Relative Contributions of Apolipoprotein(a) and Apolipoprotein-B to the Development of Fatty Lesions in the Proximal Aorta of Mice
Abstract Transgenic mice expressing transgenes for both human apolipoprotein B-100 (h-apoB) and apolipoprotein(a) [apo(a)] were fed a high-fat, atherogenic diet for 14 weeks to examine the effect of lipoprotein(a) [Lp(a)] on the development of aortic fatty lesions. The extent of lesions in the proximal region of the aorta of Lp(a) mice was measured by use of a computer-assisted image analysis of 20 sections per animal and compared with that of nontransgenic mice as well as mice expressing either the apo(a) or h-apoB transgene. The control (n=23) and apo(a) (n=22) transgenic mice had very small mean lesion areas (607 versus 128 μm2 per section). The h-apoB–expressing mice (n=20) had significantly higher mean lesion areas (3288 μm2 per section) than either the control or apo(a) transgenic animals. Coexpression of apo(a) and h-apoB transgenes resulted in only a modest increase in lesion area (4678 μm2 per section, n=19). Thus, the expression of human apo(a) in C57BL/6/SJL hybrid mice fed an atherogenic diet failed to significantly potentiate the development of aortic fatty lesions in the absence or presence of high levels of h-apoB.
- Received July 10, 1995.
- Accepted August 25, 1995.
Lipoprotein(a) [Lp(a)] is a cholesteryl ester–rich lipoprotein that circulates in human plasma and is composed of a large glycoprotein, apo(a), attached by a disulfide linkage to the apoB of LDL.1 Numerous cross-sectional and prospective studies in humans have found an association between high plasma concentrations of Lp(a) and both coronary and peripheral atherosclerosis.2 The molecular mechanisms responsible for these associations have not been conclusively elucidated.
Apo(a) closely resembles the plasma zymogen plasminogen.3 Apo(a), like plasminogen, has two major domains: a kringle-rich domain and a protease domain. In plasminogen, there are five copies of a cysteine-rich motif that has been referred to as a kringle (K). Apo(a) does not contain the first three K repeats of plasminogen (K1 through K3) but has a variable number of tandem copies of a sequence that resembles K4. In both proteins, K4 is followed by a single K5 motif and a protease domain. Tissue plasminogen activator (TPA) cleaves plasminogen within the K5 sequence to release the protease domain of plasminogen (plasmin) from the K domain. Plasmin plays a key role in fibrinolysis as well as in the activation of growth factors.2 In apo(a), there is an amino acid substitution corresponding to the site of cleavage of plasminogen by TPA and, therefore, apo(a) is not processed by TPA. Although apo(a) contains a sequence at its C-terminus that shares 94% sequence identity with the protease domain of plasminogen, apo(a) has not been conclusively demonstrated to have protease activity.
It is not known whether the primary pathological effect of apo(a) is atherogenic or thrombogenic. In in vitro assays, Lp(a) has been shown to compete with plasminogen for activation and binding to endothelial cells, and in this way it may indirectly interfere with thrombolysis.4 5 Pathological evaluation of human atherosclerotic lesions has revealed that apo(a) is present in the arterial wall in direct proportion to its plasma concentration.6 Lp(a) binds to numerous components of the extracellular matrix and may accumulate passively in lesions.7 More recently, it has been proposed that Lp(a) may promote atherosclerosis by indirectly stimulating the proliferation of smooth muscle cells in the arterial wall.8 9 10 TGFβ1, which inhibits smooth muscle cell proliferation, is secreted as a propeptide and becomes biologically active after it is cleaved by plasmin. Apo(a) has been shown to inhibit generation of the active form of TGFβ1 both in vitro and in vivo.9 10
Some of the difficulties in defining the role of apo(a) in atherosclerosis have been encountered because, until recently, there was not a convenient animal model in which to study the atherogenicity of apo(a). Apo(a) has an unusual species distribution. It is a major cholesterol-carrying lipoprotein in the hedgehog but is not present in the plasma of any other species except old-world monkeys, great apes, and humans.11 12 Apo(a) has been expressed in the plasma of mice by introduction of an h-apo(a) transgene under the control of the mouse transferrin promoter.13 In humans, apo(a) in plasma is covalently linked to LDL, but in the transgenic mouse, plasma apo(a) is not covalently attached to lipoproteins.13 In spite of this, when these apo(a) transgenic mice were placed on a high-fat diet, they developed 20-fold larger oil red O–positive intimal lesions in their proximal aortas than did nontransgenic littermate control animals.14 Immunocytochemical analysis of the aortic lesions revealed focal colocalization of apo(a) and mouse apoB within the fatty lesions.
To develop a more physiological animal model in which to study the atherogenicity of Lp(a), mice have been produced that coexpress the h-apo(a) and h-apoB transgene. In these mice, all of the plasma apo(a) is covalently linked to h-apoB within the LDL fraction.15 16 To determine whether Lp(a) is more atherogenic than apo(a) alone and whether Lp(a) is more atherogenic than h-apoB, animals expressing the apo(a) transgene alone, the h-apoB-100 transgene alone, or both transgenes together were fed an atherogenic diet for 14 weeks, and the areas of oil red O–positive lesions in their proximal aortas were measured and compared.
Mice and Diets
The C57BL/6 and SJL mice were obtained from the Jackson Laboratory (Bar Harbor, Me). Two mouse lines expressing either an h-apo(a)13 or h-apoB15 transgene (621-1) were maintained by crossing transgene-positive offspring into (C57BL/6×SJL)F1 mice. The transgenic mice used in these studies were hemizygous for either one or both of the transgenes and are referred to as follows: apo(a) transgenic mice, apo(a) mice; h-apoB transgenic mice, h-apoB mice; and mice hemizygous for both the apo(a) and h-apoB transgenes, Lp(a) mice. The nontransgenic littermates served as controls.
The mice were housed in a conventional animal facility with free access to food and water on a 14-hour/10-hour light/dark cycle. All of the litters were weaned between 21 and 28 days and maintained on a chow diet (Teklad 7001) until they reached 8 to 10 weeks of age. At that time, 31 apo(a) mice (16 males, 15 females), 21 h-apoB mice (12 males, 9 females), 24 Lp(a) mice (14 males, 10 females), and 34 littermate controls (19 males, 15 females) were placed on a high-fat, high-cholesterol diet (Harlan Teklad) that contained 7.5% casein, 2.5% dextrose, 1.625% sucrose, 1.625% dextran, 7.5% cocoa butter, 1.25% cholesterol, 0.5% sodium cholate, and 1.25% cellulose, as well as vitamins and minerals (AIN-76). Five apo(a) mice (4 females, 1 male), 5 Lp(a) mice (1 female, 4 males), and 5 h-apoB (5 females) mice were maintained on a regular mouse chow diet. The mice were maintained on either the high-fat or regular chow diet for 14 weeks, at which time they were sent to the Gladstone Institute of Cardiovascular Disease (University of California at San Francisco) for morphometric analysis of intimal lesions in the proximal aorta. The genotypes of the mice were not revealed to those performing the morphometric analyses.
Plasma Apolipoprotein and Lipid Measurements
Venous blood samples were obtained from a subset of each group of mice by retro-orbital sinus puncture before the high-fat diet was initiated and 1 week before the animals were killed. The plasma was isolated by centrifugation at 5000g for 10 minutes, and aliquots were stored at −20°C. Lipid measurements were made within 1 week of collection of the blood. Plasma cholesterol and triglyceride levels were determined enzymatically with assay kits obtained from Boehringer Mannheim Corp (Biochemical Div) and Sigma Chemical Co, respectively. To qualitatively estimate changes in the concentrations of the apolipoproteins, aliquots of the d<1.215 g/mL fraction were size-fractionated on a 4% to 12% gradient sodium dodecyl sulfate–polyacrylamide minigel and stained with Coomassie brilliant blue R-250.17
Apo(a) levels were determined by a double monoclonal antibody–based enzyme immunoassay.18 The capture antibody used in the assay, A6, is specific for the K4 type 2 repeat of apo(a), and the detection antibody, A40, recognizes an epitope in the penultimate K4 repeat in the tandem array, corresponding to K436.3 The results are expressed as total Lp(a) mass. Because apo(a) is not covalently attached to apoB-100 in the apo(a) transgenic mice, the plasma apo(a) level can be estimated by multiplying the Lp(a) concentration by 0.138, since apo(a) comprises this fraction of the total mass of Lp(a). The total plasma h-apoB levels (including both apoB and apoB-48) were determined by use of a solid-phase competitive radioimmunoassay.15 19
Preparation of Aortic Sections
The areas of the oil red O–positive lesions in the proximal aortas were quantified as described previously.19 Briefly, mice were anesthetized using methoxyflurane (Pitman-Moore, Inc) and the hearts were first perfused with phosphate buffered saline and then with 10% neutral-buffered formalin before they were removed from the thorax together with the aortic arch. The tissues were postfixed overnight at 4°C in a phosphate-buffered, 10%-formalin solution. We isolated the aortic root by severing the heart approximately 1 mm below the aortic valve. The tissues were embedded in Tissue-Tek OCT compound (Miles, Inc) in cryostat molds. Samples were frozen by immersion in liquid nitrogen and stored at −20°C. Sequential sections, each 10 μm thick, were cut by use of a Reichert 2800 cryostat. Every other section was collected onto glass slides starting at the aortic sinus, which was identified by the appearance of the semilunar valve leaflets, and extending over the next 1.2 mm. Thus, sections were obtained from the entire aortic sinus and 400 μm of the proximal aorta. The slides were incubated in 0.5% oil red O in propylene glycol for 4 hours and counterstained in Mayer’s hematoxylin for 1 minute. The area of fatty lesion revealed by oil red O staining was measured in 20 sections from the proximal aorta of each mouse.19 Both the mean cross-sectional area of oil red O–positive lesions and the total lesion area were determined by automated pixel counting by use of a computerized image 1/at image analysis system (software version 4.03a, Universal Imaging Corp). In addition, 5 sections that were uniformly spaced to sample the entire sinus region were also evaluated in all of the female mice from each of the four groups. The transition between the sinus and the proximal aorta was used as a starting site for analyses in either direction: distally into the proximal aorta or proximally into the sinus.
Means, standard errors, and variance were determined and Students’ t tests were run with the statview ii program (Abacus Concepts, Inc). Pair-wise comparisons between the groups were made using two-factor repeated-measure ANOVA.
The plasma lipid levels from a subset of each of the four groups of mice were compared before and after the mice were fed a high-fat, high-cholesterol diet for a 14-week period (the Table⇓). In each group, the mean plasma levels of cholesterol were higher in male than in female mice.15 Expression of h-apoB, with or without coexpression of apo(a), was associated with a significant increase in both cholesterol and triglyceride levels, as was previously reported.15 19 After consumption of the high-fat diet, there was a significant increase in plasma cholesterol level as well as a decrease in plasma triglyceride level in all four groups. In both the Lp(a) and h-apoB mice, there was an approximately twofold increase in the plasma levels of total h-apoB, as determined by use of a radioimmunoassay, after the high-fat diet. Immunoblot analysis of the plasma by use of an h-apoB–specific antibody that recognizes both apoB-48 and apoB-100 revealed that much of the apoB was apoB-48 rather than apoB-100 (data not shown). This was to be expected, since ≈75% of human apoB-100 is edited in the livers of the h-apoB transgenic mice.19 The plasma concentrations of apo(a) were higher in male than in female mice in both the apo(a) and Lp(a) mice and increased after ingestion of the high-fat diet in both groups. There was an ≈20% decrease in the amount of apo A-I after the high-fat diet, as assessed by use of Coomassie blue staining of the d<1.215 g/mL fraction (data not shown).
The areas of lesions that stained with oil red O in the proximal aorta were quantified in 22 apo(a) mice (11 males, 11 females), 19 Lp(a) mice (10 males, 9 females), and 20 h-apoB mice (11 females, 9 males) and in 23 nontransgenic littermate controls (13 males, 10 females) after the mice were fed the atherogenic diet for 14 weeks. The mean lesion areas for each animal are given in Fig 1⇓. Although it has been noted previously that female mice tend to be more susceptible to atherosclerosis than males are,19 20 both sexes were included in the present study because a significant increase in aortic lesions in both male and female apo(a) transgenic mice had been found in a previous study.14 A total of 5 mice from each group were maintained on a chow diet during the dietary challenge period, and no atherosclerotic lesions were found in the aortas of these mice (data not shown).
Most of the apo(a) transgenic mice and the control mice had no detectable oil red O staining within the proximal aorta. The control mice had a higher mean lesion area per section than the apo(a) transgenic mice, although the difference was not statistically significant (607 versus 128 μm2 per section). Female mice tended to have larger lesion areas than the males, which is consistent with prior studies.19 20 Both the h-apoB and the Lp(a) mice had significantly larger lesion areas than the apo(a) and control mice (3288 μm2 and 4678 μm2, respectively). Interestingly, there was no significant difference in mean lesion area between animals that expressed the h-apoB transgene and those that expressed both the h-apoB and apo(a) transgenes.
These results from the proximal aorta were contrary to prior studies that found a significant increase in fatty lesions in response to an atherogenic diet with use of the same line of transgenic apo(a) mice in the same genetic background.14 21 We examined the possibility that the precise localization of lesions in the aortic root might have varied between the two studies. To explore this possibility, five sections from the sinus region were examined in the subset of female mice from the four different experimental groups. The results of these studies are shown in Fig 2⇓. The area of lipid staining in both the control and apo(a) mice was not significantly different (1071±783 μm2 and 2898±825 μm2; P<.13, two-tailed t test). Both the apoB and Lp(a) transgenic mice had higher mean lesion areas (11 061±2880 and 26 973±6656 μm2), with the Lp(a) mice having approximately a twofold-higher mean lesion area than the apoB mice, which was of borderline statistical significance (P<.04, two-tailed t test).
A panel of images representative of the degree of atherosclerosis seen in the sinus and proximal aortic regions of nontransgenic control, apo(a), and Lp(a) mice is shown in Fig 3⇓. The lesions found in the proximal aortas of some control (Fig 3A⇓) and apo(a) (Fig 3C⇓) mice were extremely small (<1000 μm2 per section). Many animals from these groups had no lesions at all. The modest lesions seen in the proximal aortas of Lp(a) mice (Fig 3E⇓) were of similar size to those seen in h-apoB mice (not shown). The sinus regions revealed slightly larger lesion areas than the proximal aorta, but these lesions were still minimal in the case of control (Fig 3B⇓) and apo(a) (Fig 3D⇓) mice. The Lp(a) mice (Fig 3F⇓) had fairly large lesions in the sinus regions, as did the h-apoB mice (not shown).
The major finding of the present study is the absence of a significant difference in the amount of aortic lipid staining in the proximal aorta after ingestion of an atherogenic diet in mice expressing h-apoB compared with mice expressing Lp(a). Expression of h-apoB was associated with a significant increase in lesion area compared with control animals, and coexpression of apo(a) was not associated with any further increase in mean lesion area in the proximal aorta. However, in the sinus region there was an approximately twofold-higher mean lesion area in Lp(a) mice when compared with h-apoB mice, which was statistically significant (P<.04). No significant differences were found between apo(a) transgenic mice and littermate controls in either the sinus region or in the proximal aorta; both of these groups of mice had little or no oil red O staining in the proximal aorta.
Ingestion of a high-fat, synthetic diet had the expected effects on plasma lipid levels in control as well as in transgenic animals.22 23 24 Plasma cholesterol levels increased significantly, and plasma triglyceride levels fell. The reduction in plasma triglyceride levels in association with ingestion of a synthetic, high-fat diet in mice is well described,19 22 23 although the mechanism responsible has not been elucidated. There was a significant increase in plasma apo(a) levels in both apo(a) and Lp(a) mice after the high-fat diet. This was an unexpected finding because the apo(a) transgene is under the control of the mouse transferrin promoter, the expression of which is not known to be influenced by dietary factors.13 It has not been determined if the increase in plasma apo(a) concentration was due to an increase in apo(a) synthesis or a decrease in its clearance. High-fat diets are associated with downregulation of LDL receptor activity. In humans, the LDL receptor does not appear to play a major role in the removal of apo(a) from plasma.25 The mouse, however, does not normally express apo(a), and the sites of removal of apo(a) from plasma have not been determined. Expression of the apo(a) transgene in mice in which the LDL receptor has been inactivated by homologous recombination26 results in a one-and-a-half-fold to twofold increase in plasma apo(a) levels (data not shown), implicating the LDL receptor either directly or indirectly in the clearance of h-apo(a) from mouse plasma.
The results of these studies contrast sharply with prior studies by others14 21 that reported a dramatic increase in the development of oil red O–positive lesions in the proximal aorta of apo(a) transgenic mice after consumption of an atherogenic diet. These conflicting results cannot be attributed to differences in the apo(a) transgene. The animals used in the present studies are from the same line as that used in the prior studies14 21 and have been maintained by breeding the apo(a) transgenic–positive animals into (C57BL/6XSJL)F1 animals and thus should contain ≈50% C57BL/6 genes and ≈50% SJL genes. Subtle differences in the proportion of genes from these two strains, which vary dramatically in their propensity for the development of fatty lesions in the proximal aorta,20 could account for the observed differences in results. The C57BL/6 strain is highly susceptible to the development of aortic lesions, especially if the high-fat diet is supplemented with cholic acid.22 In contrast, the SJL strain is very resistant to the development of aortic lesions.22 The importance of genetic background to the development of fatty lesions in mice is dramatically illustrated by comparing the mean lesion area of lipid staining in the h-apoB transgenic mice in the present study to the results of a prior study19 in which the mice had been backcrossed twice into C57BL/6 and thus had ≈87.5% C57BL/6 genes. These latter mice had much more complicated and extensive aortic lesions,19 perhaps owing in part to a diet that was higher in saturated fat (16% versus 7.5%) and that was administered for a longer period (18 versus 14 weeks).
The discrepancy between our findings in apo(a) transgenic mice and those of prior studies14 21 cannot be attributed to differences in the region of the aorta that was analyzed, since analysis of the more proximal aorta also failed to disclose a significant difference in oil red O staining (Fig 2⇑). Other potential explanations for the observed differences between our results and the results of prior studies are subtle dietary or environmental differences. There might have been differences in the amounts of modified or oxidized lipids fed to the two groups of mice. There might also have been uncontrolled differences in overall cleanliness or pathogen exposure that contributed to lesion development, although the animals in at least one of the prior studies14 that used apo(a) transgenic mice were housed in the same facility. Pathogen exposure may have an impact on the expression of a number of diseases, which may include atherosclerosis. If any of these factors were responsible for the observed differences in results, it would be expected that the differences would be reflected in the controls as well as in the apo(a) transgenic mice. However, the control mice in this study had an approximately 10-fold-greater mean lesion area (607 μm2 versus 58 μm2).14
The results of this study should not be interpreted to mean that apo(a) is not atherogenic. This study strongly suggests, however, that the effects of apo(a) on lesion development in the mouse are modest in comparison with the effects of h-apoB. The amount of lipid staining found in association with apo(a) expression in this study or in prior studies14 21 is significantly less than observed after other genetically altered mice were challenged with an atherogenic diet. For example, in mice in which the apo E gene was inactivated, the mean lesion area was 260 642 μm2 per section after ingestion of a western-type diet.27 Moreover, coincident with the evaluation of these mice, we quantified the lesion area in the proximal aorta of a 129/C57BL/6 mouse in which the LDL receptor had been inactivated (LDLR −/− mouse) after ingestion of the same diet for 14 weeks, and the lesion area was 189 542 μm2 per section.
It is possible that we were unable to detect an effect of apo(a) on atherosclerosis because of relatively low levels of gene expression. Plasma concentrations of Lp(a) in animals fed the atherogenic diet exceeded the 95th percentile of Lp(a) when compared with human levels, the only exception being the apo(a) female mice.28 It may be that higher concentrations of plasma apo(a) or more prolonged exposure is required to manifest an atherogenic effect in a species such as mice that is relatively resistant to atherosclerosis. In mice expressing apo E3-Leiden, a mutant form of apolipoprotein E that is associated with type III hyperlipidemia in humans, fatty lesions did not develop when mice expressed low levels of apo E3-Leiden.29 Only mice with high concentrations of apo E3-Leiden in their plasma developed significant fatty lesions after being challenged with an atherogenic diet. Efforts are now being directed toward developing an animal model in which there is a pronounced effect of apo(a) expression on the development of aortic lesions, which can be used to evaluate the effects of other factors, both pharmacological and dietary, on lesion development.
Selected Abbreviations and Acronyms
|TGFβ1||=||transforming growth factor β-1|
|TPA||=||tissue plasminogen activator|
This work was supported by grants from the National Institutes of Health (No. HL-20948 to Dr Hobbs, No. HL-30086 to Dr Marcovina, and No. HL-41633 to Dr Young) and by the Perot Family Foundation. Dr Hobbs is an Established Investigator for the American Heart Association.
Utermann G; Scriver CR, Beaudet AL, Sly WS, Valle D, eds. Lipoprotein(a): The Metabolic Basis of Inherited Disease. New York, NY: McGraw Hill Inc; 1995:1887-1912.
Rath M, Niendorf A, Reblin T, Dietel M, Krebber HJ, Beisiegel U. Detection and quantification of lipoprotein(a) in the arterial wall of 107 coronary bypass patients. Arteriosclerosis. 1989;9:579-592.
Van der Hoek YY, Sangrar W, Côté GP, Kastelein JJP, Koschinsky ML. Binding of recombinant apolipoprotein(a) to extracellular matrix proteins. Arterioscler Thromb. 1994;14:1792-1798.
Kojima S, Harpel PC, Rifkin DB. Lipoprotein(a) inhibits the generation of transforming growth factor β: an endogenous inhibitor of smooth muscle cell migration. J Cell Biol. 1991;113:1439-1445.
Grainger DJ, Kirschenlohr HL, Metcalfe JC, Weissberg PL, Wade DP, Lawn RM. Proliferation of human smooth muscle cells promoted by lipoprotein(a). Science. 1993;260:1655-1658.
Laplaud PM, Beaubatie L, Rall SC Jr, Luc G, Saboureau M. Lipoprotein(a) is the major apoB-containing lipoprotein in the plasma of a hibernator, the hedgehog (Erinaceus europaeus). J Lipid Res. 1988;29:1157-1170.
Chiesa G, Hobbs HH, Koschinsky ML, Lawn RM, Maika SD, Hammer RE. Reconstitution of lipoprotein(a) by infusion of human low density lipoprotein into transgenic mice expressing human apolipoprotein(a). J Biol Chem. 1992;267:24369-24374.
Linton MF, Farese RV, Chiesa G, Grass DS, Chin P, Hammer RE, Hobbs HH, Young SG. Transgenic mice expressing high plasma concentrations of human apolipoprotein B100 and lipoprotein(a). J Clin Invest. 1993;92:3029-3037.
Callow MJ, Stoltzfus LJ, Lawn RM, Rubin EM. Expression of human apolipoprotein B and assembly of lipoprotein(a) in transgenic mice. Proc Natl Acad Sci U S A. 1994;91:2130-2134.
Marcovina SM, Albers JJ, Gabel B, Koschinsky ML, Gaur VP. The effect of the number of apolipoprotein(a) kringle 4 domains on the immunochemical measurement of lipoprotein(a). Clin Chem. 1995;41:246-255.
Purcell-Huynh DA, Farese RV Jr, Johnson DF, Flynn LM, Pierotti V, Newland DL, Linton MF, Sanan DA, Young SG. Transgenic mice expressing high levels of human apolipoprotein B develop severe atherosclerotic lesions in response to a high-fat diet. J Clin Invest. 1995;95:2246-2257.
Liu AC, Lawn RM, Verstuyft JG, Rubin EM. Human apolipoprotein A-l prevents atherosclerosis associated with apolipoprotein(a) in transgenic mice. J Lipid Res. 1994;35:2263-2267.
Nishina PM, Verstuyft JG, Paigen BA. Synthetic low and high fat diet for the study of atherosclerosis in the mouse. J Lipid Res. 1990;31:859-869.
LeBoeuf RC, Caldwell M, Kirk E. Regulation by nutritional status of lipids and apolipoproteins A-I, A-II, and A-IV in inbred mice. J Lipid Res. 1994;35:121-133.
Rader DJ, Mann WA, Cain W, Kraft HG, Usher D, Zech LA, Hoeg JM, Davignon J, Lupien P, Grossman M, Wilson J, Brewer HB Jr. The low density lipoprotein receptor is not required for normal catabolism of Lp(a) in humans. J Clin Invest. 1995;95:1403-1408.
Ishibashi S, Brown MS, Goldstein JL, Gerard RD, Hammer RE, Herz J. Hypercholesterolemia in low density lipoprotein receptor knockout mice and its reversal by adenovirus-mediated gene delivery. J Clin Invest. 1993;92:883-893.
Linton MF, Atkinson JB, Fazio S. Prevention of atherosclerosis in apolipoprotein E-deficient mice by bone marrow transplantation. Science. 1995;267:1034-1037.
Marcovina SM, Albers JJ, Jacobs DR Jr, Perkins LL, Lewis CE, Howard BV, Savage P. Lipoprotein(a) concentrations and apolipoprotein(a) phenotypes in Caucasians and African Americans. Arterioscler Thromb. 1993;13:1037-1045.
Hoegenesch H, van Vlijmen BJM, van den Moogdenberg AMJM, Gijbels MJJ, van der Boom H, Frants RR, Hofker MH, Havekes LM. Diet-induced hyperlipoproteinemia and atherosclerosis in apolipoprotein E3-Leiden transgenic mice. J Clin Invest. 1994;93:1403-1410.