Modification of LDL is directly related to the initiation and progression of atherosclerosis. Modified LDL can promote subendothelial accumulation of cholesterol and activates the chronic inflammatory response characteristic of atheromatous lesions [1
]. Electronegative low-density lipoprotein (LDL(−)) is a minor modified LDL subfraction that is present in circulation [2
]. LDL(−) has some potentially atherogenic characteristics compared to native LDL. LDL(−) has a higher content of non-esterified fatty acids (NEFA) and lysophosphatidylcholine (LPC) [3
], possibly related to its platelet-activating factor acetylhydrolase (PAF-AH) activity, due to the presence of the PAF-AH enzyme [5
]. LDL(−) also presents a phospholipase C (PLC)-like activity [6
], whose origin is still unknown. It has been suggested that it could be related to other atherogenic traits of LDL(−) [7
], such as increased susceptibility to aggregation [8
], abnormal apolipoprotein B (apoB) conformation [9
] or high binding to proteoglycans [10
]. LDL(−) also promotes the release of cytokines, including MCP1, IL6, IL8, IL10 and GRO, in endothelial cells [11
], monocytes and lymphocytes [12
In a previous study, we found that HDL counteracted the cytokine release induced by LDL(−) in monocytes [12
]. However, the mechanism by which HDL inhibited the inflammatory action of LDL(−) was not clearly defined. This effect could be partly due to the exchange of NEFA between LDL(−) and HDL, since it was previously reported that NEFA are involved in the inflammatory effect of LDL(−). It is also possible that other molecules, generated by the PLC-like activity of LDL(−), could be involved in the inflammatory action of LDL(−) [3
]. The PLC-like activity in LDL(−) hydrolyzes the polar head of choline-containing phospholipids and preferentially degrades LPC, with medium efficiency sphingomyelin (SM) and with lower efficiency phosphatidylcholine (PC). The products of this hydrolysis are phosphocoline (Pchol), monoacylglycerol (MAG), ceramide (CER) and diacylglycerol (DAG). Pchol is water-soluble and presumably leaves the LDL particle. In contrast, MAG, CER and DAG, which are hydrophobic, would remain retained in the LDL particles [14
]. An increased content of MAG, CER and DAG in LDL modifies the surface structure of the particles, leading to their aggregation and fusion by hydrophobic associations [14
]. PLC-like activity of LDL(−) is mainly present in a subfraction of aggregated LDL(−) [8
], suggesting that this activity is involved in LDL aggregation.
Even though LPC is rapidly degraded by the PLC-like activity, MAG would be scarce in LDL, since the amount of LPC is much lower (2%–3% of total phospholipids in LDL) than PC (70%) and SM (20%). CER and DAG are considered as bioactive and inflammatory molecules that promote cell signal transduction. CER participates in the regulation of cell proliferation and differentiation, inflammation and apoptosis in a wide variety of cells, including cells of the immune system [16
]. It is also the central core in sphingolipid metabolism, since it is a precursor of other bioactive sphingolipids [19
]. DAG stimulates protein kinase C activity and adenylcyclase, which generates cAMP, a pivotal molecule in many biological processes [20
]. DAG is also essential for propagation of the downstream signals required for NFkB activation. CER and DAG can be generated in cells in response to extracellular stimuli, such as cytokine, LPS [21
] and LDL modified by oxidation or acetylation [22
]. Besides activating the SM-CER pathway in cells, these modified LDLs can also carry ceramide itself [24
The aim of the current study was to assess whether the products of the PLC-like activity generated during incubation of LDL(−) with cells, particularly CER and DAG, are involved in the inflammatory effect of LDL(−).
3. Experimental Section
3.1. Lipoprotein Isolation and Separation of LDL Subfractions
Plasma samples from healthy normolipemic subjects (total cholesterol < 5.2 mmol/L, triglyceride < 1 mmol/L) were obtained in EDTA-containing Vacutainer tubes. LDL (density range 1.019–1.050 g/mL) and HDL (1.100–1.210 g/mL) were isolated by sequential flotation ultracentrifugation at 4 °C. Total LDL was subfractionated in LDL(+) and LDL(−) by preparative anion-exchange chromatography in an ÄKTA-FPLC system (GE Healthcare, Chalfont Sant Giles, UK) [29
]. In all experiments, the proportion of LDL(−) ranged from 4% to 6% of total LDL, and the major characteristics of both LDL subfractions were similar to those previously reported [25
] (data not shown). Both subfractions were concentrated by centrifugation with Amicon microconcentrators. All the analyses with these subfractions were performed within 3 days of isolation.
3.2. LDL Incubation with HDL
LDL(+) and LDL(−) (0.5 g/L apoB) were incubated with HDL (0.5 g/L apoAI) at 37 °C for 2 h in phosphate saline buffer (PBS) and shaken gently in the presence of 20 μM butylated hydroxytoluene (BHT) to avoid oxidation. After incubation, LDL and HDL were re-isolated according to their density. Half of each sample was kept at 4 °C, and the other half was incubated at 37 °C for 20 h to reproduce the conditions of cytokine release experiments with cells. Afterwards, samples were kept at 4 °C until the physicochemical characterization.
3.3. LDL Modification
To selectively enrich LDL with PLC-like activity products, liposomes containing these different compounds were used as donor and native LDL was used as acceptor, as described [30
]. To form the liposomes, total lipids were extracted from native LDL (1 mg apoB), according to the Bligh and Dyer method, and the compounds were added to lipids from the chloroform phase. N
-sphingosine (ceramide, CER), 1,2-diacyl-sn-glycero-3-phosphocholine (diacylglycerol, DAG) and sphingomyelin (SM) were added at 5 and 10 μM. All compounds were from Sigma. Afterwards, the lipid extract was dried under a nitrogen stream, resuspended in 200 μL of a KBr solution (density 1.019 g/mL) and sonicated in a water bath until the mixture was translucent, which indicates that liposomes are formed. The acceptor (LDL at 1 mg apoB) was then incubated with the donor vesicle (liposomes) at 37 °C for 45 min in the presence of 20 μM BHT. After the incubation, the suspension was overlaid with KBr solution (density 1.019 g/mL), and liposomes and LDL were re-isolated by ultracentrifugation. Liposomes, which are less dense than 1.019 g/mL, were recovered floating at the surface, while enriched-LDL (CER-LDL, DAG-LDL, SM-LDL) was recovered in the lower phase. Non-enriched LDL was processed in parallel and incubated with non-enriched liposomes. Lipid extracts of the samples were separated by thin layer chromatography (TLC), to assess the increase of DAG, CER and SM.
LDL(+) (0.5 g/L) was modified with PLC (Sigma, Madrid, Spain) by incubation with the enzyme at 50 and 100 U/L for 1 h at 37 °C. The reaction buffer was 5 mM HEPES, 2 mM CaCl2, 5 mM MgCl2, 140 mM NaCl at pH 7.4. The reaction was stopped with 10 mM EDTA. LDL(+) was aggregated in vitro by vortexing for increasing times.
3.4. Lipid and Apoprotein Composition
Cholesterol, triglyceride, apoB, apoAI (Roche Diagnostic, Basel, Switzerland), phospholipid and NEFA (Wako Chemicals, Richmond, VA, USA) content of the samples were determined in a Hitachi 917 autoanalyzer. The results were expressed as the percentage of lipoprotein mass and, in the case of NEFA, as mol NEFA/mol apoB for LDL and mol NEFA/mol apoAI for HDL.
Other minor lipid components, including CER, DAG and MAG, were evaluated by extracting lipids from LDLs (250 μg apoB), according to the Bligh and Dyer method. Lipid extracts were resuspended in 20 μL chloroform and applied to the TLC plate. The silica gel plates were developed using three sequential mobile phases, as follows. Phase 1: chloroform/methanol/water (v/v/v 65:40:5) to 5 cm. Phase 2: toluene/diethilether/ethanol (v/v/v 60:40:3) to 13 cm. Phase 3: heptane to 17 cm. Lipids were stained by dipping the plates in a solution containing 5% phosphomolibdic and 5% sulfuric acid in ethanol for 1 min, and then dried at 100 °C for 10 min.
3.5. Aggregation Level and Oxidation Tests
Aggregation was monitored by measuring absorbance of LDL and HDL at 450 nm (0.5 g/L apoB or apoAI for basal aggregation and 0.2 g/L for susceptibility to aggregation). Lipoproteins were vortexed at increasing times up to 1 min [12
The peroxide content of lipoproteins was measured following the Auerbach assay, using 20 μg of apoB for LDL or apoAI for HDL [31
3.6. Enzymatic Activity Measurements
PAF-AH activity of the samples at 0.2 g/L apoB (LDL) and apoAI (HDL) was evaluated by a commercial colorimetric assay based on degradation of 2-thio-PAF (Cayman Chemical, Michigan, MI, USA) [5
]. PLC-like activity of LDL(+), LDL(−), HDL and apoAI was evaluated in the different incubation conditions [6
]. Briefly, we used a commercial fluorimetric assay based on enzyme-coupled reactions and final detection of fluorescent Amplex red. Sphingomyelin (SM) and lysophosphatidylcholine (LPC) were used as substrates, and 30 μg of lipoproteins (apoB in LDLor apoAI in HDL) were assayed. Fluorescence was monitored for 3 h and activity was calculated from the maximum slope.
To corroborate the SMase activity and avoid possible interferences of the Amplex red method, the degradation of SM labeled with the fluorescent probe bodipy (bodipy-SM) (Molecular Probes, Leiden, The Netherlands) was also measured. As described, samples (50 μg) were incubated with bodipy-SM for 3 h at 37 °C, followed by lipid extraction and separation by TLC [6
3.7. Cytokine Release Experiments in Monocytes
Human monocytes were isolated according to their density from the blood of normal volunteers who gave their written informed consent [13
]. Monocytes were cultured in 12-well plates (106
cells/well) for 1 day and incubated in the experiments, with LDL previously filtered in sterility, at 150 mg apoB/L for 20 h, in the case of LDL. After incubation, the supernatant was collected, and the concentrations of IL6, IL10 and MCP1 were quantified by ELISA (Bender Medsystems, Burlingame, CA, USA) [13
3.8. Statistical Analysis
Results are expressed as mean ± SD. A Sigma Stat 2.0 statistical package was used. Differences between groups were tested with Wilcoxon’s T-test (for paired data).
Our results show a relationship between the content of CER in LDL(−) and the release of cytokines in monocytes. In contrast, in our experimental conditions, DAG-LDL did not promote a significant increase of IL6 release compared to native LDL. LDL(−) pre-incubated with HDL decreased its PLC-activity and CER content. This effect could explain the described counteracting effect of HDL on LDL(−) inflammatory action [12
]. It is known that CER acts as a second messenger in signal transduction pathways in cells, including biosynthesis of cytokines [18
]. A relationship between the content of CER in LDL and the development of inflammatory diseases, including atherosclerosis, has been described [24
]. Retained lesional LDL also accumulates CER [15
]. The involvement of CER in cytokine release promoted by LDL(−) could be related to the recent observation of CD14 as the main receptor of LDL(−) in monocytes [34
]. The binding of LDL(−) to CD14 could be mediated by CER, as it has been reported that CD14 binds CER [35
]. This binding activates the TLR4 assembly and cytokine induction in monocytes. Interestingly, CD14 also binds NEFA [36
]. The observation that both CER and NEFA increase their content in LDL(−) after 20 h of incubation at 37 °C supports the role of CD14 in the activation of the inflammatory response mediated by LDL(−) in monocytes.
Besides promoting hydrolysis of its own phospholipids, the high PLC-like activity of LDL(−) could induce a degradation of phospholipids in the membrane of monocytes. This could promote an increase of the intracellular CER content and its derivates. This effect was also reported for oxLDL [37
]. In addition, LDL(−) could activate cell SMase activity, since it induces Fas expression in monocytes [38
], an action known to promote SM breakdown and CER production [39
Our data show that CER-LDL released more cytokines than native LDL, but not as much as the release induced by LDL(−). CER-LDL did not reproduce all the characteristics of LDL(−), suggesting the involvement of other compounds in the inflammatory effect of LDL(−). Besides NEFA, CER metabolites, such as sphingosine and sphingosine-1-phosphate, could also be involved [40
]. It has been proposed that HDL prevents NF-kB activation and synthesis of inflammatory proteins by decreasing sphingosine-1-phosphate [44
]. The decrease in CER content in LDL(−) when incubated with HDL could also cause a decrease in these CER metabolites.