Hepatic Lipase Deficiency Increases Plasma Cholesterol but Reduces Susceptibility to Atherosclerosis in Apolipoprotein E-deficient Mice*

, The effect of hepatic lipase (HL) deficiency on the susceptibility to atherosclerosis was tested using mice with combined deficiencies in HL and apoE. Mice lack- ing both HL and apoE ( hhee ) have a plasma total cholesterol of 917 (cid:54) 252 mg/dl ( n (cid:53) 24), which is 184% that of mice lacking only apoE ( HHee ; 497 (cid:54) 161 mg/dl, n (cid:53) 20, p < 0.001). The increase in cholesterol was mainly in (cid:98) -migrating very low density lipoproteins, although high density lipoprotein cholesterol (HDLc) was also increased (53 (cid:54) 37 versus 20 (cid:54) 13 mg/dl, p < 0.01). Despite the increase in plasma cholesterol, we found that HL deficiency significantly decreased aortic plaque sizes in female mice fed normal chow (31 (cid:51) 10 3 (cid:54) 22 (cid:51) 10 3 (cid:109) m 2 in hhee versus 115 (cid:51) 10 3 (cid:54) 69 (cid:51) 10 3 (cid:109) m 2 in HHee , p < 0.001). Reduction of plaque sizes was also observed in female heterozygous apoE-deficient mice fed an atherogenic diet (2 (cid:51) 10 3 (cid:54) 2.5 (cid:51) 10 3

In circulation, nascent lipoproteins are remodeled by removal of core lipids and transfer of surface proteins and lipids prior to their removal through receptor-mediated mechanisms (1). Hepatic lipase (HL) 1 is a 60-kDa lipolytic enzyme involved in the processing of chylomicrons, intermediate density lipopro-teins (IDL), and high density lipoproteins (HDL) (2). HL efficiently hydrolyzes both triglycerides (TG) and phospholipids, while lipoprotein lipase is mainly responsible for hydrolysis of plasma TG. The preferred enzymatic substrates for HL are intermediate-size particles, such as IDL, and HDL 2 . Changes observed in lipoprotein profiles in humans with congenital HL deficiencies are the presence of ␤-migrating very low density lipoproteins (␤-VLDL) and larger HDL (3). The ␤-VLDL that accumulate in HL deficiency are more TG-rich than ␤-VLDL in typical type III hyperlipoproteinemia subjects (4). We previously reported that HL-deficient mice have a mild dyslipidemia with increased cholesterol and phospholipid, and the plasma contains large HDL floating in the 1.02-1.04 g/ml density range (5). Consistent with these observations, overexpression of HL decreases HDL cholesterol and HDL particle size in mice (6) and decreases HDL cholesterol and IDL in rabbits (7).
Both HL and cholesteryl ester transfer protein (CETP) participate in the metabolism of HDL (8). Both enzymes decrease HDL core lipids, apparently making lipid-poor apoA-I available for dissociation or transfer to other particles (8 -10). Addition of HL to native plasma or during perfusion through isolated liver from rat produces small pre-␤-migrating HDL (11). HL-mediated lipolysis may be the primary source of the circulating HDL subfractions, pre-␤-HDL, with a pre-␤ electrophoretic mobility on agarose gel. Pre-␤ 1 -HDL particles are thought to be the acceptors of cellular cholesterol in the early steps of reverse cholesterol transport (12,13). Disruption of this pathway could be detrimental to cholesterol homeostasis in peripheral tissues.
Mice provide a useful model for studying both HDL metabolism and the role of HDL in atherogenesis. Because mice lack CETP (14), one pathway for reverse cholesterol transport, e.g. transfer of cholesteryl ester from HDL to apoB-containing particles, is not operating. Also, specific components involved in lipoprotein remodeling, such as HL, can be genetically eliminated in mice by gene targeting. This allows specific mutations affecting lipoprotein metabolism to be related to atherogenesis. Certain genetically engineered mice have atherosclerotic plaques similar to those found in humans. For example, mice completely lacking apoE develop atherosclerosis on a regular chow diet (15)(16)(17)(18), while mice carrying a single copy of the apoE gene are susceptible to diet-induced atherosclerosis (19).
In this study, we evaluated the effect of HL deficiency on plasma lipoprotein distribution and atherogenesis in mice with apoE deficiency. We found that HL deficiency increases plasma cholesterol levels in female apoE-deficient mice, but their aortic plaque sizes were reduced compared with mice deficient in apoE, but normal in HL.
To test the effects of HL on lipoproteins and atherosclerosis, three experiments were designed. In the first experiment, female 129/B6 F2 mice deficient in HL (hhEE; n ϭ 10) and their wild-type littermates (HHEE; n ϭ 10) were fed the atherogenic diet for 4 months. In the second experiment, female mice heterozygous for apoE deficiency, which develop atherosclerosis only when fed an atherogenic diet (18), were tested in the presence (HHEe; n ϭ 7) or absence (hhEe; n ϭ 10) of a normal HL gene. The atherogenic diet contained 15.8% (w/w) fat, 1.25% (w/w) cholesterol, and 0.5% (w/w) sodium cholate ad was from Teklad Premier (Madison, WI). This diet was fed to the mice for 4 months, and the mice were killed at 6 months of age. In the third experiment, mice homozygous for apoE deficiency in the presence (HHee; n ϭ 14, 5 males and 9 females) or absence (hhee; n ϭ 18, 4 males and 14 females) of HL expression were maintained on normal chow and killed at 4 months of age.
Mice were maintained on 12-h dark/light cycles and allowed access to food and water ad libitum. The animals were handled following the National Institutes of Health guidelines for the care and use of experimental animals.
Lipids and Lipoprotein Analysis-Following an overnight fast, 200 -300 l of blood were collected by retro-orbital bleeding into tubes containing EDTA (final concentration of 2 mM). Plasma samples obtained by centrifugation (8000 ϫ g for 10 min at 4°C) were supplemented with aprotinin and gentamicin (final concentrations of 0.1 IU and 1 g/ml, respectively). Plasma was kept at 4°C until use. Total cholesterol and TG were determined immediately using reagents from Sigma. HDL cholesterol and phospholipids were determined within 3 days by using a kit from Wako Bioproducts (Richmond, VA).
For lipoprotein analysis, blood was collected in the morning from fasted mice under anesthesia with a lethal dose of Avertin (2,2,2tribromoethanol). Plasma samples (100 l) were subjected to gel filtration using a Superose 6HR column (Pharmacia, Uppsala), and lipid analysis of the fractions was carried out as described (20). Lipoproteins from 1 ml of plasma combined from three mice were fractionated by sequential density ultracentrifugation (20) at 70,000 rpm at 4°C in a Beckman TL-100 ultracentrifuge. Lipoprotein fractions in the density ranges of d Ͻ 1.006, d ϭ 1.006 -1.02, d ϭ 1.02-1.04, d ϭ 1.04 -1.06, d ϭ 1.06 -1.08, d ϭ 1.08 -1.10, and d ϭ 1.10 -1.12 g/ml were obtained following successive centrifugation for 5 h. The d ϭ 1.12-1.21 g/ml fraction was isolated following overnight centrifugation. Lipoproteins from each fraction were recovered by tube slicing and dialyzed in 10 mM Tris buffer, pH 7.4, 150 mM NaCl, and 1 mM EDTA. Lipoproteins were separated by agarose gel electrophoresis using commercially available precast gels (Ciba Corning, Palo Alto, CA). The apolipoprotein composition of each fraction was determined by SDS-polyacrylamide gel electrophoresis followed by Coomassie Brilliant Blue staining of the proteins.
HDL Analysis by Two-dimensional Electrophoresis-ApoA-I distribution in plasma lipoproteins was determined by two-dimensional nondenaturing gel electrophoresis (13). The first dimension was carried out on 0.75% agarose in 50 mM barbital buffer, pH 8.6, on GelBond (FMC Corp. BioProducts) at 4°C (200 V) for 100 min. The second dimension was carried out on 2-15% gradient polyacrylamide gels at 120 V for 16 h. Thereafter, the gel content was transferred to a nitrocellulose sheet in buffer containing 0.025 M Tris and 0.192 M glycine, pH 8.3, for 100 min at 392 mA under semidry conditions. ApoA-I-containing fractions were detected by chemiluminescence using an anti-mouse apoA-I antibody labeled with peroxidase and an ECL kit (Amersham Life Science, Inc.).
Cholesterol Efflux by HDL-Cellular cholesterol efflux from rat Fu5AH hepatoma cells was determined following the procedure described (21). Briefly, 2 ml of medium containing cells at a concentration of 25,000 cells/ml were plated in 2.4-cm multiwell plates and grown until confluence in minimal essential medium supplemented with 5% calf serum. Two days after plating, cellular cholesterol was labeled during a 48-h incubation with [ 3 H]cholesterol (1 Ci/well). To allow equilibration of the label, the cells were rinsed and incubated for 24 h in minimal essential medium containing 0.5% bovine serum albumin. For determination of cholesterol efflux, the cells were washed with 0.1 M phosphate-buffered saline and incubated at 37°C for 4 h with 100 g of HDL protein that had been isolated by fast protein liquid chromatography gel filtration. After recovery of the incubation medium, the remaining cellular cholesterol was then extracted. Radioactivity was measured in both medium and cells, and percentage of cholesterol efflux was calculated.
Evaluation of Atherosclerotic Lesions-The mice were killed with an overdose of Avertin. The heart and vascular tree were perfused with 4% phosphate-buffered paraformaldehyde, pH 7.4. Segments containing aortic sinus area were embedded, serially sectioned, and stained. Four sections were used for morphologic evaluations (16,19). The average of the sizes of lesions in the four sections was used to represent lesion size of each animal's aorta. Morphometric evaluation of the lesions was conducted using the Image measure/IP IM 2500 morphometry system (Phoenix Technology, Federal Way, MA).
Statistical Analysis-Paired and unpaired Student's t tests were performed for the plasma lipid concentrations. The Mann-Whitney U test was used to compare lesion sizes.

RESULTS
Plasma Lipid and Lipoproteins in Mutants-Fasted plasma lipid levels of mice with both HL and apoE deficiencies (hhee) were significantly elevated compared with mice with only apoE deficiency (HHee), as shown in Table I. Total cholesterol levels in hhee mice were nearly twice those in HHee controls. While a significant reduction in HDLc levels was observed in apoEdeficient animals (HHee) compared with wild-type mice (HHEE), HDLc levels in hhee mutants were similar to those in wild-type mice. HDLc levels in hhee mice showed a 2-fold increase over those in HHee mice. Triglycerides were also significantly higher in hhee compared with HHee mice. Male mice heterozygous for HL and homozygous for apoE (Hhee) had intermediate levels of plasma cholesterol (765 Ϯ 110 mg/dl). An ϳ75% increase in total cholesterol was observed in hhee compared with HHee mice. In contrast, only a 26% increase in total cholesterol was observed in hhEE compared with HHEE mice. Therefore, the increase in total cholesterol levels seen in HL deficiency is enhanced by the absence of apoE.
Female mice heterozygous for apoE with and without HL (HHEe and hhEe, respectively) responded to an atherogenic diet with a significant increase in total plasma cholesterol. Degrees of increase in HHEe and hhEe mice were similar, suggesting that apoE (but not HL) is the primary factor affecting the plasma lipid increase in response to dietary change. The atherogenic diet reduced TG by 65-77% and HDLc by 17-23% in HHEe and hhEe mice, respectively.
Plasma from different mutants was fractionated by gel filtration chromatography (Fig. 1). In contrast to HHEE and hhEE mice, in which most of the cholesterol was in the HDL range, the cholesterol in apoE-deficient mutants (HHee and hhee) was distributed mostly among larger lipoproteins in the VLDL and IDL/LDL range. The hhee mutants had further increased cholesterol content in these potentially atherogenic particles compared with HHee mice. In addition, the HDL fractions in hhee mice were increased, resulting in similar HDLc levels as observed in wild-type mice (HHEE). Consistent with previous results (5), HL deficiency increased the size of HDL particles as revealed in the earlier elution of HDL fractions from the gel permeation column (Fig. 1). However, the shift of the HDL fraction peak was not obvious in hhee compared with HHee mice. Fractionation of lipoproteins by sequential ultracentrifugation also showed that the distribution of cholesterol and phospholipids is determined mainly by apoE genotype (Fig. 2). Most of the cholesterol in apoE-deficient mice was within the density range of 1.006 -1.04 g/ml. In contrast, the cholesterol in HHEE and hhEE mice was present primarily within the density range of 1.06 -1.21 g/ml. The distribution of cholesterol in HL-deficient mice was shifted toward lower buoyant density compared with that in wild-type mice. In contrast, no significant difference among different groups of mice was observed in TG distribution, which was mostly in VLDL fractions (Fig. 2).
Agarose gel electrophoresis of lipoproteins in density fractions revealed a dramatic increase in the d Ͻ 1.02 g/ml fractions (␤-VLDL) in HHee mice, which was further increased in hhee mutants (Fig. 3). Consistent with gel permeation data, there was also a significant increase in the ␣-migrating particles in hhee mice, similar to the amount observed in HHEE mice. In hhee mice, lipoproteins with ␣-mobility were present in fractions with d Ͻ 1.06 g/ml. Similar particles were present at much lower levels in hhEE mutants, but they were absent in HHEE and HHee mutants. Previous studies showed that the d ϭ 1.02-1.04 g/ml fraction of hhEE mice contains ␣-migrating HDL that reacts with anti-apoE and anti-apoA-I antisera, but not anti-apoB antiserum (5). The absence of apoE is associated with increased apoB-48 and apoB-100 in the d Ͻ 1.006 g/ml fraction. SDS-polyacrylamide gel electrophoresis of lipoproteins showed that the amount of apoB in the various density fractions from hhee and HHee mice was not altered (data not shown), suggesting that the particle number had not significantly changed, but that the increase in size was mostly due to enrichment of cholesterol.
Lesion Development-To investigate the effect of HL deficiency on atherosclerosis development, three experiments were carried out. In the first experiment, hhEE mice and their wild-type littermates were fed an atherogenic diet for 4 months. No significant plaques were found in animals of either group, suggesting that hepatic lipase deficiency does not by itself increase susceptibility to diet-induced atherosclerosis in mice (data not shown).
In the second experiment, we evaluated diet-induced atherosclerosis using animals heterozygous for apoE deficiency. We fed the atherogenic diet to HHEe and hhEe females for 4 months. Animals were killed at 6 months of age, and aortic plaque sizes were evaluated (Fig. 4a). We found that the extent of lesions in hhEe mice was only 3% (2.0 ϫ 10 3 Ϯ 2.5 ϫ 10 3 m 2 , n ϭ 7) of that in HHEe mice (56 ϫ 10 3 Ϯ 49 ϫ 10 3 m 2 , n ϭ 10) despite the higher levels of plasma cholesterol in HHEe mice, shown in Table I. In the third experiment, we evaluated the effect of HL deficiency on spontaneous atherosclerosis caused by the complete lack of apoE. Mice homozygous for apoE deficiency in the presence or absence of HL were maintained on a regular chow diet and killed at 4 months of age. All apoE null mutants developed atherosclerotic lesions (Fig. 4b). Strikingly, the plaque sizes in hhee females were 25% (31 ϫ 10 3 Ϯ 21 ϫ 10 3 m 2 , n ϭ 14) of those in HHee control females (115 ϫ 10 3 Ϯ 69 ϫ 10 3 m 2 , n ϭ 9, p Ͻ 0.001). In contrast, no significant difference in plaque size was observed in the small number of males examined (20 ϫ 10 3 Ϯ 5.5 ϫ 10 3 m 2 (n ϭ 4) versus 21 ϫ 10 3 Ϯ 12 ϫ 10 3 m 2 (n ϭ 5)) despite higher plasma cholesterol levels as compared with HHee mice (Table I). Lesion sizes in male and female hhee mice were not significantly different. Maturity of plaques was directly correlated with their size, and no particular differences FIG. 1. Fast protein liquid chromatography gel filtration of mouse plasma. Equal volumes of plasma (n ϭ 5) from wild-type and mutant mice fed a regular chow diet were pooled, and 100 l were subjected to gel filtration chromatography using a Superose 6HR column. Fractions (0.5 ml) were collected, and cholesterol concentration was measured in each of the following fractions: a, the wild type (HHEE (q)) and hepatic lipase deficiency (hhEE (E)); and b, apoE deficiency (HHee (f)) and double mutants for HL and apoE deficiencies (hhee (Ⅺ)).
in plaque components between mice with and without HL were apparent by light microscopic evaluation.
Cholesterol Efflux and HDL Particles-To determine if HDL function is altered in HL-deficient mice, we assessed their capacity to promote efflux of cellular cholesterol (Fig. 5). HDL from both hhee and HHee mice had reduced capacity to promote cholesterol efflux (55 and 26%, respectively) compared with HDL from wild-type mice (100%). Both male and female HHee mice had equal capacity to promote cholesterol efflux, but HDL from male hhee mice had 50% higher capacity compared with HDL from female hhee mice, consistent with the fact that male hhee mice have a higher plasma HDLc level compared with female hhee mice (Table I).
We then examined whether the effect on cholesterol efflux was related to qualitative changes in HDL particles. Plasma samples from different mice were separated by two-dimensional nondenaturing gel electrophoresis and blotted with an apoA-I antibody (Fig. 6). The general electrophoretic pattern of HDL was similar in all genotypes and very close to that ob-served in humans (13). However, pre-␤ 1 -HDL from all mutant mice displayed increased levels and slightly larger sizes compared with wild-type mice. The ␣ 1 -HDL fraction was markedly reduced in HHee mutants and less in hhee mutants. A new fraction of ␣-mobility, ␣ 2 , was found in hhee mice and much less so in HHee mice. This fraction is larger than ␣ 1 in size. Electrophoretic patterns of male and female mutants were indistinguishable (data not shown). DISCUSSION The most striking change in the lipid profiles of mice deficient in both HL and apoE is a marked increase in cholesterol present mostly in ␤-VLDL. Although this increase primarily results from the lack of apoE, the comparison between hhee and HHee mice clearly demonstrates additional and independent effects of HL deficiency. The absence of HL therefore increases Mice were fed an atherogenic diet (high-fat diet (HF)) (a) or a regular chow diet (b). Lesion size for each mouse aorta is taken as the mean lesion size of four sections measured as described (16,19). Results are expressed as mean lesion size in each group Ϯ S.D. the accumulation of atherogenic ␤-VLDL already present in apoE deficiency. The increase in lipids is, however, not associated with an apparent increase in apoB-48 or apoB-100, suggesting that hhee mice accumulate larger apoB-containing particles. This is consistent with a role of HL in the remodeling of remnant particles.
In addition to lipid abnormalities in HL-deficient humans, several studies have suggested that HL plays a role in chylomicron remnant removal (22,23). Acute inhibition of HL activity by antibodies in primates or in a rat liver perfusion system causes the accumulation of TG-rich lipoproteins (22). Ji et al. (23) have shown in vitro that cell surface-bound HL enhances binding of remnant lipoproteins to the cells. This process does not require apoE since binding of ␤-VLDL isolated from apoEdeficient mice and binding of apoE-containing ␤-VLDL from rabbits were equally enhanced. HL has been shown to bind to the LDL receptor-related protein and promote ␤-VLDL uptake (24). These interactions are inhibited by heparin, suggesting that HL may act as a bridge between remnant lipoproteins and cell-surface heparan sulfate proteoglycans. However, HL-deficient mice display no apparent impairment in chylomicron clearance (5), suggesting that this pathway is not crucial for their clearance in mice under normal physiological conditions. In contrast, in apoE deficiency, when the major ligand mediating removal is not operating, the absence of HL causes a further increase in the accumulation of cholesterol-rich remnants. The increase in remnant lipoproteins in hhee mice may result from an impairment in HL-facilitated remnant uptake by the liver or delayed conversion of remnants to smaller particles, preventing their effective removal.
Another effect related to HL deficiency is an increase in HDL levels. Large ␣-migrating HDL enriched in apoE but devoid of apoB were observed on agarose gel electrophoresis of d ϭ 1.02-1.04 g/ml fractions in HL-deficient mice. These particles contain apoA-I, cholesterol, and phospholipids (5). In this study, we found larger HDL in HL-deficient mouse plasma fractionated by gel permeation chromatography or ultracentrifugation. In addition, two-dimensional nondenaturing gel electrophoresis showed the presence of large HDL particles in the ␣ 2 -fraction in the hhee mutants. We also observed a slight increase in apoA-I in the ␣ 1 -migrating fraction of hhee mutant HDL compared with HHee mutant HDL. In contrast to results reported by Koo et al. (25) suggesting that apoE is required for the formation of large HDL, our observations indicate that apoE is not essential for this process.
The origin of large ␣ 2 -migrating HDL in HL-deficient mice is not clear. The most simple explanation is that these particles are generated as a consequence of sequential lecithin:cholesterol acyltransferase-mediated enlargement of small HDL 3 first to HDL 2 and then to HDL 1 (1). HDL 1 are normally enriched in apoE and are thought to be removed through apoEmediated LDL receptor interaction (26). The absence of apoE in hhee mutants may contribute to the accumulation of these particles by delaying their removal by hepatic receptors. Further studies are necessary for determining the exact origin of these large ␣-migrating fractions.
It has been proposed that only a small portion of apoA-Icontaining particles are active in cholesterol removal from peripheral cells and transport back to liver, a process known as reverse cholesterol transport (27). These lipid-poor HDL have pre-␤ electrophoretic mobility, and HL and CETP are thought to be important in their formation (8). HL activity has been found to contribute to the formation of pre-␤-migrating HDL as observed in in vitro and liver perfusion experiments (8,11). However, the distributions of pre-␤-migrating HDL in the plasma of hhee, HHee, and hhEE mice are similar to that in wild-type mice. Because hhee mice lack CETP, HL, and apoE, other enzymes must be sufficient for the formation of pre-␤-HDL. One alternative mechanism to generate these particles involves phospholipid transfer protein, which is believed to promote fusion of HDL particles, liberating small lipid-poor particles analogous to pre-␤ 1 -HDL (28). The absence of HL may favor this process by increasing phospholipid transfer protein substrate. Another possible mechanism may involve the recently described HDL receptor or scavenger receptor type BI, which has been reported to be expressed in a variety of tissues including the liver (29). This receptor mediates selective uptake of cholesteryl esters without endocytosis, perhaps making apoA-I available to form smaller particles similar to FIG. 6. Two-dimensional gel electrophoresis of plasma lipoproteins. Plasma from individual mice were separated by agarose gel electrophoresis (first dimension) and then by nondenaturing polyacrylamide gradient gel electrophoresis on 4 -15% gradient gels (second dimension). HDL electrophoretic fractions were detected by Western blotting using an anti-mouse apoA-I antibody. ␣or ␤-particles are indicated according to their electrophoretic mobility. pre-␤ 1 -HDL.
HL deficiency increased ␤-VLDL cholesterol levels in apoEdeficient mice. This would be expected to have an atherogenic effect. However, the female hhee mutants developed smaller atherosclerotic lesions compared with the HHee mutants. No difference was observed in the small number of males examined. Reduction in plaque size was also observed in female hhEe mice compared with HHEe mice when fed an atherogenic diet. Since genetic background is clearly a factor that contributes to atherosclerosis, it is important to use experimental and control groups of mice with comparable genetic backgrounds. The mice used in this study carried ϳ87% of their genome from the C57BL/6J strain, with the rest being from 129/Ola. Offspring from several families were used to help ensure that the region from 129/Ola would be randomly distributed. The only systemic difference between the two groups (HHee and hhee, or HHEe and hhEe) is the region near the HL gene: this region in hh mice is from 129/Ola, while in HH mice, it is from C57BL/6J. We cannot therefore exclude a possibility that some loci linked to HL, but not HL deficiency itself, are the cause of the protection in hhee mice. However, since no major loci that affected atherosclerosis are known to be linked to the HL gene, this systemic difference is unlikely to have influenced our results.
An increase in ␤-VLDL particles but reduced atherosclerosis was recently observed in apoE-deficient mice overexpressing human apoA-IV (30). The reason for these findings is not clear, but they may be the result of a more favorable balance in extrahepatic cholesterol by either decreased influx or increased efflux. Although the size of ␤-VLDL was not directly measured, our data are consistent with an increase in ␤-VLDL size. Larger ␤-VLDL particles have a lesser ability to enter the artery wall, and their conversion by HL to smaller particles may increase their atherogenecity, as suggested by previous studies in rabbits (31,32). Furthermore, we observed an increase in HDL cholesterol in hhee mutants and a higher capacity of hhee mutant HDL to promote cholesterol efflux in vitro compared with HHee mutant HDL. However, protection against atherosclerosis in mice may not solely depend on the capacity to promote cholesterol efflux since male hhee mice had a higher efflux than female hhee mice, yet did not appear to display a decrease in atherosclerosis. Further studies including determination of the sizes of ␤-VLDL particles in males and females with HL deficiency are necessary to explain the sexrelated differences.
It is premature to conclude that the inhibition of HL may help reduce formation of atherosclerotic plaques in humans. Physiological differences between the two species, such as the absence of CETP activity in mice and apoB editing in the liver of mice, but not in humans, may influence differently the role of HL in atherogenesis. Consequently, HL deficiency in humans may be associated with premature atherosclerosis (3). In addition, while men are at higher risk for developing atherosclerosis than women, the gender effects appear to be opposite in mice. Previously, we observed that chow-fed female apoE-deficient mice (33) and female heterozygous apoE-deficient mice fed an atherogenic diet (34) develop larger plaques in their aortic sinus area compared with males. Similar observations were made by Paigen et al. (35) in the diet-induced atherosclerosis of inbred C57BL/6J mice. Further investigations in HLdeficient mice may help us to understand the cause of this gender difference.
In conclusion, our study clearly shows that HL is involved in the catabolism of ␤-VLDL/IDL and HDL by a mechanism independent of apoE. We found that neither HL nor CETP is essential for the formation of murine pre-␤ 1 -HDL and that HL deficiency is associated with decreased susceptibility to atherosclerosis in female mice despite higher plasma cholesterol levels. This beneficial effect was associated with increased cholesterol efflux by HDL from HL-deficient mice.