sate for lipoprotein remnant clearance when the LDL receptor is absent. A likely other candidate for lipo-
protein remnant clearance is the LRP. LRP is a multi- functional receptor, that binds a broad range of
ligands, including apoE-enriched lipoproteins, lipo- protein lipase, and a
2
-macroglobulin-proteinase com- plexes [3,4]. A 39 k-Da receptor associated protein
RAP was also identified that binds with high affinity to the multiple binding sites of LRP, thereby blocking
the receptor-ligand interaction [16]. Recently, Willnow et al. demonstrated that transient inactivation of LRP
in vivo in wild-type mice by injection of an adenoviral vector containing RAP did not induce an increase in
lipoprotein remnants, whereas overexpression of RAP in the liver of LDL receptor-deficient mice resulted in a
massive accumulation of VLDL and LDL cholesterol [17]. Furthermore, liver-specific inactivation of LRP in
LDL receptor-deficient mice by conditional gene target- ing induced accumulation of mainly LDL cholesterol
and to a lesser extent VLDL cholesterol, whereas in wild-type animals this caused a compensatory up regu-
lation of the LDL receptor in the liver [18]. Herz et al. further elucidated that the initial hepatic removal of
remnant lipoproteins is largely independent of either the LDL receptor or LRP [19]. Endocytosis of surface
bound remnants in wild-type mice is predominantly mediated by the LDL receptor and in LDL receptor-
deficient animals by slow endocytosis via LRP. The current model for liver uptake of lipoproteins involves
two successive stages, including sequestration of lipo- protein remnants by proteoglycans on parenchymal
liver cells and subsequent internalization by either the LDL receptor or LRP. For uptake via LRP further
enrichment with exogenous apoE during the phase of sequestration is suggested to be required [20,21].
Although the liver is the major source of apoE synthesis, apoE is also produced by a wide variety of
other cell types, including macrophages [22 – 24]. Re- cently, by transplantation of wild-type bone marrow
into apoE-deficient mice we [25] and others [26 – 28] found that macrophage-derived apoE can reduce hyper-
cholesterolemia in apoE-deficient mice due to increased recognition and uptake of lipoprotein remnants by
parenchymal liver cells, leading to a decreased suscepti- bility to atherosclerosis.
The aim of the present study was to investigate the role of the hepatic LDL receptor in macrophage apoE
induced reduction in serum cholesterol levels and atherosclerosis. To address this question, mice solely
deficient for apoE and mice deficient for both apoE and LDL receptors are reconstituted with apoE positive
bone
marrow. Our
results demonstrate
that macrophage-derived apoE is unable to normalize hy-
percholesterolemia and
decrease susceptibility
to atherosclerosis in apoE-deficient mice, lacking the LDL
receptor, despite high levels of apoE in the circulation. Therefore, it is concluded that for efficient clearance of
remnant lipoproteins that are enriched with extrahep- atic apoE, a functional hepatic LDL receptor is
essential.
2. Materials and methods
2
.
1
. Animals ApoE knock-out apoE − − .LDLr + + , LDL
receptor knock-out apoE + + .LDLr − − , apo- E.LDL
receptor double
knock-out apoE − −
.LDLr − − , and wild-type C57Bl6NCrlBR apoE + +
.LDLr + + mice
were used.
ApoE − − . LDLr + + mice were created as previously described
[11]. ApoE + + .LDLr − − mice were obtained from the Jacksons Laboratory Bar Harbour, ME, USA as
mating pairs and were bred in the Gaubius Laboratory, Leiden,
The Netherlands.
ApoE − − .LDLr − − mice were obtained by cross-breeding of apoE − −
.LDLr + + and apoE + + .LDLr − − mice. All knock-out mice were hybrids between C57Bl6 and 129
Sv strains at least three backcrosses to C57Bl6. C57Bl6NCrlBR apoE + + .LDLr + + mice were
purchased from Broekman Institute B.V., Someren, The Netherlands.
All mice used for bone marrow transplantation ex- periments were housed in sterilized filter-top cages and
given free access to food and water. Animals were maintained on sterilized regular chow SRM-A, con-
taining 5.7
w w
fat Hope Farms, Woerden, The Netherlands and drinking water was supplied with
antibiotics 83
mgl ciprofloxacin
and 67
mgl polymyxin B sulfate and 6.5 gl sugar.
Animal procedures were performed at the Sylvius Laboratories of the LeidenAmsterdam Center for
Drug Research in accordance with the national laws. All experimental protocols were approved by the Ethics
Committee for
Animal Experiments
of Leiden
University.
2
.
2
. Irradiation and bone marrow transplantation Female ApoE − − .LDLr + + and ApoE − −
.LDLr − − mice were used as recipients for the bone marrow transplantation studies. To induce bone mar-
row aplasia, the recipient mice age 5 – 6 weeks were exposed to a single dose of 13 Gy 0.28 Gymin, 200
kV, 4 mA total body irradiation, using an Andrex Smart 225 Ro¨ntgen source Andrex Radiation Products
AS, Copenhagen, Denmark with a 4-mm aluminium filter, 1 day before the transplantation as previously
described [25,29]. Donor bone marrow cell suspensions were isolated by flushing the femurs and tibias with
phosphate-buffered saline. Single-cell suspensions were
prepared by passing the cells through a 30-mm nylon gauze. Irradiated recipients received 1.0 × 10
7
bone marrow cells by intravenous injection into the tail
vein.
2
.
3
. Serum cholesterol and triglyceride analysis After an overnight fasting-period, approximately 100
ml blood was drawn from each individual mouse by tail bleeding. The concentrations of total cholesterol in
serum were determined using enzymatic procedures Boehringer Mannheim, Germany. Precipath stan-
dardized serum; Boehringer Mannheim, Germany was used as an internal standard.
The distribution of cholesterol over the different lipoproteins in serum was determined by loading of 30
ml serum of each mouse onto a Superose six column 3.2 × 30 mm, Smart-system, Pharmacia, Uppsala, Swe-
den. Serum was fractionated at a constant flow rate of 50 mlmin, using phosphate-buffered saline. Total
cholesterol content in the effluent was determined enzymatically.
2
.
4
. Quantification of apoE ApoE was measured using a sandwich ELISA spe-
cific for mouse apoE, as described earlier [25]. Briefly, for determination of apoE, a rabbit-anti-mouse apoE
polyclonal antibody SB Rabbit 67-AH, SmithKline Beecham, Harlow, UK was used as primary antibody,
biotinylated rabbit-anti-mouse apoE polyclonal anti- body was used as secondary antibody SB Rabbit 67-
AH-biotin, SmithKline Beecham, Harlow, UK, and finally biotinylated HRP conjugated streptavidin was
applied. HRP was detected by incubation with 3,3,5,5- tetramethylbenzidin TMB; Pierce, USA for 30 min at
room temperature. The reaction was stopped with 2 moll H
2
SO
4
and the absorbance was read at 450 nm. Pooled serum from C57Bl6 mice, with known apoE
level, was used as standard.
2
.
5
. b
VLDL isolation and characterization bVLDL was isolated from pooled serum of three
mice from each transplantation group at 4 weeks after BMT by discontinuous KBr gradient ultracentrifuga-
tion at 250 000 × g for 18 h, as described by Redgrave et al. [30]. The fraction of d B 1.006 gml was isolated,
dialyzed against PBS1 mmoll EDTA, and subse- quently characterized with respect to the free choles-
terol, cholesteryl ester, phospholipid, and triglyceride content,
using enzymatic
procedures Boehringer
Mannheim, Germany. The protein content was deter- mined according to Lowry et al. [31].
2
.
6
. Histological analysis of hearts and aortas for atherosclerosis
To analyze the development of atherosclerosis, mice were sacrificed at 4 months after BMT. Hearts and
aortas were perfused in situ with oxygenated Krebs buffer 37°C, 100 mm Hg for 20 – 30 min via a cannula
in the left ventricle, followed by a post-perfusion fixa- tion with 3.7 neutral-buffered formalin Formal-fixx,
Shandon Scientific, England and subsequent storage in formalin. To evaluate the development of atheroscle-
rotic lesions, the aortas were separated from the hearts. Hearts were bisected at the level of the atria and the
base of the heart plus aortic root were taken for analysis. Cryostat 10 mm cross sections of the aortic
root were made and stained with oil red O BDH, England. The atherosclerotic lesion area in the sections
was quantified using a light microscope connected with a 24-bits full color video camera and Optimas 6.1 image
analysis software BioScan, Edmonds, WA. Mean le- sion area was calculated in mm
2
from 10 sections, starting at the appearance of the tricuspid valves as
described previously [32].
2
.
7
. Statistical analysis Statistically significant differences among the means
of the different populations were tested by ANOVA. To compare pairs of groups, the Student-Newman-
Keuls multiple comparison test was performed after ANOVA.
3. Results