Metabolomics in atherosclerosis

Citation for the original published paper (version of record):

Djekic, D., Nicoll, R., Novo, M., Henein, M. (2015)

Metabolomics in atherosclerosis.

International Journal of Cardiology Metabolic & Endocrine, 8: 26-30

http://dx.doi.org/10.1016/j.ijcme.2014.11.004

Access to the published version may require subscription.

N.B. When citing this work, cite the original published paper.

Permanent link to this version:

Metabolomics in atherosclerosis

Demir Djekic

_{, Rachel Nicoll}

_{, Mehmed Novo}

_{, Michael Henein}

⁎

a

Department of Public Health and Clinical Medicine, Umeå University And Heart Centre, Umeå, Sweden

b

Department of Community Medicine and Rehabilitation, Umeå University, Sweden

a b s t r a c t

a r t i c l e i n f o

Article history: Received 20 June 2014

Received in revised form 13 November 2014 Accepted 25 November 2014

Available online 3 December 2014 Keywords:

Atherosclerotic cardiovascular disease Metabolomics

Proteomics Lipidomics Atherosclerosis

It is well established that atherosclerotic cardiovascular disease (ACD) is a leading cause of death in the West. There are several predisposing factors for ACD, which can be divided into two groups:firstly modifiable risk factors, including hypertension, dyslipidaemia, type 2 diabetes mellitus, obesity, smoking and a sedentary lifestyle and secondly the unmodifiable risk factors such as age, gender and heredity. Since single biomarkers are unable to provide sufficient information about the biochemical pathways responsible for the disease, there is a need for a holistic approach technology, e.g., metabolomics, that provide sufficiently detailed information about the metabolic status and assay results will be able to guide food, drug and lifestyle optimisation. Rather than investigating a single pathway, metabolomics deal with the integrated identification of biological and pathological molecular pathways. Mass spectrometry (MS) and nuclear magnetic resonance (NMR) spectroscopy are the two most commonly used techniques for metabolite profiling. This detailed review concluded that metabolomics investigations seem to have great potential in identifying small groups of disturbed metabolites, which if put together should draw various metabolic routs that lead to the common track pathophysiology. The current evidence in using metabolomics in atherosclerotic cardiovascular disease is also limited, and more well-designed studies remain to be established, which might significantly improve the comprehension of atherosclerosis pathophysiology and consequently management.

© 2014 The Authors. Published by Elsevier Ireland Ltd. This is an open access article under the CC BY-NC-ND 4.0 license (http://creativecommons.org/licenses/by/4.0/).

1. Introduction

It is well established that atherosclerotic cardiovascular disease (ACD) is a leading cause of death in the West[1]. It is a chronic inflammatory dis-order[2]that begins in early childhood with fatty streak formation in the coronary arteries[3]and remains silent until late adulthood, when lipids start to accumulate in the intima forming atherosclerotic plaques[4]. There are several predisposing factors for ACD, which can be divided into two groups:first, modifiable risk factors, including hypertension, dyslipidaemia, type 2 diabetes mellitus, obesity, smoking and a sedentary lifestyle and, second, the unmodifiable risk factors such as age, gender and heredity[5]. Ageing is the predominant risk factor for ACD, and it renders many of the modifiable risk factors more prevalent and more severe[6]. Thus, the aetiology of atheroma formation is multifactorial with a syner-gistic effect of many risk factors. At present, cardiologists remain using conventional risk factors, derived from the Framingham studies[7]. Although helpful, those risk factors do not detect all instances of CAD, and indeed a significant percentage of patients with myocardial infarction has no conventional risk factors[8]. Other markers such as coronary calci-fication have been found to improve risk straticalci-fication but still fall a long

way short of 100% prediction[8]. Conventional coronary angiography is invasive, time consuming and expensive and subjects the patient to radiation, while computed tomography coronary angiography (CTCA) also suffers from many of the same limitations.

On the other hand, single disease biomarkers are always desired to identify risk factors or people affected by a disease to evaluate progress or to monitor an intervention in treating the disease. This principle has been shown useful for, e.g., bacterial infections, where a specific mole-cule or groups of molemole-cules are characteristic of the disease state and largely distinctive within the matrix being sampled. The commonly investigated biomarker in ACD is cholesterol. However, serum choles-terol levels may fail to identify the exact pathway to their abnormal levels since sources of increased cholesterol levels are not only food but also endogenous biosynthesis or slow conversion to bile acids[9]. Since single biomarkers are unable to provide sufficient information about the biochemical pathways responsible for the disease, there is a need for a holistic approach technology, e.g., metabolomics, that provide sufficiently detailed information about the metabolic status and assay results will be able to guide food, drug and lifestyle optimisation[9].

The metabolome refers to the complete set of small molecule (low molecular weight (b1500 Da)) metabolites in a cell, tissue, organ or organism[10], while metabolomics is the comprehensive analysis and quantification of these small molecules based on biofluids and tissue

IJC Metabolic & Endocrine 8 (2015) 26–30

⁎ Corresponding author.

E-mail address:Michael.Henein@medicin.umu.se(M. Henein).

http://dx.doi.org/10.1016/j.ijcme.2014.11.004

2214-7624/© 2014 The Authors. Published by Elsevier Ireland Ltd. This is an open access article under the CC BY-NC-ND 4.0 license (http://creativecommons.org/licenses/by/4.0/). Contents lists available atScienceDirect

IJC Metabolic & Endocrine

The term metabolic profile was first introduced by Hornings et al. in early 1970s[16]. They suggested that metabolic profiles may be valuable for characterising both normal and disease state. The‘-omics’ approach includes genomics, transcriptomics, proteomics and the rapidly emerging metabolomics[17].

Mass spectrometry (MS) and nuclear magnetic resonance (NMR) spectroscopy are the two most commonly used techniques for metabo-lite pro_filing[18]. Each of these techniques has important strengths and weaknesses. For example, MS is innately more sensitive than NMR but requires a prior separation of metabolites using chromatography (liquid or gas chromatography) or capillary electrophoresis (CE)[19]. Further, the ionisation or ion suppression effect could impair the analytic quan-tification. The MS can, with standard techniques, detect smaller metab-olites than NMR (picomolars for MS and micromolars for NMR), and those metabolites detected with NMR need to contain a hydrogen atom. By using NMR sample, recovery is non-destructive, and the sample is analysed in only one measurement while for MS only a small amount of sample is used that eventually will be destructed[20]. Both techniques can be used to characterise metabolic data in a targeted or non-targeted aspect. For the targeted approach, the investiga-tor focuses on a limited set of metabolites of known identity. Untargeted approach deals with the investigation of as many peaks as possible, with unknown underlying identities of the species. Therefore, untargeted anal-yses are considered more sensitive and more likely to discover new bio-markers, while target analyses are used in biomarker validation[21]. Because of data complexity, NMR can assign definitive metabolite identi-ties to only a subset of peaks arising from the sample. Although MS can generate sometimes thousands of peak metabolites from the biofluids, chromatographic extraction time and mass to change ratio are often in-sufficient to confidently assign peak intensities[21]. The organisation of the human metabolome database (HMDB) is a comprehensive online da-tabase that gathers small molecule metabolites found within the human body. The HMDB second version, produced in 2009, was able to identify 6500 metabolites[22]and the latest version available from 2013 has sig-nificantly expanded to identify more than 40.000 metabolites[23].

During the last decade, there has been growing interest in the appli-cation of metabolomics for early disease prevention and to potentiate drug development and therapy monitoring[8,9]. In the setting of coro-nary atherosclerosis, metabolomics has been shown to identify a num-ber of disordered biomarkers, some of which may be susceptible to modification. The development of a set of metabolites which could aid prediction of ACD would certainly be welcomed. Furthermore, it is pos-sible that urine or salivary samples may be used for metabolic analysis, which even avoids the need for phlebotomy. In this review, we will discuss the latest metabolomic approaches within thefield of ACD and its related risk factors.

2. Atherosclerosis 2.1. Animal studies

Lyso-phosphatidylcholine (LPC) is a class of phospholipids that are intermediates in the metabolism of lipids. LPC is generated by an enzyme found on oxidised LDL called lipoprotein-associated phospholi-pase A2 from phosphotidylcholine[24]. Increased levels of this enzyme

In another study by Kleemann et al. [28], ApoE Leiden mice were fed a low vs. high cholesterol diet. The amount of dietary cholesterol posi-tively correlated with development of atherosclerosis. With increasing dietary cholesterol intake, the liver switched from a mainly resilient to a predominantly in_{flammatory state, which is associated with early} lesion formation. The high cholesterol evoked changes involving specif-ic transcriptional master regulators, some of whspecif-ich are established, others newly identi_{fied, several of these regulators control both lipid} metabolism and inflammation and thereby link the two processes. The liver and plasma were analysed with HPLC/MC identified disturbance in di-and triglycerides, phosphatidylcholines, lysophosphatidylcholines and cholesterol esters.

The effect of different diets on atheroma formation has been also studied. Martin et al. investigated the effect of several diets fed to hyper-lipidaemic hamsters and observed that aortic cholesteryl ester, assessed by NMR spectroscopy, was an early accumulator in atherogenic plaques. The lowest atherogenicity was obtained with the plant-oil cheese diet, followed by the dairy fat cheese diet, while the greatest atherogenicity was observed with the butter diet. Aortic cholesteryl ester was positive-ly correlated with very low density lipoprotein (VLDL), cholesterol and N-acetylglycoproteins and negatively correlated with trimethylamine-N-oxide (TMAO) and albumin lysyl[29].

Likewise, Jove et al. observed that a high-fat diet caused an increase in ceramide and docosahexaenoic acid (DHA) in plasma and a tissue sample from the aorta. Free cholesterol in the aorta was positively correlated with taurocholic acid, suggesting that it could be a biomarker for early atherogenesis[30]. Furthermore, a high-fat cholesterol choate (HFCC) diet has been shown to alter plasma and urinary metabolism in LDL-receptor-deficient mice using H-NMR spectroscopy. The HFCC diet caused a significant perturbation in choline metabolism, notably the choline oxidation pathway a significant reduction in the urinary excretion of taurine, betaine and dimethylglycine[31].

In addition to these studies, recent ones showed that not only diet and liver have a role in the development of atherosclerosis but also gutfloras have demonstrated great importance. It has been suggested that microbiome, i.e., the collective genomes of the microorganisms that reside in the gut, can increase cardiovascular risk either via metabo-lism ofL-carnitine[32]or phosphatidylcholine (Wang at al 2011). Mice

were given either a control diet or a diet rich in choline. The gut microflora was suppressed in half of the mice given broad-spectrum antibiotics[33]. Upon analyzing the plasma, TMAO was suppressed to a non-detectable level. The development of the aortic root lesion was increased 3-fold in mice not treated with antibiotics, and hence with preserved gut micro flo-ra, and in those fed a choline-enhanced diet. Thus, Wang et al. supposed that dietary supplements of choline, TMAO and betaine (which are all metabolites of phosphatidylcholine) could enhance the development of atherosclerosis. Furthermore, it was proposed that these metabolites are involved in the upregulation of scavenger receptors on macrophages hence the formation of foamy cells.

In another study, Stöhr et al.[34]also linked atherosclerosis to the metabolism of carnitine. By using a targeted metabolomic approach, the group studied the plasma of genetically modified mice susceptible to atherosclerosis. Interestingly, they observed a decrease in the blood concentration of free carnitine, acetylcarnitine and glutarylcarnitine/3-hydroxy-hexanoylcarnitine.

2.2. Human studies

Brindle et al.[35]were the_{first to investigate the use of} metabolo-mics for the diagnosis of ACD. They could not only confirm the presence of ACD but also distinguish its severity by using H-NMR spectroscopy on human sera. A supervised partial least squares discriminant analysis to orthogonal signal-corrected data sets allowedN90% of subjects with stenosis of all three major coronary vessels to be distinguished from subjects with angiographically normal coronary arteries, with a speci fic-ity ofN90%. In general, the regression coefficients, or loadings, most influential for the triple vessel disease (TVD) samples lie aroundδ0.86

(due mainly to CH3 groups from fatty acid side chains in lipids, in partic-ular, LDL and VLDL) andδ1.26,δ1.3 andδ1.34 (due mainly to (CH2)n

groups from fatty acid side chains in lipids, in particular VLDL and LDL). The loadings most influential for the normal coronary artery (NCA) sam-ples lie aroundδ1.22 (due mainly to (CH2)n groups from fatty acid side

chains in lipids, in particular, HDL) andδ3.22 (due to choline-N(CH3)3+).

A lipidomic approach was used by Sun et al. to study the plasma metabolome of patients with unstable angina and atherosclerosis con-trols and suggested that 16 potential biomarkers could aid in diagnosis of unstable angina. Phytosphingosine and phosphatidylcholine were elevated, and phosphatidylglycerol was reduced relative to the athero-sclerosis control[36].

Similarly, Stübigeret al. [37]used targeted lipidomics to study the plasma in young patients with familial hyperlipidemia that are at higher risk to develop ACD. Significant alteration of sphingomyelin (SM)/phophatidylcholine (PC) and phosphatidylcholine (PC)/lyso-phosphatidylcholine (LPC) and positive correlation of SM with LDL-C and LPC with VLDL-C were found in the familiar hyperlipidaemic group in contrast to normolipidaemics. Further, a positive correlation of oxidised PC with IMT and HDL-C but negative correlation with oxidised LDL was observed.

Using quantitative MS-based metabolic profiling in 117 Caucasians with strong family history of premature ACD, Shah et al.[38]showed significantly increased ketone bodies (B-hydroxybutarate), several amino acids, e.g., glutamate and glutamine, free fatty acids (arachidonic acid) and most notably acylcarnitine. Acylcarnitine facilitates the entry of long-chain fatty acids into the mitochondrion via the carnitine shut-tle, which is critical for its use by the myocardium for ß-oxidation. Where there are fatty acid oxidation defects, acylcarnitine species accumulate and are released into the circulation. In a follow-up study, the authors distinguished the metabolic profile of 314 ACD patients from controls based on differences in branched-chain amino acids and acylcarnitines between patients and controls. Interestingly, the dicarboxylacylcarnitine was associated with the incidence of CV events[39].

Comparative use of techniques has also been tested by Teul et al.[40], who investigated the plasma of 9 patients with stable carotid atheroscle-rosis and 10 healthy individuals. They showed that a combination of both gas chromatography–mass spectrometry (GC-MS) and NMR resulted in a broader collection of deranged metabolites than using either method alone. They also reported alteration of many metabolomics path-ways, such as amino acid metabolism, decrease in metabolites of Krebs cycle and pyruvate and an elevation of ketone bodies and 2,3,4 trihydroxybutarate (THD). Most of the changes can be associat-ed with alterations of the metabolism characteristics of insulin resis-tance that can be strongly related to the metabolic syndrome.

To investigate the effect of age on the progression of ACD, Rizza et al.

[41]using MS were able to profile 49 metabolites in elderly with a high rate of previous ACD. They suggested that the metabolic profile in elderly was associated with mitochondrial dysfunction and damage and those specific metabolites could aid in the prediction of major CV events.

Zheng et al.[42]investigated the variability on human serum meta-bolic profile in communities at risk of atherosclerosis. They calculated intraclass correlation coefficients (ICC) for 178 metabolites detected

by untargeted method in 60 patients. They observed that pathways of lipid and amino acid metabolism had a relatively high ICCs and that me-tabolites in carbohydrate pathway showed relatively low ICCs. 2.3. Lipidomics and atherosclerosis

The predictive value of metabolomics has also been studied. Studies of lipid profiles indicate that individuals with smaller particle LDL have a greater CV risk compared to those with larger particle LDL. McMahan et al.[4]studied the LDL subclasses using H-NMR spectroscopy with carotid intimal-medial thickness (IMT) detected by ultrasound and found that both forms of LDL were significantly associated with IMT, but the large and small LDL particles were inversely correlated. Furthermore, Shah et al.[43], using MS in over 2000 patients undergoing cardiac cath-eterisation, identified five metabolites associated with a higher mortal-ity, namely, the medium chain acylcarnitines, short- and long-chain dicarboxylacylcarnities, branched-chain amino acids and fatty acids.

In 4309 healthy sera of individuals, Wurtz et al. observed using NMR that increasing concentrations of VLDL, intermediate-density lipopro-tein (IDL) and LDL subclasses and low concentrations of HDL were asso-ciated with phenotypes that were at the greatest risk for atherosclerosis, findings that mirror biochemistry[44]. Likewise, the urine profile of 4630 patients from four populations groups (Japan, China, UK, USA) showed that alanine correlated positively while hippurate correlated inversely with blood pressure[45]. The same technique has been shown to differentiate between high- versus low-risk individuals based on con-ventional atherosclerosis risk factors (cholesterol, triglycerides, LDL and HDL). Low-risk individuals had high 3-hydroxybutarate and low levels of threonine, whereas higher risk individuals were associated with low level of a ketoglutare and dimethylglycine[46]. The HDL in human plasma has recently been shown to play a central role in atheroprotection. Lipidomic approach revealed that the abundance of PC, LPC PS and PA was elevated in small, dense in contrast to large, light HDL; the inverse occurred for sphingomyelin and ceramide. Interestingly, several compo-nents of HDL were strongly correlated with antioxidative, antithrombotic, anti-inflammatory and antiapoptotic activity[47]. Intraplaque lipids have also been analysed by Stegemann et al.[48], who showed a higher concentration of polyunsaturated cholesteryl esters with long-chain fatty acids and certain sphingomyelin.

The effect of inducible myocardial ischaemia was also tested in 36 patients using MS, and 23 metabolites were significantly altered in the ischemic group but not in controls, while six metabolites were related to the Krebs cycle, including citric acid with a high degree of accuracy. Oxaloacetate and citruline were significantly reduced in the ischemic group and correlated with the severity of ischemia[49].

2.4. Proteomics and atherosclerosis

Proteomics have also been studied in individuals and patients with atherosclerosis. In 359 urine samples from individuals with severe ACD using capillary electrophoresis coupled to ESI-TOF-MS, Zimmerli et al.[50] identified 15 metabolites, which discriminated between patients and healthy controls with a 98% sensitivity and 83% specificity as well as with the level of exercise after coronary intervention. Subse-quently, the same authors reported 238 discriminatory biomarkers in another sample of ACD patients with a sensitivity of 79% and specificity of 88%. Among these markers were fragments ofα1-antitrypsin, collagen type 1 and 3, granin-like neuroendocrine peptide precursors, membrane-associated progesterone receptor component 11, sodium potassium ATPase gamma chain andfibrinogen alpha chain[51].

Early proteomic markers of ACD have also been studied by Delles et al. using (CE-MS) in ApoE-deficient mice fed a high-fat diet compared to a low fat diet. Polypeptide fragments of alpha1-antitrypsin, epidermal-like growth factor and collagen allowed identification of atherosclerosis with a sensitivity of 90% and 100% specificity. Furthermore, using immunohis-tochemistry,α1-antitrypsin, EGF and collagen type I were shown to be

as 95 proteins (involved with immune defense, inflammation, growth and coagulation) were discerned in patients vs controls. Even in patients with acute coronary syndrome, Dardé et al.[55]identified similar altered proteins at day 0, 4, 60 and 180 using 2D gel electrophoresis.

2.5. Association of metabolites with atherosclerosis risk factors

Spijkers et al. studied the plasma metabolome and isolated arterial tissue from spontaneous hypertensive rats (SPR) and Wistar–Kyoto (WKY) normotensive rats and found altered sphingolipid metabolism in the former group. They also observed endothelium-dependent contraction in the arteries of SHR but not in WKY upon administration of a sphingosine kinase inhibitor (SKI) or sphingomylinase. These contractions were mediated by ceramide, which was elevated in the plasma and correlated with severity of hypertension[56]. Ceramide is known to cause apoptosis[57]and vascular dysfunction[58].

In humans, plasma lipidomics were studied in 19 hypertensive and 51 normotensive males. Graessler et al.[59]observed a signi_{ficant reduction} of ether phosphatidylcholines and ether phosphatidylethanolamies and suggested that these ether lipids could contribute to development of hypertension. More speci_{fically those reduced ether lipids constituted} arachidonic acid. In addition, among the obese subjects, there was a significantly increased level of saturated triacylglycerides (TAG) and diacylglycerol (DAG).

Furthermore, testing of Mexican Americans at risk of dyslipidaemia and insulin resistance demonstrated that higher systolic blood pressure was significantly associated with DAG and that the diastolic blood pressure was associated with elevated levels of monohexosylcermide, phosphatidylcholine and DAG [60]. Considering that DAG acts on TRPC6 channel mediating vasoconstriction[61], it was proposed as a potential marker for hypertension. Also, a comparative study between African Americans and Caucasians showed significant reduction in palmitic, oleic, pamitoleic, aracidonic and linoleic free fatty acids in the latter compared to the former[62].

In a recent study, Zheng et al.[63]investigated the serum of patients with incident hypertension. Upon using gas chromatography–mass spectrometry and liquid chromatography, they found a significant association between six steroids and the risk of incident hypertension (highest versus lowest quintile hazard ratio, 1.72; 95% confidence interval, 1.05-2.82; P for trend, 0.03), in both men and women. 3. Conclusion

The current state of understanding atherosclerosis pathophysiology and its relationship to risk factors is not entirely comprehensive neither satisfactory, particularly in various groups of patients who do notfit within the one big box. Metabolomics investigations seem to have great potential in identifying small groups of disturbed metabolites, which if put together should draw various metabolic routs that lead to the common track pathophysiology. Nevertheless, this association does not necessarily imply that these disturbed metabolites have caused the atherosclerosis; other potential explanations include the atheroscle-rosis causing the disturbed metabolites or that the metabolites are simply markers for the disease. The current evidence in using metabolo-mics in atherosclerotic cardiovascular disease is limited and more well

References

[1]D'Agostino RB, Russell MW, Huse DM, Ellison RC, Silbershatz H, Wilson PW, et al. Primary and subsequent coronary risk appraisal: new results from the Framingham study. Am Heart J 2000;139(2 Pt 1):272–81.

[2]Yu XH, Jiang N, Zheng XL, Cayabyab FS, Tang ZB, Tang CK. Interleukin-17A in lipid metabolism and atherosclerosis. Clin Chim Acta 2014.

[3]McMahan CA, Gidding SS, McGill HC. Coronary heart disease risk factors and atheroscle-rosis in young people. J Clin Lipidol 2008;2(3):118–26.

[4]McMahan CA, Gidding SS, Viikari JS, Juonala M, Kähönen M, Hutri-Kähönen N, et al. Association of Pathobiologic Determinants of Atherosclerosis in Youth risk score and 15-year change in risk score with carotid artery intima-media thickness in young adults (from the Cardiovascular Risk in Young Finns Study). Am J Cardiol 2007; 100(7):1124–9.

[5]Badimon L, Vilahur G, Padro T. Nutraceuticals and atherosclerosis: human trials. Cardiovasc Ther 2010;28(4):202–15.

[6]Majeed F, Kelemen MD. Acute coronary syndromes in the elderly. Clin Geriatr Med 2007;23(2):425–40 2.

[7]Kannel WB, Doyle JT, McNamara PM, Quickenton P, Gordon T. Precursors of sudden coronary death. Factors related to the incidence of sudden death. Circulation 1975; 51(4):606–13.

[8]Nicoll R, Henein MY. Arterial calcification: friend or foe? Int J Cardiol 2013;167(2): 322–7.

[9]German JB, Hammock BD, Watkins SM. Metabolomics: building on a century of biochemistry to guide human health. Metabolomics 2005;1(1):3–9.

[10]Schmidt C. Metabolomics takes its place as latest up-and-coming "omic" science. J Natl Cancer Inst 2004;96(10):732–4.

[11]Griffiths WJ, Koal T, Wang Y, Kohl M, Enot DP, Deigner HP. Targeted metabolomics for biomarker discovery. Angew Chem Int Ed Engl 2010;49(32):5426–45.

[12]Goodacre R, Vaidyanathan S, Dunn WB, Harrigan GG, Kell DB. Metabolomics by numbers: acquiring and understanding global metabolite data. Trends Biotechnol 2004;22(5):245–52.

[13]Peng J, Gygi SP. Proteomics: the move to mixtures. J Mass Spectrom 2001; 36(10):1083–91.

[14]Gates SC, Sweeley CC. Quantitative metabolic profiling based on gas chromatography. Clin Chem 1978;24(10):1663–73.

[15]Saito M. History of supercriticalfluid chromatography: instrumental development. J Biosci Bioeng 2013;115(6):590–9.

[16]Horning EC, Horning MG. Metabolic profiles: gas-phase methods for analysis of metabolites. Clin Chem 1971;17(8):802–9.

[17]Goldsmith P, Fenton H, Morris-Stiff G, Ahmad N, Fisher J, Prasad KR. Metabonomics: a useful tool for the future surgeon. J Surg Res 2010;160(1):122–32.

[18]Gebregiworgis T, Powers R. Application of NMR metabolomics to search for human disease biomarkers. Comb Chem High Throughput Screen 2012;15(8):595–610.

[19]Lindon JC, Nicholson JK. The emergent role of metabolic phenotyping in dynamic patient stratification. Expert Opin Drug Metab Toxicol 2014;1–5.

[20]Lindon JC, Nicholson JK. Spectroscopic and statistical techniques for information recovery in metabonomics and metabolomics. Annu Rev Anal Chem (Palo Alto, Calif) 2008;1:45–69.

[21]Rhee EP, Gerszten RE. Metabolomics and cardiovascular biomarker discovery. Clin Chem 2012;58(1):139–47.

[22]Wishart DS, Knox C, Guo AC, Eisner R, Young N, Gautam B, et al. HMDB: a knowledgebase for the human metabolome. Nucleic Acids Res 2009;37(Database issue):D603–10.

[23]Wishart DS, Jewison T, Guo AC, Wilson M, Knox C, Liu Y, et al. HMDB 3.0—The Human Metabolome Database in 2013. Nucleic Acids Res 2013;41(Database issue):D801–7.

[24]Colley KJ, Wolfert RL, Cobble ME. Lipoprotein associated phospholipase A(2): role in atherosclerosis and utility as a biomarker for cardiovascular risk. EPMA J 2011;2(1): 27–38.

[25]Khakpour H, Frishman WH. Lipoprotein-associated phospholipase A2: an independent predictor of cardiovascular risk and a novel target for immunomodulation therapy. Cardiol Rev 2009;17(5):222–9.

[26]Glass CK, Witztum JL. Atherosclerosis. the road ahead. Cell 2001;104(4):503–16.

[27]Clish CB, Davidov E, Oresic M, Plasterer TN, Lavine G, Londo T, et al. Integrative biological analysis of the APOE*3-leiden transgenic mouse. OMICS 2004;8(1):3–13.

[28]Kleemann R, Verschuren L, van Erk MJ, Nikolsky Y, Cnubben NH, Verheij ER, et al. Atherosclerosis and liver inflammation induced by increased dietary cholesterol intake: a combined transcriptomics and metabolomics analysis. Genome Biol 2007;8(9):R200.

[29]Martin JC, Canlet C, Delplanque B, Agnani G, Lairon D, Gottardi G, et al. 1H NMR metabonomics can differentiate the early atherogenic effect of dairy products in hyperlipidemic hamsters. Atherosclerosis 2009;206(1):127–33.

[30]Jové M, Ayala V, Ramírez-Núñez O, Serrano JC, Cassanyé A, Arola L, et al. Lipidomic and metabolomic analyses reveal potential plasma biomarkers of early atheromatous plaque formation in hamsters. Cardiovasc Res 2013;97(4):642–52.

[31]Cheng KK, Benson GM, Grimsditch DC, Reid DG, Connor SC, Griffin JL. A metabolomic study of the LDL receptor null mouse fed a high-fat diet reveals profound perturba-tions in choline metabolism that are shared with ApoE null mice. Physiol Genomics 2010.

[32]Ussher JR, Lopaschuk GD, Arduini A. Gut microbiota metabolism ofL-carnitine and cardiovascular risk. Atherosclerosis 2013;231(2):456–61.

[33]Wang Z, Klipfell E, Bennett BJ, Koeth R, Levison BS, Dugar B, et al. Gutflora metabolism of phosphatidylcholine promotes cardiovascular disease. Nature 2011;472(7341): 57–63.

[34]Stöhr R, Cavalera M, Menini S, Mavilio M, Casagrande V, Rossi C, et al. Loss of TIMP3 exacerbates atherosclerosis in ApoE null mice. Atherosclerosis 2014; 235(2):438–43.

[35]Brindle JT, Antti H, Holmes E, Tranter G, Nicholson JK, Bethell HW, et al. Rapid and noninvasive diagnosis of the presence and severity of coronary heart disease using 1H-NMR-based metabonomics. Nat Med 2002;8(12):1439–44.

[36]Sun M, Gao X, Zhang D, Ke C, Hou Y, Fan L, et al. Identification of biomarkers for unstable angina by plasma metabolomic profiling. Mol Biosyst 2013;9(12):3059–67.

[37]Stübiger G, Aldover-Macasaet E, Bicker W, Sobal G, Willfort-Ehringer A, Pock K, et al. Targeted profiling of atherogenic phospholipids in human plasma and lipoproteins of hyperlipidemic patients using MALDI-QIT-TOF-MS/MS. Atherosclerosis 2012; 224(1):177–86.

[38]Shah SH, Hauser ER, Bain JR, Muehlbauer MJ, Haynes C, Stevens RD, et al. High heritability of metabolomic profiles in families burdened with premature cardiovas-cular disease. Mol Syst Biol 2009;5:258.

[39]Shah SH, Bain JR, Muehlbauer MJ, Stevens RD, Crosslin DR, Haynes C, et al. Associa-tion of a peripheral blood metabolic profile with coronary artery disease and risk of subsequent cardiovascular events. Circ Cardiovasc Genet 2010;3(2):207–14.

[40]Teul J, Rupérez FJ, Garcia A, Vaysse J, Balayssac S, Gilard V, et al. Improving metabo-lite knowledge in stable atherosclerosis patients by association and correlation of GC-MS and 1H NMRfingerprints. J Proteome Res 2009;8(12):5580–9.

[41]Rizza S, Copetti M, Rossi C, Cianfarani MA, Zucchelli M, Luzi A, et al. Metabolomics signature improves the prediction of cardiovascular events in elderly subjects. Atherosclerosis 2014;232(2):260–4.

[42]Zheng Y, Yu B, Alexander D, Couper DJ, Boerwinkle E. Medium-term variability of the human serum metabolome in the Atherosclerosis Risk in Communities (ARIC) study. OMICS 2014;18(6):364–73.

[43]Shah SH, Sun JL, Stevens RD, Bain JR, Muehlbauer MJ, Pieper KS, et al. Baseline metabolomic profiles predict cardiovascular events in patients at risk for coronary artery disease. Am Heart J 2012;163(5):844–50 [e1].

[44]Würtz P, Soininen P, Kangas AJ, Mäkinen VP, Groop PH, Savolainen MJ, et al. Characterization of systemic metabolic phenotypes associated with subclinical atherosclerosis. Mol Biosyst 2011;7(2):385–93.

[45]Holmes E, Loo RL, Stamler J, Bictash M, Yap IK, Chan Q, et al. Human metabolic phenotype diversity and its association with diet and blood pressure. Nature 2008;453(7193):396–400.

[46]Bernini P, Bertini I, Luchinat C, Tenori L, Tognaccini A. The cardiovascular risk of healthy individuals studied by NMR metabonomics of plasma samples. J Proteome Res 2011;10(11):4983–92.

[47]Camont L, Lhomme M, Rached F, Le Goff W, Nègre-Salvayre A, Salvayre R, et al. Small, dense high-density lipoprotein-3 particles are enriched in negatively charged phos-pholipids: relevance to cellular cholesterol efflux, antioxidative, antithrombotic,

anti-inflammatory, and antiapoptotic functionalities. Arterioscler Thromb Vasc Biol 2013;33(12):2715–23.

[48]Stegemann C, Drozdov I, Shalhoub J, Humphries J, Ladroue C, Didangelos A, et al. Comparative lipidomics profiling of human atherosclerotic plaques. Circ Cardiovasc Genet 2011;4(3):232–42.

[49]Sabatine MS, Liu E, Morrow DA, Heller E, McCarroll R, Wiegand R, et al. Metabolomic identification of novel biomarkers of myocardial ischemia. Circulation 2005;112(25): 3868–75.

[50]Zimmerli LU, Schiffer E, Zürbig P, Good DM, Kellmann M, Mouls L, et al. Urinary proteomic biomarkers in coronary artery disease. Mol Cell Proteomics 2008;7(2): 290–8.

[51]Delles C, Schiffer E, von Zur Muhlen C, Peter K, Rossing P, Parving HH, et al. Urinary proteomic diagnosis of coronary artery disease: identification and clinical validation in 623 individuals. J Hypertens 2010;28(11):2316–22.

[52]von zur Muhlen C, Schiffer E, Sackmann C, Zürbig P, Neudorfer I, Zirlik A, et al. Urine proteome analysis reflects atherosclerotic disease in an ApoE−/− mouse model and al-lows the discovery of new candidate biomarkers in mouse and human atherosclerosis. Mol Cell Proteomics 2012;11(7) [M111.013847].

[53]Mayr M, Chung YL, Mayr U, Yin X, Ly L, Troy H, et al. Proteomic and metabolomic analyses of atherosclerotic vessels from apolipoprotein E-deficient mice reveal alterations in inflammation, oxidative stress, and energy metabolism. Arterioscler Thromb Vasc Biol 2005;25(10):2135–42.

[54]Donahue MP, Rose K, Hochstrasser D, Vonderscher J, Grass P, Chibout SD, et al. Discovery of proteins related to coronary artery disease using industrial-scale proteomics analysis of pooled plasma. Am Heart J 2006;152(3):478–85.

[55]Dardé VM, de la Cuesta F, Dones FG, Alvarez-Llamas G, Barderas MG, Vivanco F. Analysis of the plasma proteome associated with acute coronary syndrome: does a permanent protein signature exist in the plasma of ACS patients? J Proteome Res 2010;9(9): 4420–32.

[56]Spijkers LJ, van den Akker RF, Janssen BJ, Debets JJ, De Mey JG, Stroes ES, et al. Hypertension is associated with marked alterations in sphingolipid biology: a po-tential role for ceramide. PLoS One 2011;6(7):e21817.

[57]Cuvillier O, Pirianov G, Kleuser B, Vanek PG, Coso OA, Gutkind S, et al. Suppression of ceramide-mediated programmed cell death by sphingosine-1-phosphate. Nature 1996;381(6585):800–3.

[58]Zhang QJ, Holland WL, Wilson L, Tanner JM, Kearns D, Cahoon JM, et al. Ceramide mediates vascular dysfunction in diet-induced obesity by PP2A-mediated dephosphor-ylation of the eNOS-Akt complex. Diabetes 2012;61(7):1848–59.

[59]Graessler J, Schwudke D, Schwarz PE, Herzog R, Shevchenko A, Bornstein SR. Top-down lipidomics reveals ether lipid deficiency in blood plasma of hypertensive patients. PLoS One 2009;4(7):e6261.

[60]Kulkarni H, Meikle PJ, Mamtani M, Weir JM, Barlow CK, Jowett JB, et al. Plasma lipidomic profile signature of hypertension in Mexican American families: specific role of diacylglycerols. Hypertension 2013.

[61]Wang Y, Deng X, Hewavitharana T, Soboloff J, Gill DL. Stim, ORAI and TRPC channels in the control of calcium entry signals in smooth muscle. Clin Exp Pharmacol Physiol 2008;35(9):1127–33.

[62]Wikoff WR, Frye RF, Zhu H, Gong Y, Boyle S, Churchill E, et al. Pharmacometabolomics reveals racial differences in response to atenolol treatment. PLoS One 2013;8(3): e57639.

[63]Zheng Y, Yu B, Alexander D, Mosley TH, Heiss G, Nettleton JA, et al. Metabolomics and incident hypertension among blacks: the atherosclerosis risk in communities study. Hypertension 2013;62(2):398–403.