Omega 3 Improves Both apoB100-containing Lipoprotein Turnover and their Sphingolipid Profile in Hypertriglyceridemia

Abstract

Context

Evidence for an association between sphingolipids and metabolic disorders is increasingly reported. Omega-3 long-chain polyunsaturated fatty acids (n-3 LC-PUFAs) improve apolipoprotein B100 (apoB100)-containing lipoprotein metabolism, but their effects on the sphingolipid content in lipoproteins remain unknown.

Objectives

In subjects with hypertriglyceridemia, we analyzed the effect of n-3 LC-PUFAs on the turnover apoB100-containing lipoproteins and on their sphingolipid content and looked for the possible association between these lipid levels and apoB100-containing lipoprotein turnover parameters.

Methods

Six subjects underwent a kinetic study before and after n-3 supplementation for 2 months with 1 g of fish oil 3 times day containing 360 mg of eicosapentaenoic acid (EPA) and 240 mg of docosahexaenoic acid (DHA) in the form of triglycerides. We examined apoB100-containing lipoprotein turnover by primed perfusion labeled [5,5,5-²H₃]-leucine and determined kinetic parameters using a multicompartmental model. We quantified sphingolipid species content in lipoproteins using mass spectrometry.

Results

Supplementation decreased very low-density lipoprotein (VLDL), triglyceride, and apoB100 concentrations. The VLDL neutral and polar lipids showed increased n-3 LC-PUFA and decreased n-6 LC-PUFA content. The conversion rate of VLDL1 to VLDL2 and of VLDL2 to LDL was increased. We measured a decrease in total apoB100 production and VLDL1 production. Supplementation reduced the total ceramide concentration in VLDL while the sphingomyelin content in LDL was increased. We found positive correlations between plasma palmitic acid and VLDL ceramide and between VLDL triglyceride and VLDL ceramide, and inverse correlations between VLDL n-3 LC-PUFA and VLDL production.

Conclusion

Based on these results, we hypothesize that the improvement in apoB100 metabolism during n-3 LC-PUFA supplementation is contributed to by changes in sphingolipids

n-3 long chain polyunsaturated fatty acid, hypertriglyceridemia, apolipoprotein B100, sphingomyelin, ceramide, kinetic analysis, modeling

High low-density lipoprotein (LDL) cholesterol (1), low high-density lipoprotein (HDL) cholesterol, and hypertriglyceridemia (2-4) are strong biomarkers of cardiovascular disease; however, these risk factors account for only a fraction of atherosclerotic complications (5). Other bioactive lipids, such as sphingolipids, although less abundant, are involved as signaling molecules in various biological processes and are related to several metabolic and chronic diseases (6). De novo synthesis of sphingolipids begins in the endoplasmic reticulum via serine-palmitoyl transferase (SPT) (7) but data on their plasma transport, especially in lipoprotein, are very scarce. Labelling de novo biosynthesized sphingolipid pathways showed the amount of sphingomyelin label and mass in very low-density lipoprotein (VLDL) was small, suggesting that most of the sphingomyelin that is in LDL may be from sources other than synthesis within the hepatocyte, likely red blood cells and the plasma membranes rich in sphingomyelins (8).

Microsomal triglyceride transfer protein (MTP) is required for the assembly of apolipoprotein B100 (apoB100)-containing lipoproteins and their secretion into the blood circulation (9). Sphingolipids synthesized in the liver are likely incorporated into VLDL by MTP, then recovered in intermediate-density lipoprotein (IDL) and in LDL after VLDL triglyceride lipolysis.

Iqbal et al. have shown that plasma ceramide and sphingomyelin concentrations were significantly reduced in patients with abetalipoproteinemia with MTP deficiency, suggesting that MTP might regulate plasma secretion of ceramides and sphingomyelins (9). Other studies have shown that excess of cellular palmitate activates the de novo biosynthesis pathway of sphingolipids (8, 10) and increases the formation of diacylglycerols and ceramides in C2C12 muscle cells (11). Choline deficiency or incubation of the cells with fumonisin caused a reduction in the synthesis and secretion of sphingolipids (8). Ceramides induce insulin resistance by antagonizing insulin signaling (11-13) and are associated with inflammation (14) whereas, sphingomyelins, the major sphingolipid components of LDL (15), are considered to be proatherogenic (16-19). Inhibition of SPT activity leads to reduced circulating concentration of ceramides, dyslipidemia, and insulin resistance (20, 21).

Omega-3 long-chain polyunsaturated fatty acids (n-3 LC-PUFAs) improve triacylglycerol metabolism (22-24) and affect the availability of sphingolipids by moderating the rate-limiting step in their de novo synthesis (25). We have shown that n-3 LC-PUFA supplementation decreased the concentrations of ceramides and their precursors in the liver and muscles of hamsters fed a high-fat diet supplemented with n-3 LC-PUFAs. We also found a positive correlation between fasting blood glucose concentration or hepatic triacylglycerol secretion and hepatic ceramide concentration (26).

Based on these findings and on the probable role of liver lipoprotein secretion in the plasma concentrations of ceramides and sphingomyelins, we hypothesized that n-3 LC-PUFA supplementation impacts both the secretion of apoB100-containing lipoproteins and their sphingolipid content. We also hypothesized that the plasma concentrations of these bioactive lipids would be correlated with apoB100 turnover parameters in subjects with hypertriglyceridemia. Therefore, we studied the effect of n-3 LC-PUFA supplementation on the metabolism of apoB100-containing lipoproteins by kinetic and modeling studies and by analyzing their sphingolipid content.

Material and Methods

Subjects

As a pilot study, 6 male subjects with hypertriglyceridemia were recruited to this clinical intervention for 8 weeks (Fig. 1). None of the study subjects had diabetes mellitus, proteinuria, or hypothyroidism nor had been taking any medication or herbal drugs that could affect carbohydrate or lipid metabolism. Patients were advised not to change their eating habits from the inclusion visit into the protocol until the end of the study, and especially to limit their consumption of fish to once per week and alcohol consumption to 2 glasses of wine per day or equivalent. Selected relevant clinical characteristics of the subjects are presented in Table 1. This randomized clinical protocol was approved by the Ethics Committee of Nantes University Hospital, and it was registered by AFFSAPS as no. 060575-35. Written informed consent was obtained from each subject.

Table 1.

Clinical characteristics at baseline and after fish oil supplementation for 8 weeks

	Baseline (n = 6)	n-3 LC-PUFA (n = 6)
Age (years)	48,33 ± 11,04	48,33 ± 11,04
Body mass index (kg/m2)	26.9 ± 2.3	26.9 ± 2.3
Plasma triglycerides (g/L)	2.66 ± 0.76	1.74 ± 0.55 *
Plasma cholesterol (g/L)	2.18 ± 0.23	2.08 ± 0.37
LDL cholesterol (g/L)	1.22 ± 0.34	1.34 ± 0.33
HDL cholesterol (g/L)	0.42 ± 0.09	0.40 ± 0.09
ApoAI (g/L)	1.44 ± 0.23	1.34 ± 0.23
ApoB100 (g/L)	1.21 ± 0.25	1.150.25

Data are presented as mean ± SD.

Abbreviations: n-3 LC-PUFA, omega-3 long-chain polyunsaturated fatty acid.

*P value statistically significant at P < 0.05.

Figure 1.

Flowchart of the n-3 long-chain polyunsaturated fatty acids (n-3 LC-PUFA) after supplementation trial illustrating the participants including in the present study.

Study design

The protocol was similar to that described in a previous study (23). Briefly, patients underwent a basal kinetic study (described below) the day before the start of the treatment with n-3 LC-PUFAs. They received 1 soft capsule 3 times a day containing 1 g of fish oil. Each capsule consisted of saturated 60% omega-3 polyunsaturated fatty acids, namely 360 mg of eicosapentaenoic acid (EPA) (36%) and 240 mg of docosahexaenoic acid (DHA) (24%) in the form of triglycerides. This dose of EPA and DHA has been set to achieve nutritional adequacy for these fatty acids according to the recommendation provided by many health agencies worldwide (27). Indeed, the evidence base supports a dietary recommendation of 500 mg/day of EPA and DHA for cardiovascular disease risk reduction. As patients are advised to eat fish once a week, omega-3 supplementation provided in this study should bring intakes closer to the recommended level.

The kinetic protocol was conducted as previously described (23). Briefly, the endogenous labeling of apoB100 was carried out by constant infusion of [5,5,5-²H₃]-leucine in subjects fasted overnight for 12 hours prior to the study and remained fasting during the entire procedure. Each subject received intravenously a bolus of 10 μmol/kg tracer immediately followed by a constant tracer infusion (10 μmol/kg per hour) for 14 hours. Venous blood samples were drawn into EDTA tubes (Venoject, Paris, France) at baseline and at 15, 30, and 45 minutes, 1, 1.5, 2, and 2.5 hours, and then hourly until 14 hours. Sodium azide, an inhibitor of bacterial growth, and Pefabloc SC (Interchim, Montluçon, France), a protease inhibitor, were added to blood samples at a final concentration of 1.5 and 0.5 mmol/L, respectively.

Analytical procedures

Isolation and measurement of enrichment of lipoprotein containing apoB100

Isolation of lipoproteins and measurement of leucine enrichment in apoB100 have previously been described (23). Briefly, lipoproteins were separated by sequential ultracentrifugation, and apoB100 was isolated by sodium dodecylsulfate polyacrylamide gel electrophoresis. Apolipoproteins bands were dried under a nitrogen vacuum and then hydrolyzed. The amino acids were purified by cation exchange chromatography and then esterified and derivatized. Electron-impact gas chromatography-mass spectrometry (GC-MS) was performed on a 5891 A gas chromatograph connected to a 5971 A quadrupole mass spectrometer (Hewlett Packard Co., Palo Alto, CA, USA). The isotopic ratio was determined by selected ion monitoring at m/z 282 and 285. Calculations of apoB100 kinetic parameters were based on the tracer:tracee mass ratio.

Modeling

Kinetic analysis of tracer:tracee ratios of apoB100 was achieved using computer software for simulation, analysis, and modeling and the model (Fig. 2) previously developed was applied (23). In this model, a forcing function, corresponding to the time course of plasma leucine enrichment, was used to drive the appearance of leucine tracer in the apoB100 of the different lipoprotein fractions. Since hypertriglyceridemia is accompanied by increased heterogeneity within VLDLs, we looked at VLDL1 and VLDL2 for a better fit and more accurate calculation of VLDL turnover constants. ApoB100 enters into plasma through VLDL secretion and the direct production of IDL and LDL. Direct removal of apoB100 occurs from VLDL1 (k(0,10)), VLDL2 (k(0,20)), VLDL remnant (VLDLR) (k(0,11)), IDL (k(0,30)), and LDL (k(0,40)). ApoB100 transfer to IDL and LDL occurs by conversion for VLDL1 by lipolysis and transfer cascade (k(20,10), k(11,10)), VLDL2 (k(30,20), k(40,20)), and IDL (k(40,30)). The fractional catabolic rate (FCR) corresponds to the sum of constant turnover of each compartment. The methods provided identified values and standard deviations as obtained by iterative least squares fitting for individual kinetic parameters. The use of more complex models did not provide a significant improvement in the fitting from the F test and Akaike information criterion (28).

Figure 2.

Model of apolipoprotein B100-containing lipoprotein metabolism. VLDL, very low density lipoprotein; VLDL R, very low-density lipoprotein Remnant; IDL, intermediate density lipoprotein; LDL, low-density lipoprotein. Details are given in the “Methods.” This model provides a better fit than the same model that does not include the VLDL1 remnant compartment and the VLDL2 to LDL delipidation shunt as shown by the Akaike information criterion value respectively of 1.29 and 2.40.

As all our patients were overweight, pools of apoB100 in the plasma or in VLDL, IDL, and LDL were calculated by multiplying the apoB100 concentration by 0.037 (l/kg), assuming a plasma volume of 3.7% of body weight (29). The apoB100 production rate in mg/kg per hour represented the product of the fractional catabolic rate and pool size of apoB100 in the lipoprotein fractions.

Sphingolipid analysis

Sample preparation was described in a previous work (26). Briefly, internal standards (C17 ceramide and sphingomyelin were added to 100 µL of lipoproteins. Lipids were twice extracted using an acidified isopropanol/cyclohexane mixture (40:60, v/v, 0.6% formic acid) then acidified methanol with formic acid and spun at 3000 rpm for 10 minutes at 4°C. The upper phase was dried under nitrogen at room temperature. All samples were diluted into 150 µL of acidified isopropanol/cyclohexane mixture.

Sphingolipid were separated by an NH₂ column (SPE) and fractions were collected with a different elution mixture following the adapted method described by Bodennec et al. (30). Briefly, the fraction of ceramides was collected with 1.6 mL of CHCl₃/MeOH (23:1 v/v). A 2-mL bolus of CHCl₃/MeOH (2:1 v/v) allowed one to elute the sphingomyelins in the next fraction. These 2 fractions were dried under nitrogen and removed in an ultra performance liquid chromatography (UPLC) mobile phase mixture B.

The identification of different sphingolipids was performed as previously described (26). The different extracted fractions of sphingolipids were analyzed on a Waters Xevo TQD triple quadrupole mass spectrometer combined with an Acquity H-Class UPLC system (Waters Corporation, Milford, USA). A Waters C18 BEH column (2.1 mm × 50 mm, 1.7-µm particles) equipped with a guard column was used for ceramide and sphingomyelin analyses. The column and the autosampler temperature were maintained respectively at 43°C and at 10°C. The nonlinear separation gradient used 5 mM ammonium formate in water (0.1% formic acid) as solvent A and 5 mM ammonium formate in MeOH (0.1% formic acid) as solvent B. Mobile phase flow rate was maintained at 0.6 mL/minute. Two microliters of sample was injected into the UPLC system. The ESI⁺ was operated at 150°C. Nitrogen was used as the desolvatation gas with a flow rate of 650 L/hour at 350°C, and a capillary voltage set at 3.2 kV. The extractor voltage was fixed to 3 V and the radiofrequency voltage to 2.5 V. The cone voltage varied from 18 to 38 V, depending on the sphingolipid investigated. Argon was used as the collision gas, and collision energies varied from 10 to 50 eV. Detection was done by multiple reaction monitoring (MRM). Integration and quantification were performed using the Waters TargetLinks™ and QuantLinks™ software.

Fatty acid analysis

The fatty acid profile was determined in plasma and lipoproteins. To extract lipid, 2 mL of a CHCl₃/MeOH (1:1 v/v) mixture was added to 200 µL of plasma or lipoproteins, vortexed, and stirred for 1 hour. Then 1 mL of 0.9% NaCl was added and centrifuged at 2000 rpm for 10 minutes at 4°C. The lower phase was removed, dried under nitrogen at room temperature, and diluted into 100 µL of CHCl₃.

Lipid fractions were isolated using a NH₂ column (SPE) and neutral lipids were eluted by 2 mL of chloroform and 2 mL of acetone in fraction 1; then polar lipids were eluted by 1 mL of MeOH/CHCl₃ (6:1 v/v) and 1 mL of sodium acetate 0.05 M with MeOH/CHCl₃ (6:1 v/v) in fraction 2. The 2 fractions were dried under nitrogen and dissolved into 1 mL of H₂O and 1 mL of ethanol. A 250-µL bolus of 1 M NaOH was added and incubated in a water bath for 16 hours at 95°C. Free fatty acids (FFAs) were then extracted twice with 3 mL of hexane and dried under N_2. A derivatization reaction was performed after addition of 25 µL of pentafluorobenzylbromide (50% in acetonitrile) and 25 µL of N,N-diisopropylethylamide (50% in acetonitrile) and incubation for 20 minutes at room temperature. The derivatization reagent was dried and FFAs were resuspended in hexane. Two microliters of extracted and derivative samples were injected into GC-MS (Agilent Technologies 5973i quadrupole mass spectrometer coupled with a 6890 gas chromatograph system, USA) at 250°C in splitless mode using a Hewlett Packard capillary column (HP-5MS, 30 m × 0.25 mm i.d. × 0.25 µm film thickness). Helium was used as the carrier gas at a flow of 1 mL/min. The negative chemical ionization mode (NCI) was used with methane as the reactant gas.

Biochemical analysis

Cholesterol and triacylglycerol concentrations were measured in plasma using commercially available enzymatic kits (Boehringer Mannheim GmbH, Mannheim, Germany) at 3 different sampling times. ApoB100 concentrations were obtained in lipoprotein fractions by combining selective precipitation and mass spectrometry (31).

Statistical analyses

Statistical analyses were performed using Statview software (SAS Institute Inc., SAS campus drive, Cary, NC, USA). The effect of n-3 LC-PUFA supplementation on kinetic parameters and plasma and lipoproteins components (sphingolipids, cholesterol, triglycerides, fatty acids, and apolipoprotein B100) versus basal state was evaluated using the nonparametric Wilcoxon signed rank-test t. The present work was based on our previous study (23); we set the minimum number of subjects recruited to 6 to observe statistical differences for kinetic parameters. To have a high probability of studying at least 5 subjects as in our previous study (23), 18 subjects were prescreened.

Spearman correlations were performed to assess correlations between the different kinetic parameters and different lipids (sphingolipids, cholesterol, triglycerides, and fatty acids) in plasma and lipoproteins. The results are expressed as means ± SD or SEM. Differences were considered statistically significant at P < .05.

Results

Clinical parameters

Supplementation with n-3 LC-PUFA did not affect the body mass index, homeostasic model assessment of insulin resistance, or plasma concentrations of cholesterol, LDL, or HDL cholesterol. As expected, plasma triacylglycerols decreased with treatment, 1.74 ± 0.55 g/L versus 2.66 ± 0.76 (P = .031) (Table 1). Treatment with n-3 LC-PUFA also decreased VLDL1 triacylglycerols and apoB100 content, respectively, 1.85 ± 0.82 versus 2.45 ± 0.69 g/L (P = .045) and 33.67 ± 3.04 versus 80.76 ± 27.01 mg/L (P = .028) (data not shown).

Fatty acid content

To highlight the efficiency of n-3 LC-PUFA supplementation in patients with hypertriglyceridemia, we measured the content of each fatty acid family as a percentage of the total amount of fatty acid. The total amount of fatty acid takes into account the plasma unesterified fatty acids and those esterified in other lipid fractions (triglycerides, cholesterol, and phospholipids) and in the different circulating lipoproteins. Analysis of the plasma fatty acid profile showed an increase in the plasma n-3 LC-PUFA after 8 weeks of supplementation (3.40 ± 0.51% vs 1.20 ± 0.27%, P = .04) (data not shown). We also analyzed the fatty acid profile in VLDL neutral lipids. An increase in total n-3 LC-PUFA content (P = .031) and decrease in total n-6 LC-PUFA content were measured in VLDL (P = .031) (Fig. 3A). Supplementation particularly increased the EPA and DHA content of VLDL neutral lipids (P = .031; Fig. 3B). The total content of saturated fatty acids, and in particular palmitic acid, in VLDL (C16:0), showed a significant decrease (P = .031; Fig. 4) after n-3 LC-PUFA supplementation.

Figure 3.

Long-chain polyunsaturated fatty acid (LC-PUFA) content in VLDL neutral lipids before and after fish oil supplementation. (A) n-3 LC-PUFA and n-6 LC-PUFA content in VLDL neutral lipids before (K1) and after (K2) fish oil supplementation. (B) EPA and DHA in VLDL neutral lipids before (K1) and after (K2) fish oil supplementation. Data are expressed as the mean ± SEM of n = 6 subjects/group. *P value statistically significant at P < .05 using Wilcoxon’s signed rank t-test.

Figure 4.

Saturated fatty acid (SFA) content and palmitic acid (C16:0) in VLDL before (K1) and after (K2) fish oil supplementation. Data are expressed as the mean ± SEM of n = 6 subjects/group. *P value statistically significant at P < .05 using Wilcoxon’s signed rank t-test.

Lipoprotein kinetics

ApoB100 concentrations in VLDL, IDL, and LDL were measured before and at different times during stable isotope infusion. As no significant variations were observed between measurements, all subjects were considered to be in the metabolic steady state over the study. Fitted lines of the time course of labeled leucine enrichment in VLDL, IDL, and LDL apoB100 (Fig. 5) and of lipoprotein apoB100 concentration (Fig. 6) before and after n-3 LC-PUFA treatment for subject 1 showed close agreement with experimental points.

Figure 5.

Time course of enrichment of VLDL 1 (■), VLDL 2 (♦), intermediate-density lipoprotein (IDL, ▲), and LDL (●) apolipoprotein B100 in a representative subject (number 1) before (A) and after (B) supplementation with n-3 polyunsaturated fatty acids.

Figure 6.

Time course of lipoprotein apolipoprotein B100 concentration in VLDL 1 (MV1, ●), VLDL 2 (MV2, ▲), IDL (MI, ♦), and LDL 1 (ML, ■) in a representative subject (number 3) before (A) and after (B) supplementation with n-3 polyunsaturated fatty acids.

Kinetic data are presented in Table 2. Compared to the baseline, treatment with n-3 LC-PUFA reduced the overall absolute production rate (PR) of apoB100 (0.64 ± 0.34 mg/kg/hour vs 0.48 ± 0.24; P = .047), mainly by decreasing VLDL1 apoB100 production (0.390 ± 0.227 vs 0.238 ± 0.114 mg/kg/hour; P = .027). Associated with this decrease, fish oil supplementation significantly increased the conversion rate of VLDL1 to VLDL2 (0.116 ± 0.065 per hour vs 0.094 ± 0.104 per hour , P = .027) and of VLDL2 to LDL (0.099 ± 0.11 per hour vs 0.114 ± 0.134 per hour , P = .027). In contrast, the conversion rate of VLDL2 to IDL (0.116 ± 0.065 per hour vs 0.059 ± 0.039 per hour P = .027) and of IDL to LDL (0.024 0.029 vs 0.015 vs 0.021 P = .027) decreased with the supplementation.

Table 2.

Kinetic parameters before and after fish oil supplementation

Before treatment	ApoB100	VLDL1			VLDL2				IDL			LDL
	Total production	Production	Uptake rate	CR VLDL2	Production	Uptake rate	CR to IDL	CR to LDL	Production	Uptake rate	CR to LDL	Production	FCR
	g/kg/h	g/kg/h	h–1	h–1	g/kg/h	h–1	h–1	h–1	g/kg/h	h–1	h–1	g/kg/h	h–1
S1	1.13	0.72	0.11	0.10	0.67	0.18	0.10	0.04	0.29	0.010	0.03	0.31	0.008
S2	0.97	0.61	016	0.10	0.59	0.27	0.20	0.24	0.10	0.005	0.01	0.31	0.007
S3	0.2	0.13	0.04	0.06	0.15	0.04	0.09	0.00	0.21	0.000	0.008	0.10	0.007
S4	0.44	0.258	0.01	0.08	0.40	0.40	0.20	0.24	0.14	0.000	0.01	0.22	0.014
S5	0.58	0.35	0.08	0.02	0.29	0.15	0.06	0.03	0.15	0.006	0.08	0.12	0.006
S6	0.52	0.27	0.01	0.06	0.44	0.16	0.06	0.04	0.16	0.000	0.006	0.21	0.009
Mean	0.64	0.39	0.07	0.07	0.42	0.20	0.12	010	0.18	0.004	0.02	0.21	0.009
SD	0.35	0.23	0.06	0.03	0.19	0.12	0.06	0.11	0.07	0.005	0.03	0.09	0.003
After treatment
S1	0.70	0,35	0.14	0.12	0.29	0.08	0.12	0.04	0.14	0.007	0.01	0.34	0.008
S2	0.78	0.37	0.21	0.15	0.50	0.001	0.05	0.27	0.10	0.001	0.01	0.47	0.005
S3	0.14	0.07	0.00	0.06	0.13	0.03	0.08	0.01	0.09	0.002	0.006	0.08	0.002
S4	0.37	0.17	0.04	0.10	0.24	0.001	0.005	0.27	0.02	0.001	0.001	0.28	0.014
S5	0.38	0.21	0.16	0.02	0.15	0.04	0.05	0.03	0.07	0.002	0.06	0.13	0.010
S6	0.49	0.26	0.09	0.09	0.19	0.04	0.05	0.08	0.06	0.000	0.004	0.26	0.011
Mean	0.48	0.24	0.11	0.09	0.25	0.03	0.06	0.11	0.08	0.002	0.01	0.26	0.008
SD	0.24	0.11	0.08	0.04	0.13	0.03	0.04	0.13	0.04	0.002	0.02	0.14	0004
	P = .047	P = .027	NS	P = .027	NS	NS	P = .027	P = .027	P = .027	NS	P = .027	P = 0.046	NS

Abbreviations: CR, conversion rate; FCR, fractional catabolic rate; ns, not significant.

Sphingolipid content in apoB100-containing lipoproteins

Ceramide content in VLDL decreased significantly with n-3 LC-PUFA supplementation, with large decreases being observed for specific ceramide species, C20:0, C22:0, C24:0, and C24:1 (Fig. 7A). Neither the LDL ceramide (Fig. 7B) nor the VLDL sphingomyelin content (Fig. 7C) showed changes, but there was an increase in the sphingomyelin content of LDL (Fig. 7D). This increase was related to a higher content of all measured sphingomyelin species except sphingomyelin 16:1. n-3 LC-PUFA supplementation did not appear to cause any change in IDL sphingolipid content (data not shown).

Figure 7.

Ceramides (A, B) and sphingomyelins (C,D) profiles in VLDL and LDL respectively before (K1) and after fish oil supplementation (K2). Data are expressed as the mean ± SEM of n = 6 subjects/group. *P value statistically significant at P < .05 given by a Wilcoxon’s signed rank-test t.

Associations between sphingolipid content and apoB100-containing lipoprotein metabolism

We found positive correlations of VLDL triacylglycerol content with VLDL ceramide content (r = 0.71, P = .01) and VLDL sphingomyelin content (r = 0.71, P = .009). A positive correlation was observed between n-6 LC-PUFA and ceramide contents in VLDL (r = 0.59, P = .04) and between plasma palmitic acid and ceramide VLDL concentrations (r = 0.81, P = .04). The VLDL n-3 LC-PUFA content was inversely correlated with apoB100 VLDL production (r = -0.70, P = .01) and positively correlated with the lipolysis parameter (r = 0.60, P = .03).

Discussion

The aim of this pilot study was to examine the effect of 8 weeks of supplementation with n-3 LC-PUFA on the metabolism of apoB100-containing lipoproteins and their sphingolipid content in fasting subjects with hypertriglyceridemia. We hypothesized that n-3 LC-PUFA supplementation improves the turnover of apoB100 and sphingolipid profiles of apoB100-containing lipoproteins, and that the level of these bioactive lipids is associated with the change in apoB100-containing lipoprotein metabolism. Lipoprotein turnover was examined by using stable isotopes and multicompartmental modeling. Sphingolipid species were determined by mass spectrometry based on a targeted lipidomic approach. We found that n-3 LC-PUFA supplementation lowered plasma triacylglycerol concentrations by increasing the conversion rate from the VLDL1 to VLDL2 and from VLDL2 to LDL and decreasing total apoB100 production mainly by a reduced VLDL1 apoB100 production. These results are supported by the finding that n-3 LC-PUFA supplementation led to significant changes in the fatty acid and sphingolipid profiles, mainly a decrease in VLDL ceramide and an increase in LDL sphingomyelin concentrations.

To our knowledge, this is the first study analyzing the effect of omega-3 supplementation on the distribution of sphingolipids in lipoproteins containing apolipoprotein B100 or on the link between this composition and the metabolism of these lipoproteins. In the present study, treatment efficacy is supported by significant omega-3 enrichment in plasma and VLDL. Since we used fish oil rich in EPA and DHA, the observed effects are related to the presence of these 2 fatty acids.

The present study was performed in the fasting state in order to rule out the acute effect of intestinal absorbed n-3 fatty acids on metabolism of apoB100-containing lipoprotein. Indeed, it is well known that chylomicrons and their remnants enriched in n-3 fatty acids (32, 33) or a direct delivery of dietary n-3 PUFA to the liver (34) reduce hepatic VLDL synthesis. The decrease in plasma triglycerides in supplemented subjects observed in the current study is related to an increase in VLDL lipolysis and a decrease in their production. The increase in VLDL lipolysis is essentially related to a higher conversion rate of VLDL1 to VLDL2 and to a lesser extent to a direct transformation of VLDL2 to LDL. This result is consistent with previously published data (23, 24).

The effect of n-3 fatty acids on lipoprotein lipase (LPL) activity was reported in healthy human subjects (35, 36) or insulin-resistant human subjects (37). Dietary n-3 PUFAs have been also shown to enhance chylomicron triacylglycerol clearance by increasing LPL activity in hypertriacylglycerolemia patients (36). An increase in post-heparin LPL activity and the level of LPL mRNA expression in the adipose tissue of subjects with an atherogenic lipoprotein phenotype has been reported with n-3 fatty acids (38). In accordance with other studies (23, 24), omega-3 fatty acids induce a shift in the channeling of apoB100 from large VLDL to small VLDL and LDL in the present study. This may be explained by the secretion of small VLDLs with a lower triacylglycerol:apoB ratio, which are rapidly converted to LDL (39). In the current study, however, we measured a decrease in the lipolysis of VLDL2 to IDL and IDL to LDL. This decrease did not counteract the hypotriglyceridemic effect of the omega 3, probably because it corresponds to low lipolysis fluxes (transformation flux of VLDL2 to IDL and IDL to LDL) compared with those of VLDL1 to VLDL2 and VLDL2 to LDL. This decrease was not reported in published studies (23, 24), but these 2 studies were conducted in nondiabetic subjects and using a different level and form of omega-3 supplementation. Further studies are needed to understand these discrepancies.

The measured decrease of triglyceridemia after 8 weeks of omega 3 supplementation is also related to a fall in apolipoprotein B100 production, mainly in VLDL1 production. These data are concordant with previous studies (23, 24, 37). Hepatic VLDL production is governed by the availability of triacylglycerols and apoB100, which are required for hepatic VLDL assembly (40). It has been reported that the n-3 LC-PUFA inhibit VLDL secretion by enhancing fatty acid oxidation (41), decreasing the activity of acyl-CoA:1.2-diacylglycerol acyltransferase, and promoting apoB100 degradation (42). We measured enrichment in n-3 LC-PUFAs and a depletion of saturated fatty acids, particularly palmitic acid, in VLDL. In a previous report, treatment with palmitate reduced expression of the tricarboxylic acid cycle and oxidative phosphorylation mitochondrial genes (43, 44). Based on these data and on the observed lower VLDL content of saturated fatty acids and palmitic acid, we suggest that n-3 LC-PUFA supplementation competitively decreased the availability of palmitic acid for hepatic triacylglycerol synthesis and secretion.

Another mechanism supporting the lower VLDL production is the liver availability of ceramides, which are precursors of glycosphingolipids and sphingomyelins. Ceramides are primarily synthesized by 3 pathways: (1) de novo synthesis from serine and palmitoyl-CoA catalyzed by SPT, (2) transformation of sphingosine by ceramide synthases, and (3) transformation from sphingomyelins by specific sphingomyelinases via the salvage pathway. Hepatic VLDL apoB100 production may be stimulated by ceramides. A previous study has shown that the reduction of ceramide and sphingomyelin synthesis by myriocin induced a decrease in liver apoB100 production in LDLR^–/– mice (45) and hamsters (46). We have shown that hamsters fed a diet supplemented with n-3 LC-PUFA (EPA and DHA) exhibited lower VLDL triacylglycerol production (47) and decreased sphingosine, sphinganine, and ceramide contents in the liver compared with controls (26). The findings suggest a possible inhibiting effect of n-3 LC-PUFAs on the de novo ceramide synthesis pathway.

In the present study, we measured decreases of 49% in VLDL ceramide concentrations in response to supplementation with n-3 LC-PUFAs. Reduced palmitic acid content measured in VLDL neutral lipids with supplementation suggests that the de novo synthesis of ceramide may be decreased. These results are strengthened by the positive correlation between VLDL ceramide and palmitic acid contents. Palmitic acid is a major substrate of SPT and stimulates mRNA expression and activity of this enzyme (8, 14). Moreover, we measured a positive correlation between total VLDL ceramide and VLDL triacylglycerol concentrations. In our study, we showed a higher conversion rate of VLDL to IDL and LDL and reduced ceramide concentration in VLDL. We did not measure any correlation between this lipolysis parameter and any VLDL sphingolipids while Ng et al. reported an inverse association between plasma ceramide concentration with VLDL apoB100 FCR in men with the metabolic syndrome during high-dose rosuvastatin treatment (48). This apparent discrepancy could be linked to the mechanism of action involved by the treatment used, accelerated catabolism with rosuvastatin and decreased production with of n-3 LC-PUFAs. Furthermore, in our study we analyzed the sphingolipids of VLDL and nonplasma sphingolipids.

Additionally, we assessed the plasma and apoB100-containing lipoprotein concentrations of sphingomyelins and found that sphingomyelins were transported mainly in plasma LDL and to a lesser extent in VLDL. These results are in accordance with data reported by Wiesner et al. (15). In contrast, Hammad et al. measured a larger amount of sphingomyelins in VLDL (49). These conflicting results may be due to differences in analysis methods. Data on the localization, distribution, and role of sphingomyelin lipid species in lipoprotein particles are still scarce. Consumption of n-3 LC-PUFAs led to significant increases in almost all sphingomyelin species in LDL, with no significant change in VLDL. Sphingomyelins have been reported to be proatherogenic (16, 19, 50). In several studies, LDL, VLDL, and IDL formed aggregates and exhibited increased affinity for macrophage uptake when exposed to sphingomyelinase (12, 51). The question is whether higher plasma sphingomyelin concentration would be a risk factor for Cardiovascular Diseases (CAD) and would indicate worse prognosis, remains controversial. After 5 years of follow-up of participants free of clinical CAD at baseline, Yeboah et al. (52) showed no association between plasma sphingomyelin levels and incident CAD events while Sigruener et al. (53) found long-chain saturated sphingomyelins (23:0, 24:0)] seemed to be associated with a protective effect on cardiovascular mortality. In our study, we measured a higher LDL sphingomyelin (24:0) content. Nevertheless, it is unclear whether the effects of sphingomyelins in lipoproteins in the arterial wall during atherogenesis are directly related to the sphingomyelin content of lipoproteins and/or to sphingolipid messengers (ie, ceramide, ceramide-1-phosphate, sphingosine, and sphingosine-1-phosphate) generated from sphingomyelin hydrolysis by sphingomyelinase. Li et al. have reported that endogenous ceramides contributed to the transcytosis of oxidized LDL across endothelial cells and promoted the initiating step of atherosclerosis (54) while Walters and Wrenn (55) showed that sphingomyelinase induces LDL aggregation due to generation of ceramide and subsequent hydrophobic interactions. Meanwhile, acid sphingomyelinase and the de novo ceramide synthesis inhibitor myriocin significantly inhibited transcytosis of oxidized LDL. Activated secretory sphingomyelinase was associated with acute systemic inflammation (56) and early atherogenesis (57), whereas the inhibition of acid sphingomyelinase activity suppressed the release of inflammatory cytokines from macrophages (58) and its deficiency is associated with reduced atherogenesis in aSMase^–/–/Apoe^–/– and aSMase^–/–/LDLR^–/– mice (56, 59). Based on these observations, we speculate that the increase in LDL sphingomyelin content is related to an inhibiting effect of n-3 LC-PUFA on acid sphingomyelinase activity. This possibility could be supported by reduced sphingomyelinase activity in endothelial progenitor cells (60), in the hippocampus (61), and in isolated carotid arteries (62) from diabetic, aged, and spontaneously hypertensive rats, respectively, that have been supplemented with n-3 LC-PUFA. It has also been shown that elevated human retinal endothelial membrane DHA content disrupted NFκB signaling and downregulated acid sphingomyelinase expression and activity. This suppression of sphingomyelinase activity is associated with a decline in ceramide, an inflammatory and proapoptotic lipid (63).

In conclusion, we demonstrated that n-3 LC-PUFA (EPA and DHA) supplementation improves apoB100-containing lipoprotein metabolism and leads to a specific pattern of changes in the sphingolipid profile in patients with hypertriglyceridemia characterized by a decrease in VLDL ceramides and an increase in LDL sphingomyelin content. We found significant associations between kinetic parameters and ceramide content in VLDL. These data suggest that improvement of metabolism of apoB100-containing lipoproteins is associated with altered lipoprotein sphingolipid content. Further targeted investigations are needed to confirm or refute these data and to understand the involvement of sphingolipids in lipoprotein metabolism. Finally, in this pilot study only 6 subjects were recruited, which might limit this study. However, each patient was their own control, reducing the important human variability between subjects and increasing statistical robustness. Nevertheless, it will be interesting to expand this study with a larger sample size.

Abbreviations

Abbreviations
 
• apoB100

  apolipoprotein B100

 
• APR

  absolute production rate

 
• CR

  conversion rate

 
• DHA

  docosahexaenoic acid

 
• EPA

  eicosapentaenoic acid

 
• FCR

  fractional catabolic rate

 
• FFA

  free fatty acid

 
• GC-MS

  gas chromatography-mass spectrometry

 
• IDL

  intermediate-density lipoprotein

 
• LDL

  low-density lipoprotein

 
• n-3 LC-PUFA

  omega-3 long-chain polyunsaturated fatty acid

 
• MTP

  microsomal triglyceride transfer protein

 
• SPT

  serine-palmitoyl transferase

 
• VLDL

  very low-density lipoprotein

 
• UPLC-MS/MS

  ultra performance liquid chromatography-tandem mass spectrometry

Acknowledgments

We thank Emmanuel Desnot for her technical assistance and Eliane Hivernaud for the progress of the clinical trial.

Financial Support: This work was supported by the Pierre Fabre Laboratories

Author Contributions: V.F.R. performed the targeted lipidomic analysis and metabolite determinations in plasma and lipoprotein and contributed to the writing of the manuscript; A.A. performed the fatty acid analysis; Y.Z. and M.K. conducted and supervised the original clinical trial; K.O. designed and coordinated the study, performed kinetic analysis, wrote the manuscript and had primary responsibility for the final content.

Clinical Trial Information: Clinical trial registered by AFFSAPS as no. 060575-35 (2006).

Additional Information

Disclosure Summary: The authors declare no competing interests.

Data Availability

The datasets generated during and/or analyzed during the current study are not publicly available but are available from the corresponding author on reasonable request.