cells (passage 12) were cultured with 0, 5, 10, 15, and 20 pM adrenaline under adherent conditions. Pulse rate (beats/min) was recorded on day-15 of incubation. Linear regression analysis showed a correlation of pulse rate with adrenaline concentration (y=l.lx+21.8, R^=0.541, P<0.001). One-Way ANOVA analysis confirmed a significant effect of adrenaline on pulse rate in P19CL6 cells (P<0.001). a, b, and c represent groups (adrenaline concentrations) with significant difference in the pulse rate of P19CL6 cells.

3.7 Microscopic examination of P19CL6 cells under adherent culture conditions for nuclei staining

3.7.1 Rationale

Cardiac and skeletal muscle cells display characteristic cell morphologies which can be assessed by light and/or fluorescence microscopy (Skerjanc, 1999; Rudnicki et ak, 1990; Wilcox, 1993). P19-derived cardiomyocytes display a mononuclear morphology, whereas P19-derived skeletal muscle cells fuse to form multinucleated myotubes with time in culture (Skeijanc, 1999; van der Heyden et ak, 2003). As the outcome of previous PCR-based experiments to measure gene expression levels were equivocal in defining P19CL6 cells as cardiomyocytes, it was thought that microscopy may help to resolve their identity.

100 3.7.2 Results

P19CL6 cells, cultured on coverslips under both adherent and non-adherent conditions, showed a mononuclear cell morphology with no cell fusion throughout 15 days of incubation as observed by phase-contrast microscopy (Fig. 3.14) (Kimes and Brandt, 1976; Skerjanc, 1999). The mononuclear nature of the P19CL6 cells was confirmed by fluorescence microscopy, using a nuclear stain (Fig. 3.15).

0

0 Figure 3.14: Phase-contrast microscopy of P19CL6 and H9C2(2-1) ceils. PI9CL6 cells were cultured on petri dishes in the presence of 1% DMSO. After four days, the aggregated cells were cultured on coverslips, which were collected on different time points. Cells were fixed using para-formaldehyde and light micrographs were taken using 4Qx magnification (A) on day 6, and (B) day 15 of the incubation period. Single cells are shown in enlarged boxes to show cardiomyocyte morphology. H9C2(2-1) cells were cultured in 37 mm diameter culture plates for 15 days under adherent conditions. Phase contrast micrographs were taken using 4Qx magnification (C) on day-6 before cell fusion and (D) day-15, when cells displayed multinucleated myotubes.

101 Conversely, under phase-contrast microscopy, H9C2(2-1) cells cultured under adherent conditions propagated as myoblasts and, with time, began to form tubular structures (Fig. 3.14). However, the H9C2(2-1) cells cultured on coverslips did not form fused tubular structures and therefore the expected multinucleated morphology of these cells, a characteristic of skeletal muscle cells, could not be confirmed by fluorescence microscopy.

Figure 3.15: Nuclear staining of P19CL6 cells cultured under adherent conditions.P19CL6 cells were cultured on eoverslips under adherent conditions in the presence of %1 DMSO. Coverslips were collected on different time points and fixed using para-formaldehyde. Coverslips were then immersed in Hoechst dye to stain nuclei. Coverslips were washed once with PBS and mounted on slides. Mieroscopic examination was performed at 400x magnification using a Zeiss Axiovert 135 TV fluorescent microscope (Carl Zeiss AG, Oberkochen, Germany) setting the excitation and emission filters at 365 nm and 420 nm, respectively. Slides were photographed using a Contax SLR-NX camera (Kyocera Yashica Ltd., Reading, UK). All P19CL6 cells showed a mononuclear morphology, as shown on (A) day-6 and (B ) day-15 of the incubation.

102 3.8 Discussion

The P19CL6 cell line, a derivative of P19 embryonal carcinoma cells, is widely used as an in vitro model of cardiovascular cells (Young et al., 2004; van der Heyden and Defize, 2003; van der Heyden et al., 2003; Peng et al., 2002; Anisimov et al., 2002; Paquin et al., 2002; Ridgeway et al., 2000; Monzen et al., 2001) that can differentiate into beating cardiomyocytes on exposure to DMSO (Habara-Ohkubo, 1996). During the course of various incubations of P19CL6 cells it was noted that the '"beating" phenotype was inconsistent between cultures. As one of the principal aims of this thesis was to investigate the effects of fatty acids and fibrates in cardiomyocytes, it was imperative that the cell-type being used was both correctly characterised and cultured under the correct conditions. Thus, the aim of this chapter was to characterise the P19CL6 cells in greater depth.

A review of the literature, focusing on P19CL6 and P19 cell culture conditions, showed that two separate methods are commonly used (Habara-Ohkubo, 1996; Morley and Whitfield, 1993; van der Heyden et al., 2003; van der Heyden and Defize, 2003; Skeijanc, 1999). The original paper that described the isolation of P19CL6 cells and characterised them as cardiomyocytes used adherent culture conditions (Habara-Ohkubo, 1996). However, a substantial number of papers, especially those relating to P19 cells (van der Heyden et al., 2003; van der Heyden and Defize, 2003; Skeijanc, 1999; McBumey, 1993), used non-adherent cell culture conditions {Le. growth on agar plates prior to re-plating under adherent conditions) (Smith et al., 1987; van der Heyden et al., 2003; Skeijanc, 1999). In order to understand and define conditions under which the P19CL6 cells display a cardiomyocyte phenotype, different culture conditions were investigated. The expression of cardiac- and skeletal muscle-specific genes was investigated under both adherent and non-adherent culture conditions in order to define specific conditions for the culture of P19CL6 cardiomyocytes. Initially, the H9C2(2-1) skeletal muscle cell-line was used as a positive control for skeletal muscle phenotype (Kimes and Brandt, 1976), and later, mouse heart and skeletal muscle tissue were also examined.

Under adherent cell culture conditions, the detection of RT-PCR products showed that P19CL6 cells expressed the cardiac-specific a-MHC gene (but not p-MHC), whereas H9C2(2-1) cells did not expressed the skeletal muscle-specific MyoD and Myogenin

103 genes (Fig. 3.1). Since the H9C2(2-1) cell-line has been previously characterised as a skeletal muscle cell-line (Kimes and Brandt, 1976) and therefore would be expected to express the skeletal muscle-specific genes, MyoD and Myogenin, it was likely that the mRNA levels were too low to be detected by RT-PCR, and therefore a second round of PCR amplification (nested PGR) was performed. The detection of nested PCR products showed that P19CL6 cells expressed a-MHC and P-MHC as well as MyoD and Myogenin, whereas H9C2(2-1) cells displayed expression of only MyoD and Myogenin under the same conditions (Fig. 3.2). Expression of the cardiac-specific a-MHC and p-MHC mRNA in the P19CL6 cells was expected as these cells have been previously characterised as cardiomyocytes (Habara-Ohkubo, 1996). Similarly, the expression of MyoD and Myogenin in H9C2(2-1) cells was also expected and this, to some extent, confirms the skeletal muscle phenotype of these cells (Kimes and Brandt, 1976; Edmondson and Olson, 1989; Sabourin and Rudnicki, 2000; Yun and Wold, 1996; Fujisawa-Sehara et al., 1990). However, in addition to the expression of a-MHC and p-MHC, expression of MyoD and Myogenin was also detected in P19CL6 cells (Fig. 3.2), which initially suggested that these cells may have a heterogeneous phenotype.

As the P19CL6 cells cultured under adherent conditions did not display the cardiac gene expression pattern that was expected, it was thought that aggregation might be required for the differentiation of P19CL6 cells into cardiomyocytes. Aggregation of P19 cells is apparently required for their differentiation into cardiac, skeletal muscle, or neural cells (Armour et al., 1999; McBumey, 1993; Skeijanc, 1999; Skeijanc et al., 1994; Smith et al., 1987; van der Heyden and Defize, 2003; van der Heyden et al., 2003), and these cells do not differentiate efficiently if cultured on adherent solid glass or plastic surfaces (McBumey, 1993; Smith et al., 1987). Under non-adherent culture conditions, consistent expression of cardiac- (a-MHC and P-MHC) and skeletal muscle-specific genes (MyoD and Myogenin) was observed under different culture conditions after 15 days in the presence or absence of DMSO (Fig. 3.3). Since a mixed phenotype was observed under all conditions, the time-dependent experiment of these genes was investigated. The purpose of this experiment was to determine whether the P19CL6 cells express only a-MHC and P-MHC, and not MyoD and Myogenin, at the same time point during 15 days of incubation.

104 P19CL6 cells consistently expressed a-MHC and p-MHC as well as MyoD and Myogenin throughout the incubation period, increasing up to day-9 and then decreasing. Moreover, on day-3 and day-9 the lowest and the highest relative expression levels were observed with all genes. In addition, the relative expression of a-MHC and P-MHC was higher than that of MyoD and Myogenin at all time points. Though these data could not be statistically validated (n=l), they indicate that the P19CL6 cells do not express only the cardiac-specific genes, i.e. a-MHC and P-MHC, at any time point of the incubation.

The effect of different culture conditions (adherent and non-adherent), incubation period (time-course) and presence or absence of DMSO showed that the P19CL6 cells may display a mixed phenotype. Therefore, other characteristics were investigated in order to validate the cardiac phenotype of the P19CL6 cells. The micronutrient levels in serum used to make the cell culture media could affect cell growth and differentiation (van der Heyden and Defize, 2003; Skeijanc, 1999; Takahashi et al., 2003; Nag et al., 1985; Wilton and Skeijanc, 1999), therefore the effect of FBS on “beating" characteristic of P19CL6 cells was examined. There was no significant difference in pulse rates with media prepared with four different FBS samples (Table 3.2 and Fig. 3.11). It is therefore reasonable to conclude that FBS itself does not directly affect the “beating" characteristics of P19CL6 cells. However, passage number of the P19CL6 cells had a significant effect on the beating phenotype with an inverse correlation between passage number and pulse rate for passages 12 to 16 (Table 3.2 and Fig. 3.12). Above passage 16 no beating was observed in any cultures, suggesting that the P19CL6 cells lose their “beating" characteristic with increasing passage number and therefore the P19CL6 cells with the lower passage number are more likely to display a cardiac phenotype. However, Two-Way ANOVA showed no significant effect of FBS on passage number, i.e. there was no significant effect of FBS on pulse rate in P19CL6 cells fi-om different passage numbers.

Vitamins and hormones act as potent inducers of cell differentiation and beating in cardiomyocytes (Habara-Ohkubo, 1996; van der Heyden and Defize, 2003; Skeijanc, 1999; Wobus et al., 1994; Takahashi et al., 2003; Nag et al., 1985; Rodriguez et al., 1994; Pennec et al., 2002). In particular, adrenaline has been shown to affect beating of the P19CL6 cells (Habara-Ohkubo, 1996; Pennec et al., 2002; Wobus et al., 1994), and therefore the effect of adrenaline on the “beating" phenotype of PI9CL6 cells cultured

105 under adherent conditions was investigated. Beating was observed in the P19CL6 cells cultured with different concentrations of adrenaline, even in the absence of adrenaline, suggesting that P19CL6 cells may display “beating” characteristic of cardiomyocytes, since contraction in cardiomyocytes is triggered by changes in intracellular Ca^"^ concentration and does not require adrenaline (Katz and Lorell, 2000; Katz, 2001). In addition, a significant positive correlation between pulse rate and adrenaline concentrations was observed (Table 3.3 and Fig. 3.13) that can be explained by an increase in Ca^^ conductance on exposure to adrenaline (Pennec et al., 2002).

The detection of cardiac- and skeletal muscle-specific genes in the P19CL6 cells together with the loss of the “beating” characteristics with increasing passage number suggests a cell-type with a mixed cardiac and skeletal muscle phenotype. However, drawing this conclusion requires validation of the data in two ways: (i) using the positive and/or negative controls, and (ii) confirming the identity of PCR product DNA sequences to known gene sequences. In order to compare the P19CL6 cells phenotype with positive and/or negative control samples, mouse heart and skeletal muscle tissues were used as representatives of the cardiac and skeletal muscle phenotypes.

Consistent expression of a-MHC, p-MHC, MyoD and Myogenin was detected in heart as well as skeletal muscle tissue (Fig. 3.7). This result was not expected, as heart tissue should express only a-MHC and P-MHC (Gulick et al., 1991; Metzger et al., 1995), whereas skeletal muscle only expresses MyoD, and Myogenin (Sabourin and Rudnicki, 2000; Buckingham, 2001). Therefore, initially the expression of other cardiac-specific genes, such as atrial natriuretic factor (ANF) and myosin light chain-2 polypeptide (MLC2) was used to characterise the cardiac and skeletal muscle phenotypes, with cardiac (CaCh-h) and skeletal muscle (CaCh-m) isoforms of voltage-dependent L-type calcium channel being included later in the study (Fassler et al., 1996). The expression of ANF and MLC2 was detected in heart as well as skeletal muscle tissue but not in P19C16 cells (Fig. 3.8). In addition, mouse heart tissue and P19CL6 cells consistently expressed only the CaCh-h (cardiac isoform), whereas skeletal muscle tissue showed expression of both CaCh-h and CaCh-m isoforms under the same conditions (Fig. 3.8 and Fig. 3.9). These results suggested that the expression of ANF, MLC2 and CaCh-h is not exclusive to cardiomyocytes and can be detected in skeletal muscle tissue. This conclusion is in line with our previous observations from a-MHC, p-MHC, MyoD

106 and Myogenin gene expression that indicate both the cardiomyocytes and skeletal muscle myotubes may express cardiac- (a-MHC and P-MHC) and skeletal muscle-specific genes MyoD and Myogenin.

In order to confirm that PCR products were correct, they were extracted and sequenced. Comparison of GAPDH, a-MHC, P-MHC, MyoD, and Myogenin PCR product sequences to known gene sequences displayed 93%-100% identity, confirming that the DNA sequences of the PCR products were identical to those of the expected genes (Table 3.1). In other words, expression of supposedly cardiac- and skeletal muscle-specific genes was confirmed in both the heart and skeletal muscle tissues. This result may have implications for other work and needs to be rigorously investigated and confirmed, but this was not within the scope of this study. A bioinformatics approach utilising the Human Genome Bio-Information database (GeneCards, http://bioinfo.weizmann.ac.il/cards/index.shtml/) was used to explain the observed results. This database shows the expression pattern of a given gene in twelve different human tissues, based on microarray experiments (Shmueli et al., 2003; Yanai et al., 2004), and confirms that both heart and skeletal muscle tissue express a-MHC, p-MHC, MyoD and Myogenin (Fig. 3.16 and Fig. 3.17). The a-MHC and P-MHC mRNA expression level is detectable only in cardiomyocytes if a laboratory procedure such as RT-PCR is used, and therefore these genes are considered as cardiac-specific genes in the literature (Gulick et al., 1991; Metzger et al., 1995; Morkin, 2000). A similar explanation could also apply for MyoD and Myogenin as skeletal muscle-specific genes (Sabourin and Rudnicki, 2000; Buckingham, 2001; Fassler et al., 1996). However, more sensitive techniques, such as microarrays, showed that both heart and skeletal muscle tissue express a-MHC, P-MHC, MyoD and Myogenin {Human Genome Bio-Information database) (Fig. 3.16 and Fig. 3.17). Since the nested PCR technique used in this study was more sensitive than the technique used in the literature, i.e. RT-PCR, it is therefore reasonable to observe expression of MyoD and Myogenin in the heart tissue (and P19CL6 cells) or to detect a-MHC and p-MHC in the skeletal muscle cells (Fig. 3.3 and Fig. 3.7).

107 GeneNote - expression arrays 1 probe-set 10000 - a-M H C

1000 -

100-

10 -

1- 0 - a ■r BMR SPL TMS BRN SPC HRT MSL LVR PNC PST KDN LNG Immune Nervous Muscle Secretary Other

GeneNote - expression arrays 1 probe-set 10000 P-MHC

1000

BMR SPL TMS BRN SPC HRT MSL LVR PNC PST KDN LNG Immune Nervous Muscle Secretary Other

Figure 3.16: Tissue-specific gene expression patterns obtained from the GeneCards database information for the expression of (A) a-MHC and (B) P-MHC in different human tissues. The expression of a given gene in twelve different human tissues is shown based on duplicate microarray analysis. BMR: born marrow; SPL: spleen; TMS: thymus; BRN: brain; SPC: spinal cord; HRT: heart; MSL: skeletal muscle; LVR: liver; PNC: pancreas; PST: prostate; KDN: kidney; LNG: lung.

108 GeneNote - expression arays 1 probe-set MyoD

100-

10 - f f f f BMR SPL TMS BRN SPC HRT MSL LVR PNC PST KDN LNG Immune Nervous Muscle Secretary Other

GeneNote - expression arrays 1 probe-sert Myogenin B

100 -

10-

T BMR SPL TMS BRN SPC HRT MSL LVR PNC PST KDN LNG Immune Nervous Muscle Secretary Other

Figure 3.17: Tissue-specific gene expression patterns obtained from the GeneCards database information for the expression of (A) MyoD and (B) Myogenin in different human tissues. The expression of a given gene in twelve different human tissues is shown based on duplieate microarray analysis. BMR: born marrow; SPL: spleen; TMS: thymus; BRN: brain; SPC: spinal cord; HRT: heart; MSL: skeletal muscle; LVR: liver; PNC: pancreas; PST: prostate; KDN: kidney; LNG: lung.

Since the PCR-based experiments to investigate tissue-specific gene expression profiles in P19CL6 cells were inconclusive and did not clearly confirm the cardiac or skeletal muscle phenotype of P19CL6 cell-line, it was thought that the examination of cell morphology by light and fluorescence microscopy may provide some answers in this respect. A mononuclear cell morphology with no cell fusion throughout the incubation period was observed by phase-contrast microscopy in P19CL6 cells (Fig. 3.14) that was later confirmed by fluorescence microscopy, using a nuclear stain (Fig. 3.15). In contrast, H9C2(2-1) myoblasts cultured under adherent conditions, fused with time and formed tubular multinucleated structures, the characteristic morphology of skeletal muscle cells

109 (Fig. 3.14) (Skeijanc, 1999; van der Heyden et al., 2003). These microscopy observations showed that the P19CL6 cells display a mononuclear morphology, as described for P19-derived cardiomyocytes (Skeijanc, 1999; Rudnicki et al., 1990; Wilcox, 1993), and therefore confirmed, to some extent, a cardiac phenotype of P19CL6 cells.

In conclusion, the techniques used here have not provided a “specific” answer in terms of whether P19CL6 cells show a cardiac, skeletal muscle or mixed phenotypes, and with respect to the tissue-specific gene expression profiles and microscopy observations of P19CL6 cells, it cannot be concluded that the P19CL6 cell-line is a pure cardiac cell-line. Therefore, as a next step, it was thought to compare global transcriptome levels in P19CL6 and H9C2(2-1) cells, as well as mouse cardiac and skeletal muscle tissues using microarray analysis to characterise the P19CL6 cells in greater depth.

110 Chapter 4

Characterisation of the P19CL6 cell-line by microarray analysis

111 Chapter 4:

Characterisation of the P19CL6 cell-line by microarray analysis

4.1 Introduction

The P19CL6 cell-line, a sub-clone of P19 mouse embryonal carcinoma cells (Habara-Ohkubo, 1996), has been proposed to spontaneously contract after several days in culture with DMSO and display a pure cardiac muscle phenotype (Habara- Ohkubo, 1996; Hiroi et al., 2001; Hosoda et al., 2001; Young et al., 2004), but in the present study, spontaneous contraction was not consistently observed. Thus, the cardiac phenotype of P19CL6 cells was further investigated by determination of the expression level of cardiac- (a-MHC and p-MHC) and skeletal muscle-specific genes MyoD and Myogenin by PCR-based experiments, but the consistent expression of both cardiac- and skeletal muscle-specific genes was detected in the P19CL6 cells (Fig. 3.2; Fig. 3.3; Fig. 3.4). In addition, a significant reduction in pulse rate was observed in P19CL6 cells with increasing passage number, suggesting that the P19CL6 cells lose their “beating” characteristic with increasing passage number (Fig. 3.12 and Table 3.2). Collectively, these observations did not clearly define whether the P19CL6 cells show a cardiac, skeletal muscle or heterogeneous phenotype. Moreover, the literature on the differentiation of P19 mouse embryonal cells into cardiomyocytes showed that the P19-derived cardiomyocytes are embryonic in nature (Skeijanc et al., 1998; van der Heyden and Defize, 2003; Rudnicki et al., 1990; Skeijanc, 1999), and therefore, P19CL6 cell-line could potentially display an embryonic phenotype as well. The aim of this chapter was to use cDNA microarray analysis as a tool to facilitate the further characterisation of the P19CL6 cells by investigation of global transcriptome levels with specific objectives and rationales as follows:

Objective 1: To establish whether the P19CL6 cell-line display global transcriptome levels similar to those of the H9C2(2-1) cell-line or mouse heart and skeletal muscle tissue.

12 cDNA microarray can potentially be used to measure the expression level of thousands of genes and provides a global analysis of gene expression at the level of transcription (Goldsmith and Dhanasekaran, 2004; Simon et al., 2002). Phenotypic differences among cells of different types are determined by changes in the global expression profiles or “transcriptome” (Napoli et al., 2003; Goldsmith and Dhanasekaran, 2004; Simon et al., 2002; Henriksen and Kotelevtsev, 2002), which provide evidence for the characterisation of different cell types (Goldsmith and Dhanasekaran, 2004; Clarke et al., 2004; Kawasaki, 2004). Therefore, a microarray analysis was performed to compare the global transcriptome in the P19CL6 cell-line with those of the H9C2(2-1) cells, characterised previously as skeletal muscle cells (Kimes and Brandt, 1976), and mouse cardiac and skeletal muscle tissues.

Objective 2: To compare gene expression profiles in the P19CL6 cell-line with those of mouse embryonic and adult heart tissues.

There is a wealth of evidence to confirm that embryonic and adult cardiomyocytes display different metabolic pathways (Lehman and Kelly, 2002; Makinde et al., 1998; Stanley et al., 2005; Lavrentyev et al., 2004), e.g. embryonic heart relies on glucose as a major source of energy whereas adult heart predominantly oxidise fatty acids (Lehman and Kelly, 2002; Makinde et al., 1998; Stanley et al., 2005; Lavrentyev et al., 2004) embryonic and adult heart therefore display different transcriptome profiles. Since the P19-derived cardiomyocytes are embryonic in nature (Skeijanc et al., 1998; van der Heyden and Defize, 2003; Rudnicki et al., 1990; Skeijanc, 1999), it was of interest to establish whether P19CL6 cells display a transcriptome similar to that of mouse embryonic or adult heart tissue.

113 4.2 Validation of microarray procedures

4.2.1 Rationale

To optimise microarray normalisation and filtering procedures, a self-hybridisation experiment was performed by hybridisation of RNA isolated from P19CL6 cells labelled separately with Cy3 and Cy5 to a microarray slide. Normalisation and filtering were first performed on the self-hybridisation experiment, as indicated in Materials and Methods, and was subsequently applied to the other microarray hybridisations (Yu et al., 2002). This process standardised the microarray data and enabled comparison of relative gene expression levels (Wit and McClure, 2004; Fang et al., 2003).

4.2.2 Results

Total-RNA isolated fi-om P19CL6 cells was divided into two aliquots, labelled separately with Cy3 and Cy5 and hybridised on a microarray slide. Images were captured and the data analysed. Normalisation and sequential filtering was then performed on the self-hybridisation array to eliminate spots of low quality. Normalisation also removed the background noise, systemic bias, and artefact signals, as described in Materials and Methods. Scatter plots were generated to compare gene expression patterns before and after the normalisation and filtering process (Fig. 4.1). Scatter plots showed a compact distribution with nearly all spots between the lines representing ±2-fold changes in gene expression level (Fig. 4.1). A linear regression analysis was performed and showed a significant correlation of Cy3 and Cy5 signals (slope=0.96, R^=0.98, P<0.001) indicating optimisation of normalisation and filtering processes.

14 Self hybridisation of P19CL6 Before normalisation & filtering . . v

“ P19CL6-Cy5

Self hybridisation of P19CL6 After normalisation & filtering

P I i n f, << P I 1C . ", :t 'iM'ftii 100 MO P10CL6-dy5 «

Figure 4.1: Scatter plots for the self-hybridisation of P19CL6 cell-line (A) before, and (B) after data normalisation and filtering process. Two RNA aliquots of P19CL6 were labelled with Cy3 and Cy5 dye, mixed and hybridised on a microarray. Data was nomialised and filtered and the scatter plots were generated to demonstrate the gene expression profiles of (A) before, and (B) after normalisation and filtering. Scatter plots showed the distribution of spots according to the gene expression levels with lines representing ±2-fold changes in gene expression levels.

115 4.3 Comparison of global transcriptome levels in the P19CL6 and H9C2(2-1) cell-lines, as well as mouse cardiac and skeletal muscle tissues using microarray analysis

4.3.1 Rationale

Since the results of the PCR-based experiments (Chapter-3) were inconclusive in terms of confirming a cardiac, skeletal muscle or heterogeneous phenotype of the P19CL6 cell-line, microarray analysis was used for further characterisation of P19CL6 cells by investigation of global transcriptome levels in P19CL6 and H9C2(2-1) cells, as well as mouse cardiac and skeletal muscle tissues.

4.3.2 Results

In order to reduce the number of arrays required to compare global transcriptome levels in the P19CL6 cell-line with those of the H9C2(2-1) cell-line and mouse cardiac and skeletal muscle tissues, a loop experiment was designed (Fig. 4.2). Microarray data was normalised and filtered, as described in Materials and Methods (Section 2.2.8.5). The end-point of noimalisatipn and filtering was 2162 spots that were considered to be acceptable. Comparison of transcriptome levels in the P19CL6 cell-line with those of the H9C2(2-1) skeletal muscle cells showed 327 genes over-expressed in P19CL6 cells whereas 382 genes were down-regulated >2-fold (Fig. 4.3A). However, these two cell lines displayed similar expression levels for 1453 genes. Comparison of global transcriptome levels in H9C2(2-1) cells with mouse skeletal muscle tissue showed à relatively compact distribution (Fig. 4.3B). Although the majority of genes (1886 genes) were detected between ±2-fold changes in gene expression levels, 100 genes were up-regulated in H9C2(2-1) cells, whereas 176 genes were down-regulated, i.e. over-expressed in muscle tissue.

116 B S1 S2 S3 84

P19CL6-Cy3 H9C2(2-1)-Cy3 Meuse Meuse Muscle-Cy3 Heart-Cy3

Mouse Meuse Müscle-Cy5 Heart-CyS P19CL6-Cy5

Figure 4.2; Experimental loop design to compare global transcriptome levels in P19CL6 and H9C2(2-1) cells, as well as mouse cardiac and skeletal muscle tissues. (A) loop design showing the four samples (nodes) to be compared. The sample at the start of the arrow was labelled with Cy3 and the “target” sample with Cy5. Samples numbered 1 to 4 are the P19CL6 and H9C2(2-1) cell-lines, mouse skeletal muscle and heart tissue, respectively. (B) samples and Cy-dye labelling used on different microarray slides, based on the loop design. S1-S4: slide numbers.

The data obtained from the hybridisation of mouse skeletal muscle and heart tissues indicated that the expression levels of relatively a few genes changed >2-fold in these tissues. Skeletal muscle tissue displayed up-regulation of 105 genes, whereas 98 genes were over-expressed in the heart (Fig. 4.3C). Data also showed 1959 genes with relatively close Cy3:Cy5 signal ratio lined between ±2-fold changes in gene expression lines. Hybridisation of mouse heart tissue versus the P19CL6 cell-line on the other hand, showed a wide distribution of gene expression levels with 332 genes over-expressed in the heart tissue and 320 genes in the P19CL6 cell-line (Fig. 4.3D).

17 ÛÛ IIIc % I - o o> 1 u o OR § ■

T3 CL C3g

S "o' Æ |« . £ c Ï

■■. O X

% ^ 'k : ■■ • • iik \ o

• • ^ 5 ^ 6 ^ I"O i - s - s i 2 CD % m : i |§ o . - a !& eAo-apsnuu- ■.•i.ii.- vtaM""3Psn>'j I i II II 3la) o 1 |

D£ 3 E a Ii After normalisation of microarray data, linear modelling was performed (MARIEiExGen; http://exgen.ma.umist.ac.uk/) in order to convert data fi-om the loop experiment into a linear model (Wit and McClure, 2004; Dalgaard, 2002; Smyth, 2004). The data produced by linear modelling was imported into a web-based clustering programme (http://rana.lbl.gov/) (Eisen et al., 1998) to generate a pair-wise linkage cluster. Analysis formed a hierarchical cluster based on the phylogenetic algorithm, which was visualised with TreeView software (http://rana.lbl.gov/) (Eisen et al., 1998). Complete-linkage cluster analysis showed that the H9C2(2-1) cell-line and heart tissue were closely related (Fig. 4.4A). In addition, the H9C2(2-1) cell-line was more closely related to the heart than skeletal muscle tissue (Fig. 4.4A). Cluster analysis also showed that the H9C2(2-1) cell-line and heart tissue were more closely related than the P19CL6 cell-line and heart tissue (Fig. 4.4B). Furthermore, mouse heart and skeletal muscle tissues were more closely related to each other than to the P19C16 cell-line (Fig. 4.4C).

Similar results were obtained using principal components analysis (PCA), confirming that the H9C2(2-1) cell-line is more closely related to the mouse heart tissue than the P19CL6 cell-line (Fig. 4.5). As indicated in Table 4.1, the correlation coefficient between H9C2(2-1) and heart (0.73) was greater than that between P19CL6 and heart (0.41). Collectively, the degree of correlation obtained for the P19CL6 and H9C2(2-1) cell-lines, and mouse tissues (Table 4.2), indicated a greater correlation between the H9C2(2-1) cell-line and heart tissue as compared to the P19CL6 cell-line.

119 1 / i o CD (O ■C CVJ 1 o _ l _J t r CO CO O 0) CM Ü o œ CD CD O) CD i X X CO G) ; ^ w q I q I ■ Sfl CO CO Q) CD _J _J Ü ■ c Ü Ü Ü o W CO Ü CO CD CD (O 0) X K CL CL

Figure 4.4: Hierarchical cluster analysis of mRNA expression levels in microarrays comparing the P19CL6 and H9C2(2-1) cell-lines, as well as mouse heart and skeletal muscle tissue. The comparisons were clustered according to the degree of pair-wise similarity. The colour gradient represents green for down-regulation and red for up-regulation. Each gene is represented by a single row and comparisons are represented by columns. (A), (B) and (C) demonstrate cluster analysis results for different sets of comparisons. Blue, red and yellow circles show the close relationship between comparisons, as discussed in the text. H9C2: H9C2(2-1) cell-line.

120 1.0 H9C2(2-1) vs muscle > Heart vs muscle

CM 0 .5 - H9C2(2-1 ) vs heart H9C2(2-1 ) vs PI 9 C I^ \ oc c Heart vs P19CI6 â P19CI6 vs muscle E o 0.0

Q. Muscle vs P19CI6< 19CI6 vs heart P19CI6 vsH9C2(2-1) Heart vsH9C2(2-1) -0.5 -

Muscle vs heart Muscle vs H9C2(2-1) - 1.0

- 1.0 -0 .5 0.0 0.5 1.0 Principal component 1 Figure 4.5: Principal component analysis (PCA) of global transcriptome levels in P19CL6 and H9C2(2-1) cells, and mouse heart and skeletal muscle tissue. Circles indicate a close relationship between samples, as discussed in the text. PCI and PC2 represent 70% and 21% of total variance, respectively.

12 '2 ■C (S G (0 u 3 0) s 1 i n c o 00 00 o ir> CO 00 00 o • g 2 o c o CM m o o c o CM lO o > 1 N - CM p 1 N CM p U » o d d d ■ d d d d 0 \ u 1 3<0 P h 'S S

- w (N r (0 -O . s 0) ç co c c X

" g S c o m o 00 00 lO I f o 00 c o o h - o i n c o o N o m § CM Tl" p CM CM CM p CM ü - f N 1 1 o d d d d d O- ' d 3 SL 1 O o • î 2 05ü E i X È u CM a ) X 0 C « U co

< s s 1 « m h . o -S" 00 m N o 00 s o o o o N CM 00 O o h - CM c B -S" 1 p N -S- 1 p M . o d d o d d d o d & 0) * © O) CL CM i ü ? s 05 U c X P h t \ - x "C JO t : CL, CA • PN W5 ÈZ 0) % l_ X .2 o o o o CM CM o o O p 3 o CM o 00 c o o CM o 00 c o p 00 CD CM 1 p 00 CD CM ' 1 .5 g d d d d d d d d • f h ^

Sample Mouse Heart Mouse Skeletal Muscle

P19CL6 cell-line 0.405 0.133

H9C2(2-1) cell-line 0.728 0.258

The data obtained from linear modelling was used to perform principal component analysis. The data is a summary of information indicated in Table 4.1.

4.4 Microarray analysis of global transcriptome levels in the P19CL6 cell-line, mouse embryo and heart tissue

4.4.1 Rationale

The investigation of global transcriptome levels in P19CL6 cells showed that the P19CL6 cell-line did not display a close correlation with the adult heart tissue (Section 4.3). However, as P19-derived cardiomyocytes have been characterised as embryonic cells (Skeijanc et al., 1998; van der Heyden and Defize, 2003; Rudnicki et al., 1990; Skeijanc, 1999), it was of interest to establish whether the P19CL6 cell-line, a sub-clone of PI 9 cells (Habara-Ohkubo, 1996), was similar to mouse embryo tissue, as compared to adult heart tissue.

4.4.2 Results

A loop experiment was performed to compare global transcriptome levels in the P19CL6 cell-line, mouse embryo and heart tissues. Total-RNA isolated from the samples was labelled according to the experimental plan (Fig. 4.6) and hybridised to microarrays. Microarray data was normalised and filtered, as discussed in Materials and Methods (Section 2.2.8.5). The end-point of normalisation and filtering was 4121 spots that were considered as acceptable. Scatter plots were generated to demonstrate the distribution of gene expression levels based on the Cy3:Cy5 signal ratio (Fig. 4.7).

12: S2 S3

P19CL6-Cy3 Mouse Mouse Embryo-Cy3 Heart-Cy3

Mouse Mouse Embryo-Cy5 Heart-Cy5 P19CL6-Cy5

Figure 4.6: Experimental loop design to compare global transcriptome levels in P19CL6, as well as mouse embryo and heart tissues. (A) loop design showing the three samples (nodes) to be eompared. The sample at the start of the arrow was labelled with Cy3 and the “target” sample with Cy5. Samples numbered 1 to 3 are the P19CL6 cell-line, mouse embryo and heart tissue, respectively. (B) samples and Cy-dye labelling used on different mieroarray slides, based on the loop design. SI-S3: slide numbers.

Comparison of global transcriptome in the P19CL6 cell-line with mouse embryo showed that the majority of genes (3263 genes) were not differentially expressed. While 351 genes were found to be over-expressed (>2-fold) in the P19CL6 cell-line, 507 genes were down-regulated, i.e. over-expressed in embryo (Fig. 4.7A). Comparison of global transcriptome levels in mouse embryo with heart confirmed that 335 genes were up-regulated in the embryo whereas heart over-expressed 386 genes (Fig. 4.7B). However, the expression levels of the remaining 3400 genes were not significantly different in these tissues. A wide distribution of gene expression levels was observed in the comparison of mouse heart with the P19CL6 cell-line, with over-expression of 642 genes in the heart and 690 genes in the P19CL6 cell-line (Fig. 4.7C).

The data from this loop experiment was converted to a linear model using LLAMA linear modelling programme (Wit and McClure, 2004; Dalgaard, 2002; Smyth, 2004). Complete-linkage cluster analysis was then performed to demonstrate the degree of pair-wise similarity in gene expression levels between the samples (Eisen et al., 1998). Cluster analysis confirmed that the correlation between mouse embryo and heart was greater than that between the P19CL6 cell-line and heart (Fig. 4.8). The relationship observed between the embryo and heart was also closer than the relationship between embryo and the P19CL6 cell-line. In addition, the P19CL6 cell-line was more closely related to the embryo than heart tissue (Fig. 4.8). Similar results were obtained using

24 PCA, confirming that mouse embryo and heart tissues are more closely related than the P19CL6 cell-line and heart (Fig. 4.9). As summarised in Table 4.3, the correlation coefficient between the P19CL6 cell-line and embryo (0.67) was greater than that between the P19CL6 cell-line and heart (0.05).

Embry

Heart vs P19CL6 Figure 4.7: Scatter plots of samples: (A) P19CL6 cell-line vs mouse embryo, (B) mouse embryo vs heart tissue, and (C) heart tissue vs P19CL6 cell-line. Green lines indicate equivalent expressions (middle line) and 2 fold up- or down-regulation. A colour gradient shows red spots indicating genes that are up-regulated in the Cy3-labelled samples (Y-axis) while green spots represent genes that are over-expressed in the Cy5-labelled sample (X-axis).

P19CL6-Cy5

125 _ l tl o o cts G) 3 (D JD o X E CL O) LU o O tl JDi CO E 0 E LU X LU

Figure 4.8: Hierarchical cluster analysis of mRNA expression levels in microarrays comparing the P19CL6 cell-line and mouse embryo and heart tissue. The comparisons were clustered according to the degree of pair-wise similarity. The colour gradient represents green for down-regulation and red for up-regulation. Each gene is represented by a single row and comparisons are represented by columns.

2 6 1D-

E m b rw vs P19CL6

O.S. Heart vs1R19CL6 o Em bryo vs H eart CM c o O 0 .Ü - a E o o 15 H eart vs Em bryo o c Q.

-10 -05 □ JO Q.S 1.0 Principal component 1

Figure 4.9: Principal component analysis (PCA) of global transcriptome levels in the P19CL6 cell-line, mouse embryo and heart tissue. Circle shows the close correlation between samples, as discussed in the text. Principal components 1 and 2 represent 66% and 34% of total variance, respectively.

Table 4.3: Correlation matrix obtained by principal component analysis (PCA) of mRNA expression levels in the P19CL6 cell-line and mouse embryo and heart tissue.

Correlation range Comparisons Heart vs P19CL6 Embryo vs Heart Embryo vs P19CL6

Heart vs P19CL6 1.000 0.666

Embryo vs Heart 0.666 1.000 0.046

Embryo vs P19CL6 0.046 1.000

P19CL6 vs Heart 1 000 0.666

Heart vs Embryo 0.666 1.000 0,046

P19CL6 vs Embryo 0.046 1.000 The data obtained from linear modelling was used to perform principal component analysis. Principal components 1 and 2 represent 66% and 34% of total variance, respectively.

127 4.5 Microarray analysis of global transeriptome levels in the P19CL6 and H9C2(2-1) celHines, mouse skeletal muscle and mouse embryonic and adult heart tissue

4.5.1 Rationale

Previous experiments (Sections 4.3 and 4.4) showed that the P19CL6 cell-line is not closely related to the mouse adult heart or skeletal muscle tissue. The correlation between the heart and embryo tissues was also closer than that of the heart and the P19CL6 cell-line. However, a comparison of global transeriptome levels in the P19CL6 cell-line with those of embryonic heart has not previously been made. Therefore, an experiment was designed to compare global transeriptome levels in the P19CL6 cell-line with those of mouse embryonic heart. The H9C2(2-1) cell-line as well as mouse adult heart and skeletal muscle tissues were also included in the experiment to validate the previous results.

4.5.2 Results

An interwoven loop experiment (Fig. 4.10) was performed in order to compare global transeriptome levels in the P19CL6 and H9C2(2-1) cell-lines, and skeletal muscle, embryonic heart and adult heart tissue (Simon et al., 2002; Dobbin and Simon, 2002). Total-RNA from P19CL6 (two different biological samples) and the other samples were hybridised to microarray slides (Table 4.4 and Fig. 4.10). In addition, a reference RNA sample was prepared by mixing equivalent amounts of all total-RNA samples (Simon et al., 2002; Dobbin and Simon, 2002). Microarray data was normalised and filtered, and scatter plots generated (Fig. 4.11; Fig. 4.12; Fig. 4.13).

128 Figure 4.10: Representation of dual-hybridisation optimal interwoven loop design for microarray experiment. Loop design showing the seven samples (nodes) to be compared. The sample at the start of the arrow was labelled with Cy3 and the “target” sample with Cy5. Nodes 1-7 indicate total-RNA sources as follows: 1: P19CL6 (1); 2: skeletal muscle; 3: embryonic heart; 4: P19CL6 (2); 5: Reference total-RNA; 6: H9C2(2-1); 7: adult heart. Reference RNA sample was prepared by mixing equivalent amounts of. all total-RNA samples (Simon et al., 2002; Dobbin and Simon, 2002).

Table 4.4: Samples and Cy-dye labelling used on different microarray slides, based on the loop design.

Array No. Cy3 CyS 1 P19CL6 cell-line (2) Adult heart

2 Reference total-RNA^ Adult heart 3 Embryonic heart H9C2(2-1) cell-line 4 Embryonic heart Reference total-RNA 5 P19CL6 cell-line (1) Skeletal muscle 6 Reference total-RNA P19CL6 cell-line (2) 7 Reference total-RNA Skeletal muscle 8 Skeletal muscle Adult heart 9 Adult heart H9C2(2-1) cell-line 10 P19CL6 cell-line (1) Reference total-RNA 11 P19CL6 cell-line (1) Embryonic heart 12 Skeletal muscle Embryonic heart 13 Adult heart P19CL6 cell-line (1) 14 P19CL6 cell-line (2) H9C2(2-1) cell-line 15 H9C2(2-1) cell-line P19CL6 cell-line (1) 16 P19CL6 cell-line (2) Skeletal muscle 17 P19CL6 cell-line (2) Embryonic heart

18 H9C2(2-1) cell-line Reference total-RNA Î The reference sample was prepared by mixing equivalent amounts of all total-RNA samples. The P19CL6 cell-lines 1 and 2 represent two different biological samples.

129 (D J = 1 I Ô-Lo II '%, m s - : II5 2 i ; i f wU ON U 0> rs î eAo-éouaja^ay es i l ^ Ln pë" "Q Ê U a i g « I .S bû % V?&. J- :' «1 'W. III - II il

■ § g g g i2 ^ 3 -5

■lî = S S î & u I eAo-aouaja^ay 9^0-9106 Id ON . a. §- 2^ dH W o S" (N cd.1 (50 %, f ! o r - ^ II û-.Si il o _a> H 0 S ^ - 8

eAo-iJBaL|-LU3 II ii # oT S .= S T3O ^

bûIf 1 1 .s X s i I f 8 a C/5 c/2 ■®-w % £Âo-a|osn|y\| 11(D 111 1 2 I ^ lA | 1 1'^ £ £ i f " © T 3 o s k. "3 o b û 1 "% 3 2^ 1 O X © O

i T . a < o s oi s 3 g _o i % % S 1 § & o 1 g ¥ 3 % S© 1 b û S O o ~a £Âo-a|osn|/\i 9^0-9136 W u 1c/5 o 2^ S s CL so CL P 3 U T3 OS -SO S" < P4 n 3 • o s % - • g i ’H. 1 _ © m .,z S g o & IS (K © •m; © — 'S T3 'E _ © 0 5 £ ? "a. > > “ c/5 © C y i S . 2 - 2 c d 'Ô 5 CL u c / 3 1 a 1 9- C/5 r i S ■8 u 3 — os 3 £Ao-aouaja^ay S > 1 w ’5 b ê S ' u III Q. I i 3 in O à l i 5 0: % <3 w J= f s ^ i (U I fi ÛÛ^ !î îî sII e li B 1 C /5 — €Ao -(V-3)306H 9/^3-(LZ)Z36H « II o X 5 u CL LU M o % 7 v ü <0 s © aT "'S > O — O Z "S o w 3 o so u 3 s o U i ■£ o s o (U u - o o s PL S bû P n M' 3 Sr 3 O >. i T ~ o S a % e 3 & 9^0-9136 W 9/^3-9136 W ©S? (N Xi S T3 2 a o s B 3 3 - 0> O 3 _ ç 1 % i S - 2 M > 1) M % QC o s "%;;. CL & Ï E !/) ge 3 t/5 u _g ' x c*-i s o ? O 2 > U Os X 1 0) ï CL Ph CL k £ 3 g ê *3 O T3 U ©

1

Figure 4.14: Representation of theoretical loops 1 and 2 derived from optimal interwoven loop design for microarray experiment. Loop designs showing the seven samples (nodes) to be compared. The sample at the start of the arrow was labelled with Cy3 and the “target” sample with Cy5. Nodes 1-7 indicate total-RNA sources as follows: 1: P19GL6 (1); 2: skeletal muscle; 3: embryonic heart; 4: P19CL6 (2); 5: Reference total-RNA; 6: H9C2(2-1); 7: adult heart. P19CL6 (1) and (2) represent two different biological samples. Reference RNA sample was prepared by mixing equivalent amounts of all total-RNA samples (Simon et al., 2002; Dobbin and Simon, 2002).

133 1.0 / P19CL6 vs H9C2(2-1), Muscle vs H9C2(2-1) 0.8 O I ° \ Em-heart vs H9C2(2-1 ) . 0.6 ° o U \ Ad-heart vs H9C2(2-1) Muscle vs Ad-heart ; \ P19CL6 vs Em-heart 0.4 ° O i O ! Muscle vs Em-head PI 9CL6 vs Ad-heart I c 0.2 Em-heart vs Ad-heart 0 O # u s c le vs P19CL6 o 0.0 \ o ______-— Q. P19CL6 vs Muscle O Ad-heart vs Em-heart o -0.2 ;Ad-heart vs P19CL6 : Em-heart vs Muscle -0.4 o Em-heart vs P19CL6 Ad-heart vs Muscle .. H9C2(2-1) vs Ad-heart I

- 0.6 ° o ...... H9C2(2-1) vs Em-heart

- 0.8 H9C2(2-1) vs Muscle I , H9C2(2-1) VSP19CL6

- 1.0

- 1.0 -0.5 0.0 0.5 1.0 Component 1

1.0 g P19CL6 vsH9C2(2-1)

Muscle vs H9C2(2-1) PI 9CL6 vs Em-heart . Ad-heart vs H9C2(2-1) o 0.5 ..9.P O Em-heart vs H9C2(2-1) . P19CL6 vs Ad-heart O Muscle vs Em-heart O CM P19CL6 vs Muscle. C ^ O Ad-heart vs Em-heart 0 C O Ad-heart vs Muscle o 0.0 - M uscle vs Ad-heart % Q. n 1 oE Em-heart vs Ad-heart O / Muscle vs PI 9CL6 ^ O Em-heart vs Muscle O O Ad-heart vs PI 9CL6 H9C2(2-1) vs Em-heart

-0.5 - o O o o H9C2(2-1) vs Muscle! ' Em-heart vs P19CL6 H9C2(2-1 ) vs Ad-heart /

H9C2(2-1) VSP19CL6 j

- 1.0

- 1.0 -0.8 -0.6 -0.4 - 0.2 0.0 0.2 0.4 0.6 0.8 1.0 Component 1 Figure 4.15: Principal components analysis (PCA) of global transeriptome levels in the P19CL6 and H9C2(2-1) cell-lines, mouse skeletal muscle and embryonic and adult heart tissue obtained from theoretical loops 1 and 2. The PGA-derived data (A) and (B) correspond to the loop 1 and 2, respectively. Goloured circles represent different clusters of comparisons to be determined. Ad-heart: adult heart; Em-heart: embryonic heart

3 4 ns «0 g S 5 r .2 CM CD CD t'­ o CM CD CD t'a o g « g ? CO Tf CD CM en CD CO at CD CM CO CD £ VO 1 1 CM 1 in N CO q 1 X 1 t^ ry 1 m ta. CO q CD CD CD CD CD O CD CD d d d U 1 = s o \ LU -Ç i PL W w 2 tr -c m 05 00 CD cn CD C'­ fa- cn 00 CD cn CD ta- 5 2 CM in in 00 CD en CM in in 00 at CD co 1 M- 1 CM CO 1 q CO 1 ta. 1 CM CO CO 1 q CO i .s ^ Ô CD CD CD o CD CD o CD CD d d E < «g o . lU 1 CO 2 § V

I ^_o> § .2 o N CD m O CD CD |a- o in at CD CD £ ,® ■*1^ CO Tf CD •M- CD CM CO at o at CD CM N 1 1 CO in N q 1 1^ CM 1 1 CO in N R 1 t'a O CD CD CD CD CD CD CD C) CD CD d II 1 i | 05 -a X

" 2 < 5 2 1 0) 2 'Cf 00 cn CD at cn CD at 00 cn CD a t cn CD S s w 0) CO TT CD CD a t at CD CO at CD CD at at CD > 1 1 CO CM 1 CO q a t CO in 1 1 CO CM 1 CO q CO in CD CD CD C3 CD CD CD CD CD CD d d S 6 § 1 1 ■ 1 < A (g 'O u o u .2 CU

S i I r 1 T 2 00 o CO CD cn in CD OO CD CO CD cn m CD Ph o 3 - CO CO CD CD 00 a t CO CO CD CD 00 0) 1 CO 1 in q CO CO 1 1 CO 1 in q CO CO 1 > O)

è e 2 r 73 « (o 5 m CM CD CM o a t cn m CM CD CM CD at cn > 3 - •M- CD CO CO in a t a t CD CO CO in 1 2 4 -1 £ CO Ü - 9 q œ 1 CO CO 1 t>^ 1 CO t^ q 5 1 CO CO 1 ta. 1 05 T3 C5 CD CD CD CD CD CD CD CD CD d d 1 • i 1 2 g z < S o " 3 k es S t .1 3 Ü CO CD CM CD N ta- CO O CM o ta­ ta. CO S u 3 | CO CD ■M- in in CM CO o at in in CM A » % Ü "? CO q N CO CM 1 1 1 CO q ta. CO CM 1 1 1 CO CD CD CD CD CD CD CD CD d a | i -o -W w u 03 es p M H-» E a> © s 1 ^ (O g - ^ V —I o CO in OO CD OO CD CD CO in 00 CD 00 CD o CO Tf CO in CO O CO at CO in at CO CD CO CO in in 1 CM 1 1 q CO CO in m 1 CM 1 1 “S i CD c> o CD CD CD CD CD CD CD d d l05 i X a 1 1 t t g t CD t CD CO r 3 _o t _l CD ■c o 1 s U s O 3 3 CL O V) s 1 u s CO CL Ü 3 3 3 <35 b: to •9 3 •9 05 ; Ü ■9 U> 3 Ü ■9 s 05 o 3 ■9 ■9 § o! 05 s s o E t ! i Ê S < < cn a a! LU 05 s LU < X < q! X • » /^S •Ë S X s è 3 S s s s s s s S s s s s i S s g *9 T CL t s s g IIu o T f

Comparisons Mouse embryonic heart Mouse adult heart

P19CL6 cell-line 0.732 0.248

H9C2(2-1) cell-line 0.720 0.444

Ad-heart: mouse adult heart; Em-heart: mouse embryonic heart. The data is a summary of information indicated in Table 4.5.

4.6 Discussion

The present study shows that the P19CL6 cell-line does not display a pure cardiac phenotype, according to the global transeriptome profiles, as analysed here with cDNA microarrays. Collectively, the results indicate that although the P19CL6 cell-line displayed some characteristics of cardiomyocytes and, to some extent, showed a correlation with mouse adult heart tissue, global transeriptome profiles of the P19CL6 cell-line were more closely related to the mouse embryonic heart tissue. In contrast, the H9C2(2-1) cell-line was more closely related to the adult heart tissue than the P19CL6 cell-line. Therefore, based on these observations it seems that the H9C2(2-1) cell-line is more suitable cardiomyocyte model system than the P19CL6 cell-line.

Analysis of data fi'om different experiments indicated that the mouse adult heart and skeletal muscle tissues are closely related (Fig. 4.3C and Fig. 4.4C) while the P19CL6 cell-line did not show a high degree of correlation with adult heart tissue (Fig. 4.3D). In addition, the correlation between the heart and skeletal muscle was greater than that between heart and the P19CL6 cell-line (Fig. 4.4C and Table 4.1). The correlation between the heart and skeletal muscle was also greater than that the P19CL6 cell-line and skeletal muscle tissue (Fig. 4.4C and Table 4.1). Therefore, the global transeriptome profiles observed in the P19CL6 cell-line were not indicative of a pure cardiac or skeletal muscle phenotype. Furthermore, the P19CL6 cell-line were more closely related to the mouse embryo tissue than heart tissue (Fig. 4.8 and Table 4.3). The global transeriptome in the P19CL6 cell-line also showed a greater correlation with mouse embryonic heart (Table 4.5 and Table 4.6). In addition, the correlation coefficient between the P19CL6

136 cell-line and embryonic heart (0.73) was greater than that between P19CL6 and adult heart (0.25). Therefore, based on the data obtained here the P19CL6 cell-line displays a phenotype closer to an embryonic heart rather than adult heart. This conclusion is in line with other studies indicating that the P19-derived cardiomyocytes are embryonic in nature (Skeijanc et al., 1998; van der Heyden and Defize, 2003; Rudnicki et al., 1990; Skeqanc, 1999). Therefore, the P19CL6 cell-line is not a pure adult heart model system, as it has been claimed (Habara-Ohkubo, 1996), and more likely displays an embryonic heart phenotype.

Analysis of microarray data showed that the H9C2(2-1) cell-line was closely related to the mouse skeletal muscle tissue (Fig. 4.3B). The mouse adult heart and skeletal muscle tissues were also closely related (Fig. 4.3C and Fig. 4.4C). Therefore, it can be concluded that the H9C2(2-1) cell-line is closely related to the mouse heart tissue (Fig. 4.3 and Fig. 4.4B). The correlation between the H9C2(2-1) cell-line and heart tissue was also greater than that between the H9C2(2-1) cell-line and skeletal muscle tissue (Fig. 4.4A; Fig. 4.5; Table 4.2). These observations were supported by other studies confirming that although the H9C2(2-1) cell-line was initially described to have skeletal muscle cell properties (Kimes and Brandt, 1976) subsequent investigations showed that these cells also display cardiomyocyte properties such as expression of the cardiomyocyte-specific subtype of calcium channel (Hescheler et al., 1991; van Nieuwenhoven et al., 1998). Consequently, the H9C2(2-1) cell-line is now being widely used as a cardiomyocyte model system (Mestril et al., 1994; Nyayapati et al., 1996; van der Putten et al., 2002; Kanaya et al., 2005; van Nieuwenhoven et al., 1998; Minakawa et al., 2003). Moreover, the H9C2(2-1) cell-line showed a correlation with embryonic heart tissue (Table 4.5 and Table 4.6). As indicated in the previous studies, the H9C2(2-1) cell-line is a rat embryonic cardiac-derived cell-line (Mestril et al., 1994; Kimes and Brandt, 1976; van der Putten et al., 2002; Brostrom et al., 2000) and this may explain the embryonic properties of this cell-line.

The inspection of global transeriptome levels revealed a close correlation between the skeletal muscle tissue and the H9C2(2-1) cell-line (Fig. 4.3B). Skeletal muscle tissue was also closely related to heart tissue (Fig. 4.3C), which led us to draw the conclusion that the H9C2(2-1) cell-line and adult heart tissue, were closely related. In contrast, the P19CL6 cell-line did not display a close gene expression pattern to the heart tissue (Fig. 4.3D). These observations suggest that the H9C2(2-1) cell-line is a better

137 cardiomyocyte model system than the P19CL6 cell-line. Similar results were also obtained by analysis of data from different experiments indicating that although both the P19CL6 and H9C2(2-1) cell-lines were more closely correlated to heart tissue than skeletal muscle, the H9C2(2-1) cell-line and adult heart tissue were more closely related than the P19CL6 cell-line and heart tissue (Fig. 4.4B; Fig. 4.5; Table 4.1; Table 4.5). Therefore, it is reasonable to conclude that the H9C2(2-1) cell-line is a better cardiac model system than the P19CL6 cell-line. This observation is in the agreement with previous studies that used the H9C2(2-1) cell-line as a cardiomyocyte model system (Mestril et al., 1994; Nyayapati et al., 1996; van der Putten et al., 2002; Kanaya et al., 2005; van Nieuwenhoven et al., 1998; Minakawa et al., 2003), even though the H9C2(2-1) cell-line was initially described as skeletal muscle cell-line (Kimes and Brandt, 1976).

In conclusion, these data indicate that the P19CL6 cell-line is more closely related to the embryonic heart than the adult heart tissue. However, the H9C2(2-1) cell-line consistently displayed a greater correlation to the heart than of the P19CL6 cell-line. Therefore, based on the global transeriptome profiles it can be concluded that the H9C2(2-1) cell-line is a better cardiomyocyte model system than the P19CL6 cell-line.

138 5

Investigation of the effect of long-chain fatty acids and fibrates on gene transcription in P19CL6 cell-line

139 Chapter 5:

Investigation of the effect of long-chain fatty acids and fibrates on gene transcription in P19CL6 cell-line

5.1 Introduction

Fatty acids have various roles in physiology that impact on lipid, carbohydrate and protein metabolism as well as growth and differentiation, namely; (i) functioning as a source of energy, (ii) serving as structural components, and (iii) functioning as a regulator of gene expression (Jump, 2004; Pegorier et al., 2004; Clarke, 2000). The regulatory function of long-chain fatty acids (LCFAs) on gene expression is achieved by controlling the abundance and activity of transcription factors such as SREBP, PPARa, PPARp, PPARy, HNF4-a, and LXR (Sampath and Ntambi, 2004; Khan and Vanden Heuvel, 2003; Clarke, 2001; Jump, 2002a; Guan et al., 2002; Mandard et al., 2004). Previous studies have shown that the characteristics of LCFA, e.g. chain length and degree of unsaturation, differentially affect gene expression levels in cardiomyocytes (Biàgi et al., 1999; Bordoni et al., 1996; Dyerberg et al., 2004; Williams et al., 2004; Cameron-Smith et al., 2003; Belury, 2002; van der Lee et al., 2000a; Hickson-Bick et al., 2000; Clavel et al., 2002). In particular, n-3 and n-6 PUF A supplementation has been implicated in the induction of FA uptake, transport and p-oxidation (Cameron-Smith et al., 2003; Clavel et al., 2002; Ntambi and Bene, 2001; Jump, 2004; Clarke, 2000), thereby modulating blood lipid profiles and preventing the progression of atherosclerosis (Fruchart, 2001; Ferre, 2004; Guan et al., 2002; Thies et al., 2003; Mandard et al., 2004).

Several cell culture and animal model systems have been used to study the effects of LCFAs and fibrates on cardiomyocytes at the gene expression level (Biagi et al., 1999; Gilde et al., 2003; Stavinoha et al., 2004; Dyerberg et al., 2004; Cameron-Smith et al., 2003; Clavel et al., 2002; Marti et al., 2002; Gamier et al., 1993; Kong and Rabkin, 2004). These ligands are known to alter gene expression levels. However, the extent of alteration and up- or down-regulation of gene expression is dependent on the gene itself, incubation/intervention period, cell type and fatty acid properties (van der Lee et al..

140 2000a; Cook et al., 2001; Marti et al., 2002; Gamier et al., 1993; Gilde et al., 2003; De Vries et al., 1997; van Bilsen et al., 1997; Stavinoha et al., 2004; Chang et al., 2001).

Therefore, based on the effects of LCFAs and fibrates on the other cardiomyocyte model systems it was hypothesised that the mRNA levels of lipid metabolism-related genes such as H-FABP, PPARa, PPARp, and PPARy would be affected by these ligands in mouse P19CL6 cells. In addition, since the effects of fatty acids differ based on their chain length and degree of unsaturation (Ntambi and Bene, 2001; Jump, 2004) it was hypothesised that LCFAs would differentially affect the expression levels of these proteins. Furthermore, the effects of LCFAs and fibrates on global transeriptome levels was investigated in order to obtain a broader picture of the effects of these compounds on gene expression levels in the P19CL6 cell-line. Therefore, the aim of this chapter was to investigate the effects of LCFAs including palmitic (saturated FA), oleic (monounsaturated FA), linoleic (n-6 PUFA), and a-linolenic (n-3 PUFA) acids and clofibrate on gene expression levels in the P19CL6 cell-line with the specific objectives as follows:

Objective 1: Determine the effect of LCFAs and clofibrate on the expression of H-FABP, PPARa, PPARp, and PPARy in the P19CL6 cell-line, using PCR-based analysis.

Objective 2: Investigation of the effect of LCFAs and clofibrate on global transeriptome levels in P19CL6 cells, using microarray analysis.

141 5.2 Determination of LCFAs and clofibrate effects on the expression of H-FABP, PPARa, PPARP, and PPARy in P19CL6 cells

5.2.1 Rationale

LCFAs and fibrates are known to affect the expression of a panel of lipid metabolism-related genes (van der Lee et ah, 2000a; van der Lee et ah, 2000b; Brandt et ah, 1998; Stavinoha et ah, 2004; Young et aL, 2001; Gilde et ah, 2003; Ouali et ah, 2000). In particular, we were interested in the effect of these ligands on the expression of PPARa, PPARP, PPARy and H-FABP in the P 19CL6 cell-line.

5.2.2 Results

P19CL6 cells were cultured with DMSO, total-RNA isolated and the mRNA expression levels of H-FABP, PPARa, PPARp, and PPARy determined. The detection of RT-PCR products showed expression of GAPDH, a housekeeping gene (Ullmannova and Haskovec, 2003), in P19CL6 cells cultured under adherent conditions, thus confirming the presence of mRNA in the sample (Fig. 5.1). The expression of H-FABP and PPARp was also detected in P19CL6 cells, whereas expression of PPARa and PPARy was not observed under the conditions used (Fig. 5.1). Similar results were obtained when the experiment was repeated (data not shown).

142 GAPDH H-FABP PPARa PPARp PPARy

500 bp ;

400 bp "

300 bp

200 bp

m Figure 5.1: Expression of H-FABP, PPARa, PPARp and PPARy in P19CL6 cells cultured under adherent conditions on exposure to 1% DMSO, as detected by RT-PCR. P19CL6 cells were cultured with 1% DMSO for 15 days, total-RNA isolated and mRNA levels determined by RT-PCR using gene-specific primers (Table 2.1). RT-PCR products were separated on 2% agarose gel containing ethidium bromide and the gel photographed. Theoretical sizes of gene-specific PCR products were GAPDH: 362 bp; H-FABP: 242 bp; PPARa: 880 bp; PPARP: 360 bp; PPARy: 856 bp.

Since there is no published data on the expression of PPARs in the P19CL6 cell-line, the THP-1 human monocyte-like cell-line was used as a positive control to optimise primer specificity and PCR conditions. THP-1 cells differentiate into macrophages and express PPARs upon exposure to phorbol myristate acetate (PMA) (Fu et aL, 2002; Neve et al., 2001; Kohro et al., 2004). Phase-contrast microscopy of THP-1 monocytes cultured for 24, 48, and 72 hours with 100 nM PMA showed an adherent monolayer, which was interpreted to indicate differentiation into macrophages-like cells. Analysis of gene expression levels in THP-1 cells cultured with 100 nM PMA showed expression of GAPDH at all the time points throughout the incubation period (Fig. 5.2), whereas PPARa and PPARy were not detected at any time point (Fig. 5.2). However, after a second round of PCR with nested primers, the expression of PPARa and PPARy in differentiated THP-1 cells was detected (Fig. 5.3), confinning the requirement for nested PCR (second amplification) to detect expression of these genes. Likewise, the detection of nested PCR products confinued the expression of H-FABP, PPARa, PPARp and PPARy in P19CL6 cells (Fig. 5.4).

143 PPARa PPARyGAPDH

300 bp

Figure 5.2: Expression of PPARa and PPARy in THP-1 cells, as detected by RT-PCR. THP-1 cells were cultured with 100 nM PMA for 24, 48, and 72 hours, RNA isolated and gene expression levels determined by RT-PCR using gene-specific primers (Table 2.1). RT-PCR products were separated on 2% agarose gel containing ethidium bromide and the gel photographed. Theoretical sizes of gene-specific PCR products were GAPDH: 362 bp; PPARa: 880 bp; PPARy: 856 bp

Q£ 0 1 c<0 % ■ a 11 I I

800 bp

60 0 bp 500 bp 40 0 bp Î 300 bp

Figure 5.3: Expression of PPARa and PPARy in THP-1 cells, as detected by RT-PCR and nested PCR. THP-1 cells were cultured with 100 nM PMA for 48 hour, RNA isolated and gene expression levels determined by RT-PCR followed by nested PCR using gene-specific primers (Table 2.2). PCR products from both RT-PCR and nested PCR steps were separated on 2% agarose gel containing ethidium bromide and the gel photographed. Theoretical sizes of gene-specific PCR products were GAPDH: 362 bp; PPARa: 880 bp; PPARy: 856 bp (for RT-PCR) and GAPDH: 362 bp; PPARa: 350 bp; PPARy: 525 bp (for nested PCR).

144 Io XII Q . CLf CLI

500 bp 400 bp 300 bp

200 bp

Figure 5.4: Expression of H-FABP, PPARa, PPARp, and PPARy in P19CL6 cells cultured under adherent conditions on exposure to 1% DMSO, as detected by nested PCR. P19CL6 cells were cultured with 1% DMSO for 15 days, RNA isolated and gene expression levels determined by RT-PCR followed by nested PCR using gene-specific primers (Table 2.2). PCR products were separated on 2% agarose gel containing ethidium bromide and the gel photographed. Theoretical sizes of gene-specific PCR products were GAPDH: 280 bp; H-FABP: 242 bp; PPARa: 350 bp; PPARP: 247 bp; PPARy: 525 bp.

In order to determine the effects of LCFAs and fibrate on gene expression levels, P19CL6 eells were cultured with 0.4 mM LCFAs or clofibrate, as discussed in the Materials and Methods. The mRNA expression of H-FABP, PPARa, PPARP and PPARy in P19CL6 cells was detected by RT-PCR followed by nested PCR techniques. Semi-quantitative gene expression levels, relative to GAPDH expression, were then calculated (Table 5.1). A One-Way ANOVA analysis was conducted to explore the impact of LCFAs or clofibrate on gene expression. The expression of H-FABP and PPARp did not display significant changes with LCFAs or clofibrate (Table 5.1). However, LCFAs had a significant effect (P<0.05) on the expression of PPARa and PPARy in P19CL6 eells (Table 5.1). In addition, levels of PPARa and PPARy mRNA were significantly increased with all the LCFAs, but not clofibrate (Table 5.1). As this data is at best semi-quantitative it should viewed as preliminary, although the results may be indicative of the underlying biological response.

145 Table 5.1: Comparison of relative H-FABP, PPARa, PPARp, and PPARy gene expression levels in P19CL6 cells cultured with different LCFAs or clofibrate.

Gene Expression Relative to GAPDH (Mean ± SD) Gene BSA PA OA LA ALA Clo

H-FABP 1.22±0.11 1.15+0.18 1.27+0.12 1.22+0.27 1.27+0.23 1.23+0.16

PPARa 1.27+0.14 1.70+0.40® 1.89±0.19^ i.ooto.io"^ 2.04+0.24"^ 1.49+0.15

PPARP 0.97+0.37 1.09+0.16 1.16±0.23 1.37+0.33 1.35±0.17 1.11+0.17

PPARy 1.21+0.23 1.78±0.21® 1.96+0.22*^ 2.14+0.26"^ 1.80+0.26® 1.24+0.20 P19CL6 cells were cu tured with LCFAs or clofibrate (0.4 mM) under adherent conditions, Gene expression levels were measured in three biological repeats using RT-PCR followed by nested PCR. PCRprodi lets were separated on 2% agarose gels containing ethidium bromide and the gel photographed. S emi-quantitative estimation of gene expression was calculated relative to GAPDH. One-Way Ab vlOVA analysis was conducted to explore the impact of LCFAs on gene expression. Data is si lown as Mean ± Standard Deviation. Significant differences in gene expression levels relati ve to BSA treatment were shown as ^P<0.05, and ^ P<0.001. ALA: a-linolenic acid; BSA: 1 lovine serum albumin; Clo: clofibrate; LA: linoleic acid; 0 A: oleic acid: PA: palmitic acid.

146 5.3 Determination of LCFA and clofibrate effects on global transeriptome levels in P19CL6 cells, using microarray analysis

5.3.1 Rationale

In the previous section, it was shown that PPARa and PPARy mRNA levels were increased by LCFAs in the P19CL6 cell-line. Therefore, in order to get a wider prospective on the effects of LCFAs and clofibrate in P19CL6 cells microarray analysis was used to investigate global transeriptome profiles in P19CL6 cells cultured with these ligands.

5.3.2 Results

P19CL6 cells were cultured with 1% DMSO for 14 days and serum-free media for day 15. Cells were then treated with LCFAs or clofibrate (0.4 mM) for 24 hours and total-RNA isolated. A loop experiment was designed (Fig. 5.5) to investigate the effects of LCFAs and clofibrate on global transeriptome levels in the P19CL6 cell-line (Simon et al., 2002; Dobbin and Simon, 2002). cDNA was synthesised, labelled according to the experimental plan (Fig. 5.5) and hybridised to microarrays. After scanning the microarrays, the data was normalised and filtered, as described in the Materials and Methods (Section 2.2.S.5). The end-point of normalisation and filtering was 2213 spots that were considered as acceptable. Scatter plots were generated, which showed that the majority of genes were not differentially expressed (Fig. 5.6). The data from this loop experiment was converted into a linear model using the LLAMA linear modelling programme (MARIEiExGen; http://exgen.ma.umist.ac.uk/) (Wit and McClure, 2004; Dalgaard, 2002; Smyth, 2004). The data produced by linear modelling was imported into Gene Cluster (http://rana.lbl.gov/) (Eisen et al., 1998) and the obtained hierarchical cluster visualised with TreeView (http://rana.lbl.gov/) (Eisen et al., 1998).

147 A BSA

Clo PA

OAALA

LA

B S1 S2 S3 8 4 8 5 86 BSA-CyS PA-Cy3 0A-Cy3 LA-Cy3 ALA-Cy3 Clo-Cy3

► W-----

PA-CyS OA-CyS LA-CyS ALA-Cy5 Clo-CyS BSA-CyS

Figure 5.5: Experimental loop design to compare global transeriptome levels in P19CL6 cells cultured with LCFAs or clofibrate. (A) loop design showing the six samples (nodes) to be compared. The sample at the start of the arrow was labelled with Cy3 and the “target” sample with Cy5. (B) samples and Cy-dye labelling used on different microarray slides, based on the loop design. S1-S6: slide numbers. ALA: a-linolenic acid; BSA: bovine serum albumin; Clo: clofibrate; LA: linoleic acid; OA: oleic acid; PA: palmitic acid.

Complete-linkage hierarchical cluster analysis of microarray data showed that expression levels in cells treated with LA or ALA were similar (Fig. 5.7) and closely related to the general effect observed with clofibrate-treated cells (Fig. 5.7). A similar pattern was also shown with PCA confirming that LA, ALA and clofibrate were closely related to each other (Fig. 5.8) but differed from that of PA and OA (Fig. 5.7 and Fig. 5.8).

148 €AO-P!oe ojaip

%

1R> !5 O i 2.2 I ^

GAo-ppB opiiuiBj eAo-ppe aiuaioujT-x)

0> D

€Ao-vsa c/2 (U Figure 5.7: Hierarchical cluster analysis of mRNA expression levels in microarrays comparing the P19CL6 cells cultured with LCFAs and clofibrate. The comparisons were clustered according to the degree of pair-wise similarity. The colour gradient represents green for down-regulation and red for up-regulation. Each gene is represented by a single row and comparisons are demonstrated by columns. ALA: a-linolenic acid; BSA: bovine serum albumin; Clo: clofibrate; LA: linoleic acid; OA: oleic acid; PA: palmitic acid.

1 .0 -

0.5 - ALA yg BSA Clo yg BSA

o l a yg BSA

BSA yg LAO

O OA yg BSA BSA yg Clo O q BSA ygA l^

O PA yg BSA

1.0 -0.5 0.0 0.5 1.0 Principal component 1 Figure 5.8: Principal components analysis (PCA) of global transeriptome levels in P19CL6 cultured with LCFAs or clofibrate. The red circle shows close relation between the samples. PCI and PC2 represent 66% and 19% of total variance, respectively. ALA: a-linolenic acid; BSA: bovine serum albumin; Clo: clofibrate; LA: linoleic acid; OA: oleic acid; PA: palmitic acid. 150 5.4 Discussion

In order to determine the effect of LCFAs and clofibrate on gene expression levels in P19CL6 cells the mRNA expression levels of lipid metabolism-related genes, /.e. H-FABP, PPARa, PPARp and PPARy, was initially investigated by PCR analysis in a focused study, while global transeriptome levels were subsequently studied by microarray analysis.

A semi-quantitative estimation of gene expression levels, using RT-PCR, in P19CL6 cells cultured with LCFAs or clofibrate (0.4 mM) showed no significant difference in the relative expression levels of H-FABP (Table 5.1). Previous studies have shown that LCFAs or fibrates up-regulate H-FABP in cardiomyocytes (Chang et al., 2001; van der Lee et aL, 2000a; van Bilsen et al., 1997; Gamier et aL, 1993; Clavel et al., 2002). However, other studies showed no changes in H-FABP mRNA after a high-fat diet (Veerkamp and van Moerkerk, 1993; Gamier et al., 1993; Coe and Bemlohr, 1998). Similar contradictory observations have also been reported on the expression of H-FABP in skeletal muscle cells, indicating no significant response in cellular H-FABP content to high-fat diet for one week (Roepstorff et aL, 2004; Veerkamp and van Moerkerk, 1993), whereas long-term (4 weeks) supplementation significantly increased expression of H-FABP mRNA (Roepstorff et al., 2004). Therefore, the non-response of H-FABP expression levels, as observed here, can probably be explained by the relatively short exposure time (24 hours). This leads to the conclusion that H-FABP gene expression responds slowly to LCFAs and clofibrate, and incubations for longer than 24 hours would probably be required before any effects of these ligands on H-FABP mRNA expression would be observed.

P19CL6 cells cultured with LCFAs or clofibrate also showed no significant changes in the relative expression level of PPARp, whereas the expression of PPARa and PPARy was significantly increased in the presence of LCFAs, regardless of the chain length and degree of unsaturation (Table 5.1). Therefore, the observations made in this study are in agreement with previous work which showed that dietary fat and fibrates increased PPARa and PPARy expression levels in different tissues (Vidal-Puig et aL, 1996; Bonilla et aL, 2000; Akbiyik et aL, 2004a; Groubet et aL, 2003; Moral et aL, 2004; Chambrier et aL, 2002; Cabrero et aL, 2003; Madsen et aL, 2005). Unlike the significant increase observed in PPARa and PPARy expression levels, incubation of P19CL6 cells with

151 LCFAs or clofibrate did not affect the expression of PPARp (Table 5.1). This observation is consistent with previous reports showing that the ingestion of dietary fat has no effect on PPARp expression levels, as shown in rat colon mucosa and human endothelial cells (Delage et al., 2005; Ye et al., 2002).

In this study, PI9CL6 cells cultured with clofibrate showed no statistically significant change in the expression levels of the PPAR isoforms (Table 5.1), although the induction of PPARa mRNA by clofibrate has been shown in other model systems (Akbiyik et al., 2004a; Ibabe et al., 2005; Guan et al., 2003; Akbiyik et al., 2004b; Amsaguine-Safir et al., 2003). The non-response of PPARa gene expression level, as observed here, can probably be explained by the short exposure time (24 hours) as long-term incubations have been mainly used in other studies (Amsaguine-Safir et al., 2003; Akbiyik et al., 2004a; Guan et al., 2003).

In addition to the effect of LCFAs and clofibrate on H-FABP, PPARa, PPARP and PPARy expression level, the effects of these ligands on global transeriptome levels in the P19CL6 cell-line was also investigated by cDNA microarray analysis. LA, ALA and clofibrate showed similar effects, as shown by cluster analysis (Fig. 5.7) and PCA (Fig. 5.8), whereas PA and OA appeared to have a different effect (Fig. 5.7 and Fig. 5.8). These observations are in agreement with previous studies showing that dietary SFA and PUFA have different effects on gene expression (Weatherill et al., 2005; Lee et al., 2004; Siculella et al., 2004; Giudetti et al., 2003a; Giudetti et al., 2003b; Garcia-Pelayo et al., 2004; Zhao et al., 2004; Weigert et al., 2004), which could be explained by differences in the affinity of SFA and PUFA to activate nuclear receptors (Zhao et al., 2004; Gilde and van Bilsen, 2003; Khan and Vanden Heuvel, 2003; Forman et al., 1995; Diniz et al., 2004). The effect of clofibrate on global transeriptome profiles was also more closely related to LA and ALA than PA or OA (Fig. 5.7 and Fig. 5.8). Therefore, the effect of PUFAs appeared to be similar to that of clofibrate, but differed from that of SFA or MUFA (monounsaturated fatty acid).

In conclusion, both RT-PCR and microarray experiments showed that culturing of P19CL6 cells with LCFAs or clofibrate affected gene expression. PCR experiments confirmed that LCFAs affect the abundance of PPARa and PPARy, but not PPARp or H-FABP, under the conditions used. Microarray experiments showed that the effects of

152 LA, ALA and clofibrate were similar but differed from those of PA and OA. In addition, the effect of clofibrate appeared to be similar to those of PUFAs.

153 Chapter 6

Conclusion

154 Chapter 6:

6.1 Conclusion

The P19CL6 cell-line, a sub-clone of PI 9 murine embryonal carcinoma cells, is used widely as an in vitro model of cardiovascular cells (Young et ah, 2004; van der Heyden and Defize, 2003; van der Heyden et ah, 2003; Peng et ah, 2002; Anisimov et ah, 2002; Paquin et ah, 2002; Ridgeway et ah, 2000; Monzen et ah, 2001). P19CL6 cells are reported to differentiate into spontaneously contracting cardiomyocytes on exposure to DMSO and have been proposed to display a pure cardiac muscle phenotype (Habara- Ohkubo, 1996; Hiroi et ah, 2001; Hosoda et ah, 2001; Young et ah, 2004). In this study P19CL6 cells displayed some characteristics of cardiomyocytes that included the expression of cardiac-specific markers, an increase in pulse rate with adrenaline and the display of mononuclear cardiomyocyte morphology. However, spontaneous contraction, a characteristic of the P19CL6 cell-line (Habara-Ohkubo, 1996), was inconsistent between the cultures and the pulse rate decreased significantly with increasing passage number, with no “beating” being recorded after passage 16. In addition, the expression of markers characteristics of cardiac- (a-MHC, p-MHC and CaCh-h) and skeletal muscle (MyoD and Myogenin) cells was observed in the P19CL6 cells, which led us to conclude that these cells have a heterogeneous phenotype. A comparison of the P19CL6 cell-line was therefore made to the H9C2(2-1) cell-line, characterised as skeletal muscle cell-line (Kimes and Brandt, 1976), and mouse cardiac and skeletal muscle tissues. Unlike the detection of both cardiac- and skeletal muscle-specific genes in P19CL6 cells, the expression of only MyoD and Myogenin, which are markers of skeletal muscle cells, was detected in H9C2(2-1) cells. In addition, H9C2(2-1) cells fused with time and formed tubular multinucleated structures, a morphology characteristic of skeletal muscle cells (Skeijanc, 1999; van der Heyden et al., 2003). Since this study was initiated to investigate the effects of LCFAs and fibrates on cardiac tissue, it was important to use a well-characterised model system. In this study, cDNA microarrays were used to characterise the P19CL6 cell-line in more depth. Analysis of data indicated that the global transcriptome profiles observed in the P19CL6 cell-line were not indicative of a pure adult cardiac or skeletal muscle phenotype and more closely related to mouse embryonic heart tissue. Therefore, the data obtained here

155 showed that the P19CL6 cell-line displays a phenotype closer to embryonic heart than adult heart. This conclusion is in accordance with other studies which indicate that the P19-derived cardiomyocytes are embryonic in nature (Skeijanc et al., 1998; van der Heyden and Defize, 2003; Rudnicki et al., 1990; Skeijanc, 1999) and is in contrast to the claim that the P19CL6 cell-line is an adult heart model system (Habara-Ohkubo, 1996). In contrast, the H9C2(2-I) cell-line showed a closer correlation with mouse adult heart tissue than skeletal muscle tissue or the P19CL6 cell-line. These observations lead to the conclusion that the H9C2(2-1) cell-line is a more suitable cardiomyocyte model system than the P19CL6 cell-line, and are in agreement with the utilization of the H9C2(2-1) cell-line as a cardiomyocyte model system (Mestril et al., 1994; Nyayapati et al., 1996; van der Putten et al., 2002; Kanaya et al., 2005; van Nieuwenhoven et al., 1998; Minakawa et al., 2003).

Since one of the aims of this study was to determine the effects of LCFAs and fibrates on gene expression levels in P19CL6 cells, the mRNA expression levels of lipid metabolism-related genes, i.e. H-FABP, PPARa, PPARp and PPARy, was initially investigated by PGR analysis in a focused study. Global transcriptome levels were subsequently studied by cDNA microarray analysis.

P19CL6 cells cultured with LCFAs or clofibrate showed no significant changes in H-FABP and PPARp expression, whereas the expression of PPARa and PPARy was significantly increased in the presence of LCFAs, regardless of the chain length and degree of unsaturation. The non-response of H-FABP expression levels, as observed here, can be explained by the relatively short exposure time (24 hours) since longer incubations with LCFAs or fibrates have been linked to increased H-FABP expression (Chang et al., 2001; Gamier et al., 1993; Clavel et al., 2002; Veerkamp and van Moerkerk, 1993; Coe and Bemlohr, 1998; Cameron-Smith et al., 2003; Roepstorff et al., 2004). The observation that the expression of PPARa and PPARy increased in the presence of LCFAs is in agreement with previous studies which showed that the administration of dietary fat increases PPARa and PPARy expression levels in different tissues (Vidal-Puig et al., 1996; Bonilla et al., 2000; Akbiyik et al., 2004a; Groubet et al., 2003; Moral et al., 2004; Chambrier et al., 2002). However, LCFAs did not affect the expression of PPARp in P19CL6 cells, which may suggest that incubations longer than 24 hours are necessary for the induction of PPARp expression by these ligands.

156 In this study, clofibrate showed no statistically significant changes in the expression levels of the PPAR isoforms, although the induction of PPARa mRNA by clofibrate has been shown in other model systems (Akbiyik et ah, 2004a; Ibabe et ah, 2005; Guan et al., 2003; Akbiyik et al., 2004b; Amsaguine-Safir et al., 2003). Since longer incubation times have been mainly used in previous studies, incubations in excess of 24 hours would probably be required before any effects of clofibrate on PPARa mRNA expression would be observed in PI9CL6 cells.

Investigation of the effects of LCFAs and clofibrate on global transcriptome levels in the P19CL6 cell-line showed that LA, ALA and clofibrate display similar effects, which differ from those of PA and OA. Similar observations have been previously reported showing that SFA and PUFA have different effects on gene expression (Weatherill et al., 2005; Lee et al., 2004; Siculella et al., 2004; Giudetti et al., 2003a; Giudetti et al., 2003b; Garcia-Pelayo et al., 2004; Zhao et al., 2004; Weigert et al., 2004), which could be explained by differences in the ability of SFA and PUFA to activate nuclear receptors (Zhao et al., 2004; Glide and van Bilsen, 2003; Khan and Vanden Heuvel, 2003; Forman et al., 1995; Diniz et al., 2004). In addition, the effect of clofibrate on global transcriptome profiles was more closely related to LA and ALA than PA or OA. Therefore, the effect of PUFAs appeared to be similar to that of clofibrate, but differed from that of SFA or MUFA.

In conclusion, the H9C2(2-1) cell-line is probably a better cardiomyocyte model system than the P19CL6 cell-line. In addition, LA, ALA and clofibrate have a similar effect, which differ from those of PA and OA, on gene expression levels in P19CL6 cells. These observations are in agreement with the effect of these ligands on other cell-lines, tissues, animals and human, and reflect the differential outcomes of SFA, MUFA and PUFA on metabolic processes related to CHD.

157 6.2 Recommendations for future work:

The works that have been carried out for this thesis has increased our knowledge of the P19CL6 cell-line and yielded some interesting results, but has also raised issues that need to be addressed:

• The pulse rate in P19CL6 cells decreased significantly with increasing passage number until no “beating” was eventually recorded. Investigation of the potential changes occurring by increasing passage number at the transcriptional level would probably reveal the mechanisms underlying losing the “beating” phenotype in P19CL6 cells.

• PC A and cluster analysis showed that the effects of PUFA on global transcriptome levels in P19CL6 cells differ from those of SFA and MUFA. Dietary SFAs are considered as atherogenic fat, whereas PUFAs are cardioprotective (Ascherio, 2002; Mori and Beilin, 2001; Wijendran and Hayes, 2004; Dyerberg et al., 2004; Calder, 2004a; Mesa et al., 2004; Calder, 2004b). Therefore, further experiments are required to identify differentially regulated genes that relate to the novel functions and metabolism of SFA, MUFA and PUFA in better cardiomyocyte model systems or animals.

• Publications of the results obtained in this study.

158 Bibliography

159 Reference List

Abumrad, N. A., Cobum, C. T., and Ibrahimi, A. Membrane proteins implicated in long- chain fatty acid uptake by mammalian cells: CD36, F ATP and FABPpm. Biochim. Biophys. Actal441'A-\2> {\999)

Abumrad, N. A., Harmon, C. M., and Ibrahimi, A. Membrane transport of long-chain fatty acids: evidence for a facilitated process. J. Lipid Res. 39:2309-2318 (1998)

Adida, A., and Spener, F. Intracellular lipid binding proteins and nuclear receptors involved in branched-chain fatty acid signaling. Prostaglandins Leukot. Essent. Fatty Acids 61:9\-9^ {2002)

Akbiyik, F., Cinar, K., Demirpence, E., Ozsullu, T., Tunca, R., Haziroglu, R., Yurdaydin, C., Uzunalimoglu, O., and Bozkaya, H. Ligand-induced expression of PPARa and activation of fatty acid oxidation enzymes in fatty liver. Eur. J. Clin. Invest. 34:429-435 (2004a)

Akbiyik, F., Ray, D. M., Bozkaya, H., and Demirpence, E. Ligand- and species- dependent activation of PPARa. Cell. Physiol. Biochem. 14:269-276 (2004b)

Amsaguine-Safir, S., Bianchi, A., Collet, P., Huin-Schohn, C., Jeanvoine, A., Becuwe, P., Kremarik-Bouillaud, P., Domenjoud, L., Keller, J. M., Schohn, H., and Dauca, M. Induction of the expression of the PPARa by clofibrate in jerboa tissues. Microsc. Res. Tech. 61:185-190 (2003)

Angerer, P., and von Schacky, C. n-3 polyunsaturated fatty acids and the cardiovascular systQm. Curr. Opin. Clin. Nutr. Metab. Care 3A39-AA5 {2000)

Anisimov, S. V., Tarasov, K. V., Riordon, D., Wobus, A. M., and Boheler, K. R. SAGE identification of differentiation responsive genes in P19 embryonic cells induced to form cardiomyocytes in vitro. Mech. Dev. 117:25-74 (2002)

Aranda, A., and Pascual, A. Nuclear hormone receptors and gene expression. Physiol. Rev. 81:1269-1304(2001)

Arkell, J., and Jackson, C. J. Constitutive secretion of MMP9 by early-passage cultured human endothelial cells. Cell Biochem. Funct. 21:381-386 (2003)

Armour, C., Garson, K., and McBumey, M. W. Cell-cell interaction modulates myoD- induced skeletal myogenesis of pluripotent PI9 cells in vitro. Exp. Cell Res. 251:79-91 (1999)

Ascherio, A. Epidemiologic studies on dietary fats and coronary heart disease. Am. J. Med. 113:9-12 (2002)

Assmann, G., Cullen, P., Jossa, F., Lewis, B., and Mancini, M. Coronary heart disease, reducing the risk : the scientific background to primary and secondary prevention

160 of coronary heart disease. A worldwide view. Arterioscler. Thromb. Vase. Biol. 19:1819-1824(1999)

Asztalos, B. F., and Schaefer, E. J. HDL in atherosclerosis: actor or bystander. Atheroscler. Suppl. 4:2\-29 (2003)

Augustus, A. S., Kako, Y., Yagyu, H., and Goldberg, I. J. Routes of FA delivery to cardiac muscle: modulation of lipoprotein lipolysis alters uptake of TG-derived FA.^m./. 284:E331-E339 (2003)

Barish, G. D., and Evans, R. M. PPARs and LXRs: atherosclerosis goes nuclear. Trends Endocrinol. Metab. 15:158-165 (2004)

Bartlett, K., and Eaton, S. Mitochondrial p-oxidation. Eur. J. Biochem. 271:462-469 (2004)

Belke, D. D., Betuing, S., Tuttle, M. J., Graveleau, C., Young, M. E., Pham, M., Zhang, D., Cooksey, R. C., McClain, D. A., Litwin, S. E., Taegtmeyer, H., Severson, D. L., Kahn, C. R., and Abel, E. D. Insulin signaling coordinately regulates cardiac size, metabolism, and contractile protein isoform expression. J. Clin. Invest. 109:629-639 (2002)

Belke, D. D., Larsen, T. S., Lopaschuk, G. D., and Severson, D. L. Glucose and fatty acid metabolism in the isolated working mouse heart. Am. J. Physiol. 277:R1210- R1217(1999)

Belury, M. A. Dietary conjugated linoleic acid in health: physiological effects and mechanisms of action. Rev. Awfr. 22:505-531 (2002)

Benko, S., Love, J. D., Beladi, M., Schwabe, J. W., and Nagy, L. Molecular determinants of the balance between co-repressor and co-activator recruitment to the retinoic acid receptor. J. Biol. Chem. 278:43797-43806 (2003)

Benoit, G., Malewicz, M., and Perlmanh, T. Digging deep into the pockets of orphan nuclear receptors: insights from structural studies. Trends Cell Biol. 14:369-376 (2004)

Berger, J. P., and Moller, D. E. The mechanisms of action of PPARs. Annu. Rev. Med. 53:409-435(2002)

Bemeis, K., and Rizzo, M. LDL size: does it matter. Riv/ 5 5 . Med. Wkly. 134:720-724 (2004)

Bhatnagar, D. Lipid-lowering drugs in the management of hyperlipidaemia. Pharmacol. 79:205-230 (1998)

Biagi, P. L., Bordoni, A., Lorenzini, A., Horrobin, D. F., and Hrelia, S. Essential fatty acid metabolism in long term primary cultures of rat cardiomyocytes: a beneficial effect of n-6/n-3 fatty acids supplementation. Mech. Ageing Dev. 107:181-195 (1999)

161 Biljmohun, R. S., Hutten, B. A., Kastelein, J. J., and Stroes, E. S. Efficacy and safety of high-density lipoprotein cholesterol-increasing compounds: a meta-analysis of randomized controlled trials. J. Co//. CcrrJ/o/. 45:185-197 (2005)

Black, B. L., Martin, J. P., and Olson, E. N. The mouse MRF4 promoter is trans­ activated directly and indirectly by muscle-specific transcription factors. J. Biol C/zom. 270:2889-2892 (1995)

Blanquart, C., Barbier, O., Fruchart, J. C., Staels, B., and Glineur, C. Peroxisome proliferator-activated receptors: regulation of transcriptional activities and roles in inflammation. J. Rfero/âfR/oc/zom. Mo/. R/o/. 85:267-273 (2003)

Bolon, C., Gauthier, C., and Simonnet, H. Glycolysis inhibition by palmitate in renal cells cultured in a two-chamber system. Am. J. Physiol 273:01732-01738 (1997)

Bonen, A., Oampbell, S. E., Benton, O. R., Ohabowski, A., Ooort, S. L., Han, X. X., Koonen, D. P., Glatz, J. F., and Luiken, J. J. Regulation of fatty acid transport by fatty acid translocase FAT/OD36. Proc. Nutr. Soc. 63:245-249 (2004)

Bonilla, S., Redonnet, A., Noel-Suberville, 0., Pallet, V., Garcin, H., and Higueret, P. High-fat diets affect the expression of nuclear retinoic acid receptor in rat liver. Rr./. Awfr. 83:665-671 (2000)

Bordoni, A., Lopez-Jimenez, J. A., Spano, 0., Bjagi, P. L., Horrobin, D. F., and Hrelia, S. Metabolism of linoleic and a-linolenic acids in cultured cardiomyocytes: effect of different n-6 and n-3 fatty acid supplementation. Mol Cell Biochem. 157:217-222 (1996)

Brandt, J. M., Djouadi, F., and Kelly, D. P. Fatty acids activate transcription of the muscle carnitine palmitoyltransferase I gene in cardiac myocytes via the PPARa. J. Biol Chem. 213:13186-13192 (\998)

Bremer, J. The biochemistry of hypo- and hyperlipidémie fatty acid derivatives: metabolism and metabolic effects. Prog. Lipid Res. 40:231-268 (2001)

Brinkmann, J. F., Abumrad, N. A., Ibrahimi, A., van der Vusse, G. J., and Glatz, J. F. New insights into long-chain fatty acid uptake by heart muscle: a crucial role for fatty acid translocase/CD36. R/oc/zem./. 367:561-570 (2002)

Brostrom, M. A., Reilly, B. A., Wilson, F. J., and Brostrom, C. O. Vasopressin-induced hypertrophy in H9C2 heart-derived myocytes. Int. J. Biochem. Cell Biol. 32:993- 1006(2000)

Brown,T.A. 2001. Gene cloning and DNA analysis: An introduction. Blackwell Science, Oxford.

Buckingham, M. Skeletal muscle formation in vertebrates. Cuir. Opin. Gen. Dev. 11:440-448 (2001)

Burr, M. L., Fehily, A. M., Gilbert, J. F., Rogers, S., Holliday, R. M., Sweetnam, P. M., Elwood, P. C., and Deadman, N. M. Effects of changes in fat, fish, and fibre

162 intakes on death and myocardial reinfarction: diet and reinfarction trial (DART). Lancet 2:151-16\(\989)

Cabrero, A., Cubero, M., Llaverias, G., Jove, M., Planavila, A., Alegret, M., Sanchez, R., Laguna, J. C., and Carrera, M. V. Differential effects of peroxisome proliferator- activated receptor activators on the mRNA levels of genes involved in lipid metabolisrh in primary human monocyte-derived macrophages. Metabolism 52:652-657(2003)

Cabrero, A., Laguna, J. C., and Vazquez, M. Peroxisome proliferator-activated receptors and the control of inflammation. Curr. Drug Targets. Inflamm. Allergy 1:243-248 (2002)

Caggiula, A. W., and Mustad, V. A. Effects of dietary fat and fatty acids on coronary artery disease risk and total and lipoprotein cholesterol concentrations: epidemiologic studies, yfm./. C/m. Awfr. 65:S1597-S1610 (1997)

Calder, P. C. Polyunsaturated fatty acids, inflammation, and immunity. Lipids 36:1007- 1024(2001)

Calder, P. C. Long-chain n-3 fatty acids and inflammation: potential application in surgical and trauma patients. Braz. J. Med. Biol. Res. 36:433-446 (2003)

Calder, P. C. Long-chain n-3 fatty acids and cardiovascular disease: further evidence and insights. Aw^r. Re5. 24:761-772 (2004a)

Calder, P. C. n-3 Fatty acids and cardiovascular disease: evidence explained and mechanisms explored. C//n. Rc/. 107:1-11 (2004b)

Calder, P. C., and Grimble, R. F. Polyunsaturated fatty acids, inflammation and immunity. Rwr./. C//Z7. Awfr. 56:S14-S 19 (2002)

Calder, P. C., and Miles, E. A. Fatty acids and atopic disease. Pediatr. Allergy Immunol. 11:29-36 (2000)

Camejo, G. PPAR agonists in the treatment of insulin resistance and associated arterial disease. Int. J. Clin. Pract. Suppl. 36-44 (2003)

Cameron-Smith, D., Burke, L. M., Angus, D. J., Tunstall, R. J., Cox, G. R., Bonen, A., Hawley, J. A., and Hargreaves, M. A short-term high-fat diet up-regulates lipid metabolism and gene expression in human skeletal muscle. Am. J. Clin. Nutr. 77:313-318(2003)

Campbell, S. E., Tandon, N. N., Woldegiorgis, G., Luiken, J. J., Glatz, J. F., and Bonen, A. A novel function for FAT/CD36: Involvement in long chain fatty acid transfer into the mitochondria. J. Biol. Chem. 279:36235-36241 (2004)

Castrillo, A., and Tontonoz, P. PPARs in atherosclerosis: the clot thickens. J. Clin. Rzv&yf. 114:1538-1540(2004)

163 Chabowski, A., Coort, S. L., Calles-Escandon, J., Tandon, N. N., Glatz, J. F., Luiken, J. J., and Bonen, A. Insulin stimulates' fatty acid transport by regulating the expression of FAT/CD36 but not FABPpm- Am. J. Physiol. 287:E781-E789 (2004)

Chahoud, G., Aude, Y. W., arid Mehta, J. L. Dietary recommendations in the prevention and treatment of coronary heart disease: do we have the ideal diet yet. Am. J. CWW. 94:1260-1267 (2004)

Chambrier, C., Bastard, J. P., Rieusset, J., Chevillotte, E., Bonnefont-Rousselot, D., Therond, P., Hainque, B., Riou, J. P., Laville, M., and Vidal, H. Eicosapentaenoic acid induces mRNA expression of PPARy. Obes. Res. 10:518-525 (2002)

Chang, W., Rickers-Haunerland, J., and Haunerland, N. H. Induction of cardiac FABP gene expression by long chain fatty acids in cultured rat muscle cells. Mol. Cell Biochem. 221:121-132 {2001)

Chapman, M. J. Fibrates in 2003: therapeutic action in atherogenic dyslipidaemia and future perspectives, vlr/zgro-yc/erojzj 171:1-13 (2003)

Chawla, A., Repa, J. J., Evans, R. M., and Mangelsdorf, D. J. Nuclear receptors and lipid physiology: opening the X-files. Science 294:1866-1870 (2001)

Chen, Y. E., Fu, M., Zhang, J., Zhu, X., Lin, Y., Akinbami, M. A., and Song, Q. Peroxisome proliferator-activated receptors and the cardiovascular system. Vitam. Aozm 66:157-188 (2003)

Chilton, R. J. Pathophysiology of coronary heart disease. Am. Osteopath. Assoc. 104:5-8 (2004)

Churchill, G. A. Fundamentals of experimental design for cDNA microarrays. Nat. Genet. 32:490-495 (2002)

Clarke, P. A., Poele, R., and Workman, P. Gene expression microarray technologies in the development of new therapeutic agents. Eur. J. Cancer 40:2560-2591 (2004)

Clarke, S. D. Polyunsaturated fatty acid regulation of gene transcription: a molecular mechanism to improve the metabolic syndrome. J. Nutr. 131:1129 (2001)

Clarke, S. D. Polyunsaturated fatty acid regulation of gene transcription: a mechanism to improve energy balance and insulin resistance. Br. J. Nutr. 83:S59-S66 (2000)

Clavel, S., Farout, L., Briand, M., Briand, Y., and Jouanel, P. Effect of endurance training and/or fish oil supplemented diet on cytoplasmic fatty acid binding protein in rat skeletal muscles and heart. Eur. J. Appl. Physiol. 87:193-201 (2002)

Claycomb, W. C., Lanson, N. A., Stallworth, B. S., Egeland, D. B., Delcarpio, J. B., Bahinski, A., and Izzo, N. J. HL-1 cells: a cardiac muscle cell line that contracts and retains phenotypic characteristics of the adult cardiomyocyte. Proc. Nat. Acad. Sci. U. S. A. 95:2979-2984 (1998)

Coe, N. R., and Bemlohr, D. A. Physiological properties and functions of intracellular fatty acid-binding proteins. Biochim. Biophys. Acta 1391:287-306 (1998)

164 Coe, N. R., Smith, A. J., Frohnert, B. L, Watkins, P. A., and Bemlohr, D. A. The fatty acid transport protein (FATPl) is a very long chain acyl-CoA synthetase. J. Biol C/zem. 274:36300-36304 (1999)

Cook, G. A., Edwards, T. L., Jansen, M. S., Bahouth, S. W., Wilcox, H. G., and Park, E. A. Differential regulation of camitine palmitoyltransferase-I gene isoforms (CPT- la and CPT-ip) in the rat heart. J. Mol Cell Cardiol 33:317-329 (2001)

Corton, J. C., Anderson, S. P., and Stauber, A. Central role of peroxisome proliferator activated receptors in the actions of peroxisome proliferators. Annu. Rev. Pharmacol. Toxicol. 40:A9\-5\8 {2000)

Cullen, P. Triacylglycerol-rich lipoproteins and atherosclerosis: where is the link. Biochem. Soc. Trans. 31'.\080-\084 {2003)

Dalgaard,P. 2002. Introductory statistics with R. Springer, London.

Das, U. N. Beneficial effect(s) of n-3 fatty acids in cardiovascular diseases: but, why and how. Prostaglandins Leukot. Essent. Fatty Acids 63:351-362 (2000)

De Lorgeril, M., Salen, P., Martin, J. L., Monjaud, I., Boucher, P., and Mamelle, N. Mediterranean dietary pattem in a randomized trial: prolonged survival and possible reduced cancer rate. Arch. Intern. Med. 158:1181-1187 (1998)

De Roos, N. M., Bost, M. L., and Katan, M. B. Replacement of dietary saturated fatty acids by trans fatty acids lowers semm HDL cholesterol and impairs endothelial function in healthy men and women. Arterioscler. Thromb. Vase. Biol 21:1233- 1237(2001)

De Vries, J. E., Vork, M. M., Roemen, T. H., De Jong, Y. F., Cleutjens, J. P., van der Vusse, G. J., and van Bilsen, M. Saturated but not mono-unsaturated fatty acids induce apoptotic cell death in neonatal rat ventricular myocytes. J. Lipid Res. 38:1384-1394(1997)

DeGrella, R. F., and Light, R. J. Uptake of fatty acids by isolated heart myocytes. Basic Res. Cardiol W .lO l-no {1985)

Delage, B., Bairras, C., Buaud, B., Pallet, V., and Cassand, P. A high-fat diet generates alterations in nuclear receptor expression: prevention by vitamin A and links with cyclooxygenase-2 and p-catenin. Int. J. Cancer 116:839-846 (2005)

Desvergne, B., and Wahli, W. Peroxisome proliferator-activated receptors: nuclear control of metabolism. Endocr. Rev. 20:649-688 (1999)

Diniz, Y. S., Cicogna, A. C., Padovani, C. R., Santana, L. S., Faine, L. A., and Novelli, E. L. Diets rich in saturated and polyunsaturated fatty acids: metabolic shifting and cardiac health. Nutrition 20:230-234 (2004)

Dobbin, K., and Simon, R. Comparison of microarray designs for class comparison and class discovery. Bioinformatics 18:1438-1445 (2002)

165 Dressel, U., Allen, T. L., Pippal, J. B., Rohde, P. R., Lan, P., and Muscat, G. E. The PPARp/ô agonist, GW501516, regulates the expression of genes involved in lipid catabolism and energy uncoupling in skeletal muscle cells. Mol Endocrinol 17:2477-2493 (2003)

Duplus, E., and Forest, C. Is there a single mechanism for fatty acid regulation of gene transcription. Biochem. Pharmacol 64:893-901 (2002)

Duplus, E., Glorian, M., and Forest, C. Fatty acid regulation of gene transcription. J. Biol Chem. 115:30149 (2000)

Duval, C., Fruchart, J. C., and Staels, B. PPARa, fibrates, lipid metabolism and inflammation, yfrc/z. Ma/. Coewr 97:665-672 (2004)

Dyerberg, J., Eskesen, D. C., Andersen, P. W., Astrup, A., Buemann, B., Christensen, J. H., Clausen, P., Rasmussen, B. F., Schmidt, E. B., Tholstrup, T., Toft, E., Toubro, S., and Stender, S. Effects of trans- and n-3 unsaturated fatty acids on cardiovascular risk markers in healthy males. An 8 weeks dietary intervention study. Rwr. J. C//n. Aw^r. 58:1062-1070 (2004)

Eckel, R. H., Grundy, S. M., Appel, L. J., Byers, T., Campos, H., Cooney, G. J., Denke, M. A., Howard, B. V., Kennedy, E., Krauss, R. M., Kris-Etherton, P. M., Lichtenstein, A. H., Marckmann, P., Pearson, T. A., Riccardi, G., Rudel, L. L., Rudrum, M., Sacks, F., Stein, D. T., Tracy, R. P., Ursin, V., Vogel, R. A., Zock, P. L., Bazzarre, T. L., and Clark, J. Summary of the scientific conference on dietary fatty acids and cardiovascular health. Circulation 103:1034-1039 (2001a)

Eckel, R. H., Grundy, S. M., Appel, L. J., Byers, T., Campos, H., Cooney, G. J., Denke, M. A., Kennedy, E., Marckmann, P., Pearson, T. A., Riccardi, G., Rudel, L. L., Rudrum, M., Stein, D. T., Tracy, R. P., Ursin, V., Vogel, R. A., Zock, P. L., Bazzarre, T. L., and Clark, J. AHA scientific statement: summary of the scientific conference on dietaiy fatty acids and cardiovascular health. J. Nutr. 131:1322- 1326 (2001b)

Edmondson, D. G., and Olson, E. N. A gene with homology to the myc similarity region of MyoD 1 is expressed during myogenesis and is sufficient to activate the muscle differentiation program. Dev. 3:628-640 (1989)

Eisen, M. B., Spellman, P. T., Brown, P. J., and Botstein, D. Cluster analysis and display of genome-wide expression patterns. Proc. Nat. Acad. Scl U. S. A. 95:14863- 14868(1998)

Elkeles, R. S., Diamond, J. R., Poulter, C., Dhanjil, S., Nicolaides, A. N., Mahmood, S., Richmond, W., Mather, H., Sharp, P., and Feher, M. D. Cardiovascular outcomes in type 2 diabetes. A double-blind placebo-controlled study of bezafibrate: the SENDCAP study. Diabetes Care 21:641-648 (1998)

Escher, P., and Wahli, W. Peroxisome proliferator-activated receptors: insight into multiple cellular functions. Mut. Res. 448:121-138 (2000)

166 Evans, R. M., Barish, G. D., and Wang, Y. X. PPARs and the complex journey to obesity. MeR. 10:355-361 (2004)

Fang, Y., Brass, A., Hoyle, D. C., Hayes, A., Bashein, A., Oliver, S. G., Waddington, D., and Rattray, M. A model-based analysis of microarray experimental error and normalisation. 31 :e96 (2003)

Fassler, R., Rohwedel, J., Maltsev, V. A., Bloch, W., Lentini, S., Guan, K., Gullberg, D., Hescheler, J., Addicks, K., and Wobus, A. M. Differentiation and integrity of cardiac muscle cells are impaired in the absence of pi integrin. J. Cell Set. 109:2989-2999(1996)

Fazio, S., and Linton, M. F. The role of fibrates in managing hyperlipidemia: mechanisms of action and clinical efficacy. Curr. Atheroscler. Rep. 6:148-157 (2004)

Ferdinandusse, S., Denis, S., Dacremont, G., and Wanders, R. J. Studies on the metabolic fate of n-3 polyunsaturated fatty acids. J. Lipid Res. 44:1992-1997 (2003)

Ferre, P. The biology of peroxisome proliferator-activated receptors: relationship with lipid metabolism and insulin sensitivity. Diabetes 53:S43-S50 (2004)

Finck, B. N., and Kelly, D. P. PPARa signaling in the gene regulatory control of energy metabolism in the normal and diseased heart. J. Mol. Cell. Cardiol. 34:1249-1257 (2002)

Fonarow, G. C., and Horwich, T. B. Prevention of heart failure: effective strategies to combat the growing epidemic. Rev. Cardiovasc. Med. 4:8-17 (2003)

Forman, B. M., Tontonoz, P., Chen, J., Brun, R. P., Spiegelman, B. M., and Evans, R. M. 15-Deoxy-A 12, 14-prostaglandin J 2 is a ligand for the adipocyte determination factor PPARy. Ce//83:803-812 (1995)

Francis, G. A., Fayard, E., Picard, F., and Auwerx, J. Nuclear receptors and the control of metabolism, yfnnw. Rev. R/zy5/o/. 65:261-311 (2003)

Frayn, K. N. The glucose-fatty acid cycle: a physiological perspective. Biochem. Soc. 31:1115-1119 (2003)

Fredenrich, A., and Bayer, P. Reverse cholesterol transport, high density lipoproteins and HDL cholesterol: recent data. Diabetes Metab. 29:201-205 (2003)

Frohnert, B. L, and Bemlohr, D. A. Regulation of fatty acid transporters in mammalian cells. Prog. Lipid Res. 39:83-107 (2000)

Frohnert, B. I., Hui, T. Y., and Bemlohr, D. A. Identification of a functional peroxisome proliferator-responsive element in the murine fatty acid transport protein gene. J. Biol. Chem. 274:3970-3977 (1999)

Fmchart, J. C. PPARa activation and high-density lipoprotein metabolism. Am. J. CarRW. 88:24-29 (2001)

167 Fruchart, J. C., Duriez, P., and Staels, B. Molecular mechanism of action of the fibrates. J. Roc. R/o/. 193:67-75 (1999)

Fruchart, J. C., Staels, B., and Duriez, P. PPARs, metabolic disease and atherosclerosis. Pharmacol. Res. 44:345-352 (2001)

Fu, Y., Luo, N., Lopes-Virella, M. F., and Garvey, W. T. The adipocyte lipid binding protein (ALBP/aP2) gene facilitates foam cell formation in human THP-1 rmcvoiflmgQS. Atherosclerosis 165:259-269 (2001)

Fujii, S., Kawaguchi, H., and Yasuda, H. Purification and characterization of fatty acid- binding protein from rat kidney. Arch. Biochem. Biophys. 254:552-558 (1987)

Fujisawa-Sehara, A., Nabeshima, Y., Hosoda, Y., Obinata, T., and Nabeshima, Y. Myogenin contains two domains conserved among myogenic factors. J. Biol. C/zcm. 265:15219-15223 (1990)

Gao, H., Bodine, P. V., Murrills, R., Bex, F. J., Bilezikian, J. P., and Morris, S. A. PTH- dependent adenylyl cyclase activation in SaOS-2 cells: passage dependent effects on G protein interactions. J. Cell. Physiol. 193:10-18 (2002)

Gao, J., and Serrero, G. Adipose differentiation-related protein (ADRP) expressed in transfected COS-7 cells selectively stimulates long chain fatty acid uptake. J. Biol. C/zew. 274:16825-16830 (1999)

Garcia-Pelayo, M. C., Garcia-Peregrin, E., and Martinez-Cayuela, M. Differential translational effects of myristic acid and eicosapentaenoic acid on 3-hydroxy-3- methylglutaryl-CoA reductase from Reuber H35 hepatoma cells. Exp. Biol. Med. 229:781-786 (2004)

Gamier, A., Poizat, C., Keriel, C., Cuchet, P., Vork, M. M., de Jong, Y. F., and Glatz, J. F. Modulation of fatty acid-binding protein content of adult rat heart in response to chronic changes in plasma lipid levels. Mol. Cell Biochem. 123:107-112 (1993)

Ghisla, S. p-oxidation of fatty acids. A century of discovery. Eur. J. Biochem. 271:459- 461 (2004)

Giguere, V. Orphan nuclear receptors: from gene to function. Endocr. Rev. 20:689-725 (1999)

Glide, A. J., and van Bilsen, M. Peroxisome proliferator-activated receptors (PPARS): regulators of gene expression in heart and skeletal muscle. Acta Physiol. Scand. 178:425-434 (2003)

Glide, A. J., van der Lee, K. A., Willemsen, P. H., Chinetti, G., van der Leij, F. R., van der Vusse, G. J., Staels, B., and van Bilsen, M. PPARa and PPARp/ô, but not PPARy, modulate the expression of genes involved in cardiac lipid metabolism. Circ. Res. 92:518-524 (2003)

Gimeno, R. E., Ortegon, A. M., Patel, S., Punreddy, S., Ge, P., Sun, Y., Lodish, H. F., and Stahl, A. Characterization of a heart-specific fatty acid transport protein. J. Biol. C/zem. 278:16039 (2003)

168 GISSI-Trial. Dietary supplementation with n-3 polyunsaturated fatty acids and vitamin E after myocardial infarction: results of the GISSI-Prevenzione trial. Lancet 354:447-455(1999)

Giudetti, A. M., Beynen, A. C., Lemmens, A. G., Gnoni, G. V., and Geelen, M. J. Hepatic fatty acid metabolism in rats fed diets with different contents of C l8:0, C l8:1 cis and C l8:1 trans isomers. Br. J. Nutr. 90:887-893 (2003a)

Giudetti, A. M., Sabetta, S., di Summa, R., Leo, M., Damiano, F., Siculella, L., and Gnoni, G. V. Differential effects of coconut oil- and fish oil-enriched diets on tricarboxylate carrier in rat liver mitochondria. J. Lipid Res. 44:2135-2141 (2003b)

Glatz, J. F., Schaap, F. G., Binas, B., Bonen, A., van der Vusse, G. J., and Luiken, J. J. Cytoplasmic fatty acid-binding protein facilitates fatty acid utilization by skeletal voxxscIq. Acta Physiol. Scand. 118:361-31 \ (2003)

Glatz, J. F., and Storch, J. Unravelling the significance of cellular fatty acid-binding proteins. Curr. Opin. Lipidol. 12:267-274 (2001)

Glatz, J. F., Vanbreda, E., Keizer, H. A., Dejong, Y. F., Lakey, J. R., Rajotte, R. V., Thompson, A., van der Vusse, G. J., and Lopaschuk, G. D. Rat heart fatty acid- binding protein content is increased in experimental diabetes. Biochem. Biophys. Res. Commun. 199:639-646 {\99A)

Goldsmith, Z. G., and Dhanasekaran, N. The microrevolution: applications and impacts of microarray technology on molecular biology and medicine. Int. J. Mol. Med. 13:483-495 (2004)

Green, L. M., Murray, D. K., Tran, D. T., Nelson, G. A., Shah, M. M., and Luben, R. A. À spontaneously arising mutation in connexin32 with repeated passage of FRTL-5 cells coincides with increased growth rate and reduced thyroxine release. J. Mol. Endocrinol. 27:145-163 (2001)

Griffin, B. A., and Fielding, B. A. Postprandial lipid handling. Cw t . Opin. Clin. Nutr. Metab. Care 4:93-98 {200\)

Grimaldi, P. A. Regulatory role of PPAR5 in muscle metabolism. A new target for Metabolic Syndrome treatment. Biochimie. 87:5-8 (2005)

Groubet, R., Pallet, V., Delage, B., Redonnet, A., Higueret, P., and Cassand, P. Hyperlipidic diets induce early alterations of the vitamin A signalling pathway in rat colonic mucosa. Endocr. Regul. 37:137-144 (2003)

Grynberg, A., and Demaison, L. Fatty acid oxidation in the heart. J. Cardiovasc. Pharmacol. 28:S11-S17 (1996)

Guan, J. Z., Tamasawa, N., Murakami, H., Matsui, J., Yamato, K., and Suda, T. Clofibrate, a peroxisome-proliferator, enhances reverse cholesterol transport through cytochrome P450 activation and oxysterol generation. Tohoku J. Exp. Med. 201:251-259 (2003)

169 Guan, Y., Zhang, Y., and Breyer, M. D. The role of PPARs in the transcriptional control of cellular processes. D r w g A e w i ' 15:147-154 (2002)

Gulick, J., Subramaniam, A., Neumann, J., and Robbins, J. Isolation and characterization of the mouse cardiac myosin heavy chain genes. J. Biol. Chem. 266:9180-9185 (1991)

Guo, Q., Wang, P. R., Milot, D. P., Ippolito, M. C., Hernandez, M., Burton, C. A., Wright, S. D., and Chao, Y. Regulation of lipid metabolism and gene expression by fenofibrate in hamsters. Biochim. Biophys. Acta 1533:220-232 (2001)

Habara-Ohkubo, A. Differentiation of beating cardiac muscle cells from a derivative of PI9 embryonal carcinoma cells. Cell Struct. Funct. 21:101-110 (1996)

Hajri, T., and Abumrad, N. A. Fatty acid transport across membranes: relevance to nutrition and metabolic pathology. Annu. Rev. Nutr. 22:383-415 (2002)

Hamilton, J. A., and Kamp, F. How are free fatty acids transported in membranes? Is it by proteins or by free diffusion through the lipids. Diabetes 48:2255-2269 (1999)

Handschin, C., and Meyer, U. A. Regulatory network of lipid-sensing nuclear receptors: roles for CAR, PXR, LXR, and FXR. Arch. Biochem. Biophys. 433:387-396 (2005)

Harmon, C. M., and Abumrad, N. A. Binding of sulfosuccinimidyl fatty acids to adipocyte membrane proteins: isolation and amino-terminal sequence of an 88-kD protein implicated in transport of long-chain fatty acids. J. Membr. Biol. 133:43- 49(1993)

Haunerland, N. H., and Spener, F. Fatty acid-binding proteins: insights from genetic manipulations. Prog. Lipid Res. 43:328-349 (2004)

Hauton, D., Bennett, M. J., and Evans, R. D. Utilisation of triacylglycerol and non- esterified fatty acid by the working rat heart: myocardial lipid substrate preference. Biochim. Biophys. Acta 1533:99-109 (2001)

Henriksen, P. A., and Kotelevtsev, Y. Application of gene expression profiling to cardiovascular disease. CorR/ovoi-c. Ro5. 54:16-24 (2002)

Hertzel, A. V., and Bemlohr, D. A. Regulation of adipocyte gene expression by polyunsaturated fatty acids. Mol. Cell. Biochem. 188:33-39 (1998)

Hertzel, A. V., and Bemlohr, D. A. The mammalian fatty acid-binding protein multigene family: molecular and genetic insights into function. Trends Endocrinol. Metab. 11:175-180 (2000)

Hescheler, J., Meyer, R., Plant, S., Krautwurst, D., Rosenthal, W., and Schultz, G. Morphological, biochemical, and electrophysiological characterization of a clonal cell (H9C2) line from rat heart. Circ. Res. 69:1476-1486 (1991)

170 Hickson-Bick, D. L., Buja, M. L., and McMillin, J. B. Palmitate-mediated alterations in the fatty acid metabolism of rat neonatal cardiac myocytes. J. Mol. Cell. Cardiol. 32:511-519(2000)

Hiroi, Y., Kudoh, S., Monzen, K., Ikeda, Y., Yazaki, Y., Nagai, R., and Komuro, I. Tbx5 associates with Nkx2-5 and synergistically promotes cardiomyocyte differentiation. Gezzef. 28:276-280(2001)

Hosoda, T., Monzen, K., Hiroi, Y., Oka, T., Takimoto, E., Yazaki, Y., Nagai, R., and Komuro, I. A novel myocyte-specific gene Midori promotes the differentiation of P19CL6 cells into cardiomyocytes. J. Biol. Chem. 276:35978-35989 (2001)

Hu, F. B., Bronner, L., Willett, W. C., Stampfer, M. J., Rexrode, K. M., Albert, C. M., Hunter, D., and Manson, J. E. Fish and omega-3 fatty acid intake and risk of coronary heart disease in women. J. Am. Med. Assoc. 287:1815-1821 (2002)

Ibabe, A., Herrero, A., and Cajaraville, M. P. Modulation of peroxisome proliferator- activated receptors (PPARs) by PPARa- and PPARy-specific ligands and by 17^- estradiol in isolated zebrafish hepatocytes. Toxicol. In Vitro 19:725-735 (2005)

Issemann, I., and Green, S. Activation of a member of the steroid hormone receptor superfamily by peroxisome proliferators. Aa^wre 347:645-650 (1990)

Jepsen, K., and Rosenfeld, M. G. Biological roles and mechanistic actions of co­ repressor complexes./. Ce//Rcz. 115:689-698 (2002)

Jones-Villeneuve, E. M., Rudnicki, M. A., Harris, J. F., and McBumey, M. W. Retinoic acid-induced neural differentiation of embryonal carcinoma cells. Mol. Cell. Biol. 3:2271-2279 (1983)

Jump, D. B. Dietary polyunsaturated fatty acids and regulation of gene transcription. CuiT. Opin. Lipidol. \3:\55-\64 (lOOld)

Jump, D. B. The biochemistry of n-3 polyunsaturated fatty acids. J. Biol. Chem. 277:8755-8758 (2002b)

Jump, D. B. Fatty acid regulation of gene transcription. Crit. Rev. Clin. Lab. Sci. 41:41- 78(2004)

Jump, D. B., and Clarke, S. D. Regulation of gene expression by dietary fat. Annu. Rev. A/rr. 19:63-90 (1999)

Jump, D. B., Clarke, S. D., Thelen, A., and Liimatta, M. Coordinate regulation of glycolytic and lipogenic gene expression by polyunsaturated fatty acids. J. Lipid R&y. 35:1076-1084 (1994)

Kamp, F., Hamilton, J. A., Kamp, F., Westerhoff, H. V., and Hamilton, J. A. Movement of fatty acids, fatty acid analogues, and bile acids across phospholipid bilayers. Biochemistiy 31:\\01 A-W086 {\993)

Kamp, F., Zakim, D., Zhang, F., Noy, N., and Hamilton, J. A. Fatty acid flip-flop in phospholipid bilayers is extremely fast. Biochemistry 34:11928-11937 (1995)

171 Kanaya, H., Takeya, R., Takeuchi, K., Watanabe, N., Jing, N., and Sumimoto, H. Fhos2, a novel formin-related actin-organizing protein, probably associates with the nestin intermediate filament. Genes Cells 10:665-678 (2005)

Kato, H., Tillotson, J., Nichaman, M. Z., Rhoads, G. G., and Hamilton, H. B. Epidemiologic studies of coronary heart disease and stroke in Japanese men living in Japan, Hawaii and California. Am. J. Epidemiol. 97:372-385 (1973)

Katz, A. M. A growth of ideas: role of calcium as activator of cardiac contraction. Cardiovasc. Res. 52:8-13 (2001)

Katz, A. M., and Lorell, B. H. Regulation of cardiac contraction and relaxation. Circulation 102:IV-69-IV-74 (2000)

Kawasaki, E. S. Microarrays and the gene expression profile of a single cell. Ann. N. Y. JcflR. Rd. 1020:92-100 (2004)

Kersten, S. Effects of fatty acids on gene expression: role of PPARa, LXRa and sterol regulatory element-binding protein-Ic. Proc. Nutr. Soc. 61:371-374 (2002)

Keys, A. Coronary heart disease in seven countries. Nutrition 13:250-252 (1997)

Khan, S. A., and Vanden Heuvel, J. P. Role of nuclear receptors in the regulation of gene expression by dietary fatty acids. J. Nutr. Biochem. 14:554-567 (2003)

Kim, J. J., and Battaile, K. P. Burning fat: the structural basis of fatty acid P-oxidation. Cun\ Opin. Stmct. Biol. 12:721-728 (2002)

Kimes, B. W., and Brandt, B. L. Properties of a clonal muscle cell line from rat heart. Exp. Cell Res. 98:367-381 (1976)

/ Kohro, T., Tanaka, T., Murakami, T., Wada, Y., Aburatani, H., Hamakubo, T., and Kodama, T. A comparison of differences in the gene expression profiles of phorbol 12-myristate 13-acetate differentiated THP-1 cells and human monocyte- derived macrophage. Jivfr/zgroWer. TTzromA 11:88-97(2004)

Kong, J. Y., and Rabkin, S. W. Reduction of palmitate-induced cardiac apoptosis by fenofibrate. Mo/. Cell Biochem. 258:1-13 (2004)

Kota, B. P., Huang, T. H., and Roufogalis, B. D. An overview on biological mechanisms of PPARs. Pharmacol. Res. 51:85-94 (2005)

Kris-Etherton, P. M., Harris, W. S., and Appel, L. J. Fish consumption, fish oil, omega-3 fatty acids, and cardiovascular disease. Arterioscler. Thromb. Vase. Biol. 23:E20- E30(2003)

Kromhout, D. Fatty acids, antioxidants, and coronary heart disease from an epidemiological perspective. Lipids 34:S27-S31 (1999)

Kromhout, D., Menotti, A., Bloemberg, B., Aravanis, C., Blackburn, H., Buzina, R., Dontas, A. S., Fidanza, F., Giampaoli, S., and Jansen, A. Dietary saturated and

172 trans fatty acids and cholesterol and 25-year mortality from coronary heart disease: the Seven Countries Study. Prev. Med. 24:308-315 (1995)

Lavrentyev, E. N., He, D., and Cook, G. A. Expression of genes participating in regulation of fatty acid and glucose utilization and energy metabolism in developing rat hearts, /. 287:H2035-H2042 (2004)

Lee, C. H., Chawla, A., Urbiztondo, N., Liao, D., Boisvert, W. A., and Evans, R. M. Transcriptional repression of atherogenic inflammation: modulation by PPAR5. Science 302:453-457 (2003a)

Lee, C. H., Olson, P., and Evans, R. M. Minireview: lipid metabolism, metabolic diseases, and peroxisome proliferator-activated receptors. Endocrinology , 144:2201 (2003b)

Lee, J. Y., Zhao, L., Youn, H. S., Weatherill, A. R., Tapping, R., Feng, L., Lee, W. H., Fitzgerald, K. A., and Hwang, D. H. Saturated fatty acid activates but polyunsaturated fatty acid inhibits Toll-like receptor 2 dimerized with Toll-like receptor 6 or 1./RzW. C/zezzz. 279:16971-16979 (2004)

Lehman, J. J., and Kelly, D. P. Transcriptional activation of energy metabolic switches in the developing and hypertrophied heart. Clin. Exp. Pharmacol. Physiol. 29:339- 345(2002)

Leigh-Firbank, E. C., Minihane, A. M., Leake, D. S., Wright, J. W., Murphy, M. C., Griffin, B. A., and Williams, C. M. Eicosapentaenoic acid and docosahexaenoic acid from fish oils: differential associations with lipid responses. Br. J. Nutr. 87:435-445 (2002)

Lemberger, T., Desvergne, B., and Wahli, W. Peroxisome proliferator-activated receptors: a nuclear receptor signaling pathway in lipid physiology. Annu. Rev. Cell Dev. Biol. 12:335 {\996)

Leonhardt, M., and Langhans, W. Fatty acid oxidation and control of food intake. Physiol. Behav. 83:645-651 (2004)

Lewin, T. M., and Coleman, R. A. Regulation of myocardial triacylglycerol synthesis and metabolism. Biochim. Biophys. Acta 1634:63-75 (2003)

Li, A. C., and Glass, C. K. PPAR- and LXR-dependent pathways controlling lipid metabolism and the development of atherosclerosis. J. Lipid Res. 45:2161-2173 (2004)

Li, Y., Lambert, M. H., and Xu, H. E. Activation of nuclear receptors: a perspective from structural genomics. Structure 11:741-746 (2003)

Lindsay,G., and Gaw,A. 2001. Coronary heart disease prevention. Churchill Livingstone, London.

Linsel-Nitschke, P., and Tall, A. R. HDL as a target in the treatment of atherosclerotic cardiovascular disease. Nat. Rev. Drug Discov. 4:193-205 (2005)

173 Luiken, J. J., Coort, S. L., Koonen, D. P., Bonen, A., and Glatz, J. F. Signalling components involved in contraction-inducible substrate uptake into cardiac myocytes. Proc. Aw/r. Roc. 63:251-258 (2004)

Madsen, L., Petersen, R. K., and Kristiansen, K. Regulation of adipocyte differentiation and function by polyunsaturated fatty acids. Biochim. Biophys. Acta 1740:266-286 (2005)

Maglich, J. M., Sluder, A., Guan, X., Shi, Y., McKee, D. D., Garrick, K., Kamdar, K., Willson, T. M., and Moore, J. T. Comparison of complete nuclear receptor sets from the human, Caenorhabditis elegans and Drosophila genomes. Genome Biol. 2:R29.1-R29.7(2001)

Makinde, A. O., Kantor, P. F., and Lopaschuk, G. D. Maturation of fatty acid and carbohydrate metabolism in the newborn heart. Mol. Cell. Biochem. 188:49-56 (1998)

Manco, M., Calvani, M., and Mingrone, G. Effects of dietary fatty acids on insulin sensitivity and secretion. Diabetes Obes. Metab. 6:402-413 (2004)

Mandard, S., Muller, M., and Kersten, S. PPARa target genes. Cell. Mol. Life Sci. 61:393-416 (2004)

Mardy, K., Belke, D. D., and Severson, D. L. Chylomicron metabolism by the isolated perfused mouse heart. Am. J. Physiol. 281:E357-E364 (2001)

Marti, A., Vaquerizo, J., Zulet, M. A., Moreno-Aliaga, M. J., and Martinez, J. A. Down- regulation of heart HFABP and UCP2 gene expression in diet-induced (cafeteria) obese rats. J. Physiol. Biochem. 58:69-74 (2002)

Marx, N., Duez, H., Fruchart, J. C., and Staels, B. Peroxisome proliferator-activated receptors and atherogenesis: regulators of gene expression in vascular cells. Circ. Rcj. 94:1168-1178(2004)

Massolini, G., and Calleri, E. Survey of binding properties of fatty acid-binding proteins: chromatographic methods. J. Chromatogr. B 797:255-268 (2003)

Mata, L. P., and Ortega, R. M. Omega-3 fatty acids in the prevention and control of cardiovascular disease. Eur. J. Clin. Nutr. 57:S22-S25 (2003)

Mathers, J. C. What can we expect to leam from genomics. Proc. Nutr Soc. 63:1-4 (2004)

Mauger, J. F., Lichtenstein, A. H., Ausman, L. M., Jalbert, S. M., Jauhiainen, M., Ehnholm, C., and Lamarche, B. Effect of different forms of dietary hydrogenated fats on LDL particle size. Am. J. Clin. Nutr. 78:370-375 (2003)

McBumey, M. W. P19 embryonal carcinoma cells. Int. J. Dev. Biol. 37:135-140 (1993)

McBumey, M. W., Jones-Villeneuve, E. M., Edwards, M. K., and Anderson, P. J. Control of muscle and neuronal differentiation in a cultured embryonal carcinoma cell line. Nature 299:165-167 (1982)

174 McPherson,M.J., Hames,B.D., and Taylor,G.R. 1995. PCR2: A Practical Approach. Oxford University Press Inc., New York.

Mesa, M. D., Buckley, R., Minihane, A. M., and Yaqoob, P. Effects of oils rich in eicosapentaenoic and docosahexaenoic acids on the oxidizability and thrombogenicity of low-density lipoprotein. Atherosclerosis 175:333-343 (2004)

Mestril, R., Chi, S. H., Sayen, M. R., O’Reilly, K., and Dillmann, W. H. Expression of inducible stress protein 70 in rat heart myogenic cells confers protection against simulated ischemia-induced injury. J. Clin. Invest. 93:759-767 (1994)

Metzger, J. M., Lin, W. I., Johnston, R. A., Westfall, M. V., and Samuelson, L. C. Myosin heavy chain expression in contracting myocytes isolated during embryonic stem cell cardiogenesis. Circ. Res. 76:710-719 (1995)

Minakawa, M., Takeuchi, K., Ito, K., Tsushima, T., Fukui, K., Takaya, S., and Fukuda, I. Restoration of sarcoplasmic reticulum protein level by thyroid hormone contributes to partial improvement of myocardial function, but not to glucose metabolism in an early failing heart. Eur. J Cardiothorac. Surg. 24:493-501 (2003)

Mitcheson, J. S., Hancox, J. C., and Levi, A. J. Cultured adult cardiac myocytes: future applications, culture methods, morphological and electrophysiological properties. Cardiovasc. Res. 39:280-300 (1998)

Monzen, K., Hiroi, Y., Kudoh, S., Akazawa, H., Oka, T., Takimoto, E., Hayashi, D., Hosoda, T., Kawabata, M., Miyazono, K., Ishii, S., Yazaki, Y., Nagai, R., and Komuro, I. Smads, TAKl, and their common target ATF-2 play a critical role in cardiomyocyte differentiation./. Cell Biol. 153:687-698 (2001)

Moral, R., Solanas, M., Manzanares, E. M., Haro, D., and Escrich, E. Influence of DMBA-induced mammary cancer on the liver CPT-I, mitochondrial HMG-CoA synthase and PPARa mRNA expression in rats fed a low or high com oil diet. Int. J. Mol. Med. 14:283-287 (2004)

Mori, T. A., and Beilin, L. J. Long-chain omega 3 fatty acids, blood lipids and cardiovascular risk reduction. Curr. Opin. Lipidol. 12:11-17 (2001)

Morkin, E. Control of cardiac myosin heavy chain gene expression. Microsc. Res Tech. 50:522-531 (2000)

Morley, P., and Whitfield, J. F. The differentiation inducer, dimethyl sulfoxide, transiently increases the intracellular calcium ion concentration in various cell types. J. Cell. Physiol. 156:219-225 (1993)

Nag, A. C., Ingland, M., and Cheng, M. Factors controlling embryonic heart cell proliferation in serum-ffee synthetic media. In Vitro Cell Dev. Biol. 21:553-562 (1985)

Nagy, L., and Schwabe, J. W. Mechanism of the nuclear receptor molecular switch. Trends Biochem. Sci. 29:317-324 (2004)

175 Nakamura, M. T., and Nara, T. Y. Structure, function, and dietary regulation of A-6, A-5, and A-9 desaturases. Annu. Rev. Nutr. 24:345-376 (2004)

Napoli, C., Lerman, L. O., Sica, V., Lerman, A., Tajana, G., and De Nigris, F. Microarray analysis: a novel research tool for cardiovascular scientists and physicians. Aeflr/89:597-604 (2003)

Neve, B. P., Corseaux, D., Chinetti, G., Zawadzki, C., Fruchart, J. C., Duriez, P., Staels, B., and Jude, B. PPARa agonists inhibit tissue factor expression in human monocytes and macrophages. C/rcw/flRozz 103:207-212 (2001)

Newton,C.A., and Graham,A. 1997. PCR. BIOS Scientific Publishers Ltd, Oxford.

Niu, Y. G., Hauton, D., and Evans, R. D. Utilization of triacylglycerol-rich lipoproteins by the working rat heart: routes of uptake and metabolic fates. J. Physiol. 558:225-237 (2004)

Norman, H. A., Go, V. L., and Butrum, R. R. Review of the international research conference on food, nutrition, and cancer. V. Nutr. 134:3391 S-3393 (2004)

Ntambi, J. M., and Bene, H. Polyunsaturated fatty acid regulation of gene expression. J. Mol. Neurosci. 16:213-118 {200\)

Nyayapati, S., Afshan, G., Lomitzo, F., Byrnes, R. W., and Petering, D. H. Depletion of cellular iron by BPS and ascorbate: effect on toxicity of adriamycin. Free Radie. Biol. Med. 20:319-329 (1996)

Ockner, R. K., Manning, J. A., Poppenhausen, R. B., and Ho, W. K. A binding protein for fatty acids in cytosol of intestinal mucosa, liver, myocardium, and other tissues. Rc/ewce 177:56-58 (1972)

Olpin, S. E. Implications of impaired ketogenesis in fatty acid oxidation disorders. Prostaglandins Leukot. Essent. Fatty Acids 70:293-308 (2004)

Oomen, C. M., Feskens, E. J., Rasanen, L., Fidanza, F., Nissinen, A. M., Menotti, A., Kok, F. J., and Kromhout, D. Fish consumption and coronary heart disease mortality in Finland, Italy, and the Netherlands. Am. J. Epidemiol. 151:999-1006 (2000)

Ouali, F. E., Djouadi, F., Merlet-Benichou, C., Riveau, B. E., and Bastin, J. Regulation of fatty acid transport protein and mitochondrial and peroxisomal P-oxidation gene expression by fatty acids in developing rats. Pediatr. Res. 48:691-696 (2000)

Packard, C. J. Triacylglycerol-rich lipoproteins and the generation of small, dense low- density lipoprotein. R/oc/zew. Soc. Trazzj. 31:1066-1069 (2003)

Panagiotakos, D. B., Chrysohoou, C., Pitsavos, C., Menotti, A., Dontas, A. S., Skoumas, J., Stefanadis, C., and Toutouzas, P. Forty-years (1961-2001) of all-cause and coronary heart disease mortality and its determinants: the Corfu cohort fi*om the Seven Countries Study. Int. J. Cardiol. 90:73-79 (2003)

176 Paquin, J., Danalache, B. A., Jankowski, M., McCann, S. M., and Gutkowska, J. Oxytocin induces differentiation of P19 embiyonic stem cells to cardiomyocytes. Proc. Nat. Acad. Sci. U. S. A. 99:9550-9555 (2002)

Pegorier, J. P., May, C. L., and Girard, J. Control of gene expression by fatty acids. J. Nutr. 134:24448-2449 (2004)

Peng, C. P., Wei, Y., Levsky, J. M., McDonald, T. V., Childs, G., and Kitsis, R. N. Microarray analysis of global changes in gene expression during cardiac myocyte differentiation. R/zyi'/o/. 9:145-155 (2002)

Pennec, J. P., Gallet, M., Gioux, M., and Dorange, G. Cell culture of bivalves: tool for the study of the effects of environmental stressors. Cell Mol. Biol. 48:351-358 (2002)

Plutzky, J. The vascular biology of atherosclerosis. Am. J. Med. 115:55-61 (2003)

Pohl, J., Ring, A., Ehehalt, R., Herrmann, T., and Stremmel, W. New concepts of cellular fatty acid uptake: role of fatty acid transport proteins and of caveolae. Proc. Nutr. Roc. 63:259-262 (2004a)

Pohl, J., Ring, A., Ehehalt, R., Schulze-Bergkamen, H., Schad, A., Verkade, P., and Stremmel, W. Long-chain fatty acid uptake into adipocytes depends on lipid raft function. Rzoc/zczzzzj^zg; 43:4179-4187 (2004b)

Pohl, J., Ring, A., Hermann, T., and Stremmel, W. Role of FATP in parenchymal cell fatty acid uptake. Rzoc/zzm. Rzqp^j. 1686:1-6 (2004c)

Pohl, J., Ring, A., Korkmaz, U., Ehehalt, R., and Stremmel, W. FAT/CD36-mediated long-chain fatty acid uptake in adipocytes requires plasma membrane rafts. Mol. Biol. Cell. 16:24-31 (2005)

Poulter, N. Global risk of cardiovascular disease. Heart 89:2-5 (2003)

Poulter, N. Coronary heart disease is a multifactorial disease. Am. J. Hypert. 12:92-95 (1999)

Puddu, P., Puddu, G. M., and Muscari, A. Peroxisome proliferator-activated receptors: are they involved in atherosclerosis progression. Int. J. Cardiol. 90:133-140 (2003)

Rader, D. J. Regulation of reverse cholesterol transport and clinical implications. Am. J. Cardiol. 92:42-49 (2003)

Ramirez, M., Amate, L., and Gil, A. Absorption and distribution of dietary fatty acids from different sources. Early Hum. Dev. 65:895-8101 (2001)

Rangwala, 8. M., and Lazar, M. A. PPARy in diabetes and metabolism. Trends R/zflZ'/zzflco/. Rcz. 25:331-336 (2004)

177 Reddy, J. K., and Hashimoto, T. Peroxisomal p-oxidation and peroxisome proliferator- activated receptor: an adaptive metabolic system. Annu. Rev. Nutr. 21:193-230 (2001)

Renaud, J. P., and Moras, D. Structural studies on nuclear receptors. Cell Mol Life Scl 57:1748-1769(2000)

Renaud, S. C., and Lanzmann-Petithory, D. Coronary heart disease: dietary links and ^dAhogQUQsis. Public Health Nutr. 4A59-A1A {20Q\)

Richieri, G. V., Ogata, R. T., Zimmerman, A. W., Veerkamp, J. H., and Kleinfeld, A. M. Fatty acid binding proteins from different tissues show distinct patterns of fatty acid interactions, 39:7197-7204 (2000)

Ricote, M., Huang, J. T., Welch, J. S., and Glass, C. K. The PPARy as a regulator of monoc5de/macrophage function. J. TcwAroc. 5/o/. 66:733-739 (1999)

Ridgeway, A. G., Wilton, S.j and Skeqanc, I. S. Myocyte enhancer factor 2C and myogenin up-regulate each other's expression and induce the development of skeletal muscle in P19 cells. J. Biol Chem. 275:41-46 (2000)

Robbins, J., Horan, T., Gulick, J., and Kropp, K. The chicken myosin heavy chain family. J. Biol Chem. 261:6606-6612 (1986)

Robinson-Rechavi, M., Carpentier, A. S., Duffraisse, M., and Laudet, V. How many nuclear hormone receptors are there in the human genome. Trends Genet. 17:554- 556(2001)

Rodriguez, E. R., Tan, C. D., Onwuta, U. S., Yu, Z. X., Ferrans, V. J., and Parrillo, J. E. 3,5,3'-Triiodo-L-thyronine induces cardiac myocyte differentiation but not neuronal differentiation in PI9 teratocarcinoma cells in a dose-dependent manner. Biochem. Biophys. Res. Commun. 205:652-658 (1994)

Roepstorff, C., Helge, J. W., Vistisen, B., and Kiens, B. Studies of plasma membrane fatty acid-binding protein and other lipid-binding proteins in human skeletal muscle. Proc. Nutr. Soc. 63:239-244 (2004)

Ross, R. Atherosclerosis: an inflammatory disease. N. Engl J. Med. 340:115-126 (1999)

Ross, R. The pathogenesis of atherosclerosis: a perspective for the 1990s. Nature 362:801-809(1993)

Rudnicki, M. A., Jackowski, G., Saggin, L., and McBumey, M. W. Actin and myosin expression during development of cardiac muscle from cultured embryonal carcinoma cells. Dev. Biol 138:348-358 (1990)

Sabourin, L. A., and Rudnicki, M. A. The molecular regulation of myogenesis. Clinical Genetics SI:\6-25 (lOOÇi)

Sacchettini, J. C., Gordon, J. I., and Banaszak, L. J. Crystal structure of rat intestinal fatty-acid-binding protein. Refinement and analysis of the Escherichia coli- derived protein with bound palmitate. J. Mol Biol 208:327-339 (1989)

178 Sacks, F. M., and Katan, M. B. Randomized clinical trials on the effects of dietary fat and carbohydrate on plasma lipoproteins and cardiovascular disease. Am. J. Med. 113:13-24(2002)

Sampath, H., and Ntambi, J. M. Polyunsaturated fatty acid regulation of gene expression. Nutr. Rev. 62:333-339 (2004)

Schaffer, J. E., and Lodish, H. F. Expression cloning and characterization of a novel adipocyte long chain fatty acid transport protein. Cell 79:427-436 (1994)

Schulman, I. G., and Heyman, R. A. The flip side: identifying small molecule regulators of nuclear receptors. CAem. Rfo/. 11:639-646 (2004)

Shmueli, O., Hom-Saban, S., Chalifa-Caspi, V., Shmoish, M., Ophir, R., jamin-Rodrig, H., Safran, M., Domany, E., and Lancet, D. GeneNote: whole genome expression profiles in normal human tissues. C. R. Biol. 326:1067-1072 (2003)

Siculella, L., Sabetta, S., Damiano, F., Giudetti, A. M., and Gnoni, G. V. Different dietary fatty acids have dissimilar effects on activity and gene expression of mitochondrial tricarboxylate carrier in rat liver. FEBS Lett. 578:280-284 (2004)

Simon, R., Radmacher, M. D., and Dobbin, K. Design of studies using DNA microarrays. Genet. Epidemiol. 23:21-36 (2002)

Simopoulos, A. P., Leaf, A., and Salem, J. Workshop statement on the essentiality of and recommended dietary intakes for omega-6 and omega-3 fatty acids. Prostaglandins Leukot. Essent. Fatty Acids 63:119-121 (2000)

Skeijanc, I. S. Cardiac and skeletal muscle development in P19 embryonal carcinoma cells. Trends Cardiovasc. Med. 9:139-143 (1999)

Skeijanc, I. S., Petropoulos, H., Ridgeway, A. G., and Wilton, S. Myocyte enhancer factor 2C and Nkx2-5 up-regulate each other's expression and initiate cardiomyogenesis in P19 cells. J. Biol. Chem. 273:34904-34910 (1998)

Skeijanc, I. S., Slack, R. S., and McBumey, M. W. Cellular aggregation enhances MyoD-directed skeletal myogenesis in embryonal carcinoma cells. Mol. Cell Biol. 14:8451-8459(1994)

Sladek, R., and Giguere, V. Orphan nuclear receptors: an emerging family of metabolic regulators. Adv. Pharmacol. 47:23-87 (2000)

Smith, S. A. Peroxisome proliferator-activated receptors and the regulation of mammalian lipid metabolism. Biochem. Soc. Trans. 30:1086-1090 (2002)

Smith, S. C., Reuhl, K. R., Craig, J., and McBumey, M. W. The role of aggregation in embryonal carcinoma cell differentiation. J. Cell. Physiol. 131:74-84 (1987)

Smyth, G. K. Linear models and empirical bayes methods for assessing differential expression in microarray experiments. Stat. Appl. Genet. Mol. Biol. 3:1-26 (2004)

179 Soltys, C. L., Buchholz, L., Gandhi, M., Clanachan, A. S., Walsh, K., and Dyck, J. R. Phosphorylation of cardiac protein kinase B is regulated by palmitate. Am. J. 283:H1056-H1064 (2002)

Spady, D. K. Reverse cholesterol transport and atherosclerosis regression. Circulation 100:576-578 (1999)

Spratt, K. A. Reducing the risk of coronary heart disease via lipid reduction. J. Am. Osteopath. Assoc. 1{S4:^9-S\2> (lOQA)

St-Onge, M. P., and Jones, P. J. Physiological effects of medium-chain triglycerides: potential agents in the prevention of obesity. J. Nutr. 132:329-332 (2002)

Stahl, A., Gimeno, R. E., Tartaglia, L. A., and Lodish, H. F. Fatty acid transport proteins: a current view of a growing family. Trends Endocrinol. Metab. 12:266-273 (2001)

Stanley, W. C., Recchia, F. A., and Lopaschuk, G. D. Myocardial substrate metabolism in the normal and failing heart. Physiol. Rev. 85:1093-1129 (2005)

Stavinoha, M. A., RaySpellicy, J. W., Hart-Sailors, M. L., Mersmann, H. J., Bray, M. S., and Young, M. E. Diurnal variations in the responsiveness of cardiac and skeletal muscle to fatty acids. Am. J. Physiol. 287:E878-E887 (2004)

Stender, S., and Dyerberg, J. Influence of trans fatty acids on health. Ann. Nutr. Metab. 48:61-66 (2004)

Stocker, R., and Keaney, J. F. Role of oxidative modifications in atherosclerosis. Physiol. Rev. 84:1381-1478 (2004)

Storch, J., and Thumser, A. E. The fatty acid transport function of fatty acid-binding proteins. Biochim. Biophys. Acta 1486:28-44 (2000)

Stremmel, W., Strohmeyer, G., Borchard, F., Kochwa, S., and Berk, P. D. Isolation and partial characterization of a fatty acid binding protein in rat liver plasma membranes. Proc. Nat. Acad. Sci. U. S. A. 82:4-8 (1985)

Szapary, P. O., and Rader, D. J. The triglyceride-high-density lipoprotein axis: an important target of therapy. Am. Heart J. 148:211-221 (2004)

Takahashi, T., Lord, B., Schulze, P. C., Fiyer, R. M., Sarang, S. S., Gullans, S. R., and Lee, R. T. Ascorbic acid enhances differentiation of embryonic stem cells into cardiac myocytes. Circulation 107:1912-1916 (2003)

Talmud, P. J. How to identify gene-environment interactions in a multifactorial disease: CHD as an example. Proc. Nutr. Soc. 63:5-10 (2004)

Talmud, P. J., and Stephens, J. W. Lipoprotein lipase gene variants and the effect of environmental factors on cardiovascular disease risk. Diab. Obes. Metab. 6:1-7 (2004)

Tan, N. S., Shaw, N. S., Vinckenbosch, N., Liu, P., Yasmin, R., Desvergne, B., Wahli, W., and Noy, N. Selective cooperation between fatty acid binding proteins and

180 peroxisome proliferator-activated receptors in regulating transcription. Mol. Cell. R/o/. 22:5114(2002)

Tandon, N. N., Lipsky, R. H., Burgess, W. H., and Jamieson, G. A. Isolation and characterization of platelet glycoprotein IV (CD36). J. Biol. Chem. 264:7570- 7575(1989)

Tham, D. M., Wang, Y. X., and Rutledge, J. C. Modulation of vascular inflammation by AR.S. Drug News Perspect. \6:\09-\\6

Thies, F., Garry, J. M., Yaqoob, P., Rerkasem, K., Williams, J., Shearman, C. P., Gallagher, P. J., Calder, P. C., and Grimble, R. F. Association of n-3 polyunsaturated fatty acids with stability of atherosclerotic plaques: a randomised controlled trial. The Lancet 361:477-485 (2003)

Tholstrup, T., Ehnholm, C., Jauhiainen, M., Petersen, M., Hoy, C. E., Lund, P., and Sandstrom, B. Effects of medium-chain fatty acids and oleic acid on blood lipids, lipoproteins, glucose, insulin, and lipid transfer protein activities. Am. J. Clin. Vwrr. 79:564-569 (2004)

Thompson, J., Winter, N., Terwey, D., Bratt, J., and Banaszak, L. J. The crystal structure of the liver fatty acid-binding protein: a complex with two bound oleates. J. Biol. C/zem. 272:7140-7150(1997)

Thorp, J. M., and Waring, W. S. Modification of metabolism and distribution of lipids by ethyl chlorophenoxyisobutyrate. Nature 194:948-949 (1962)

Trigatti, B. L., Anderson, R. G., and Gerber, G. E. Identification of caveolin-I as a fatty acid binding protein. Biochem. Biophys. Res. Commun. 255:34-39 (1999)

Tsuchiya, S., Yamabe, M., Yamaguchi, Y., Kobayashi, Y., Konno, T., and Tada, K. Establishment and characterization of a human acute monocytic leukemia cell line (THP-1). Thr. J: Ca/zcgr 26:171-176 (1980)

Tulenko, T. N., and Sumner, A. E. The physiology of lipoproteins. J. Nucl. Cardiol. 9:638-649 (2002)

Ullmannova, V., and Haskovec, C. The use of housekeeping genes (HKG) as an internal control for the detection of gene expression by quantitative real-time RT-PCR. Folia Biol. 49:211-216 (2003)

Vakkilainen, J., Steiner, G., Ansquer, J. C., Aubin, F., Rattier, S., Foucher, C., Hamsten, A., and Taskinen, M. R. Relationships between low-density lipoprotein particle size, plasma lipoproteins, and progression of coronary artery disease: the DAIS study. Chrw/arioM 107:1733-1737 (2003) van Bilsen, M., De Vries, J. E., and van der Vusse, G. J. Long-term effects of fatty acids on cell viability and gene expression of neonatal cardiac myocytes. Prostaglandins Leukot. Essent. Fatty Acids 57:39-45 (1997)

181 van Bilsen, M., Smeets, P. J., Gilde, A. J., and van der Vusse, G. J. Metabolic remodelling of the failing heart: the cardiac burn-out syndrome. Cardiovasc. Res. 61:218-226(2004)

van der Heyden, M. A., and Defize, L. H. Twenty one years of P19 cells: what an embryonal carcinoma cell line taught us about cardiomyocyte differentiation. Cardiovasc. Res. 58:292-302 (2003)

van der Heyden, M. A., van Kempen, M. J., Tsuji, Y., Rook, M. B., Jongsma, H. J., and Opthof, T. PI 9 embryonal carcinoma cells: a suitable model system for cardiac electrophysiological differentiation at the molecular and functional level. Cardiovasc. Res. SS’.AlOAll

van der Lee, K. A., Vork, M. M., De Vries, J. E., Willemsen, P. H., Glatz, J. P., Reneman, R. S., van der Vusse, G. J., and van Bilsen, M. Long-chain fatty acid-induced changes in gene expression in neonatal cardiac myocytes. /. Lipid Res. 41:41-47 (2000a)

van der Lee, K. A., Willemsen, P. H., van der Vusse, G. J., and van Bilsen, M. Effects of fatty acids on uncoupling protein-2 expression in the rat heart. FASEB J. 14:495- 502 (2000b)

van der Putten, H. H., Joosten, B. J., Klaren, P. H., and Everts, M. E. Uptake of tri­ iodothyronine and thyroxine in myoblasts and myotubes of the embryonic heart cell line H9C2(2-1). J. Endocrinol. 175:587-596 (2002) van der Vusse, G. J., van Bilsen, M., and Glatz, J. F. Cardiac fatty acid uptake and transport in health and disease. Cardiovasc. Res. 45:279-293 (2000) van Nieuwenhoven, F. A., Luiken, J. J., De Jong, Y. F., Grimaldi, P. A., van der Vusse, G. J., and Glatz, J. F. Stable transfection of fatty acid translocase (CD36) in a rat heart muscle cell line (H9C2). J. Lipid Res. 39:2039-2047 (1998) van Nieuwenhoven, F. A., Willemsen, P. H., van der Vusse, G. J., and Glatz, J. F. Co­ expression in rat heart and skeletal muscle of four genes coding for proteins implicated in long-chain fatty acid uptake. Int. J. Biochem. Cell Biol. 31:489-498 (1999)

Veerkamp, J. H., and Maatman, R. G. Cytoplasmic fatty acid-binding proteins: their structure and genes. Prog. Lipid Res. 34:17-52 (1995)

Veerkamp, J. H., and van Moerkerk, H. T. Fatty acid-binding protein and its relation to fatty acid oxidation. Mol. Cell Biochem. 123:101-106 (1993)

Ventura-Clapier, R., Gamier, A., and Veksler, V. Energy metabolism in heart failure. J. RAyjW. 555:1-13 (2004)

Verges, B. Clinical interest of PPAR ligands. Diabetes Metab. 30:7-12 (2004)

Vidal-Puig, A., Jimenez-Linan, M., Lowell, B. B., Hamann, A., Hu, E., Spiegelman, B., Flier, J. S., and Moller, D. E. Regulation of PPARy gene expression by nutrition and obesity in rodents. J. Clin. Invest. 97:2553-2561 (1996)

182 Vinciotti, V., Khanin, R., D'Alimonte, D., Liu, X., Cattini, N., Hotchkiss, G., Bucca, G., De Jesus, O., Rasaiyaah, J., Smith, C. P., Kellam, P., and Wit, E. An experimental evaluation of a loop versus a reference design for two-channel microarrays. Bioinformatics 21'A92-50\ (2005)

Wahli, W. Peroxisome proliferator-activated receptors (PPARs): from metabolic control to epidermal wound healing. Swiss. Med. Wdy. 132:83-91 (2002)

Wang, M., and Briggs, M. R. HDL: the metabolism, function, and therapeutic importance. C/icm. Rev. 104:119-137 (2004)

Wang, Y. X., Lee, C. H., Tiep, S., Yu, R. T., Ham, J., Kang, H., and Evans, R. M. PPARÔ activates fat metabolism to prevent obesity. Cell 113:159-170 (2003)

Weatherill, A. R., Lee, J. Y., Zhao, L., Lemay, D. G., Youn, H. S., and Hwang, D. H. Saturated and polyunsaturated fatty acids reciprocally modulate dendritic cell functions mediated through TLR4. J. Immunol. 174:5390-5397 (2005)

Weigert, C., Brodbeck, K., Staiger, H., Kausch, C., Machicao, P., Haring, H. U., and Schleicher, E. D. Palmitate, but not unsaturated fatty acids, induces the expression of interleukin-6 in human myotubes through proteasome-dependent activation of nuclear factor-KB. J Biol. Chem. 279:23942-23952 (2004)

Whatling, C., Bjork, H., Gredmark, S., Hamsten, A., and Eriksson, P. Effect of macrophage differentiation and exposure to mildly oxidized LDL on the proteolytic repertoire of THP-1 monocytes. J. Lipid Res. 45:1768-1776 (2004)

White, S. M., Constantin, P. E., and Claycomb, W. C. Cardiac physiology at the cellular level: use of cultured HL-1 cardiomyocytes for studies of cardiac muscle cell structure and function. Am. J. Physiol. 286:H823-H829 (2004)

Wijendran, V., and Hayes, K. C. Dietary n-6 and n-3 fatty acid balance and cardiovascular health. Annu. Rev. Nutr. 24:597-615 (2004)

Wilcox, J. N. Fundamental principles of in situ hybridization. J. Histochem. Cytochem. 41:1725-1733 (1993)

Williams, C. M., Bateman, P. A., Jackson, K. G., and Yaqoob, P. Dietary fatty acids and chylomicron synthesis and secretion. Biochem. Soc. Trans. 32:55-58 (2004)

Willson, T. M., Brown, P. J., Stembach, D. D., and Henke, B. R. The PPARs: from orphan receptors to drug discovery. J. Med. Chem. 43:527-550 (2000)

Willson, T. M., Lambert, M. H., and Kliewer, S. A. PPARy and metabolic disease. Annu. Rev. Biochem. 70:341-367 (2001)

Wilton, S., and Skeijanc, I. S. Factors in serum regulate muscle development in P19 cells. In Vitro Cell. Dev. Biol. Anim. 35:175-177 (1999)

Wit,E., and McClure,J. 2004. Statistics for microarrays: design, analysis, and inference. Jone Wiley & Sons Ltd., Chichester, UK.

183 Wobus, A. M., Kleppisch, T., Maltsev, V. A., and Hescheler, J. Cardiomyocyte-like cells differentiated in vitro from embryonic carcinoma cells PI9 are characterized by functional expression of adrenoceptors and Ca2+ channels. In Vitro Cell Dev. R/o/. 30:425-434 (1994)

Wolfram, G. Dietary fatty acids arid coronary heart disease. Eur. J. Med. Res. 8:321-324 (2003)

Xu, H. E., Lambert, M. H., Montana, V. G., Plunket, K. D., Moore, L. B., Collins, J. L., Oplinger, J. A., Kliewer, S. A., Gampe, R. T., McKee, D. D., Moore, J. T., and Willson, T. M. Structural determinants of ligand binding selectivity between the peroxisome proliferator-activated receptors. Proc. Nat. Acad. Sci. U. S. A. 98:13919-13924 (2001)

Yanai, I., Benjamin, H., Shmoish, M., Chalifa-Caspi, V., Shklar, M., Ophir, R., Bar-Even, A., Hom-Saban, S., Safran, M., Domany, E., Lancet, D., and Shmueli, O. Genome-wide midrange transcription profiles reveal expression level relationships in human tissue specification. Bioinformatics. 21:650-659 (2004)

Yaqoob, P. Fatty acids and the immune system: from basic science to clinical applications. Rroc. M/fr. 5'oc. 63:89-104 (2004)

Yaqoob, P. Fatty acids as gatekeepers of immune cell regulation. Trends Immunol. 24:639-645(2003)

Ye, P., Hu, X., and Zhao, Y. The increase in plasminogen activator inhibitor type-1 expression by stimulation of activators for peroxisome proliferator-activated receptors in human endothelial cells. Chin. Med. Sci. J. 17:112-116 (2002)

Young, D. A., Gavrilov, S., Pennington, C. J., Nuttall, R. K., Edwards, D. R., Kitsis, R. N., and Clark, I. M. Expression of metalloproteinases and inhibitors in the differentiation of P19CL6 cells into cardiac myocytes. Biochem. Biophys. Res. Commun. 322:759-765 (2004)

Young, M. E., Goodwin, G. W., Ying, J., Guthrie, P., Wilson, C. R., Laws, F. A., and Taegtmeyer, H. Regulation of cardiac and skeletal muscle malonyl-CoA decarboxylase by fatty acids.Am. J. Physiol. 280:E471-E479 (2001)

Yu, H., Cook, T. J., and Sinko, P. J. Evidence for diminished functional expression of intestinal transporters in Caco-2 cell monolayers at high passages. Pharm. Res. 14:757-762 (1997)

Yu, J., Othmari, M. L, Faijo, R., Zareparsi, S., MacNee, S. P., Yoshida, S., and Swaroop, A. Evaluation and optimization of procedures for target labeling and hybridization of cDNA microarrays. Mol. Vis. 8:130-137 (2002)

Yun, K., and Wold, B. Skeletal muscle determination and differentiation: story of a core regulatory network and its context. Cwr. Opin. Cell Biol. 8:877-889 (1996)

Zanotti, G. Muscle fatty acid-binding protein. Biochim. Biophys. Acta 1441:94-105 (1999)

184 Zhang, J., Fu, M., Zhu, X., Xiao, Y., Mon, Y., Zheng, H., Akinbami, M. A., Wang, Q., and Chen, Y. E. PPAR5 is up-regulated during vascular lesion formation and promotes post-confluent cell proliferation in vascular smooth muscle cells. J. Biol C/zem. 277:11505-11512 (2002)

Zhang, L., and Chawla, A. Role of PPARy in macrophage biology and atherosclerosis. Trends Endocrinol Metab. 15:500-505 (2004)

Zhang, X., and Young, H. A. PPAR and immune system: what do we know. Int. Immunopharmacol I:\029-1044 (2002)

Zhang, Z., Burch, P. E., Cooney, A. J., Lanz, R. B., Pereira, F. A., Wu, J., Gibbs, R. A., Weinstock, G., and Wheeler, D. A. Genomic analysis of the nuclear receptor family: new insights into structure, regulation, and evolution fi-om the rat genome. Genome Res. 14:5^0-590 {200A)

Zhao, A., Yu, J., Lew, J. L., Huang, L., Wright, S. D., and Cui, J. Polyunsaturated fatty acids are FXR ligands and differentially regulate expression of FXR targets. DNA Ce//R/o/. 23:519-526 (2004)

Zimmerman, A. W., van Moerkerk, H. T., and Veerkamp, J. H. Ligand specificity and conformational stability of human fatty acid-binding proteins. Int. J. Biochem. Ce//R/e/. 33:865-876 (2001)

185