Apolipoprotein E2 Transgenic Rabbits

Transgenic rabbits were produced that expressed high plasma levels (30–70 mg/dl) of human apolipoprotein (apo) E2(Cys-158), an apoE variant associated with the human genetic disorder type III hyperlipoproteinemia (HLP). Male transgenic rabbits fed normal chow had up to 8-fold (289 ± 148 mg/dl) and 15-fold (697 ± 452 mg/dl) increases in plasma total cholesterol and triglycerides, respectively, compared with nontransgenic males. Female transgenic rabbits had only a modest hyperlipidemia (total cholesterol, 140 ± 46 mg/dl; total triglycerides, 174 ± 66 mg/dl). Both sexes displayed the hallmarks of type III HLP: β-migrating very low density lipoproteins (β-VLDL) (intestinal and hepatic remnant lipoproteins) and significantly increased VLDL and intermediate density lipoproteins. Apolipoprotein E2-containing VLDL particles were cleared from the circulation more slowly and were more resistant to lipoprotein lipase-mediated lipolysis than normal VLDL. Only females had increased high density lipoproteins (HDL) (40%), which were shifted from typical small HDL to larger HDL1. Plasma apoE2 was predominantly associated with β-VLDL in males and with HDL in females. To ascertain reasons for the phenotypic gender difference, we treated the male transgenic rabbits with 17α-ethinyl estradiol. Estrogen treatment for 10 days dramatically decreased total cholesterol (73%) and triglycerides (89%) and converted β-VLDL to pre-β-migrating VLDL. Concomitantly, lipoprotein lipase and hepatic lipase activities increased by 90%, low density lipoprotein receptor activity was stimulated significantly, apoE2 was redistributed to HDL, and HDL were converted to HDL1. Conversely, ovariectomy in female transgenic rabbits significantly increased total cholesterol (75%), triglycerides (117%), and β-VLDL, while decreasing lipoprotein lipase and hepatic lipase activities by 35% and redistributing apoE2 to the β-VLDL. Thus, estrogen status appears to be responsible for much of the gender difference of the lipoprotein phenotype, mainly by modulating both lipase and low density lipoprotein receptor activities. Furthermore, transgenic rabbits fed normal chow for 11 months developed fatty streaks, and some had more advanced atherosclerotic lesions, especially around the aortic arch and proximal abdominal aorta. The lesions were more extensive in males, roughly correlating with the magnitude of the hyperlipidemia. Therefore, high plasma levels of human apoE2 in transgenic rabbits result in a type III HLP phenotype, in which males have both more severe hyperlipidemia and more extensive atherosclerosis than females.

Type III hyperlipoproteinemia (HLP) 1 is a genetically determined disorder of lipid metabolism in humans that is characterized by both hypercholesterolemia and hypertriglyceridemia. There is an associated accumulation of abnormal plasma lipoproteins, namely ␤-migrating very low density lipoproteins (␤-VLDL), which are cholesterol-enriched remnant lipoproteins derived from both intestine and liver (for a review, see Ref. 1). Affected subjects are predisposed to premature atherosclerosis (1,2). Type III HLP can be caused either by receptor binding-defective variants of apolipoprotein (apo) E (1,3,4), most commonly apoE2(Cys-158) (1), or by apoE deficiency (5,6). Apolipoprotein E normally functions as a ligand for remnant uptake by lipoprotein receptors, mainly in the liver (for reviews, see Refs. 3 and 7-9). In type III HLP, defective apoE causes impaired receptor-mediated lipoprotein catabolism that leads to ␤-VLDL accumulation in the plasma (and thus hyperlipidemia). Type III HLP caused by apoE2 homozygosity is an adult-onset disease with a striking predominance in males: almost all affected women are postmenopausal, which suggests that estrogen status may modulate the expression of type III HLP in humans (1).
For the study of the pathogenesis of this disorder, models of type III HLP have been created in transgenic mice expressing apoE(Arg-112, Cys-142) (9) or apoE-Leiden (10), both of which are associated with dominant transmission of the disease in humans (11,12), or apoE2(Cys-158) (13,14). These transgenic mice are yielding important information for our understanding of the mechanisms of this disorder. These models have lipoprotein profiles similar to those in human type III HLP, including hypercholesterolemia, hypertriglyceridemia, and ␤-VLDL accumulation in plasma (9, 10) but lack the gender differences and spontaneous atherosclerosis seen in humans with the disease (15,16). In contrast, apoE knockout mice (a model of human apoE deficiency) develop spontaneous atherosclerosis, but their lipoprotein phenotype differs from human type III HLP in that their plasma triglyceride levels are only slightly increased, and the markedly cholesterol-enriched remnant li-poproteins that accumulate are atypical of ␤-VLDL in human type III HLP (17,18). Recently, we generated an animal model of recessive type III HLP by expressing human apoE2 in transgenic mice in which both hypo-and hyperlipidemia developed, depending on the expression levels of apoE2 (13). However, because of the general resistance of mice to the development of atherosclerosis and because of some significant differences in lipoprotein metabolism between mice and humans, additional animal models of type III HLP that resemble the human disease more closely may be required to investigate the pathogenesis of this disorder and the associated susceptibility to atherosclerosis.
Rabbits have been used successfully to express several transgenes (19 -22). As an experimental model, rabbits have several advantages over mice. Rabbits have higher levels of apoBcontaining lipoproteins than mice (23), a lipoprotein profile more like that of humans, and a pattern of hepatic apoB100 and intestinal apoB48 synthesis resembling that of humans. Like humans and unlike mice, rabbits have cholesteryl ester transfer protein, which has been reported to be elevated in human type III HLP patients (24). The larger plasma volumes in rabbits permit metabolic studies of lipoprotein subclasses and facilitate lipoprotein turnover studies. Furthermore, rabbits are very susceptible to the development of atherosclerosis, and the lesions can resemble those seen in human atherosclerosis (25,26). For these reasons, we generated transgenic rabbits that express human apoE2 in the liver. Here, we report that high expression levels of apoE2 in transgenic rabbits lead to a type III HLP phenotype with a significant gender difference in which estrogen plays a major role. The transgenic rabbits also develop spontaneous atherosclerosis.

EXPERIMENTAL PROCEDURES
Materials-New Zealand White rabbits were purchased from Jackson Laboratories (Bar Harbor, ME). Plasmid pBSSK was purchased from Pharmacia (Uppsala, Sweden). A Superose 6 column (Pharmacia) was used on a Pharmacia fast performance liquid chromatography system. The Centricon concentration filters were from Amicon (Lexington, MA). Cholesterol and triglyceride standards were from Abbott (North Chicago, IL) and Boehringer Mannheim, respectively. The automated system for lipid analysis (Kinetic Microplate Reader) was from Molecular Devices (Menlo Park, CA). All of the reagents for lipoprotein agarose gels were from Ciba Corning (Palo Alto, CA). The ECL chemiluminescence detection kit for Western blots was purchased from Amersham Life Sciences (Little Chalfont, Buckinghamshire, United Kingdom).
DNA Construct-The DNA construct used to generate transgenic rabbits was pHEG1LEcys158 (13). It contained the complete human apoE2 gene together with 5 kilobase pairs of its 5Ј-flanking sequence and 1.7 kilobase pairs of its 3Ј-flanking sequence, which was ligated to a 3.8-kilobase pair downstream fragment containing the hepatic control region for this gene (13,27). In both transgenic mice and rabbits, this construct yields high expression of apoE in the liver and little expression in any other tissue (9,19,27).
Preparation of Transgenic Rabbits-Transgenic rabbits were prepared by microinjecting the above construct into New Zealand White rabbit embryos (19,20). At 6 weeks of age, the resulting rabbits were weaned and maintained on a normal chow diet. The presence of the transgene was detected by Southern blotting of 10 g of DNA with a human apoE cDNA probe (28) and by immunoblotting of 1 l of plasma with a human-specific anti-apoE polyclonal antiserum (29). In the Western blot assay, rabbit apoE and human apoE2 were semiquantitated by comparing the densitometric readings of the sample bands with those of different concentrations of purified rabbit or human apoE. All experiments were performed under protocols approved by the Committee on Animal Research, University of California, San Francisco.
Lipoprotein Separation and Analysis-Blood was collected from the intermedial auricular artery of 4 -11-month-old rabbits that had been fasted overnight. EDTA was used as anticoagulant at a final concentration of 10 mM. Plasma was obtained by centrifugation at 14,000 rpm (microcentrifuge) for 10 min at 4°C, and samples were stored for no more than 2 days at 4°C in the presence of 1 mM phenylmethylsulfonyl fluoride as a protease inhibitor. Lipoproteins in 200 l of plasma were separated by fast performance liquid chromatography on a Superose 6 column, as described previously (29,30). The major lipoprotein classes eluted from the column were pooled  and concentrated with Centricon filters (fractions 16 -18, VLDL; fractions 19 -22, intermediate density lipoproteins (IDL); fractions 23-27, low density lipoproteins (LDL) and a subclass of high density lipoproteins (HDL) called HDL 1 ; and fractions 28 -33, HDL). For agarose gel electrophoresis, 2-l aliquots of concentrated lipoproteins were run on precast agarose gels (1%) for 1 h at 90 V. The gels were dried and stained with Fat Red 7B. In some cases, to analyze apolipoprotein content or distribution of apoE2 in various lipoproteins, the pooled samples representing different lipoprotein classes were separated by 12% SDS-polyacrylamide gel electrophoresis, followed by detection with either Coomassie staining or anti-human apoE immunoblotting.
Cholesterol and triglycerides were measured on total plasma and on chromatographic fractions by an enzymatic colorimetric method adapted for use with a microplate reader (9,30). The cholesteryl ester content of VLDL and IDL was determined by subtracting the free cholesterol from the total cholesterol value. The HDL cholesterol concentrations were measured after precipitation of apoB-containing lipoproteins by heparin/manganese chloride (31) (Wako Pure Chemical Industries, Osaka, Japan).
VLDL Clearance-The VLDL (d Ͻ 1.006 g/ml) were isolated by ultracentrifugation from normal or apoE2 transgenic male rabbits and iodinated by the iodomonochloride method (32). The 125 I-labeled normal or apoE2 VLDL (25 g of protein/kg of body weight) were injected into the marginal ear vein of normal male rabbits (three rabbits for each group). At designated times, blood samples were drawn from the middle artery of the ear into tubes containing EDTA, the plasma was separated in a microcentrifuge (14,000 rpm for 10 min at 4°C), and radioactivity was measured as described (33).
Lipolysis of VLDL in Vitro-To determine the ability of normal and apoE2 VLDL to serve as substrates for lipase-mediated lipolysis, 30 g of VLDL triglyceride was incubated for 30 min at 37°C with 10 l of VLDL-depleted postheparin rabbit plasma, which was collected from normal rabbits 10 min after intravenous injection of heparin (50 units/ kg). The incubation was performed in either the presence (to measure hepatic lipase) or absence (to measure total lipolytic activity) of 1.2 M NaCl. After incubation, the levels of released free fatty acids were determined by an enzymatic colorimetric method (34) (Wako Chemicals U. S. A., Richmond, VA). The lipoprotein lipase (LPL)-mediated lipolysis was calculated as the difference between total lipolysis and hepatic lipase-mediated lipolysis (35).
Estrogen Studies in Male ApoE2 Transgenic Rabbits-Three male nontransgenic and three male apoE2 transgenic rabbits at 8 months of age were injected intramuscularly with 17␣-ethinyl estradiol (Sigma) at a dose of 100 g/kg/day for 10 days (36). After 0, 5, and 10 days of estrogen treatment, preheparin plasma was collected for plasma lipid and lipoprotein analysis (see above). Postheparin plasma was collected 10 min after injection of 150 units of heparin/kg before and after 10 days of estrogen treatment. The postheparin LPL and hepatic lipase activities were determined with a triolein/Triton X-100 emulsion assay system, as described previously (35).
Monoclonal antibody 9D9 is cleared rapidly from plasma as a result of specific binding to the rabbit cell surface LDL receptors, indirectly indicating the level of these receptors in vivo (37). To investigate the effect of estrogen treatment on LDL receptor activity, IgG of the 9D9 antibody was purified from mouse ascites fluid, labeled with 125 I, and administered intravenously to rabbits. Its removal from plasma was monitored as described previously (38).
Estrogen Studies in Female ApoE2 Transgenic Rabbits-To investigate further the effect of sex hormones on lipids and lipoproteins, three female nontransgenic and three female apoE2 transgenic rabbits at 9 months of age underwent a bilateral ovariectomy under ketamine-HCl (50 mg/kg)/acepromazine (1 mg/kg) anesthesia, as described previously (39). At 0, 5, and 10 days after ovariectomy, preheparin plasma was collected for plasma lipid and lipoprotein analyses (see above). Postheparin plasma was collected 10 min after injection of 150 units of heparin/kg before and 10 days after ovariectomy. Postheparin LPL and hepatic lipase activities were determined as described (35).
Quantitation of Atherosclerotic Lesions-For the atherosclerosis studies, seven male and seven female transgenic rabbits and four male and four female nontransgenic rabbits were maintained on a normal chow diet after weaning. At 11 months, the rabbits were euthanized with an overdose of pentobarbital. After thoracotomy and laparotomy, the heart, the entire aorta, and the common iliac, external iliac, and femoral arteries were carefully dissected free and immediately immersed in cold Dulbecco's phosphate-buffered saline. After the adven-titial fat had been completely removed, the aortas were opened longitudinally from the arch, pinned out flat on styrofoam sheets, and fixed with 3% paraformaldehyde in Dulbecco's phosphate-buffered saline for at least 24 h. After staining with Sudan IV (40) to visualize the atherosclerotic lesions, the aortas were photographed on Kodak Ektachrome Lumiere film. The slides were scanned, and large-format 24-bit color image files were produced by Edward Herderick (Laboratory of Vascular Diseases, Ohio State University, Columbus, OH). Atherosclerosis was measured by computer-based quantitative morphometry of the area of sudanophilic lesions relative to the aortic surface area. Each aorta was analyzed in five nonoverlapping regions: arch (extending from its origin at the heart to the level of the ductus scar), thoracic aorta, proximal abdominal aorta, distal abdominal aorta, and terminal aorta (including the iliofemoral arteries). Morphometry was performed with an Image 1/AT image analysis system (Universal Imaging Corp., West Chester, PA). Networked Silicon Graphics Iris Indigo computers linking the Gladstone Institute of Cardiovascular Disease and the Laboratory of Vascular Diseases were used to prepare probability of occurrence maps (41), which display differences in the distribution (and area) of sudanophilic lesions between experimental groups in the study. For qualitative characterization of the lesions, segments of aorta from selected animals were cross-sectioned in a cryostat and stained with Oil Red O for lipids (42) and with trichrome for collagen (43).
Statistical Analysis-Mean lipid levels are reported as the mean Ϯ S.D. Data for the extent of atherosclerotic lesions in the pinned-out aortas are reported as the mean Ϯ S.E. Differences in cholesterol levels, triglyceride levels, and the extent of atherosclerosis in the pinned-out aortas were evaluated by analysis of variance. Differences in VLDL composition and lipid changes in response to estrogen treatment and ovariectomy were assessed by t test.

Effects of ApoE2 Expression on Plasma Lipid Levels-Trans-
genic rabbits (F1) expressing high levels of apoE2 (30 -70 mg/ dl) were generated from one male founder whose plasma apoE2 concentration was 43 mg/dl. Table I summarizes the plasma lipid levels in both the F1 apoE2 transgenic rabbits and their nontransgenic littermates at 5 months of age. Nontransgenic rabbits had minor gender differences in their lipid values, with females having somewhat higher total and HDL cholesterol levels than males. Nontransgenic rabbits expressed 3.8 Ϯ 0.9 mg/dl of endogenous apoE (as determined in nine rabbits by Western blot with purified rabbit apoE as a standard; data not shown). All of the apoE2 transgenic rabbits were hyperlipi-demic, and there were dramatic gender differences. Male transgenic rabbits had a very significant hyperlipidemia, with about 8-and 15-fold increases in plasma total cholesterol and triglyceride levels, respectively, compared with nontransgenic males, whereas female transgenic rabbits had about 3-and 6-fold higher plasma total cholesterol and triglyceride levels, respectively, than nontransgenic females. Furthermore, apoE2 levels correlated strongly with both plasma cholesterol and triglyceride levels in males but only with plasma cholesterol levels in females (Fig. 1). Even at the higher apoE2 levels, the females were resistant to the development of severe hypertriglyceridemia and much less prone to the development of hypercholesterolemia. The HDL cholesterol levels in the female transgenic rabbits were 200% higher than in males (p Ͻ 0.05) and about 40% higher than in nontransgenic female rabbits (Table I).
These results indicate that female rabbits are more resistant to apoE2-induced hyperlipidemia than males.
Analysis of Lipoproteins and Apolipoproteins-Plasma lipoproteins in the transgenic rabbits were analyzed by gel filtration chromatography on a Superose 6 column (Fig. 2). Compared with a nontransgenic littermate, in which most of the cholesterol was in the HDL and LDL fractions ( Fig. 2A), a male transgenic rabbit with an apoE2 plasma concentration of 63 mg/dl had dramatically increased VLDL and IDL cholesterol levels (Fig. 2B). A female rabbit with a similar plasma concentration of apoE2 (59 mg/dl) had much less severe hyperlipidemia, with lower VLDL and IDL but higher HDL (Fig. 2C), than the male transgenic rabbit (Fig. 2B). The ratio of triglyceride to cholesterol in VLDL was about 2:1 in the male transgenic rabbit and about 1:1 in the female transgenic rabbit. Another significant gender difference was that the plasma HDL elution peak was shifted from fraction 30 (typical HDL) to fraction 26 (HDL 1 ) in the female transgenics but not in the males.
Column fractions (as shown in Fig. 2A) representing different lipoproteins were pooled, concentrated, and analyzed by agarose gel electrophoresis (Fig. 3). The nontransgenic plasma lipoproteins showed typical pre-␤-migrating VLDL, a slightly slower migrating IDL, ␤-migrating LDL, and ␣-migrating HDL (Fig. 3A). In contrast, both male and female transgenic rabbits had ␤-migrating VLDL in both the VLDL and IDL fractions (Fig. 3, B and C). The HDL 1 levels were only slightly higher in male transgenic than in male nontransgenic rabbits; this difference was much more pronounced in the females. In addition, in the female transgenic rabbits, both the HDL 1 and HDL fractions had particles that migrated at the ␣2 position (Fig.  3C). The ␣2-migrating HDL were present in all female transgenic rabbits analyzed.
Apolipoprotein E2 distribution among the various lipoprotein classes was analyzed for both male and female transgenic rabbits. Pooled Superose 6 lipoprotein fractions of each rabbit representing the major lipoprotein classes were separated by  12% SDS-polyacrylamide gel electrophoresis, and apoE2 was detected by anti-human apoE immunoblotting (Table II). Male transgenic rabbits had much more apoE2 in the VLDL and IDL fractions than female transgenic rabbits (49 versus 21%). Plasma concentrations of endogenous rabbit apoE were unaffected by the expression of apoE2 (3.8 Ϯ 0.9 versus 3.5 Ϯ 0.8 mg/dl for 9 nontransgenic and 12 transgenic rabbits, respectively), and most of the endogenous rabbit apoE was present in the HDL fractions in both males and females (data not shown).
The Superose 6-isolated ␤-VLDL from both male and female transgenic rabbits were characterized in more detail (Table  III). The ratio of apoE2 to rabbit apoE in ␤-VLDL was about 7:1 in transgenic males but only 4:1 in transgenic females. The higher triglyceride levels in ␤-VLDL from males suggested by Fig. 2 were confirmed by determining the ratio of cholesterol (or cholesteryl esters) to triglycerides, which in males was about half of that in females. The ␤-VLDL from the female transgenics also had a lower apoB48:apoB100 ratio (0.15 Ϯ 0.04 versus 0.35 Ϯ 0.04). Negative stain electron microscopy revealed that the ␤-VLDL particles from males and females had similar diameters (34 Ϯ 12 versus 37 Ϯ 13 nm), but the particles from females had many surface protrusions that were not apparent in particles from males (data not shown). We presume these protrusions represent excess surface resulting from rapid lipolysis of the core, which would be consistent with the markedly lower triglyceride content of ␤-VLDL from females. This analysis demonstrates a substantial gender difference in apoE2 transgenic rabbits, with females having less severe hyperlipidemia, lower ␤-VLDL, lower ratios of triglycerides to total cholesterol and of apoB48 to apoB100 in the ␤-VLDL, and higher levels of ␣2-migrating HDL 1 .
Effect of ApoE2 on VLDL Clearance and Lipolysis-To understand the mechanism of remnant accumulation in transgenic rabbits, the effect of apoE2 expression on VLDL clearance was investigated. For this study, 125 I-labeled normal or apoE2 VLDL (d Ͻ 1.006 g/ml) were injected into normal rabbits, and the clearance of these particles was monitored. As shown in Fig. 4A, the presence of apoE2 dramatically slowed the clearance of transgenic VLDL (compared with normal VLDL) from the plasma of normal rabbits, suggesting that the hyperlipidemia in transgenic rabbits was at least partly caused by  defective interaction of apoE2 with lipoprotein receptors, most likely the LDL receptor. Since apoE2 transgenic rabbits developed not only hypercholesterolemia but also hypertriglyceridemia and since increasing evidence indicates that apoE2 might impair lipolysis of triglyceride-rich lipoproteins (44,45), the effect of apoE2 on lipolysis of rabbit VLDL was also investigated. As shown in Fig.  4B, apoE2 in the transgenic VLDL inhibited LPL-mediated lipolysis by 77% compared with normal VLDL, while hepatic lipase-mediated lipolysis was not affected. These results, taken together with the VLDL clearance data, suggest that the hypertriglyceridemia in transgenic rabbits is caused by apoE2induced impairment of both VLDL clearance and lipolysis.

Effects of Estrogen on Lipids and Lipoproteins in Male ApoE2
Transgenic Rabbits-To investigate the potential role of sex hormones in modulating type III HLP and to explain some of the gender differences seen in apoE2 transgenic rabbits, we treated male transgenic rabbits with 17␣-ethinyl estradiol. The results are summarized in Table IV. Estrogen treatment (100 g of 17␣-ethinyl estradiol/kg/day) of three male transgenic rabbits resulted in 68 and 73% decreases in total cholesterol after 5 and 10 days of treatment, respectively, without significant changes in plasma levels of apoE2. The decrease in triglycerides was even more pronounced, with 86 and 89% reductions after 5 and 10 days of treatment, respectively. Superose 6 chromatography revealed that the decrease in total cholesterol and triglycerides was due to a dramatic decrease in VLDL and IDL (␤-VLDL) (data not shown). Agarose gel electrophoresis demonstrated a complete "conversion" of ␤-VLDL to normal pre-␤-migrating VLDL in estrogen-treated male apoE2 rabbits (Fig. 5). The HDL cholesterol level increased from 24 Ϯ 4 to 40 Ϯ 6 mg/dl in the male transgenic rabbits treated with estrogen for 10 days (Table IV). The increased HDL were shifted from the typical HDL to the larger HDL 1 , which migrated at the ␣2 position (Fig. 5). This profile was very similar to that observed in untreated female apoE2 transgenic rabbits (compare Figs. 3C and 5). Equally interesting, the majority of the apoE2 shifted from the ␤-VLDL and IDL fractions (49%) to the HDL 1 and HDL fractions (89%) after 10 days of estrogen treatment (Table II). Plasma concentration and lipoprotein distribution of the endogenous rabbit apoE were relatively unaffected by estrogen treatment (data not shown). Similar results for total cholesterol and triglycerides were also obtained in estrogen-treated nontransgenic rabbits but with an overall lesser response (Table IV). Superose 6 chromatography revealed that estrogen treatment of the male nontransgenic rabbits for 10 days significantly decreased VLDL, IDL, and LDL and slightly decreased HDL (data not shown), as reported previously (36).
The effect of estrogen treatment on postheparin lipolytic activity was determined in these animals. Postheparin plasma LPL and hepatic lipase activities in the male apoE2 transgenic rabbits increased by 93 and 94%, respectively, after 10 days of estrogen treatment (Table IV). The lipase values in the estrogen-treated male transgenic rabbits at 10 days were similar to those in the female apoE2 transgenic rabbits before ovariectomy (Table V).
The effect of estrogen treatment on LDL receptor activity was determined by measuring the clearance of LDL receptor antibody 9D9 from estrogen-treated and untreated male rabbits. As shown in Fig. 6, estrogen treatment for 10 days significantly stimulated clearance of the 9D9 antibody in both non-

FIG. 4. Effects of apoE2 on VLDL clearance and VLDL lipolysis.
A, 125 I-Labeled nontransgenic or apoE2 transgenic VLDL (25 g of protein/kg of body weight) were injected intravenously into normal male rabbits, and their clearance was monitored as described under "Experimental Procedures." Results are presented as mean Ϯ S.D., n ϭ 3. B, 30 g of nontransgenic or apoE2 transgenic VLDL triglycerides was incubated with 10 l of VLDL-depleted postheparin rabbit plasma for 30 min at 37°C in the absence or presence of 1.2 M NaCl. Lipase activities were calculated as described under "Experimental Procedures." Results are presented as mean Ϯ S.D. of determinations in four rabbits. *, p Ͻ 0.001 versus nontransgenic VLDL by t test. FFA, free fatty acid. transgenic and apoE2 transgenic male rabbits, suggesting probable LDL receptor up-regulation by estrogen treatment. Moreover, the apoE2 transgenic rabbits displayed a trend toward slower 9D9 clearance (lower LDL receptor expression) compared with nontransgenic rabbits (Fig. 6). Taken together, these data indicate that estrogen exerted its major lipid-lowering effect by stimulating both lipolytic and LDL receptor activities.

Effects of Estrogen on Lipids and Lipoproteins in Female ApoE2
Transgenic Rabbits-To determine if a deficiency of endogenous estrogen production would enhance the type III HLP phenotype, female apoE2 transgenic rabbits were ovariectomized. Removal of ovaries in three female apoE2 transgenic rabbits led to 48 and 75% increases in total cholesterol 5 and 10 days after ovariectomy, respectively, without significant changes in plasma levels of apoE2 (Table V). The increase in plasma triglycerides was even more pronounced, with 112 and 119% elevations 5 and 10 days after ovariectomy, respectively. Superose 6 chromatography demonstrated that ovariectomy in female apoE2 transgenic rabbits significantly increased VLDL and IDL (Fig. 7), both of which were ␤-migrating (data not shown). The ratio of triglycerides to cholesterol in the VLDL was about 2:1 (Fig. 7) and was very similar to that of the male apoE2 transgenic rabbits. The HDL cholesterol in the ovariectomized female apoE2 transgenic rabbits was slightly decreased (Table V) and shifted from larger HDL 1 to smaller, more typical HDL (Fig. 7). Furthermore, the percentage of the apoE2 in the ␤-VLDL and IDL fractions increased from 21 to 41% 10 days after ovariectomy. This apoE2 distribution pattern was very similar to that seen in male transgenic rabbits (Table II). Both the plasma concentration and the lipoprotein distribution of the endogenous rabbit apoE were relatively unaffected by ovariectomy (data not shown). On the other hand, while plasma total cholesterol levels in the ovariectomized nontransgenic female rabbits did not change significantly, plasma triglyceride levels increased substantially 5 and 10 days after ovariectomy (Table V). Furthermore, the HDL cholesterol levels decreased from 41 Ϯ 9 to 27 Ϯ 8 mg/dl in the nontransgenic female rabbits 10 days after ovariectomy ( Table  V). The VLDL, IDL, and LDL in the nontransgenic animals increased significantly as demonstrated by Superose 6 chromatography (data not shown).
The effect of ovariectomy on postheparin plasma lipolytic activity was also determined in these animals. Both postheparin plasma LPL and hepatic lipase activities decreased by about 35% 10 days after the ovariectomy (Table V). These levels were similar to those of apoE2 transgenic male rabbits before estrogen treatment (Table IV). A similar change in postheparin plasma lipase activities was also observed in the nontransgenic female rabbits (Table V). These findings complement the results in the estrogen-treated male rabbits and indicate that estrogen status is at least partly responsible for the gender differences in the apoE2 transgenic rabbits.
Characterization of the Atherosclerosis-An atherosclerosis susceptibility study was carried out on both male and female apoE2 transgenics. The rabbits were maintained on a normal rabbit chow diet after weaning and were sacrificed at 11 months. Plasma lipid levels were measured at 4, 5, 6, 8, 10, and 11 months. The mean lipid levels of the six determinations for each rabbit are shown in Table VI. Although the lipid levels varied widely among individual rabbits, the total cholesterol and triglyceride levels were higher in four of the male transgenic rabbits than in any of the females. In contrast, HDL cholesterol levels were higher in females than in males, which resulted in a much higher ratio of total cholesterol to HDL cholesterol in males than in females.
Sudan IV staining of aortas from the transgenic rabbits revealed obvious lesions, especially in the aortic arch and the upper part of the abdominal aorta. Although male transgenics had more extensive lesions than females, the distribution of the lesions was very similar in both sexes (Fig. 8A). Quantitative analysis of lesion areas demonstrated that the nontransgenic rabbits essentially had no stained lesions, whereas the male transgenics had significant involvement of the aortic arch (24%) and the upper part of the abdominal aorta (10%) and the female transgenics had less involvement (10% aortic arch, 5% upper abdominal aorta) (Fig. 8B). Since the plasma lipid levels differed between males and females (Table VI), the gender difference in the extent of the lesions may reflect the lower total cholesterol and higher HDL cholesterol levels in females.  Larger numbers of animals will need to be examined to correlate the extent of atherosclerosis with specific lipid parameters.
Examination of sections of male transgenic aortas revealed both mature fatty streaks and some more advanced lesions, but there were no complicated lesions. Generally, layers of foam cells were deposited in the aortic intima, and considerable amounts of collagen were distributed around the foam cells (data not shown). Sections of female transgenic aorta showed mostly fatty streaks (data not shown). Thus, the apoE2 transgenic rabbits develop spontaneous atherosclerosis, with males having in general more advanced lesions (and more lesion area) than females. DISCUSSION We have generated transgenic rabbits that express high levels (30 -70 mg/dl) of human apoE2 (i.e. apoE2 (Cys-158)). These rabbits present a lipid and lipoprotein phenotype resembling that of human type III HLP, including its biochemical hallmark-accumulation of ␤-VLDL in plasma. The severity of the hyperlipidemia in the transgenic rabbits, especially the hypercholesterolemia, appears to depend on the expression levels of apoE2, as in apoE2 transgenic mice (13). High levels of apoE2 overexpression in the rabbits would be expected to create a situation biologically analogous (but not identical) to apoE2/2 homozygosity (i.e. a preponderance of apoE2 that effectively excludes or overwhelms the biological activity of the endogenous apoE). Supporting this presumption is the demonstration that the apoE2-containing ␤-VLDL are removed from circulation at a slower rate than normal VLDL and are a poor substrate for lipolysis. The lower rate of lipolysis appears to be specific for LPL. Since hepatic lipase is thought to be involved in the final conversion of IDL to LDL in the lipolytic cascade, our results suggest that the impediment occurs earlier in the cascade. Therefore, if the lipolytic defect caused by the presence of apoE2 contributes to the lower LDL levels seen in human type III HLP, LPL may be primarily involved. Previously, we demonstrated in vitro that apoE2 impairs the lipolytic conversion of VLDL to LDL (44).
Although spontaneous hyperlipidemia has also been found in transgenic mice expressing apoE(Arg-112, Cys-142) (9), apoE-Leiden (10), or apoE2(Cys-158) (13,14), the rabbit model of apoE2 overexpression more closely resembles human type III HLP. For example, the ␤-VLDL from apoE2 transgenic rabbits  and humans have a much higher cholesterol:triglyceride ratio and cholesteryl ester content than ␤-VLDL from transgenic mice (9,10,13). In human type III HLP, the VLDL cholesterol: plasma triglyceride ratio is typically Ͼ0.3 (1). This diagnostic ratio of Ͼ0.3 was achieved in the apoE2 transgenic rabbits but not in the previously reported transgenic mouse models (9,10,13). This difference may be explained by the presence of cholesteryl ester transfer protein activity in rabbits (46) and its absence in mice (47). In rabbits, as in humans, cholesteryl ester transfer protein facilitates the transfer of cholesteryl esters from HDL to VLDL and LDL (48) and may be one of the factors that determine the cholesteryl ester content of ␤-VLDL.
Noteworthy in this study is the significant gender difference in the expression of type III HLP in apoE2 transgenic rabbits, with females having less severe hyperlipidemia. Estrogen treatment markedly reduced the hyperlipidemia in male apoE2 transgenic rabbits and converted the lipoprotein profile toward a normal pattern. Conversely, ovariectomy increased the hyperlipidemia in female apoE2 transgenic rabbits and induced a more overt type III HLP phenotype. Therefore, these gender differences are at least partly due to the estrogen status of these animals and clearly indicate that estrogen modulates the expression of the hyperlipidemia in the type III HLP transgenic rabbits. These results help to explain why women who are homozygous for apoE2 usually develop type III HLP only after menopause, whereas men can develop the overt type III HLP profile at a younger age (1).
There are several possible mechanisms for the estrogen effect in the transgenic rabbits. First, estrogen may stimulate both lipase and LDL receptor activities, most likely by affecting transcription of the relevant genes, although this remains to be proven. The involvement of estrogen in stimulating these activities has been demonstrated (36,49,50), and our studies show the dramatic effect it has in modulating type III HLP. Second, estrogen's ability to lower plasma lipid levels (and account for the lesser severity of hyperlipidemia in females) may be due principally to its ability to overcome two blockades to catabolism caused by the presence of apoE2. Consistent with the higher lipolytic activity in both nontransgenic and transgenic females are the lower triglyceride content of female transgenic rabbit ␤-VLDL and the surface protrusions on their ␤-VLDL particles, presumably indicating an excess surface resulting from rapid lipolysis of ␤-VLDL. Furthermore, the higher HDL levels in the female transgenic rabbits may be due to a stimulated apoAI production rate by estrogen (51) and could serve as a sink to "attract and sequester" the apoE2 from the remnant lipoproteins, leading to a decreased distribution of apoE2 in the ␤-VLDL and a further decrease in the already impaired clearance of remnant lipoproteins. In our studies, estrogen treatment of male apoE2 transgenic rabbits shifted the distribution of apoE2 from ␤-VLDL to HDL fractions and increased HDL, supporting this possible mechanism. Conversely, ovariectomy of female transgenic rabbits redistributed apoE2 from HDL to ␤-VLDL. The redistribution of apoE2 and the changes in HDL cholesterol were attenuated in the ovariectomized females compared with the estrogen-treated males (implying that the changes may occur at a slower rate). While this may be due primarily to the pharmacologic doses of estrogen given to the male rabbits, it is possible that the higher amount of apoE2 on female transgenic HDL may be interfering with a processing event leading to plasma HDL changes. Based on our studies so far, it seems that the apoE2 transgenic rabbits provide a good animal model for the study of other secondary factors and mechanisms that modulate the expression of type III HLP.
The status of LDL receptors in type III HLP is a matter of some interest. One hypothesis is that these receptors are upregulated in response to a "perceived" deficiency in cholesterol FIG. 8. Atherosclerotic lesions in apoE2 transgenic rabbits. Both nontransgenic and apoE2 transgenic rabbit aortas were stained with Sudan IV to visualize the lesion area, as described under "Experimental Procedures." A, pinned-out aortas from a male and a female apoE2 transgenic rabbit. B, quantitative analysis of lesion area in rabbit aortas. Values (mean Ϯ S.E.) are presented as percentage of aortic surface covered by lipid staining. The aortic arch is defined as ending at the level of the ductus scar. nonTg, nontransgenic (n ϭ 8); male Tg, male transgenic (n ϭ 7); female Tg, female transgenic (n ϭ 7). Differences were evaluated by analysis of variance. p Ͻ 0.05, transgenic rabbits (both genders) versus nontransgenic rabbits (both genders) for both aortic arch and abdominal aorta; p Ͼ 0.05, male versus female transgenic rabbits for both aortic arch and abdominal aorta. delivery because of the defective apoE2 ligand on circulating lipoproteins, resulting in low plasma LDL levels, which do occur in type III HLP (1). Another possibility is that the LDL receptors are down-regulated because they become saturated with the overwhelming levels of circulating lipoproteins (␤-VLDL) in type III HLP. Our results on the plasma clearance of antibody 9D9 (Fig. 6) suggest that the latter might be the case. The 9D9 clearance data showed a trend (but did not reach statistical significance) toward expression of fewer LDL receptors (certainly not more) in the type III HLP rabbits than in normal rabbits. Down-regulation of LDL receptors associated with apoE2 overexpression could lead to an increase in LDL levels; however, LDL levels were not increased. Therefore, we suspect that the lipolytic defect associated with apoE2, rather than LDL receptor status, is the primary contributor to low LDL levels in type III HLP. More data of this type may help in reaching a final conclusion. Rabbits have been used widely for experimental atherosclerosis studies, generally by feeding with a high-fat, high-cholesterol diet (25,26). In addition, a model of spontaneous atherosclerosis has also been established in Watanabe heritable hyperlipidemic (WHHL) rabbits, a model of familial hypercholesterolemia due to LDL receptor deficiency (52,53). In WHHL rabbits, as in rabbits fed a high cholesterol diet, atherosclerotic lesions mainly involve the aortic arch and thoracic aorta, with less involvement of the abdominal aorta (54,55). In apoE2 transgenic rabbits, however, the atherosclerotic lesions mainly involved the aortic arch and abdominal aorta, with much less involvement of the thoracic aorta. Also, there may be differences in the qualitative nature of the lesions, since the lesions in this study were not particularly complex. Whether qualitative and/or distribution differences in lesions between WHHL and apoE2 transgenic rabbits reflect a significant difference related to the type of hyperlipidemia needs to be further evaluated. Nevertheless, the development of spontaneous atherosclerosis in transgenic rabbits expressing high plasma levels of apoE2 makes them a valuable animal model. The rabbits breed normally and require only hemizygosity for human apoE2; thus, they represent an attractive alternative model to WHHL and cholesterol-fed rabbits. Furthermore, these studies have again demonstrated the atherogenic potential of ␤-VLDL (for review, see Refs. 1, 56, and 57). Because these rabbits develop hyperlipidemia, as well as spontaneous atherosclerotic lesions, both detrimental and protective genetic and environmental factors can be investigated. Since lipid metabolism in rabbits resembles that in humans rather closely, and seemingly more closely in some regards than that in transgenic mice, this rabbit model should be very useful in future studies to investigate other factors that modulate the expression of type III HLP and that influence the development of atherosclerosis.