|Home | About | Journals | Submit | Contact Us | Français|
In vitro data indicate that human LDL modified by Group V secretory phospholipase A2 (GV sPLA2) is pro-atherogenic. Consistent with this, gain and loss of function studies demonstrated that GV sPLA2 promotes atherosclerosis in LDLR-/- mice. The current study investigates whether GV sPLA2 promotes atherosclerotic processes in apoE-/- mice.
LDL (d = 1.019-1.063) from apoE-/- and LDLR-/- mice fed chow or Western diet were hydrolyzed by GV sPLA2. Phosphatidylcholine on LDL from LDLR-/- mice fed either a chow or Western diet was hydrolyzed to a greater extent (61.1±0.4% and 45.3±4.6%) than the corresponding fractions from apoE-/- mice (41.7±3.6% and 39.4±1.2%). ApoE-/- LDL induced macrophage foam cell formation in vitro without modification by GV sPLA2, whereas hydrolysis of LDLR-/- LDL was a pre-requisite for foam cell formation. In contrast to findings in LDLR-/- mice, GV sPLA2 deficiency did not significantly reduce atherosclerosis in apoE-/- mice, although collagen content was significantly reduced in lesions of apoE-/- mice lacking GV sPLA2.
The ability of GV sPLA2 to promote atherosclerotic lipid deposition in apoE-/- and LDLR-/- mice may be related to its ability to increase the atherogenic potential of LDL from these mice as assessed in vitro.
Secretory phospholipase A2 (sPLA2)1 enzymes hydrolyze the fatty acid at the sn-2 position of glycerophospholipids 1. Ten sPLA2 isoforms have been described in mammals, and at least seven of these can be detected in human atherosclerotic lesions2. These enzymes have been suggested to promote atherosclerosis through their hydrolyzing activities in the arterial intima 3-5. GV, GX and GIII sPLA2 have been shown to effectively hydrolyze LDL 6-9. Our laboratory has shown that GV sPLA2 hydrolysis of human LDL produces smaller LDL that are susceptible to aggregation 7, modifications that enhance LDL retention in the vessel wall. In addition, GV sPLA2 hydrolysis promotes LDL uptake by macrophages through a pathway that is independent of macrophage scavenger receptors and that involves cell surface proteoglycans 10.
Consistent with in vitro findings, overexpression of GV sPLA2 in bone marrow-derived cells results in increased atherosclerosis in LDLR-/- mice, whereas deficiency of the enzyme results in reduced atherosclerosis 11. However, several considerations led us to question whether GV sPLA2 plays a major role in atherosclerotic lipid deposition in apoE-/- mice. First, in vitro studies have shown that LDL isolated from apoE-/- mice are oxidatively modified and hence induce macrophage foam cell formation in a CD36-dependent manner without the requirement for further modification 12. Secondly, lipoprotein phospholipid (PL) content is altered in apoE-/- mice, such that the SM to PC ratio is relatively high due to increased SM production and decreased SM degradation13. Since previous in vitro data demonstrate that GV sPLA2 hydrolysis of PC in liposomes and human LDL is inhibited by SM 6, 14-16, we speculated that particles from apoE-/- mice may be relatively poor substrates for GV sPLA2 hydrolysis.
In this study, we investigated whether GV sPLA2 contributes to atherogenic processes in apoE-/- mice. Our results show that in contrast to its effects on LDLR-/- mouse or human LDL, GV sPLA2 does not impact the ability of apoE-/- mouse LDL to induce macrophage foam cell formation in vitro. Consistent with these in vitro findings, targeted deletion of GV sPLA2 did not alter atherosclerotic lesion area in apoE-/- mice. Nevertheless, lesional collagen deposition was reduced in apoE-/- mice lacking GV sPLA2, suggesting that the role of this enzyme in atherosclerotic processes in vivo is complex.
C57BL/6, LDLR-/- and apoE-/- mice were obtained from the Jackson Laboratory. CD36-/- mice were kindly provided by R.L. Silverstein 17. The generation of mice lacking syndecan-4-/- has already been described 18. GV sPLA2 deficient were generously provided by Dr. Jonathan Arm 19.
Atherosclerosis was quantified in the aortic arch and the aortic root as described previously 20.
Plasma levels of total cholesterol and triglycerides were determined using enzymatic assays (Wako). PC and SM were measured using colorimetric enzymatic assays (Cayman chemical). Lipoprotein cholesterol distribution was determined by separating pooled plasma samples or lipoprotein fractions on a Superose 6 column 21.
RNA was isolated using TRIzol reagent. Real-time RT-PCR was performed using the standard curve method and normalized to 18S.
LDL (d = 1.019–1.063) were isolated from mouse plasma.
Data are expressed as mean ± S.E. Results were analyzed by Student’s t test or two-way analysis of variance followed by Bonferroni’s post-test. Values of p < 0.05 were considered statistically significant.
Further detailed Experimental Procedures are provided in supplemental materials, available online at http://atvb.ahajournals.org.
Lipoprotein fractions with a density conventionally defined as LDL (d = 1.019-1.063) were isolated from 3 separate pools (~5 ml) of plasma from apoE-/-, LDLR-/-, and apoE-/-xGV sPLA2-/-2 (DKO) mice fed a normal diet, and from a single pool of plasma from apoE-/- and LDLR-/- mice fed a Western diet for 9 weeks. LDL isolated from LDLR-/- mice fed either diet eluted from size exclusion columns as a relatively homogeneous fraction, whereas apoE-/- LDL eluted as two discrete subfractions (supplemental Fig. 1).
Each LDL fraction was analyzed to determine PC and SM content (Table 1). Deficiency of GV sPLA2 in apoE-/- mice did not result in significant alterations in LDL PC or SM compared to LDL from apoE-/- mice. Although there was a trend for reduced PC on LDL from LDLR-/- mice compared to apoE-/- or DKO mice fed a normal diet, this did not reach statistical significance. In contrast, SM content was significantly higher (~3-fold) on LDL from apoE-/- and DKO mice compared to LDL from LDLR-/- mice fed chow diet. As a result, the PC/SM ratio was ~2-fold higher for LDLR-/- LDL compared to apoE-/- LDL. After high fat diet feeding, the SM content of apoE-/- and LDLR-/- LDL was increased ~2.5-fold, whereas the PC content remained relatively unaltered. This led to markedly reduced PC/SM ratios for both LDL fractions compared to their normal diet counterparts. Thus, LDL isolated from apoE-/- mice fed a Western diet was the most SM-enriched fraction, while LDL from LDLR-/- mice fed chow was relatively SM-poor. The PC/SM ratio for LDL from LDLR-/- mice fed Western diet and apoE-/- mice fed chow diet was intermediate.
We investigated whether alterations in SM content are associated with differences in the ability of GV sPLA2 to hydrolyze lipoprotein PC. Aliquots of LDL corresponding to 100 nmol PC were incubated with 10, 25 or 50 U of enzyme for 6 h and the extent of hydrolysis was assessed by measuring the amount of FFA released. For particles from chow-fed mice, significantly more PC was hydrolyzed on LDLR-/- LDL compared to apoE-/- LDL at each enzyme concentration (Fig. 1A). Interestingly, the SM-enriched LDL isolated from either LDLR-/- or apoE-/- mice after Western diet feeding appeared to be poorer substrates for hydrolysis compared to the corresponding fractions from chow-fed mice. Whereas 6 h incubation with 10U of enzyme was sufficient to achieve almost maximal hydrolysis of LDL isolated from chow-fed animals (Fig. 1A), 50 U of GV sPLA2 was relatively less effective in hydrolyzing LDLR-/- and apoE-/- mouse LDL after high fat diet (Fig. 1B). Similar differences in hydrolysis of LDL from the two strains of mice were also evident in time course experiments with 25 U GV sPLA2 (Fig. 1C,D). In order to achieve maximal hydrolysis, LDL particles were incubated overnight with 500 U/mg GV sPLA2. For two separate hydrolysis experiments, the average PC hydrolysis (expressed as the percent of the total PC in the reaction) was ~68% for LDL from LDLR-/- mice fed chow and ~49% for LDL from apoE-/- mice fed chow. LDL from Western diet-fed mice were hydrolyzed to a lesser extent: ~62% hydrolysis for LDLR-/- LDL and ~42% hydrolysis for apoE-/- LDL. Taken together, our data is consistent with previous reports that particles with increased SM content appear to be refractive to hydrolysis by GV sPLA2 6, 15, 16
To investigate whether increased SM content, and not other factors such as the lack of apoE, is primarily responsible for the reduced activity of GV sPLA2 towards apoE-/- LDL, hydrolysis experiments were carried out using apoE-/- and LDLR-/- LDL with and without pre-treatment with SMase. SM depletion of apoE-/- LDL significantly increased the hydrolysis of these particles by GV sPLA2, to a level similar to hydrolysis of SMase-treated LDLR-/- LDL (supplemental Fig. 2).
It was of interest to compare the effect of GV sPLA2 hydrolysis on the inherent atherogenicity of LDLR-/- and apoE-/- LDL. Accordingly, mouse peritoneal macrophages were incubated with maximally hydrolyzed LDL (“GV-LDL”) or control LDL (“mock-LDL”) prepared from the two strains of mice after normal or high fat diet feeding. Hydrolysis of LDL from LDLR-/- mice fed a normal rodent diet resulted in a significant 1.8-fold increase in macrophage CE accumulation after 48 h compared to the corresponding mock-LDL (Fig. 2A). Thus, GV sPLA2 hydrolysis significantly increased the atherogenecity of LDLR-/- LDL, analogous to our findings with human LDL 7. Notably, apoE-/- LDL GV were significantly more effective in inducing foam cell formation compared to control LDLR-/- LDL even without modification, and GV sPLA2 hydrolysis did not further increase macrophage accumulation of CE (Fig. 2A). For both LDLR-/- and apoE-/- mice, high fat diet feeding resulted in the generation of particles that were relatively more atherogenic (Fig. 2B). Similar to LDL from chow-fed mice, GV sPLA2 hydrolysis significantly increased macrophage accumulation of LDL from high fat fed LDLR-/- mice, but not apoE-/- mice. Taken together, our findings indicate that apoE-/- LDL are more atherogenic compared to LDLR-/- LDL and do not require further modification to induce foam cell formation.
Mouse peritoneal macrophages were co-incubated with 50 μg/ml Alexa Fluor-488 labeled apoE-/- LDL (green) and Alexa Fluor-568 LDLR-/- LDL (red) (Fig. 3, upper panels) or these same LDL fractions after hydrolysis by GV sPLA2 (Fig. 3, lower panels). Consistent with measurements of CE accumulation, unmodified apoE-/- LDL, but not LDLR-/- LDL were readily taken up by macrophages from wild type mice (Fig. 3, upper left), whereas both LDL fractions were internalized when hydrolyzed by GV sPLA2 (Fig. 3, lower left).
Peritoneal macrophages from CD36-/- mice exhibited reduced uptake of apoE-/- LDL compared to wild type macrophages (Fig. 3, upper panels), providing evidence that these unmodified particles are endocytosed at least in part through a CD36-dependent mechanism, confirming a previous report 12. Interestingly, CD36-/- cells mediated uptake of both apoE-/- and LDLR-/- LDL after modification by GV sPLA2 (Fig. 3, lower right panel). These data suggest that unmodified apoE-/- LDL, but not LDLR-/- LDL, are bound and internalized by CD36, leading to macrophage foam cell formation. Hydrolysis of either apoE-/- or LDLR-/- LDL exposes a cryptic site for macrophage binding and subsequent uptake that is independent of CD36.
To investigate whether GV sPLA2 contributes to vascular lipid deposition in apoE-/- mice in vivo, 7-8 week-old male and female apoE-/- and DKO animals were fed a Western diet for 9 or 12 weeks. Plasma total cholesterol, triglyceride and lipoprotein cholesterol distributions were not significantly different in DKO mice compared to the corresponding apoE-/- mice for any experiment (supplemental Table 1 and supplemental Fig. 3).
As expected, atherosclerotic lesion area on the intimal surface of the aortic arch increased in both male and female mice with longer duration of Western diet feeding (Fig. 4A, B). However, deficiency of GV sPLA2 did not alter lesion area in either male or female mice fed diet for 9 or 12 weeks. After 9 weeks of diet feeding, the CE content of aortas from male apoE-/- mice (4.0±1.0 μg/mg) was similar to DKO mice (4.0±0.8 μg/mg), indicating that deficiency of GV sPLA2 did not result in alterations in atherosclerotic plaque volume (supplemental Fig. 4A). Interestingly, aortic PC, but not SM, content was significantly higher in DKO mice compared to apoE-/- mice, consistent with reduced GV sPLA2 hydrolytic activity in the intima of DKO mice (supplemental Fig. 4 B,C).
Lesion area in aortic root sections was quantified after Oil red O staining of neutral lipids. There was no significant difference in mean lesion area in the aortic root of DKO mice compared to apoE-/- mice in any experiment (Figure 4C-F), or macrophage content of lesions (Fig. 5A,B). The lack of effect of GV sPLA2 deficiency on the extent of atherosclerosis in apoE-/- mice does not appear to be due to a compensatory upregulation of GX sPLA2 expression in lesions of DKO mice (supplemental Fig. 4D).
We previously reported that overexpression of GV sPLA2 in bone marrow-derived cells significantly increased collagen deposition in atherosclerotic lesions of LDLR-/- mice 11. To assess whether GV sPLA2 can modulate collagen deposition in apoE-/- mice, aortic root sections from male mice fed the Western diet for 9 weeks were stained with picrosirius red, and then visualized by polarized light microscopy. Collagen area, calculated as a percentage of atherosclerotic lesion area, was significantly decreased (~50%) in mice lacking GV sPLA2 (Fig. 5C,D).
A major finding in the present study is that unlike what occurs in LDLR-/- mice, the targeted deletion of GV sPLA2 in apoE-/- mice does not ameliorate atherosclerotic lesion area when these animals are fed a high fat diet. We provide insights into potential mechanisms underlying the different results in LDLR-/- and apoE-/- mice, highlighting evidence that distinct pathogenic mechanisms may prevail in the development of atherosclerosis in these two major animal models of the human disease.
The analysis of LDL isolated from apoE-/- and LDLR-/- mice indicated that apoE-/- particles are relatively SM-enriched compared to LDLR-/- particles. When fed normal rodent diet, the PC/SM ratio was 2-fold lower for apoE-/- LDL compared to LDLR-/- LDL. Similar findings were previously reported and attributed to both an increase in SM synthesis as well as a defect in SM clearance 13. Interestingly, after 9 weeks on high fat diet, the PC/SM ratio decreased in both LDLR-/- and apoE-/- mice, largely due to an increase in the SM content of particles rather than a relative depletion in PC. However, the relative SM enrichment was much more pronounced in apoE-/- mice, resulting in a 3-fold difference in the PC/SM ratio for the 2 strains after high fat diet feeding. PC and SM content of LDL from apoE-/- mice lacking GV sPLA2 (DKO mice) was not significantly different from apoE-/- mice, suggesting that endogenous GV sPLA2 has little effect on the PL composition of circulating lipoproteins. Interestingly, a significant increase in PC was detected in the aortas of DKO mice compared to apoE-/- mice, consistent with differences in sPLA2 activity in the arterial intima for the two strains.
We investigated whether differences in PL composition influence GV sPLA2 hydrolysis. Our data indicate that LDL isolated from LDLR-/- mice is more effectively hydrolyzed compared to apoE-/- mouse LDL. Furthermore, Western diet feeding had a negative impact on the ability of GV sPLA2 to hydrolyze LDL from the 2 strains. Thus, decreased PC/SM ratios for the various mouse LDL fractions was associated with a decrease in GV sPLA2 hydrolysis. Our analysis of VLDL fractions (d < 1.019) from LDLR-/- and apoE-/- mice showed a similar relationship between PC/SM ratio and hydrolysis (data not shown). However, the size heterogeneity of apoE-/- and LDLR-/- VLDL complicates the interpretation of this result, since the impact of particle size on GV sPLA2 activity is not well established 22. Depletion of SM by SMase effectively increased GV sPLA2 hydrolysis of apoE-/- LDL, in agreement with a previous study of human lipoprotein fractions 6. The mechanism by which SM inhibits GV sPLA2 hydrolysis is not clearly understood, although this effect has been attributed to segregation of PC and sPLA2 between disordered and ordered SM/free cholesterol/PC lipid phases 16.
Our data that apoE-/- LDL is a relatively poor substrate for GV sPLA2 hydrolysis compared to LDLR-/- LDL seems unlikely to completely account for the total lack of effect of GV sPLA2 deficiency on atherosclerotic lipid deposition in apoE-/- mice. After overnight incubations, >40% of the PC on apoE-/- LDL was hydrolyzed by GV sPLA2, which might be expected to increase its atherogenicity. We have shown, for example, that hydrolysis of 15-20% of the PL on human LDL is sufficient to induce structural alterations in the particle to promote aggregation 7. To directly assess whether GV sPLA2 increases the atherogenic potential of apoE-/- and LDLR-/- LDL, we quantified macrophage CE content after incubations with LDL fractions. In the case of particles from chow-fed mice, incubations with unmodified apoE-/- LDL resulted in ~2.5-fold more CE accumulation compared to the modest amount induced by unmodified LDLR-/- LDL. In parallel studies with fluorescently-labeled LDLs, we determined that macrophages from wild-type mice effectively internalize unmodified apoE-/- LDL, but not LDLR-/- LDL. Uptake of unmodified apoE-/- LDL was markedly reduced for macrophages lacking CD36. This data is in agreement with a previous study showing that apoE-/- LDL is a high-affinity ligand for CD36, presumably due to its increased state of oxidation 12.
Interestingly, hydrolysis of LDL from chow-fed apoE-/- mice had little effect on macrophage CE accumulation, whereas hydrolysis of LDLR-/- LDL resulted in an almost 2-fold increase. Confocal microscopy studies clearly showed that macrophages internalize GV sPLA2-modified LDLR-/- LDL in a CD36-independent manner to a greater extent than the corresponding unmodified LDL. We interpret our results to suggest that hydrolysis of LDLR-/- LDL brings about structural changes in the particle that alter its interaction with macrophages, leading to enhanced uptake through a CD36-independent mechanism. Thus, LDLR-/- LDL appears to behave similarly to human LDL, where GV sPLA2 hydrolysis promotes macrophage foam cell formation through a CD36-independent process that involves cellular proteoglycans 10. We recently reported that syndecan-4, a cell-surface proteoglycan, mediates macrophage uptake of GV sPLA2-hydrolyzed human LDL 23. The interaction of GV-LDL with proteoglycans is likely due to conformational changes in apoB100 on the phospholipid-depleted particle that expose proteoglycan-binding sites 24. A similar alteration in structure may also occur with apoE-/- LDL, which contains primarily apoB-48, since these particles taken up by CD36-/- macrophages after hydrolysis by GV sPLA2. However, the CD36-independent pathway appears to play only a minor role in foam cell formation induced by apoE-/- LDL, since GV sPLA2 hydrolysis did not significantly increase CE accumulation in wild-type macrophages.
Interestingly, unmodified LDL isolated from both LDLR-/- and apoE-/- mice after Western diet feeding were significantly more effective in promoting foam cell formation compared to the corresponding fractions from chow-fed animals. This increased ability to promote CE accumulation may be due at least in part to the increased cholesterol content of the particles after high fat feeding. When normalized to protein, the total cholesterol of apoE-/- and LDLR-/- LDL was increased 3.0-fold and 5.8-fold, respectively. It is notable that the atherogenicity of LDL from LDLR-/- mice fed Western diet was even further increased after GV sPLA2 hydrolysis, whereas macrophage CE accumulation induced by apoE-/- LDL with and without hydrolysis was similar.
Taken together, our data from in vitro experiments are in striking agreement with results from atherosclerosis studies in apoE-/- and LDLR-/- mice. First, the enhanced macrophage accumulation of unmodified apoE-/- LDL compared to LDLR-/- LDL after normal diet feeding may relate to the spontaneous development of atherosclerosis that only occurs in apoE-/- mice. Second, the inability of GV sPLA2 to enhance macrophage accumulation of apoE-/- LDL in vitro is consistent with the lack of effect of GV sPLA2 deficiency on atherosclerotic lipid deposition in apoE-/- mice. To date, a number of studies have investigated the role of CD36 in atherogenesis in apoE-/- mice 25-29. Although the results from these studies have been inconsistent, the general consensus is that CD36 contributes to atherosclerotic lipid deposition in apoE-/- mice, but the effect may be region-specific 30. On the other hand, studies implicating CD36 in atherosclerotic lesion progression in LDLR-/- mice have not been reported. Based on the current data, it is tempting to speculate that pathways leading to foam cell formation that are independent of CD36 may be critical for lesion formation in the LDLR-/- model. However, the role of CD36 in both apoE-/- and LDLR-/- mice requires further study, given the recent recognition that this receptor plays a role in atherogenic processes that are independent of lipid uptake 31, 32.
Interestingly, despite the lack of effect of GV sPLA2 deficiency on atherosclerotic lesion area in apoE-/- mice, collagen deposition in lesions of DKO mice was significantly reduced compared to apoE-/- mice. Although the mechanisms linking sPLA2 and collagen deposition are not known, previous studies have reported increased collagen deposition in lesions of LDLR-/- mice with macrophage over-expression of either GV 11 or GIIA 33, 34 sPLA2. Our experiments are the first to show that deficiency of sPLA2 leads to decreased lesional collagen area through a mechanism that is independent of its effects on lesional lipid deposition/macrophage foam cell formation.
During the preparation of this manuscript, Shaposhnik et al. reported that a non-selective sPLA2 inhibitor effectively reduced atherosclerotic lesion area and increased the size of the fibrous cap in apoE-/- mice fed a Western diet for 12 weeks 35. Our data would suggest that the drug used in this study may have effects beyond GV sPLA2 inhibition. In summary, our data indicate that an intervention targeting GV sPLA2 modification of LDL is more effective in reducing atherosclerotic lipid deposition in LDLR-/- mice compared to apoE-/- mice. This finding is consistent with in vitro data showing that apoE-/- LDL, unlike LDLR-/- LDL, is sufficiently modified to induce macrophage foam cell formation. Thus, pathways leading to macrophage foam cell formation, a key event in atherogenesis, do not appear to be equivalent for these two animal models of the human disease. In addition, the finding that GV sPLA2 modulates collagen deposition in lesions of both apoE-/- and LDLR-/- mice suggests that mechanisms driving atherosclerotic lesion deposition and plaque complexity may be distinct.
This work was supported by National Institutes of Health grants R01 HL071098 and P01 HL080100 (to N.R.W) and United States Department of Agriculture Fellowship Grant n2005-38420-15825 (to M.Z.).
We thank William Bailey for his technical assistance.
Authors have no conflicts of interest do disclose.