Isoform-Specific Effects of Apolipoprotein E on Atherogenesis
Gene Transduction Studies in Mice
Background We recently used a bone marrow–based gene therapy approach to show that small amounts of retrovirus-derived human apolipoprotein E3 (apoE3) produced by macrophages are protective against early atherosclerosis in apoE-deficient mice.
Methods and Results In the present study, we evaluated whether the effect produced by macrophage-derived apoE3 is related to its ability to bind cellular membranes. To this end, we used apoE2 and apoEcys142, dysfunctional human variants with reduced binding to the LDL receptor or to heparan sulfate proteoglycans, respectively. ApoE-deficient mice, 5 weeks of age, received transplants of apoE−/− bone marrow cells transduced with either parental retrovirus or apoE3, apoE2, or apoEcys142 retroviral vectors. Human apoE was detected by ELISA in the serum of apoE3, apoE2, and apoEcys142 mice as early as 4 weeks after bone marrow transplantation, and at 8 weeks, plasma apoE levels were 55.5±20.3, 50.5±8.7, and 15.3±7.3 μg/dL, respectively. In all groups, cholesterol levels increased with age but were not affected by apoE expression. As previously demonstrated, the lesion area in male apoE3 mice (3808±2224 μm2/section) was 40% smaller than that in control mice (6503±3475 μm2/section). In apoE2 mice, however, the lesion area was similar to that of controls (5991±2771 μm2/section), and apoEcys142 mice showed an unexpected and significant increase in lesion size (10 320±6128 μm2/section). Thus, transplantation with marrow transfected with receptor binding–defective apoE variants did not replicate the antiatherogenic effect of apoE3.
Conclusions These data provide in vivo evidence suggesting that macrophage-derived apoE delays development of atherosclerosis through a receptor-dependent pathway.
Received July 2, 2001; revision received September 7, 2001; accepted September 21, 2001.
Macrophages and macrophage-derived foam cells are active players in the initiation and progression of atherosclerosis.1 Several lines of evidence have demonstrated that macrophage-derived apolipoprotein E (apoE) protects against the development of atherosclerosis, especially in the early stages of foam cell formation.2–4 The mechanism of its antiatherogenic effects, however, remains unclear.
ApoE-deficient (apoE−/−) mice have prominent hypercholesterolemia and develop early atherosclerotic lesions.5,6 When bone marrow from wild-type mice was transplanted into apoE−/− mice, spontaneous development of atherosclerosis was prevented through a dramatic reduction in serum cholesterol levels.7 The association of macrophage apoE with circulating lipoproteins and the consequent induction of their hepatic clearance may be a major cause of the antiatherogenic effects of macrophage apoE in this model. Studies in transgenic mice have shown that macrophage apoE can be protective against atherosclerosis because of its local effects in the vessel wall, irrespective of changes in the levels of circulating lipoproteins.2,3 We also recently used a bone marrow–based gene therapy approach with a retroviral system to provide clear evidence that small amounts of macrophage-derived apoE can delay foam cell formation.4 When apoE−/− bone marrow cells were transfected with a retrovirus expressing either mouse apoE or human apoE3 and transplanted into 5-week-old apoE−/− recipient mice, there was a significant reduction in foam cell formation, seen at euthanization at 13 weeks of age, without changes in serum cholesterol concentrations. This effect, however, was lost at later stages of atherogenesis.
The antiatherogenic mechanism of macrophage-derived apoE can be related to several pathways linked to lipid metabolism (such as the internalization of trapped lipoproteins or the induction of cholesterol efflux), or it may be the result of one or more of its other activities, such as the inhibition of platelet aggregation, antiproliferative effects on T lymphocytes, effects on oxidation, and effects on inflammatory processes.8,9 Because apoE can function either as a soluble protein or by binding to membrane receptors, it is important to determine whether disruption of binding to the LDL receptor (LDLR) or to heparan sulfate proteoglycans (HSPGs) reduces or eliminates known effects of normal apoE in the vessel wall.
Several variants of apoE have been identified that have different degrees of defective binding to either the LDLR or HSPG.10 These apoE isoforms provide an excellent tool to dissect the role of cell-surface binding in the local function of apoE. We chose to use human apoE2 and apoEcys142 because of their different properties in binding the LDLR and HSPG. ApoE2 differs from apoE3 by having a cysteine instead of an arginine at residue 158. It is highly defective in LDLR-binding activity (<1% of normal apoE3 activity) but displays normal binding to HSPG.11 ApoEcys142 contains 2 amino acid substitutions: an arginine substitution for cysteine in the receptor-binding region at residue 142 and an arginine for cysteine substitution located at residue 112. ApoEcys142 is markedly defective in HSPG binding, despite having 25% to 45% of the LDLR-binding activity of apoE3.11
Using apoE2 and apoEcys142 in retroviral vectors, we were able to stably express low levels of these variants in macrophages of apoE−/− mice. Neither of these variants had the beneficial effects of apoE3 on atherogenesis, and apoEcys142 actually worsened lesion progression. Our results indicate that the role of macrophage-derived apoE in the artery wall is dependent on its LDLR-binding capability as well as its ability to bind to HSPG.
All mice used for these experiments were on a C57BL/6 background and were purchased from Jackson Laboratories, Bar Harbor, Maine. Mice were maintained in microisolator cages and given a rodent chow diet (No. 5010; Purina Mills, Inc) containing 4.5% fat and autoclaved water acidified to pH 2.8. The Western-type diet given to mice used for immunocytochemical analyses contained 21% fat and 0.15% cholesterol (Teklad). Animal care and experimental procedures were performed according to the regulations of Vanderbilt University’s Institutional Animal Care and Usage Committee.
Cloning of ApoE Variants and Construction of Retroviral Vectors
The human apoE cDNAs for apoE3, apoE2, and apoEcys142 were amplified from an expression vector driven by the cytomegalovirus promoter and cloned into the pLXSN vector as described previously.4 The apoE inserts were confirmed by sequencing with the ThermoSequenase kit (Amersham).
Retroviral Bone Marrow Transplantation
Parental (control) and human apoE3-, apoE2-, and apoEcys142-expressing producer cell lines were prepared and assayed for the presence of recombinant wild-type virus by the S+ L− assay as previously described.4 ApoE−/− bone marrow was collected from hind leg femur and tibia of donor mice and incubated for 48 hours in medium containing interleukin-3 and interleukin-6. After this incubation, marrow was divided equally into 4 groups and plated onto producer cell lines containing either the parental retroviral vector, apoE3, apoE2, or apoEcys142 and incubated for an additional 48 hours. The transduced marrow was collected and counted, and 1×106 to 3×106 cells were injected via the tail vein into lethally irradiated (900 rads) recipient apoE−/− mice.
Blood samples were collected by retro-orbital venous plexus puncture. Serum was separated by centrifugation, preserved with 1 mmol/L PMSF (Sigma), and stored at −20°C until use. Serum total cholesterol levels were determined with clinical reagents on a microtiter plate assay as previously described.4
Plasma lipoproteins were separated by ultracentrifugation and by fast performance liquid chromatography (FPLC). Pooled serum samples (200 μL) were adjusted to d=1.019 g/mL with KBr and ultracentrifuged. The upper phase was removed, and the bottom fraction was further ultracentrifuged to produce the fractions of d=1.019 to 1.040, d=1.040 to 1.210, and d>1.210 g/mL. FPLC analyses were performed as previously described.12
ApoE Western Blotting Analyses
Serum samples were tested for the presence of human apoE by Western blot analyses as described.4 Similar Western blot analyses were performed on serum after separation of lipoprotein fractions by ultracentrifugation and FPLC after the fractions had been desalted by centrifugation with Amicon Microcon-100 columns (Millipore Co).
ELISA for Human ApoE
Serum levels of human apoE in mice receiving apoE-transduced bone marrow were determined by a specific ELISA system as follows. Ninety-six–well ELISA plates were coated with anti-human apoE polyclonal antibody (BioDesign) at a concentration of 1 μg/mL. Plates were then washed with PBS/0.5% Tween-20 (PBS-T) and blocked with 10% FBS/PBS. The diluted serum standard and samples were then added to each well. A human serum sample (apoE concentration ≈5 mg/dL) was diluted to obtain a standard curve that was linear from 0.31 to 1000 μg/dL. Standards and samples were incubated overnight at 4°C. Plates were washed and then incubated with biotin-conjugated anti-human apoE (BioDesign). After washing, 0.25 μg/well (diluted in 10% FBS/PBS) of avidin-conjugated peroxidase (Sigma) was added and incubated for 30 minutes. Plates were washed 8 times, after which 100 μL of 2,2′-azino-bis-(3-ethylbenzthiazoline-6-sulfonic acid) (Sigma) was added to each well. Color development was allowed to proceed for 15 to 20 minutes. The plates were then read on a microplate reader (Molecular Devices) at a wavelength of 405 nm. Serum apoE levels were calculated by comparison to the standard curve and are expressed as μg/dL.
Lesion Area Quantification and Immunohistochemical Analyses for ApoE
For quantification of lesion area of the proximal aorta, 10-μm frozen sections were cut, stained with oil red O, and quantified as described.4 Immunocytochemical analysis of tissue samples for macrophages and mouse apoE was performed essentially as described.4
Statistical analyses were performed with a 1-way ANOVA followed by the Student-Newman-Keuls post hoc test.
Construction of Human ApoE–Containing Retroviral Vectors
Retroviral constructs containing either no insert (control), apoE3, apoE2, or apoEcys142 were produced. ApoE cDNA inserts were sequenced and amplified and digested with HhaI to confirm base-pair changes in the receptor-binding region by the genotyping method of Hixson et al13 (data not shown). Five-week-old apoE−/− mice received transplants of apoE−/− bone marrow cells transduced with parental retrovirus (control animals) or vector containing the apoE3, apoE2, or apoEcys142 cDNAs. In control mice, human apoE was not detected before or after bone marrow transplantation (BMT) (Figure 1). Before BMT (baseline), serum samples from all groups showed no apoE in serum. In recipients of apoE3-, apoE2-, and apoEcys142-transduced marrow, the expression of human apoE was observed at 4 weeks after BMT and continued until euthanization at 8 weeks after BMT. Serum levels of apoE3, apoE2, and apoEcys142 are shown in Table 1. Although apoEcys142 levels showed a trend toward being lower in male mice, this difference did not reach statistical significance. In contrast, both apoE2 and apoEcys142 levels were significantly lower than apoE3 levels in female animals (P<0.05).
Serum cholesterol levels were measured before BMT and after transplantation (Table 2). In all groups, cholesterol levels were elevated 8 weeks after BMT relative to baseline values (P<0.001); however, there were no differences in post-BMT serum total cholesterol levels between groups. FPLC analyses of post-BMT serum samples showed no difference in the distribution of lipoproteins in animals receiving apoE3, apoE2, or apoEcys142 (Figure 2). The apoE from all 3 groups was largely associated with the bulk of the lipoprotein mass, which consisted of VLDL- and IDL/LDL-size lipoproteins (a representative sample of apoE2 is shown in Figure 2). These results were confirmed on serum samples that were separated by ultracentrifugation (Figure 3). Most apoE3 was associated with the VLDL (d<1.019, lane 3), with lesser amounts as IDL/LDL (d=1.019 to 1.040, lane 4) and HDL (d=1.040 to 1.210, lane 5). No apoE was associated with the bottom fraction (d>1.210) in any group.
All animals were euthanized at 13 weeks of age and analyzed for atherosclerotic lesion area. In male animals, the extent of arterial lesions in apoE3 mice (n=16) at 8 weeks after BMT was 40% smaller than those in controls (n=13) (3808±2224 versus 6503±3475 μm2/section) (Figure 4A). The lesion area in recipients of apoE2-transduced marrow (5991±2771 μm2/section, n=9) was similar to that of controls, whereas the apoEcys142 mice showed significantly larger plaque area than all other groups (10 320±6128 μm2/section, n=13, P<0.05). Results in female mice were slightly different from those of male animals (Figure 4B), with apoE3 (n=11) and control (n=11) animals having similar lesion areas (5235±769 and 6271±846 μm2/section, respectively). The lesion areas of apoEcys142 female mice (11 976±2199 μm2/section, n=9) and apoE2 female mice (9731±3204 μm2/section, n=10) were significantly greater than those of both control and apoE3-expressing mice (P<0.05).
Because lesion area was low and contained few macrophages, immunocytochemical detection of apoE in lesions was difficult. To alleviate this, additional male mice in each group received transplants and were placed on a high-fat diet to accelerate lesion progression. These animals were euthanized, and proximal aorta sections were immunostained for either the macrophage marker MOMA-2 or human apoE (Figure 5). Areas of MOMA-2 staining as well as apoE staining were quantified by digitizing morphometric analysis. It was found that 30% to 50% of the total macrophage area was also positive for apoE in both apoE3- and apoE2-expressing mice (Figure 5C through 5F). In contrast, apoEcys142 expression was much lower (Figure 5H). When sections were permeabilized with 0.3% Triton X-100, apoEcys142 became detectable within the lesion (data not shown).
We previously demonstrated that low levels of macrophage-derived apoE are able to protect against the development of atherosclerotic lesions in apoE−/− mice.4 In this study, we determined whether the protective effect of apoE was due to its ability to bind surface receptors by expressing human apoE3, apoE2, or apoEcys142 in arterial macrophages through a retroviral BMT technique. Confirming our previous results,4 apoE3 expression in male mice caused a 40% reduction in lesion area compared with control mice receiving parental virus–transduced marrow. The lesion size in recipient mice reconstituted with apoE2-producing marrow, however, was similar to that of controls. Surprisingly, low-level expression of apoEcys142 produced aortic plaques that were significantly larger than those seen in controls, indicating that simple point mutations in apoE can modulate atherogenesis through a wide spectrum that includes beneficial and worsening effects.
Studies using both apoE2 transgenic mice and adenovirus-induced apoE2 expression have demonstrated that when apoE2 is expressed at levels high enough to reduce serum cholesterol levels, atherosclerotic lesion area is concomitantly reduced.14,15 In contrast, mice expressing high levels of apoEcys142 develop severe hypercholesterolemia, accumulate β-VLDL, and are susceptible to accelerated lesion formation.16,17 These studies, combined with our data, indicate a protective effect of apoE2 only when accompanied by serum cholesterol reductions but a proatherogenic effect of apoEcys142 even in the absence of serum cholesterol changes. This provides evidence that the local role of apoE in the artery wall may be in part due to its HSPG-binding capability.
Expression of apoEcys142 was lower both in serum samples and in aortic lesions. This may be due to the lower viral titer of the apoEcys142-producing cell line (2.5×105 cfu/mL) than that of the other cell lines (4×105 and 4.75×105 cfu/mL for apoE3 and apoE2, respectively). A difference in processing and/or clearance of apoEcys142-containing lipoprotein particles, however, cannot be ruled out. With regard to arterial macrophage apoE expression, it should be noted that the immunocytochemical method used detects mostly extracellular or cell-surface antigen expression. This is an important distinction, for 2 reasons. First, the 30% to 50% apoE staining detected in the apoE3- and apoE2-expressing mice seems high, considering that only 1% to 5% of macrophages are secreting the protein. Given the HSPG-binding capability of these apoE isoforms and the secluded milieu of the macrophage-rich lesion, the apoE is most likely secreted from a small percentage of the arterial macrophages and then trapped by HSPGs on other macrophages present in the lesion. Second, considering that when tissue sections were permeabilized with Triton X-100, apoE staining was detectable within lesions from apoEcys142-expressing mice (data not shown), the lack of apoEcys142 detection in the artery wall may reflect not a lack of secretion from arterial macrophages but rather an absence of membrane binding.
In this study, we found large and significant differences between male and female mice. In our previous study, we demonstrated that male mice expressing apoE from 5 to 13 weeks of age had a significant reduction in lesion area compared with controls and that female mice expressing apoE from 10 to 26 weeks of age had no changes in lesion area.4 We attributed these differences to the time course of apoE expression and to the inability of low-level apoE expression to affect more complex lesions. We could not exclude the possibility, however, that the sex difference could account for the reduced lesion progression in the second group. Sex differences in the development of atherosclerosis in mice have also been reported by other groups in unrelated studies.18–20
Several possible mechanisms by which apoE may reduce foam cell formation have been proposed. A working model includes effects on reverse cholesterol transport,8 a mechanism by which the cell regulates the balance between intake and output of cholesterol and through which it opposes the atherogenic pressures that turn it into a foam cell. Although no convincing evidence indicates that the apoE isoforms may affect cholesterol efflux differently, it is noteworthy that the cell surface HSPGs appear to be intimately interrelated with the ability of apoE to induce efflux.21,22 The lack of protective effects of apoE2 or apoEcys142 in this study may be at least partly related to the inability of these variants to bind to membrane structures and drive cholesterol efflux from a physiologically appropriate location. Lin et al23 recently demonstrated that when macrophages lose the surface pool of apoE, the cell responds with a significant accumulation of cholesterol. Thus, the difference in the macrophage cell-surface pool of apoE can underlie the observed difference in the lipid deposition in the artery wall between our study groups. The apparent worsening effect of apoEcys142 on atherogenesis may indicate the larger physiological role played by HSPGs in the control of cholesterol efflux.
Several recent studies have indicated that apoE may have effects unrelated to plasma lipid changes but with direct beneficial repercussions on vascular health. Inhibition of platelet aggregation, as well as antiproliferative, antioxidant, and anti-inflammatory actions, have been presented as putative physiological effectors of the modulating role of apoE in atherogenesis.8,9 Whether these other properties of apoE require receptor and/or HSPG binding remains to be determined. Our observation that single amino acid changes affecting HSPG and LDLR binding eliminate the protective effect of apoE3 demonstrates that apoE binding to cell surfaces is crucial to its function as an antiatherogenic agent within the artery wall.
This work was supported in part by NIH grants HL-53989 and HL-57986. Drs Hasty and Major are recipients of NIH postdoctoral fellowships. Drs Fazio and Linton are Established Investigators of the AHA. This work was part of Dr Hasty’s doctoral dissertation.
The first 2 authors contributed equally to this work.
Shimano H, Ohsuga J, Shimada M, et al. Inhibition of diet-induced atheroma formation in transgenic mice expressing apolipoprotein E in the arterial wall. J Clin Invest. 1995; 95: 469–476.
Bellosta S, Mahley RW, Sanan DA, et al. Macrophage-specific expression of human apolipoprotein E reduces atherosclerosis in hypercholesterolemic apolipoprotein E-null mice. J Clin Invest. 1995; 96: 2170–2179.
Hasty AH, Linton MF, Brandt SJ, et al. Retroviral gene therapy in apoE-deficient mice: apoE expression in the artery wall reduces early foam cell lesion formation. Circulation. 1999; 99: 2571–2576.
Nakashima Y, Plump AS, Raines EW, et al. ApoE-deficient mice develop lesions of all phases of atherosclerosis throughout the arterial tree. Arterioscler Thromb. 1994; 14: 133–140.
Reddick RL, Zhang SH, Maeda N. Atherosclerosis in mice lacking apoE: evaluation of lesional development and progression. Arterioscler Thromb. 1994; 14: 141–147.
Linton MF, Atkinson JB, Fazio S. Prevention of atherosclerosis in apoE deficient mice by bone marrow transplantation. Science. 1995; 267: 1034–1037.
Curtiss LK. ApoE in atherosclerosis: a protein with multiple hats. Arterioscler Thromb Vasc Biol. 2000; 20: 1852–1853.
Fazio S. Recent insights into the pathogenesis of type III hyperlipoproteinemia. Trends Cardiovasc Med. 1993; 3: 191–195.
Ji Z-S, Fazio S, Mahley RW. Variable heparan sulfate proteoglycan binding of apolipoprotein E variants may modulate the expression of type III hyperlipoproteinemia. J Biol Chem. 1994; 269: 13421–13428.
Linton MF, Fazio S. Bone marrow transplantation normalizes lipid and lipoprotein levels in apolipoprotein E deficient mice. Circulation. 1994; 90 (suppl I): I-133.Abstract.
Hixson JE, Vernier DT. Restriction isotyping of human apolipoprotein E by gene amplification and cleavage with HhaI. J Lipid Res. 1990; 31: 545–548.
Huang Y, Schwendner SW, Rall SCJr, et al. Hypolipidemic and hyperlipidemic phenotypes in transgenic mice expressing human apolipoprotein E2. J Biol Chem. 1996; 271: 29146–29151.
Fazio S, Lee Y-L, Ji Z-S, et al. Type III hyperlipoproteinemic phenotype in transgenic mice expressing dysfunctional apolipoprotein E. J Clin Invest. 1993; 92: 1497–1503.
Fazio S, Sanan DA, Lee Y-L, et al. Susceptibility to diet-induced atherosclerosis in transgenic mice expressing a dysfunctional human apolipoprotein E (Arg112, Cys142). Arterioscler Thromb. 1994; 14: 1873–1879.
Tangirala RK, Rubin EM, Palinski W. Quantitation of atherosclerosis in murine models: correlation between lesions in the aortic origin and in the entire aorta, and differences in the extent of lesions between sexes in LDL receptor-deficient and apolipoprotein E-deficient mice. J Lipid Res. 1995; 36: 2320–2328.
Smith JD, Trogan E, Ginsberg M, et al. Decreased atherosclerosis in mice deficient in both macrophage colony-stimulating factor (op) and apolipoprotein E. Proc Natl Acad Sci U S A. 1995; 92: 8264–8268.
Obunike JC, Pillarisetti S, Paka L, et al. The heparin-binding proteins apolipoprotein E and lipoprotein lipase enhance cellular proteoglycan production. Arterioscler Thromb Vasc Biol. 2000; 20: 111–118.
Lin CY, Duan HW, Mazzone T. Apolipoprotein E-dependent cholesterol efflux from macrophages: kinetic study and divergent mechanisms for endogenous versus exogenous apolipoprotein E. J Lipid Res. 1999; 40: 1618–1626.