ASSESSMENT OF THE ROLE OF PLASMALOGEN IN THE MODULATION OF OXIDATIVE STRESS AND INFLAMMATION IN ATHEROSCLEROSIS
Aliki Anadena Rasmiena
Bachelor of Science (Honours)
This thesis is submitted in total fulfilment of the degree of Doctor of Philosophy
December 2015
Department of Biochemistry and Molecular Biology Faculty of Medicine, Dentistry and Health Sciences University of Melbourne and Metabolomics Laboratory, Baker IDI Heart & Diabetes Institute
i
ABSTRACT
Background: Oxidative stress is a contributing factor to atherosclerosis. Circulating levels of plasmalogens (phospholipids with potential anti-oxidant properties) have been shown to be negatively associated with coronary artery disease, suggesting an elevated level of oxidative stress in these patients. We hypothesised that: (1) oxidative stress affects lipoprotein lipid composition and function; and (2) regulation of the level of plasmalogen can influence atherosclerosis progression and inflammation. Method/results: Low density and high density lipoproteins (LDL and HDL) were oxidised with copper chloride at different time points. Liquid chromatography combined with tandem mass spectrometry analysis showed a myriad of changes in the lipid composition of lipoproteins during oxidation. Plasmalogen was one of the lipids that were most affected, early in the oxidation of lipoproteins. Incubation of THP1- derived macrophages with HDL of differing levels of oxidative stress showed that the capacity of mildly- and heavily oxidised HDL to efflux cholesterols was significantly reduced as compared to native HDL. Similarly, the ability of HDL to delay LDL oxidation and to accept oxidised lipids from oxidised LDL deteriorated progressively under mild- and heavy oxidative stress. To investigate the effect of plasmalogen enrichment in atherosclerosis, ApoE- and ApoE/glutathione peroxidase 1-deficient (ApoE-/- and ApoE-/-GPx1-/-) mice were fed a high-fat diet with or without 2% batyl alcohol (BA, precursor to plasmalogen synthesis) for 12 weeks. Mass spectrometry analysis of lipids in plasma, heart, liver and adipose tissue showed that plasmalogen concentration was increased in all tissues of the BA-treated ApoE-/- and ApoE-/-GPx1-/- mice. Oxidation of plasmalogen in the treated mice was apparent by the increase in sn-2 lysophosphatidylcholine in circulation. En face analysis showed that compared to the untreated mice, aortic plaque accumulation in the BA-treated ApoE-/- and ApoE-/-GPx1- /- mice was significantly reduced (70%). Immunohistochemistry of the aortic sinus and aorta indicated that the levels of the inflammatory marker, VCAM-1 and the oxidative stress marker, nitrotyrosine were reduced only in the BA-treated ApoE-/-GPx1-/- mice. Treatment with BA also resulted in a decrease in the body weight gain and fasting blood glucose without any effect on the fasting insulin level in these mice. Further lipidomic analysis demonstrated that diacyl- and triacylglycerols in the liver were lowered whereas that in the plasma was increased. Flow cytometry analysis of the peripheral
ii whole blood of C57/BL6 mice showed that treatment with an alkylglycerol mix for 12 weeks lowered the levels of total monocytes and neutrophils. Conclusion: Oxidation affected lipids in both the surface layer and core of the lipoproteins and this translated to a deterioration of the lipoprotein function with increasing level of oxidative stress. In addition, the modulation of plasmalogen levels via treatment with alkylglycerol alleviated atherosclerosis in vivo potentially via a plethora of mechanisms involving inflammation and oxidative stress, and the regulation of glucose and body weight. Plasmalogen modulation represents a potential therapy to prevent atherosclerosis and reduce cardiovascular disease risk.
iii
DECLARATION
This thesis comprises only of original work towards the degree of Doctor of Philosophy except where indicated in the preface. Acknowledgements have been made in text to other material that has previously published. This thesis is no more than 100,000 words exclusive of tables, figures, references, and appendices.
Aliki Anadena Rasmiena
December 2015
iv
PREFACE
The following technical work was conducted by individuals in the Baker IDI Heart and Diabetes Institute other than myself. I am sincerely thankful for their time to contribute this work to my thesis:
1) Dr. Judy B. de Haan (Oxidative Stress Laboratory) for providing transgenic ApoE/GPx1-deficient mice (Chapter 5).
2) Mr Annas Al-Sharea (Vascular Pharmacology Laboratory) for performing the mice tail bleed and flow cytometry analysis (Chapter 6).
3) Miss Natalie Mellett (Metabolomics Laboratory) for preparing alkylglycerol mix and the animal technicians at the Baker IDI (Miss Samantha Sacca, Miss Elisha Lastavec, Miss Megan Haillay) for performing the daily gavage of the alkylglycerol mix (Chapter 6).
v
ACKNOWLEDGEMENTS
I am indebted to A/Prof. Peter Meikle, the principal supervisor of my PhD candidature, for first taking me on as his PhD student, and for the guidance and support throughout my candidature. His gift for imparting knowledge on lipidomics has been invaluable to me, as has the opportunities he has given to me to travel to various conferences. It has been instrumental in helping me expand my professional network and to continue to fuel my passion for Science. Furthermore, his calm collectedness and perseverance during challenging times have shown me what it takes to be a leader in the scientific field, and has inspired me to have a future career in science.
I thank Dr. Dedreia Tull, co-supervisor and Prof. Malcolm McConville, chair of my PhD committee, for their invaluable feedback and discussion as well as words of encouragement throughout my candidature.
I thank Dr. Judy de Haan for providing transgenic ApoE/GPx1-deficient mice. I am also grateful for her guidance both in the analysis of immunohistochemistry data and throughout the study. Her optimism during the manuscript submissions and revisions really encouraged us to pull through.
To the past and present members of Metabolomics Laboratory, particularly Dr. Christopher Barlow, I thank him for the time he took to discuss concepts with me, to educate me in the use of complex Excel formulas, which proved to be an instrumental skill for handling large lipidomic data and for the training in mass spectrometry. I also acknowledge Dr. Theodore Wai Ng to inspire me to do PhD in the first place and for providing me training in sequential ultracentrifugation early in my PhD. I thank Kevin Huynh and Ricardo Tan for their technical assistance throughout the animal study and for making animal cull days an enjoyable one. I am grateful for the time Jacqui Weir and Natalie Mellett took to answering my technical questions on mass spectrometry and lipidomics, and for making sure that the laboratory was always neat and in order so I could perform my experiments. Anmar Anwar and Kang Yu Peng, fellow PhD students, warrant special mention for being wonderful office mates. Special thanks to Dr. Husna
vi
Begum, a very good friend out of the lab and practically a surrogate big sister to me. I thank her for her advice and for being there for me through tough times in PhD.
To many staff members in the Baker IDI Heart and Diabetes Institute, particularly Nada Stefanovic and Dr. Arpeeta Sharma, I thank them for providing me training in organ dissection, en face, and immunostaining.
I thank the University of Melbourne and Baker IDI for funding my PhD scholarship and the travel award I have received. Also I thank the organisers of various conferences including EAS, AAS, ASMR and Royal Society of Victoria for giving me the opportunities to present my work.
To student committee members and friends in the Department and Baker IDI, I am grateful to get to know all of them; I thank them for making my candidature a colourful one and for the friendships we have built as we became more matured in our own journeys to completing our PhDs. I am deeply thankful to Kai Lin Giam, a close friend, for her words of wisdom and emotional support throughout our journey together since Honours. Special thanks to Camelia Quek whose cheerful attitude and positive outlook on life is contagious and got me through challenging times in PhD.
I wish to acknowledge the emotional support of my close circle of friends, Joe Lim, Kelvin Yong, Joyce Yong, and Jesse Hon throughout my PhD. I thank them for always making my weekends and breaks from PhD a fun one. Special mention to Shou Farn Chung, meeting him towards the late stages of my PhD has been a blessing; I am grateful for the mini adventures he has introduced to me that taught me that sometimes there is more to life than doing a PhD.
Most importantly, I acknowledge the incredible support and endless love from my parents, Dedy Djubaedi and Ratna Mediawati as well as my sister, Gratia Anadena. I am forever grateful for them for believing in me, their prayers, and for inspiring me to do my best. I dedicate this thesis to my dad who once said to me, …”go out there and be useful to your community”. Also, I dedicate this thesis to my grandparents, Kwik Ping Hoo and Kang Haw Nie who passed away from the very disease I studied. May these studies I conducted during my PhD contribute however small to the advancement of
vii medical science and so there is hope for better treatment and quality of life for our future generation.
viii
PUBLICATIONS
Publications arising from the research conducted for this thesis at the time of submission include the following:
1) A.A. Rasmiena, T.W. Ng, and P.J. Meikle, Metabolomics and ischaemic heart disease. Clinical Science, 2013. 124(5): 289-306.
2) A.A. Rasmiena, C. Barlow, N. Stefanovic, K. Huynh, R. Tan, A. Sharma, D. Tull, J.B. deHaan, P.J. Meikle. Attenuation of atherosclerosis in ApoE- and ApoE/GPx1- deficient mice by plasmalogen modulation. Atherosclerosis, 2015. 243(2): 598-608.
3) A.A. Rasmiena, C. Barlow, T.W. Ng, and P.J. Meikle. High-density lipoproteins transfer surface but not internal oxidised lipids from oxidised low-density lipoproteins. Biochimica et Biophysica Acta - Molecular and Cell Biology of Lipids, 2016. 1861(2): 69-77.
ix
TABLE OF CONTENTS
ABSTRACT…………………………………………………………………………………..…ii
PREFACE…………………………………………………………………...………………..…v
ACKNOWLEDGEMENTS……...……………………………………………….…………....vi
PUBLICATIONS……………………………………………………………………………….ix
LIST OF TABLES…………………………………………………………………………..…xv
LIST OF FIGURES………………………………………………………………………….xvii
ABBREVIATIONS…………………………………………………………………………….xx
1 LITERATURE REVIEW ...... 2 1.1 OVERVIEW OF ISCHAEMIC HEART DISEASE ...... 2 1.1.1 Epidemiology and pathophysiology of ischaemic heart disease ...... 2 1.1.2 Risk factors associated with ischaemic heart disease ...... 3 1.1.2.1 Major risk factors ...... 4 1.1.2.1.1 Low density lipoprotein-cholesterol (LDL-C) ...... 4 1.1.2.1.2 Triglycerides ...... 5 1.1.2.1.3 High density lipoprotein-cholesterol (HDL-C) ...... 6 1.1.2.1.4 Other major risk factors ...... 6 1.1.2.2 Emerging risk factors ...... 8 1.1.3 Primary and secondary prevention of ischaemic heart disease ...... 8 1.1.3.1 Statins ...... 8 1.1.3.2 Omega-3 fatty acids ...... 10 1.1.3.3 Anti-oxidant vitamins ...... 13 1.2 OXIDATIVE STRESS AND INFLAMMATION AS CONTRIBUTING FACTORS TO ATHEROSCLEROSIS ...... 16 1.2.1 Oxidative stress and inflammation in the cardiovascular system ...... 16 1.2.1.1 Major sources of reactive oxygen species ...... 16 1.2.1.1.1 NADPH oxidase ...... 16 1.2.1.1.2 Endothelial nitric oxide synthase ...... 16 1.2.1.1.3 Xanthine oxidase ...... 17 1.2.1.1.4 Myeloperoxidase ...... 17 1.2.1.2 Markers of inflammation ...... 18
x
1.2.1.2.1 Vascular adhesion molecule-1 and intracellular adhesion molecule - 1 ... 18 1.2.1.2.2 Monocyte chemoattractant protein - 1 ...... 19 1.2.2 Impact of oxidative stress and inflammation on cellular function ...... 19 1.2.3 Animal models of atherosclerosis ...... 20 1.3 INTRODUCTION TO LIPOPROTEIN METABOLISM ...... 21 1.3.1 Lipoprotein composition ...... 22 1.3.1.1 Apolipoproteins associated with lipoproteins ...... 22 1.3.1.1.1 Apolipoprotein A ...... 25 1.3.1.1.2 Apolipoprotein B ...... 25 1.3.1.1.3 Apolipoprotein E ...... 26 1.3.1.2 Enzymes and lipid transfer proteins associated with lipoproteins ...... 27 1.3.1.2.1 Plateler-activating factor acetylhydrolase ...... 27 1.3.1.2.2 Lechithin:cholesterol acetyltransferase ...... 28 1.3.1.2.3 Paroxonase-1 ...... 29 1.3.1.2.4 Phospholipid transfer protein ...... 29 1.3.1.2.5 Cholesteryl ester transfer protein ...... 30 1.3.1.3 Lipid composition of lipoproteins ...... 32 1.3.2 Overview of lipoprotein metabolism and functions ...... 34 1.3.2.1 Chylomicron and very low density lipoprotein ...... 34 1.3.2.2 Low density lipoprotein ...... 34 1.3.2.3 High density lipoprotein ...... 35 1.4 PLASMALOGENS: BIOCHEMISTRY, DISEASE ASSOCIATIONS AND PLASMALOGEN MODULATION ...... 38 1.4.1 Structure and oxidation chemistry of plasmalogens ...... 38 1.4.2 Biosynthestic pathway of plasmalogen and the lipid distribution in cells and tissues ...... 40 1.4.3 Biological functions of plasmalogens ...... 44 1.4.3.1 Plasmalogens as membrane components ...... 44 1.4.3.2 Plasmalogen as a potential anti-oxidant ...... 45 1.4.3.3 Plasmalogens as lipid mediators...... 45 1.4.4 Association of plasmalogen to cardiovascular disease ...... 46 1.4.5 Modulation of plasmalogen in vitro and in vivo ...... 47 1.5 LIPIDOMICS AS AN APPROACH TO STUDY THE PATHOGENESIS OF ISCHEAMIC HEART DISEASE ...... 48 1.5.1 Lipidomics ...... 48 1.5.2 Lipidomics in studies of ischaemic heart disease ...... 50
xi
1.5.2.1 Lipidomics of inflammation and oxidative stress ...... 50 1.5.2.2 Lipidomic assessment of atherosclerotic plaque ...... 53 1.5.2.3 Lipidomics and lipid metabolism ...... 54 1.6 HYPOTHESES AND STUDY OBJECTIVES ...... 57 1.6.1 Statement of research questions ...... 57 1.6.2 Hypotheses and study aims ...... 58 2 MATERIALS AND GENERAL METHODS ...... 61 2.1 GENERAL CHEMICALS ...... 61 2.2 GENERAL METHODS ...... 62 2.2.1 Isolation and handling of lipoproteins ...... 62 2.2.1.1 Isolation of blood plasma ...... 62 2.2.1.2 Fractionation of lipoproteins from human plasma ...... 62 2.2.2 Lipoprotein oxidation and measurement of oxidation ...... 63 2.2.2.1 Lipoprotein oxidation ...... 63 2.2.2.2 Conjugated diene measurement ...... 64 2.2.2.3 Thiobarbituric acid reactive substances (TBARS) assay ...... 64 2.2.3 Cell culture ...... 64 2.2.4 Bicinchoninic acid (BCA) protein assay ...... 65 2.2.5 Extraction and analysis of lipids ...... 65 2.2.5.1 Lipid extraction ...... 65 2.2.5.2 Liquid chromatography electrospray ionisation tandem mass spectrometry .. 66 2.2.5.3 Lipid analysis ...... 67 2.2.6 Mouse models of atherosclerosis ...... 68 2.2.7 Preparation of tissues for lipidomic analysis ...... 68 2.2.8 Statistical analyses ...... 69 3 CHARACTERISATION OF CHANGES IN THE LIPID COMPOSITION OF LIPOPROTEINS DUE TO OXIDATIVE STRESS ...... 73 3.1 ABSTRACT ...... 73 3.2 INTRODUCTION ...... 74 3.3 METHODS ...... 75 3.3.1 Lipoprotein oxidation ...... 75 3.3.2 Lipid analysis ...... 75 3.3.3 Identification of oxPC and oxCE by untargeted lipidomics ...... 76 3.4 RESULTS ...... 76 3.4.1 Conjugated diene and TBARS production in oxidised LDL and HDL ...... 76
xii
3.4.2 Changes in lipid class composition associated with the outer layer of lipoproteins following oxidation ...... 77 3.4.3 Changes in lipid species associated with the outer layer of lipoproteins following oxidation...... 84 3.4.4 Changes in the composition of lipid classes associated with the core of lipoproteins...... 91 3.4.5 Analyses of molecular species of lipids associated with the core of lipoproteins 92 3.4.6 Identification of major oxPC and oxCE species in oxidised LDL ...... 94 3.4.7 Oxidised lipids in LDL and HDL ...... 95 3.5 DISCUSSION ...... 103 3.6 LIMITATIONS ...... 107 4 CHARACTERISATION OF HDL FUNCTION FOLLOWING OXIDATIVE STRESS ...... 110 4.1 ABSTRACT ...... 110 4.2 INTRODUCTION ...... 111 4.3 METHODS ...... 112 4.3.1 Cholesterol efflux assay ...... 112 4.3.2 Assessment of the ability of HDL to delay LDL oxidation ...... 114 4.3.3 Assessment of the ability of HDL to accept oxidised lipids from oxLDL ...... 115 4.3.4 Data analysis and statistics ...... 116 4.4 RESULTS ...... 117 4.4.1 Oxidation of HDL led to impaired cholesterol efflux ability ...... 117 4.4.2 The ability of HDL to protect LDL against oxidation was dependent on its oxidative state ...... 118 4.4.3 HDL acceptance of oxidised lipids from oxLDL ...... 119 4.4.3.1 Oxidised phosphatidylcholine (oxPC) ...... 120 4.4.3.2 Oxidised cholesteryl ester (oxCE) ...... 120 4.4.3.3 7-ketocholesterol and 7β-hydroxycholesterol ...... 123 4.4.3.4 Transfer of lysophosphatidylcholine from oxLDL to HDL ...... 123 4.4.3.5 Sphingomyelin to phosphatidylcholine ratio in native and oxidised HDL ... 128 4.4.3.6 Phosphatidylserine in the native and oxidised HDL ...... 128 4.5 DISCUSSION ...... 132 4.6 LIMITATIONS ...... 138 5 PLASMALOGEN MODULATION ATTENUATES ATHEROSCLEROSIS IN APOE- AND APOE/GPX1-DEFICIENT MICE…………………………………..….141 Addendum……………………………………………………………………………..………152 Supplementary material…………...... ………………………...……………....……………..154
xiii
6 EFFECT OF PLASMALOGEN MODULATION ON TISSUE LIPID METABOLISM AND INFLAMMATION…………………..…...... …………….170 6.1 ABSTRACT………………………………………………………………………...170 6.2 INTRODUCTION…………………………………………………………………..171 6.3 METHODS………………………………………………………………………….172 6.3.1 Animal groups and diet study…………………………………………………172 6.3.2 Tissue homogenisation and lipid analysis………………………………….…173 6.3.3 Flow cytomtery analysis………………………………………………………174 6.3.4 Statistics……………………………………………………………………….174 6.4 RESULTS…………………………………………………………………………...175 6.4.1 Effects of plasmalogen modulation in plasma and liver………………………175 6.4.2 Effects of plasmalogen modulation in adipose tissue…………………………179 6.4.3 Plasmalogen modulation via alkylglycerol mix treatment lowered monocytes and neutrophils levels in vivo………………………………………………...180 6.5 DISCUSSION………………………………………………………………………185 6.6 LIMITATIONS……………………………………………………………………..189 7 GENERAL DISCUSSION………………………………………...... ………………….192 7.1 Oxidation leads to altered lipid composition and function in lipoproteins…………192 7.2 Plasmalogen modulation attenuates oxidative stress and inflammation in atherosclerosis……………….………………………………………………………194 7.3 Potential role of plasmaloge metabolites….………………………………………..195 7.4 Conclusion and future directions…………………………………………………...196 8 REFERENCES………………………………………………………………………….198 9 APPENDICES…………………………………………………………………………...228 (Supplementary Table 2.1 to 6.3 and Supplementary Figure 3.1 to 4.1)……………...…228- 265
High density lipoprotein efficiently accepts but surface but not internal oxidised lipids from oxidised low density lipoprotein………………………………………………………………266
xiv
LIST OF TABLES
Chapter 1 Table 1.1 Major and emerging risk factors based on the US National 4 Cholesterol Education Program, Adult Treatment Panel III Table 1.2 Clinical characteristics of individuals at risk of ischaemic 7 heart disease Table 1.3 Studies on statins and their findings 10 Table 1.4 Studies on omega-3 fatty acids and their findings 12 Table 1.5 Studies on anti-oxidant vitamins and their principal findings 15 Table 1.6 Characteristics of the lipoprotein family 24 Table 1.7 Characteristics of major apolipoproteins, enzymes, and 31 transfer proteins of lipoprotein metabolism. Table 1.8 Lipid composition of lipoproteins in healthy individuals. 33 Table 1.9 Cell and tissue distribution of plasmalogens 44 Table 1.10 Major findings from lipidomic studies of ischaemic heart 56 disease
Chapter 2 Table 2.1 General chemicals 61 Table 2.2 Preparation of density solutions for sequential 63 ultracentrifugation Table 2.3 Internal standards used in lipid extraction and mass 66 spectrometry analysis. Table 2.4 Conditions for liquid chromatography electrospray ionisation 70 tandem mass spectrometry analysis of lipid species.
Chapter 3 Table 3.1 Comparison of the observed and exact masses for the oxidised 97 lipid species identified in the untargeted LC-MS analysis. Table 3.2 The effect of oxidation on the level of oxidised 99 phosphatidylcholine in low density lipoprotein Table 3.3 The effect of oxidation on the level of oxidised cholesteryl 100 ester in low density lipoprotein Table 3.4 The effect of oxidation on the level of oxidised 101 phosphatidylcholine in high density lipoprotein Table 3.5 The effect of oxidation on the level of oxidised cholesteryl 102 ester in high density lipoprotein
Chapter 4 Table 4.1 Changes in the level of oxidised lipids in the re-isolated LDL 126 after co-incubation Table 4.2 Changes in the levels of oxidised lipids in re-isolated HDL 127 with or without co-incubation with oxLDL
Chapter 5 Table 1 Plasma lipid measurements of mice on a high fat diet 144 supplemented with/without 2% BA
xv
Chapter 6 Table 6.1 Differences in the level of plasma lipids in the mice treated 177 with/without 2% batyl alcohol Table 6.2 Differences in the level of hepatic lipids in the mice treated 178 with/without 2% batyl alcohol Table 6.3 Differences in the level of lipids in adipose tissue in mice 182 treated with/without 2% batyl alcohol
xvi
LIST OF FIGURES
Chapter 1 Figure 1.1 Progression of atherosclerosis 3 Figure 1.2 Schematic representation of a lipoprotein structure 23 Figure 1.3 Schematic representation of lipoprotein metabolism 37 Figure 1.4 Subclasses of glycerophosphoethanolamine 39 Figure 1.5 Proposed radical reactions of plasmalogen 40 Figure 1.6 Biosynthetic pathway of plasmalogens 43 Figure 1.7 Typical workflow of lipidomic studies 50
Chapter 3 Figure 3.1 Conjugated diene and TBARS formation in oxidised low 79 density and high density lipoproteins. Figure 3.2 The effect of oxidation on alkyl- and alkenylphospholipids 80 associated with outer layer of low-density lipoproteins Figure 3.3 The effect of oxidation on alkyl- and alkenylphospholipids 81 associated with outer layer of high-density lipoproteins Figure 3.4 The effect of oxidation on lysophosphatidylcholine in low 82 density lipoproteins Figure 3.5 The effect of oxidation on lysophosphatidylcholine in high 83 density lipoproteins Figure 3.6 The effect of oxidation on sphingomyelin and free 84 cholesterol associated with outer layer of lipoproteins Figure 3.7 The effect of oxidation on alkyl- and alkenylphospholipids 87 containing monounsaturated- and polyunsaturated fatty acids in the outer layer of low-density lipoproteins. Figure 3.8 The effect of oxidation on alkyl- and alkenylphospholipids 88 containing monounsaturated- and polyunsaturated fatty acids in the outer layer of high-density lipoproteins. Figure 3.9 The effect of oxidation on lysophosphatidylcholine with 89 saturated- and polyunsaturated fatty acids in low density lipoprotein. Figure 3.10 The effect of oxidation on lysophosphatidylcholine with 90 saturated- and polyunsaturated fatty acids in high density lipoprotein. Figure 3.11 The effect of oxidation on sphingomyelin containing 91 monounsaturated- and polyunsaturated fatty acids in the outer layer of lipoproteins Figure 3.12 The effect of oxidation on lipids associated with the core of 93 the lipoproteins Figure 3.13 The effect of oxidation on triacylglycerol species containing 94 monounsaturated and polyunsaturated fatty acids Figure 3.14 Untargeted lipidomic analysis of native and oxidised low 96 density lipoprotein Figure 3.15 Extracted ion chromatogram (m/z = 772.5440 - 772.5540) of 98 the oxidised lipid corresponding to PC(34:3(O)) in low density lipoprotein. Figure 3.16 Production of oxidised phosphatidylcholine and cholesteryl 98
xvii
esters in oxidised low density- and high density lipoproteins
Chapter 4 Figure 4.1 Cholesterol efflux capacity of native and oxidised HDL 117 Figure 4.2 Plasmalogen and oxidised lipids in the re-isolated HDL 121 following co-oxidation with LDL. Figure 4.3 Plasmalogen and oxidised lipids in the re-isolated LDL 122 following co-oxidation with HDL. Figure 4.4 Effective transfer of oxidised PC, but not oxidised CE from 125 oxidised LDL by HDL. Figure 4.5 Transfer of lysophosphatidylcholine from oxLDL to native 129 and oxHDL. Figure 4.6 Levels of isomers of lysophosphatidylcholine in LDL and 130 HDL following co-incubations. Figure 4.7 Level of sphingomyelin relative to total phosphatidylcholine 131 as a contributing factor to the ability of HDL to transfer oxidised lipids. Figure 4.8 Phosphatidylserine in the native and oxidised HDL. 131
Chapter 5 Figure 1 Level of alkyl- and alkenylphospholipids in plasma and heart 145 of mice following 12 weeks of high fat diet supplemented with/without batyl alcohol Figure 2 Level of lysophosphatidylcholine and 146 lysophosphatidylethanolamine in plasma and heart of mice following 12 weeks of high fat diet supplemented with/without batyl alcohol Figure 3 Atherosclerotic plaque in the aorta of mice fed a high fat diet 147 supplemented with/without batyl alcohol. Figure 4 Aortic lesion and inflammation as well as oxidative stress in 148 mice fed a high fat diet supplemented with/without batyl alcohol. Figure 2- Level of lysophosphatidylcholine and 153 Addendum lysophosphatidylethanolamine in plasma and heart of mice following 12 weeks of high fat diet supplemented with/without batyl alcohol
Chapter 6 Figure 6.1 Level of hepatic alkyl- and alkenylphospholipids in mice 176 following 12 weeks of a high fat diet supplemented with/without batyl alcohol Figure 6.2 Level of alkyl- and alkenylphospholipids in adipose tissue of 180 mice following 12 weeks of a high fat diet supplemented with/without batyl alcohol Figure 6.3 Weekly body weights of C57/BL6 mice orally treated with 181 lecithin carrier or alkylglycerol mix for 12 weeks Figure 6.4 Plasma level of alkyl- and alkenylphospholipids in mice 183 following 12 weeks of oral treatment of lecithin carrier or alkylglycerol mix
xviii
Figure 6.5 Gating strategies for the analysis of monocytes and 184 neutrophils using flow cytometry Figure 6.6 Level of monocytes and neutrophils of peripheral whole 185 blood of C57/BL6 mice treated with/without alkylglycerol mix
xix
ABBREVIATIONS
AADHAP-R Acyl/alkyl dihydroacetone phosphate -reductase AAG3P-AT Acyl/alkyl-glycero-3-phosphate-acyltransferase ABCA1 ATP-binding cassette transporter A1 ABCG1 ATP-binding cassette sub family G member 1 transporter ACS Acute coronary syndrome AGE Advanced glycation end products ALA α-linoleic acid ADHAP-S Alkyl dihydroacetone phosphate synthase ApoAI Apolipoprotein AI ApoE Apolipoprotein E ApoE-/- Apolipoprotein E-deficient mouse ApoE-/-GPx1-/- Apolipoprotein E- and glutathione peroxidase-1-deficient mouse BMI Body mass index CAD Coronary artery disease CE Cholesteryl ester Cer Ceramide CETP Cholesteryl ester transfer protein CID Collision-induced dissociation COH Free cholesterol C-PT Choline phosphotransferase CVD Cardiovascular disease CVE Cardivascular event DALY Disability adjusted life years DG Diacylglycerol DHA Docosahexaenoic acid DHAP Dihydroacetone phosphate DHAP-AT Dihydroacetone phosphate acyltransferase DHC Dihexosylceramide dhCer Dihydroceramide eNOS Endothelial nitric oxide synthase EPA Eicosapentaenoic acid E-PT Ethanolamine phosphotransferase ESI-MS/MS Electrospray ionisation tandem mass spectrometry GM3 GM3 ganglioside GPx1 Glutathione peroxidase-1 HDL High density lipoprotein HDL-C High density lipoprotein-cholesterol HUVEC Human umbilical vascular endothelial cells ICAM-1 Intracellular adhesion molecule-1 IDL Intermediate density lipoprotein IHD Ischaemic heart disease LC Liquid chromatography LCAT Lecithin:cholesterol acyltransferase LDL Low density lipoprotein
xx
LDL-C Low density lipoprotein-cholesterol LDL-R Low density lipoprotein receptor Lp-PLA2 Lipoprotein-associated phospholipase A2 LPC Lysophosphatidylcholine LPE Lysophosphatidylethanolamine MAPK Mitogen-activated protein kinases MCP-1 Monocyte chemoattractant protein-1 MDA Malondialdehyde MHC Monohexosylceramide MS Mass spectrometry MUFA Monounsaturated fatty acids n-3 FA Omega-3 fatty acid NADPH Nicotinamide adenine dinucleotide phosphate NCEP The US National Cholesterol Education Program NCEP-ATP III The US National Cholesterol Education Program, Adult Treatment Panel III NF-KB Nuclear factor kappa-light-chain-enhancer of activated B cells NO Nitric oxide NOX Nicotinamide adenine dinucleotide phosphate oxidase oxCE Oxidised cholesteryl ester oxHDL Oxidised high density lipoprotein oxLDL Oxidised low density lipoprotein oxPC Oxidised phosphatidylcholine PAH Platelet-activating factor PC Phosphatidylcholine PC(O) Alkylphosphatidylcholine PC(P) Alkenylphosphatidylcholine or phosphatidylcholine plasmalogen PE Phosphatidylethanolamine PE(O) Alkylphosphatidylethanolamine PE(P) Alkenylphosphatidylethanolamine or phosphatidylethanolamine plasmalogen PH Phosphohydrolase PI Phosphatidylinositol PLC Phospholipase C PLTP Phospholipid transfer protein PON-1 Paraoxonase-1 PS Phosphatidylserine PUFA Polyunsaturated fatty acid ROS Reactive oxygen species SFA Saturated fatty acids SFA+MUFA CE Cholesteryl ester containing saturated and monounsaturated fatty acids SFA+MUFA PC Phosphatidylcholine containing saturated and monounsaturated fatty acids SM Sphingomyelin SM/PC Ratio of sphingomyelin to phosphatidylcholine
xxi
sn-1 LPC Lysophosphatidylcholine with fatty acid at the sn-1 position sn-2 LPC Lysophosphatidylcholine with fatty acid at the sn-2 position SRB-1 Scavenger receptor class B member 1 TBARS Thiobarbituric acid reactive substances TG Triacylglycerol or triglyceride THC Trihexosylceramide VCAM-1 Vascular adhesion molecule-1 VLDL Very low density lipoprotein
xxii
CHAPTER 1 – LITERATURE REVIEW
1
CHAPTER 1 – LITERATURE REVIEW
1 LITERATURE REVIEW
1.1 OVERVIEW OF ISCHAEMIC HEART DISEASE
1.1.1 Epidemiology and pathophysiology of ischaemic heart disease
Ischaemic heart disease (IHD) is a major contributor to global mortality, morbidity as well as to the burden of cardiovascular disease (CVD). Of the 57 million global deaths in 2008, approximately 30% (17.3 million deaths) were attributable to CVD and 42.5% of the CVD deaths were caused by IHD [1]. Disability adjusted life years (DALY) is a measure of overall disease burden, expressed as the number of years lost due to ill-health, disability or premature death. The World Health Organisation estimates a global increase in IHD burden from 47 million DALY in 1990 to 82 million DALY by 2020 [2]. In Australia, 34% (48,456) of all deaths were caused by CVD in 2008 and IHD contributed to 49% of these deaths [3]. Economic and healthcare costs in both the prevention and treatment of heart disease are substantial (CVD costs the Australian health system approximately $6 billion per year [4]) and constitutes a major concern for public health policy.
Atherosclerosis underscores the manifestation of IHD. It is characterised by the narrowing of the vessel lumen and compensatory enlargement of coronary arteries due to an accumulation of atheromatous plaque within the intima of the arterial walls [5]. The pathogenesis of atherosclerosis is illustrated in Figure 1.1. Atherosclerosis begins to develop early in life and progresses with time. However, the rate of progression is, to a large extent, unpredictable and differs markedly among seemingly comparable individuals. One of the early events leading to atherosclerosis is the formation of “fatty streaks” within the intima of arterial walls [6]. Endothelial dysfunction causes changes to the permeability of the endothelial cells such that it allows high levels of low density lipoprotein (LDL) particles, which function as major cholesterol transporter, to enter and accumulate in the arterial walls [6]. A micro-environment of free radicals in the intima causes oxidation of the LDL particles; This increases the vessel permeability to monocytes which then travel to the sub-endothelial space and engulf the oxidised LDL [6]. Fatty streaks are typically characterised by deposits of monocytes, macrophages, foam cells and lipids within the intima. Some, but not all, fatty streaks progress into fibrolipid plaques, which are distinguished by the presence of vascular smooth muscle cells and increased extracellular fibres within the intima. These sub-clinical
2
CHAPTER 1 – LITERATURE REVIEW
atherosclerotic lesions can progress to become advanced complex plaques via plaque necrosis, which is formed by a combination of macrophage apoptosis and dysfunctional phagocytic clearance of the apoptotic cells [7]. Complex plaques can become unstable as a result of thinning of the protective fibrous cap and smooth muscle cell layer over the plaque by the action of collagenases such as matrix metalloproteinase I [5, 8]. Unstable plaques may rupture, leading to thrombosis, myocardial infarction and stroke with the associated morbidity and mortality.
Figure 1.1 The progression of atherosclerosis. Endothelial dysfunction causes changes to the permeability of the endothelial layer such that it allows low density lipoproteins (LDL) to enter, accumulate, and become oxidised by free radicals in the intimal layer (1). Subsequently, monocytes adhere to and infiltrate the endothelial layer (2), and differentiate to macrophages that engulf the oxidised LDL, forming foam cells (3). Fibroblasts and vascular smooth muscle cells migrate to the sub-endothelial space and proliferate (4), forming a fibrolipid plaque. Over time, the lack of phagocytic clearance as well as macrophage apoptosis result in the formation of necrotic plaque covered with a fibrous cap (5). The thinning of the protective fibrous cap as a result of the activation of matrix metalloproteinase can result in plaque rupture and thrombosis.
1.1.2 Risk factors associated with ischaemic heart disease
The etiology of IHD is complex and still not fully understood. It consists of multiple factors that may interact with one another. Previous studies have identified risk factors that are associated with the disease. The US National Cholesterol Education Program, Adult Treatment Panel III (NCEP-ATP III) [9] classified the risk factors into
3
CHAPTER 1 – LITERATURE REVIEW
major, and emerging risk factors, and they are subdivided into lipid and non-lipid risk factors. These risk factors are summarised in Table 1.1. The discussion below focuses on the principal components, which are relevant to this project.
Table 1.1 Major and emerging risk factors based on the US National Cholesterol Education Program, Adult Treatment Panel III.
Major risk factors Emerging risk factors Lipid LDL-cholesterol Lipoprotein remnants Triglyceride Lipoprotein (a) HDL-cholesterol Small LDL particles Atherogenic dyslipidemia HDL subspecies Apolipoproteins Total cholesterol/HDL-C ratio Non-lipid Hypertension Homocysteine Diabetes Thrombogenic/hemostatic factors Overweight/obesity Inflammatory markers Atherogenic diet Impaired fasting glucose Cigarette smoking Physical inactivity Age Gender Family history of premature IHD
1.1.2.1 Major risk factors
The major risk factors are cardinal features of IHD; they were identified to have strong causal/inverse relationships with IHD and therefore were incorporated into current risk score assessment systems in clinical practice as diagnostic and prediction tools.
1.1.2.1.1 Low density lipoprotein-cholesterol (LDL-C)
Low density lipoprotein-cholesterol (LDL-C) level above 4.1 mmol/l (160 mg/dl) is considered atherogenic and is a hallmark feature in patients at high risk of IHD (Table 1.2). Observational studies on several populations have indicated a log- linear relationship between serum total cholesterol and the risk of IHD [10, 11]. Serum total cholesterol is often a good indicator of LDL-C. Using an algorithmic model based on the US National Cholesterol Education Program (NCEP)’s LDL and total cholesterol
4
CHAPTER 1 – LITERATURE REVIEW
categories, the levels of the aforementioned lipids are directly related to the new-onset risk of IHD of both males and females who were initially without the disease [12]. Such an association is further strengthened in studies of individuals with familial hypercholesterolemia. Familial hypercholesterolemia is an autosomal dominant genetic disorder characterised with elevated levels of LDL-C commonly due to mutations in the LDL-receptor genes. As a result, individuals with familial hypercholesterolemia are predisposed to premature IHD, even in the absence of other risk factors [13, 14].
1.1.2.1.2 Triglycerides
Early studies did not support a causal relationship of triglycerides and IHD primarily because triglycerides are linked intrinsically to cholesterol through mutual lipoprotein carriers as part of lipoprotein metabolism [15]. Thus, the elevation of serum triglycerides may be confounded by the increase in total cholesterol and/or LDL-C levels and therefore triglyceride was not seen as an independent risk factor for IHD. Interest in elevated triglycerides was renewed in prospective epidemiological studies [16, 17] of mostly western populations. In these studies, the link of fasting triglycerides and IHD was significant, even after adjustment for other risk factors. Prospective studies [18, 19] on non-fasting (postprandial) triglycerides in response to normal food intake found a significant correlation of the aforementioned lipid to IHD [20], thus providing a clarification to the pre-existing controversy on the atherogenic potential of fasting and postprandial elevated triglycerides. Other epidemiological evidence of remnant like particles in IHD studies [21, 22] suggested that some lipoprotein remnants such as small very low density lipoproteins (VLDL) and intermediate density lipoproteins (IDL) are triglyceride rich and potentially atherogenic [23]. The measurements for remnant like particles alone were not suggested to add prognostic value to triglycerides because their correlations overlapped [24] and so it is not the level of triglyceride per se that is indicative of the risk of IHD, but rather the level of triglyceride rich lipoproteins particularly the remnants like particles. Lipid measurement of triglycerides is included in the risk assessment systems by NCEP, with triglycerides levels more than 2.3 mmol/l (200 mg/dl) being considered as hypertriglyceridemia and at increased risk of IHD (Table 1.2).
5
CHAPTER 1 – LITERATURE REVIEW
1.1.2.1.3 High density lipoprotein-cholesterol (HDL-C)
Studies have shown strong epidemiological evidences of low level of high density lipoprotein-cholesterol (HDL-C) as an independent risk factor for IHD, even after correction for other risks [25, 26]. A low level of HDL-C may be a sign of insulin resistance and other associated metabolic risk factors [27]. In fact, low HDL-C level is often observed with elevated triglycerides and small LDL particles [28-30] such that the three lipid abnormalities are called the lipid triad, a characteristic of atherogenic dyslipidemia. Based on NCEP-ATPIII, HDL-C below 1 mmol/l (40 mg/dl) is considered low and constitutes a risk factor of IHD (Table 1.2).
The mechanistic details of low HDL-C level and the pathogenesis of atherosclerosis are not fully elucidated, though many studies [31-33] have suggested impaired high density lipoprotein (HDL) anti-oxidative and anti-inflammatory roles. The level of plasma HDL-C did not always reflect the lipoprotein function as demonstrated in animal, genetic and population studies [34]. In addition, a recent human trial (ACCELERATE) on lipid modulating and cholesteryl ester transfer protein inhibitor agent, Evacetrapib, [35] failed to show sufficient efficacy in the improvement of patient risk to cardiovascular event (CVE) [36], thus highlighting a dissociation in the current therapies between improvements in the HDL function and HDL-C levels.
1.1.2.1.4 Other major risk factors
The other major and non-lipid risk factors of IHD include obesity, diabetes, age and gender. Below we discuss these components very briefly.
Obesity is defined as a body mass index (BMI - weight in kg divided by square of height in metres) of >30kg/m2 (Table 1.2) [37]. Studies have confirmed visceral adiposity as a major predictor of type II diabetes [38] as well as IHD [39]. Diabetes is defined as a state of hyperglycemia with a fasting blood glucose level of >7mmol/l (126 mg/dl) [40] (Table 1.2). Large prospective epidemiological studies [41, 42] showed that the risk of IHD was increased up to three-fold among diabetics compared to non- diabetics, after adjustment for other risk factors. Hyperglycemia in diabetes leads to the modification of macromolecules and the formation of advanced glycation end products (AGE), as a result of non-enzymatic addition of carbohydrates to proteins [43]. The binding of AGE to receptors for AGE as well as the glycation of transcription factors
6
CHAPTER 1 – LITERATURE REVIEW
result in endothelial dsyfunction [44]. This is accompanied by the augmentation of inflammatory pathways and the production of cytokines and reactive oxygen species (ROS) in the vascular endothelial cells that further deteriorates cellular function [44, 45]. The endothelial dysfunction promotes leukocyte adhesion, infiltration of monocytes, and the accumulation of LDL in the arterial intima as discussed in section 1.1.1. Furthermore, AGE mediates the modification of extracellular matrix by increasing the matrix volume and cross linking of matrix proteins, and reducing the matrix flexibility [44]. AGE also reduces the activity of matrix metalloproteinase [46], leading to negative vascular remodelling and calcification which are characteristics of stable plaque [44]. Although the independence of diabetes and obesity as IHD risk factors is not clear, the NCEP categorises them as separate risk factors because of the substantial line of evidence linking diabetes and obesity to IHD.
Age is an absolute and non-modifiable risk factor of IHD; It is a reflection of the accumulative exposure to known and unknown IHD risk factors, that in turn reflects the cumulative progression of atherosclerosis [9]. Age and gender play important roles in contributing to IHD risk. Population based studies showed that men were at higher risk of developing IHD as compared to women [47, 48], and the differences in the risk between the genders could not be explained entirely by the standard lipid risk factors such as LDL-C, HDL-C, and triglycerides [49]. With increasing age, the IHD risk for both men and women is elevated, but the risk attributed to sex difference alone is markedly diminished [48].
Table 1.2 Clinical characteristics of individuals at risk of ischaemic heart disease.
Risk factors Status LDL-cholesterol >4.1 mmol/l (160 mg/dl) Triglycerides >2.3 mmol/l (200 mg/dl) HDL-cholesterol <1.0 mmol/l (40 mg/dl) Blood pressure ≥ 140/90 mmHg or on antihypertensive medication Body mass index ≥ 30 kg/m2 Diabetes Type I or type II diabetes, without any glycemic control Cigarette smoking Current Age and gender Male: ≥45 years; Female: ≥ 55 years Family history of Myocardial infarction or sudden IHD death before 55 years of premature CVD age in father and 65 years of age in mother or other male/female first-degree relative
7
CHAPTER 1 – LITERATURE REVIEW
1.1.2.2 Emerging risk factors
Unknown risk factors that influence the major risk factors and cumulatively contribute to increasing IHD risk are yet to be identified. In recent years, potential risk factors have been identified through various intervention and epidemiological studies [50-52] hence they are termed "emerging risk factors". However, the absolute risk of IHD imparted by these factors is yet to be validated. These emerging risk factors may serve to explain the residual risk that cannot be accounted for by the major risk factors. It is probable that the emerging risk factors improve the prediction of IHD in individuals. The current state of the emerging risk factors, however, do not support their use in research as the standardised measurements for the risk factors mainly because of their costs and unavailability in clinical practice [9].
1.1.3 Primary and secondary prevention of ischaemic heart disease
Some of the most studied agents for primary and secondary prevention of IHD include statins, omega-3 fatty acids, and anti-oxidant vitamins.
1.1.3.1 Statins
Statins (HMG CoA inhibitors) are the most effective drugs for the treatment of hypercholesterolemia particularly as a result of increased circulating LDL-C. They are the most commonly prescribed agents for IHD due to their tolerability and potency in lowering LDL-C [53]. The currently available statins include atorvastatin, fluvastatin, pravastatin (administered as acid form), lovastatin, and simvastatin (administered as inactive form or lactone), and rosuvastatin [54].
HMG CoA reductase is the enzyme involved in converting HMG-CoA into mevalonic acid which is a cholesterol precursor. Statins inhibit the action of HMC-CoA reductase in hepatocytes in the liver, which produce the majority of cholesterol in the body. Statins are specific, high affinity, competitive inhibitors for the active site of HMG-CoA reductase; they reversibly bind as well as alter the conformation of the enzyme [55]. Subsequently, the reduction in the intracellular cholesterol is suggested to induce the gene expression of LDL receptors, which decrease the level of circulating LDL and its precursors (VLDL and IDL) [56]. Statins were also shown to potentially alter the mass and activity of apolipoproteins and enzymes associated with the lipoprotein metabolism
8
CHAPTER 1 – LITERATURE REVIEW
[57, 58]. Studies into the action of statins on inflammatory pathways have shown a reduction in the migration of neutrophils towards the chemoattractant fMLP [59] and a decrease in the absolute and relative number of circulating endothelial progenitor cells in treated IHD patients [60]. In addition, statins appeared to reduce the levels of C- reactive protein, which is a marker of inflammation [61]. Analysis of plasma lipid composition of the RADAR study participants (n=80) revealed that the administration of different types of statins lead to differences in lipid metabolism; Rosuvastatin increased, whereas atorvastatin decreased the plasma concentration of phosphatidylcholine [62]. In addition, both statins reduced the concentration of plasma sphingomyelin [62]. Overall, rosuvastatin was found to be more effective in lowering the ratio of plasma sphingomyelin-to-phosphatidylcholine (SM/[SM+PC]), a marker of atherogenesis, as compared to atorvastatin [62]. Although the mechanistic details that underscore the discordance in the ratio (SM/[SM+PC]) are unknown, they may reflect differences in the clinical outcomes of these statins [62].
In recent decades, clinical trials such as IDEAL, ASTEROID and RADAR have clearly demonstrated that statin is the most efficacious therapy for IHD (Table 1.3). However, statin only reduces approximately 30 % of the plaque burden, and even prolonged or high doses of statin will not eliminate atherosclerosis completely [5]. Moreover, differences in the biochemistry, pharmacokinetics, and efficacy of the various types of statins led to questions of their mechanisms and therapeutic equivalence. Statins have also been reported to cause muscle toxicity and elevation of creatine kinase, albeit at a low incidence level [53]. Thus although statins have been one of the most successful interventions for IHD, statins alone are not the entire answer to this burgeoning health problem. In addition to the need to understand the residual risk of atherosclerosis, new therapies are also required to further reduce this residual risk.
9
CHAPTER 1 – LITERATURE REVIEW
Table 1.3 Studies on statins and their findings.
Name of study Main findings References
ESTABLISH Early atorvastatin treatment (20 mg/day) reduced LDL-C [63] level and plaque volume in treated patients with acute coronary syndrome as compared to controls. MRC/BHF Simvastatin (40 mg/day) reduced 22 – 33% of CVE risk for [64] Heart diabetic individuals with, and without diagnosed occlusive Protection arterial disease, or with pre-treatment of LDL-C. REVERSAL Intensive lipid lowering regimen (80 mg/day of [65] atorvastatin) reduced percentage change in atheroma volume compared to moderate lipid lowering group (40 mg/day of pravastatin). Both regiments reduced CRP levels. PROSPER Pravastatin (40 mg/day) reduced coronary mortality in [66] elderly males and females by 24 %. IDEAL Patients with history of acute MI and average baseline [67] LDL-C of 3.1mmol/l under intensive treatment (80 mg/day atorvastatin) achieved no significance difference in CVE or all-cause mortality compared to those under standard treatment (20 – 40 mg/day simvastatin). ASTEROID Rosuvastatin (40 mg/day) decreased LDL-C and increased [68] HDL-C. The treatment also reduced percent of atheroma volume, total and nominal atheroma volume. PROVE-IT Early intensive therapy (80 mg/day atorvastatin) after acute [69] TIMI 22 coronary syndromes reduced the hazard risk of all-cause mortality by 16% compared to the standard therapy (40 mg/day pravastatin). RADAR Rosuvatstatin (10mg/day, 20mg/day, 40mg/day) were more [70] effective in reducing LDL-C and LDL-C/HDL-C ratio in IHD patients and patients with low HDL-C compared to atorvastatin (20mg/day, 40mg/day, 80mg/day) at 6, 12, and 18 weeks of treatment.
1.1.3.2 Omega-3 fatty acids
The most common omega-3 fatty acids (n-3 FA) are plant-derived α-linoleic acid (ALA, 18:3n-3), and fish-oil derived eicosapentaenoic acid (EPA, 20:5n-3) and docosahexaenoic acid (DHA, 22:6n-3) [71]. The three n-3 FAs are not synthesised in vertebrates, and therefore, have to be acquired from diet. Studies investigating the modification of IHD risk via dietary intake of the n-3 FA have shown mixed results [72], although a compelling number of studies support the anti-atherogenic effects of n- 3 FA. EPA and DHA either in the form of fish oils, dietary fish or supplements showed
10
CHAPTER 1 – LITERATURE REVIEW
pleiotropic effects such as lowering of blood pressure, heart rate, and triglyceride levels, increasing of endothelial relaxation, atherosclerotic plaque stability, as well as decreasing of platelet aggregation, the production of inflammatory eicosanoids, chemoattractants and thrombosis [71, 73]. However, other studies showed that increasing the intake of both EPA and DHA or fish, and the intake of ALA were not associated with the reduction in risk of myocardial infarction [74, 75] (as summarised in Table 1.4).
Although the precise mechanisms of n-3 FA (EPA and DHA in particular) in lowering the triglyceride concentration is not fully understood [71], studies have suggested that n-3 FA may be involved in lipid and lipoprotein metabolism. It was suggested that the triglyceride lowering effect of n-3 FA is achieved by lowering the synthesis of VLDL- triglycerides, and increasing the level of clearance of chylomicrons and VLDL by increasing the level of lipoprotein lipase [71]. EPA and/or DHA were reported to increase intracellular degradation of apolipoprotein B [76], which is associated with non-HDL particles (chylomicrons, VLDL, IDL, and LDL). Also, EPA and/or DHA were associated with decreased hepatic lipogenesis [77], thus resulting in lower level of triglyceride in the liver; This involves either the elevation of peroxisome proliferator- activated nuclear receptor-α mediated β-oxidation or the suppression of sterol regulatory element-binding proteins transcription factors [77]. Furthermore, EPA and/or DHA were proposed to increase peroxisome proliferator-activated nuclear receptor-γ-induced triglyceride hydrolase or lipoprotein lipase [71] gene expression and activity in adipose tissue [78], that potentially lead to lower level of triglyceride in the adipose tissue. Kinetic studies support the idea that fish oils reduce plasma level of triglyceride in addition to increasing HDL-C among insulin-resistant obese males; This may be due to the decrease of the fractional catabolic rate and of concomitant production of HDL apolipoprotein AI and apolipoprotein AII [79]. Evidence for the effect of n-3 FA on lipoprotein metabolism also came from a lipidomic analysis which reported the decrease in the concentration of bioactive lipids ceramides, lysophosphatidylcholine, and diayclglycerol after 8-week of fatty fish consumption [80]. Overall, n-3 FA appears to reduce the risk of IHD by predominantly lowering the level of plasma triglyceride; studies are ongoing to fully elucidate the exact mechanisms of n-3 FA in IHD.
11
CHAPTER 1 – LITERATURE REVIEW
Table 1.4 Studies on omega 3-fatty acids and their findings.
Forms of Findings References omega-3 fatty acid supplementation Dietary fish ≥ 2 servings of fish or ≥1 serving of tuna or dark fish per week delayed the progression of stenosis compared to [81] lower fish intakes among postmenopausal women with IHD; Higher fish consumption resulted in fewer new lesion, minimum coronary artery diameter, and lower concentration of inflammatory marker VCAM-1 in all diabetic and non-diabetic women. High fish intakes (> 1 serving per week or >20g/day) reduced risk of non fatal coronary events, but not for fatal [82] coronary events among Japanese middle-aged men. 8-week of fatty fish consumption (4 fish per week) significantly decreased bioactive lipids (ceramides, [80] lysophosphatidylcholine, and diacylglycerol), which are related to inflammation and insulin resistance compared to controls and participants who ate lean fish. Increasing intakes of EPA+DHA and fish (up to ≥5 fish meals/week) did not reduce the risk of coronary events [74] among US males.
Dietary High degree of inter-individual variability in lipid metabolism (prostaglandin E2, 12-HETE, thomboxane B2) was [83] supplements found among healthy subjects post 6 weeks of 1.9 g/day EPA and 1.5 g/day DHA supplementation. 1.8 g/day of EPA and statins reduced major coronary event compared to controls (statins only) among [84] hypercholesterolaemic patients. 1.8 g/day of EPA reduced carotid intimal media thickness and brachial-ankle pulse wave velocity among type II [85] diabetes patients during 2 years of study. Compared to control and 3.3 g/day EPA treatment, 3.7 g/day DHA treatment increased total cholesterol levels via [86] E4 carriers, which resulted from elevated LDL-C levels among healthy normolipidaemic males. Other sources Intake of ALA via food did not reduce the risk of IHD among Dutch elderly. [75] Indo-Mediterranean dietary intervention (rich in ALA) resulted in reduced sudden cardiac deaths, non-fatal MI, [87] and lower cholesterol concentration.
12
CHAPTER 1 – LITERATURE REVIEW
1.1.3.3 Anti-oxidant vitamins
The association of oxidative stress with atherosclerosis has led to a hypothesis that the disease progression can be delayed or prevented if oxidative stress is reduced. To test the hypothesis, an extensive number of intervention studies mainly involving vitamin C, vitamin E, or a combination of the anti-oxidant vitamins were conducted. The results from several cellular [88], and animal studies [89-91] and small set of human studies [92, 93] have been encouraging. However, the positive effects could not be translated to clinical application as the outcomes of human trial have been equivocal, and comprised mainly of negative results such as those observed in WACS, and HOPE studies (Table 1.5).
Several reasons for the failure in the clinical trials have been proposed. These include inappropriate isoforms, dose and/or duration of treatment used during these intervention studies [94]. The efficacy of anti-oxidant vitamins observed in the animal studies may be due, in part, to the higher doses used in those studies as compared to that in the clinical trials. Moreover, there is an individual variability in the absorption of the vitamins as well as an ongoing controversy on the optimal dose of the anti-oxidant vitamins [94]. The primary outcome of the animal studies was, in most studies, the prevention of the earliest forms of atherosclerosis [95]. The level of anti-oxidant efficacy observed in animal studies may not be translated to human studies because of the complexities of existing disease in humans, and the limited duration of the trials. In addition, clinical trials often used easily available agents, which are of questionable potency. Natural vitamin E is composed of eight different isoforms [94] and the different levels of potency of the isoforms warrant further investigations and may have influenced the outcome of clinical trials.
The precise mechanisms of vitamin C and vitamin E in the attenuation of atherosclerosis are unknown. However, given their biochemistry, studies have suggested that vitamin C protects membrane lipids from peroxidation by acting as a scavenger of ROS. In addition, vitamin C acts as an one-electron reducing agent of lipid hydroxyperoxyl radicals via the vitamin E redox cycle [96].Vitamin C is able to maintain or enhance the bioavailability of nitric oxide, an important vasodilator for endothelial vasomotor functions. This is important for the improvement of endothelial
13
CHAPTER 1 – LITERATURE REVIEW
dysfunction as a result of defective endothelium-dependent vasodilation [97, 98]. Vitamin E or α-tocopherol is a potent peroxyl radical (ROO.) scavenger, which has a higher rate of reactivity with ROO. than with polyunsaturated fatty acids [99]. The hydroxyl group of α-tocopherol reacts with ROO. to form the corresponding lipid hydroperoxide and tocoheryl radicals. Tocopheryl radicals are resonance-stabilised and do not react with oxygen to propagate more free radicals, thus vitamin E is termed a chain-breaking anti-oxidant. In the presence of vitamin C, tocopheryl radicals are returned to their reduced state [100].
14
CHAPTER 1 – LITERATURE REVIEW
Table 1.5 Studies on anti-oxidant vitamins and their principal findings.
Name of Study type Intervention Findings References study The St. 1,005 asymptomatic healthy men and women, Atorvastatin 20mg/day, Treatment significantly reduced total [101] Francis aged 50 to 70 years, with coronary calcium vitamin C 1g/day, and cholesterols, LDL-C and triglycerides but Heart Study score ≥ 80th percentile of their gender and vitamin E 1000U/day. did not significantly reduce progression of age; Mean treatment of 4.3 years. coronary calcium score. VEAPS 353 men and women aged ≥ 40 years, with Vitamin E 400IU/day. Vitamin E did not reduce the progression [102] LDL-C ≥ 3.37 mmol/l, and no clinical of intima media thickness, though it symptoms of IHD; Followed for 3 years. significantly raised plasma vitamin E levels, reduced LDL-C and LDL oxidisability. The 14,641 US males, aged ≥ 50 years, Vitamin C 500mg/day Neither vitamin E nor vitamin C [103] Physicians's including 754 men with prevalent IHD; and vitamin E 400IU significantly reduced the risk of major Health study Followed for 10 years. every other day. CVE and myocardial infarction. Vitamin II E was associated with increased risk of hemorrhagic stroke. WACS Women, aged ≥40 years, with a history of Vitamin C 500mg/day, There were no overall effects of the anti- [104] IHD or ≥ 3 risk factors of IHD; 2x2x2 vitamin E 600IUand β- oxidants on the risk of CVE. factorial study; Followed for 9.4 years. carotene 50mg every other day. HOPE 3,654 patients, aged ≥ 55 years with diabetes Vitamin E 400IU/day and Vitamin E treatment had no effect on the [105] and IHD and/or additional risk factors; 2x2 Ramipril 10mg/day. cardiovascular outcomes. factorial study; Followed for 4.5 years. ASAP 520 smoking and non-smoking men, and Slow releasing vitamin C Treatment significantly reduced carotid [106, 107] postmenopausal women, aged 45 to 69 years 350mg/day and vitamin E artery intima-media thickness in men, but with hypercholestolemia; Followed for 3 and 136IU/day. not in women in the 3 and 6 years follow- 6 years. up.
15
CHAPTER 1 – LITERATURE REVIEW
1.2 OXIDATIVE STRESS AND INFLAMMATION AS CONTRIBUTING FACTORS TO ATHEROSCLEROSIS
1.2.1 Oxidative stress and inflammation in the cardiovascular system
In aerobic cells, oxidants are normal by-products of cellular metabolism. However, the production rate of these oxidants is often elevated in diseased states. Oxidative stress is defined as a disproportion in the level of oxidants and anti-oxidants, in favour of the oxidants, thus potentially causing damage [108]. Inflammation is a complex biological response to a harmful stimulus. Participants and/or markers of the inflammatory process in atherosclerosis include vascular adhesion molecule-1 (VCAM- 1) and intracellular adhesion molecule-1 (ICAM-1), which are receptors on the vascular endothelial cells for immune cells including monocytes and leukocytes, as well as monocyte chemoattractant protein-1 (MCP-1).
1.2.1.1 Major sources of reactive oxygen species
Major sources of ROS in the cardiovascular system include nicotinamide adenine dinucleotide phosphate (NADPH) oxidase (NOX), dysfunctional endothelial nitric oxide synthase (eNOS), xanthine oxidase [109], and myeloperoxidase.
1.2.1.1.1 NADPH oxidase
NOX has been proposed as the predominant producer of superoxide in endothelial and smooth muscle cells [109]. The expression or activity of NOX and/or the generation of ROS were shown to increase in human atherosclerotic arteries [110], and patients with high cardiovascular risk [111]. There are different isoforms of NOX including NOX 1, NOX 2 and NOX 4. Studies of mouse models of diabetes-accelerated atherosclerosis demonstrated that pharmacological inhibition of both NOX 1 and NOX 4 [112], and genetic deletion of NOX 1, but not NOX 4 [113], improved atherosclerotic plaques and significantly reduced the formation of ROS and the levels of inflammatory markers including VCAM-1 and MCP-1. These studies highlight the potentially different contribution of isoforms of NOX to atherosclerosis progression.
1.2.1.1.2 Endothelial nitric oxide synthase
eNOS transfers electrons from donor NADPH to a prosthetic heme group, catalysing the reaction of L-arginine and the cofactor 5,6,7,8-tetrahydrobiopterin to L-
16
CHAPTER 1 – LITERATURE REVIEW
citrulline and nitric oxide (NO) [109, 114]. NO is a vasoprotective molecule that stimulates the dilation of blood vessels, and inhibits the adhesion of leukocytes to the vascular walls, a feature of atherosclerosis progression [115]. Studies have proposed that during vascular disease states, the expression of NOX and eNOS are upregulated; Their respective products, superoxide and NO, react readily to form peroxynitrite, which in turn oxidises eNOS’ co-factor, 5,6,7,8-tetrahydrobiopterin and causes oxidative damage to the zinc-thiolate cluster of eNOS. As a result, eNOS is ''uncoupled'' and the level of superoxide increases [115]. Additionally, in conditions where the substrate L- arginine becomes limited or in the presence of native [116] or oxidised LDL [117], the same effect may occur. Peroxynitrite is able to oxidise proteins at the tyrosine residues to form a more stable oxidative product, nitrotyrosine [118].
1.2.1.1.3 Xanthine oxidase
Xanthine oxidase exists in endothelial cells, but not in smooth muscle cells [114]. The enzyme readily donates electrons to molecular oxygen to generate superoxide and hydrogen peroxide [109, 115]. The activity and expression of xanthine oxidase is induced by the presence of interferon-γ, which is an inflammatory mediator [119]. In studies of hyperlipidemic animals [120] and hypercholesterolemic patients [121], inhibition of xanthine oxidase by oxypurinol were shown to reduce superoxide and to improve impaired vasodilation. These studies suggest a role of xanthine oxidase in endothelial dysfunction particularly in dyslipidemia.
1.2.1.1.4 Myeloperoxidase
Myeloperoxidase, a glycosylated heme protein, is released from monocytes and activates polymorphonuclear leukocytes at sites of inflammation [122]. It catalyses the conversion of hydrogen peroxide and Cl- to hypochloric acid; Hypochloric acid can then . oxidise nitrite to form the radical NO2, which in turn reacts with tyrosine to form 3- nitrotyrosine [123]. Clinical studies showed the association of myeloperoxidase to atherosclerosis; It was demonstrated that increased level of myeloperoxidase in blood and/or white blood cells and its oxidative products were predictive of future CVE [122, 124]. In addition, its level in serum was predictive of endothelial dysfunction in humans [125]. It was also suggested that myeloperoxidase uses NOX-derived hydrogen peroxide to produce hypochloric acid and chlorinated species, that exacerbate hydrogen
17
CHAPTER 1 – LITERATURE REVIEW
peroxide-induced vascular injuries by impairing the endothelium-dependent relaxation [123, 126], thus contributing to atherosclerosis.
1.2.1.2 Markers of inflammation
VCAM-1, ICAM-1 and MCP-1 are participants in inflammatory response and their increased levels are used as markers of inflammation particularly in diseased states. Below we briefly discuss the roles of these markers in inflammation.
1.2.1.2.1 Vascular adhesion molecule-1 and intracellular adhesion molecule - 1
VCAM-1 and ICAM-1 belong to the cytokine-inducible IgG gene superfamily. These molecules are expressed during inflammation and often associated with atherosclerosis. Peripheral arterial disease indicates a systemic inflammation and is a common manifestation of atherosclerosis. The levels of circulating VCAM-1 and ICAM-1 were demonstrated to be significantly higher in patients with peripheral arterial disease compared to the healthy control group before and after a treadmill test [127], thus indicating: (1) a sustained increased level of the inflammatory markers in the patients compared to the healthy controls, and (2) an increase in VCAM-1 and ICAM-1 during hemodynamic stress. In other studies, it was demonstrated that VCAM-1 and ICAM-1 were expressed at the atherosclerotic lesion-prone sites of endothelial and intimal cells of animal models of atherosclerosis [128, 129]. In addition, these molecules promoted monocyte rolling and attachment to the carotid arteries [130].
The mechanism, which triggers the expression of VCAM-1 and ICAM-1, is unclear. A study on VCAM-1 domain 4-deficient and ICAM-1-deficient mice demonstrated that VCAM-1 but not ICAM-1 was involved in early atherosclerosis progression [131]. Other studies identified that low shear stress and triglyceride rich lipoproteins increased tumor necrosis factor-α-induced expression of VCAM-1 in endothelial cells [132]. Whereas, membrane-shed submicron particles from the atherosclerotic lesions could induce the expression of ICAM-1 and cause monocyte adhesion in cell culture and isolated perfused mouse carotid artery [133]. Additionally, resistin, an adipocyte- specific hormone was also found to induce the expression of the adhesion molecules via a p38 mitogen-activated protein kinases (MAPK)-dependent pathway [134].
18
CHAPTER 1 – LITERATURE REVIEW
1.2.1.2.2 Monocyte chemoattractant protein - 1
MCP-1 belongs to the CC chemokine superfamily and is characterised by adjacent cysteine residues in the proximity of the N-terminus of the protein and several clusters of genes [135]. MCP-1 is an agonist for monocytes; it attracts and regulates their migration and infiltration. The expression of MCP-1 can be induced via several stimulants including tumor necrosis factor-α, interleukins IL-1 and IL-4, platelet- derived growth factor, and interferon-γ [136]. MCP-1 can be produced in endothelial, epithelial, smooth muscle cells, and fibroblasts. However, the major sources of MCP-1 include monocytes and/or macrophages [137]. MCP-1 has a pathophysiological role in the atherosclerotic progression. As a result of vascular insults, dysfunctional endothelial cells secrete MCP-1 which tethers at the proteoglycans facing the vessel lumen [138]. CC chemokine receptor 2 is seven transmembrane G-coupled protein receptors on the membrane of monocytes, but they can also be expressed in endothelial cells [139, 140]. The binding of monocytic CC chemokine receptor 2 to the ligand MCP-1 results in the monocyte rolling and adherence to the endothelial cells [141], as well as activation of signalling events which attracts monocytes to the site of inflammation [138, 142]. Early studies have reported an elevated expression of MCP-1 in human atherosclerotic lesions [143] and vascular smooth muscle cells of hypercholesterolemic primates [144] as well as the association of elevated MCP-1 levels in circulation with increased risk of myocardial infarction in coronary artery disease (CAD) patients [145].
1.2.2 Impact of oxidative stress and inflammation on cellular function
Oxidative stress and/or inflammation cause cellular dysfunction, which can exacerbate disease progression. The treatment of vascular smooth muscle cell cultures with the pro-inflammatory cytokine IL-6 resulted in the increased production of ROS, while the same treatment to C57/BL6J mice led to impaired endothelium-dependent vasodilatation [146]. In addition, increased oxidative stress resulted in premature senescence of vascular smooth muscle cells. It has been demonstrated that vascular smooth muscle cells in human fibrous plaques showed telomere shortening, DNA oxidative damage, and disrupted cell regulation that are characteristics of cellular senescence. It was also reported that the degree of telomere shortening was associated with the severity of atherosclerosis [147]. Similarly, endothelial cells obtained from
19
CHAPTER 1 – LITERATURE REVIEW
patients with high CVD risk exhibited premature senescence, telomere shortening, as well as markers of cell damage and lipid peroxidation [148].
Under normal conditions, endothelial progenitor cells are released from the bone marrow in response to an inflammation stimulant and are mobilised to the sites of injury to mediate neovascularisation and tissue repair [149]. However, chronic levels of inflammatory stimulation such as by tumor necrosis factor-α and glucose were shown to result in a decrease of endothelial progenitor cells number that was mediated by p38 MAPK [150]. In patients with CVD, endothelial progenitor cell functionality and number were shown to be reduced [151]. Together these studies highlight that oxidative stress and inflammation lead to cellular dysfunction primarily by disrupting cellular regulation and as well as causing cell death.
1.2.3 Animal models of atherosclerosis
Mice with C57/BL6 background that are deficient in apolipoprotein E (ApoE-/-) are often used for animal experimentation in atherosclerosis research. Apolipoprotein E (apoE) is a ligand required for the interactions of lipoproteins including VLDL and LDL with LDL-receptor (LDL-R) and Lipoprotein-Related Protein for their hepatic clearance. The lack of apoE in the ApoE-/- mouse results in the reduction of lipoprotein hepatic clearance, and thus an increased level of triglycerides as well as a state of hypercholesterolemia in the mouse. Atherosclerosis spontaneously develops in ApoE-/- mouse with chow diet feeding within 10 weeks [152], and is greatly accelerated with western type diet (typically 21% fat, 0.15% cholesterol) feeding.
Although the use of murine model of atherosclerosis has greatly enhanced our understanding of the disease process, the caveat of using such model is the biological and physiological differences to humans. Compared to humans, mice in general have low levels of LDL and do not express cholesteryl ester transport protein, one of the key proteins involved in the lipoprotein metabolism. In mice such as C57/BL6, HDL makes up the predominant lipoprotein in circulation and the major carrier of plasma cholesterol and so these mice are not susceptible to atherosclerosis [152]. Moreover, mouse heart rate averages 300 beats/min whereas human heart rate is normally within the range of 50 - 90 beats/min. The postural differences between the two species may also affect hemodynamics and their susceptibility to atherosclerosis development [153, 154]; these
20
CHAPTER 1 – LITERATURE REVIEW
postural differences may also explain the difference in the lesional distribution between man and mice [152]. Mice are models of atherosclerosis progression, but not plaque instability and rupture that resemble normal pathology in humans. A recent study has shown that plaque instability in mice could be achieved surgically [155].
Human prospective studies had previously identified the association of low level of glutathione peroxidase-1 (GPx-1) activity with the increased risk of atherosclerosis [156]. GPx-1 is an anti-oxidant enzyme which is expressed ubiquitously in cells; it reduces and detoxifies hydrogen peroxide and lipid hydroperoxide, although its mechanistic action on lipid hydroperoxide was proposed to work consecutively with -/- -/- -/- phospholipase A2 [157]. A lack of GPx-1 in ApoE mice (ApoE GPx1 ) was shown to accelerate atherosclerosis [158] and diabetes-associated atherosclerosis [159]. Compared to ApoE-/- mice, ApoE-/-GPx1-/- mice were shown to have higher levels of oxidative stress markers such as nitrotyrosine and superoxide [158]. The use of ApoE-/- and ApoE-/-GPx1-/- mice in atherosclerosis studies may provide insight into the roles of oxidative stress in atherosclerosis and allow us to assess the efficacy of various anti- oxidant treatments on these mice models, which have differing levels of oxidative stress.
1.3 INTRODUCTION TO LIPOPROTEIN METABOLISM
Lipids play vital biological functions in metabolism, nutrition, and health. Lipid abnormalities underlie the development of IHD. Although the exact roles of lipids in IHD have not been fully elucidated, their associations in IHD have been established and this will be discussed further in section 1.5.3. Several lipids including triglycerides, fatty acids, phospholipids, and cholesterol play important roles in the lipid metabolism. Triglycerides and fatty acids are important as energy storage; phospholipids and cholesterols maintain the integrity of cellular membranes. Cholesterol is also a precursor to multiple steroids, hormone and cell signalling molecules. Moreover, lipids help in the absorption of fat-soluble vitamins. Due to the amphipathic properties of lipids, with the exception of free fatty acids, they are transported in plasma to and from tissues by macromolecular complexes called lipoproteins.
Lipoproteins display heterogeneity but they share similar features. Lipoproteins are composed of lipids and amphipathic proteins, known as apolipoproteins. The outer layer of lipoproteins consists mainly of phospholipids and free cholesterols; whereas
21
CHAPTER 1 – LITERATURE REVIEW
hydrophobic cholesteryl esters and triglycerides constitute the core of the lipoproteins (Figure 1.2). In human plasma, lipoproteins are classified based on their hydrated density, particle size, floatation rate, and electrophoretic mobility. The major lipoprotein classes include chylomicrons, VLDL, IDL, LDL, and HDL. Chylomicrons have the largest particle size and the lowest density in the lipoprotein family as they have the lowest and highest proportions of proteins and triglycerides respectively. In contrast, HDL has the smallest particle size and highest density. The characteristics of each class of the lipoprotein family are summarised in Table 1.6. LDL and HDL are divided into subclasses. LDL subclasses are distinguished by their particle sizes and hydrated density and are generally known as large LDL, medium LDL, and small LDL [160]. Similarly, HDL is sub-divided into HDL2 and HDL3 classes due to their hydrated densities and can be classified further as HDL2b, HDL2a, HDL3a, HDL3b, HDL3c because of their decreasing sizes and differing mobility in gel electrophoresis [161]. Another level of HDL heterogeneity has been uncovered in its metabolic functions such as cholesterol efflux, anti-oxidative, and anti-inflammation [161].
1.3.1 Lipoprotein composition
1.3.1.1 Apolipoproteins associated with lipoproteins
Apolipoproteins play important roles in interacting with lipoprotein-associated enzymes and regulating plasma lipid metabolism, as well as providing structural support of the lipoprotein complex. To date, there are at least 10 classes of apolipoproteins, which have been identified. However, only five of them have been well studied and documented (Table 1.7). These include apolipoprotein A, B, C, D, and E. Other types of apolipoproteins designated F, H, J, and L-I, M and more recently O have been described to be associated with HDL functionality. However, their quantitative contribution in the lipoprotein metabolism is still largely unknown. The discussion below will focus only on major components of A, B, and E of the apolipoprotein family that are relevant to this thesis. The function and the location of the apolipoproteins in the subclasses of lipoproteins are summarised in Table 1.7.
22
CHAPTER 1 – LITERATURE REVIEW
Figure 1.2 Schematic representation of a lipoprotein structure. Lipoproteins are composed of lipids and apolipoproteins. The outer layer of lipoproteins consists mainly of phospholipids and free cholesterols whereas cholesteryl esters and triglycerides constitute the core of lipoproteins. Phospholipids consist of several subclasses depending on the headgroups (for instance, ethanolamine or choline), as well as the different bonds (ester, vinyl ether, or ether) at the sn-1 position of the glycerol backbone. Figure was obtained from cardiologydoc.wordpress.com and then modified; Lipid molecules were drawn by ChemDraw.
23
CHAPTER 1 – LITERATURE REVIEW
Table 1.6 Characteristics of the lipoprotein family.
Lipoproteins Hydrated density Sf b Mean particle Electrophoretic Apolipoprotein (g/ml) a diameter (nm) c mobility d content
Chylomicrons < 0.95 >400 220 Origin B48, C, E VLDL 0.95 - 1.006 20 - 400 29 - 140 α2 B100,C, E IDL 1.006 - 1.019 12 - 20 25 α2 - β B100, E LDL 1.019 - 1.063 0 - 12 19 -22 β B100 HDL 1.063 - 1.21 0 - 9 7.5 - 11.5 α1 AI, AII, C a based on methodology by Havel et al. [162] and Chapman et al. [163]; b Sf, Svedberg floatation rates adapted from Crownwell and Kruger, and Patsch et al. [164, 165]; c determined by NMR [166]; d adapted from Cornwell and Kruger [164].
24
CHAPTER 1 – LITERATURE REVIEW
1.3.1.1.1 Apolipoprotein A
Apolipoprotein AI (apoAI) and AII (apoAII) are primarily associated with HDL structure and function. Both apoAI and apoAII are synthesised predominantly in the liver and intestine.
ApoAI constitutes approximately 70% of total HDL protein [167]. It is a 28kDa amphipathic protein. The conformation of apoAI allows lipid binding to the protein, and the formation of mature and stable micellar complexes as well as the transfer of the protein between lipoprotein classes [168]. ApoAI is involved in activating lecithin:cholesterol acyltransferase (LCAT) [167] which esterifies cholesterols in the formation of mature HDL particles. The redox status of apoAI is one of the major determinants of HDL3 anti-oxidative capacity in mediating protection of LDL against free-radical induced oxidation; the methionine residues of apoAI are responsible for reducing phosphatidylcholine hydroperoxides to phosphatidylcholine hydroxides [33]. Discoid HDL was shown to have two apoAI molecules, whereas spherical HDL can have more than three (four to five) apoAI molecules, particularly in the large HDL subpopulations (size of 9.4 to 14 nm) [168].
ApoAII is the second most abundant protein in HDL, constituting 15%-20% of total HDL protein [31]. Unlike apoAI which is believed to be contained in all HDL particles, apoAII may only be present in half of them [169]. ApoAII has a homodimeric structure and a molecular weight of 17kDa [168]. The roles of apoAII in the lipoprotein metabolism have not been fully elucidated and the findings so far have been complex and controversial [170]. Nonetheless, apoAII was demonstrated to modulate the activity of enzymes involved in the lipoprotein metabolism including LCAT activity via scavenger receptor class B member 1 (SRB-1)-dependent pathway [171, 172] and the lipid hydrolytic activity of hepatic lipase [173]. In addition, apoAII may regulate the metabolism of VLDL by inhibiting the lipoprotein phosphatidylcholine hydrolysis [173, 174].
1.3.1.1.2 Apolipoprotein B
Apolipoprotein B (apoB) plays a key role in the lipoprotein transport system. The two main isoforms of the apoB family include apoB100 and apoB48; they are
25
CHAPTER 1 – LITERATURE REVIEW
involved chiefly in the metabolism of triglyceride-rich lipoproteins [175]. In contrast to other classes of apolipoprotein, apoB proteins do not undergo exchange between the lipoproteins [176]; they stay in the associated lipoproteins throughout their metabolic fate.
ApoB100 is an amphipathic monomeric protein with a molecular weight of approximately 540kDa and is produced in the liver. The protein structure of apoB100 allows high-affinity binding to lipids in VLDL and LDL [177, 178]. The mechanism of the assembly of apoB containing lipoproteins is complex, but can be described in two major stages that follow the "lipid pocket model" [176, 179]. In the first stage, nascent polypeptide of the apoB (residues 1-1000 in α1 domain) interacts with microsomal triglyceride-transfer protein, producing an intermediate complex. Subsequently, this allows the accretion of lipid, creating a lipid nidus for the assembly of the lipoproteins. In the second stage, the transcription and translation of apoB occur concurrently with the lipoprotein assembly such that the C-terminal of the protein is synthesised on the ribosomes by endoplasmatic reticulum while the N-terminal is involved in the formation of nascent lipoprotein particles. The addition of β-sheets that line the lipid pocket then allow a pocket expansion by the accretion of more lipids, thus creating a mature lipoprotein particle [176, 179]. ApoB100 contains LDL-R binding domains responsible for the uptake of plasma LDL back into the liver.
ApoB48, so named as it has a molecular weight of approximately 48% of apoB100. The protein is synthesised in the intestine and secreted in chylomicrons and is also found in chylomicron remnants [179, 180] . The association of apoB48 to lipoproteins appears to be specific to chylomicron metabolism as the protein was not detected in fasting individuals [180]. Moreover, apoB48 is not involved in the uptake of VLDL into the liver [180] as the protein does not possess LDL-R binding domains [181].
1.3.1.1.3 Apolipoprotein E
ApoE is a 34kDa polymorphic glycoprotein. ApoE has three isoforms designated apoE2, E3, and E4 arising from three alleles of the same gene locus on human chromosome 19. The three isoforms have unique binding properties contributed by the difference in the amino acid sequence at position 112 and 158 [182, 183]. As a result, each of the isoforms has different preferential lipoprotein binding; apoE4
26
CHAPTER 1 – LITERATURE REVIEW
preferentially associates with VLDL and LDL, whereas apoE3 with HDL [184]. Both apoE3 and E4 were reported to show similar affinity for LDL-R, whereas that of apoE2 was 50 - 100 times weaker [185]. The lower LDL-R binding capacity of apoE2 may result in delayed lipoprotein clearance [186], particularly if the level of the isoform dominates in circulation.
ApoE plays multifunctional roles in the lipoprotein metabolism. It directs the transport of endogenous triglycerides and cholesterols in VLDL or dietary forms of the lipids in chylomicrons to the extrahepatic cells and liver respectively. ApoE delivers the lipid via two pathways: the LDL-R and the LDL-R related protein pathway [182]. As previously mentioned, the absence of apoE in ApoE-/- mice has been reported to result in hypercholesterolemia and the development of atherosclerotic plaques, even if they were on a chow diet.
1.3.1.2 Enzymes and lipid transfer proteins associated with lipoproteins
The lipoprotein transport system consists of a complex network of receptors, enzymes, and proteins that allow the uptake, exchange, and transfer of lipids and apolipoproteins to and from the sources of synthesis and target cells, via the circulation. The implication of the enzymes and transfer proteins may extend beyond their primary roles as studies showed that their dysfunctions might contribute to the development of atherosclerosis. The primary biological functions of the enzymes and transfer proteins are discussed below and summarised in Table 1.7.
1.3.1.2.1 Plateler-activating factor acetylhydrolase
Platelet-activating factor acetylhydrolase or also termed lipoprotein-associated phospholipase A2 (Lp-PLA2) is an N-glycosylated 45kDa enzyme, which is secreted by monocyte-derived macrophages, mast cells and T-lymphocytes. These cells contribute to the majority of circulating Lp-PLA2 [187]. The enzyme is closely associated with small dense LDL and, to a lesser extent, HDL. However, only approximately 0.1% of the lipoproteins are enriched with the enzyme [188]. Lipoprotein (a), an atherogenic lipoprotein and a preferential carrier of oxidised phospholipids in human plasma appears to be laden with Lp-PLA2 [189, 190].
27
CHAPTER 1 – LITERATURE REVIEW
Unlike other isoforms of PLA2, Lp-PLA2 has a unique Ser/Asp/His catalytic triad [188] which allows the hydrolysis of various substrates including platelet-activating factor (PAF), oxidised phospholipids, and esterified isoprostanes [191, 192]. The calcium- independent enzyme regulates the biological action of PAF by hydrolysing the sn-2 ester bonds of PAF to produce biologically inactive lyso-PAF [188]. Lp-PLA2 is also involved in the production of lysophosphatidylcholine and oxidised non-esterified fatty acids which are proposed to be pro-apoptotic and pro-inflammatory [193, 194]. Lysophosphatidylcholine in particular, interacts with the G-protein-coupled receptor G2A that is able to regulate macrophage and T-cell migration as well as macrophage activation [195]. These lysophosphatidylcholine effects may contribute to the initiation of atherosclerosis. The elevated expression of Lp-PLA2 and increased lysophosphatidylcholine production were found in advanced [196] and symptomatic atherosclerotic plaques [197]. Thus, these features have been associated with the risk of
IHD. This is further supported in studies which demonstrated that increased Lp-PLA2 mass or activity was associated with CAD and elevated risk of CAD [198, 199]. Lp-
PLA2 was also suggested to catalyse the hydrolysis of pro-inflammatory oxidised phospholipids to form lysophosphatidylcholine. Recent analysis on oxidised LDL led to the identification of species of short and long chain of oxidised phosphatidylcholine
(oxPC) as the substrates of Lp-PLA2 and saturated and mono-unsaturated lysophosphatidylcholine as the major products. These species were then validated in human atherosclerotic lesions [200]. Despite the atherogenic potential of increased level of Lp-PLA2, clinical trials with Darapladib, a selective and potent oral inhibitor of
Lp-PLA2, showed that the drug failed to significantly reduce the risk of major events including cardiovascular death in stable CAD patients [201] and of recurrent events in patients with acute coronary syndrome [202].
1.3.1.2.2 Lechithin:cholesterol acetyltransferase
LCAT is secreted primarily in the liver. The majority of LCAT in circulation (75%) is associated with HDL [168]. LCAT catalyses the esterification of cholesterol to cholesteryl esters in lipoproteins [203], particularly in HDL. Specifically, with apoAI as a co-factor and physiological activator, LCAT transfers an acyl group of lecithin (phosphatidylcholine) to the free hydroxyl group of cholesterol to form a more
28
CHAPTER 1 – LITERATURE REVIEW
hydrophobic cholesteryl ester which will be retained in the lipoprotein core and produces lysophosphatidylcholine in the process [204, 205].
Deficiencies in functional LCAT in humans were reported to increase the risk of atherosclerosis [206] and cause arterial stiffness [207] by impairing the lipoprotein metabolism including lowering the levels of HDL-C, elevating triglycerides and C- reactive protein 1 [206] which is an acute phase protein and a marker of inflammation. Co-incubation of recombinant human LCAT and plasma from patients with familial LCAT deficiency resulted in the normalisation of lipoprotein profiles including increased levels of cholesteryl ester, HDL-C, and induction of the maturation of preβ- HDL to mature circulating HDL [208]. These studies highlight the importance of LCAT in the lipoprotein metabolism.
1.3.1.2.3 Paroxonase-1
Human paraoxonase-1 (PON-1) is a member of the paraoxonase family. It is a lactonase and arylesterase which is primarily synthesised in the liver and almost exclusively associated with HDL [168]. PON-1 is known to possess anti-oxidative and anti-inflammatory mechanism; PON-1 was reported to inhibit the THP1 monocyte - to - macrophage differentiation [209], and to directly suppress macrophage pro- inflammatory responses such as the production of ROS [210]. The expression of PON-1 is associated with the expression of SRB1 and SRB1-mediated HDL cytoprotection against necrotic macrophage apoptosis [210]. Furthermore, the arylesterase activity of PON-1 was shown to have a greater stimulation towards phosphatidylcholine- containing oxidised chain at the sn-2 position compared to the paraoxonase activity [211]. Purified PON-1 was demonstrated to inhibit LDL and HDL oxidation, as indicated by a reduction in the conjugated diene production; furthermore, PON-1 was shown to reduce inflammation as indicated by its effect on lowering the production of lysophosphatidylcholine and ICAM-1 level in the presence of oxidised HDL [212].
1.3.1.2.4 Phospholipid transfer protein
Phospholipid transfer protein (PLTP) belongs to the lipid transfer/lipopolysaccharide-binding protein and is primarily expressed in the liver. PLTP from the liver contributes to 25% of all the PLTP activity in circulation [213]. PLTP mediates the transfer and exchange of phospholipids from apoB-containing
29
CHAPTER 1 – LITERATURE REVIEW
lipoproteins to HDL [214], but was also reported to efficiently transfer α-tocopherols, diacylglycerols, and lipopolysaccharides [215]. Although PLTP is required for normal functioning and metabolism of lipoproteins, its elevated expression was associated with risk factors of IHD including diabetes and insulin resistance [216]; it has been reported that PLTP was widely distributed in tissues as well as in atherosclerotic lesions and macrophages [217]. However, decreased expression of PLTP has also been associated with peripheral atherosclerosis [218].
1.3.1.2.5 Cholesteryl ester transfer protein
Cholesteryl ester transfer protein (CETP) is expressed most abundantly in the liver, adipose tissue, and spleen. CETP mediates the bidirectional exchange of cholesteryl ester from HDL to VLDL or LDL for triglyceride (or triacylglycerol). Triglyceride-rich HDL is then metabolised by hepatic lipase [219] leading to the shedding of apoAI.
Deficiency of CETP, either genetically or by pharmacotherapy inhibitors (torcetrapid or evacetrapid) have been reported to enhance the capacity of HDL to promote ATP- binding cassette sub family G member 1 (ABCG1)-mediated cholesterol efflux from macrophages [220] and to decrease LDL-C and increase HDL-C levels in patients with dyslipidaemia [221], suggesting the contribution of CETP to atherosclerosis. However, the exact role of CETP in atherosclerosis still warrants further investigation as CETP is clearly required for normal metabolism and functioning of lipoprotein. Furthermore, inhibition of CETP did not result in the reduction of atherosclerosis risk in humans [36, 222].
30
CHAPTER 1 – LITERATURE REVIEW
Table 1.7 Characteristics of major apolipoproteins, enzymes, and transfer proteins of lipoprotein metabolism.
Description Molecular Lipoprotein Biological functions size (kDa) Apolipoprotein ApoAI 28 HDL Lipid binding; interaction with ABCA1/ABCG1 and SRB1; activation of LCAT; anti-oxidant property ApoAII 17 HDL Modulate LCAT and HDL activity; reduce cholesterol efflux; inhibit phosphatidylcholine hydrolysis ApoB100 540 VLDL, IDL, LDL VLDL assembly; ligand for LDL-R ApoB48 259 CM, CR Chylomicron assembly and secretion ApoCIII 8.8 CM, VLDL, HDL Inhibit lipoprotein lipase and hepatic uptake of triglyceride-rich lipoprotein ApoE 34 CM, VLDL, LDL, HDL Ligand for several receptors including LDL-R and Lipoprotein-Related Protein; potential anti-atherogenic property Enzymes
PAH-AH (Lp-PLA2) 45 LDL, HDL Regulate PAF; hydrolyse oxidised phospholipids LCAT 64 HDL Expansion of HDL lipid core via esterification of cholesterol; potential anti- oxidant property PON-1 43 HDL Potential anti-oxidant and anti-inflammatory property Lipid transfer proteins PLTP 81 Transfer of phospholipids from apoB-containing lipoproteins to HDL CETP 74 Mediate bidirectional exchange of cholesteryl esters from HDL to VLDL or LDL for triglycerides
31
CHAPTER 1 – LITERATURE REVIEW
1.3.1.3 Lipid composition of lipoproteins
Mass spectrometry has allowed researchers to identify a diversity of lipids in human plasma [223], lipoproteins [224] and other biological samples. As compared to a decade ago, the high throughput technology has enhanced our understanding of the compositional and molecular details of lipids in circulation. With this knowledge, it is possible to profile differences in circulating lipids between healthy and diseased states and subsequently identify lipid biomarkers associated with metabolic abnormalities, which underlie disease progression.
Lipidomic studies on healthy individuals revealed that approximately 60% of glycerophospholipids in plasma such as phosphatidylcholine and phosphatidylethanolamine, and alkenylphosphatidylethanolamine or PE plasmalogen were found in the HDL fraction. The HDL fraction has also been reported to contain 87% of the plasma lysophosphatidylcholine [224]; although in this study the HDL was not separated from the albumin fraction. According to a study in our laboratory, albumin fraction contains approximately 70% of lysophosphatidylcholine in circulation. In contrast, the majority of the sphingolipids including sphingomyelin [224], hexosylceramides, and lactosylceramide (approximately 50% respectively) [225] and ceramide (60%) were found in the LDL fraction [224]. In addition, approximately 90% of, sphingosine 1-phosphate, known as an anti-apoptotic molecule [226], was found in the HDL fraction [224, 225]; studies on the sub-populations of HDL revealed that compared to the largest HDL particle, HDL2b, sphingosine 1-phosphate was enriched three-fold in the small dense HDL3c. However, there was no significant difference in the sphingosine 1-phosphate content between HDL3a, 3b, and 3c sub-fractions. Both sphingomyelin and free-cholesterol in HDL3c were shown to be two-fold lower compared to those in HDL2b. Concomitant progressive enrichment of phospholipids and decrease of sphingomyelin and ceramide were parallel across the increasing density and decreasing size of HDL sub-fractions (HDL2b, HDL2a, HDL3a, HDL3b, HDL3c). The differences in the lipid ratio across the HDL sub-fractions were demonstrated to affect the level of enzymatic activities, anti-oxidative, and anti-apoptotic capacity of the HDL sub-fractions, with HDL3c being the most functionally efficient in the aforementioned capacities [226, 227]. In a study where intermediates of cholesterol
32
CHAPTER 1 – LITERATURE REVIEW
metabolism in the lipoproteins were analysed, it was found that approximately 40% of total 24S-hydroxycholesterol in plasma were in LDL and HDL each and 10-20% was in VLDL; whereas, half of total 27-hydroxycholesterol and cholesterol were found in HDL and LDL, respectively [228]. The lipid composition of the lipoproteins from healthy subjects is summarised in Table 1.8. For the rest of Chapter 1, we will focus on the lipid of interest in this project, plasmalogen (section 1.4), and discuss on the lipidome of the diseased state, including in oxidised lipoproteins (section 1.5).
Table 1.8 Lipid composition of lipoproteins in healthy individuals.
Lipid class/name VLDL (%)e LDL (%)e HDL (%)e
Phosphatidylcholinea 8.1 29.9 62.0 Lysophosphatidylcholinea 1.7 11.0 87.3 Phosphatidylethanolaminea 18.6 21.3 60.1 Phosphatidylethanolamine- 11.6 28.5 59.9 plasmalogena Sphingomyelina 7.2 50.4 42.6 Ceramide d18:1a 15.6 60.3 24.1 Hexosylceramideb 8.0 49.1 42.0 Lactosylceramideb 8.2 46.4 44.4 Sphingosine 1-phosphate d18:1b 1.7 6.5 91.8 Sphingosine 1-phosphate d18:0b 1.9 3.3 94.7 Cardiolipinc 11.0 67.0 17.0 Phosphatidylserinec 9.0 81.0 7.5 24S-hydroxycholesterold 12.0 38.0 43.0 27-hydrocholesterold 5.0 31.0 47.0 Cholesterold 9.0 54.0 33.0
a determined by Wiesner et al. [224] b determined by Scherer et al. [225] c determined by Deguchi et al. [229] d determined by Buckard et al. [228] e Values are percentage of the sum of lipid species/class in respective lipoprotein fractions relative to sum of the lipid in circulation.
33
CHAPTER 1 – LITERATURE REVIEW
1.3.2 Overview of lipoprotein metabolism and functions
Knowledge of the physiology of the lipoprotein transport system is critical in understanding abnormalities of the lipoprotein metabolism. The lipoprotein transport system is divided into the exogenous and endogenous pathways. Lipoprotein metabolism is illustrated in Figure 1.3 and the functions of individual lipoproteins class are further discussed in the sections below. The exogenous pathway involves the transport of dietary lipids in chylomicrons from the intestine into circulation, whereas the endogenous pathway involves the hepatic production of lipoproteins (VLDL and nascent HDL) as well as the interaction between the lipoproteins in circulation.
1.3.2.1 Chylomicron and very low density lipoprotein
Nascent chylomicrons are produced in the enterocytes and lymph. Chylomicrons enter the circulation and then acquire apoC and E from other lipoproteins [230]. It is possible that there is an exchange of apoAI and apoAII of the nascent chylomicron for apoC and apoE, thus forming a mature chylomicron [231, 232]. In the endogenous pathway, VLDL is produced in the liver. It acquires apoC and apoE from HDL [233, 234], as well as cholesteryl ester via the action of CETP. In circulation, both chylomicron and VLDL undergo lipolysis catalysed by lipoprotein lipase thereby delivering fatty acids to tissues [235] such as adipose tissue and muscles for energy storage. Further lipolysis results in the production of smaller and denser chylomicrons, termed chylomicron remnants, which are removed from the circulation via the interaction of apoB/apoE and hepatic LDL-R or LDL-R-related protein. Only a fraction (about 50%) of the VLDL remnants is converted to LDL via lipoprotein lipase. The remaining VLDL remnants are taken up by the liver via VLDL-receptor.
1.3.2.2 Low density lipoprotein
LDL is the major carrier of cholesterol in circulation between the liver and extrahepatic tissues. LDL is metabolised from VLDL; the lipolysis of triglycerides in VLDL and the increase of cholesteryl ester content from HDL via CETP [236] results in the formation of mature LDL. LDL can be taken up by the liver and/or peripheral tissues by the interaction of apoB100 and LDL-R to maintain cholesterol homeostasis. In the condition where there is a low cellular level of cholesterol, the expression of LDL-R is increased [237]. When the cholesterol demand is met, the expression of LDL-
34
CHAPTER 1 – LITERATURE REVIEW
R is negatively regulated; this involves the suppression of HMG-CoA reductase activity and the increased esterification of cholesterols [236].
Studies reported that oxidised or minimally modified LDL but not native LDL induced the adhesion of monocytes to endothelial cells [238, 239], which is a cardinal feature of the onset of atherosclerosis. However, prolonged exposure of endothelial cells to pathophysiological levels (>180mg/dl) of native LDL was found to also cause monocytes adhesion via the induction of ICAM-1 [240]. These highlight the roles of the levels of oxidation, exposure and concentration of LDL in atherosclerosis.
1.3.2.3 High density lipoprotein
The majority of mature spherical HDL is produced via intravascular processes. Nascent HDL is synthesised primarily in the liver and intestine as lipid-free apoAI or lipid poor pre-β HDL which is a bi-layered phospholipid disc wrapped with two ring- shaped apoAI molecules. The nascent HDL readily acquires lipids from the peripheral cells via ATP-binding cassette transporter A1 (ABCA1). Cholesterol of pre-β HDL is then esterified by LCAT, thereby forming a neutral lipid core of cholesteryl ester and decreasing the density of the lipoprotein to form HDL3. Further ABCG1-mediated uptake of cholesterol and their esterification leads to the formation of larger HDL2 particles. In turn, HDL2 can be converted back to HDL3 via CETP-mediated transfer of cholesteryl ester to apoB-containing lipoproteins; via the endothelial- and hepatic lipases-mediated hydrolysis of the core triglycerides; and via SRB1-mediated selective uptake of HDL cholesteryl ester by the peripheral cells. In turn, smaller and/or lipid poor HDL is produced and it allows the shedding of apoAI for the next cycle of lipidation [168]. HDL can also be taken up by the liver via hepatic SRB1 or holoparticle HDL receptors for clearance from the circulation. Reverse cholesterol transport is a process whereby cholesterol is removed from peripheral tissues and transferred to HDL, which is transported back to the liver. This includes cholesterol efflux from macrophages in the intimal of arteries [241]. Reverse cholesterol transport contributes to the anti-atherogenic role of HDL and dysfunctional reverse cholesterol transport may contribute to the progression of atherosclerosis.
HDL has been suggested to be multifactorially atheroprotective, with several roles in addition to the reverse cholesterol transport, such as anti-oxidation, anti-inflammatory,
35
CHAPTER 1 – LITERATURE REVIEW
anti-thrombotic, and anti-apoptotic with the smallest and densest HDL3c being the most potent [227]. These functions typically involve HDL-associated enzymes, which are discussed in previous sections (1.3.1.2). HDL can protect LDL against oxidation; the anti-oxidative capacity was shown to be regulated by the lipoprotein phospholipid monolayer surface rigidity and the methionine residues of the apoAI [33]. Also, HDL was shown to limit the expression of pro-inflammatory signalling molecules tumour- necrosis factor-α and interleukin-1, which promote endothelial cell adhesion and lesion development [242]. Moreover, HDL was reported to prevent the apoptosis of endothelial cells that was induced by oxidised LDL [243]. While these findings are compelling evidence of the anti-atherogenicity of HDL, studies showed that HDL can turn pro-inflammatory when the lipid and protein composition were altered during disease states such as found in patients with type II diabetes [244] and end-stage renal disease [245] and during acute inflammation [246]. Modified HDL which is produced in vitro either via the oxidation of HDL with copper sulphate [247] or myeloperoxidase [248] as well as in subjects with CVD [249] showed a pro-atherogenic property where its ability to perform ABCA1-mediated reverse cholesterol transport was impaired. These findings highlight the role of oxidative stress in modifying lipid and protein components of lipoproteins under chronic inflammatory conditions that subsequently result in modulated functions.
36
CHAPTER 1 – LITERATURE REVIEW
Figure 1.3 Schematic representation of lipoprotein metabolism. In the intestine, dietary fats are absorbed and packaged as triglycerides in chylomicrons. Chylomicrons are transported to blood capillaries then undergo lipolysis to release triglycerides, which are taken up into adipose tissue and muscle cells for energy storage. The chylomicron remnants are taken up by the liver for clearance. ApoB-containing lipoproteins (VLDL) are synthesised by the liver and released into circulation where they mature to become LDL. LDL is taken up for clearance by hepatic LDL-R. HDL is produced by the liver and intestine as lipid poor apoAI or pre-β HDL. The nascent HDL acquires cholesterol from the peripheral tissues via ATP-binding cassette transporter A1 (ABCA1). The cholesterol is then esterified by LCAT, thereby forming mature HDL (HDL3). HDL3 can acquire more lipids from the peripheral cells via ATP-binding cassette sub family G member 1 transporter (ABCG1). The expression of ABCA1 and ABCG1 are regulated by liver X receptor (LXR). Further accumulation of esterified cholesterols leads to HDL2 formation. HDL2 can be converted back to HDL3 or the nascent HDL form by hepatic or endothelial lipase. Exchange of lipids between HDL and apoB-containing lipoproteins is mediated by PLTP and CETP. CETP-mediated removal of esterified cholesterols from mature HDL can produce triglyceride-rich HDL, which is a substrate for hepatic lipase. HDL is also taken up by the liver for clearance via SRB1 in the process of reverse cholesterol transport. Reverse cholesterol transport involving the hepatic uptake of HDL and lipolysis of HDL allows the shedding of apoAI.
37
CHAPTER 1 – LITERATURE REVIEW
1.4 PLASMALOGENS: BIOCHEMISTRY, DISEASE ASSOCIATIONS AND PLASMALOGEN MODULATION
1.4.1 Structure and oxidation chemistry of plasmalogens
Plasmalogens are a subclass of glycerophospholipids. They are characterised by a hydrophilic head group (ie. ethanolamine, choline, serine, or inositol) which is attached to the sn-3 position of the glycerol backbone via a phosphodiester bond; the alkyl chain at the sn-1 position is attached via a vinyl ether bond, which is a cis double- bond located adjacent to an ether bond, while the fatty acid chain at the sn-2 position is attached via an ester bond. Examples of the ethanolamine glycerophospholipid subclasses are shown in Figure 1.4. The alkenyl chain at the sn-1 position consists predominantly of 16:0, 18:0 or 18:1 carbon chains [250]. The vinyl ether linkage is acid labile and susceptible to ROS due to its relatively lower bond dissociation energies as compared to allylic and alkyl linkages [251]. The long chain fatty acids at the sn-2 position are predominantly polyunsaturated [252] in nature including DHA and arachidonic acid [250]. The bis-allylic linkage in a polyunsaturated fatty acid (PUFA) chain is formed from two double-bonded carbon units (─C═C─) separated by a single- bond carbon unit (─C─C─) [253]. The bis-allylic hydrogens which are attached to the single bond carbon unit has a lowest C-H bond dissociation energy in the fatty acid chain and is comparable to that of the vinyl ether bond [251].
Plasmalogens are considered to play a role as anti-oxidants in various diseases where oxidative stress is implicated such as in atherosclerosis [254], cancer [250, 255], as well as in aging [256]. However, the details of the lipid oxidation chemistry are not fully understood. A previous study has proposed the radical reactions of the lipid; this involves a vinyl ether peroxidation via a rapid initial formation of plasmalogen hemiacetal hydroperoxy radical, and a subsequent slow propagation of the radical reactions. The slow propagation is mainly due to the need of the bis-allylic methylene group of the fatty acid chain at the sn-2 position to be at a close proximity to the hydroperoxy radical for a hydrogen abstraction to occur [251] (Figure 1.5). The radical reaction can also be initiated at the bis-allylic methylene groups of the fatty acid chain at the sn-2 position, thus forming a hydroperoxy radical, which may react rapidly with the vinyl ether linkages of other plasmalogen molecules. However, the same requirement for propagation applies and so the reaction is proposed to be a slow process
38
CHAPTER 1 – LITERATURE REVIEW
[251]. Following the cleavage of the peroxyl radical at the vinyl ether group, potentially by circulating peroxidase, fatty aldehydes and lysophospholipids are produced [252, 257]. These oxidative by-products have been associated with increased inflammation and their elevated levels were detected in human atherosclerotic lesions [258].
Figure 1.4 Subclasses of glycerophosphoethanolamine. Phosphatidylethanolamine, alkylphosphatidylethanolamine and alkenylphosphatidylethanolamine are similar in their structures. However they can be differentiated via the bond linking the alkyl chain at the sn-1 position of the glycerol backbone; Phosphatidylethanolamine has an ester bond, alkylphosphatidylethanolamine has an ether bond, whereas alkenylphosphatidylethanolamine has a vinyl ether bond. Figure is adapted from Nagan and Zoeller, Progress of Lipid Research (2001), 40:199-229.
39
CHAPTER 1 – LITERATURE REVIEW
Figure 1.5 Proposed radical reactions of plasmalogen. Peroxidation of plasmalogen that occurs at the vinyl ether bond is a relatively fast reaction, resulting in the initial formation of plasmalogen hemiacetal hydroperoxy radical. The subsequent propagation is a slow process because it requires the hydroperoxy radical to come into close proximity with the bis-allylic methylene of the fatty acid chain at the sn-2 position of the glycerol backbone for hydrogen abstraction to occur.
1.4.2 Biosynthestic pathway of plasmalogen and the lipid distribution in cells and tissues
The biosynthesis of plasmalogens involves two main organelles, the peroxisome and the surface of endoplasmic reticulum as outlined in Figure 1.6. Plasmalogen biosynthesis starts with the esterification of dihydroacetone phosphate (DHAP) to a long chain acyl coA. The reaction is catalysed by DHAP-acyltransferase (DHAP-AT) [259]. The product, 1-acyl-DHAP, is then converted by alkyl DHAP-synthase (ADHAP-S) to 1-alkyl-DHAP, which is then transported to the surface of the endoplasmic reticulum. The long chain fatty alcohol which is needed in the second step
40
CHAPTER 1 – LITERATURE REVIEW
either comes from dietary intake or gets supplied by fatty acyl-coA reductase (Far1 or Far2), which uses NADPH as a co-factor [250]. Far1 enzyme is negatively regulated by the level of cellular plasmalogens [250, 260], and therefore, the formation and supply of long chain fatty alcohol is believed to be the rate limiting step of the biosynthetic pathway. The first two steps of plasmalogen biosynthesis are carried out by peroxisomal enzymes whereas the remaining steps are carried out on the surface of the endoplasmic reticulum.
Using NADPH as a co-factor, acyl/alkyl DHAP-reductase (AADHAP-R) reduces the ketone group at the sn-2 position of 1-alkyl-DHAP to form 1-alkyl-2-lyso-sn-glycero-3- phosphate (1-alkyl-G3P), the ether linked analogue of lyso-phosphatidate (1-acyl-2- lyso-sn-glycero-3-phosphate) [259]. This third step is a fusion point between the syntheses of plasmalogens and diacyl phospholipids. A line of evidence has shown that the AADHAP-R is localised in both the cytosolic sides of the peroxisome and endoplasmic reticulum, thus suggesting its importance for both pathways [250, 259, 261]. A long chain fatty acid is then esterified at the sn-2 position of 1-alkyl-G3P by acyl/alkyl-G3P-acyltransferase (AAG3P-AT), producing 1-alkyl-2-acyl-G3P in which the phosphate group is then removed by phosphohydrolase (PH) to form 1-alkyl-2-acyl- glycerol [250]. Addition of choline or ethanolamine head groups to 1-alkyl-2-acyl- glycerol is catalysed either by choline or ethanolamine phosphotransferase (C-PT or E- PT) to form 1-alkyl-2-acylglycerophosphocholine (1-alkyl-2-acyl-GPC) or 1-alkyl-2- acylglycerophosphoethanolamine (1-alkyl-2-acyl-GPE). Subsequently 1-alkyl-2-acyl- GPE is converted to alkenylphosphatidylethanolamine or PE plasmalogen by desaturation of the alkyl chain to form an alkenyl linkage. The biosynthesis of alkenylphosphatidylcholine (PC plasmalogen) is not as well characterised as PE plasmalogen; It is suggested that the former is derived from the latter [250, 259]. PC plasmalogen is synthesised by the reaction of CDP-choline and alkylglycerol (1- alkenyl-2-acyl-sn-glycerol), which is produced by the catalysed conversion of PE plasmalogen by PE plasmalogen specific phospholipase-C (PLC). It is also possible that PC plasmalogen was synthesised by the methylation of the head group of PE plasmalogen phosphatidylethanolamine N-acetyltransferase [262, 263].
The early steps of the biosynthetic pathway in the peroxisome can be bypassed via a dietary intake of 1-alkyl glycerol or alkylglycerol (Figure 1.6). This was first shown in
41
CHAPTER 1 – LITERATURE REVIEW
an in vitro study where Chinese Hamster Ovary cells that were made deficient in DHAP-AT and in turn deficient in plasmalogens, were incubated with 1-O-hexadecyl- sn-glycerol (0 - 20 µmol/l). This resulted in a dose-dependent accumulation of plasmalogen in the cells [264].
Plasmalogens are abundant in mammalian biological membranes such that they make up about 8 - 20 % of human total phospholipid mass [259, 265]. Plasmalogens compose 50 - 55 % of the total phospholipids in human heart [266, 267], brain [262, 268], mammalian spermatozoa [269, 270], as well as inflammatory cells [252, 262, 265, 271]. The proportions of PE and PC plasmalogens differ depending on the specific tissues. Fifty to seventy percent of ethanolamine glycerophospholipids in myelin sheath of the brain is made up of PE plasmalogen [259, 268]. Human heart contains similar levels of PE and PC plasmalogen, approximately 20 % for each plasmalogen relative to the total phospholipids [272]. In blood cells, PE plasmalogens content is higher in cells of the innate immunity such as neutrophils, than that of the specific immunity such as lymphocytes [252, 271]. Human plasma was reported to have 5 % of plasmalogen content of the total phospholipids [273]. Up to 4.5 % and 60 % of phosphatidylcholine and phosphatidylethanolamine respectively were found to consist of plasmalogens in the whole plasma, LDL, and HDL of normolipidemic donors; Forty two percent of the total plasmalogen in plasma was associated with LDL whereas, 36 % was associated with HDL [274]. In a recent study, it was reported that 28.5 % of PE plasmalogen in human plasma was associated with LDL, whereas 60 % was associated with HDL (Table 1.7) [224]. These findings are summarised in Table 1.9.
42
CHAPTER 1 – LITERATURE REVIEW
Figure 1.6 Biosynthetic pathway of plasmalogens. Plasmalogen synthesis involves the peroxisome and endoplasmic reticulum. The biosynthesis starts with the conversion of dihydroacetone phosphate (DHAP) to 1-acyl-DHAP and its further conversion to 1-alkyl-DHAP. 1-alkyl-DHAP is then transferred to the peroxisome and converted to 1-alkyl glycero-3-phosphate (1-alkyl-G3P) via acyl/alkyl DHAP-reductase (AADHAP-R) which resides in the cytosolic side of both peroxisome and endoplasmic reticulum. Multiple enzymatic steps occur along the surface of endoplasmic reticulum to form 1-alkyl-2-acyl-glycerolphosphatidylcholine or 1-alkyl-2-acyl- glycerolphosphatidylethanolamine. The latter is converted to PE plasmalogen by 1-desaturase. Subsequently, PC plasmalogen can be synthesised from PE plasmalogen by the reaction of CDP- choline and 1-alkenyl-2-acyl-sn-glycerol, which is produced by the catalysed conversion of PE plasmalogen by PE plasmalogen specific phospholipase-C (PLC). It is also proposed that phosphatidylethanolamine N-acetyltransferase (PEMT) can convert PE plasmalogen to PC plasmalogen by methylation of the head group. *Far1 is negatively regulated by plasmalogens. Several steps in the biosynthetic pathway can be bypassed via dietary intake of long chain fatty alcohol or 1-alkyglycerol.
43
CHAPTER 1 – LITERATURE REVIEW
Table 1.9 Cell and tissue distribution of plasmalogens.
Cells / tissues Distribution of plasmalogensa References Human heart 32 - 50 % [266, 267] (PC and PE plasmalogen make up approx. [272] 20% each)
Mammalian 55 % [270, 275] spermatozoa Human brain 20 - 50 % [259, 262, 268]
Inflammatory cells Up to 50 % [252, 262, PE plasmalogen content in cells of innate 265, 271] immunity > cells of specific immunity
Human plasma 5 % [224, 273, 28.5 % of PC and 60 % of PE are 274] plasmalogens 42 % and 36 % of total plasmalogens are associated with LDL and HDL, respectively a Distribution of plasmalogens based on total % of phospholipids of respective tissues/cells.
1.4.3 Biological functions of plasmalogens
1.4.3.1 Plasmalogens as membrane components
Plasmalogens play an important role as a structural membrane component and contribute to the membrane dynamics. Plasmalogen in particular the PE species composes a major lipid constituent in membranes and cells that undergo rapid membrane fusion such as synaptic vesicles. The lipid has a marked propensity for adopting a conformation which lowers the activation energy required for the formation of putative fusion intermediates; It was demonstrated that vesicles containing equimolar mixtures of PE plasmalogen and phosphatidylcholine allowed more rapid membrane fusion events than those containing equimolar mixtures of the diacyl analogues (i.e. phosphatidylethanolamine and phosphatidylcholine). The rate was particularly faster when vesicles containing PE plasmalogen with arachidonic acid were used [276].
Plasmalogens were also shown to be crucial for vesicular cholesterol transport and homeostasis as studies demonstrated a disruption to these functions in cells that were deficient in plasmalogens; A study using NRel-4 cells, which are defective in DHAP- AT, showed that the dysfunctional cholesterol transport phenotype was due to
44
CHAPTER 1 – LITERATURE REVIEW
deficiency in PE plasmalogen. The cellular function was restored when PE plasmalogen level was normalised by transfecting the cells with c-DNA of DHAP-AT gene [277]. It was demonstrated that the impairment in the cholesterol transport system occurred in the transport of cholesterol from cellular membrane to acetyl co-enzyme A acetyltransferase in the endoplasmic reticulum [277]. The precise roles of plasmalogen in the cholesterol metabolism still require further investigations as acetyl co-enzyme A acetyltransferase activity and its lipid milieu does not require PE plasmalogens.
1.4.3.2 Plasmalogen as a potential anti-oxidant
Plasmalogen is proposed to be a potential anti-oxidant because of three main characteristics: (1) the enhanced electron density of the vinyl ether bond at the sn-1 position that makes it more susceptible to ROS as compared to other types of linkages; (2) the position of the vinyl ether linkage which is proposed to be in the hydrophillic domain of the membrane [252], and so is a target to ROS; and (3) the proposed slow propagation of the plasmalogen hemiacetal hydroperoxy radicals [251] (Figure 1.5). Macrophage like cells, RAW.12 and RAW.108 that were mutant in the peroxisomal systems demonstrated susceptibililty to chemical hypoxia and ROS generators [278]. Both types of cells showed resistance to the insults after they were supplemented with alkylglycerol which normalised their cellular plasmalogen levels [264, 278]. In vitro studies demonstrated that enrichment of LDL with PE or PC plasmalogen resulted in the delay in the copper-catalysed conjugated diene formation of LDL [242, 279]. In addition, the oxidation of PUFA of diacyl phospholipids were delayed in the presence of plasmalogens; It was shown that the products of plasmalogen degradation did not propagate the oxidation of PUFA [280]. Alpha-fatty aldehyde is an oxidative by-product of plasmalogen from the cleavage of alkenyl chain at the sn-1 position. Studies showed that α-fatty aldehydes can form Schiff base adducts with phosphatidylethanolamine; It was suggested that the oxidative by products could potentially harm cells in the in vivo setting [281]. The effects of plasmalogen oxidative by products remain controversial and warrant further investigation.
1.4.3.3 Plasmalogens as lipid mediators
Plasmalogens are proposed to act as sinks for PUFA in the sn-2 position of the glycerol backbone [259]. Mediated by phospholipase A2 especially during hypoxia or
45
CHAPTER 1 – LITERATURE REVIEW
ischemia [259], the release of the fatty acids (DHA or arachidonic acid), and the concomitant production of lyso-plasmalogens, act as lipid mediators for further cellular signalling activities. Arachidonic acid serves as a precursor for the formation of eicosanoid [252], that leads to the production of thromboxane, prostaglandins, and leukotrienes which are essential in the immune and inflammatory regulatory functions [250]. DHA is precursor for anti-inflammatory lipid mediators resolvins and protectins which help to terminate acute inflammation in tissues by the removal of chemokines and regulation of leukocyte infiltration [282, 283]. Resolvins and protectins have also been demonstrated to alleviate obesity-induced insulin resistance and hepatic steatosis [284]. DHA is proposed to play a role in the vesicle formation during the release of neurotransmitters [268], whereas lyso-plasmalogens induce changes in the membrane permeability, allowing the influx of Ca2+ via plasma channel. Subsequently, this enables the translocation of Ca2+-dependent enzymes and other second-messengers for further cellular signal transduction [252].
1.4.4 Association of plasmalogen to cardiovascular disease
Earlier studies showed that there was an accumulation of plasmalogen-derived oxidation products such as α-chloro fatty aldehydes (2-chlorohexadecanal) and unsaturated lysophosphatidylcholine [258], as well as lysophosphatidylcholine- chlorohydrins [285] in human atherosclerotic lesions as compared to normal aortas. It was also shown that 2-chlorohexadecanal was derived from HDL which was oxidised via a myeloperoxidase system-generated hypochlorite [286]. These findings have provided an association of plasmalogen oxidative by-products to CVD. However, the association of plasmalogen itself with the disease was lacking. This association of plasmalogen with CAD was demonstrated by Meikle et al. where analyses on plasma lipids of patients with either stable angina or acute coronary syndrome (ACS) was conducted; It was shown that PC plasmalogen was 20 % lower in the stable angina patients as compared to those in the healthy controls. In addition, the level of PE plasmalogen was 19 % lower in the ACS patients (ie. unstable angina) as compared to those with stable angina [254]. This study suggested a depletion of plasmalogens in the patients as compared to the healthy controls and might reflect the differing levels of oxidative stress between patients with stable angina and ACS.
46
CHAPTER 1 – LITERATURE REVIEW
1.4.5 Modulation of plasmalogen in vitro and in vivo
Oxidative stress underlies the progression of various aging diseases including atherosclerosis. As mentioned above, the study by Meikle et al. suggested that there was an elevated level of oxidative stress in the CAD patients and that plasmalogens as potential endogenous anti-oxidants, in these patients were depleted. It is also possible that the synthesis of plasmalogen is downregulated with age due to increased oxidative stress and this makes individuals become more susceptible to disease progression. In light of the earlier studies, it can be postulated that by regulating the levels of plasmalogen, oxidative stress levels and in turn, disease progression can both be attenuated. This may be achieved when the levels of plasmalogen and oxidative stress are in equilibrium. Alternatively, the levels of plasmalogens shall be more than sufficient to counter-act the effects of oxidative stress. In this case, plasmalogen can act as a first layer of protection for the cells from the damaging effects of ROS.
Alkylglycerol, a precursor to plasmalogen, is naturally abundant in shark liver oil [287], consisting of 10 % the total esterified deep sea shark liver oil [288]. Analysis of the shark liver oil alkyglycerol fraction revealed that 1-O-octadecenylglycerol (18:1) was the predominant lipid species which made up 76 % of the fraction, followed by 1-O- hexadecenylglycerol (16:1), 1-O-hexadecylglycerol (16:0), and 1-O-octadecylglycerol (18:0) which made up 6 %, 5 %, and 3 % of the fraction, respectively [288]. A cell culture study on human pulmonary arterial endothelial cells, which were supplemented with 1-O-hexadecylglycerol demonstrated an approximately two-fold increase in cellular plasmalogen levels, as well as increased resistance to ROS such as hydrogen peroxide, hyperoxia, and the superoxide generator plumbagin [287, 289]. In humans and rodents, dietary alkylglycerols were absorbed and metabolised in the intestinal mucosal cells without cleavage of the ether bond, and they were found to be incorporated into tissue plasmalogens [290]. Modulation of plasmalogen levels in vivo as a disease remedy was demonstrated whereby the administration of shark liver oil- enriched diets to rats with IHD and hypertension resulted in improved lipid profile, immunological response, and clinical symptoms [287]. Furthermore, there was a reduction in the ischemia reperfusion injury to the hearts of Sprague Dawley rats that was attributed to increased plasmalogen biosynthesis when prior to ischemia, they were supplemented with 1-O-hexadecylglycerol [291].
47
CHAPTER 1 – LITERATURE REVIEW
1.5 LIPIDOMICS AS AN APPROACH TO STUDY THE PATHOGENESIS OF ISCHEAMIC HEART DISEASE
The majority of this section of the literature review has been published - Rasmiena, A.A, Ng, T.W, Meikle, P.J. Metabolomics and ischaemic heart disease (2013). Clinical Science, 124(5), 289-306.
1.5.1 Lipidomics
Metabolomics is now a fast growing field in modern integrated science. It represents a directional shift in metabolic research from approaches which concentrated on single pathways to those which attempt to gain a comprehensive understanding of complex metabolic networks [292]. Metabolomics can provide another viewpoint of disease mechanisms, in the form of metabolite levels and flux, which can be integrated into the existing knowledge gained via genomic and proteomic approaches. Metabolomics is well suited to the study of chronic disease at a population level where samples are often analysed using mass spectrometry. Mass spectrometry (MS) detects and quantifies metabolites based on their mass-to-charge (m/z) ratio and signal intensity of their gas-phase ions. The development of MS technology over the past decade has largely contributed to the current advanced state of metabolomics and its growing application in clinical screening and diagnostics. The high-throughput nature of the technology enables its application in epidemiological studies to identify biomarkers and define metabolic pathways related to disease [293].
Lipidomics is a subset of metabolomics that specialises in the characterisation of the lipid complement in biological systems. The direct association between dyslipidemia and altered lipid metabolism with metabolic diseases, including type II diabetes and CVD, has prompted multiple studies in this field. Improvements in MS instrumentation have reduced analysis time and facilitated the high throughput profiling of the lipidome in larger sample sets. A number of approaches have been used for the analysis of the lipidome: (1) "shotgun lipidomics", which involves direct infusion of the sample followed by the application of multiple scan modes to obtain unbiased detection of lipids with high sample throughput [294, 295]; (2) ''targeted lipidomics'', which employs multiple reaction monitoring and stable isotope internal standards to obtain accurate and precise measurements of known lipids of interest [223, 296]; (3) ''untargeted lipidomics'', which utilises high mass accuracy to identify previously unknown lipid
48
CHAPTER 1 – LITERATURE REVIEW
species. In our laboratory we have developed a targeted lipidomics approach which can quantify over 300 lipid species from 10 l of human plasma [297].
Tandem mass spectrometry employs collision-induced dissociation (CID) to obtain fragments of ions that are characteristic of the lipid of interest. These fragments then allow the identification of a lipid class by precursor or neutral loss scanning [298]. For glycerophospholipids, CID in the positive ion mode principally results in the removal of the polar head group as a neutral or charged species. Phosphatidylcholine and phosphatidylethanolamine are usually analysed by the product ion scan of 184 Da and neutral loss of 141 Da, respectively. The same method can be applied to the analysis of PC and PE plasmalogens. However, this methodology lacks the ability to differentiate diacylphospholipids or alkylphospholipids with plasmalogens in a complex lipid mixture. PE plasmalogens particularly do not show a strong loss of 141 Da during CID; this is likely due to the vinyl ether substituent altering the favourable CID mechanism of the neutral loss of 141 Da [299]. Furthermore, plasmalogens are not able to form a 6- membered ring transition state that allows the carbonyl oxygen from the sn-1 and sn-2 position to attack the sn-3 methylene carbon that would otherwise result in the neutral loss of phosphoethanolamine [300]. To counteract this problem, a previous study has investigated two prominent CID fragment ions of the [M+H]+ of PE plasmalogens that were characteristic of the sn-1 and sn-2 positions for various plasmalogen species. These two ions were subsequently used to detect specific molecular species of PE plasmalogens [299]. However, PC plasmalogens produce primarily a single product ion of m/z 184, and very little fragment ions characteristic of the sn-1 and sn-2 positions [301]. In order to obtain structural detail of the alkenyl and acyl chains in PC plasmalogens and diacylphosphatidylcholine, lithium adducts are used which provide stronger signals of fragment ions characteristic of the sn-1 and sn-2 positions [302].
Advanced lipidomic technologies generate vast amounts of data. The extraction of data from chromatograms to data matrices, statistical analysis, visualisation, data/result interpretation and integration of “omics” datasets with other datasets are growing challenges and have been reviewed extensively by Oresic et al. and Eliasson et al. [303, 304]. The typical workflow of lipidomic analysis is outlined in Figure 1.7.
49
CHAPTER 1 – LITERATURE REVIEW
Figure 1.7 Typical workflow of lipidomic studies. Samples are collected and, where necessary, lipids are extracted often with the addition of internal standards and/or derivatisation. Lipids are then quantified (liquid or gas chromatography coupled with MS and/or NMR spectroscopy). Raw data are extracted, used for lipid identification and further processed prior to statistical analysis. Bioinformatic tools are used to identify associations with disease states and outcomes, determine significant correlations, characterise metabolic signatures and to integrate results with existing biological knowledge.
1.5.2 Lipidomics in studies of ischaemic heart disease
There are several studies that will be discussed in details below and the key findings of the studies are summarised in Table 1.10.
1.5.2.1 Lipidomics of inflammation and oxidative stress
Based on immunological studies [305, 306], oxidative stress has been shown to result in the modification of the lipid and protein components of LDL. This modification translates to atherogenic and dysfunctional forms of LDL, which are subsequently taken up by macrophages via scavenger receptors, such as CD36 [307], thus leading to foam cell formation and progression to atherosclerotic plaque as previously discussed. Some of the early lipidomic studies involved the identification of oxidation products from human plasma. Nine species of oxysterols (cholesterol oxidation products) were detected and identified in human plasma from 31 subjects
50
CHAPTER 1 – LITERATURE REVIEW
using gas chromatography-MS [308]. Using the same approach, 7ß-hydroxycholesterol, a predominant cholesterol oxidation product in membranes and lipoproteins, as well as seven other cholesterol oxidation products, were detected in serum from 20 subjects from the Kuopio Atherosclerosis Prevention Study (KAPS) with progressive carotid atherosclerosis, further supporting the association of lipid oxidation with atherogenesis [309]. More recently, 16 species of phosphatidylcholine and lysophosphatidylcholine oxidation products were identified in human plasma from five male subjects with alcoholic liver disease (also associated with increased oxidative stress) using quadrupole-time of flight MS [310].
To further elucidate the mechanism of lipid accumulation and foam cell formation, several metabolomic studies were focusing on the macrophage scavenger receptor CD36. CD36 is a glycosylated membrane protein, expressed on the surface of cells, such as microvascular endothelial cells and macrophages [311, 312] where it is involved in the uptake of oxidised lipoproteins (oxLDL and oxHDL) [312, 313]. A study demonstrated that a novel class of oxidised phospholipids with sn-2 acyl group, possessing terminal γ-hydroxy (or oxo)-α-β-unsaturated carbonyl groups (oxPCCD36) serve as high-affinity ligands for CD36 [307]. Further in vitro analyses of the oxPCCD36 showed that they promoted CD36-dependent macrophage binding and foam cell formation when incorporated into cholesterol-laden particles [314] as well as activation of platelets at pathophysiological levels [315]. In addition, their levels were markedly elevated in plasma of humans with low HDL levels [315]. The oxPCCD36 were also found to be enriched in atherosclerotic lesions in rabbits [314]. Therefore, oxPCCD36 appear to function in atherogenesis by mediating the recognition and uptake of oxidised forms of LDL via CD36 on macrophages.
Lp-PLA2 is of interest in atherosclerosis studies because of its increased expression in vulnerable atherosclerotic lesions [316, 317], and the hypothesised pro-inflammatory nature of its enzymatic products such as lysophosphatidylcholine and oxidised non- esterified fatty acids [318, 319]. Previous studies have shown elevated levels of lysophosphatidylcholine in both sera of patients with atherosclerosis [320] and in human atherosclerotic lesions [258]. Twenty nine different species of oxidised phosphatidylcholine (oxPC) and lysophosphatidylcholine were identified from oxidised human LDL that made up the substrates and/or products for Lp-PLA2 by employing a
51
CHAPTER 1 – LITERATURE REVIEW
direct infusion (shotgun) approach on an electrospray ionisation tandem mass spectrometry (ESI-MS/MS) platform [200]. These species were differentiated into three classes: (i) short-chain oxPC, which were substrates for Lp-PLA2; (ii) long-chain oxPC, which were also substrates for Lp-PLA2, but were less efficiently hydrolysed; and (iii) saturated and monounsaturated lysophosphatidylcholine, which were predominant phosphatidylcholine products of Lp-PLA2. The concentrations of the three classes of lipids were shown to increase during LDL oxidation. To assess the physiological relevance of these findings, they identified 26 out of the 29 different species in human carotid artery plaque samples (n = 5), though their quantities differed between plaque samples [200]. Secondary attack on lysophosphatidylcholine by HOCl can lead to the oxidative production of lysophosphatidylcholine-chlorohydrin [252]. Messner et al. identified lysophosphatidylcholine-chlorohydrin (16:0) and (18:0), and demonstrated that their concentrations were increased by 67-fold and 82-fold respectively in human atherosclerotic tissue (n = 2) as compared to normal tissue (n = 2) [321]. These findings further support the concept of increased levels of oxidative stress within atherosclerotic lesions leading to elevated levels of oxidative by-products.
Cholesteryl esters are neutral lipids that make up a major component of LDL particles [224, 322]. Recent metabolomic studies have identified oxidised cholesteryl esters (oxCE) in atherosclerotic lesions [323, 324]. By employing ESI-MS/MS, seven groups of abundant oxCE species, including novel oxidation products, were identified from human atheromata (n = 6) [324]. One of the three most abundant oxCE species was identified as the CE (20:4)-derived 15-hydroxy-eicosatetraenoate (15-HETE) [324]. Consistent with this finding, Gertow et al. identified 15-HETE as the most abundant arachidonic acid-derived oxidative product of 15-lipoxygenase. They also determined that the 15-lipoxygenase mRNA was more highly expressed in atherosclerotic lesions from symptomatic subjects (n = 102) compared with asymptomatic subjects (n = 30) [323]. However, in contrast to Gertow et al., Hutchins et al. did not propose the implication of 15-lipoxygenase in atherosclerosis based on their findings of regio- and stereospecificiy studies. Instead, a non-enzymatic mechanism involving free radicals was suggested to dominate during atherosclerosis development [324]. Further studies are warranted to clarify the aforementioned mechanisms, which may not be mutually exclusive.
52
CHAPTER 1 – LITERATURE REVIEW
These studies exemplify how lipidomics can define the relationship of oxidation and inflammation with IHD through the comprehensive and systematic characterisation of oxidised lipids and the subsequent identification and quantification of these lipids in biological samples. The findings of the studies are summarised in Table 1.10.
1.5.2.2 Lipidomic assessment of atherosclerotic plaque
Atherosclerotic plaque represents an end-product of the metabolic abnormalities associated with IHD and as such, it provides a metabolic readout of the altered metabolism. There have been a number of studies aimed to characterise the plaque metabolome. By employing ESI-MS/MS, targeted quantitative analysis of atherosclerotic plaque identified three predominant species out of 10 free fatty acids present in plaques, linoleic acid (C18:2), oleic acid (C18:1), and palmitic acid (C16:0) [325]. Consistent with this finding, Mas et al. utilised MS-imaging techniques to characterise both the location and composition of non-esterified fatty acids and they found significantly elevated levels of C18:2, C18:1, and C16:0 in the neointima of atheromatous plaques of type II diabetic patients as compared to non-diabetics (total n = 40) [326]. The local enrichment in plaque of non-esterified fatty acids, especially
C18:2, was accompanied by increased expression of Lp-PLA2 and MCP-1 [326], which are implicated in inflammation and tissue injury. In addition, C18:2 triggered the activation of nuclear factor KB (NF-KB)as well as the expression of Lp-PLA2 and NF-
KB, suggesting a bidirectional relationship between inflammation and non-esterified fatty acids [326]. Together, these findings support the link between local inflammation, which is accompanied by altered metabolite composition, and atherosclerotic plaque progression.
The majority of myocardial infarcts result from the rupture of the atherosclerotic fibrous cap [327, 328]. Much is known about the cellular and pathophysiological causes of this rupture, such as the production of collagenases triggered by inflammatory mediators interferon-γ [329], and the different levels of shear stress in the carotid artery that differentiate stable/unstable areas of a maximum stenosis [330]. However, the molecular mechanism that precedes plaque instability is not fully understood. In an attempt to shed light on this, a study aimed to differentiate the lipid compositions of atherosclerotic plaques from carotid endarterectomy samples of symptomatic and asymptomatic patients, and stable and unstable areas of the same symptomatic
53
CHAPTER 1 – LITERATURE REVIEW
atherosclerotic lesions [331]. Employing a shotgun lipidomic approach on a nanoflow ESI-MS/MS system, 150 lipid species from nine different classes were identified and quantified from the atherosclerotic plaque samples. Compared to control radial arteries (n = 3), plaque samples (n = 3) were characterised by the enrichment of cholesteryl ester, phosphatidylcholine, lysophosphatidylcholine and certain sphingomyelin species. The difference in the lipid composition between the unstable and stable areas of plaques (n = 8) was statistically significant for several species including CE(18:0), CE(20:3), SM(d18:1/15:0), and PC(36:4) [331]. System-wide analysis revealed plaque specific lipid signatures containing 19 lipid species for the asymptomatic-symptomatic lesions (n = 6 per group), and 12 lipid species in the stable-unstable plaque areas [331].
The study by Stegemann et al. represents a milestone in our understanding of plaque biology by detailing the dynamic nature of plaque composition. Future studies in this area may ultimately give us an insight to the lipid signatures defining plaque vulnerability/stability, as well as the metabolites involved in the inflammatory cascade of atherogenesis. The key findings from the studies above are summarised in Table 1.10.
1.5.2.3 Lipidomics and lipid metabolism
The above studies demonstrate that oxidation of lipids within plaque is a major metabolic process and that these same lipids can also be isolated from plasma suggesting that they can leave the plaque tissue. However, the primary contributing source of plaque lipids is from circulation in the form of lipoproteins that cross the arterial wall to the intima. As such, the analysis of plasma lipids may provide important insight into metabolic processes that can contribute to plaque progression.
Using liquid chromatography (LC) coupled with tandem mass spectrometry (ESI- MS/MS), a targeted lipidomics study by Meikle et al. has profiled 305 known lipid species in plasma from 220 individuals representing stable angina (n = 60); unstable coronary syndrome or ACS (n = 80) and matched healthy controls (n = 80) [254]. Multiple lipid species that were associated with either stable disease (relative to healthy control) or unstable disease (relative to stable disease) were identified. In addition to cholesteryl ester and triglycerides, ceramide, phosphatidylinositol and phosphatidylethanolamine species were shown to be positively associated with stable
54
CHAPTER 1 – LITERATURE REVIEW
CAD while lysophosphatidylcholine, alkylphosphatidylcholine and PC plasmalogen species were negatively associated with stable disease. In contrast, we observed phosphatidylinositol to be negatively associated with unstable CAD along with the alkylphosphatidylethanolamine and PE plasmalogen species (Table 1.10). These findings amongst previous reports point to the involvement of a number of lipid metabolic pathways in the inflammation and oxidative stress that may be contributing to disease onset and progression. Interestingly, while lysophosphatidylcholine, generated by the action of Lp-PLA2, has been observed to be elevated in plaque, Meikle et al. observed this lipid to be negatively associated with both stable and unstable CAD in circulation. One possible explanation for this apparent contradiction may be the source of the lysophosphatidylcholine, which in circulation is partially derived from the action of LCAT during the process of reverse cholesterol transport. A recent study from Duivenvoorden et al. reported that LCAT deficiency was associated with accelerated atherogenesis and the findings by Meikle et al. may also reflect decreased LCAT activity in those with stable and unstable CAD [332].
As discussed in section 1.4, plasmalogen is proposed to play a role as an anti-oxidant because the vinyl ether linkage and the high proportion of PUFA at the sn-2 position of the glycerol backbone make it susceptible to oxidation. The negative association of PC and PE plasmalogen observed with stable and unstable CAD in this study, thus suggested that there was an elevated level of oxidative stress in these patients. Indeed, the decreased levels of these PE plasmalogen species in the unstable CAD group and in turn, increased levels of lysophosphatidylethanolamine species, LPE 20:4 and LPE 22:6 suggested the action of ROS on the plasmalogens [254]. These findings highlight the metabolism of plasmalogen in atherosclerosis. But more importantly, together with other previous reports, these findings raise the question of the role of plasmalogen as lipid anti-oxidants in the context of atherosclerosis. We therefore postulate that plasmalogens levels may affect oxidation of other lipids and inflammation pathways involved in atherosclerosis and that regulation of plasmalogen level can modulate disease progression and outcome.
55
CHAPTER 1 – LITERATURE REVIEW
Table 1.10 Major findings from lipidomic studies of ischaemic heart disease.
Area of Samples Principal findings References investigation analysed Oxidation Plasma 16 species of oxPC and lysophosphatidylcholine found [310] in plasma of patients with alcoholic liver disease 9 species of oxysterols found in healthy volunteers [308] 7ß-hydroxycholesterol found in KAPS subjects with a [309] fast progression of carotid atherosclerosis OxLDL 8 major oxPC species derived from 1-palmitoyl-2- [314] arachidonyl-PC and 1-palmitoyl-2-linoleoyl-PC that together constitute oxPCCD36 that are recognised by macrophage CD36 scavengers. Elevated levels of 29 species of saturated and mono- [200] unsaturated lysophosphatidylcholine, short chain oxPC and long chain oxPC that make up the substrates/products of Lp-PLA2 Plaque OxPCCD36 which is recognised by macrophage CD36 [307] scavengers were enriched in rabbit atherosclerotic lesions [200] Elevated levels of 26 out of 29 species of Lp-PLA2- specific saturated and mono-unsaturated lysophosphatidylcholine, short chain oxPC and long chain oxPC found in human atherosclerotic lesions 67 fold and 82 fold increase of [321] lysophosphatidylcholine -chlorohydrin (16:0) and (18:0) respectively in human atherosclerotic tissues as [323] compared to normal tissues [324] Increased level of oxylipin 15-HETE in human atheromata 7 groups of oxCE found human atheromata
Inflammation Plasma Increase of palmitic acid (C16:0, 8-fold), stearic acid [326] (C18:0, 3-fold), and 1-monolinoleoylglycerol (C18:2- glycerol, 3-fold) respectively in patients with stable atherosclerosis than in healthy subjects Plaque Increase in linoleic acid (C18:2), oleic acid (C18:1), [325] palmitic acid (C16:0) in human atherosclerotic plaques
Symptomatic/ Plasma Comparison of patients with unstable CAD vs. stable [254] asymptomatic CAD showed: increase in total ceramide, ischaemic dihexosylceramide, trihexosylceramide, odd chain heart disease phosphatidylcholine, phosphatidylethanolamine, free cholesterol and diacylglycerol; and decrease in total monohexosylceramide, phosphatidylglycerol, phosphatidylinositol, alkylphosphatidylcholine, PC plasmalogen, alkylphosphatidylethanolamine, PE plasmalogen, lysophosphatidylcholine, lysophosphatidylethanolamine, cholesteryl ester and triacylglycerol. Plaque 19 lipid species signature in [331] symptomatic/asymptomatic lesions and 12 lipid species signature in stable/unstable plaque areas Enrichment of cholesteryl ester, lysophosphatidylcholine, phosphatidylcholine and certain sphingomyelin in carotid endarterectomies as compared to control radial arteries
56
CHAPTER 1 – LITERATURE REVIEW
1.6 HYPOTHESES AND STUDY OBJECTIVES
1.6.1 Statement of research questions
Oxidised LDL is one of the contributing factors to atherosclerosis progression. While some studies have identified several lipids as oxidative by-products of the lipoprotein oxidation, they failed to demonstrate the global change to the major lipid classes in the lipoprotein. At present, the molecular detail of the lipoprotein oxidation is still not fully understood. Specifically, major oxidised lipids and changes to the native lipids in the lipoproteins have not been identified. Furthermore, the initial failures with HDL-raising therapy to improve atherosclerosis highlight that the quantity of the lipoproteins did not necessarily reflect their function; “Quality over quantity” of the lipoproteins matters in this case. Lipids make up the bulk of lipoprotein structures and so lipids are thought to play a role in the lipoprotein function. However there is a lack of understanding of the relationship between the lipid composition and lipoprotein function. In addition, the effect of oxidative stress associated with the disease on lipoprotein structure and function is not clear.
Plasmalogens have been previously identified in our laboratory as lipids that were negatively associated with CAD. The potential anti-oxidant and atheroprotective roles of plasmalogens and their decreased level in circulation in CAD suggested that there was a heightened level of oxidative stress in these patients. The effects of plasmalogen modulation have been demonstrated in cancer, peroxisomal disorder and lipoprotein cholesterol efflux and anti-oxidant capacity, but its effect on atherosclerosis, oxidative stress and inflammation associated with the disease is still unknown.
My PhD project will address these research questions using comprehensive experimental designs, suitable murine models and cell culture, combined with advanced mass spectrometric based lipidomic methodologies. Better understanding on the effect of oxidation on lipoprotein lipid composition and how this relates to functions will provide the opportunity for the development of new HDL therapeutics where the lipid composition of such formulations may influence functionality and thereby efficacy. In addition, insight into the effect of plasmalogen modulation on inflammation, oxidative
57
CHAPTER 1 – LITERATURE REVIEW
stress and atherosclerosis will provide the relevance for the use of plasmalogen as a potential therapy to reduce CVD risk.
1.6.2 Hypotheses and study aims
Our first hypothesis was that oxidative stress in atherosclerosis affects lipoprotein structure and function. Here we aimed to characterise changes in the lipid composition and function of oxidised lipoproteins (oxLDL and oxHDL). Lipidomics offers a global view of the changes in the lipid composition of biological systems. This methodology provides a wealth of information and valuable insight into the lipid metabolism during oxidation and/or inflammation leading up to atherosclerosis or in the context of treatment. One of the primary objectives of this study was to identify major oxidised lipids and changes to the native lipids in the lipoproteins due to oxidation and how these changes related to the lipoprotein function. To identify new major oxidised lipids, I used an untargeted lipidomic approach to analyse lipids in oxidised LDL. Metabolic changes as a result of oxidation to known native lipids in the lipoprotein were analysed using an established in-house LC-ESI MS/MS based targeted analysis. I also developed a methodology to assess the ability of HDL to transfer oxidised lipids from oxidised LDL, and used it along with other existing methodologies to assess the effect of oxidation on HDL function. Our second hypothesis was that regulation of the level of plasmalogen can influence atherosclerosis progression. Here we aimed to assess the ability of plasmalogen to prevent atherosclerosis in mouse models with differing levels of oxidative stress. We used murine models of atherosclerosis with differing levels of oxidative stress such as ApoE-/- and ApoE-/-GPx1-/- mice and modulated their level of plasmalogen via oral administration of alkylglycerol. I then assessed the effect of the treatment on several end-points including the accumulation of atherosclerotic plaques, and the levels of aortic oxidative stress and inflammatory markers.
Our third hypothesis was that the regulation of plasmalogen level can modulate inflammation. Here we aimed to assess the effect of the modulation of plasmalogen level on the tissue lipid metabolism and circulating immune cells. I used the aforementioned lipidomic mass spectrometric-based targeted methodology to analyse the lipid composition of the mouse tissues harvested from the study above. In a separate
58
CHAPTER 1 – LITERATURE REVIEW
study, I used flow cytometry technique to analyse the circulating immune cells including monocytes and neutrophils in alkylglycerol-treated and untreated mice.
59
CHAPTER 2 - MATERIALS AND GENERAL METHODS
60
CHAPTER 2 – MATERIALS AND GENERAL METHODS
2 MATERIALS AND GENERAL METHODS
2.1 GENERAL CHEMICALS
Table 2.1 General chemicals.
Reagents Supplier [3H]-cholesterol Bioscientific P/L 2,2,2-tribromethanol Sigma Aldrich 2-thiobarbituric acid Sigma Aldrich Ammonium formate Sigma Aldrich Butanol Merck Butylated hydroxytoluene (BHT) SAFC Chloroform Merck Copper (II) chloride Sigma Aldrich Copper (II) sulphate Sigma Aldrich Ethylene diaminetetra-acetic acid di-sodium salt (EDTA) Ajax Finechem Hydrogen peroxide 30% LabServ Mayer's haematoxylin Sigma Aldrich Methanol Merck Phorbol 12-Myristate 13-Acetate (PMA) Sigma Aldrich Potassium bromide (KBr) Sigma Aldrich Potassium chloride (KCl) Merck
Potassium phosphate dibasic (KH2PO4) Sigma Aldrich RPMI 1640 media Gibco Sodium bromide (NaBr) Sigma Aldrich Sodium chloride (NaCl) Amresco
Sodium phosphate monobasic (Na2HPO4) Sigma Aldrich Sucrose Sigma Aldrich Sudan IV ProSciTech Tert-amyl alcohol Sigma Aldrich Tetrahydrofuran Sigma Aldrich Thrichloroacetic acid BDH, Analar TO-901317 Sigma Aldrich Tris Amresco
61
CHAPTER 2 – MATERIALS AND GENERAL METHODS
2.2 GENERAL METHODS
2.2.1 Isolation and handling of lipoproteins
2.2.1.1 Isolation of blood plasma
Five healthy non-diabetic normolipidemic male (n = 4) and female (n = 1) volunteers aged between 25 to 65 years were recruited for the study. All subjects provided informed written consent. They fasted overnight (≥12 h) prior to the blood collection. Venous blood was collected from the antecubital vein into sterile EDTA tubes (1.8 mg/ml of blood) by the venipuncture butterfly technique. Blood was centrifuged within 2 h of collection at 3,041 xg for 15 min to separate the plasma and the red blood cells. Plasma from the different volunteers were measured for levels of cholesterol, LDL-C, HDL-C, triglycerides, and fasting blood glucose using commercial enzymatic kits on a COBAS Integra 400 Plus blood chemistry analyser (Roche Diagnostics, Australia). Plasma was then pooled and stored at -80oC.
2.2.1.2 Fractionation of lipoproteins from human plasma
Plasma was fractionated into VLDL, LDL, HDL by sequential ultracentrifugation using a method adapted from Havel et al. [162]. Briefly, plasma was thawed overnight at 4oC. Subsequently sucrose was added to plasma to give a final concentration of 0.6% (w/v) to prevent lipoprotein aggregation during ultracentrifugation [333]. EDTA was added to plasma to a final concentration of 2 mmol/l to prevent oxidation. NaCl/KBr/NaBr solutions were made up to the following densities: 1.019 g/ml, 1.063 g/ml, 1.21 g/ml, 1.346 g/ml, and 1.488 g/ml by following the protocol in Table 2.2. Their densities were measured using a densitometer (Densito 30PX, Metler Toledo, Victoria, Australia).
For each centrifuge tube, 44 ml of plasma was used. The density of plasma was first adjusted to 1.019 g/ml by adding 1.75 ml of the NaCl/KBr solution with density of 1.346 g/ml. Subsequently, the sample was overlayed gently with 24.25 ml of 1.019 g/ml density solution to make up a total volume of 70 ml. Plasma was centrifuged in an Optima LE-90K centrifuge with a Type 45 Ti fixed angle rotor (Beckman Coulter, New South Wales, Australia) at 45,000 rpm (234,998 xg), 12oC for 22 h. The top layer (20 ml) corresponding to the VLDL fraction was then recovered and the next 20 ml corresponding to excess density solution was discarded. The density of the remaining
62
CHAPTER 2 – MATERIALS AND GENERAL METHODS mixture was then adjusted to 1.063 g/ml by the addition of 4.66 ml of the NaCl/KBr solution (d = 1.346 g/ml) and subsequently overlayed with 35.34 ml of 1.063 g/ml solution. The sample was then centrifuged (45,000 rpm, 12oC, 22 h). The top 20 ml layer corresponding to LDL fraction was aspirated and the next 20 ml aspirated and discarded. The density of the remaining mixture was adjusted for the HDL fractionation by adding 15.86 ml of 1.488 g/ml NaBr solution. The mixture was then overlayed with 24.14 ml of 1.21 g/ml solution and centrifuged (45,000 rpm, 12oC, 22 h). The HDL was aspirated in the top 20 ml and the remaining density solution was discarded. Immediately upon isolation, VLDL, LDL and HDL were dialysed against phosphate- buffered saline (PBS, 137 mmol/l NaCl, 2.7 mmol/l KCl, 8.1 mmol/l Na2HPO4, 1.5 mmol/l KH2PO4, pH 7.4) containing 5 µmol/l EDTA with three buffer changes (225x sample volume) over 24 h at 4oC. Sucrose was added to aliquots of lipoprotein fractions to give a final concentration of 10% (w/v); this was to preserve the lipoprotein function [334, 335] prior to storage at -80oC.
Table 2.2 Preparation of density solutions for sequential ultracentrifugation.
Target Protocol density (g/ml) 1.488 225.00 g of NaBr (MW 102.89 g/mol) was dissolved in deionised water until the total weight reached 500 g. 1.346 45.90 g of NaCl (MW 58.44 g/mol) and 106.2 g KBr (MW 119.01 g/mol) were mixed and made up to 300 ml with deionised water. 1.21 29.63 g NaBr was mixed well with 100 ml of 0.15 mol/l NaCl.
1.182 20.75 g NaBr was made up to 100 ml with deionised water.
1.063 20.5 ml of density solution (1.346 g/ml) was mixed with 100 ml of 0.15 mol/l NaCl. 1.019 4.28 ml of density solution (1.346 g/ml) was mixed with 100 ml of 0.15 mol/l NaCl.
2.2.2 Lipoprotein oxidation and measurement of oxidation
2.2.2.1 Lipoprotein oxidation
All lipoprotein oxidation conducted in this study was carried out using copper chloride unless stated otherwise in the chapter’s method section; this method was adapted from
63
CHAPTER 2 – MATERIALS AND GENERAL METHODS
Kleinveld et al. [333]. Typically LDL (0.1 mg protein/ml) and HDL (0.3 mg protein/ml) were oxidised with 1 or 5 µmol/l copper chloride in PBS at 37oC between 60 to 300 min. The protein ratio of 3:1 (HDL: LDL) was used to provide approximately equal amounts of HDL and LDL lipids in the assay system. The oxidation was terminated by adding EDTA to a final concentration of 2 or 10 μmol/l (2x the concentration of copper chloride) and by lowering the temperature of the samples to 4oC.
2.2.2.2 Conjugated diene measurement
Conjugated diene was used as a measure of fatty acid oxidation [333]. LDL or oxLDL (0.1 mg protein/ml), and HDL or oxHDL (0.3 mg protein/ml) were measured for the absorbance of conjugated diene at 234 nm on a DU800 spectrophotometer (Beckman Coulter, New South Wales, Australia).
2.2.2.3 Thiobarbituric acid reactive substances (TBARS) assay
Malondialdehyde (MDA) is an end product of lipid peroxidation. Its level is measured in the TBARS assay [336]. Aliquots of LDL/oxLDL (100 μl, 0.1 mg protein/ml) or HDL/oxHDL (100 μl, 0.3 mg protein/ml) were mixed with 200 μl of 10% (w/v) trichloroacetic acid to precipitate the protein and 300 μl of 1% (w/v) of thiobarbituric acid. The samples were incubated in boiling water bath for 30 min and then centrifuged (15,588 xg, 5 min). The supernatant was collected and the absorbance at 540 nm measured. The concentrations of MDA in the samples were determined by comparison to a calibration curve (0 to 100 μmol/l 1,1,3,3-tetramethoxypropane) and expressed relative to the amount of protein in the samples.
2.2.3 Cell culture
Tamm-Horsfall protein-1 (THP1) monocytes (a gift from Prof. Dmitri Sviridov) were cultured with (Roswell Park Memorial Institute) RPMI 1640 medium containing 2 mmol/l L-glutamine (Gibco, Thermo Fisher Scientific, Scoresby, Victoria, Australia) and 10% fetal bovine serum. Human umbilical vascular endothelial cells (HUVEC, passages 2-6, Lonza, Basel, Switzerland) were cultured with Endothelial Cell Growth Medium-2 (EBM2) medium with EGM growth supplements (Lonza, Basel, Switzerland) containing 2% fetal bovine serum. Both cell media contained 100 U/ml
64
CHAPTER 2 – MATERIALS AND GENERAL METHODS antibiotic/antimitotic (Life Technologies, Victoria, Australia). The cultured cells were o maintained at 37 C in 5% CO2.
2.2.4 Bicinchoninic acid (BCA) protein assay
Measurement of total protein concentration was performed using the BCA assay (Pierce Biotechnology BCA Protein Assay kit, Life Technologies, Victoria, Australia) according to the manufacturer’s kit instructions [337]. Briefly, 25 µl of sample (at least at 0.1 mg/ml) and of the BSA standards were aliquoted into 96 well plates followed by the addition of 200 µl of working reagent per well. The samples were incubated at 37oC for 30 min and the absorbance at 562 nm measured. The protein concentrations of samples were determined by comparison to a calibration curve (0 - 1 mg/ml BSA).
2.2.5 Extraction and analysis of lipids
2.2.5.1 Lipid extraction
Lipids were extracted as previously described [297]. Briefly, samples were mixed with 20 volumes of chloroform:methanol (2:1) and internal standards (20 μl, Table 2.2). The lipid standards were either commercially available as deuterated lipids or of very low abundance in biological samples (ie. non-physiological). The mixtures were briefly vortexed, mixed for 10 min (on a rotary mixer), sonicated for 30 min and then allowed to stand at room temperature for 20 min before they were centrifuged at 16,000 xg for 10 min at room temperature. The supernatant was dried either under a stream of nitrogen gas at 40oC or in vacuum at room temperature. Subsequently they were reconstituted in a mixture of water saturated butanol and methanol (1:1 v/v) containing 5 mmol/l ammonium formate.
To ensure unbiased analysis, samples were randomised prior to the lipid extraction. To assess analytical performance both within and between analytical runs, each lipid extraction included a number of quality control (QC) samples: (1) QC1, a “reagent blank” (10 μl milliQ water with no internal standards) every 95 samples; (2) QC2 which is QC1 mixed with 10 μl of internal standard mix, for every 40 samples; (3) QC3, a THP1 monocyte quality control (2 mg protein/ml) for every analytical run; (4) QC4, a plasma QC (10 μl of pooled healthy human plasma sample previously profiled) every 20 samples; and (5) a technical plasma QC (10 μl of lipid extract of pooled QC4s for the
65
CHAPTER 2 – MATERIALS AND GENERAL METHODS entire study) every 20 samples. These QC samples were monitored across each analytical run to assess the variance within each run. Median coefficient of variance of less than 15% across all lipid species indicates a good lipid extraction and analytical run.
Table 2.3 Internal standards used in lipid extraction and mass spectrometry analysis.
Internal standards pmol a 7-ketocholesterol (d7) 100 Cardiolipin 14:0/14:0/14:0 100 Ceramide 17:0 100 Cholesterol (d7) 10,000 Cholesterol ester 18:0 (d6) 1,000 Diacylglycerol 15:0/15:0 200 Dihexosylceramide 16:0 (d3) 50 Dihydroceramide 8:0 50 Glyceryl triheptadecanoate 17:0/17:0/17:0 100 Lysophosphatidylcholine 13:0 100 Lysophosphatidylethanolamine 14:0 100 Monoacylglycerols 17:0 100 Monohexosylceramide 16:0 (d3) 50 Phosphatidylcholine 13:0/13:0 100 Phosphatidylethanolamine 17:0/17:0 100 Phosphatidylglycerol 17:0/17:0 100 Phosphatidylserine 17:0/17:0 100 Sphingomyelin C12:0 200 Trihexosylcermide 17:0 50 aAmount of internal standard per sample.
2.2.5.2 Liquid chromatography electrospray ionisation tandem mass spectrometry
Approximately 375 lipids were quantified routinely using multiple reaction monitoring in positive ion mode on an Agilent 1200 HPLC system coupled to a QTrap 4000 triple quadrupole mass spectrometer (Applied Biosystems, Massachusetts, USA) using methodology similar to that described previously [254]. LC separation was performed on a 2.1 x 100 mm C18 Poroshell column (Agilent Technologies, California, USA) at 300 μl/min. The following gradient conditions were used: 10 % B to 100 % B over 13
66
CHAPTER 2 – MATERIALS AND GENERAL METHODS min, 100 % B over 3 min, and a return to 10 % B over 1 min, followed by 10 % B over 3 min. Solvents A and B consisted of water:tetrahydrofuran:methanol in the ratio of 60:20:20 and 5:75:20 respectively, both containing 10 mmol/l ammonium formate. The column was heated to 50oC and the auto-sampler regulated to 25oC. The conditions for the tandem mass spectrometry of each lipid class are provided in Table 2.4. Different lipid species under the same lipid class were analysed the same way. The LC-ESI- MS/MS method was develop to target lipid classes and subclasses in plasma, cells, and animal tissues. These lipid classes and subclasses are listed in Table 2.3 and include alkylglycerols (batyl alcohol); sterols (cholesteryl ester; and free cholesterol); oxysterols (7-ketocholesterol; and 7β-hydroxycholesterol); diacylglycerophospholipids (phosphatidylcholine; phosphatidylethanolamine; phosphatidylserine; and phosphatidylinositol), 1-alkyl-2-acylglycerophospholipids or alkylphospholipids (alkylphosphatidylcholine; and alkylphosphatidylethanolamine); 1-alkenyl-2- acylglycerophospholipids or alkenylphospholipids (alkenylphosphatidylcholine or PC plasmalogen; and alkenylphosphatidylethanolamine or PE plasmalogen); monoacylglycerophospholipids (lysophosphatidylcholine; and lysophosphatidylethanolamine); sphingolipids (dihydroceramide; ceramide; monohexosylceramide; dihexosylceramide; trihexosylceramide; sphingomyelin; and
GM3 ganglioside), and glycerolipids (diacylglycerol; and triacylglycerol) [254, 297, 338].
2.2.5.3 Lipid analysis
Lipid peak integration was carried out using MultiQuant v.2.1.1 (ABSciex, Massachusetts, USA). The lipid species with peak intensity less than three times the average background level of the QC2 were not analysed. Comparative lipid concentrations were calculated by relating the peak area of each species to the peak area of the corresponding internal standard. This ratio was then multiplied by the amount of internal standard added into the sample. In a number of cases described by Weir et al. previously [297], correction factors for lipid species were applied. The corrections applied for individual lipid classes are summarised in Supplementary Table 2.1.
Relative lipid concentrations were expressed as pmol/ml of plasma or pmol/mg protein of tissues, or normalised to total level of phosphatidylcholine in the sample. Total lipids
67
CHAPTER 2 – MATERIALS AND GENERAL METHODS of each class were obtained by the sum of relative concentration of individual lipid species [297].
2.2.6 Mouse models of atherosclerosis
All aspects of animal care and experimentation in this study were approved by the Alfred Medical Research and Education Precinct Animal Ethic Committee and conformed to guidelines laid down by the National Health and Medical Research Council of Australia (E/1345/2013/B). The animals were housed in standard conditions with unrestricted access to food and water at the Precinct Animal Centre of the Baker IDI Heart and Diabetes Institute. They were maintained on a 12 h light and dark cycle in a pathogen free environment. ApoE-/- mice are models of atherosclerosis on a C57/BL6 background. They were obtained from the Animal Resources Centre, Western Australia. ApoE-/-GPx1-/- mice are models of atherosclerosis with elevated oxidative stress level and are also on a C57/BL6 background. They were a gift from Dr. Judy de Haan at the Baker IDI Heart and Diabetes Institute in Melbourne, Australia. Details of an animal study using these models are described in Chapter 5 of this thesis. A total of 58 mice completed the study. Prior to the cull, animals were anaesthetised by Avertin (2,2,2- tribromoethanol) IP (0.3 ml of 2.5% solution per 20 g mouse; Sigma Chemical Co, USA) following food withdrawal for 3 h. Plasma and tissues including heart, aorta, liver, and adipose tissue were obtained and snap frozen. Details of subsequent tissue and lipid analyses of plasma and heart are described in Chapter 5, whereas analyses of liver and adipose tissue are described in Chapter 6.
2.2.7 Preparation of tissues for lipidomic analysis
Tissues were snap frozen in liquid nitrogen following organ dissection. Approximately 50 to 100 mg of tissue was homogenised in 200 - 300 μl of ice cold phosphate buffered saline, pH 7.4 containing 100 µmol/l butylated hydroxytoluene using a Polytron electric homogeniser for 10 sec and then with a mini probe sonicator for 15 sec at amplitude 23. The homogenates were subsequently stored at -80C.
The protein concentrations of the tissue homogenates were measured using a standard BCA assay as described in section 2.2.4. Lipidomic analysis was performed on the homogenates (20 - 50 µg protein) using liquid chromatography electrospray ionisation tandem mass spectrometry as described in section 2.2.5.2.
68
CHAPTER 2 – MATERIALS AND GENERAL METHODS
2.2.8 Statistical analyses
Statistical analyses were used to compare biochemical, lipidomic and immunohistological data between sample groups and they are fully described in individual result chapters (Chapter 3 to Chapter 6) of this thesis. Generally, parametric data were analysed by Student t-tests while non-parametric data were analysed via Mann Whitney U test. P-values were adjusted for multiple comparisons using the Benjamini Hochberg method [339]. P-value of less than 0.05 were considered statistical significant.
69
CHAPTER 2 – MATERIALS AND GENERAL METHODS
Table 2.4 Conditions for liquid chromatography electrospray ionisation tandem mass spectrometry analysis of lipid species.
b No. of Parent Voltage settings (V) Lipid class Internal standard Fragmentation a species ion DP EP CollE CXP Alkylglycerol (batyl alcohol) 1 Monoacylglycerols 17:0 [M+H]+ PI, 93.1 Da 80 10 15 8 Alkylphosphatidylcholine 19 Lysophosphatidylcholine 13:0 [M+H]+ PI, m/z 184.1 100 10 45 11 Alkylphosphatidylethanolamine 11 Phosphatidylethanolamine 17:0/17:0 [M+H]+ NL, 141 Da 80 10 31 7 Ceramide 6 Ceramide 17:0 [M+H]+ PI, m/z 264.3 50 10 35 12 + Cholesterol 1 Cholesterol (d7) [M+NH4] PI, m/z 369.3 55 10 17 12 + Cholesteryl ester 25 Cholesterol ester 18:0 (d6) [M+NH4] PI, m/z 369.3 30 10 20 12 + Diacylglycerol 17 Diacylglycerol 15:0/15:0 [M+NH4] NL, fatty acid 55 10 30 22 Dihexosylceramide 6 Dihexosylceramide 16:0 (d3) [M+H]+ PI, m/z 264.3 100 10 65 12 Dihydroceramide 6 Dihydroceramide 8:0 [M+H]+ PI, m/z 284.3 90 30 28 10 + GM3 ganglioside 6 Trihexosylcermide 17:0 [M+H] PI, m/z 264.3 155 10 105 16 Lysoalkylphosphatidylcholine 9 Lysophosphatidylcholine 13:0 [M+H]+ PI, m/z 285.2 90 10 42 5 Lysophosphatidylcholine 22 Lysophosphatidylcholine 13:0 [M+H]+ PI, m/z 184.1 100 10 45 11 Lysophosphatidylethanolamine 6 Lysophosphatidylethanolamine 14:0 [M+H]+ NL, 141 Da 80 10 31 7 + Lysophosphatidylinositol 3 Phosphatidylethanolamine 17:0/17:0 [M+NH4] NL, 277 Da 51 10 33 14 Monohexosylceramide 6 Monohexosylceramide 16:0 (d3) [M+H]+ PI, m/z 264.3 77 10 50 12 Oxysterol (7-ketocholesterol) 1 7-ketocholesterol (d7) [M+H]+ PI, m/z 81 60 10 57 8 Oxysterol (7β- 1 7-ketocholesterol (d7) [M+H]+ PI, m/z 159 80 10 35 24 hydroxycholesterol) Phosphatidylcholine 51 Phosphatidylcholine 13:0/13:0 [M+H]+ PI, m/z 184.1 100 10 45 11 Phosphatidylcholine 14 Phosphatidylcholine 13:0/13:0 [M+H]+ PI, m/z 184.1 100 10 45 11 plasmalogen Phosphatidylethanolamine 20 Phosphatidylethanolamine 17:0/17:0 [M+H]+ NL, 141 Da 80 10 31 7 Phosphatidylethanolamine 11 Phosphatidylethanolamine 17:0/17:0 [M+H]+ NL, 141 Da 80 10 31 7 plasmalogen + Phosphatidylinositol 15 Phosphatidylethanolamine 17:0/17:0 [M+NH4] NL, 277 Da 51 10 43 14 Phosphatidylserine 5 Phosphatidylserine 17:0/17:0 [M+H]+ NL, 185 Da 86 10 29 16
70
CHAPTER 2 – MATERIALS AND GENERAL METHODS
b No. of Parent Voltage settings (V) Lipid class Internal standard Fragmentation a species ion DP EP CollE CXP Sphingomyelin 20 Sphingomyelin C12:0 [M+H]+ PI, m/z 184.1 65 10 35 12 Glyceryl triheptadecanoate Triacylglycerol 35 [M+NH ]+ NL, fatty acid 95 10 30 12 17:0/17:0/17:0 4 Trihexosylceramide 5 Trihexosylceramide 17:0 [M+H]+ PI, m/z 264.3 130 10 73 12 a NL - neutral loss, PI - product ion b CollE - collision energy, CXP - collision cell exit potential, DP - declustering potential, EP - exit potential
71
CHAPTER 3 - CHARACTERISATION OF CHANGES IN THE LIPID COMPOSITION OF LIPOPROTEINS DUE TO OXIDATIVE STRESS
72
CHAPTER 3 – CHARACTERISATION OF CHANGES IN THE LIPID COMPOSITION OF LIPOPROTEINS DUE TO OXIDATIVE STRESS
3 CHARACTERISATION OF CHANGES IN THE LIPID COMPOSITION OF LIPOPROTEINS DUE TO OXIDATIVE STRESS
3.1 ABSTRACT
Background: The primary source of atherosclerotic plaque lipids is circulating lipoproteins that cross the arterial wall into the intima layer and accumulate over time. The intima micro-environment is conducive to lipoprotein oxidation. Oxidised lipoproteins subsequently contribute to the progression of atherosclerosis. However, the progressive changes in the lipid composition of lipoproteins during oxidation are not well defined. Therefore, the primary aim of this chapter was to characterise changes in the lipid composition of lipoproteins due to differing levels of oxidative stress. Methods/results: Isolated human LDL and HDL were oxidised with copper chloride for up to 300 min at 37oC and subsequently analysed using LC-MS/MS. Relative to SFA+MUFA PC, oxidation reduced the levels of plasmalogen and PUFA-containing alkylphospholipids in the outer layer of lipoproteins, and cholesteryl ester and PUFA- containing triacylglycerol in the core of lipoproteins; whereas oxidation increased the ratio of sphingomyelin to phosphatidylcholine and the level of lysophosphatidylcholine. SFA- and MUFA- containing lysophosphatidylcholine and PUFA containing lysophosphatidylcholine were produced differently during the oxidation and the regioisomers gave us a clue to the major sources of the lysophosphatidylcholine. Lipids in LDL were generally more susceptible to oxidation as compared to those of HDL particularly cholesteryl ester and triacylglycerol in the core of lipoprotein. Using an untargeted lipidomic approach, we identified 14 major species of oxidised phosphatidylcholine and cholesteryl ester which were increased relative to SFA+MUFA PC in oxidised LDL and HDL. Conclusions: Oxidation resulted in a myriad of changes to the lipid composition of lipoproteins. Lipids in both the outer layer and core of lipoproteins were affected by oxidation within the same amount time, but more so in LDL than in HDL. These findings highlight the different susceptibility of the lipids in LDL and HDL to oxidation.
73
CHAPTER 3 – CHARACTERISATION OF CHANGES IN THE LIPID COMPOSITION OF LIPOPROTEINS DUE TO OXIDATIVE STRESS
3.2 INTRODUCTION
The primary source of atherosclerotic plaque lipids is circulating lipoproteins that cross the arterial wall into the intima layer and accumulate over time. The intima micro- environment is conducive to lipoprotein oxidation. Oxidised lipoproteins subsequently contribute to the progression of atherosclerosis. However, the mechanism of lipoprotein oxidation is poorly defined. Previously, Meikle et al. demonstrated a negative association of circulating plasmalogen and positive association of circulating lysophosphatidylethanolamine with CAD patients [254]. Other studies have also shown that there was a significantly higher level of lysophosphatidylcholine containing saturated fatty acids (SFA) and monounsaturated fatty acids (MUFA) in both oxLDL and human atherosclerotic lesions [197, 200]. As previously mentioned in Chapter 1, lysophospholipids can be derived from either the oxidation of alkenylphospholipids
(plasmalogens) or the action of Lp-PLA2, an independent predictor of CAD [199, 340], on fatty acyl chain at the sn-2 position of the glycerol backbone of particularly oxidised diacylphospholipids. Plasmalogens have been proposed to have anti-oxidant properties [264, 278]. Therefore, decreased levels of plasmalogens, combined with the increased levels of lysophospholipids suggest an elevated level of oxidative stress in CAD patients. However, these lipids provided only a restricted view on the metabolic changes associated with advanced disease. The progressive changes in the lipid composition of lipoproteins during oxidation have not been fully investigated. Therefore, the primary aim of this chapter was to characterise changes in the lipid composition of lipoproteins due to differing levels of oxidative stress. The first part of this aim involved the identification of major oxidised lipids including oxPC and oxCE in lipoproteins and the measurement of their levels during oxidation. Using an untargeted lipidomic approach, we identified major oxidised lipid species from oxLDL. We combined this with an established in-house targeted lipidomic approach to quantify lipids of interests in LDL and HDL fractions and assess if and when plasmalogen, together with other lipid classes were affected by oxidation. These lipidomic approaches enabled us to gain a more complete picture of native lipids in the lipoproteins that were depleted and the oxidised lipids produced during oxidation. We hypothesised that lipids on the outer layer of lipoproteins such as plasmalogen and alkylphospholipids would be affected early in the oxidation process whereas lipids in the core of lipoproteins such as cholesteryl ester and triacylglycerols would be affected later in the process.
74
CHAPTER 3 – CHARACTERISATION OF CHANGES IN THE LIPID COMPOSITION OF LIPOPROTEINS DUE TO OXIDATIVE STRESS
3.3 METHODS
3.3.1 Lipoprotein oxidation
Isolated human LDL (0.1 mg protein/ml) and HDL (0.3 mg protein/ml) were oxidised with 5 μmol/l copper chloride for up to 300 min at 37oC. The oxidation was terminated by the addition of 10 μmol/l EDTA and cooling the samples to 4oC. The lipoprotein oxidation was performed in three independent experiments.
The lipoprotein oxidation was monitored via the production of conjugated diene and malondialdehyde. These were measured via the absorbance of conjugated diene at 234 nm in a spectrophotometer and the TBARS assay, performed as described in Chapter 2 (section 2.2.2.2 and 2.2.2.3). The conjugated diene measurement was done in three independent oxidation experiments, whereas the TBARS assay was carried out once with n = 3.
3.3.2 Lipid analysis
Following oxidation, aliquots of the lipoprotein samples (100 μl at approximately 0.1 mg protein/ml LDL and 0.3 mg protein/ml HDL) were lyophilised and subsequently reconstituted in 10 μl of deionised water. The lipids were then extracted as described in Chapter 2 (section 2.2.5.1). The lipids were analysed by LC-ESI-MS/MS using multiple reaction monitoring in a positive ion mode as described in Chapter 2 (section 2.2.5.2 and 2.2.5.3).
Comparative lipid concentrations were calculated by relating the peak area of each species to the peak area of the corresponding internal standard (Table 2.4, Chapter 2). Peak integration was carried out using MultiQuant software v.2.1.1. Total lipids of each class were obtained by the sum of relative concentration of individual lipid species. Levels of oxPC, oxCE, alkenylphospholipids, alkylphospholipids, lysophosphatidylcholine, cholesteryl ester and triacylglycerol were normalised to the level of phosphatidylcholine containing SFA and MUFA (SFA+MUFA PC). These species were found to be resistant to oxidation and so represent a stable factor for normalisation. Other lipids such as sphingomyelin were expressed relative to levels of total phosphatidylcholine (PC) to reflect the relative contribution these lipid classes made to the surface lipids of the lipoprotein particles.
75
CHAPTER 3 – CHARACTERISATION OF CHANGES IN THE LIPID COMPOSITION OF LIPOPROTEINS DUE TO OXIDATIVE STRESS
Lysophosphatidylcholine species were analysed by their fragmentation ions at m/z 184. The regioisomers (sn-1 LPC and sn-2 LPC where the fatty acid is at the sn-1 position and sn-2 position of glycerol backbone, respectively) were distinguished by their retention time on the chromatography; sn-2 isomers were eluted slightly earlier than sn- 1 isomers [341, 342]. The LC-ESI-MS/MS method as described above allows the separation and measurement of both the sn-1 and sn-2 isomers of lysophosphatidylcholine.
3.3.3 Identification of oxPC and oxCE by untargeted lipidomics
To identify the major oxidised lipids in oxLDL, a time-course oxidation of LDL (1 mg protein/ml) was performed with 8 µmol/l copper sulphate for 1, 2, 4, 8, and 24 h at 37oC. The lipids were subsequently extracted as described in Chapter 2 (section 2.2.5.1) and analysed using an Agilent 1200 liquid chromatography system with an Agilent 6520 quadrupole-time of flight mass spectrometer (LC-QTOF MS). Lipid extracts (5 µl) were injected and run on a 2.1 x 100 mm C18 column (Zorbax-Eclipse, Agilent, USA) at 250 μl/min. The following gradient conditions were used: 0 % B to 40 % B over 4 min, 40 % B to 63.6 % B over 13 min, 63.6 % B to 100 % B over 12 min, 100% B over 3 min, and a return to 0 % B over 1 min, followed by 0 % B over 7 min. Solvents A and B consisted of water:terahydrofuran:methanol in the ratio of 60:20:20 and 5:75:20 respectively, both containing 10 mmol/l ammonium formate. The mass spectrometer was operated in positive ion mode from m/z 100 to 1500 at a scan rate of 1.36 scan sec-1. Two reference ions at m/z 121.0509 and 922.009 were used throughout the analysis. The source temperature was set to 325oC, with 7L min-1 drying gas and a nebuliser pressure of 40 psig. The fragmentor, skimmer, and octopole voltages were set to 120 V, 30 V and 750 V, respectively. The data files are subsequently converted to .mzdata using Mass Hunter prior to analysis using MZmine v.2.10.
3.4 RESULTS
3.4.1 Conjugated diene and TBARS production in oxidised LDL and HDL
Oxidation of LDL showed three distinct phases of conjugated diene production: the lag phase, the exponential phase which led to maximum diene production, and the plateau phase (Figure 3.1A). Maximum diene production in LDL was reached after 84 min of oxidation. The TBARS formation corresponded well to the diene production in the first
76
CHAPTER 3 – CHARACTERISATION OF CHANGES IN THE LIPID COMPOSITION OF LIPOPROTEINS DUE TO OXIDATIVE STRESS
50 min of oxidation, after which the TBARS formation slowed compared to the diene production. Maximum TBARS was formed after 120 min of oxidation (Figure 3.1A).
HDL showed a general increase in conjugated diene production throughout the oxidation and showed less distinct phases compared to LDL. Maximum diene production was obtained at 219 min of oxidation. The TBARS formation of HDL also showed a general increasing trend but lagged behind the diene profile (Figure 3.1B). Maximum TBARS was formed at 300 min of oxidation (Figure 3.1B).
3.4.2 Changes in lipid class composition associated with the outer layer of lipoproteins following oxidation
Relative to SFA+MUFA PC, a decrease in the total level of alkylphosphatidylcholine in both LDL and HDL occurred after approximately 45 min of oxidation (Figure 3.2A and 3.3A). The steepest drop in the lipid level was between 45 to 90 min in LDL (Figure 3.2A) and between 45 to 120 min in HDL (Figure 3.3A).
Relative to SFA+MUFA PC, we observed no effect on the total alkenylphosphatidylcholine or PC plasmalogen level for the first 30 min of oxidation in both LDL and HDL, and then a rapid decrease from 30 to 90 min in LDL (Figure 3.2B) and a more gradual decrease from 30 to 300 min in HDL (Figure 3.3B). Following 300 min of oxidation, the levels of PC plasmalogen relative to SFA+MUFA PC were decreased by 83% (P = 0.11) and 77% (P = 0.07) in LDL and HDL, respectively (Figure 3.2B and 3.3B).
At 300 min of oxidation, the levels of alkylphosphatidylcholine relative to SFA+MUFA PC in LDL and HDL were 79% and 80% lower (P<0.05 for both), respectively compared to their levels at 0 min (Figure 3.2A and 3.3A). Whereas, complete depletion of alkylphosphatidylethanolamine relative to SFA+MUFA PC in LDL and HDL was observed after 90 min and 300 min of oxidation, respectively (Figure 3.2C and 3.3C).
The levels of alkenylphosphatidylethanolamine or PE plasmalogen relative to SFA+MUFA PC in LDL and HDL showed a similar but more complete decrease; PE plasmalogen in LDL was completely depleted after 90 min of oxidation (Figure 3.2D), whereas that of HDL showed a slower decline but was depleted after 240 min oxidation (Figure 3.3D).
77
CHAPTER 3 – CHARACTERISATION OF CHANGES IN THE LIPID COMPOSITION OF LIPOPROTEINS DUE TO OXIDATIVE STRESS
After 30 min and 60 min of oxidation in LDL and HDL, we observed a steady increase in the total levels of lysophosphatidylcholine relative to SFA+MUFA PC (Figure 3.4A and 3.5A). Relative to SFA+MUFA PC, the level of lysophosphatidylcholine was increased approximately 15-fold (P<0.05) and 12-fold (P = 0.07) in LDL and HDL, respectively at 300 min of oxidation compared to those at 0 min. We then analysed the levels of regioisomers of lysophosphatidylcholine species, sn-1 and sn-2 LPC to determine which isomer contributed predominantly to the total lysophosphatidylcholine level. We observed that relative to SFA+MUFA PC, sn-1 LPC was more abundant than sn-2 LPC in both non-oxidised (control) LDL and HDL (Figure 3.4B and 3.5B). Oxidation resulted in an increase of both sn-1 and sn-2 LPC levels relative to SFA+MUFA PC in LDL and HDL (Figure 3.4C and 3.5C).
Relative to total PC, the level of sphingomyelin was increased by 4.0-fold (P<0.05) in LDL and 3.0-fold (P<0.05) in HDL (Figure 3.6A-B). Compared to the non-oxidised samples, we did not observe any difference in the levels of cholesterol relative to SFA+MUFA PC in LDL and HDL throughout the oxidation (Figure 3.6C-D).
78
CHAPTER 3 – CHARACTERISATION OF CHANGES IN THE LIPID COMPOSITION OF LIPOPROTEINS DUE TO OXIDATIVE STRESS
A B
Figure 3.1 Conjugated diene and TBARS formation in oxidised low density and high density lipoproteins. Oxidation of (A) LDL; and (B) HDL with copper chloride was monitored via the production of conjugated diene and TBARS as described in section 3.3.1. Conjugated diene production in oxidised lipoproteins is shown in blue diamonds whereas that of non-oxidised lipoproteins is shown in green. Conjugated diene is expressed as nmol/mg protein of LDL or HDL. The conjugated diene profile in LDL is divided into three distinct phases: (a) the lag phase; (b) the exponential phase; and (c) the plateau phase. Less distinct phases were observed in HDL. TBARS formation in oxidised lipoproteins is shown in red squares; data represent mean ± SD for n = 3, and is expressed as nmol of malondialdehyde (MDA) per mg protein of LDL or HDL.
79
CHAPTER 3 – CHARACTERISATION OF CHANGES IN THE LIPID COMPOSITION OF LIPOPROTEINS DUE TO OXIDATIVE STRESS
A B
C D
Figure 3.2 The effect of oxidation on alkyl- and alkenylphospholipids associated with outer layer of low density lipoproteins. Changes in the levels of (A) alkylphosphatidylcholine; (B) alkenylphosphatidylcholine or PC plasmalogen; (C) alkylphosphatidylethanolamine; and (D) alkenylphosphatidylethanolamine or PE plasmalogen in LDL from 0 min to 300 min of oxidation with copper chloride. Non-oxidised lipoproteins are shown as blue diamonds and oxidised lipoproteins are shown as red squares. Data represent mean ± SD, expressed as nmol/μmol of PC containing saturated- and monounsaturated fatty acids, n = 2 per sample except for non-oxidised LDL (n = 1). Data were analysed using Student t-test, * indicates P<0.05 compared to the level of lipid at 0 min.
80
CHAPTER 3 – CHARACTERISATION OF CHANGES IN THE LIPID COMPOSITION OF LIPOPROTEINS DUE TO OXIDATIVE STRESS
A B
C D
Figure 3.3 The effect of oxidation on alkyl- and alkenylphospholipids associated with outer layer of high density lipoproteins. Changes in the levels of (A) alkylphosphatidylcholine; (B) alkenylphosphatidylcholine or PC plasmalogen; (C) alkylphosphatidylethanolamine; and (D) alkenylphosphatidylethanolamine or PE plasmalogen in HDL from 0 min to 300 min of oxidation with copper chloride. Non-oxidised lipoproteins are shown as blue diamonds and oxidised lipoproteins are shown as red squares. Data represent mean ± SD, expressed as nmol/μmol of PC containing saturated- and monounsaturated fatty acids, n = 2 per sample. Data were analysed using Student t-test, * indicates P<0.05 compared to the level of lipid at 0 min.
81
CHAPTER 3 – CHARACTERISATION OF CHANGES IN THE LIPID COMPOSITION OF LIPOPROTEINS DUE TO OXIDATIVE STRESS
A
B
C
Figure 3.4 The effect of oxidation on lysophosphatidylcholine in low density lipoproteins. Changes in the levels of (A) lysophosphatidylcholine in unoxidised (blue diamonds) and oxidised (red squares) LDL with time; (B) sn-1 LPC and sn-2 LPC in non-oxidised LDL; and (C) sn-1 LPC and sn-2 LPC in oxidised LDL. Data is expressed as mean ± SD, relative to nmol/μmol PC containing saturated- and monounsaturated fatty acids, n = 2 per sample, except for non-oxidised LDL (n = 1). Level of lysophosphatidylcholine was analysed for each time point using Student t-test, * indicates P<0.05 compared to the level of lipid at 0 min.
82
CHAPTER 3 – CHARACTERISATION OF CHANGES IN THE LIPID COMPOSITION OF LIPOPROTEINS DUE TO OXIDATIVE STRESS
A
B
C
Figure 3.5 The effect of oxidation on lysophosphatidylcholine in high density lipoproteins. Changes in the levels of (A) lysophosphatidylcholine in unoxidised (blue diamonds) and oxidised (red squares) HDL with time; (B) sn-1 LPC and sn-2 LPC in non-oxidised HDL; and (C) sn-1 LPC and sn-2 LPC in oxidised HDL. Data is expressed as mean ± SD, relative to nmol/μmol PC containing saturated- and monounsaturated fatty acids, n = 2. Level of lysophosphatidylcholine for each time point was compared to the level of lipid at 0 min and analysed using Student t-test.
83
CHAPTER 3 – CHARACTERISATION OF CHANGES IN THE LIPID COMPOSITION OF LIPOPROTEINS DUE TO OXIDATIVE STRESS
A B
C D
Figure 3.6 The effect of oxidation on sphingomyelin and free cholesterol associated with outer layer of lipoproteins. Changes in the levels of (A-B) sphingomyelin; and (C-D) free cholesterol in LDL (left panels) and HDL (right panels) from 0 min to 300 min of oxidation with copper chloride. Non-oxidised lipoproteins are shown as blue diamonds and oxidised lipoproteins as red squares. Data represent mean ± SD, expressed as nmol/μmol of PC containing saturated- and monounsaturated fatty acids for lysophosphatidylcholine and free cholesterol, and of total PC for sphingomyelin , n = 2 per sample except for non-oxidised LDL (n = 1). Data were analysed using Student t-test, * indicates P<0.05 compared to the level of lipid at 0 min.
3.4.3 Changes in lipid species associated with the outer layer of lipoproteins following oxidation.
Plasmalogens containing MUFA and PUFA including PC(P-34:1), PC(P-38:5), and PE(P-16:0/22:5) are some of the most abundant species in LDL and HDL. Relative to SFA+MUFA PC, the levels of these plasmalogens were generally decreased throughout the oxidation (Figure 3.7A-C and Figure 3.8A-C). Other molecular species of plasmalogens showed similar trends (Supplementary Table 3.1 and 3.2). PC or PE plasmalogens containing PUFA were depleted faster than that containing SFA+MUFA; PC(P-34:1) levels started to drop steeply after 30 min of oxidation in LDL (Figure
84
CHAPTER 3 – CHARACTERISATION OF CHANGES IN THE LIPID COMPOSITION OF LIPOPROTEINS DUE TO OXIDATIVE STRESS
3.7A) and after 90 min in HDL (Figure 3.8A). The levels reached a plateau in LDL after 240 min of oxidation whereas in HDL, it continued to decrease until the end of oxidation (Figure 3.7A and 3.8A). On the other hand, the level of PC(P-38:5) was decreased after 30 min of oxidation in both LDL and HDL and it was completely depleted by the end of oxidation (Figure 3.7B and 3.8B). PE(P-16:0/22:5) was depleted after 75 min of oxidation in LDL, but it was depleted at a slower rate in HDL (Figure 3.7C and 3.8C).
Relative to total SFA+MUFA PC, we observed no difference in the level of alkylphospholipid species containing MUFA such as PC(O-34:1) compared to the non- oxidised lipoproteins (Figure 3.7D and 3.8D). Similar to plasmalogens, alkylphospholipids containing PUFA such as PC(O-38:5) and PE(O-18:2/20:3) were steadily reduced after 15-30 min of oxidation until they were completely depleted by the end of the oxidation (Figure 3.7E-F and 3.8E-F). Similar trends were observed with other molecular species of alkylphospholipids containing MUFA and PUFA (Supplementary Table 3.3 and 3.4).
Lysophosphatidylcholine species containing SFA and PUFA have differential changes in their levels during oxidation. Relative to SFA+MUFA PC, the levels of lysophosphatidylcholine containing SFA such as LPC 18:0 were increased up to 11 fold (P<0.05) and 9.5 fold (P = 0.07) in LDL and HDL, respectively. (Figure 3.9A and 3.10A and Supplementary Table 3.5 and 3.6) at 300 min, compared to at 0 min. In contrast, the level of lysophosphatidylcholine containing PUFA such as LPC 18:2 was steadily increased up to 1.3 fold at 120 min of LDL oxidation, after which it was decreased steadily until the end of oxidation (Figure 3.9B). In HDL, LPC 18:2 was increased slightly from 0 min to 100 min of oxidation, after which it reached a plateau where its level remained the same (Figure 3.10B). These differential trends with the oxidation of LPC containing SFA or MUFA and PUFA were also observed in other molecular species (Supplementary Table 3.5 and 3.6).
Analyses of the regioisomers of lysophosphatidylcholine species revealed that lysophosphatidylcholine containing SFA, including LPC 18:0, was composed predominantly of sn-1 LPC isomers in LDL and HDL (Figure 3.9C and 3.10C). In contrast, lysophosphatidylcholine with PUFA such as LPC 20:4 consisted primarily of sn-2 LPC in both lipoproteins (Figure 3.9D and 3.10D); the level of sn-2 LPC 20:4
85
CHAPTER 3 – CHARACTERISATION OF CHANGES IN THE LIPID COMPOSITION OF LIPOPROTEINS DUE TO OXIDATIVE STRESS particularly in LDL generally increased up to 60 min of oxidation, after which it was steadily decreased (Figure 3.9D).
Relative to total PC, the level of SM containing SFA such as SM 34:0 were increased by 4.6-fold (P<0.05) in LDL and 3.6-fold (P<0.05) in HDL at the end of 300 min of oxidation (Figure 3.11A-B and Supplementary Table 3.7 and 3.8). SM containing PUFA such as SM 36:3 was increased by 2.4-fold (P<0.001) in LDL and by 2.0-fold (P = 0.118) in HDL (Figure 3.11C-D, and Supplementary Table 3.7 and 3.8).
86
CHAPTER 3 – CHARACTERISATION OF CHANGES IN THE LIPID COMPOSITION OF LIPOPROTEINS DUE TO OXIDATIVE STRESS
A B
*
C D
E F
Figure 3.7 The effect of oxidation on alkyl- and alkenylphospholipids containing monounsaturated- and polyunsaturated fatty acids in the outer layer of low density lipoproteins. Changes in the levels of (A) PC(P-34:1); (B) PC(P-38:5); (C) PE(P-16:0/22:5); (D) PC(O-34:1); (E) PC(O-38:5); and (F) PE(O-18:2/20:3) in LDL from 0 min to 300 min of oxidation with copper chloride. Non-oxidised lipoproteins are shown as blue diamonds and oxidised lipoproteins as red squares. Data represent mean ± SD, expressed as nmol/μmol of total saturated- and monounsaturated PC, n = 2 per sample except for non-oxidised LDL (n = 1). Data were analysed using Student t-test, * indicates P<0.05 compared to the level of lipid at 0 min.
87 CHAPTER 3 – CHARACTERISATION OF CHANGES IN THE LIPID COMPOSITION OF LIPOPROTEINS DUE TO OXIDATIVE STRESS
A B
C D
E F
Figure 3.8 The effect of oxidation on alkyl- and alkenylphospholipids containing monounsaturated- and polyunsaturated fatty acids in the outer layer of high density lipoproteins. Changes in the levels of (A) PC(P-34:1); (B) PC(P-38:5); (C) PE(P-16:0/22:5); (D) PC(O-34:1); (E) PC(O-38:5); and (F) PE(O-18:2/20:3) in HDL from 0 min to 300 min of oxidation with copper chloride. Non-oxidised lipoproteins are shown as blue diamonds and oxidised lipoproteins as red squares. Data represent mean ± SD, expressed as nmol/μmol of total saturated- and monounsaturated PC, n = 2 per sample except for non-oxidised LDL (n = 1). Data were analysed using Student t-test, * indicates P<0.05 compared to the level of lipid at 0 min.
88 CHAPTER 3 – CHARACTERISATION OF CHANGES IN THE LIPID COMPOSITION OF LIPOPROTEINS DUE TO OXIDATIVE STRESS
A B
C
D
Figure 3.9 The effect of oxidation on lysophosphatidylcholine with saturated- and polyunsaturated fatty acids in low density lipoprotein. Changes in the levels of (A) LPC 18:0; (B) LPC 18:2; (C) regioisomers of LPC 18:0; (D) regioisomers of LPC 20:4 from 0 to 300 min of oxidation in LDL. In (A) and (B), non-oxidised lipoproteins are shown as blue diamonds and oxidised lipoproteins as red squares. All data is expressed as mean ± SD, relative to nmol/μmol PC containing saturated- and monounsaturated fatty acids, n = 2, except for non-oxidised LDL (n = 1). Levels of LPC 18:0 and LPC 18:2 were analysed for each time point using Student t-test, * indicates P<0.05, ** indicates P<0.01, compared to the level of lipid at 0 min.
89 CHAPTER 3 – CHARACTERISATION OF CHANGES IN THE LIPID COMPOSITION OF LIPOPROTEINS DUE TO OXIDATIVE STRESS
A B
C
D
Figure 3.10 The effect of oxidation on lysophosphatidylcholine with saturated- and polyunsaturated fatty acids in high density lipoprotein. Changes in the levels of (A) LPC 18:0; (B) LPC 18:2; (C) regioisomers of LPC 18:0; (D) regioisomers of LPC 20:4 from 0 to 300 min of oxidation in HDL. In (A) and (B), non-oxidised lipoproteins are shown as blue diamonds and oxidised lipoproteins as red squares. All data is expressed as mean ± SD, relative to nmol/μmol PC containing saturated- and monounsaturated fatty acids, n = 2. Levels of LPC 18:0 and LPC 18:2 for each time point were compared to the level of lipid at 0 min and analysed using Student t-test.
90 CHAPTER 3 – CHARACTERISATION OF CHANGES IN THE LIPID COMPOSITION OF LIPOPROTEINS DUE TO OXIDATIVE STRESS
A B
C D
Figure 3.11 The effect of oxidation on sphingomyelin containing monounsaturated- and polyunsaturated fatty acids in the outer layer of lipoproteins. Changes in the levels of (A-B) SM 34:0; (C-D) SM 36:3 in LDL (left panels) and HDL (righ panels) from 0 min to 300 min of oxidation with copper chloride. Non-oxidised lipoproteins are shown as blue diamonds and oxidised lipoproteins as red squares. Data represent mean ± SD, expressed as nmol/μmol of total PC, n = 2 per sample except for non-oxidised LDL (n = 1). Data were analysed using Student t-test, * indicates P<0.05, ** indicates P<0.01, *** indicates P<0.001 compared to the level of lipid at 0 min.
3.4.4 Changes in the composition of lipid classes associated with the core of lipoproteins
Relative to SFA+MUFA PC, we observed a general decrease in total levels of cholesteryl ester in LDL and HDL throughout the oxidation (Figure 3.12A-B). However, the depletion in cholesteryl ester in LDL was steeper after approximately 60 min of oxidation (Figure 3.12A) compared to that in HDL (Figure 3.12B). The level of cholesteryl ester relative to SFA+MUFA PC was decreased by 71% in LDL (P = 0.06) and 42% in HDL (P<0.05) at 300 min compared to 0 min of oxidation.
91 CHAPTER 3 – CHARACTERISATION OF CHANGES IN THE LIPID COMPOSITION OF LIPOPROTEINS DUE TO OXIDATIVE STRESS
Relative to SFA+MUFA PC, the level of triacylglycerol in LDL was steadily decreased throughout the oxidation and it was 53% (P<0.05) lower at 300 min of oxidation compared to that at 0 min (Figure 3.12C). The level of triacylglycerol relative to SFA+MUFA PC between oxidised and non-oxidised HDL was not significantly different throughout the oxidation (Figure 3.12D). The level of triacylglycerol relative to SFA+MUFA PC in oxHDL was decreased by 30% after 300 min of oxidation although this was not statistically significant (P = 0.05).
3.4.5 Analyses of molecular species of lipids associated with the core of lipoproteins
Relative to SFA+MUFA PC, we observed no difference in the level of triacylglycerol containing MUFA such as TG 16:0 16:0 18:1 in LDL and HDL throughout the oxidation (Figure 3.13A-B). In contrast, the level of PUFA-containing triacylglycerol such as TG 16:0 18:2 18:2 was reduced by 95% (P = 0.12) in LDL and 83% (P<0.05) in HDL (Figure 3.13C-D) relative to SFA+MUFA PC in respective lipoproteins after 300 min of oxidation. Similar trends with other molecular species of SFA- or MUFA and PUFA containing triacylglycerols were observed (Supplementary Table 3.9 and 3.10).
92 CHAPTER 3 – CHARACTERISATION OF CHANGES IN THE LIPID COMPOSITION OF LIPOPROTEINS DUE TO OXIDATIVE STRESS
A B
C D
Figure 3.12 The effect of oxidation on lipids associated with the core of the lipoproteins. Changes in the total levels of (A-B) cholesteryl esters; and (C-D) triacylglycerols in LDL (left panels) and HDL (right panels) from 0 min to 300 min of oxidation with copper chloride. Non- oxidised lipoproteins are shown as blue diamonds and oxidised lipoproteins as red squares. Data represent mean ± SD, expressed as nmol/μmol of PC containing saturated- and monounsaturated fatty acids, n = 2 per sample except for non-oxidised LDL (n = 1). Data were analysed using Student t-test, * indicates P<0.05 compared to the level of lipid at 0 min.
93 CHAPTER 3 – CHARACTERISATION OF CHANGES IN THE LIPID COMPOSITION OF LIPOPROTEINS DUE TO OXIDATIVE STRESS
A B
C D
Figure 3.13 The effect of oxidation on triacylglycerol species containing monounsaturated and polyunsaturated fatty acids. The relative level of (A and B) TG 16:0 16:0 18:1 ;and (C and D) TG 16:0 18:2 18:2 in LDL (left panels) and HDL (right panels) following oxidation with copper chloride for up to 300 min. Non-oxidised lipoproteins are shown as blue diamonds and oxidised lipoproteins as red squares. Data represent mean ± SD, expressed as nmol/μmol of PC containing saturated- and monounsaturated fatty acids, n = 2 per sample except for non-oxidised LDL (n = 1). Data were analysed using Student t-test, * indicates P<0.05 compared to the level of lipid at 0 min.
3.4.6 Identification of major oxPC and oxCE species in oxidised LDL
This section of the thesis has been published - A.A Rasmiena, C.K Barlow, T.W Ng, P.J Meikle. High density lipoprotein efficiently accepts surface but not internal oxidised lipids from oxidised low density lipoprotein. Biochimica et Biophysica Acta - Molecular and Cell Biology of Lipids, 2016. 1861(2): 69 - 77. (see Appendices page 266).
Upon oxidation, multiple new features were evident in the lipidomic analysis of the LDL (Figure 3.14). While complete characterisation of these new species was beyond the scope of the current work, we observed several features which correspond to the addition of one or two oxygen atoms to major species of phosphatidylcholine (PC (34:3), PC (34:2), PC (36:3), PC (36:2)), and cholesteryl ester (CE (16:1), CE (16:0), CE (18:3), CE (18:2), CE (18:1)) present in LDL (Figure 3.14, Table 3.1). The absence
94 CHAPTER 3 – CHARACTERISATION OF CHANGES IN THE LIPID COMPOSITION OF LIPOPROTEINS DUE TO OXIDATIVE STRESS of significant oxidation products arising from phosphatidylcholine containing only SFA and MUFA suggests their resistance to oxidation and justifies our normalisation of the levels of oxidised lipids to these lipids.
The experimentally determined mass of these features was within 10 mDa of the mass of the proposed oxidised lipid, and they were consistent with the previously reported masses of the oxidised lipid species (Table 3.1). The retention times of these species were earlier than their corresponding non-oxidised counterparts consistent with an increase in polarity upon oxidation. Product ion analysis of these newly identified species showed major product ions of m/z 184.1 and m/z 369.4 corresponding to the phosphocholine head group and cholesterol respectively.
Extracted ion chromatograms of these features (Figure 3.15, Supplementary Figure 3.1 and 3.2) demonstrate that the signal intensity increased with oxidation time. In many instances, oxidation led to multiple chromatographic features consistent with multiple isomeric products presumably differing in the location of the oxygen(s). We have not further characterised these isomeric species but represent them with an O or O2 in parenthesis following the sum composition of the fatty acid.
Following the initial identification of oxPC and oxCE species, we undertook to include these species in our previously established triple quadrupole based targeted lipidomics approach. Multiple reaction monitoring transitions for the oxidised lipids were established using product ions of 184.1 and 369.4 for the oxPC and oxCE, respectively. Retention times were then established using comparison of native and oxLDL samples (Table 3.1).
3.4.7 Oxidised lipids in LDL and HDL
Relative to SFA+MUFA PC, total levels of oxPC and oxCE were increased in LDL and HDL throughout the oxidation (Figure 3.16). In the LDL we observed an initial lag phase (0 - 30 min) followed by a rapid increase (30 - 90 min) and reaching a plateau at 240 min (Figure 3.16A and 3.16C). The HDL showed a more linear response over the entire 300 min of oxidation (Figure 3.16B and 3.16D). The newly identified oxidised lipids in this study comprising of seven species of oxPC and seven species of oxCE showed similar increases (Table 3.2 - 3.5).
95 CHAPTER 3 – CHARACTERISATION OF CHANGES IN THE LIPID COMPOSITION OF LIPOPROTEINS DUE TO OXIDATIVE STRESS
Figure 3.14 Untargeted lipidomic analysis of native and oxidised low density lipoprotein. LDL was oxidised for 24 h with copper sulphate as described in the methods (section 3.3.3). Lipids were extracted and untargeted lipidomic analysis was performed. Panel A shows the region of the analysis corresponding to oxPC while Panel B shows the region corresponding to oxCE.
96 CHAPTER 3 – CHARACTERISATION OF CHANGES IN THE LIPID COMPOSITION OF LIPOPROTEINS DUE TO OXIDATIVE STRESS
Table 3.1 Comparison of the observed and exact masses for the oxidised lipid species identified in the untargeted LC-MS analysis.
Oxidised Parent ion Observed Exact Error Retention Previous species a mass mass time reports (mDa) (min) b
PC (34:3(O)) [M+H]+ 772.551 772.549 2 9.8 [200]
PC (34:2(O)) [M+H]+ 774.561 774.564 -3 9.0 [200]
+ PC (34:3(O2)) [M+H] 788.537 788.544 -7 8.8 [200]
+ PC (34:2(O2)) [M+H] 790.551 790.559 -8 8.8 [200]
PC (36:3(O)) [M+H]+ 800.577 800.58 -3 10.2 [200]
PC (36:2(O)) [M+H]+ 802.590 802.596 -6 10.7 [200]
+ PC (36:3(O2)) [M+H] 816.570 816.575 -5 8.9 [200]
+ CE (16:1(O)) [M+NH4] 656.598 656.598 0 13.0 -
+ CE (16:0(O)) [M+NH4] 658.611 658.613 -2 13.1 -
+ CE (18:3(O)) [M+NH4] 680.598 680.598 0 12.8 [324]
+ CE (18:2(O)) [M+NH4] 682.605 682.613 -8 13.6 [324]
+ CE (18:1(O)) [M+NH4] 684.628 684.629 -1 13.6 -
+ CE (18:3(O2)) [M+NH4] 696.591 696.593 -2 12.9 [324]
+ CE (18:2(O2)) [M+NH4] 698.605 698.608 -3 12.8 [324]
a PC - phosphatidylcholine ; CE - cholesteryl ester b Retention time (min) of oxidised lipids was based on the comparison with native LDL. The retention time were obtained from LC run on a 2.1 x 100 mm C18 Poroshell column (Agilent, USA) at 300 μl/min. The LC condition was as described in General Methods (Chapter 2).
97 CHAPTER 3 – CHARACTERISATION OF CHANGES IN THE LIPID COMPOSITION OF LIPOPROTEINS DUE TO OXIDATIVE STRESS
Figure 3.15 Extracted ion chromatogram (m/z = 772.5440 - 772.5540) of the oxidised lipid corresponding to PC(34:3(O)) in low density lipoprotein. LDL was oxidised for 0, 1, 2, 4, 8, and 24 h with copper sulphate as described in the methods (section 3.3.3). Lipids were extracted and untargeted lipidomic analysis was performed.
A B
C D
Figure 3.16 Production of oxidised phosphatidylcholine and cholesteryl ester in oxidised low density- and high density lipoproteins. Total level of (A-B) oxidised phosphatidylcholine; and (C- D) oxidised cholesteryl esters in LDL (left panels) and HDL (right panels) from 0 to 300 min of oxidation with copper chloride (section 3.3.1). Non-oxidised lipoproteins are shown as blue diamonds and oxidised lipoproteins as red squares. Data represent mean ± SD, expressed as nmol/μmol of PC containing saturated- and monounsaturated fatty acids, n = 2 per sample except for non-oxidised LDL (n = 1). Data were analysed using Student t-test, * indicates P<0.05 compared to the level of lipid at 0 min.
98 CHAPTER 3 – CHARACTERISATION OF CHANGES IN THE LIPID COMPOSITION OF LIPOPROTEINS DUE TO OXIDATIVE STRESS
Table 3.2 The effect of oxidation on the level of oxidised phosphatidylcholine in low density lipoprotein.
Time a (min) Sample PC 34:3 (O) PC 34:2 (O) PC 34:3 (O2) PC 34:2 (O2) PC 36:3 (O) PC 36:2 (O) PC 36:3 (O2) Total oxPC LDL 0.1 3.6 0.3 0.0 1.1 3.2 0.0 8.3 0 oxLDL 0.4 ± 0.5 6.6 ± 4.3 0.4 ± 0.2 0.2 ± 0.2 6.2 ± 6.5 7.9 ± 5.4 0.0 ± 0.0 19 ± 15 LDL 0.2 4.0 0.3 0.0 0.9 1.8 0.0 7.1 15 oxLDL 2.7 ± 2.6 11 ± 9.0 2.8 ± 1.3 7.1 ± 10 9.6 ± 8.6 20 ± 11 2.0 ± 2.8 46 ± 47 LDL 0.1 3.5 0.2 0.0 0.9 2.2 0.0 7.0 30 oxLDL 5.1 ± 3.7 17 ± 15 4.7 ± 1.2 18 ± 23 18 ± 16 25 ± 16 4.6 ± 6.5 84 ± 77 LDL 0.2 4.1 0.2 0.0 2.4 0.4 0.0 7.3 45 oxLDL 12 ± 9.3 31 ± 20 8.2 ± 4.0 64 ± 74 27 ± 17 52 ± 35 14 ± 15 190 ± 160 LDL 0.5 4.4 0.5 0.0 3.1 4.7 0.0 13 60 oxLDL 22 ± 14 50 ± 25 13 ± 6.7 140 ± 97 59 ± 18 58 ± 44 31 ± 7.6 360 ± 190 LDL 0.2 4.3 0.2 0.0 1.1 2.1 0.0 8.0 75 oxLDL 37 ± 10 69 ± 11 23 ± 9.3 200 ± 10 68 ± 14 100 ± 80 51 ± 6.4 520 ± 60 LDL 0.1 4.0 0.0 0.0 2.8 0.7 0.0 7.5 90 oxLDL 46 ± 11 84 ± 13 36 ± 11 200 ± 1.8 77 ± 4.6 99 ± 85 45 ± 2.2 580 ± 51 LDL 1.5 5.8 2.5 0.0 4.2 6.1 0.0 20 120 oxLDL 64 ± 3.6 95 ± 16 72 ± 19 190 ± 6.3 100 ± 1.8 100 ± 97 56 ± 9.3 670 ± 49 LDL 1.5 6.5 2.6 0.0 3.3 5.8 0.0 20 240 oxLDL 67 ± 5.1 130 ± 21 270 ± 49 170 ± 43 120 ± 2.9 130 ± 130 110 ± 5.9 990 ± 16 LDL 1.6 7.7 3.5 0.0 7.4 8.0 0.0 28 300 oxLDL 68 ± 1.5 120 ± 17 260 ± 44 180 ± 45 120 ± 16 120 ± 140 120 ± 32 1000 ± 96
PC - phosphatidylcholine. a LDL - non-oxidised LDL; Data is represented as nmol/μmol of PC containing SFA+MUFA, n = 1; oxLDL - LDL oxidised for 0 to 300 min with copper chloride; Data is represented as mean ± SD, n = 2, expressed as nmol/μmol of PC containing SFA+MUFA.
99 CHAPTER 3 – CHARACTERISATION OF CHANGES IN THE LIPID COMPOSITION OF LIPOPROTEINS DUE TO OXIDATIVE STRESS
Table 3.3 The effect of oxidation on the level of oxidised cholesteryl ester in low density lipoprotein.
Time (min) Sample a CE 16:1 (O) CE 16:0 (O) CE 16:0 (O2) CE 18:1 (O) CE 18:2 (O) Total oxCE LDL 33 17 0.0 1200 450 1700 0 oxLDL 60 ± 29 19 ± 8.0 0.0 ± 0.0 1600 ± 70 2400 ± 2600 4100 ± 2800 LDL 28 9.5 0.0 1400 310 1700 15 oxLDL 74 ± 50 11 ± 5.6 0.0 ± 0.0 2100 ± 780 6100 ± 7900 9000 ± 9500 LDL 92 12 0.0 1700 540 2400 30 oxLDL 32 ± 10 68 ± 77 15 ± 21 2700 ± 1200 6300 ± 6900 10000 ± 9300 LDL 100 13 0.0 1700 510 2300 45 oxLDL 280 ± 390 190 ± 140 47 ± 67 4000 ± 1400 12000 ± 9000 19000 ± 13000 LDL 40 9.1 0.0 1600 510 2100 60 oxLDL 820 ± 820 350 ± 280 140 ± 110 6200 ± 2900 24000 ± 3400 37000 ± 12000 LDL 82 7.7 0.0 1400 970 2500 75 oxLDL 1400 ± 400 580 ± 232 250 ± 53 9500 ± 4000 27000 ± 7000 49000 ± 17000 LDL 33 18 0.0 1400 540 2000 90 oxLDL 1300 ± 470 648 ± 170 280 ± 97 12000 ± 6100 26000 ± 6000 51000 ± 16000 LDL 40 15 0.0 1700 1600 3400 120 oxLDL 2500 ± 50 1200 ± 49 400 ± 6.1 16000 ± 2500 30000 ± 2600 67000 ± 1800 LDL 25 11 0.0 1200 1000 2400 240 oxLDL 6600 ± 1500 2600 ± 930 600 ± 41 40000 ± 16000 33000 ± 2600 99000 ± 20000 LDL 31 10 0.0 1900 1600 4000 300 oxLDL 5800 ± 1900 2400 ± 280 710 ± 1.4 34000 ± 7200 30000 ± 5500 88000 ± 19000
CE - cholesteryl ester. a LDL - non-oxidised LDL; Data is represented as nmol/μmol of PC containing SFA+MUFA, n = 1; oxLDL - LDL oxidised for 0 to 300 min with copper chloride; Data is represented as mean ± SD, n = 2, expressed as nmol/μmol of PC containing SFA+MUFA.
100 CHAPTER 3 – CHARACTERISATION OF CHANGES IN THE LIPID COMPOSITION OF LIPOPROTEINS DUE TO OXIDATIVE STRESS
Table 3.4 The effect of oxidation on the level of oxidised phosphatidylcholine in high density lipoprotein.
Time Sample a PC 34:3 (O) PC 34:2 (O) PC 34:3 (O ) PC 34:2 (O ) PC 36:3 (O) PC 36:2 (O) PC 36:3 (O ) Total oxPC (min) 2 2 2 HDL 0.3 ± 0.0 3.9 ± 0.9 0.3 ± 0.1 0.3 ± 0.1 1.0 ± 3.3 4.9 ± 0.1 0.0 ± 0.0 11 ± 4.4 0 oxHDL 0.4 ± 0.2 4.7 ± 0.0 0.5 ± 0.4 0.5 ± 0.4 4.0 ± 0.9 4.4 ± 0.6 0.0 ± 0.0 15 ± 2.8 HDL 0.5 ± 0.1 4.2 ± 0.7 0.7 ± 0.3 0.7 ± 0.3 4.2 ± 0.5 1.5 ± 1.8 0.0 ± 0.0 12 ± 1.3 15 oxHDL 3.5 ± 1.9 12 ± 0.4 5.0 ± 4.7 13 ± 0.5 15 ± 2.7 10 ± 0.3 3.6 ± 1.4 62 ± 8.0 HDL 0.5 ± 0.2 4.4 ± 0.3 0.5 ± 0.1 0.5 ± 0.1 3.9 ± 0.2 4.4 ± 0.0 0.0 ± 0.0 14 ± 0.0 30 oxHDL 7.3 ± 3.5 30 ± 3.7 5.5 ± 3.3 39 ± 3.4 33 ± 4.1 25 ± 4.2 9.7 ± 3.2 150 ± 19 HDL 0.7 ± 0.4 4.0 ± 0.7 0.9 ± 0.4 0.9 ± 0.4 2.2 ± 2.1 3.6 ± 0.8 0.0 ± 0.0 12 ± 2.4 45 oxHDL 9.3 ± 2.6 43 ± 6.3 5.3 ± 1.7 61 ± 0.5 48 ± 2.2 44 ± 7.7 9.7 ± 2.3 220 ± 23 HDL 1.0 ± 0.5 5.2 ± 0.3 1.2 ± 0.6 1.2 ± 0.6 5.1 ± 0.2 6.2 ± 1.1 0.0 ± 0.0 20 ± 3.4 60 oxHDL 12± 0.2 55 ± 6.7 6.1 ± 0.5 66 ± 16 48 ± 1.6 46 ± 7.7 18 ± 1.6 250 ± 12 HDL 0.9 ± 0.5 4.6 ± 0.2 1.3 ± 0.8 1.9 ± 1.2 4.5 ± 0.3 4.9 ± 0.4 0.0 ± 0.0 18 ± 3.4 75 oxHDL 17 ± 2.2 73 ± 8.1 8.1 ± 0.5 100 ± 11 74 ± 11 66 ± 6.5 30 ± 6.8 370 ± 26 HDL 1.0 ± 0.6 4.8 ± 0.1 1.2 ± 0.6 1.2 ± 0.6 4.8 ± 0.9 1.5 ± 1.8 0.0 ± 0.0 15 ± 0.7 90 oxHDL 25 ± 5.8 88 ± 4.0 13 ± 1.8 150 ± 20 78 ± 2.2 78 ± 8.3 28 ± 14 460 ± 12 HDL 1.5 ± 0.9 5.0 ± 0.3 1.9 ± 1.2 1.9 ± 1.2 6.6 ± 1.6 5.1 ± 0.4 0.0 ± 0.0 22 ± 5.0 120 oxHDL 47 ± 15 110 ± 11 23 ± 11 250 ± 74 94 ± 17 87 ± 17 51 ± 19 660 ± 160 HDL 2.1 ± 1.4 5.8 ± 0.4 2.4 ± 1.3 2.4 ± 1.3 6.2 ± 1.1 8.7 ± 3.0 0.0 ± 0.0 28 ± 8.5 240 oxHDL 83 ± 1.4 170 ± 18 100 ± 17 330 ± 12 140 ± 8.8 140 ± 13 58 ± 7.9 1000 ± 42 HDL 1.4 ± 0.3 5.2 ± 0.8 1.9 ± 0.4 1.9 ± 0.4 4.5 ± 1.0 4.6 ± 1.6 0.0 ± 0.0 20 ± 0.1 300 oxHDL 86 ± 7.6 140 ± 8.1 160 ± 9.0 440 ± 25 150 ± 11 140 ± 8.1 79 ± 25 1200 ± 76
PC - phosphatidylcholine. a HDL - non-oxidised HDL; oxHDL - HDL oxidised for 0 to 300 min with copper chloride. All data is represented as mean ± SD, n = 2, expressed as nmol/μmol of PC containing SFA+MUFA.
101 CHAPTER 3 – CHARACTERISATION OF CHANGES IN THE LIPID COMPOSITION OF LIPOPROTEINS DUE TO OXIDATIVE STRESS
Table 3.5 The effect of oxidation on the level of oxidised cholesteryl ester in high density lipoprotein.
Time a Sample CE 16:1 (O) CE 16:0 (O) CE 16:0 (O2) CE 18:1 (O) CE 18:2 (O) CE 18:2 (O2) CE 18:3 (O) CE 18:3 (O2) Total oxCE (min) HDL 6.4 ± 8.7 3.8 ± 3.6 0.0 ± 0.0 670 ± 29 180 ± 32 27 ± 4.8 14 ± 4.0 0.0 ± 0.0 900 ± 74 0 oxHDL 10 ± 3.2 14 ± 12 0.0 ± 0.0 690 ± 70 280 ± 110 39 ± 0.8 25 ± 12 38 ± 41 1100 ± 110 HDL 7.5 ± 12 4.1 ± 0.4 0.0 ± 0.0 440 ± 290 230 ± 55 61 ± 9.3 63 ± 34 38 ± 27 850 ± 280 15 oxHDL 10 ± 9.1 5.1 ± 2.6 0.0 ± 0.0 740 ± 150 800 ± 41 250 ± 130 200 ± 170 340 ± 420 2300 ± 920 HDL 20 ± 9.3 4.9 ± 1.0 0.0 ± 0.0 500 ± 6.1 200 ± 57 64 ± 19 44 ± 19 47 ± 30 880 ± 5.0 30 oxHDL 16 ± 15 4.2 ± 0.9 6.6 ± 9.3 930 ± 75 2400 ± 200 640 ± 270 300 ± 180 180 ± 130 4500 ± 850 HDL 8.4 ± 1.4 3.4 ± 0.9 0.0 ± 0.0 640 ± 44 130 ± 170 40 ± 2.2 37 ± 15 46 ± 24 910 ± 92 45 oxHDL 16 ± 6.3 6.7 ± 1.4 25 ± 5.8 1200 ± 33 4300 ± 240 1400 ± 140 470 ± 75 230 ± 52 7600 ± 260 HDL 15 ± 5.2 2.5 ± 0.2 0.0 ± 0.0 650 ± 99 230 ± 11 84 ± 40 100 ± 61 120 ± 87 1200 ± 300 60 oxHDL 20 ± 7.2 6.0 ± 2.1 21 ± 8.0 1500 ± 420 6100 ± 1800 2200 ± 380 710 ± 18 330 ± 27 11000 ± 1900 HDL 9.4 ± 21 3.0 ± 3.2 0.0 ± 0.0 640 ± 67 140 ± 57 51 ± 12 46 ± 19 72 ± 51 950 ± 66 75 oxHDL 40 ± 26 28 ± 30 23 ± 10 1500 ± 59 6400 ± 930 2600 ± 360 850 ± 140 440 ± 65 12000 ± 1400 HDL 16 ± 1.1 4.7 ± 1.8 0.0 ± 0.0 690 ± 120 180 ± 34 56 ± 12 47 ± 20 49 ± 25 1000 ± 140 90 oxHDL 210 ± 87 54 ± 69 18 ± 2.3 2100 ± 220 7800 ± 1000 4900 ± 970 1200 ± 220 580 ± 174 17000 ± 2800 HDL 6.6 ± 4.3 4.1 ± 0.2 0.0 ± 0.0 680 ± 120 300 ± 22 90 ± 52 91 ± 65 150 ± 110 1300 ± 360 120 oxHDL 350 ± 100 140 ± 20 20 ± 29 2700 ± 26 7600 ± 570 9900 ± 430 1800 ± 91 1700 ± 510 24000 ± 1700 HDL 9.8 ± 7.4 11 ± 3.5 0.0 ± 0.0 830 ± 230 290 ± 14 100 ± 34 94 ± 53 190 ± 120 1500 ± 420 240 oxHDL 1100 ± 75 420 ± 110 49 ± 19 5600 ± 770 14000 ± 1600 14000 ± 610 2900 ± 58 8200 ± 1800 47000 ± 5100 HDL 20. ± 3.8 4.7 ± 4.6 0.0 ± 0.0 640 ± 15 310 ± 150 130 ± 42 82 ± 11 150 ± 31 1300 ± 110 300 oxHDL 1300 ± 390 600 ± 48 89 ± 26 9100 ± 630 16000 ± 5000 15000 ± 4400 3000 ± 1300 11000 ± 2500 55000 ± 14000
CE- cholesteryl ester. a HDL - non-oxidised HDL; oxHDL - HDL oxidised for 0 to 300 min with copper chloride. All data is represented as mean ± SD, n = 2, expressed as nmol/μmol of PC containing SFA+MUFA.
102 CHAPTER 3 – CHARACTERISATION OF CHANGES IN THE LIPID COMPOSITION OF LIPOPROTEINS DUE TO OXIDATIVE STRESS
3.5 DISCUSSION
The level of oxidation in samples can be measured in several ways. To monitor and confirm the oxidation of lipoproteins, we crudely measured their oxidation levels by the production of conjugated diene, and TBARS which consist predominantly of malondialdehyde; both conjugated diene and TBARS are markers of PUFA oxidation [343-345]. We found that both LDL and HDL were successfully oxidised over time with copper chloride. They showed lipoprotein-associated distinct patterns of the production of conjugated diene which corresponded well with the TBARS formation. This finding is consistent with a previous report [333] that demonstrated the same correlation between the two measurements.
After we confirmed the oxidation of the lipoproteins, we further investigated the associated changes in the lipid composition. Alkylphospholipids such as alkylphosphatidylcholine and alkylphosphatidylethanolamine are lipids in the outer layers of lipoproteins with alkylphosphatidylethanolamine being a precursor to the synthesis of both PE plasmalogen and PC plasmalogen in the biosynthetic pathway. Unlike the vinyl ether bonds of plasmalogens, ether bonds of alkylphosphatidylcholine and alkylphosphatidylethanolamine are relatively resistant to oxidation. While relative to SFA+MUFA PC, the total levels of alkylphospholipids were decreased during the oxidation, analyses of the individual species of the lipid subclass suggests that alkylphosphatidylcholine containing SFA or MUFA are resistant to oxidation compared to alkylphosphatidylcholine or alkylphosphatidylethanolamine containing PUFA. This finding suggests that the oxidation of alkylphospholipids is likely to occur on the double bonds of the fatty acid chains and not on the ether bond of the alkyl chain.
Plasmalogens are lipids that are proposed to have anti-oxidant properties and are also located in the outer layer of lipoproteins. Plasmalogens, both containing SFA and MUFA or PUFA were affected by oxidation as their levels relative to SFA+MUFA PC were decreased over time during the oxidation experiments. This depletion of lipid is consistent with a previous report [251] that showed that plasmalogens were susceptible to oxidation due to their vinyl ether bonds. We also observed that PE plasmalogen was affected early in the oxidation process and was depleted first compared to PC plasmalogen, suggesting that PE plasmalogen was preferentially oxidised relative to PC plasmalogen. To our knowledge, this is the first report on the susceptibility of PE
103 CHAPTER 3 – CHARACTERISATION OF CHANGES IN THE LIPID COMPOSITION OF LIPOPROTEINS DUE TO OXIDATIVE STRESS plasmalogen compared to PC plasmalogen to oxidation. Studies have shown that the choline headgroup has a greater hydration than the ethanolamine headgroup whereas the ethanolamine has an expanded hydrophobic volume [346]. We speculate that these characteristics result in PE plasmalogen sitting less closely packed in the membrane compared to PC plasmalogen and consequently renders the vinyl ether bond more accessible to free radicals. Thus PE plasmalogen is more susceptible to oxidation. Additionally, we found that relative to SFA+MUFA PC, PE plasmalogen in LDL was depleted faster than in HDL. HDL is known to possess anti-oxidative mechanisms such as the ability of apoAI to reduce lipid hydroperoxides to respective hydroxides [33] and so this capacity may help to reduce the accumulation of the oxidative intermediates (i.e. lipid hydroperoxides) in HDL and in turn, this may reduce the propagation of free radicals to other lipid species in HDL including plasmalogen.
Lysophosphatidylcholine in lipoprotein is a product of either oxidative degradation of plasmalogen or enzymatic cleavage of glycerophospholipids by Lp-PLA2. An increase in the total level of lysophosphatidylcholine is often associated with disease progression and this has been demonstrated in stable atherosclerotic plaques and oxLDL [197, 200]. Consistent with previous reports, we observed increased levels of lysophosphatidylcholine relative to SFA+MUFA PC in the oxidised lipoproteins. We then analysed the regioisomers of lysophosphatidylcholine to investigate which isomers predominantly contribute to the total lysophosphatidylcholine prior to, during and after oxidation. Sn-1 LPC contains fatty acid at the sn-1 position of the glycerol backbone and are produced by the enzymatic action of Lp-PLA2 on the fatty acyl chain at the sn-2 position of the glycerol backbone; Whereas, sn-2 LPC contains fatty acid at the sn-2 position and are produced by oxidative degradation of the vinyl ether bond of plasmalogen at the sn-1 position. Our analysis revealed that the total lysophosphatidylcholine in LDL and HDL was predominantly composed of sn-1 LPC. Interestingly, analysis of the molecular species of lysophosphatidylcholine showed different patterns of the production of SFA-or MUFA containing lysophosphatidylcholine and PUFA-containing lysophosphatidylcholine. The most abundant SFA-containing lysophosphatidylcholine in lipoproteins such as LPC 18:0 had higher level of sn-1 compared to sn-2 isomer relative to SFA+MUFA PC in both LDL and HDL. The levels of both sn-1 and sn-2 isomers of LPC 18:0 were generally increased with oxidation particularly in LDL. This suggests that during oxidation, both
104 CHAPTER 3 – CHARACTERISATION OF CHANGES IN THE LIPID COMPOSITION OF LIPOPROTEINS DUE TO OXIDATIVE STRESS oxidation at the vinyl ether bond and Lp-PLA2 play roles in producing SFA- and MUFA- containing lysophosphatidylcholine. In contrast, PUFA-containing lysophosphatidylcholine in LDL and HDL was predominantly composed of sn-2 isomers. Our findings were consistent with previous reports, which demonstrated that the predominant regioisomer in human plasma is sn-1 LPC which contains MUFA and PUFA; whereas sn-2 LPC contains predominantly only PUFA [347, 348]. Relative to SFA+MUFA PC, the level of sn-2 LPC increased initially and then declined with oxidation. This initial increase of PUFA-containing lysophosphatidylcholine suggests its production by oxidative degradation of plasmalogen, whereas the lipid steady decline after 60 min of oxidation suggests a secondary oxidation at the remaining fatty acid chain. In contrast, the level of sn-1 LPC tends to stay consistent throughout the oxidation but was decreased at a higher level of oxidation. This suggests an oxidative degradation of the lipid.
HDL has been reported to remove and inactivate lipid hydroperoxides from oxidised LDL (oxLDL) upon co-incubation of the lipoproteins [33, 349]. This ability to transfer oxidised lipids from oxLDL to HDL may play an important part of the overall HDL anti-oxidative capacity. This transfer capacity was influenced by the surface rigidity of the acceptor particle; a low ratio of SM/PC reduced surface rigidity and aided in the transfer efficiency of oxidised lipids, and in the delay of LDL oxidation [33, 168]. Once transferred the lipid hydroperoxides were subsequently reduced to their respective hydroxides by HDL-associated apoAI [350]. This process of inactivation of lipid hydroperoxides was governed by the total HDL content of apoAI and the redox status of the methionine residues of apoAI [33]. In light of this earlier study, we examined the level of sphingomyelin relative to total PC in the lipoproteins. We demonstrated that oxidation to both LDL and HDL resulted in the increase in SM/PC ratio due to a reduction in the content of the polyunsaturated PC (See Supplementary Table 3.11 – 3.14), thus suggesting a possible overall increase in the surface rigidity of the LDL and HDL upon oxidation. This altered surface lipid rigidity may contribute to the modulation of HDL anti-oxidative capacity.
We observed no difference in the level of free cholesterol relative to SFA+MUFA PC throughout the oxidation in both LDL and HDL, suggesting that the lipid was not susceptible to oxidation. It is also possible that any difference in the lipid level due to oxidation was not great enough to be seen as significant. To allow us to better assess the
105 CHAPTER 3 – CHARACTERISATION OF CHANGES IN THE LIPID COMPOSITION OF LIPOPROTEINS DUE TO OXIDATIVE STRESS effect of oxidation on cholesterol, we measured the level of oxidative by-products of cholesterol such as 7-ketocholesterol and 7β-hydroxycholesterol. This will be covered in the next chapter.
Levels of lipids in the core of lipoproteins such as cholesteryl ester and triacylglycerol were decreased during the oxidation relative to SFA+MUFA PC; the reduction in both cholesteryl ester and triacylglycerol levels were greater in LDL compared to that in HDL, suggesting that the lipids in outer layer and core of LDL were affected more by oxidation possibly due to the lack of an anti-oxidative mechanism in LDL compared to HDL as discussed above. Analyses on the molecular species of triacylglycerol showed that PUFA-containing triacylglycerol was more susceptible to oxidation compared to MUFA-containing triacylglycerol in both LDL and HDL. This was as we expected as PUFA-containing triacylglycerol contains double bonds which are susceptible to oxidation.
While our targeted lipidomic approach provided us with information on the relative reduction or modulation in the amount of the lipoprotein native lipid species, we needed to identify the products of this oxidation to give us a more complete picture of the lipoprotein oxidation and the associated altered lipoprotein composition. To this end we used an untargeted lipidomic approach on LDL prior to and following oxidation with copper sulphate. From our analyses, we found seven species of oxPC and seven species of oxCE (Table 3.1). Subsequently we validated the measurements of these oxidised lipid species in an independent LDL and HDL oxidation experiment using copper chloride. We demonstrated that relative to SFA+MUFA PC, the total oxPC and oxCE levels in LDL and HDL were increased throughout the oxidation. While characterisation of the exact structure of the oxidised lipids identified is beyond the scope of the current work and in most instances may represent isomeric mixtures of molecular species, the observed mass and chromatographic properties are consistent with these assignments. We can speculate that the oxPC is likely to arise from the oxidation of diacylglycerophospholipids or 1-alkyl-2-acylglycerophospholipids. Additionally the proposed oxPC and oxCE species are consistent with oxidised lipids previously detected in human atherosclerotic plaques [324, 351] and plasma of a rabbit model of atherosclerosis [352], thus highlighting their physiological relevance. Furthermore, these oxidised lipids can potentially be useful biomarkers for clinical measurement. Further investigation is required to determine whether these products of lipid oxidation
106 CHAPTER 3 – CHARACTERISATION OF CHANGES IN THE LIPID COMPOSITION OF LIPOPROTEINS DUE TO OXIDATIVE STRESS by copper chloride overlap with those produced by myeloperoxidase, an enzyme which catalyses lipoprotein oxidation [353] and has been shown to be associated with CAD [354].
In summary, we have demonstrated that relative to SFA+MUFA PC, oxidation reduced the levels of plasmalogen and PUFA-containing alkylphospholipids in the outer layer of lipoproteins, and cholesteryl ester and PUFA-containing triacylglycerol in the core of lipoproteins; whereas oxidation increased SM/PC and the level of lysophosphatidylcholine. SFA- and MUFA- containing lysophosphatidylcholine and PUFA containing lysophosphatidylcholine were produced differently during the oxidation and the regioisomers gave us a clue to the major sources of the lysophosphatidylcholine. In addition, no significant difference in SFA or MUFA- containing alkylphospholipids and free cholesterol relative to SFA+MUFA PC were observed with oxidation. Lipids in LDL were generally more susceptible to oxidation as compared to those of HDL particularly cholesteryl ester and triacylglycerol in the core of lipoprotein. In conclusion, oxidation resulted in a myriad of changes to the lipid composition of lipoproteins. In contrast to our hypothesis, lipids in both the outer layer and core of lipoproteins were affected by oxidation within the same amount time, but more so in LDL than in HDL. These findings highlight the different susceptibility of the lipids in the two classes of lipoproteins to oxidation.
3.6 LIMITATIONS
The time-course experiment of lipoproteins inevitably resulted in a large number of samples to process at one time. This large-scale experiment limited the number of biological and technical replicates in this study. To minimise variance between subjects, we had pooled plasma from six healthy volunteers to form one biological sample. We acknowledge that lipoproteins from different individuals may have different susceptibility to oxidation and thus further studies are required to investigate this. The lack of technical repeats has resulted in the lack of statistical power in these analyses but the dramatic changes in lipid composition are nonetheless convincing.
Although our methodology adopted from Kleinveld et al. [333] successfully oxidised the lipoproteins, most of the lipids analysed were depleted by 120 min of oxidation, half-way through the experiment. To enable us to assess changes in the lipid
107 CHAPTER 3 – CHARACTERISATION OF CHANGES IN THE LIPID COMPOSITION OF LIPOPROTEINS DUE TO OXIDATIVE STRESS composition and functions of the lipoproteins in greater details, we required a slower method of oxidation and this will be explored in the next chapter.
While the current study has revealed a myriad of changes to the lipid composition of lipoproteins, their effects on the lipoprotein functions are still not fully understood. To address these issues, the methodology to measure oxidised lipids is further used in the next chapter to investigate HDL functions including the ability of HDL to remove LDL- derived oxidised lipids and to remove cholesterol from THP1-derived macrophages.
108
CHAPTER 4 - CHARACTERISATION OF HIGH DENSITY LIPOPROTEIN FUNCTION FOLLOWING OXIDATIVE STRESS
109
CHAPTER 4 – CHARACTERISATION OF HDL FUNCTION FOLLOWING OXIDATIVE STRESS
Some parts of this chapter has been published - A.A Rasmiena, C.K Barlow, T.W Ng, P.J Meikle. High density lipoprotein efficiently accepts surface but not internal oxidised lipids from oxidised low density lipoprotein. Biochimica et Biophysica Acta - Molecular and Cell Biology of Lipids, 2016. 1861(2): 69 - 77. (see Appendices page 266).
4 CHARACTERISATION OF HDL FUNCTION FOLLOWING OXIDATIVE STRESS
4.1 ABSTRACT
Background: HDL possesses anti-atherogenic mechanisms including reverse cholesterol transport, as well as anti-oxidative properties such as the ability to delay LDL oxidation, and to remove and inactivate lipid hydroperoxides from oxLDL. The effect of altered lipid compositions on lipoprotein function is incompletely understood. In addition, our knowledge on the identities of oxidised lipids that are removed from oxLDL or those that may contribute to the impairment of the atheroprotective functions of HDL are also limited. In this chapter we aimed to investigate the effect of altered lipid composition resulting from oxidation of HDL on the cholesterol efflux capacity, the ability to delay LDL oxidation, and the ability to accept oxidised lipids from oxLDL. Methods/results: HDL obtained from pooled plasma of normolipidemic subjects (n = 5) was oxidised under mild and heavy oxidative conditions. Analysis showed that the ability of HDL to efflux [3H]-cholesterol from THP1-derived macrophages was reduced by 28% (P<0.001) and 34% (P<0.001) in mildly- and heavily-oxidised condition. Relative to SFA+MUFA PC, both oxPC and oxCE were reduced in the re-isolated LDL following co-oxidation with native HDL. Lipoprotein surface lipids, oxidised phosphatidylcholines and oxidised cholesterol (7β-hydroxycholesterol), but not internal oxidised cholesteryl esters, were effectively transferred to native HDL. Saturated and monounsaturated lysophosphatidylcholine were also transferred from oxLDL to native HDL. However, these processes were attenuated when HDL was oxidised under mild and heavy oxidative conditions. The impaired capacities were accompanied by an increase in a ratio of sphingomyelin to phosphatidylcholine and a reduction in phosphatidylserine content relative to PC in oxidised HDL. Conclusions: Mild to strong oxidation to HDL affected the lipoprotein function including its ability to efflux cholesterol from macrophages, the ability to delay LDL
110
CHAPTER 4 – CHARACTERISATION OF HDL FUNCTION FOLLOWING OXIDATIVE STRESS
oxidation and to efficiently transfer surface oxidised lipids (oxPC and oxidised cholesterol) from oxLDL to HDL.
4.2 INTRODUCTION
It is widely known that HDL possesses anti-atherogenic mechanisms including reverse cholesterol transport, anti-oxidative, anti-inflammatory, and anti-thrombotic functions. Some of these functions are better characterised than others. In this chapter, we will investigate the effect of oxidation on the cholesterol efflux and anti-oxidative functions of HDL.
The ability of HDL to efflux cholesterol from peripheral cells via ABCA1 is part of the reverse cholesterol transport system that is particularly important to protect against atherosclerosis. Oxidation to HDL has been shown to impair its ability to efflux cholesterol from foam cells [355] and fibroblasts [356]. Although the level of HDL-C represents a risk factor for CAD, its measurement alone is inadequate to determine HDL function and therapeutic efficacy [357]. Ex vivo studies showed that cholesterol efflux was inversely associated with carotid-intima media thickness in CAD patients, independent of the HDL-C level [358]. Furthermore, cholesterol efflux was shown to be inversely associated with risk of future cardiovascular events [359, 360]. It was demonstrated that the cholesterol efflux ability of HDL was impaired in humans with atherosclerosis [361] but was moderately improved with Niacin treatment [362]. These findings underline the need to look at HDL function to better assess its atheroprotective capacity. Additionally, HDL-C levels and other traditional risk factors can be used in conjunction with the measurement of HDL function to improve the discrimination and re-classification of patient risk [360].
Studies have characterised HDL anti-oxidative properties by its ability to delay LDL oxidation, typically monitored by the production of conjugated diene. There is also a growing body of evidence highlighting the defective anti-oxidative capacity of HDL in atherosclerosis. Decreased protection of LDL against oxidation [363, 364], and triglyceride and serum amyloid A enrichment of HDL [365] have all been associated with its impaired anti-oxidative capacity. Earlier studies on HDL subpopulations from normolipidemic individuals showed that small dense HDL3c exhibited the greatest potency in inhibiting LDL oxidation as well as in cholesterol efflux. Analysis of the lipid composition demonstrated that the small dense HDL3c was preferentially enriched
111
CHAPTER 4 – CHARACTERISATION OF HDL FUNCTION FOLLOWING OXIDATIVE STRESS
in sphingosine 1-phosphate, phosphatidylserine, and phosphatidic acid, and was depleted in sphingomyelin [226, 227].
HDL has also been reported to remove and inactivate lipid hydroperoxides from oxLDL upon co-incubation [33, 349]. These processes are governed by the redox status of HDL-associated apoAI, and the surface rigidity of the phospholipid monolayer of the acceptor HDL (as determined by the ratio of sphingomyelin/phosphatidylcholine) [33]. This ability to transfer oxidised lipids from oxLDL to HDL may play an important part of the overall HDL anti-oxidative capacity.
OxLDL contributes to atherosclerosis [366-368]. Lipid oxidation products including oxPC, oxCE and lysophosphatidylcholine which is produced by the action of phospholipases on oxPC, have been detected and characterised in oxLDL [200], plasma [352] and atherosclerotic lesions [324]. These oxidation products represent bioactive lipids with potential pro-inflammatory capacity that affect plaque progression and stability [200]. In Chapter 3, we observed that the levels of lipid oxidation products were elevated in LDL and HDL with oxidation. This increase was accompanied with a myriad of changes in the lipid composition of these particles. However, the effect of altered lipid compositions on lipoprotein function is incompletely understood. Previous studies have relied heavily on conjugated diene measurement and HPLC chemiluminescence coupled with UV detection measurement of lipid hydroperoxides to assess anti-oxidative capacity of HDL. Our knowledge on the identities of oxidised lipids that are transferred to HDL or those that may contribute to the impairment of the atheroprotective functions of HDL are limited. Therefore, in this chapter we aimed to oxidise lipoproteins in an artificial system using copper chloride and investigate the effect of altered lipid composition resulting from oxidation of HDL on the cholesterol efflux capacity, the ability to delay LDL oxidation, and the ability to accept oxidised lipids from oxLDL. Using the methodology developed in Chapter 3, we also aimed to identify major oxidised lipids that were implicated in impaired HDL functionality.
4.3 METHODS
4.3.1 Cholesterol efflux assay
To assess the cholesterol efflux capacity, isolated human HDL (0.3 mg protein/ml) was oxidised with 1 μmol/l copper chloride for 0, 120, and 240 min at 37oC. The oxidation
112
CHAPTER 4 – CHARACTERISATION OF HDL FUNCTION FOLLOWING OXIDATIVE STRESS
was terminated by the addition of EDTA to a final concentration of 2 μmol/l and reducing the sample temperature to 4oC.
THP1 monocytes were cultured and maintained as described in Chapter 2 (section 2.2.3). The cholesterol efflux activity of HDL was assessed as previously described [369], with slight modifications. Briefly, THP1 cells (300,000 cells/well) in 24 well plates were labeled with [3H]-cholesterol (0.5 μCi/well) and allowed to differentiate to form macrophages by the addition of Phorbol 12-Myristate 13-Acetate at 0.1 µg/ml o followed by incubation for 72 h at 37 C, 5% CO2. The cells were then washed with PBS (without Ca and Mg) and fresh serum-free media (300 μl/well) containing LXR agonist (TO-901317, 4 μmol/l) was added to upregulate the cellular ABCA1 transporters. The o cells were incubated for 18 h at 37 C, 5% CO2. Subsequently, native/oxidised HDL was added to the cells to a final concentration of 20 μg protein/ml as cholesterol acceptors o and the samples were incubated for 2 h at 37 C, 5% CO2. The media was then collected and centrifuged at 16,000 xg for 5 min to remove any cell debris. The cells were frozen at -20oC for 30 min and allowed to sit in 500 μl of MilliQ water overnight at 4oC to detach the adherent cells from the bottom of the wells. Aliquots of 100 μl of the cells and the media which was collected the day before were mixed with 5 ml of Insta-gel Plus scintillation fluid and the level of [3H]-cholesterols in the samples were measured using a LS 6500 Scintillation Counter (Beckman Coulter, New South Wales, Australia).
Cellular cholesterol efflux capacity was calculated as the proportion of the labeled cholesterol removed from the cells to the HDL acceptors in the media, with respect to the total labeled cholesterol in both the cells and the media. This is illustrated in the formula below. The calculation took into account the dilution factor (DF) of the samples (DF = 3 for the media counts and DF = 5 for the cell counts), as well as the background counts (in samples containing no HDL, known as blanks).
% efflux = (media counts x dilution factor) x 100%
((media counts x dilution factor) + (cell counts x dilution factor))
113
CHAPTER 4 – CHARACTERISATION OF HDL FUNCTION FOLLOWING OXIDATIVE STRESS
4.3.2 Assessment of the ability of HDL to delay LDL oxidation
To test the ability of HDL and oxHDL to delay the oxidation of LDL, 1 ml of isolated human HDL (0.3 mg protein/ml) was oxidised with 1 μmol/l copper chloride for 0, 60, and 200 min at 37oC. This oxidation was carried out to produce native (non-oxidised), mildly and heavily oxidised HDL. The oxidation was monitored by the production of conjugated diene at 234 nm. Subsequently, 200 μl of isolated human LDL was added into each of the HDL samples to a final concentration of 0.1 mg protein/ml and oxidation was resumed for another 200 min at 37oC. The co-oxidation of LDL and HDL was terminated by the addition of EDTA to a final concentration of 2 μmol/l and by reducing the sample temperature to 4oC. Positive and negative controls were included in these experiments; Positive control was HDL which was oxidised without LDL, and negative control was non-oxidised HDL.
Following the oxidation, lipoproteins were re-isolated from 400 μl of each sample by sequential ultracentrifugation using NaCl/KBr/NaBr solutions prepared as described in Chapter 2. The samples were processed as follows: the samples were adjusted to a density of 1.063 g/ml by the addition of 192 μl of 1.182 g/ml NaBr solution and then overlayed with 408 μl of 1.063 g/ml KBr solution to reach a total volume of 1.0 ml. The samples were centrifuged at 435,680 xg (100,000 rpm), 16oC for 3 h using TLA 120.2 rotor and Optima MAX-TL ultracentrifuge (Beckman Coulter, New South Wales, Australia). The LDL (density of 1.019 g/ml - 1.063 g/ml) was aspirated in the top 400 μl of the density gradient. The HDL (density of 1.063 g/ml to 1.21 g/ml) was isolated in the lower 400 μl of the density gradient. The intermediate layer (200 μl) in between the LDL and HDL was analysed and found to contain only 7 - 10 % of lipids, demonstrating clear separation of the LDL and HDL fractions. Further, SDS-PAGE analysis of the re-isolated oxLDL and HDL showed no cross-contamination of apoAI and apoB (data not shown). The re-isolated lipoproteins were dialysed against phosphate-buffered saline (100 x total sample volume) containing 5 μmol/l EDTA overnight. Aliquots (100 μl) of the samples were frozen at -80oC and then lyophilised. They were reconstituted in 10 μl of deionised water prior to lipid extraction. Lipid extraction and analysis using LC-MS/MS were carried out as described in Chapter 2 (section 2.2.5.1 to 2.2.5.3).
114
CHAPTER 4 – CHARACTERISATION OF HDL FUNCTION FOLLOWING OXIDATIVE STRESS
The following study has been published - A.A Rasmiena, C.K Barlow, T.W Ng, P.J Meikle. High density lipoprotein efficiently accepts surface but not internal oxidised lipids from oxidised low density lipoprotein. Biochimica et Biophysica Acta - Molecular and Cell Biology of Lipids, 2016. 1861(2): 69 - 77. (see Appendices page 266).
4.3.3 Assessment of the ability of HDL to accept oxidised lipids from oxLDL
To assess the effect of oxidation on the ability of HDL to acts as an acceptor of oxidised lipids from oxLDL, HDL was incubated with oxLDL at a protein ratio of 3:1 (HDL:LDL). This ratio provides approximately equal amounts of HDL and LDL lipids in the assay system. The lipoproteins were then re-isolated and the lipid composition was analysed by LC-ESI-MS/MS as described in Chapter 2 (section 2.2.5.3). Prior to co-incubation of the lipoproteins, 1.0 ml of HDL (1.8 mg protein/ml) was oxidised with 6 µmol/l copper chloride for 60 and 200 min to obtain mildly and heavily oxidised forms of HDL, respectively. One milliliter of LDL (0.6 mg protein/ml) was oxidised with 6 µmol/l copper chloride for 2 h to produce a minimally oxidised LDL, where the production of conjugated diene was at the late exponential phase. Both lipoprotein oxidations were terminated with the addition of EDTA to a final concentration of 12 μmol/l. Oxidised LDL and native HDL with final protein concentrations of 0.1 mg protein/ml and 0.3 mg protein/ml respectively in a total volume of 550 μl were co- incubated in the presence of 120 µmol/l EDTA (10x the concentration of EDTA used to terminate the oxidation) at 37oC for 2 h. The mildly- and heavily-oxHDL were co- incubated with oxLDL as described above. Positive and negative controls were included in these experiments; Positive control for LDL was LDL which was oxidised and not incubated with HDL, and negative control was non-oxidised LDL. Control samples for HDL were HDL with differing levels of oxidative stress that were incubated without oxLDL. The lipoproteins were re-isolated and the lipids were extracted for analysis as described in section 4.3.2 in this chapter.
Lysophosphatidylcholine species were analysed by their fragmentation ions at m/z 184 as described in Chapter 2 (section 2.2.5.2). The regioisomers of lysophosphatidylcholine such as sn-1 lysophosphatidylcholine (sn-1 LPC) where the fatty acid is at the sn-1 position of glycerol backbone were distinguished from the sn-2 isomer by their retention time on the chromatography; sn-2 isomers were eluted slightly earlier than sn-1 isomers [341, 342]. The liquid chromatography method coupled with tandem mass spectrometry
115
CHAPTER 4 – CHARACTERISATION OF HDL FUNCTION FOLLOWING OXIDATIVE STRESS
as described in Chapter 2 allow separation and measurement of sn-1 and sn-2 isomers of lysophosphatidylcholine.
The oxPC and oxCE species identified using the untargeted LC-MS approach in Chapter 3 (Table 3.1) were added to the multiple reaction monitoring list for quantification using Q3 product ion values of 184.1 and 369.4 for oxPCs and oxCEs respectively; Declustering, entrance, and cell exit potential were set at 100, 10, and 11 V, respectively for oxPC species, and 30, 10, and 12 respectively for oxCE species (Chapter 2, Table 2.3). The collision energy was set at 45 V for oxPC species, and 20 V for oxCE species (Chapter 2, Table 2.3). Retention times were established by comparing oxLDL with native LDL (Chapter 3, Table 3.1). Additionally, 7-ketocholesterol and 7β- hydroxycholesterol, known products of cholesterol oxidation [352], were included in the targeted analysis. Multiple reaction monitoring transitions and tandem mass spectrometry conditions for these lipids were established by comparison against authentic standards and quantification was achieved by comparison against a deuterium labeled standard of 7-ketocholesterol [352] (Chapter 2, Table 2.3). Tandem mass spectrometry conditions for other lipids we analysed including cholesteryl ester, cholesterol, phosphatidylcholine, lysophosphatidylcholine, sphingomyelin and phosphatidylserine were listed in Chapter 2, Table 2.3.
Peak integration was carried out using MultiQuant software v.2.1.1. Relative lipid concentrations were calculated by relating the peak area of each species to the peak area of the corresponding internal standard (Chapter 2, Table 2.3). Total lipids of each class were calculated as the sum of the relative concentration of individual lipid species within the class [297].
4.3.4 Data analysis and statistics
To correct for differences in sample recovery following re-isolation of the lipoprotein fractions, the concentration of oxidised lipids, oxPC, oxCE, oxysterols (7- ketocholesterol and 7β-hydroxycholesterol) as well as lysophosphatidylcholine were expressed relative to molar % of SFA+MUFA PC. SFA+MUFA PCs were found to be resistant to oxidation and so represent a stable factor for normalisation. We analysed a total of 50 species of phosphatidylcholine, including 15 species of SFA+MUFA PC (Supplementary Table 4.1). OxCE and oxysterols were also normalised to SFA- and MUFA-containing cholesteryl ester (SFA+MUFA CE) and cholesterol, respectively to
116
CHAPTER 4 – CHARACTERISATION OF HDL FUNCTION FOLLOWING OXIDATIVE STRESS
confirm what we observed with the earlier normalisation method (ie. relative to SFA+MUFA PC). Other lipids such as sphingomyelin and phosphatidylserine (Supplementary Table 4.2 and 4.3) were expressed relative to levels of total PC to reflect the relative contribution these lipid classes made to the surface lipids of the lipoprotein particles. Normalisation of lipid concentration to protein content was not possible in this case as the re-isolation of lipoproteins and subsequent dialysis resulted in the dilution of samples and thus low and inaccurate protein estimates. Statistical significance between sample groups was determined using one-way ANOVA, corrected for multiple comparisons using Benjamini Hochberg, followed by post-hoc analysis, corrected by Dunn-Sidak for multiple pair-wise comparisons, as well as Student t-test; P value of less than 0.05 in all of the statistical tests was considered significant.
4.4 RESULTS
4.4.1 Oxidation of HDL led to impaired cholesterol efflux ability
Non-oxidised HDL was able to efflux an average of 5.4% of [3H]-cholesterol from THP1-derived macrophages. This ability was reduced by 28% (P<0.001) in mildly- oxHDL and 34% (P<0.001) in heavily-oxHDL.
Figure 4.1 Cholesterol efflux capacity of native and oxidised HDL. Native and oxidised HDL were incubated with THP1 macrophages with [3H]-cholesterol. The ability of the HDL and oxidised HDL to efflux the [3H]-cholesterol from the cells were analysed as described in Methods (section 4.2.1). Each circle represents a replicate from three independent experiments (n = 3 - 4/ experiment). The line represents the mean, n = 11 - 12/HDL group. Data were analysed with a Student t-test, ***P<0.001 relative to the mean of % efflux by native HDL.
117
CHAPTER 4 – CHARACTERISATION OF HDL FUNCTION FOLLOWING OXIDATIVE STRESS
4.4.2 The ability of HDL to protect LDL against oxidation was dependent on its oxidative state
To assess the ability of HDL to delay oxidation of LDL, native HDL, and mildly- and heavily-oxHDL was co-oxidised in the presence of LDL. The lipoproteins were re- isolated and the amount of oxidised lipids species in the LDL and HDL were measured relative to the total levels of SFA+MUFA PC.
Compared to the negative control (non-oxidised HDL), oxidation reduced the level of alkenylphosphatidylcholine or PC plasmalogen, relative to SFA+MUFA PC, in all HDL samples with/without co-oxidation with LDL (Figure 4.2A). The level of alkenylphosphatidylcholine relative to SFA+MUFA PC was not significantly different between the oxHDL control and any of the oxidative states, post co-oxidation with LDL (ANOVA, P = 0.17, Figure 4.2A). Correspondingly, compared to the negative control, oxidation of HDL increased the level of lysophosphatidylcholine relative to SFA+MUFA PC (Figure 4.2B). One-way ANOVA showed a significant difference between the oxidised HDL fractions; Post-hoc analysis showed that this was due to elevated lysophosphatidylcholine in the heavily oxidised HDL relative to the other forms (native and mildly-oxidised) of HDL (P<0.01 for all, Figure 4.2B). Compared to the positive control (HDL which was oxidised without LDL), the level of lysophosphatidylcholine was increased significantly (71%, P<0.01) only in re-isolated heavily-oxHDL (Figure 4.2B).
The levels of oxPC and oxCE relative to SFA+MUFA PC were significantly different between the re-isolated HDL samples (ANOVA, P = 0.01 and P = 0.04, for Figure 4.2C and 4.2D, respectively). Post-hoc analysis showed that the re-isolated native HDL had a significantly lower level of oxPC relative to SFA+MUFA PC compared to the positive control (P<0.05), and heavily-oxHDL (P<0.01) (Figure 4.2C). Whereas, the level of oxCE relative to SFA+MUFA PC in the positive control and HDL of different oxidative states were not significantly different except for that between native HDL and heavily oxHDL (P<0.05, Figure 4.2D).
Oxidation of LDL reduced the level of alkenylphosphatidylcholine, relative to SFA+MUFA PC in the lipoproteins. Co-oxidation of LDL with or without native, mildly- and heavily-oxHDL did not result in difference in the levels of the lipid relative
118
CHAPTER 4 – CHARACTERISATION OF HDL FUNCTION FOLLOWING OXIDATIVE STRESS
to SFA+MUFA PC between the re-isolated LDL samples (ANOVA, P = 0.21, Figure 4.3A).
One-way ANOVA of lysophosphatidylcholine level relative to SFA+MUFA PC in the re-isolated LDL samples showed significant difference (P <0.001, Figure 4.3B). Post- hoc analysis showed that the level of lysophosphatidylcholine, relative to SFA+MUFA PC was significantly reduced in LDL that was co-oxidised with native HDL (P<0.01) and mildly oxHDL (P<0.05) compared to the positive control (Figure 4.3B). In addition, the level of lysophosphatidylcholine relative to SFA+MUFA PC in the heavily oxHDL was higher compared to that in the positive control (P<0.05), and oxLDL that was co-oxidised with native HDL (P<0.001) and mildly oxHDL (P<0.001) (Figure 4.3B).
Levels of oxPC and oxCE were significantly different among the re-isolated LDL samples (ANOVA, P<0.001 and P < 0.01 for Figure 4.3C and 4.3D, respectively). Post- hoc analysis showed that compared to oxLDL control, the level of oxPC relative to SFA+MUFA PC, was significantly decreased in the re-isolated LDL following co- oxidation with native HDL (-58%, P<0.001) and mildly oxHDL (-45%, P<0.001) (Figure 4.3C). Furthermore, the level of the lipid in LDL following co-oxidation with heavily oxHDL was significantly different to that in LDL that was co-oxidised with HDL in native (P<0.001) and mildly-oxidised states (P<0.01) (Figure 4.3C).
Similarly to oxPC profile, the level of oxCE relative to SFA+MUFA PC was significantly reduced in the re-isolated LDL following co-oxidation with native HDL compared to the positive control (-38%, P<0.01) and to that co-oxidised with heavily oxHDL (-34%, P<0.05) (Figure 4.3D).
4.4.3 HDL acceptance of oxidised lipids from oxLDL
To assess the ability of HDL to accept oxidised lipids from oxLDL, native, mildly- and heavily-oxHDL were co-incubated with oxLDL. The HDL and oxLDL were also incubated with buffer only as controls. The lipoproteins were re-isolated and the amount of lysophospholipids and oxidised lipids were measured as a percentage relative to the total level of SFA+MUFA PC.
119
CHAPTER 4 – CHARACTERISATION OF HDL FUNCTION FOLLOWING OXIDATIVE STRESS
4.4.3.1 Oxidised phosphatidylcholine (oxPC)
The levels of oxPC relative to SFA+MUFA PC were significantly different between the re-isolated LDL samples (ANOVA, P<0.001, Figure 4.4A). There was a significant net transfer of oxPC (sum of oxPC species measured) from the oxLDL to native (non- oxidised) HDL (Table 4.1). Post-hoc analysis showed that the level of oxPC relative to SFA+MUFA PC in the re-isolated oxLDL (ie. positive control) increased significantly compared to the non-oxidised LDL (P<0.01) (Figure 4.4A). In addition, the level of the lipid in oxLDL which was incubated with native HDL decreased significantly (-67%, P<0.05) compared to the oxLDL (Figure 4.4A, Table 4.1). In parallel, oxPC in the re- isolated HDL increased from 2.6% to 11% (relative to SFA+MUFA PC, P<0.001) post- incubation with oxLDL (Figure 4.4B, Table 4.2). This process was modulated when pre-oxidised HDL under mild and heavy oxidative conditions were used. Post hoc analysis also showed that compared to oxLDL, there was a 42% (P = 0.23) decrease and 27% (P = 0.73) increase of oxPC in the re-isolated oxLDL following incubation with mildly- and heavily- oxHDL, respectively (Figure 4.4A, Table 4.1). Correspondingly, there was an increase from 19% to 23% of oxPC (P = 0.09) and a decrease from 47% to 42% of oxPC (P = 0.33) in the mildly- and heavily- oxHDL, respectively (Figure 4.4B, Table 4.2).
4.4.3.2 Oxidised cholesteryl ester (oxCE)
One-way ANOVA test showed that level of oxCE (sum of all oxCE species measured) relative to SFA+MUFA PC was significantly different in the re-isolated LDL samples (P = 0.003). Post-hoc analysis showed that compared to the non-oxidised LDL, the level of the lipid was significantly different in the positive control and re-isolated LDL post incubation with native HDL, mildly-, or heavily-oxHDL (P<0.01 for all, Figure 4.4C). In contrast, there was no significant difference in the level of oxCE relative to SFA+MUFA PC in the re-isolated oxLDL, post-incubation with native HDL, mildly-, or heavily-oxHDL, compared to the oxLDL control (Figure 4.4C, Table 4.1). However, an increase in oxCE (relative to SFA+ MUFA PC) was observed in the re-isolated native HDL from 10% to 26% (P<0.05); the increase in oxCE was attenuated in the mildly- and heavily-oxHDL (73% to 77%, P = 0.78 and 159% to 160%, P = 0.91 respectively) (Figure 4.4D, Table 4.2). Similarly, relative to SFA and MUFA containing cholesteryl ester (SFA+MUFA CE), analysis of oxCE showed no significant difference
120
CHAPTER 4 – CHARACTERISATION OF HDL FUNCTION FOLLOWING OXIDATIVE STRESS
in the concentration of total oxCE in the re-isolated oxLDL post-incubation with native HDL, mildly-, or heavily-oxHDL compared to the positive control (Supplementary Figure 4.1A). In addition, a 2.5-fold increase (3.6% to 8.6%, P<0.01) in oxCE was observed in native HDL; whereas no significant change was observed in both mildly- and heavily-oxHDL (25% to 26%, P = 0.90 and 53% to 54%, P = 0.90, respectively) (Supplementary Figure 4.1B).
A B
C D
Figure 4.2 Plasmalogen and oxidised lipids in the re-isolated HDL following co-oxidation with LDL. HDL was co-oxidised with LDL and the lipoproteins were then re-isolated for lipid analyses as described in the Methods. Total levels of (A) alkenylphosphatidylcholine or PC plasmalogen; (B) lysophosphatidylcholine; (C) oxidised PC; and (D) oxidised CE, relative to total levels of PC containing saturated and monounsaturated fatty acids (SFA+MUFA PC). HDL denotes a negative control, which is non-oxidised HDL; oxHDL denotes a positive control, which is HDL that was oxidised without LDL; native HDL, mildly- and heavily ox-HDL were HDL of different oxidative stress levels that were co-oxidised with LDL. Data represents mean ± SEM, n = 3/group, except for HDL (negative control, n = 1). Data (oxHDL, native HDL, mildly oxHDL, and heavily oxHDL) were analysed by one-way ANOVA followed by post-hoc analysis where the ANOVA was significant. Post-hoc significance values are shown; *P<0.05, and **P<0.01 relative to the oxHDL samples unless indicated otherwise.
121
CHAPTER 4 – CHARACTERISATION OF HDL FUNCTION FOLLOWING OXIDATIVE STRESS
A B
C D
Figure 4.3 Plasmalogen and oxidised lipids in the re-isolated LDL following co-oxidation with HDL. LDL was co-oxidised with native HDL, and mildly- and heavily-oxHDL and the lipoproteins were then re-isolated for lipid analyses as described in the Methods. Total levels of (A) alkenylphosphatidylcholine or PC plasmalogen; (B) lysophosphatidylcholine; (C) oxidised PC; and (D) oxidised CE, relative to total levels of PC containing saturated and monounsaturated fatty acids (SFA+MUFA PC). OxLDL denotes LDL which was oxidised without LDL, oxLDL + HDL, oxLDL + mildly oxHDL, and oxLDL + heavily oxHDL denote LDL which was co-oxidised with respective HDL of differing oxidative stress levels, and LDL denotes non-oxidised LDL. Data represents mean ± SEM, n = 3/group, except for non-oxidised LDL (n = 1). Data (oxLDL + HDL, oxLDL + mildly oxHDL, and oxLDL + heavily oxHDL) were analysed by one-way ANOVA followed by post-hoc analysis where the ANOVA was significant. Post-hoc significance values are shown; *P<0.05, **P<0.01, and ***P<0.001 relative to the oxLDL samples unless indicate otherwise.
122
CHAPTER 4 – CHARACTERISATION OF HDL FUNCTION FOLLOWING OXIDATIVE STRESS
4.4.3.3 7-ketocholesterol and 7β-hydroxycholesterol
The levels of oxidised cholesterol (sum of 7-ketocholesterol and 7β-hydroxycholesterol) relative to SFA+MUFA PC were significantly different between the control samples and any of the re-isolated LDL, post-incubation with native, mildly- and heavily oxHDL (ANOVA, P = 0.02). Post-hoc analysis showed that compared to the oxLDL control, native HDL reduced the concentration of LDL-derived oxidised cholesterol (7- ketocholesterol and 7β-hydroxycholesterol) relative to SFA+MUFA PC by 48% (P = 0.91) and 66% (P = 0.04), respectively (Table 4.1). In parallel, the concentration of total oxidised cholesterol (sum of 7-ketocholesterol and 7β-hydroxycholesterol, relative to SFA+MUFA PC) in the re-isolated HDL increased from 0.004% to 2.0% (P<0.01) (Table 4.2). However, this process was attenuated when HDL was oxidised under mild and heavy oxidative conditions; compared to the oxLDL control, changes in oxLDL 7- ketocholesterol of -41% (P = 0.97) and 49% (P = 0.91) and of 7β-hydroxycholesterol of -55% (P = 0.10) and -25% (P = 0.88) post-incubation with mildly- and heavily-oxHDL respectively were observed (Table 4.1). In addition, a smaller increase from 0.4% to 2.0% (P<0.05) and a decrease from 6.3% to 4.7% (P = 0.62) in the total oxidised cholesterol levels (relative to total SFA- and MUFA-PC) in the mildly- and heavily- oxHDL, respectively was observed (Table 4.2). Consistent with these findings, analysis of the oxidised cholesterols relative to non-oxidised cholesterol revealed a reduction in the concentration of LDL-derived 7-ketocholesterol and 7β-hydroxycholesterol by -21% (P = 0.66) and -44% (P = 0.36), respectively post incubation with native HDL compared to oxLDL control; Furthermore, these levels were attenuated in oxLDL post incubation with mildly- and heavily-oxHDL (-8%, P = 0.86 and -19%, P = 0.74, of 7- ketocholesterol; and 28%, P = 0.66 and -11%, P = 0.87 of 7β-hydroxycholesterol), respectively (Supplementary Table 4.2). Correspondingly, the total concentration of oxidised cholesterols in the re-isolated native HDL increased from 0.002% to 0.6% (P<0.001), but this was attenuated in mildly- and heavily-oxHDL (0.2% to 0.7%, P<0.05 and 3% to 2%, P = 0.57) (Supplementary Table 4.3).
4.4.3.4 Transfer of lysophosphatidylcholine from oxLDL to HDL
We examined the profile of lysophosphatidylcholine species containing SFA and MUFA which has previously been shown to be implicated in inflammation and LDL oxidation [200]. The relative level of lysophosphatidylcholine containing SFA and
123
CHAPTER 4 – CHARACTERISATION OF HDL FUNCTION FOLLOWING OXIDATIVE STRESS
MUFA (LPC 16:0) showed a significant difference (ANOVA, P = 0.002) between the controls and any of the LDL samples, post-incubation with native, mildly- or heavily- oxHDL. Post-hoc analysis showed that compared to the non-oxidised LDL, there was a significant increase in the relative level of LPC 16:0 upon oxidation (oxLDL control) (P<0.05), and post-incubation with heavily-oxHDL (P<0.001) (Figure 4.5A). The same was observed in HDL with increasing oxidation levels (Figure 4.5B). Upon incubation of the oxLDL with HDL, this lysophosphatidylcholine species was effectively transferred to the HDL particles. Compared to oxLDL control, we observed a 57% reduction of LPC 16:0 (P = 0.10) in oxLDL when co-incubated with native HDL (Figure 4.5A, Supplementary Table 4.4) and this correlated with an increase of LPC 16:0 (relative to SFA+MUFA PC) in the HDL from 5.8% to 18% (P<0.001) (Figure 4.5B, Supplementary Table 4.5). This effect was diminished with the oxidised forms of HDL; we observed a 31% reduction (P = 0.71) and a 33% increase (P = 0.64) of LPC 16:0 in the re-isolated oxLDL following the co-incubation with mildly- and heavily- oxHDL, respectively compared to oxLDL control (Figure 4.5A, Supplementary Table 4.4). This also correlated with less significant differences in the amount of LPC 16:0 in mildly- and heavily-oxHDL; an increase from 17% to 26% (P = 0.10) and a decrease from 53% to 50% (P = 0.81), respectively (Figure 4.5B, Supplementary Table 4.5). These effects were also observed in other species of lysophosphatidylcholine with SFA and MUFA (Supplementary Table 4.4 and 4.5). In addition, we examined the regioisomers of lysophosphatidylcholine, sn-1 LPC and sn-2 LPC. The levels of sn-1 LPC, but not sn-2 LPC relative to SFA+MUFA PC were significantly different between the re-isolated LDL samples (ANOVA, P<0.01 and P = 0.58 for Figure 4.6A and 4.6C, respectively). Post-hoc analysis showed that compared to the non-oxidised LDL, this relative level of sn-1 LPC in LDL was increased with oxidation (oxLDL control) (P<0.01) but significantly lower after co-incubation with native HDL (P<0.05) (Figure 4.6A). In HDL, the levels of sn-1 LPC were also increased with oxidation and were further elevated after co-incubation with oxLDL (Figure 4.6B). Whereas, the levels of sn-2 LPC in HDL remained the same but was significantly lower (P<0.05) in the heavily oxHDL compared to native HDL. No significant differences were observed in sn-2 LPC following co-incubation with oxLDL (Figure 4.6D).
124
CHAPTER 4 – CHARACTERISATION OF HDL FUNCTION FOLLOWING OXIDATIVE STRESS
A B
C D
Figure 4.4 Effective transfer of oxidised PC, but not oxidised CE from oxidised LDL by HDL. Oxidised PC content in the (A) re-isolated LDL; and (B) re-isolated HDL; and oxidised CE content in (C) the re-isolated LDL; and (D) re-isolated HDL, expressed as mean ± SD (n = 3/sample) of % of total PC containing saturated and monounsaturated fatty acids (SFA+MUFA PC). In (A) and (C), LDL denotes non-oxidised LDL; oxLDL denotes oxidised LDL which was not incubated with HDL (i.e. oxLDL control); oxLDL + nHDL denotes oxidised LDL which was co-incubated with native HDL; oxLDL + mildly oxHDL denotes oxidised LDL which was co-incubated with mildly oxidised HDL; and oxLDL + heavily oxHDL denotes oxidised LDL which was co-incubated with heavily oxidised HDL. In (B) and (D), open bars indicate HDL which was incubated without oxLDL; closed bars indicate HDL which was incubated with oxLDL. Data in panel (A) and (C) were analysed by one-way ANOVA followed by post-hoc analysis where the ANOVA was significant. Whereas data in for panel (B) and (D) were analysed by Student t-test. Post-hoc and Student t-test significance values are shown; *P<0.05, **P<0.01, and ***P<0.001, relative to oxLDL control for panel (A) and (C) and relative to the corresponding HDL control samples which were not incubated with oxLDL for panel (B) and (D), unless indicated otherwise.
125
CHAPTER 4 – CHARACTERISATION OF HDL FUNCTION FOLLOWING OXIDATIVE STRESS
Table 4.1 Changes in the level of oxidised lipids in the re-isolated LDL after co-incubation.
Assigned lipid name LDL + nHDLa LDL + mildly oxHDLb LDL + heavily oxHDLc % change d P-value e % change d P-value e P-value % change d P-value e P-value (relative to (relative to nHDL) f nHDL) f Oxidised PC PC (34:3(O)) -69 0.019 -46 0.184 0.877 22 0.911 0.003 PC (34:2(O)) -65 0.021 -38 0.310 0.718 24 0.830 0.002 PC (34:3(O2)) -84 0.942 -63 0.991 1.00 158 0.355 0.052 PC (34:2(O2)) -72 0.002 -44 0.057 0.398 3 1.00 0.001 PC (36:3(O)) -63 0.008 -45 0.072 0.902 7 1.00 0.004 PC (36:2(O)) -48 0.053 -26 0.619 0.742 13 0.990 0.012 PC (36:3(O2)) -70 0.151 -48 0.542 0.993 52 0.446 0.005 Total oxPC -67 0.021 -42 0.231 0.883 27 0.732 0.002 Oxidised CE CE (16:0(O)) -5 1.00 -7 1.00 1.00 14 1.00 1.00 CE (16:1(O)) -14 1.00 -11 1.00 1.00 2 1.00 1.00 CE (18:1(O)) 3 1.00 3 1.00 1.00 19 1.00 1.00 CE (18:2(O)) 10 1.00 5 1.00 1.00 13 1.00 1.00 CE (18:2(O2)) -15 0.978 -3 1.00 0.996 -20 0.874 1.00 CE (18:3(O)) -2 1.00 -3 1.00 1.00 2 1.00 1.00 CE (18:3(O2)) -22 0.993 -18 0.998 1.00 -12 1.00 1.00 Total oxCE -2 1.00 0 1.00 1.00 1 1.00 1.00 Oxidised cholesterol 7-ketocholesterol -48 0.913 -41 0.969 1.00 49 0.905 0.214 7-β hydroxycholesterol -66 0.039 -55 0.103 1.00 -25 0.877 0.360 Total oxidised -49 0.897 -41 0.962 1.00 47 0.915 0.210 cholesterol a LDL + nHDL - oxidised LDL which was co-incubated with native HDL; b LDL + mildly oxHDL - oxidised LDL which was co-incubated with mildly oxidised HDL; c LDL + heavily oxHDL - oxidised LDL which was co-incubated with heavily oxidised HDL. d Change (%) with reference to oxLDL (oxLDL which was not incubated with HDL); e The sample groups (LDL with and without co-incubation with native, mildly oxidised and heavily oxidised HDL) were analysed using one-way ANOVA followed by post-hoc analysis where the ANOVA was significant. Post-hoc significance values are shown and those in bold indicate statistical significance (P<0.05), n = 3/sample. f Indicates the significance of the difference between the level of oxidised lipid remaining in the oxLDL compared to the level remaining following treatment with native HDL; Post-hoc significance values in bold indicate statistical significance (P<0.05), n = 3/sample.
126
CHAPTER 4 – CHARACTERISATION OF HDL FUNCTION FOLLOWING OXIDATIVE STRESS
Table 4.2 Changes in the levels of oxidised lipids in re-isolated HDL with or without co-incubation with oxLDL.
Native HDL Mildly oxHDL Heavily oxHDL without with without with without with P- P- incubation incubation incubation incubation P-valuec incubation incubation valuec valuec (%)a (%)b (%)a (%)b (%)a (%)b Oxidised PC PC (34:3(O)) 0.060 ± 0.003 1.31 ± 0.157 0.000 1.61 ± 1.02 2.46 ± 0.781 0.319 5.50 ±1.13 5.07 ± 0.686 0.604 PC (34:2(O)) 0.950 ± 0.191 4.57 ± 0.320 0.000 6.64 ± 0.952 8.25 ± 0.865 0.096 10.8 ± 1.72 10.5 ± 1.10 0.778 PC (34:3(O2)) 0.020 ± 0.015 0.430 ± 0.083 0.001 0.550 ± 0.380 0.930 ± 0.296 0.251 5.73 ± 4.03 4.59 ± 2.33 0.695 PC (34:2(O2)) 0.360 ± 0.068 1.52 ± 0.320 0.003 4.50 ± 1.21 7.49 ± 0.760 0.005 12.7 ± 4.27 10.8 ± 3.55 0.023 PC (36:3(O)) 0.511 ± 0.084 1.29 ± 0.432 0.013 1.91 ± 0.899 2.48 ± 0.863 0.434 4.17 ± 0.659 3.58 ± 1.23 0.302 PC (36:2(O)) 0.641 ± 0.120 1.44 ± 0.102 0.001 2.66 ± 0.324 2.95 ± 0.343 0.355 4.16 ± 0.556 4.02 ± 0.476 0.764 PC (36:3(O2)) 0.112 ± 0.032 0.473 ± 0.080 0.002 1.16 ± 0.266 1.27 ± 0.169 0.566 4.07 ± 1.99 3.44 ± 1.25 0.656 Total oxPC 2.65 ± 0.325 11.0 ± 1.16 0.000 19.0 ± 5.00 22.9 ± 3.84 0.345 47.1 ± 12.2 41.9 ± 7.44 0.562 Oxidised CE CE (16:0(O)) 0.009 ± 0.015 0.153 ± 0.046 0.006 0.229 ± 0.091 0.264 ± 0.100 0.616 0.822 ± 0.181 0.856 ± 0.210 0.837 CE (16:1(O)) 0.023 ± 0.020 0.214 ± 0.090 0.023 0.340 ± 0.128 0.487 ± 0.150 0.265 1.34 ±0.334 1.39 ± 0.360 0.866 CE (18:1(O)) 0.524 ± 0.034 2.67 ± 0.976 0.018 5.58 ± 1.83 6.69 ± 1.96 0.510 18.3 ± 4.54 19.0 ± 5.42 0.887 CE (18:2(O)) 8.94 ± 1.00 16.0 ± 3.44 0.018 45.9 ± 5.38 45.7 ± 7.24 0.982 57.3 ± 22.0 59.3 ± 18.6 0.772 CE (18:2(O2)) 0.508 ± 0.113 3.53 ± 1.69 0.032 11.3 ± 3.33 12.6 ± 4.61 0.699 37.6 ± 6.00 36.6 ± 9.10 0.868 CE (18:3(O)) 0.181 ± 0.045 1.54 ± 0.507 0.010 6.52 ± 3.26 6.72 ± 3.34 0.945 12.1 ± 3.17 12.6 ± 1.88 0.837 CE (18:3(O2)) 0.046 ± 0.041 1.97 ± 0.945 0.024 2.63 ± 2.10 4.22 ± 3.23 0.511 31.2 ± 12.4 30.7 ± 13.2 0.964 Total oxCE 10.2 ± 0.679 26.1 ± 6.70 0.017 72.5 ± 15.0 76.8 ± 20.1 0.783 159 ± 14.9 160 ± 21.5 0.915 Oxidised cholesterol 7-ketocholesterol 0.00 ± 0.00 2.00 ± 0.439 0.001 0.392 ± 0.382 1.93 ± 0.809 0.041 6.24 ± 4.74 4.62 ± 2.07 0.616 7-β 0.004 ± 0.001 0.031 ± 0.014 0.033 0.020 ± 0.013 0.066 ± 0.061 0.266 0.074 ± 0.049 0.076 ± 0.051 0.951 hydroxycholesterol Total oxidised 0.004 ± 0.001 2.03 ± 0.435 0.001 0.412 ± 0.390 1.99 ± 0.867 0.045 6.32 ± 4.79 4.70 ± 2.03 0.619 cholesterol a Percentage level of oxidised lipids relative to total SFA+MUFA PC in re-isolated HDL which was incubated alone. Data is expressed as mean ± SD, n = 3/sample. b Percentage level of oxidised lipids relative to total SFA+MUFA PC in re-isolated HDL which was incubated with oxLDL. . Data is expressed as mean ± SD, n = 3/sample. c The sample groups (HDL with and without co-incubation with oxLDL) were analysed using Student t-test; values in bold indicate statistical significance (P<0.05).
127
CHAPTER 4 – CHARACTERISATION OF HDL FUNCTION FOLLOWING OXIDATIVE STRESS
4.4.3.5 Sphingomyelin to phosphatidylcholine ratio in native and oxidised HDL
The ability to transfer lipid hydroperoxides has been shown to be dependent on the ratio of sphingomyelin to phosphatidylcholine (SM/PC) that contributes to the surface rigidity of phospholipid monolayer of the acceptor particle [33]. Therefore, we examined the total level of sphingomyelin (Supplementary Table 4.6) relative to total phosphatidylcholine of native and oxidised HDL to investigate whether the difference in the ratio could have affected the oxidised lipid and lysophosphatidylcholine transfer activity of HDL. The ratio of SM/PC increased with increasing oxidative conditions of the HDL (at least P<0.05 for all comparisons among the HDL samples which were not incubated with oxLDL) (Figure 4.7), with decreasing ability to accept oxidised lipids (Figure 4.4, Table 4.2). The ratio of SM/PC was further elevated in the re-isolated native HDL post-incubation with oxLDL, but there was no significant difference in the ratio of SM/PC in the mildly- and heavily-oxHDL as compared to the corresponding samples which were incubated without oxLDL (Figure 4.7).
4.4.3.6 Phosphatidylserine in the native and oxidised HDL
Phosphatidylserine is a negatively charged lipid which has been shown to induce a conformational change of apoAI and enable its interaction with phospholipid bilayers [370]. Thus, we analysed the concentration of total phosphatidylserine (Supplementary Table 4.7) to investigate whether the difference in the phosphatidylserine level could have affected the oxidised lipid and lysophosphatidylcholine transfer capacity of HDL.
In HDL, the amount of phosphatidylserine relative to total PC was progressively decreased upon more severe oxidative conditions. There was no significant difference in the level of phosphatidylserine relative to total PC in the samples which were incubated alone and incubated with oxLDL (Figure 4.8).
128
CHAPTER 4 – CHARACTERISATION OF HDL FUNCTION FOLLOWING OXIDATIVE STRESS
A B
Figure 4.5 Transfer of lysophosphatidylcholine from oxLDL to native and oxHDL. Level of LPC 16:0 in (A) re-isolated LDL; and (B) re-isolated HDL. Data is expressed as mean ± SD (n = 3/sample) of % of total PC containing saturated and monounsaturated fatty acids (SFA+MUFA PC). LDL denotes non-oxidised LDL; oxLDL denotes oxidised LDL which was not incubated with HDL (i.e. oxLDL control); OxLDL + nHDL denotes oxidised LDL which was co-incubated with native HDL; oxLDL + mildly oxHDL denotes oxidised LDL which was co-incubated with mildly oxidised HDL; and oxLDL + heavily oxHDL denotes oxidised LDL which was co-incubated with heavily oxidised HDL. Data in panel (A) were analysed using one-way ANOVA followed by post-hoc analysis where the ANOVA was significant. Whereas data in panel (B) were analysed by Student t- test. Post-hoc and Student t-test significance values are shown; *P<0.05, **P<0.01, and ***P<0.001 relative to the oxLDL control sample for panel (A) and relative to the corresponding HDL control samples which were not incubated with oxLDL for panel (B), unless indicated otherwise.
129
CHAPTER 4 – CHARACTERISATION OF HDL FUNCTION FOLLOWING OXIDATIVE STRESS
A B
C D
Figure 4.6 Levels of isomers of lysophosphatidylcholine in LDL and HDL following co- incubations. Sn-1 LPC (top panels) in the re-isolated (A) LDL; and (B) HDL and sn-2 LPC (bottom panels) in the re-isolated (C) LDL; and (D) HDL. Data is expressed as mean ± SD (n = 3/sample) of % of total PC containing saturated and monounsaturated fatty acids (SFA+MUFA PC). LDL denotes non-oxidised LDL; oxLDL denotes oxidised LDL which was not incubated with HDL (i.e. oxLDL control); OxLDL + nHDL denotes oxidised LDL which was co-incubated with native HDL; oxLDL + mildly oxHDL denotes oxidised LDL which was co-incubated with mildly oxidised HDL; and oxLDL + heavily oxHDL denotes oxidised LDL which was co-incubated with heavily oxidised HDL. Data in panel (A) and (C) were analysed using one-way ANOVA followed by post-hoc analysis where the ANOVA was significant. Whereas data in panel (B) and (D) were analysed by Student t-test. Post-hoc and Student t-test significance values are shown; *P<0.05, **P<0.01, and ***P<0.001 relative to the oxLDL control sample for panel (A) and (C) and relative to the corresponding HDL control samples which were not incubated with oxLDL for panel (B) and (D), unless indicated otherwise.
130
CHAPTER 4 – CHARACTERISATION OF HDL FUNCTION FOLLOWING OXIDATIVE STRESS
Figure 4.7 Level of sphingomyelin relative to total phosphatidylcholine as a contributing factor to the ability of HDL to transfer oxidised lipids. Open bars indicate HDL which was incubated without oxLDL. Closed bars indicate HDL which was incubated with oxLDL. Data is expressed as mean ± SD (n = 3/sample) of the ratio of total sphingomyelin to total phosphatidylcholine (SM/PC) in native, and mildly- and heavily oxidised HDL. Data were analysed using Student t-test; *P<0.05 and **P<0.01 relative to corresponding HDL samples which were not incubated with oxLDL, unless indicated otherwise.
Figure 4.8 Phosphatidylserine in the native and oxidised HDL. Re-isolated HDL contents of total phosphatidylserine, expressed as mean ± SD (n = 3/sample) of % of total PC. Open bars indicate HDL which was incubated without oxLDL. Closed bars indicate HDL which was incubated with oxLDL. Data were analysed using Student t-test; **P<0.01, and ***P<0.001 relative to corresponding HDL control sample which were not incubated with oxLDL, unless indicated otherwise.
131
CHAPTER 4 – CHARACTERISATION OF HDL FUNCTION FOLLOWING OXIDATIVE STRESS
4.5 DISCUSSION
HDL possesses many atheroprotective functions including the ability to promote cholesterol efflux from cells and anti-oxidative properties. In this chapter, we investigated if oxidation of HDL particles affected these functions. We found that mild to strong oxidation attenuated the capacity of HDL to efflux cholesterol, to prevent LDL oxidation, and to efficiently accept oxidised lipids from oxLDL.
Our studies on the ability of HDL to efflux cholesterol showed that with increasing levels of oxidative stress, the ability of HDL to efflux cholesterol was further impaired. This finding was consistent with previous reports whereby the ability of oxHDL to efflux cholesterol from foam cells [355] and fibroblasts [371] were impaired. Our finding was also consistent with another report, which demonstrated that the oxidation of HDL by myeloperoxidase attenuates ABCA1-dependent cholesterol efflux from macrophages [372]. In our study, we oxidised the HDL heavily for 120 to 240 min with 1 μmol/l copper chloride. While it is not likely that this level of oxidative stress is observed in circulation, this level may be attainable in atherosclerotic plaques [373]. Our study, together with previous reports by Zheng et al. [372] and Nishi et al. [373] used immortalised cell-derived macrophages in the experiments. Macrophages are found in atherosclerotic plaques where they engulf oxLDL to form foam cells. Thus, in this aspect, our study is physiologically relevant as HDL can interact with macrophages to efflux the cholesterols and it is possible for HDL to be oxidised heavily when they enter the intima layer [374]. The findings from our study suggest that high levels of oxidative stress to HDL lead to its impaired ability to efflux cholesterol. In the context of the vasculature, such high levels of oxidative stress along with reduced cholesterol efflux ability of HDL can contribute to atherosclerosis.
Prior to investigating changes in the anti-oxidative function of HDL, we first confirmed if the oxidation has led to differences in the lipid composition of the lipoprotein as observed previously in Chapter 3. Indeed, lower levels of alkenylphosphatidylcholine or PC plasmalogen in both LDL and HDL were observed, regardless of the level of oxidation. This supports the proposition that plasmalogen is susceptible to oxidation. Lysophosphatidylcholine can be produced from different sources, including the action of Lp-PLA2, an independent predictor of CAD [199, 340], on the sn-2 position of the glycerol backbone of oxidised phospholipids, resulting in sn-1 LPC. These sn-1 LPC
132
CHAPTER 4 – CHARACTERISATION OF HDL FUNCTION FOLLOWING OXIDATIVE STRESS
mainly contain SFA or MUFA, that are produced from truncated oxidised phosphatidylcholine; which subsequently become preferred substrates for Lp-PLA2 hydrolysis [200]. Non-oxidised phosphatidylcholine is less susceptible to Lp-PLA2 hydrolysis [375-377]. Another source of lysophosphatidylcholine is from the oxidation of plasmalogen followed by the cleavage of the vinyl ether bond at the sn-1 position of the glycerol backbone resulting in sn-2 LPC.
In this study, we observed an increase in the total level of lysophosphatidylcholine in oxHDL that is consistent with the elevated level of lysophosphatidylcholine, previously reported in atherosclerotic lesions, along with increased expression of Lp-PLA2 [197]. This finding is also consistent with the increase in circulating lysophosphatidylcholine and lysophosphatidylethanolamine observed in CAD patients [254] and patients with mild atherosclerosis [320]. Lysophosphatidylcholine is known to be a signaling molecule that can contribute to inflammation; it was demonstrated that lysophosphatidylcholine increased the production of MCP-1 in HUVEC cells [378] and subsequently attracts monocyte adhesion by chemokine release in the endothelial cells [379]. The decrease in the total level of lysophosphatidylcholine in LDL when the lipoprotein was co-oxidised with HDL suggests a protection by HDL. As non-oxidised phospholipids are less susceptible to hydrolysis by Lp-PLA2 [375-377], the reduction in lysophosphatidylcholine could also mean that less phospholipids were oxidised to become substrates for Lp-PLA2 and that less plasmalogens were oxidised. Important implications of this reduction in lysophosphatidylcholine in LDL may include less inflammation and monocyte adhesion, which are hallmarks of atherosclerosis. The protection of LDL against oxidation by HDL was further supported as oxidation to HDL resulted in an increase in lysophosphatidylcholine level in LDL, hence a loss of protection (Figure 4.3B).
To investigate the effect of oxidation on the HDL capacity to protect LDL against oxidation, we first measured the levels of oxidised lipids including oxPC and oxCE species in oxHDL. As we expected, the total level of oxPC was increased in HDL with increasing levels of oxidative stress. Interestingly, the level of oxPC in the re-isolated native HDL was significantly lower though it was oxidised with the same amount of time with oxHDL control (i.e. HDL oxidised without LDL). This finding suggests that LDL might offer some protection to HDL during oxidation. It is also possible that both LDL and HDL present as substrates for the free radical chain reaction. As a result, the
133
CHAPTER 4 – CHARACTERISATION OF HDL FUNCTION FOLLOWING OXIDATIVE STRESS
effect of oxidation was shared between the two lipoproteins and the total oxidation products for each lipoprotein were reduced. The same trend was also observed with oxCE in HDL, though the level was not significantly different to that of the control. This was possibly due to the location of cholesteryl ester in the core of the lipoproteins, therefore, they were less affected by oxidation. The production of oxPC and oxCE in LDL depended on the oxidative state of HDL and this is an important implication in the settings of atherosclerosis where high level of oxidative stress is observed in plaques. LDL that are trapped in the intima layer is likely to be oxidised because of a micro- environment of free radicals and as we observed, the delay of LDL oxidation can be achieved only if HDL is not modified by oxidation. The mechanisms to which HDL could protect LDL from oxidation and delay the progression of atherosclerosis were not fully elucidated. However, it was proposed that HDL could remove “oxidation seeding molecules” such as lipid hydroperoxides from LDL and so it renders LDL more resistant to oxidation [380]. In addition, HDL was shown to remove 7-ketocholesterol from oxLDL or macrophages via ABCG1 transporters [381]. This ability of HDL to remove oxidised lipids including 7-ketocholesterols led us to the next investigation where we looked at the impact of oxidation on the ability of HDL to accept oxidised lipids from oxidised LDL. In this study, we optimised the measurements for oxysterols including 7-ketocholesterol and 7β-hydroxycholesterol in our mass spectrometry methodology.
Using the oxidised lipids as markers, we demonstrated the ability of native HDL and oxHDL to accept oxidised lipids including oxPC and oxidised cholesterol (7- ketocholesterol and 7β-hydroxycholesterol) from oxLDL particles. This ability of HDL to act as an acceptor of oxPC and oxidised cholesterol is not limited to oxLDL particles; previous studies by Vila et al. [382, 383] demonstrated that phospholipid and cholesterol-derived hydroperoxides could be transferred spontaneously between cell membranes and LDL, while Terasaka et al. [381] showed that 7-ketocholesterol, but not cholesterol was exported to HDL from ABCG1-transfected 293 cells.
Greenberg et al. [384] demonstrated that oxidative truncation to the sn-2 fatty acyl chain of phospholipids resulted in the re-orientation of the lipid in the membrane and its protrusion into the aqueous phase. It would be expected that this would effectively lower the free energy of activation for transfer between lipid surfaces thereby increasing the rate of transfer. Whilst the oxPC species measured in this study were not truncated,
134
CHAPTER 4 – CHARACTERISATION OF HDL FUNCTION FOLLOWING OXIDATIVE STRESS
the addition of the oxygen would increase polarity potentially leading to a re-orientation within the surface lipid layer. This re-orientation may modulate lipid phase rigidity in the oxLDL particles resulting in a decrease in the free energy of activation and so facilitate the removal of the lipids by HDL when the lipoproteins interact. Lund-Katz and Phillips [385] showed that the half-time of cholesterol exchange from human HDL to LDL at 37oC was 2.9 min whereas the half-time for dipalmitoyl-phosphatidylcholine was 5 h. The large difference between these two molecules presumably relates to their free energy of activation, a function of size and polarity. Nonetheless the time scale is comparable to our experimental scale, particularly if the half-time of the oxPC exchange is decreased relative to non-oxidised PC as expected.
In contrast to the efficient transfer of oxPC and oxidised cholesterol from oxLDL to HDL, we observed that oxCE species were transferred inefficiently. The same findings were obtained when we analysed the levels of oxCE and oxidised cholesterols relative to SFA- and MUFA-CE (Supplementary Figure 4.1) and non-oxidised free cholesterol (Supplementary Table 4.2 and 4.3), respectively, demonstrating that the amount of oxCE transferred was low relative to other major lipid constituents of the lipoprotein particles.
The larger size and lower polarity of the cholesteryl ester compared to the phospholipids and cholesterol, result in the cholesteryl ester being primarily located within the hydrophobic core of the lipoprotein particles [386, 387]. Oxidation of the cholesteryl ester increases polarity, potentially driving the oxCE into the surface lipid layer, and lowers the free energy of activation, as previously demonstrated by the preferential transfer of cholesteryl ester hydroperoxides from HDL to Hep G2 cells relative to non- oxidised cholesteryl ester [388]. However, in this system oxidation of cholesteryl ester was insufficient to cause bulk movement of the oxCE from the LDL to the HDL particles.
Oxidation of both LDL and HDL led to an increase in lysophosphatidylcholine species containing SFA and MUFA. In contrast, no significant differences were observed in the level of lysophosphatidylcholine containing PUFA in oxLDL, HDL and oxHDL, following co-incubation (data not shown). Therefore, we focused our analyses on lysophosphatidylcholine species containing SFA and MUFA. LPC 16:0 is the most abundant species in LDL and following co-incubation with HDL, it showed similar
135
CHAPTER 4 – CHARACTERISATION OF HDL FUNCTION FOLLOWING OXIDATIVE STRESS
trends with other SFA and MUFA containing LPC species (Supplementary Table 4.3 and 4.4). Therefore, we used LPC 16:0 as a representative species to demonstrate the changes of lysophosphatidylcholine in the lipoproteins following co-incubation. As mentioned above, oxPC are known to be preferred substrates for Lp-PLA2 which cleaves the sn-2 fatty acid, typically the site of PUFA, to yield the corresponding lysophosphatidylcholine species. A significant decrease in LPC 16:0 in LDL following incubation with HDL and correspondingly an elevation of the lipid in the native HDL suggest the net transfer of lysophosphatidylcholine from oxLDL to HDL. Previous studies have demonstrated that the predominant lysophosphatidylcholine regioisomer in human plasma is sn-1 LPC which contains MUFA and PUFA; whereas sn-2 LPC contains predominantly of only PUFA [347, 348]. To confirm the finding above and to investigate which regioisomer was predominantly involved in the lipid transfer process, we analysed the levels of isomers (sn-1 and sn-2 LPC) of total lysophosphatidylcholine. We demonstrated that the trends in sn-1 LPC in both LDL and HDL are very similar to that observed with LPC 16:0 (Figure 4.5). On the other hand, no differences were observed in the levels of sn-2 LPC between the LDL samples (Figure 4.6C) and the re- isolated HDL samples (Figure 4.6D). These findings suggest that the changes in total lysophosphatidylcholine result primarily from changes in sn-1 LPC, not sn-2 LPC and so most likely result from oxidation and subsequent Lp-PLA2 action on phosphatidylcholine.
In this study we observed a smaller reduction or a non-significant increase in oxidised lipids and lysophosphatidylcholine in oxLDL upon co-incubation with more heavily- oxHDL. This can be attributed to the higher initial concentration of the oxidised lipids and lysophosphatidylcholine in the oxHDL particles, thus limiting their capacity to uptake more oxidised lipids. It is possible that the transfer of oxidised lipids and lysophosphatidylcholine can also proceed from the HDL to the LDL in situations where the oxidation of HDL is greater than the oxidation of LDL. We may also postulate that the surface lipids of both LDL and HDL reach equilibrium when they were co- incubated, thus contributing to the smaller difference observed between the particles. However, further studies using stable isotope labeled lipid species will be required to confirm this. The mechanistic details of the transfer of oxidised lipids from oxLDL to HDL are yet to be elucidated. We speculate that lipidomic components of HDL play a role in the transfer of the oxidised lipids and lysophosphatidylcholine. A previous study
136
CHAPTER 4 – CHARACTERISATION OF HDL FUNCTION FOLLOWING OXIDATIVE STRESS
using HPLC chemiluminescence and UV detection to measure lipid hydroperoxides have demonstrated that the transfer of a few molecular species of phosphatidylcholine hydroperoxides can be influenced by the availability of lipid-transfer protein (cholesteryl ester transfer protein) and the surface rigidity of the acceptor particle; a low ratio of SM/PC reduced surface rigidity and aided in the transfer efficiency of oxidised lipids, and in the delay of LDL oxidation [33, 168]. Once transferred the lipid hydroperoxides were subsequently reduced to their respective hydroxides by HDL- associated apoAI as first described by Garner et al. [350]. This process was governed by the total HDL content of apoAI and the redox status of the methionine residues of apoAI [33]. In light of this earlier study, we examined the level of sphingomyelin relative to phosphatidylcholine in our HDL acceptor particles. We demonstrated that the SM/PC ratio was increased in oxHDL due to a reduction in the content of the PUFA- containing phosphatidylcholine (data not shown), thus suggesting a possible overall increase in the surface rigidity of the HDL upon oxidation. Consistent with our finding, native HDL with the lower ratio of SM/PC most effectively accepted oxidised lipids and lysophosphatidylcholine from oxLDL as compared to mildly- and heavily-oxHDL.
Phosphatidylserine is a negatively charged lipid which has been reported to be enriched in HDL3c particles [227]. It induces a conformational change of apoAI [370], and facilitates the electrostatic interaction of the protein with polar phospholipids, allowing the penetration of the protein into the phospholipid monolayer [227, 389], of other lipoproteins. We observed a decrease in the phosphatidylserine content in HDL with increasing oxidative conditions. The high level of PUFA in phosphatidylserine relative to phosphatidylcholine makes it more susceptible to oxidation and thus is likely to contribute to the observed decrease in the phosphatidylserine to phosphatidylcholine ratio. In light of the previous studies, our findings suggest that a decrease in the phosphatidylserine content may lead to a modulation of the capacity of HDL-associated apoAI to penetrate and interact with the polar oxidised lipids in oxLDL, as well as a general shift away from features of HDL3c which was associated with the greatest potency of anti-oxidative capacity. This notion is supported by our findings where the capacities of oxHDL to transfer and uptake oxidised lipids and lysophosphatidylcholine were impaired. With this study, we conclude that surface oxidised lipids (oxPC and oxidised cholesterol) are readily transferred from oxLDL to HDL whereas lipids located in the hydrophobic core of oxLDL (oxCE) are less readily transferred. This has
137
CHAPTER 4 – CHARACTERISATION OF HDL FUNCTION FOLLOWING OXIDATIVE STRESS
important implication for the role of HDL in the prevention and/or reversal of atherosclerosis where oxCE makes up a major component of the lipid core of the atherosclerotic plaque. Also, in agreement with previous reports, the composition of the HDL particle (SM/PC and phosphatidylserine content) appears to influence the antioxidant capacity.
In conclusion, our study has demonstrated that mild to strong oxidation to HDL affected the lipoprotein function including its ability to efflux cholesterol from macrophages, the ability to delay LDL oxidation and to efficiently transfer surface oxidised lipids (oxPC and oxidised cholesterol) from oxLDL to HDL. The findings are opening the way for lipidomic assessment of HDL to quantify atheroprotective functionality which is currently not reflected in routine clinical measurement of circulating HDL-cholesterol levels. Our findings may also have relevance for the development of new HDL therapeutics where the lipid composition of such formulations may influence functionality and thereby efficacy.
4.6 LIMITATIONS
The analysis of exclusively in vitro oxidised LDL and HDL in this study limits the direct relevance of the oxidised lipids (oxPC, oxCE, 7-ketocholesterol, and 7β- hydroxycholesterol) occurring in vivo. As it is beyond the scope of this thesis, we have not reported the measurement of the oxidised lipids in vivo. In vivo measurement of the oxysterols has been demonstrated previously in plasma of rabbit models of atherosclerosis [352]. Our laboratory has been able to use this methodology to measure the oxidised lipids in plasma samples from patient cohorts.
Our current lipidomic methodology is limited to accurately discriminate between lipids that were derived from LDL or HDL. Therefore in our study, we have inferred the efficient transfer of oxidised lipids by HDL from the measurement of differences in these lipids in the lipoproteins before and after co-incubation and re-isolation. To accurately discriminate the origins and monitor the movements of these oxidised lipids, stable isotopes can be used to tag the lipids. However, the methodology is complex and the development of this method is time consuming as it would require tagging the phosphatidylcholine and cholesteryl ester species individually and subsequently oxidise them and characterise the oxidised lipids. To our knowledge, no lipidomic laboratories have successfully established or employed an LC-MS methodology for oxPC or oxCE
138
CHAPTER 4 – CHARACTERISATION OF HDL FUNCTION FOLLOWING OXIDATIVE STRESS
measurement; their methods are often limited to tagging a particular metabolite of interest such as glucose [390] and palmitate [391].
The ratio of SM/PC was previously used as an indicator of surface rigidity of the phospholipid monolayer of HDL subclasses [33]. In this study, we have adopted the method to gauge differences in the surface rigidity of the HDL due to oxidation. Phosphatidylcholine containing PUFA is more susceptible to oxidation as compared to phosphatidylcholine containing SFA and MUFA and therefore, the measurement of sphingomyelin relative to total phosphatidylcholine is limited because it may not give an accurate indication of the surface rigidity of oxidised HDL. A specialised technique using atomic force microscopy can provide a better measurement of the lipoprotein rigidity. However, this technique is not commonly used/available [392].
139
CHAPTER 5 – PLASMALOGEN MODULATION ATTENUATES ATHEROSCLEROSIS IN APOE- AND APOE/GPX1- DEFICIENT MICE
140
Atherosclerosis 243 (2015) 598e608
Contents lists available at ScienceDirect
Atherosclerosis
journal homepage: www.elsevier.com/locate/atherosclerosis
Plasmalogen modulation attenuates atherosclerosis in ApoE- and ApoE/GPx1-deficient mice
Aliki A. Rasmiena a, b, Christopher K. Barlow a, Nada Stefanovic a, Kevin Huynh a, ** * Ricardo Tan a, Arpeeta Sharma a, Dedreia Tull c, Judy B. de Haan a, , Peter J. Meikle a, b, a Baker IDI Heart and Diabetes Institute, Melbourne, VIC, Australia b Department of Biochemistry and Molecular Biology, The University of Melbourne, Parkville, Australia c Metabolomics Australia, Bio21 Institute, Parkville, Australia article info abstract
Article history: Background and aim: We previously reported a negative association of circulating plasmalogens (phos- Received 8 July 2015 pholipids with proposed atheroprotective properties) with coronary artery disease. Plasmalogen mod- Received in revised form ulation was previously demonstrated in animals but its effect on atherosclerosis was unknown. We 21 October 2015 assessed the effect of plasmalogen enrichment on atherosclerosis of murine models with differing levels Accepted 22 October 2015 of oxidative stress. Available online 26 October 2015 Methods and results: Six-week old ApoE- and ApoE/glutathione peroxidase-1 (GPx1)-deficient mice were fed a high-fat diet with/without 2% batyl alcohol (precursor to plasmalogen synthesis) for 12 weeks. Keywords: Batyl alcohol Mass spectrometry analysis of lipids showed that batyl alcohol supplementation to ApoE- and ApoE/ fi Plasmalogen GPx1-de cient mice increased the total plasmalogen levels in both plasma and heart. Oxidation of Anti-oxidant plasmalogen in the treated mice was evident from increased level of plasmalogen oxidative by-product, Alkenylphospholipid sn-2 lysophospholipids. Atherosclerotic plaque in the aorta was reduced by 70% (P ¼ 5.69E-07) and 69% Alkylphospholipid (P ¼ 2.00E-04) in treated ApoE- and ApoE/GPx1-deficient mice, respectively. A 40% reduction in plaque Lipidomics (P ¼ 7.74E-03) was also seen in the aortic sinus of only the treated ApoE/GPx1-deficient mice. Only the Atherosclerosis treated ApoE/GPx1-deficient mice showed a decrease in VCAM-1 staining ( 28%, P ¼ 2.43E-02) in the aortic sinus and nitrotyrosine staining ( 78%, P ¼ 5.11E-06) in the aorta. Conclusion: Plasmalogen enrichment via batyl alcohol supplementation attenuated atherosclerosis in ApoE- and ApoE/GPx1-deficient mice, with a greater effect in the latter group. Plasmalogen enrichment may represent a viable therapeutic strategy to prevent atherosclerosis and reduce cardiovascular disease risk, particularly under conditions of elevated oxidative stress and inflammation. © 2015 Elsevier Ireland Ltd. All rights reserved.
1. Introduction
Oxidative stress is a contributing factor to the progression of Nonstandard abbreviations and acronyms: 4-HNE, 4-hydoxynonenal; ApoE / , atherosclerosis. Glutathione peroxidase-1 is an anti-oxidant apolipoprotein E-deficient mouse; ApoE / GPx1 / , apolipoprotein E/glutathione peroxidase-1- deficient mouse; BA, batyl alcohol; CE, cholesteryl ester; LPC, lyso- enzyme which is expressed ubiquitously in mammalian cells; it phophatidylcholine; LPE, lysophoshatidylethanolamine; PC, phosphatidylcholine; detoxifies hydrogen peroxide, lipid hydroperoxide, and peroxyni- PC(O), alkylphosphatidylcholine; PC(P), phosphatidylcholine plasmalogen/alkenyl- trite [1].Deficiency of glutathione peroxidase-1 in Apolipoprotein phosphatidylcholine; PE(O), alkylphosphatidylethanolamine; PE(P), phosphatidyl- E-deficient mice (ApoE / GPx1 / ) was demonstrated to result in a ethanolamine plasmalogen/alkenylphosphatidylethanolamine; ROS, reactive significant increase in atherosclerosis after 24 weeks of high-fat oxygen species; SOD, superoxide dismutase; VCAM-1, vascular cellular adhesion molecule-1. (21% fat, 0.15% cholesterol) feeding as compared to mice which / * Corresponding author. Baker IDI Heart and Diabetes Institute, 75 Commercial were deficient in Apolipoprotein E only (ApoE ). This increase in Road, Melbourne, VIC 3004, Australia. atherosclerotic plaques was accompanied by an elevation in su- ** Corresponding author. Baker IDI Heart and Diabetes Institute, 75 Commercial peroxide formation and protein nitration in the aorta [2] as well as Road, Melbourne, VIC 3004, Australia. an increase in the expression of the pro-inflammatory markers, E-mail addresses: [email protected] (J.B. de Haan), peter.meikle@ bakeridi.edu.au (P.J. Meikle). vascular cellular adhesion molecule-1 (VCAM-1) and receptor for http://dx.doi.org/10.1016/j.atherosclerosis.2015.10.096 0021-9150/© 2015 Elsevier Ireland Ltd. All rights reserved. A.A. Rasmiena et al. / Atherosclerosis 243 (2015) 598e608 599 advanced glycation products [3]. water at the Precinct Animal Centre of the Baker IDI Heart and Plasmalogens (alkenylphosphatidylcholine, PC(P) and alkenyl- Diabetes Institute. They were maintained on a 12 h light and dark phosphatidylethanolamine, PE(P)) are subclasses of glycer- cycle in a pathogen free environment. Food intake was measured ophospholipids that are characterised by a cis vinyl ether bond weekly by food pellet consumption. Body weights were determined linking an alkyl chain to the sn-1 position of the glycerol backbone. weekly. After 12 weeks, animals were anaesthetised by Avertin Plasmalogens are synthesised from the corresponding alkylphos- (2,2,2-tribromoethanol) IP (0.3 mL of 2.5% solution per 20 g mouse; pholipids (alkylphosphatidylcholine, PC(O) and alkylphosphatidy- Sigma Chemical Co, USA) following food withdrawal for 3 h, and lethanolamine, PE(O)) by the action of a desaturase. Plasmalogens organs were rapidly dissected and snap frozen. The experiment was have been proposed to be atheroprotective, partly because of its approved and conducted in accordance to the principles devised by anti-oxidant characteristics: (1) an enhanced electron density and the Alfred Medical Research and Education Precinct Animal Ethic low bond dissociation of the vinyl ether linkage which makes them Committee under guidelines laid down by the National Health and more susceptible to reactive oxygen species (ROS) attack than Medical Research of Council of Australia (E/1345/2013/B). allylic and alkyl linkages [4]; (2) plasmalogen makes up one of the lipid components in cellular phospholipid bilayer which is a target 2.2. Clinical measurements of free-radical chemical reactions [4]; and (3) the proposed slow propagation of the plasmalogen hemiacetal hydroperoxy radicals Fasting blood glucose was measured by tail bleed using a gluc- (ie. plasmalogen oxidative intermediate) [4]. Indeed, the anti- ometer (Accu-Chek, Roche Diagnostics, Australia) following 3 h of oxidative role of plasmalogen was demonstrated in studies where food withdrawal and prior to the organ dissection. Blood was ob- supplementation of alkylglycerol (a precursor to plasmalogen tained via direct heart puncture during the organ dissection, and synthesis) to plasmalogen deficient mutant Chinese Hamster Ovary was collected into EDTA tubes. Plasma was separated from the cells and macrophage like cells, RAW.12 and RAW.108 were shown blood via centrifugation at 1485 g, room temperature for 10 min, to improve resistance against ROS insults including long- and sucrose was added (final concentration of 0.6% (v/v)) as a wavelength ultraviolet light and ROS generators [5,6]. In addition, cryoprotectant for lipoproteins [15,16] prior to storage at 80 C. plasmalogen was shown to have potential atheroprotective roles; Plasma was thawed slowly on ice and then concentrations of Plasmalogen was demonstrated to be essential for intracellular total cholesterol and triglycerides were measured using commer- cholesterol transport [7] and in high-density lipoprotein (HDL)- cial enzymatic kits on a COBAS Integra 400 Plus blood chemistry mediated cholesterol efflux [8], and recently, the inclusion of analyser (Roche Diagnostics, Australia). Fasting plasma insulin was plasmalogen into reconstituted HDL improved the lipoprotein anti- measured using an ELISA kit according to the manufacturer's in- apoptotic activity on endothelial cells [9]. structions (ALPCO, USA). We have previously reported a negative association of circu- lating plasmalogens with both stable and unstable coronary artery 2.3. Tissue homogenisation disease; In parallel, we observed a positive association with the level of plasmalogen oxidative by-product, lysophosphatidyletha- Mice hearts were cut into two halves. The bottom half was snap nolamine (LPE) [10]. These findings suggested a depletion of plas- frozen in liquid nitrogen and was subsequently homogenised in ice malogens in the patients due to oxidative degradation, implying a cold phosphate buffered saline (200 mL, pH 7.6) containing higher level of oxidative stress. 100 mmol/L butylated hydroxytoluene using a Polytron electric The modulation of plasmalogen concentration by oral admin- homogeniser for 10 s and then with a mini probe homogeniser for istration of alkylglycerol has been demonstrated in humans and 15 s at amplitude 23. The homogenate was stored at 80 C. rodents [11], but the effect of plasmalogen modulation in athero- sclerosis has not been previously investigated. Batyl alcohol (BA) is 2.4. Lipoprotein fractionation a naturally occurring alkylglycerol found in human plasma and tissues and is particularly abundant in human breast milk [12]. Lipoprotein fractionation was performed by density ultracen- Shark liver oil is a natural source of alkylglycerols that has been trifugation using a method adapted from Havel et al. [17]. Briefly, exploited as a nutraceutical for many years [13,14]. We hypoth- EDTA was added to the plasma (100 mL) to a final concentration of esised that modulation of plasmalogen concentration by BA would 2 mmol/L and the density was adjusted to 1.019 g/mL in a final attenuate atherosclerosis progression. As a proof of concept that volume of 1.0 mL. The sample was centrifuged (435,680 g, 16 C, may show the atheroprotective mechanism of plasmalogen, here 3 h) in a TLA 120.2 rotor and Optima MAX-TL ultracentrifuge we assess the effect of plasmalogen enrichment in murine models (Beckman Coulter, NSW, Australia). The top layer (400 mL) of the of atherosclerosis with differing levels of oxidative stress; ApoE / sample, corresponding to the VLDL fraction was aspirated. The and ApoE / GPx1 / mice. density of the remaining mixture was the adjusted to 1.063 g/mL and overlayed with the same density solution to a final volume of 2. Materials and methods 1.0 mL. The sample was centrifuged (435,680 g,16 C, 3 h) and the top layer (400 mL) corresponding to the LDL fraction was aspirated. 2.1. Animal groups and diet study The density of the remaining mixture was further adjusted to 1.21 g/mL and the volume made up to 1.0 mL. The samples were Six-week old male C57/BL6 (Animal Resources Centre, WA, centrifuged (435,680 g, 16 C, 16 h) and the top layer (400 mL) Australia), ApoE \ (Animal Resources Centre, WA, Australia), and corresponding to HDL fraction was aspirated. ApoE \ GPx1 \ (Alfred Medical Research and Education Precinct, VIC, Australia) mice, both on C57/BL6 background, were fed a high 2.5. Lipid extraction fat diet (22% fat, 0.15% cholesterol) (Specialty Feeds, WA, Australia), containing either 0% or 2% 1-O-octadecyl-rac-glycerol (batyl Prior to lipid extraction, samples were randomised to reduce alcohol, Tokyo Chemical Industry, Astral Scientific, Australia) for 12 bias. Lipids were extracted as previously described [18]. Briefly, weeks (N ¼ 10/group). Power analysis was conducted prior to the plasma, lipoprotein or homogenised tissue was combined with animal experiment to ensure proper sampling. The animals were internal standards (see Supplementary Table I) and the lipids were housed in standard conditions with unrestricted access to food and extracted using 20 volumes of chloroform:methanol (2:1). The 600 A.A. Rasmiena et al. / Atherosclerosis 243 (2015) 598e608 extracted lipids were dried under a stream of nitrogen at 40 C and manner. Necrotic core of the aortic lesions were analysed and subsequently reconstituted in 1:1 mixture of water saturated quantified as the percentage (%) of area of lesion. butanol and methanol containing 5 mmol/L ammonium formate. Lipid extraction was performed on 10 mL of heart homogenate 2.8. Immunohistochemistry (50 mg protein) or plasma, 50 mL aliquots of VLDL, LDL and HDL from C57/BL6 mice and HDL from the ApoE / and ApoE / GPx1 / mice Immunohistochemical methods were employed as previously or 10 mL aliquots of VLDL and LDL from the ApoE / and ApoE / described [21]. Aortic sinus was frozen to allow for VCAM-1 GPx1 / mice. staining (rat monoclonal, BD Pharmigen; 1:200) and F4/80 stain- ing (rat monoclonal, Abcam; 1:75). Briefly, 4 mm frozen sections 2.6. Liquid chromatography electrospray ionisation tandem mass were fixed with cold acetone, air dried, and then quenched in 3% spectrometry hydrogen peroxide in Tris buffered saline, pH 7.6 to prevent endogenous peroxidase activity. Sections were then blocked using Lipids were quantified using multiple reaction monitoring mode 10% rabbit serum in Tris buffered saline, and incubated with pri- on a Agilent 1200 high pressure liquid chromatography system mary antibodies overnight in a humidified chamber at 4 C. Sec- coupled to a Q/TRAP 4000 triple quadrupole mass spectrometer (AB tions were then incubated with biotinylated secondary antibodies SCIEX) using methodology described previously [18] (see anti-rat IgG (raised in rabbit, Vector Laboratories; 1:200) for Supplementary Table II). Liquid chromatography separation was 10 min at room temperature, followed by horseradish peroxidase- performed on a 2.1 100 mm C18 Poroshell column (Agilent, USA) conjugated streptavidin (Vectastain Elite ABC Staining Kit, Vector at 300 mL/min. The following gradient conditions were used: 10% B Laboratories) for 30 min at room temperature. Signals were to 100% B over 13 min, 100% B over 3 min, and a return 10% B over visualised with 3,3’-diamino-benzidine terahydrochloride/ 1 min, followed by 10% B over 3 min. Solvents A and B consisted of hydrogen peroxide (SigmaeAldrich, USA). Subsequently, sections water:tetrahydrofuran:methanol in the ratio of 60:20:20 and were counterstained in Mayer's haematoxylin, dehydrated, and 5:75:20 respectively, both containing 10 mmol/L ammonium coverslipped. formate. The conditions for the tandem mass spectrometry of each Aorta was paraffin-fixed to allow staining for nitrotyrosine lipid class are provided in Supplementary Table II. Different lipid (rabbit polyclonal, Millipore; 1:200) and 4-Hydroxynonenal (rabbit species under the same lipid class were analysed the same way. The polyclonal, Millipore; 1:200). In brief, 5 mm of the aortic sections levels of individual lipid species were measured by taking a ratio of were de-waxed, hydrated, and quenched in 3% hydrogen peroxide the area under the curve of lipid of interest to the area under the in Tris buffered saline, pH 7.6. The sections were blocked using 10% curve of internal standard of the corresponding lipid class. The ratio horse serum in Tris buffered saline, and then incubated with pri- was then multiplied by the amount of internal standards added into mary antibodies overnight in a humidified chamber at 4 C. The the sample. Lipid classes were calculated from the sum of indi- sections were incubated with biotinylated secondary anti-rabbit vidual species within each class. IgG antibodies (raised in goat, Vector Laboratories; 1:500), fol- Lysophosphatidylcholine (LPC) were analysed by their frag- lowed by horseradish peroxidase-conjugated streptavidin as mentation ions at m/z 184, whereas LPE were analysed by the described above. The sections were subsequently processed in the neutral loss of 141 Da from the parent ion. The regioisomers of same manner as the frozen sections as described above. lysophospholipids such as sn-1 lysophospholipids where the fatty acid is at the sn-1 position of glycerol backbone were distinguished 2.9. Data analysis and statistics from the sn-2 isomers by their retention time on the chromatog- raphy; sn-2 isomers were eluted slightly earlier than sn-1 isomers Mass spectrometric data were analysed using a ManneWhitney [19,20]. The liquid chromatography method coupled with tandem U test, and were adjusted for multiple comparisons using the mass spectrometry as described above allow separation and mea- Benjamini Hochberg method, P < 0.05 was considered significant. surement of sn-1 and sn-2 isomers of lysophospholipids. Data were expressed as median (interquartile range). The lipidomic data were normalised to the relative total level of phosphatidyl- 2.7. Quantification of atherosclerotic plaque choline in plasma and heart to allow a direct comparison of the relative lipid levels between the two sample types. The entire aorta was cleaned of peripheral fat under a dissecting Data from the atherosclerotic plaque assessment and immu- microscope (Olympus SZX9, Olympus Optical, Tokyo, Japan). The nohistochemistry were analysed using Student t-test, comparing aorta was then stained with Sudan IV-Herxheimer's solution (BDH, the control and BA-treated groups per individual genotype, P < 0.05 Poole UK) and the en face technique was employed to assess the was considered significant. Data were expressed as mean ± SEM. total and regional (arch, thoracic, and abdominal) plaque area as previously described [21]. Digitised photographs of the opened 3. Results aortas were obtained using the dissecting microscope equipped with a digital camera (Axiocam colour camera; Carl Zeiss, North 3.1. Body weight and plasma lipid profile Ryde, NSW, Australia). Plaque area was quantified as the proportion of aortic intimal surface area occupied by red-stained plaque using In both the ApoE \ , and ApoE \ GPx1 / genotypes, mice fed a Abode Photoshop v.6.0.1 (Adobe Systems, Chatswood, NSW, high fat diet without BA supplementation had a higher body weight Australia). in the final week of the study as compared to those that received BA Frozen sections of aortic sinus were assessed as previously supplementation (Table 1). In ApoE / mice, there was 17.4% dif- described [2,22]. Briefly, the sections (4e12 sections per mouse) ference (P < 0.001) in the final weight between the two dietary were stained with Sudan IV-Herxheimer's solution and digitised groups, whereas in ApoE / GPx1 / mice, there was a 10.7% dif- photographs were taken using a light microscope (Olympus BX-50, ference (P < 0.01) (Table 1). All mice showed an increase in body Olympus Optical, Tokyo, Japan) equipped with the digital camera. weight across the study period (Supplementary Figure I). However, The lesion area indicated by the red-stained plaque was quantified both ApoE \ and ApoE \ GPx1 / mice experienced slight weight as lesion area (mm [2]) using Image-Pro Plus v.6.0 (Media Cyber- loss after the first week on the BA supplementation, after which the netics, Rockville, USA). All assessments were made in a blinded weight increased weekly (Supplementary Figure I). Food intake A.A. Rasmiena et al. / Atherosclerosis 243 (2015) 598e608 601
Table 1 Plasma lipid measurements of mice on a high fat diet supplemented with/without 2% BA.