This study was conducted with financial support from Amarin Pharma, Inc.

Acknowledgements

17. Bays HE, Ballantyne CM, Braeckman RA, et al. Am J Cardiovasc Drugs. 2013; 13:37-46.18. Bays HE, Ballantyne CM, Kastelein JJ, et al. Am J Cardiol. 2011; 108:682-690.19. Braeckman RA, Manku MS, Bays HE, et al. Prostaglandins Leukot Essent Fatty Acids. 2013; 89:195-201.20. Satoh N, Shimatsu A, Kotani K, et al. Diabetes Care. 2007; 30:144-146.21. Satoh-Asahara N, Shimatsu A, Sasaki Y, et al. Diabetes Care. 2012; 35:2631-2639.22. Cawood AL, Ding R, Napper FL, et al. Atherosclerosis. 2010; 212:252-259.23. Dawczynski C, Martin L, Wagner A, et al. Clin Nutr. 2010; 29:592-599.24. El-Saadani M, Esterbauer H, El-Sayed M, et al. J Lipid Res. 1989; 30:627-630.25. Mason RP, Walter MF, Day CA, et al. J Biol Chem. 2006; 281:9337-9345.26. Bangham AD, Standish MM, Watkins JC. J Mol Biol. 1965; 13:238-252.27. Herbette L, DeFoor P, Fleischer S, et al. Biochim Biophys Acta. 1985; 817:103-122.

1. Baynes JW. Diabetes. 1991; 40:405-412. 2. Laakso M. Diabetes. 1999; 48:937-942. 3. Lamharzi N, Renard CB, Kramer F, et al. Diabetes. 2004; 53:3217-3225. 4. Ceriello A, Motz E. Arterioscler Thromb Vasc Biol. 2004; 24:816-823. 5. Pennathur S, Heinecke JW. Antioxid Redox Signal. 2007; 9:955-969. 6. Hunt JV, Dean RT, Wolff SP. Biochem J. 1988; 256:205-212. 7. Jacob RF, Mason RP. J Biol Chem. 2005; 280:39380-39387. 8. Mason RP, Jacob RF. Circulation. 2003; 107:2270-2273. 9. Mason RP, Tulenko TN, Jacob RF. Biochim Biophys Acta. 2003; 1610:198-207.10. Self-Medlin Y, Byun J, Jacob RF, et al. Biochim Biophys Acta. 2009; 1788:1398-1403.11. Small DM. Arterioscler Thromb Vasc Biol. 1988; 8:103-129.12. Crea F, Liuzzo G. J Am Coll Cardiol. 2013; 61:1-11.13. Tulenko TN, Chen M, Mason PE, et al. J Lipid Res. 1998; 39:947-956.14. Phillips JE, Geng YJ, Mason RP. Atherosclerosis. 2001; 159:125-135.15. Yokoyama M, Origasa H, Matsuzaki M, et al. Lancet. 2007; 369:1090-1098.16. Ballantyne CM, Bays HE, Kastelein JJ, et al. Am J Cardiol. 2012; 110:984-992.

References

EPA inhibited hyperglycemia-induced changes in membrane structure through a potent antioxidant mechanism that is related to its specific physico-chemical properties and that is enhanced when combined with ATM. The favorable effects of EPA on membrane cholesterol crystal formation and oxidative stress pathways with hyperglycemia may contribute to mechanisms of atheroprotection.

Conclusion

Summary

Highly-ordered cholesterol crystalline domains were observed to form in vehicle-treated membrane samples as a function of oxidative stress with hyperglycemia. These domains consisted exclusively of free cholesterol molecules organized as a bilayer with a width of 34 Å (Fig. 7).

EPA reduced LOOH formation by 88% and 86% at 72 and 96 hr, respectively, as compared to glucose treatment alone (p<0.001) and blocked cholesterol domain formation at all experimental time points (Fig. 7). Vitamin E had no significant effect on lipid peroxidation or cholesterol domain formation at these time points.

In the absence of enzymatic processes, the antioxidant activity of EPA is attributed to free radical scavenging mechanisms associated with its multiple conjugated double bonds and high lipophilicity.

The inhibitory effects of EPA on LOOH formation were increased by >60% (p<0.001) when combined with ATM (data not shown) and could not be reproduced with other TG-lowering agents, including fenofibrate, gemfibrozil, and niacin.

24 hr24 hr24 hr24 hr24 h24 h 48 hr48 hr48 hr48 hr48 h48 h 72 hr72 hr72 hr72 hr72 h72 h

Increasing Exposure to Oxidative Stress

+ EPA

34 Å

48 Å

52 Å

CCCCCooooonnnntttrrroooooll

Figure 7. Schematic summary of the antioxidant and membrane structural effects of EPA as determined in this study.

Results

Figure 6. Comparative effects of EPA, fenofibrate (Fenofib), niacin, and gemfibrozil (Gemfib) on glucose-induced membrane lipid peroxidation. Each agent was tested at 10.0 μM following exposure to oxidative conditions for 48 hr. Values are mean ± S.D. (N = 6). *p<0.05 and **p<0.01 versus vehicle-treated control; †p<0.05 versus glucose- treated control; §p<0.001 versus all other TG-lowering treatments (Student- Newman-Keuls multiple comparisons test; overall ANOVA: p<0.0001, F=9.679).

The Antioxidant Effects of EPA Were Superior to Those of

Other Triglyceride-Lowering Agents: Fenofibrate, Niacin,

and Gemfibrozil

Figure 5. Relative cholesterol domain peak intensity values were derived by integrating the second-order cholesterol domain peak and normalizing to total peak area associated with a given diffraction pattern. Values are mean ± S.D. (N = 3). *p<0.05 and p<0.01 versus control and vitamin E treatments, respectively (Student- Newman-Keuls multiple comparisons test; overall ANOVA: p=0.0075, F=8.849).

Quantitative Assessment of the Effects of EPA and

Vitamin E on Glucose- and Peroxidation-Induced

Cholesterol Domain Formation

Figure 4. Representative x-ray diffraction patterns collected from model membranes prepared in the presence of glucose, treated with vehicle (control), vitamin E, or EPA, and subjected to oxidative conditions for 96 hr. At 0 hr, each sample exhibited a single lipid bilayer phase with an average periodicity (d-space value) of 51.5 Å, represented by diffraction peaks 1 through 4. At 72 hr, cholesterol crystalline domains, having a characteristic d-space value of 34 Å and represented by a set of distinct diffraction peaks (shown in red fill), were also observed in control and vitamin E-treated membrane samples. At 96 hr, cholesterol domains peaks were observed in all experimental samples; however, these peaks were disproportionately greater in control and vitamin E-treated samples. Changes in membrane lipid structural organization associated with each treatment and time point are depicted schematically in the insets appended to each panel. Phospholipid and cholesterol molecules are shown in gray and red, respectively. Following oxidative damage to the membrane, cholesterol forms discrete bilayers that are of reduced width as compared to the surrounding phospholipid bilayer.

EPA Blocked Glucose-Induced Cholesterol Domain Formation in Membranes Exposed to Oxidative Conditions

Figure 3. Comparative effects of vitamin E and EPA on glucose-induced membrane lipid peroxidation. *p<0.05 versus control; †p<0.05 versus vitamin E treatment (Student-Newman-Keuls multiple comparisons test; overall ANOVA—0 hr data: p=0.0073, F=12.474; 72 hr data: p=0.0204, F=7.986; 96 hr data: p=0.0008, F=29.764). Values are mean ± S.D. (N = 3).

EPA Had Superior Antioxidant Effects as Compared to

Vitamin E

Lipid Peroxidation AnalysisAll lipid membrane samples were subjected to time-dependent autoxidation by incubating at 37°C in an uncovered water bath for 72 hr in the absence and presence of various treatments.

Lipid hydroperoxide (LOOH) formation was measured using an iodometric assay, as described by el-Saadani et al.12 This method is based on the oxidation of iodide (I–) by LOOH to form triiodide anion (I3

– ), which is directly proportional to the amount of lipid hydroperoxides present in the membrane sample. I3

– was measured spectrophotometrically based on its molar absorptivity value () of 2.46 × 104 M-1 cm-1 at 365 nm.

X-ray Diffraction AnalysisThe membrane structural effects of glucose and the various compounds examined in this study were measured at 0, 72, and 96 hr intervals. Membrane samples were oriented and analyzed using small-angle x-ray diffraction as previously described.27

The presence of cholesterol domains in a given membrane sample results in the production of distinct Bragg (diffraction) peaks, typically referred to as first- and second-order cholesterol domain peaks, that have singular periodicity (d-space) values of 34 and 17 Å, respectively (Fig. 2).25

Calculations and Statistical AnalysesData are presented as mean ± SD for (n) separate samples or experiments. Differences between groups were analyzed using the unpaired, two-tailed Student’s t-test (for comparisons between only two groups) or ANOVA followed by Student-Newman-Keuls multiple comparisons post hoc analysis (for comparisons between three or more groups). Only differences with probability values less than 0.05 were considered significant.

Materials1,2-Dilinoleoyl-sn-glycero-3-phosphocholine (DLPC; 18:2; n-6) and cholesterol were purchased from Avanti Polar Lipids, Inc. (Alabaster, AL) and dissolved in HPLC-grade chloroform.

cis-5,8,11,14,17-Eicosapentaenoic acid (EPA) was purchased from Sigma- Aldrich (St. Louis, MO). Stock solutions were prepared in ethanol under nitrogen atmosphere and stored at –20°C. Vitamin E (-tocopherol) was also purchased from Sigma-Aldrich and prepared in ethanol at 1.0 mM ( = 3.06 × 104 M-1 cm-1 at 294 nm) just prior to experimental use. Atorvastatin ortho- (o-) hydroxy (active) metabolite was purchased from Toronto Research Chemicals (North York, Ontario, Canada) and solubilized in methanol to 1.0 mM. Fenofibrate, nicotinic acid (niacin), and gemfibrozil were purchased from Sigma-Aldrich and solubilized in ethanol to 1.0 mM. All test compounds were further diluted in ethanol or aqueous buffer as needed.

Glucose was prepared in saline buffer (0.5 mM HEPES, 154 mM NaCl, pH 7.3) at 11.0 mM (200 mg/dL).

CHOD-iodide color reagent (stock) was prepared, with slight modification, as described by El-Saadani et al.12 and consisted of 0.2 M K2HPO4, 0.12 M KI, 0.15 mM NaN3, 10 μM ammonium molybdate, and 0.1 g/L benzalkonium chloride. Prior to experimental use, the CHOD reagent was activated by adding 24 μM EDTA, 20 μM butylated hydroxytoluene (BHT), and 0.2% Triton X-100.

Preparation of Membrane Lipid VesiclesMultilamellar vesicles (MLVs) were prepared from binary mixtures of DLPC and cholesterol, at a cholesterol-to-phospholipid (C/P) mole ratio of 0.6:1, as previously described.25 Component lipids (in chloroform) were transferred to borosilicate culture tubes and combined with vehicle (ethanol) or an equal volume of EPA, vitamin E, or ATM stock solution, adjusted to achieve desired treatment concentrations. Samples were then shell-dried under nitrogen gas and placed under vacuum for 3 hr to remove residual solvent.

After drying, each sample was resuspended in saline buffer (0.5 mM HEPES, 154 mM NaCl, pH 7.3) to yield final phospholipid concentrations of 1.0 or 2.5 mg/mL (for lipid peroxidation or x-ray diffraction analysis, respectively). Lipid suspensions were then vortexed for 3 min at ambient temperature to form MLVs.26

Methods

Figure 2. Schematic illustration of a typical diffraction pattern collected from a biphasic membrane sample, exhibiting sterol-poor, phospholipid bilayer domains (peaks 1, 2, and 4) and cholesterol crystalline domains (peaks 1 and 2), and its relationship to the spatial arrangement of these domains in a representative membrane bilayer.

CholesterolDomain

Cholesterol DomainDiffraction Orders (34.0 Å)

1

2

4

1́2́

0.01 0.02 0.03 0.04 0.05 0.06 0.07

3.2

3.6

4.0

4.4

4.8

5.2

Phot

ons

(Log

Sca

le)

Space-1 (Å-1)

0.04 0.05 0.06

cecc -1 (Å-ÅÅ1)

34.0 Å

We hypothesized that EPA inhibits glucose-induced lipid oxidation and cholesterol crystalline domain formation in membrane vesicles enriched in polyunsaturated fatty acids and cholesterol.

Hypothesis

The goal of this study was to test the ability of EPA to interfere with the effects of glucose on lipid peroxidation and cholesterol crystalline domain formation in membrane vesicles enriched with poly- unsaturated fatty acids. These effects were compared to those of vitamin E alone or combination treatment with EPA and atorvastatin o-hydroxy (active) metabolite (ATM). Changes in membrane lipid organization were evaluated using small angle x-ray diffraction. The antioxidant effects of EPA were also compared to those of other TG-lowering agents, including fenofibrate, niacin, and gemfibrozil.

Objective

Purified EPA has been shown to have beneficial effects on markers of inflammation and to significantly reduce triglycerides along with reduced high-sensitivity C-reactive protein (hsCRP), lipoprotein-associated phospholipase A2 (Lp-PLA2), arachidonic acid/EPA (AA/EPA) plasma ratio, and oxidized LDL levels as compared to placebo.16-21

EPA incorporates readily into advanced atherosclerotic plaques where it is believed to reduce inflammation, improve plaque stability, and decrease endothelial activation.22,23

The basis for broader vascular protection with EPA is not well understood but may be related to increased antioxidant activity associated with its conjugated double bonds, which are capable of quenching higher singlet oxygen. EPA is also highly lipophilic and readily intercalates into liposomes and cell membranes.

Hyperglycemia and hyperlipidemia have been shown to increase the risk of cardiovascular disease in patients with diabetes mellitus.1-3 Acute hyperglycemia is an independent risk factor for cardiovascular disease even in non-diabetic subjects.4,5

Under conditions of oxidative stress, glucose can undergo autoxidation to form glucose radicals or other reactive oxygen species, which readily react with cellular components, including membrane proteins and phospholipids.5,6 In membranes enriched with polyunsaturated fatty acids, glucose has been shown to accelerate changes in lipid structural organization associated with oxidative stress, including the formation of cholesterol crystalline domains.7-10

The unstable atherosclerotic lesion is characterized by extracellular lipid deposits containing cholesterol (both free and esterified), phospholipids, and triacylglycerol.11,12 Membrane-associated cholesterol crystals have been characterized in cell culture systems and tissue explants using electron microscopy.13,14 Cholesterol crystalline domains are believed to contribute to mechanisms of cell death and inflammation during atherosclerosis (Fig. 1).8

Eicosapentaenoic acid (EPA) is an omega-3 fatty acid (20:5; n-3) approved for patients with severe hypertriglyceridemia. In the Japan EPA Lipid Intervention Study (JELIS), EPA was shown to be effective in preventing coronary artery disease in hypercholesterolemic patients receiving statin treatment.15

Background

Figure 1. Schematic diagram of changes in lipid raft structure and cell function during cholesterol enrichment and atherosclerosis. Subtypes of lipid rafts enriched with sphingolipid (blue) and cholesterol (red) include caveolae () that contain caveolin protein (green) and detergent-resistant membrane domains (). With progressive cholesterol accumulation, separate cholesterol crystalline membrane domains () form and precede the development of extracellular cholesterol crystals (), which contribute to mechanisms of cell injury and death, including apoptosis. Cholesterol enrichment also increases the number of membrane caveolae, leading to inhibition of endothelial nitric oxide (eNOS) following by a reduction in nitric oxide (NO) production and associated vascular benefits. Loss of normal membrane structure and function with cholesterol enrichment is also associated with disruptions in calcium regulation and redox potential.

eNO

S

eNOS eNOS

eNO

S

eNOS

eNOS

eNOS

eNOS

L-Arginine

eNOS

NODecreased NO

Production

eNOS

eNOS

NO

Dysregulation ofVascular Tone

Ca2+

Inflammation

Apoptosis

1

Ca2+

L-Arginine

2 3

4

Disease Progression Atherosclerosis

×

×

NAD(P)H Oxidase

Redox SensitiveProteins

O2–

Synopsis: Oxidative damage to polyunsaturated fatty acids (PUFAs) leads to endothelial dysfunction, inflammation, and foam cell formation during atherogenesis. At high levels, glucose contributes to PUFA oxidation and membrane reorganization, including the development of cholesterol crystalline domains.

Objectives/Purpose: In this study, we tested the comparative effects of eicosapentaenoic acid (EPA), an omega-3 fatty acid, and vitamin E, a chain-blocking antioxidant, on glucose- induced changes in membrane lipid peroxidation and structural organization in lipid vesicles containing PUFAs.

Methods: Iodometric approaches were used to measure the effects of EPA on lipid hydroperoxide (LOOH) formation in membranes composed of cholesterol and PUFAs and treated with glucose at 200 mg/dL. Membrane samples were prepared at a C/P mole ratio of 0.6 and subjected to autoxidation at 37°C for 72-96 hr in the absence or presence of EPA or vitamin E. We also measured the antioxidant activity of EPA in combination with atorvastatin o-hydroxy (active) metabolite (ATM). Small angle x-ray diffraction was used to measure the effects of glucose-induced oxidation and pharmacologic intervention on membrane lipid structural organization.

Results: Glucose treatment significantly increased LOOH formation and cholesterol domain formation as compared to untreated controls. EPA reduced LOOH formation by 88% and 86% at 72 and 96 hr, respectively, as compared to glucose treatment alone (p<0.001) and blocked cholesterol domain formation at all experimental time points. Vitamin E had no significant effect on lipid peroxidation or cholesterol domain formation at these time points. The inhibitory effects of EPA on LOOH formation were increased by >60% (p<0.001) when combined with ATM.

Conclusion: These data demonstrate that EPA inhibits hyperglycemia-induced changes in membrane structure through a potent antioxidant mechanism that is related to its specific physico-chemical properties and that is enhanced when combined with ATM.

Abstract

Eicosapentaenoic Acid Inhibited Glucose-Induced Membrane Lipid Peroxidation and Cholesterol Crystalline Domain Formation

R. Preston Mason, Ph.D.1,2, Robert F. Jacob, Ph.D.2

1Brigham and Women’s Hospital, Harvard Medical School, Boston, MA, USA; 2Elucida Research LLC, Beverly, MA, USA