Role of Group II Secretory Phospholipase A₂ in Atherosclerosis

Materials

l-α-1-Palmitoyl-2-arachidonoyl-sn-glycero-3-phosphorylcholine was obtained from Avanti Polar Lipids Inc. or Sigma Chemical Co. Soybean lipoxygenases (SLO), sPLA₂ (Naja naja), butylated hydroxytoluene (BHT), dimyristoylphosphatidylcholine (DMPC), palmitic acid, stearic acid, oleic acid, linoleic acid, arachidonic acid, and endotoxin-free gelatin, tissue-culture grade, were obtained from Sigma. Aminopropyl solid phase extraction columns were obtained from J.T. Baker. Chloroform, methanol, isopropanol, and ethyl ether (all HPLC grade or better) were obtained from Fisher Scientific. Tissue culture media, serum, and supplements were obtained from sources previously reported.⁸ Transwells and chamber slides were obtained from Costar. The fluorescent probe DiI (1,1′-dioctadecyl-3,3,3′-tetramethyl-indo-carbocyanine perchlorate) was purchased from Molecular Probes. Human recombinant sPLA₂ was a generous gifts from Dr J. Browning, Biogen. The method of preparation of sPLA₂ has been previously described.¹⁴ Briefly, recombinant sPLA₂ was purified from medium of Chinese hamster ovary (CHO) cells transfected with an expression vector encoding human sPLA₂.

Animals

sPLA₂ transgenic mice were obtained from Chrysalis DNX Transgenic Sciences, Princeton, NJ, and had been produced as described previously.¹⁵ The atherogenic diet contained 1.25% cholesterol, 15.75% fat, and 0.5% sodium cholate (TD 90221, Harlan-Teklad). After 12 weeks on the high-fat diet, the mice were sacrificed by cervical dislocation after the administration of isoflurane. All animal procedures were conducted according to the regulations of the University of California Animal Research Committee.

Lipoprotein Isolation and Modification

Human LDL (d=1.019 to 1.069 g/mL) and HDL (d=1.069 to 1.210 g/mL) were isolated from the sera of healthy blood donors by density gradient ultracentrifugation as described¹⁶ and used within 1 to 2 weeks of isolation. Concentration of LDL is expressed in terms of protein content as determined according to the method of Lowry et al.¹⁷ The concentration of endotoxin in solutions tested in biological assays was less than 50 pg/mL (determined by a chromogenic assay), which is 40-fold less than that required to induce monocyte-endothelial interactions. LDL was modified using purified SLO and PLA₂, according to Sparrow et al.¹⁸ To remove enzymes from the reaction by centrifugation, SLO and PLA₂ were separately bound to CNBr-activated Sepharose beads as described.¹⁸ LDL (1 mg/mL) was modified by incubation with SLO (5000 U/mL) alone, SLO+PLA₂ (20 U/mL), or SLO+fatty acids (187 mmol/L) at 37°C for 24 hours or at 4°C for 48 hours.¹⁹ The enzymes were removed by centrifugation, BHT (100 μmol/L) and EDTA (0.3 mmol/L) were added, and the modified LDL was stored at 4°C and used within 2 weeks.

Mouse HDL were isolated from pooled plasma by fast protein liquid chromatography,²⁰ in the absence of EDTA to avoid inactivation of paraoxonase. HDL were incubated at concentrations of 2 mg protein/mL with recombinant human sPLA₂ (0.5 μg/mL) for 16 hours, subsequently isolated again, and then added to the cocultures at a final concentration of 250 to 350 μg/mL.

Cell Culture

Human aortic endothelial cells (HAEC) and rabbit aortic endothelial cells (RAEC) were cultured as described.⁸ Human aortic smooth muscle cells (HASMC) were isolated as previously described.²¹ Cocultures of HAEC and HASMC were grown, and monocyte binding and transmigration experiments were performed in transwells or chamber slides as previously described.²¹ In all experiments, HAEC and HASMC from the same donor were used at passages 4 to 6. Blood monocytes were obtained from a large pool of healthy donors by modification of the Recalde procedure as described previously.²²

Monocyte Adhesion Assay

Binding of human monocytes to endothelial cells was performed essentially as described previously.⁸ HAEC or RAEC were treated with MM-LDL (125 to 200 μg/mL) for 4 hours at 37°C. Treatment media were removed, the cells were rinsed twice with medium, and a suspension of human monocytes (2 to 3×10⁵/well) was added for 12 minutes. Unbound monocytes were removed, and the number of bound monocytes was determined microscopically.

Monocyte Transmigration Assay

These assays were performed essentially as described previously.²¹ Cocultures were treated with native LDL (250 to 350 mg/mL) in the absence or presence of various test compounds for 24 hours. The culture supernatants were subsequently transferred to untreated cocultures and were incubated for an additional 24 hours. Monocytes were labeled with the fluorescent probe DiI at 4°C for 10 minutes, washed, and resuspended in medium-199 at the desired cell density. Labeled monocytes were added to the treated cocultures at 2×10⁵ cells/cm² and were incubated for 60 minutes at 37°C. The medium containing nonadherent cells was removed, and the cell layers were washed to remove loosely adherent monocytes. The cocultures were fixed on polycarbonate membranes with 10% neutral buffered formalin at room temperature for 24 hours and were mounted on glass slides. The number of monocytes in the subendothelial space (beneath the endothelial cells) in a minimum of 9 fields was determined under 500× magnification and fluorescence microscopy.

Measurement of Lipid Hydroperoxides

Lipid hydroperoxide formation was measured using the method of Auerbach et al.²³

Lipid Extraction

Lipids were extracted from modified LDL and from mouse livers using a modification of the method of Bligh and Dyer.²⁴ Liver tissue samples were homogenized in chloroform/methanol (2:1) containing 0.01% BHT for 5 minutes at 0°C. After addition of water, liver suspensions were vortexed for 2 minutes and centrifuged at 1200g for 15 minutes at 4°C (Beckman J-6B; rotor, Beckman JS-4.2). In separate studies chloroform/methanol/BHT was added to modified LDL in PBS, and these suspensions were vortexed for 2 minutes. The chloroform phase was carefully removed and 5 vol of chloroform was added to the residual aqueous phase. The mixture was vortexed and centrifuged as described above, and the chloroform phase was pooled with the previous extract. Phospholipid recovery typically ranged between 96% and 98% as determined by spiking samples with known quantities of 1-palmitoyl-2-[¹⁴C]arachidonoyl-sn-glycero-3-phosphorylcholine.

Solid Phase Extraction

Phospholipids, free fatty acids, and neutral lipids were separated by the method of Kaluzny et al.²⁵ using aminopropyl solid phase extraction chromatography. Total lipid extracts were dried under nitrogen, resuspended in 200 μL of chloroform, and applied to the aminopropyl columns, which had been washed with 3 mL of methanol and preconditioned with 3 mL of hexane. Neutral lipids were eluted with 3 mL of chloroform/2-propanol (2:1), free fatty acids with 3 mL of 3% acetic acid in ethyl ether, and phospholipids with 3 mL of methanol. When phospholipids were used in tissue culture experiments, acetic acid retained in the columns was removed with pure ethyl ether before phospholipid elution. Lipid fractions were dried under nitrogen and either used immediately or resuspended in chloroform containing 0.01% BHT, covered with argon, and stored at −44°C.

Substrate Specificity Assay

OxPAPC was produced by exposing dry PAPC to air for 3 days. Oxidation products, as well as residual unoxidized PAPC, were present in these preparations. Liposomes were produced by sonicating OxPAPC for 10 minutes in Tris buffer, pH 7.4. Fifty microliters containing 500 μg of OxPAPC liposomes was incubated with either human recombinant sPLA₂ (1 μg), snake venom sPLA₂ (Naja naja; 5 U), or buffer alone for 90 minutes at room temperature in Tris buffer containing 2 mmol/L Ca²⁺. The reaction was stopped and lipids were extracted by adding 200 μL chloroform/methanol (2:1)+0.01% BHT. After vortexing and centrifugation for 10 minutes, 100 μL of the chloroform phase was transferred into a glass tube, and 500 ng of internal standard (DMPC) was added to each tube.²⁶ Samples were dried under a stream of nitrogen and redissolved in acetonitrile/water/formic acid (50:50:0.1; vol/vol/vol) and analyzed by electrospray ionization–mass spectrometry (ESI-MS).

Electrospray Ionization–Mass Spectrometry

An API III triple-quadrupole biomolecular mass analyzer (Perkin-Elmer Sciex Instruments) was used for mass analysis of phospholipids. Flow-injection analysis was performed as described previously.¹⁰ For quantitative measurements, DMPC was used as an internal standard.²⁶ Mass spectrometry of individual phospholipid molecular species represents the most sensitive, discriminating, and direct method to assess alterations in phospholipid molecular species in biological tissues.²⁶ In contrast to previously used techniques like fast atom bombardment mass spectrometry, which results in extensive fragmentation of individual phospholipid ions, electrospray ionization results in the efficient production of molecular ions of phospholipids with negligible fragmentation.²⁷ Therefore, ESI-MS offers a sensitive and reliable tool for quantitative phospholipid analysis in biological tissues. For quantitative measurements we were assuming that possible differential effects of phospholipid oxidation on extraction recovery, ionization efficiency, and molecular ion yield were negligibly small in comparison with the quantitative differences reported. In an extensive examination of phospholipid quantitation by ESI-MS, Han et al²⁶ reported that the effects of unsaturation and differential surface characteristics were minimal. Because any ¹³C effects on quantitation were constant within the comparisons made, no corrections of the molecular ion signal intensities have been performed. Within the range of concentrations measured, trial experiments indicated that the response was linear, such that increases in phospholipid content were reflected in proportional increases in phospholipid to internal standard ratio.

Statistical Analysis

Statistical analysis was performed using 1-way ANOVA; probability values <0.05 were considered statistically significant.

Evaluation of Phospholipids by ESI-MS

Previous studies have shown that group II sPLA₂ was expressed to a significantly greater extent in livers¹⁵ and aortas¹ of sPLA₂ transgenic mice when compared with nontransgenic littermates. Liver was used as a surrogate for aorta as validated in previous studies in Bl-6 mice.¹² Phospholipids were separated from a total lipid extract obtained from livers of female sPLA₂ transgenic mice and nontransgenic littermates maintained for 12 weeks on an atherogenic diet. Analysis of phospholipids was performed by ESI-MS, and quantification of the different phospholipid species was carried out by addition of DMPC as an internal standard, as described in the Methods. The levels of native phospholipids, 1-palmitoyl-2-arachidonoyl-sn-glycero-3-phosphorylcholine (PLPC, m/z 758) and PAPC, m/z 782, in livers of transgenic female mice were not significantly different from the levels in livers of nontransgenic (Figure 1A⇓). In contrast, the levels of POVPC and PGPC were significantly increased in transgenic animals versus nontransgenics (Figure 1B⇓; P<0.003 for POVPC, P<0.014 for PGPC). The bioactive molecule with m/z 828.6 was not increased in the transgenic compared with nontransgenic animals (Figure 1B⇓). Interestingly, the levels of lysophosphatidylcholine (LPC) (sn-1=palmitic acid (16:0), m/z 496, and sn-1=stearic acid (18:0), m/z 524 were not significantly different between transgenic and nontransgenic animals (Figure 1A⇓).

• Download figure
• Open in new tab
• Download powerpoint

Figure 1.

Levels of unoxidized phospholipids and LPC (A) and bioactive derivatives of oxidized PAPC (B) in livers of sPLA₂ transgenic (TG) and nontransgenic (nonTG) mice. Levels of phosphatidylcholines in livers were determined using positive-mode ESI-MS. Lipids were extracted with chloroform/methanol and phospholipids separated using solid phase extraction columns after addition of DMPC as an internal standard. The abundance of ions with m/z 782 (PAPC), m/z 758 (PLPC), m/z 496 (LPC 16:0), m/z 524 (LPC 18:0), m/z 594.3 (POVPC), m/z 610.2 (PGPC), and m/z 828.6 was determined using software supplied by PE Sciex and plotted in terms of ng DMPC equivalents/mg liver tissue. Data are presented as mean+SD (n=8 animals in each group). Levels of POVPC and PGPC were significantly increased in the livers of TG compared with nonTG animals (B) (P<0.003 for POVPC and P<0.014 for PGPC). However, there was no significant increase in native phospholipids PAPC or PLPC or of LPC in the livers of the transgenics (A).

Effect of sPLA₂ Products on LDL Phospholipid Oxidation

Our group and others have shown that treatment of LDL with sPLA₂ from Naja naja combined with SLO leads to the formation of conjugated dienes and bioactive phospholipids.^{18 19} On the basis of these findings, we hypothesized that treatment of LDL with sPLA₂ led to a limited liberation of fatty acids that, when oxidized, could cause oxidation of arachidonate-containing phospholipids, leading to the formation of bioactive phospholipids. To test this hypothesis, LDL was incubated without additions, or in the presence of SLO and linoleic acid. The formation of conjugated dienes was increased 2- to 3-fold in the phospholipid fraction of the treated LDL compared with the untreated LDL (Figure 2A⇓). An approximately 10-fold increase in the level of absorbance at 270 nm, an indication for the levels of bioactive phospholipids,⁹ was also observed (Figure 2B⇓). The treated LDL preparation was active in stimulating endothelial cells to bind monocytes, further confirming the presence of biologically active phospholipids (Figure 2C⇓). Arachidonic acid had similar effects (data not shown). To further test the hypothesis that oxidized fatty acids play a role in the formation of bioactive phospholipids, LDL was treated with oxidized linoleic acid in the absence of other additives. This treatment also stimulated the formation of bioactive phospholipids (monocyte binding increased by 2.4±0.3-fold). Oxidized fatty acids alone in the absence of LDL had no effect on monocyte binding (data not shown). Several lines of evidence suggest that lipoxygenase may play a significant role in the oxidation of LDL by cells.²⁸ Free polyunsaturated fatty acids (PUFAs) are better substrates than esterified PUFAs for some mammalian lipoxygenases. We tested the hypothesis that treatment of LDL with sPLA₂ would increase the ability of cells to oxidize LDL and thereby form bioactive phospholipids. Pretreatment of LDL with different concentrations of human recombinant sPLA₂ and subsequent incubation for 24 hours with cocultures of HAEC and HASMC resulted in increased formation of lipid hydroperoxides (Figure 3A⇓) and increased bioactivity (Figure 3B⇓).

• Download figure
• Open in new tab
• Download powerpoint

Figure 2.

Biological activity, phospholipid-associated conjugated diene formation, and absorbance at 270 nm in untreated and SLO+linoleic acid-treated LDL. LDL were incubated with SLO and linoleic acid (18:2). Lipids were extracted from LDL and phospholipids were isolated by solid phase extraction chromatography and analyzed for conjugated dienes (A) and for presence of bioactive phospholipids (B) by measuring absorbance at 234 and 270 nm, respectively. C, RAEC were incubated with LDL oxidized with SLO and linoleic acid (18:2) or untreated LDL, and then tested for induction of human monocyte adhesion. These data are representative of 4 separate experiments in which each condition was tested in quadruplicate. Values are mean+SD (n=4). *Indicates P<0.001.

• Download figure
• Open in new tab
• Download powerpoint

Figure 3.

Effect of sPLA₂ treatment of LDL on oxidation by coculture cells and induction of monocyte transmigration. Human LDL was treated with different concentrations of human recombinant sPLA₂ for 4 hours, then incubated with cocultures of endothelial cells and smooth muscle cells for 24 hours. The supernatant was then removed and tested for levels of lipid hydroperoxides (A) and tested on a second set of cocultures for ability to induce monocyte transmigration (B). Lipid hydroperoxide levels were significantly elevated with as little as 0.1 μg/mL of sPLA₂. Lower levels of sPLA₂-treated LDL versus untreated LDL were more effective in stimulating monocyte transmigration (P<0.05). These data are representative of 4 separate experiments in which each condition was tested in quadruplicate. Mean+SD (n=4).

Substrate Specificity of sPLA₂

It was clear from the above studies that sPLA₂ could both hydrolyze phospholipids and lead to the formation of bioactive phospholipids. We hypothesized that these 2 observations were compatible with the idea that other phospholipids in OxPAPC were better substrates for sPLA₂ than the bioactive oxidized phospholipids. To assess the substrate specificity among the various species of oxidized phospholipids present in OxPAPC, a human recombinant sPLA₂ and snake venom sPLA₂ were used. Liposomes consisting of OxPAPC (which contained small amounts of native PAPC) were incubated with the enzymes and analyzed by ESI-MS. The bioactive molecules POVPC (m/z 594.3), PGPC (m/z 610.2), and m/z 828.6 were poorly hydrolyzed by the treatment with human recombinant sPLA₂, whereas unoxidized PAPC (m/z 782.4) was effectively hydrolyzed (Figure 4⇓ and the Table⇓). After treatment with recombinant sPLA₂, the level of LPC was increased to a greater extent than could be accounted for by hydrolysis of native PAPC. This suggested that the excess LPC was because of hydrolysis of other oxidized phospholipids present in OxPAPC. Although the reduction of the >100 peaks (seen by MS) is not necessarily discernible in Figure 4⇓, the collective production of LPC from these lipids is highly significant. The snake venom sPLA₂ showed a preference for POVPC (m/z 594.3) and unoxidized PAPC (m/z 782.4) among the phospholipids present in OxPAPC. Representative mass spectra are shown in Figure 4⇓ and the results of these experiments are summarized in the Table⇓.

• Download figure
• Open in new tab
• Download powerpoint

Figure 4.

Substrate specificity of human recombinant sPLA₂ on liposomes of oxidized PAPC. Representative mass spectra of OxPAPC liposomes under the following conditions: untreated (A), after treatment with human recombinant sPLA₂ (B), and after treatment with snake venom sPLA₂ (C). Inset B, Difference between human recombinant sPLA₂-treated and untreated (A minus B). Inset C, Difference between snake venom sPLA₂-treated and untreated (A minus C). For these studies 50 μL containing 500 μg of OxPAPC liposomes was incubated with human recombinant sPLA₂ (1 μg), snake venom sPLA₂ (Naja naja; 5 U), or buffer alone for 90 minutes at room temperature in Tris buffer containing 2 mmol/L Ca²⁺. Lipids were extracted and analyzed by ESI-MS as described in the Methods. Relative changes of different phospholipid species were evaluated by comparing relative ion abundance to the internal standard (DMPC, m/z 678.6). (In C, the level of LPC, m/z 496, was approximately 300 as measured on an expanded scale.) These experiments were performed 3 times, and a representative experiment is shown.

HDL From sPLA₂ Transgenic Mice Fail to Protect Against LDL Oxidation

HDL has been shown in vitro to prevent LDL oxidation and accumulation of bioactive lipids.⁴ This protective HDL function may also play an important role in fatty streak formation. We examined the protective effect of HDL from sPLA₂ transgenic mice and their nontransgenic littermates using a coculture system in which cell-induced LDL oxidation stimulates the secretion of monocyte chemotactic protein-1 (MCP-1): levels of MCP-1 are assessed as monocyte chemotaxis and transmigration in the coculture.²⁹ As previously reported, the addition of HDL along with LDL to such cocultures inhibited the formation of biologically active oxidized LDL.²⁹ The HDL isolated from the plasma of sPLA₂ transgenic mice did not protect against LDL-induced monocyte transmigration, whereas HDL from nontransgenic littermates was protective (Figure 5A⇓). In fact, the HDL from the transgenic mice significantly promoted LDL-oxidation and monocyte transmigration in the coculture. The HDL from the sPLA₂ transgenics contained one-third the paraoxonase activity of HDL from nontransgenics. To further study the sPLA₂-mediated effects on HDL, we performed additional in vitro experiments (Figure 5B⇓). First, we preincubated human HDL with sPLA₂ for 3 hours with gentle mixing; the HDL was then isolated again and added to the coculture together with human LDL. Monocyte transmigration in the coculture containing sPLA₂-treated HDL was increased 2-fold as compared with the addition of LDL alone (no HDL) and ≈3-fold compared with the addition of LDL and sham-treated HDL. Therefore, sPLA₂-treated HDL not only failed to protect against LDL oxidation but also exhibited proinflammatory capacity.

• Download figure
• Open in new tab
• Download powerpoint

Figure 5.

HDL isolated from sPLA₂ transgenic mice (A) or HDL treated in vitro with sPLA₂ (B) fail to protect against LDL oxidation in a coculture model of the artery wall. A, Human aortic smooth muscle and endothelial cell cocultures were incubated for 18 hours with medium (no additions), 350 μg/mL LDL (no HDL), or LDL plus HDL (nontransgenic HDL, transgenic HDL). B, sPLA₂-treated HDL denotes HDL isolated again that had been first incubated with sPLA₂ (0.5 μg/mL) for 3 hours; sham-treated HDL, HDL pretreated in a similar fashion omitting sPLA₂. Supernatants were then transferred to new cocultures. Labeled human blood monocytes were added and allowed to adhere and migrate into the endothelial cell layer for 1 hour. After washing off nonadherent cells, cocultures were fixed and the recruited monocytes were counted using 625× magnification. The number of monocytes per high-power field are given as mean+SD.