© 1997 American Heart Association, Inc.
Articles |
Effect of Probucol Treatment on Gene Expression of VCAM-1, MCP-1, and M-CSF in the Aortic Wall of LDL Receptor–Deficient Rabbits During Early Atherogenesis
From the Department of Medicine, University of California San Diego, La Jolla, Calif.
Correspondence to Joachim Fruebis, PhD, University of California San Diego, Department of Medicine, 0682, 9500 Gilman Dr, BSB 1080, La Jolla, CA 92093-0682. E-mail jfruebis{at}ucsd.edu
Abstract
Abstract Probucol is a potent inhibitor of atherosclerosis in animal models. However, the mechanism of its antiatherogenic effect is not known. To investigate the effects of probucol on gene expression of VCAM-1, MCP-1, and M-CSF in vivo during the early stages of atherogenesis, we determined gene expression in 12 control WHHL rabbits and 12 WHHL rabbits fed 1% probucol from age 3 weeks. Three animals from each group were killed at 6, 9, 12, and 18 weeks of age. Two intimal/medial segments of the thoracic aorta, each comprising the orifices of a pair of intercostal arteries, were analyzed by semiquantitative RT-PCR using GAPDH as an internal standard. A third segment located between these two segments was studied by immunocytochemistry. A basal level of VCAM-1 gene expression was observed in lesion-free aortas of both treated and untreated WHHL rabbits (and in normal NZW aortas). Immunocytochemistry showed some VCAM-1 protein in normal arteries and confirmed that VCAM-1 protein expression generally correlated with gene expression. In the untreated WHHL rabbits, a marked upregulation of VCAM-1 expression was observed at 18 weeks. To correlate gene expression with intimal monocyte/macrophages in each animal, the macrophage area was determined by morphometry of immunostained sections. In addition, a scoring system of lesions was used. VCAM-1 expression showed a highly significant correlation with the extent of intimal macrophage presence (P<.001). A lesser degree of correlation between gene expression and macrophage accumulation was also seen for MCP-1. In contrast, M-CSF expression remained constant over the entire study period and showed no correlation with the intimal macrophage accumulation. Probucol treatment completely prevented lesion formation in all animals up to 18 weeks of age. Probucol reduced the level of basal VCAM-1 expression and prevented its upregulation. MCP-1 expression was not affected by probucol treatment, whereas M-CSF expression was significantly lowered by probucol. Our results support the idea that VCAM-1 plays an important role in early atherogenesis and suggest that the antiatherogenic effect of probucol may in part be due to a downregulation of VCAM-1. Reduction of the basal level of M-CSF gene expression by probucol treatment may also contribute to its ability to inhibit atherogenesis.
Key Words: arteriosclerosis • oxidation • antioxidants • PCR • immunocytochemistry
Substantial evidence suggests that oxidative modification of LDL plays an important role in atherogenesis. In vitro studies have shown that OxLDL and its byproducts possess numerous properties that may enhance the atherogenic process (reviewed in References 1 through 41 2 3 4 ). However, the biological importance of individual mechanisms in vivo remains largely unknown. It has been hypothesized that OxLDL acts predominantly by enhancing the recruitment, intimal retention, and phenotypic transformation of monocytes5 6 and by its rapid uptake via a family of scavenger receptors on macrophages,7 8 9 10 which results in foam cell formation. If this hypothesis is correct, one would assume that a reduction of the amount of OxLDL formed or accumulated in the intima should result in a reduction of atherosclerosis. Indeed, the most striking evidence for the atherogenic role of lipoprotein oxidation was provided by the observation that several powerful lipophilic antioxidants, eg, probucol, BHT, and diphenylphenylenediamine, significantly reduced the progression of atherosclerosis in rabbits,11 12 13 14 primates,15 and mice.16 However, studies with less powerful natural antioxidants, such as vitamin E, provided inconsistent results,17 18 19 20 and a lipophilic analogue of probucol recently also failed to reduce atherogenesis,21 even though these compounds conveyed some degree of antioxidant protection to LDL. These results may indicate that only a very extensive antioxidant protection of LDL (beyond a certain threshold) is effective in preventing the formation of OxLDL in the artery wall.20 21 On the other hand, these results may also indicate that small amounts of OxLDL or its byproducts are sufficient to trigger potent atherogenic effects and that a much greater degree of antioxidant protection is required to prevent these effects. This would also suggest that the uptake of OxLDL by macrophages may not be the primary atherogenic mechanism of OxLDL.
In addition to promoting foam cell formation, OxLDL has other
potentially atherogenic properties, including its
cytotoxicity22 and the modulation of vascular
tone.23 24 Most importantly, increasing evidence suggests
that OxLDL may induce gene expression of adhesion molecules and
cytokines in arterial wall cells (reviewed in
References 4 and 254 25 ). For example, minimally modified LDL (oxidized to
an extent that does not lead to recognition by scavenger receptors) has
been shown to induce the expression of chemotactic factors like
MCP-126 and colony-stimulating factors like
M-CSF27 28 in endothelial cells and in
smooth muscle cells. Liao et al29 30 have recently
proposed mechanisms by which lipid peroxidation products, such as
those generated during the oxidative modification of LDL, may induce
expression of inflammatory mediator genes and activate
NF-
B–like transcription factors in vascular cells. Studies using
both immunocytochemistry and in situ hybridization have previously
shown that gene expression for MCP-1, M-CSF, and their
respective proteins occur in atherosclerotic lesions, predominantly in
areas rich in macrophages and OxLDL.31 32 33 However,
the biological relevance of MCP-1 and M-CSF for monocyte recruitment
and early atherogenesis has yet to be demonstrated in vivo.
In addition to providing protection to LDL extracellularly,
antioxidants may also affect the intracellular redox potential and thus
inhibit cellular effects triggered by OxLDL that may promote
atherogenesis.34 35 36 Finally, probucol and other
antioxidants could have additional cellular effects (that may or may
not be related to their antioxidant properties) that could account for
the inhibition of atherogenesis. Changes in the cellular composition of
lesions have been described in probucol-treated WHHL
rabbits37 and nonhuman primates,38 and a
limited Northern blot analysis has suggested that probucol
affects mRNA levels of a number of growth-regulatory
molecules.38 Probucol inhibits IL-1–induced monocyte
adhesion to endothelial cells, even though it remains
controversial whether probucol affects IL-1 gene expression
itself.39 40 The effect of probucol on VCAM-1 in vivo is
unknown, but OxLDL upregulates VCAM-1 expression by cultured
endothelial cells after preincubation with
TNF-
.35 Endothelial VCAM-1 expression
has been demonstrated in vivo in atherosclerotic
lesions.41 42 43 However, as with MCP-1 and M-CSF, it is
currently unknown whether VCAM-1 expression constitutes a rate-limiting
step in early atherogenesis.
The present study assesses the effect of probucol on the expression of VCAM-1, MCP-1, and M-CSF in rabbit aortas and correlates this information with the appearance of intimal macrophages and atherosclerosis.
Methods
Animals
Two groups of 12 LDL receptor–deficient (WHHL) rabbits, matched
for litter and sex, were fed a regular rabbit diet with or without 1%
(wt/wt) probucol. Probucol (a generous gift from Merrell Dow
Pharmaceuticals) was added to the chow in diethyl ether. The control
diet was similarly treated with plain solvent. The diet was started
immediately after weaning, ie, at 3 weeks of age, and continued until
18 weeks of age. Animals were given a fixed amount of food (100
g/d) and water ad libitum. Plasma cholesterol and
triglyceride levels were determined every 3 weeks, using an
automated enzymatic technique (Boehringer Mannheim
Diagnostics). Plasma probucol levels were determined using
high-performance liquid chromatography, as
described before.21 Three animals from each group were
killed at 6, 9, 12, and 18 weeks. The animals were heparinized and
given an overdose of sodium pentobarbital. The systemic circulation was
perfused with ice-cold PBS containing 0.3 mmol/L EDTA, and
the DTA and part of the lower ABD were dissected. The aorta was kept
submerged in a bath of ice-cold PBS, cleaned thoroughly of adventitia,
opened longitudinally, and pinned to the wax layer at the bottom of the
preparation tray. Six segments of the DTA were isolated, each
containing a pair of intercostal artery orifices in the middle of the
segment. These segments were numbered one through six, beginning at the
proximal end of the DTA. In addition, a single 15-mm-long segment of
the ABD free of visible atherosclerosis was prepared
distal to the branch point of the inferior mesenteric
artery. The abdominal segment was used as a control, because on the
basis of earlier studies, it can be expected to be free of even
microscopic lesions. However, comparable aortic segments from four NZW
rabbits fed a regular rabbit diet were included as additional
lesion-free controls.
Segments 1 (DTA1) and 5 (DTA5) of the DTA, as well as the abdominal segment, were used to determine gene expression by RT-PCR and were snap frozen in liquid nitrogen until further processing. Segment 3 (DTA3) was fixed with 4% paraformaldehyde, embedded in paraffin, and used for histological analysis. Segments 2, 4, and 6 were used for additional determinations of gene expression and histology.
Because immunocytochemistry with the only available antibody to rabbit VCAM-1 cannot be performed on paraformaldehyde-fixed paraffin-embedded sections, frozen sections of a broad range of atherosclerotic lesions and nonlesioned aortas were prepared from two additional WHHL rabbits, age 3 and 10 months, and two NZW control rabbits fed a nonatherogenic diet. These lesions were used primarily to provide qualitative data on the presence and localization of VCAM-1 protein in lesions of different stages, but gene expression was also determined in the corresponding thoracic segments to confirm that the levels of gene expression in these lesions were indeed comparable to those of the time-course experiment. The arteries of these additional rabbits were perfused with ice-cold PBS, dissected, and 5-mm-wide segments of the aorta were embedded in OTC and flash frozen by immersion in liquid nitrogen–cooled isopentane.
RNA Isolation, RT, and PCR
Quantitation of the PCR products was achieved by UV
densitometry of ethidium bromide–stained agarose gels. Total RNA was
isolated from frozen tissue within 24 hours after preparation, using
RNAzol B (Tel-Test). The tissue was homogenized in 1 mL of
RNAzol B using a polytron PT1200 at setting 4 three times for 20
seconds each, and RNA was prepared according to the manufacturer’s
protocol. The amount of RNA isolated was determined by measuring the
specific absorption at 260 nm. The integrity of the RNA isolated was
confirmed by agarose gel electrophoresis, run under denaturing
conditions.
Samples of 2.5 µg total RNA were reverse transcribed into cDNA in 50 µL reaction mixtures using 200 U of recombinant M-MLV reverse transcriptase (Superscript II, GIBCO-BRL) and oligo dT15 as primer. The reaction was carried out in the RT buffer supplied with the enzyme. The final concentration of dNTP (Boehringer Mannheim) was 0.15 mmol/L, and RNAse inhibitor (Boehringer Mannheim) was added at 30 U per reaction. RT was carried out for 60 minutes at 42°C followed by an inactivation step at 94°C for 10 minutes. RT products were stored at -70°C.
Table 1
shows the
oligonucleotides of the 5' and 3' primers chosen for
each target gene. Computer-assisted primer selection (Gene Runner,
Hastings Software) was conducted following general guidelines (G+C
content around 50%, no significant secondary structures, no stretches
of polypurines or polypyrimidines, no palindromic sequences). To avoid
amplification of genomic DNA, primers stretching over several exons
were chosen. In cases in which the rabbit sequence was unknown (VCAM-1,
M-CSF), multiple alignments of human and murine genes were performed,
and primers were selected from regions of high homology. Primers were
synthesized by the Molecular Biology Core Unit of the La Jolla SCOR.
For each of the tested genes, the PCR resulted in a single band of the
predicted size.
|
GAPDH expression was quantified by competitive PCR, as described below.
VCAM-1, MCP-1, and M-CSF expression were determined by semiquantitative
PCR, using GAPDH as internal standard. PCR conditions used were those
of the target gene (see Table 1
). Two different pairs of primers were
used to detect VCAM-1 expression. One pair, VCAM-1 "A," did not
include the alternative splicing region of VCAM-144 45 and
was used for the semiquantitative determination of VCAM-1 expression.
The other pair of primers, VCAM-1 "B," did include the
alternatively spliced region and therefore could result in two
different PCR products. VCAM-1 "B" was used to
analyze the contribution of the two alternately spliced gene
products to the overall expression of VCAM-1 in the artery
wall.
PCR was performed on an Ericomp Twinblock Easycycler and was carried
out in a total volume of 25 µL containing 2.5 µL cDNA template,
0.24 µmol/L primer, 100 µmol/L dNTP, 2
mmol/L MgCl2, 100 µg/mL BSA, and 0.75 U
Taq polymerase (Promega) in the PCR buffer provided with the
enzyme. A control reaction without addition of cDNA was included in
each PCR run to test for possible contamination. All analyzed
reactions proved to be free of contamination. To obtain maximum
fidelity, hot-start conditions were applied, using ampliwax beads
(Perkin-Elmer). The PCR profile used started with a 5' denaturing phase
at 88°C, followed by a 5' initial annealing phase (Table 1). The
subsequent cycles consisted of a 1-minute elongation phase at 72°C, a
45-second denaturing phase at 94°C, and a 1-minute annealing phase.
The PCR reaction was concluded by a 10-minute elongation phase, again
at 72°C. In each PCR reaction, a 465-bp PCR product of GAPDH mRNA
was coamplified and used as internal standard. GAPDH was generally
found to be expressed at higher levels than the target genes. To ensure
that GAPDH amplification would not reach the plateau phase earlier than
the target gene, addition of GAPDH primers was delayed, as described by
Kinoshita et al.46 During the first five PCR cycles, only
the target gene primers were present. After completion of the fifth
elongation phase, the PCR reaction was halted and the reaction mixture
was spun down and cooled to 4°C. GAPDH primers were added and the
reaction was continued, starting with a 5' denaturing phase at 88°C.
Immediately after completion of the reaction, the samples were mixed
with EDTA-containing sample buffer and stored at 4°C until
analyzed by gel electrophoresis.
Similar PCR conditions were used for the quantitative determination of
GAPDH expression by competitive PCR, which was performed to verify that
GAPDH expression in tissues was constant and not affected by probucol
(a requirement for using GAPDH as internal standard for the
semiquantitative determination of VCAM-1, MCP-1, and M-CSF). In
addition to the 2.5 µL of RT product, an equal volume containing
a competitor cDNA was added. The competitor cDNA (a generous gift from
Dr H. Lukhaup, University of Heidelberg) was identical to the GAPDH PCR
product, except for a 102-bp deletion. Five increasing
concentrations of competitor (between 5 and 40 pg/µL) were
chosen after an initial wide-range analysis. The amount of PCR
product formed was determined relative to the expression of GAPDH
after separation by agarose gel electrophoresis (3:1 Nusieve, FMC). To
validate that quantitation of gene expression under the PCR conditions
reported in Table 1 was performed during the exponential phase of the
amplification reaction for both the reference and the target gene, we
determined that the amplification of both PCR products proceeded in
parallel. Indeed, the same ratio of expression was obtained when
aliquots were taken after an increasing number of PCR cycles (at least
five cycles beyond the number of cycles used for the subsequent
determinations; data not shown).
Ethidium bromide–stained gels were transilluminated at 302 nm. Electronic images were captured by using a solid state black-and-white video camera (Cohu Electronic), and the intensity of the bands was determined using Optimas 4.0 imaging software (Bioscan). Each image analysis was performed twice by the same observer. Gene expression was determined in triplicate.
To demonstrate that a strong linear correlation exists between the
amount of cDNA applied to the gel and the measured fluorescence
signal, increasing amounts of the 363-bp GAPDH cDNA (5 to 50 ng per
lane) were subjected to gel electrophoresis on a 3% agarose gel and
analyzed by UV densitometry. As shown in Fig 1, this process yielded a regression line
with an R2 of .994 (P<.001).
|
To minimize errors, for each gene analyzed, the complete set of samples was processed in parallel. Aliquots of the same PCR master mix were used for the reactions, thus assuring identical conditions for each PCR reaction. Each determination was done in triplicate. Finally, as an additional control, the complete analysis of VCAM-1 and MCP-1 gene expression in untreated animals (starting with the RT reaction) was repeated at the end of the study. This procedure yielded results identical to the preceding determinations.
Histological Techniques and Determination of
Intimal Macrophage Accumulation and Lesion Size
Segment 3 of the DTA was used for immunocytochemical
analysis. Since PCR analysis and
histological analysis cannot be done in the
same specimen, adjacent aortic segments (each containing one pair of
intercostal branch sites) were selected. The rationale for this
approach was based on the fact that in the rabbit model used, both the
size and stage of atherosclerotic lesions distal to adjacent
intercostal artery orifices were very similar and progressed at a
comparable pace. Furthermore, the two segments chosen for PCR
analysis (segments 1 and 5) were proximal and distal,
respectively, to the segment used for histology (segment 3). Thus, any
difference in atherosclerosis that might result from
the general progression of atherosclerosis in rabbits
(from the arch toward the ABD) would be compensated for by averaging
the results of the two PCR determinations.
To ensure a complete and reproducible assessment of lesions, the following strategy was used. After fixation in paraformaldehyde and paraffin embedding, the entire third segment of the DTA was serially sectioned into 7-µm-thick sections. Two sections each were transferred onto numbered microscope slides. The first section showing the orifice of an intercostal artery was located microscopically and the respective microscope slide labeled "0." The microscope slides preceding slide "0" were assigned positive numbers, starting with "+1," whereas slides following section 0 were given negative numbers starting with "-1." Slides + and -15, 20, 25, 30, 35, 40, and 45 (each containing two sections) were then used for immunocytochemistry of macrophage/foam cells. Thus, 630 µm proximal and 630 µm distal to the orifice of the first intercostal orifice were assessed.
The sections were immunostained with RAM-11, a monoclonal antibody specific for rabbit macrophages (a generous gift from Dr A.M. Gown, University of Washington),47 using an avidin/biotin/alkaline phosphatase system (Vector Labs), as previously described.48 Controls were performed with nonspecific antibodies or by omitting the primary antibody and were devoid of specific staining. Methyl green was used to counterstain nuclei. All 14 selected microscope slides were analyzed for the extent of atherosclerosis and the presence of macrophage/foam cells by the same investigator.
Two different measures of the intimal macrophage presence were obtained. The first of these consisted of the classification of lesions into four stages, on the basis of their size and relative macrophage content. Stages were defined as follows: stage 0, normal intima without any subendothelial monocyte/macrophages; stage 1, isolated monocyte/macrophages in the subendothelial space; stage 2, single monolayer of monocyte/macrophages underneath the endothelium; stage 3, presence of multiple layers of macrophages, as well as transitional and more advanced lesions. The score given to a specific section represented the most advanced lesion found in the entire section. The regions corresponding to the orifice of the left and right intercostal arteries were evaluated separately to reduce the influence of isolated larger lesions on the overall macrophage score. Thus, 28 scores were obtained for each aortic segment, including 7 evaluations of the right part of the "proximal" sections (positive slide numbers), 7 of the right part of the "distal" sections (negative slide numbers), 7 of the left part of the proximal sections, and 7 of the left part of the distal sections. For the comparison of gene expression and presence of intimal macrophages, the cumulative data from each aortic tissue segment were used.
To provide a quantitative measure of intimal macrophage presence, the RAM-11–positive intimal area was determined by computer-assisted image analysis using an imaging system previously described.20 In brief, electronic images were captured with a Sony DXC-960MD CCD color video camera mounted on a Nikon Microphot microscope. Quantitation was performed with a 586-133 PC, an Oculus TCX true color frame grabber with 4 megabytes of frame buffer memory (Coreco) controlling a separate VGA image monitor, and Optimas 4.0 software (Bioscan). RAM-11–positive areas were outlined using a digitizer tablet. This method avoids errors resulting from unstained areas within macrophage/foam cells that would affect a threshold-based measurement. Furthermore, clearly necrotic areas were not included in the measurement, because such areas are unlikely to contribute to gene expression. The entire size of atherosclerotic lesions was also determined by image analysis. All morphometric measurements were performed by the same operator.
Immunostaining of Frozen Sections for
VCAM-1
Serial 7-µm-thick sections were prepared on coated slides and
fixed for 5 minutes with acetone at -20°C immediately before
immunostaining. Sections were stained with Rb1/9 (1:100
dilution), a monoclonal antibody against rabbit VCAM-1 41
(a generous gift from Dr M.I. Cybulsky and Dr M.A. Gimbrone, Harvard
Medical School), following the procedure described in reference 41.
Biotinylated horse anti-mouse IgG (Vector Labs) was used as secondary
antibody, followed by incubation with an avidin/biotin/peroxidase
complex (Vector Labs) and a substrate for peroxidase,
3-amino-9-ethylcarbazole (Sigma). Sections were then counterstained
with Gill’s hematoxylin. Controls in which the primary antibody was
omitted were devoid of any staining.
Statistical Analysis
Results were expressed as mean±SEM. Differences between groups
were assessed by ANOVA. The statistical analyses were performed
using standard statistical software (Systat for Windows, Systat).
Results
Two groups of 12 WHHL rabbits each were fed regular rabbit chow with or without 1% (wt/wt) probucol, beginning at age 3 weeks, for up to 15 weeks. Three animals from each group were killed at 6, 9, 12, and 18 weeks of age. Total plasma cholesterol levels in control animals were 724±79 mg/dL and did not show significant differences between the different time points. As expected, cholesterol levels were initially lower in probucol-treated animals (490±29 mg/dL in animals killed at age 6 weeks) and increased gradually to 750±43 mg/dL in animals killed at 18 weeks of age. Probucol levels were similar at the different time points; the average plasma concentration was 157.0±27.6 µmol/L.
Analysis of Gene Expression by Quantitative PCR
Aortic segments were isolated and processed as described in
"Methods." Total RNA was prepared from tissue samples (18.6±3.3
mg wet weight), with an average yield of 0.82±0.23 µg RNA/mg tissue.
The integrity of the isolated RNA was demonstrated by denaturing
agarose gel electrophoresis, which showed two strong bands
representing the 18S and 28S rRNA but no degradation
products (data not shown).
Gene expression of VCAM-1, MCP-1, and M-CSF in aortic tissues was
analyzed by semiquantitative PCR, using GAPDH as internal
standard. Results were expressed as the ratio between target gene
expression and GAPDH expression. To verify that GAPDH constitutes a
suitable internal standard, expression of GAPDH in aortic tissue was
determined quantitatively by competitive PCR, using a cDNA competitor
identical to the PCR product of GAPDH, except for a 102-bp deletion
(to allow for separation by gel electrophoresis) (Fig 2A). As expected, gels showed only two
bands, representing the 465-bp PCR product of aortic
GAPDH and the truncated product of the competitor. To establish
that gene expression of GAPDH in various aortic segments was comparable
and not affected by probucol treatment, we determined GAPDH expression
in corresponding tissue samples from probucol-treated and control
animals at various time points of the study. GAPDH expression in the
same aortic segment did not change significantly over time, and data
from all time points were therefore pooled. As shown in Fig 2B, probucol treatment did not affect GAPDH expression in either the DTA or
ABD. However, a significant difference in GAPDH expression was found
between the DTA and ABD. When data from the probucol-treated and
control animals were pooled (combined data in Fig 2B), GAPDH expression
in the DTA was 1.5-fold higher than in the ABD. These results indicate
that, within the time frame of our study, GAPDH expression was
independent of the age of the animals and was not affected by probucol
treatment. Thus, GAPDH expression may be considered a good internal
standard when comparing corresponding segments of the aorta. In
contrast, differences in GAPDH expression must be taken into
consideration when comparing relative levels of target gene expression
in the DTA to those in the ABD.
|
Suitable PCR primers were developed to determine VCAM-1, MCP-1, and
M-CSF gene expression in the intima/media of anatomically defined
segments of rabbit aortas, using GAPDH as internal standard. As
described in "Methods," these PCR primers were chosen so that the
size of the PCR products of target and reference genes would be
sufficiently different to allow separation by agarose gel
electrophoresis. The size of the PCR products resulting from the
selected PCR primers is shown in Table 1. GAPDH and target gene
sequences could therefore be amplified in the same reaction tube, a
prerequisite for parallel amplification.
Expression of VCAM-1 Over Time
Initially, we assessed gene expression in WHHL rabbits over time.
Fig 3 shows the expression of the VCAM-1
gene in the first segment of the DTA (DTA1) of untreated WHHL rabbits.
Even at the earliest time point examined (6 weeks of age), VCAM-1 was
expressed in the DTA (as well as the ABD; data not shown) of WHHL
rabbits. Compared with the level of expression at age 6 weeks, the
increase in average VCAM-1 expression at age 9 and 12 weeks was not
statistically significant, but by 18 weeks, a marked upregulation of
VCAM-1 expression was seen in the DTA. Similar results were also
obtained in the DTA5 segments but not in the abdominal segments free of
lesion (not shown). At all time points examined, VCAM-1 expression was
lower in probucol-treated animals than in untreated controls, and the
marked upregulation of VCAM-1 that occurred over time in the untreated
group was completely abolished by probucol treatment.
|
Contrary to expectation, a significant upregulation of VCAM-1 gene
expression appeared to occur relatively late in our time-course
experiment (Fig 3). As shown below, the fact that no significant
increase was observed at earlier time points was in part due to the
considerable variability in the extent of atherogenesis between the
animals sacrificed at the same age (despite careful litter matching of
animals assigned to various time points).
Extent of Atherosclerosis
To obtain an overall measurement of
atherosclerosis, the intimal lesion area was determined
by image analysis of all sections. However, our primary aim was
to determine the correlation between the expression of VCAM-1, MCP-1,
and M-CSF and the intimal macrophage content, for the following
reasons. Macrophages express MCP-1 and M-CSF in
lesions.31 32 33 Thus, it is possible that gene and protein
expression of these factors are proportional to the number of intimal
macrophages. In turn, MCP-1 and M-CSF may contribute to the
recruitment of monocytes into the vessel wall and/or their survival
within the intima. VCAM-1 may also play a role in the recruitment of
monocytes into lesions, although its role in the penetration of
macrophages is less well established than for T cells.
Two different measures of intimal macrophages were used. The
first consisted of the scoring of lesions on the basis of their size
and macrophage content. Fig 4 shows
representative sections without lesions (stage 0; A);
with isolated intimal monocyte/macrophages (stage 1; B); with a
continuous monolayer of monocyte/macrophages (stage 2, C); and
with multiple layers of macrophages, as well as transitional
lesions (stage 3; D through F). For each aortic segment (DTA3), a total
of 14 slides representing 630 µm proximal and
630 µm distal to the orifice of the first intercostal orifice
were analyzed, as described in "Methods." Lesions to the
left and right of the orifice were assessed separately. The cumulative
value of these 28 determinations was then used as the overall
"lesion score" (Table 2). In
addition, the intimal macrophage area was determined by
computer-assisted morphometry, as described in "Methods." This
procedure provides a quantitative measure of intimal
macrophages but reflects primarily macrophage size,
rather than macrophage numbers. Furthermore, cumulative values
of macrophage areas of early atherosclerotic lesions are easily
skewed by a few large lesions.
|
|
At 6 weeks of age, only 4 of the 42 sections studied from the three
untreated WHHL rabbits showed any lesions, and only isolated intimal
macrophages (stage 1 lesions) were observed. At 9 weeks of age,
two of three control rabbits had similar stage 1 lesions, whereas the
third animal showed both single layers of monocyte/macrophages
(stage 2) as well as more complicated multilayered lesions (stage 3).
At 12 weeks of age, a similar nonhomogeneous picture was obtained. One
animal showed lesions of all three stages, including transitional
lesions, in particular in the vicinity of flow dividers (Fig 4F); the
second animal showed a single stage 2 lesion and one stage 3 lesion;
and the third had only stage 1 lesions. At 18 weeks, all three control
animals had developed numerous stage 2 lesions, and two animals also
had stage 3 lesions. By contrast, none of the 12 probucol-treated
rabbits showed any lesions, except for isolated macrophages in
2 of the 168 slides studied (from a 12-week-old rabbit). It was
particularly noteworthy that the aortic wall of all three
probucol-treated animals examined at 18 weeks was completely devoid of
intimal macrophages.
These results highlight the fact that atherogenesis in animals of the
same time point showed considerable variability. In particular, one
animal in the 9-week-old group showed extensive lesion formation.
Although these observations are consistent with the gene
expression (Fig 3), all subsequent analysis of gene expression
was based on the extent of intimal macrophage presence in each
animal, irrespective of the time point.
Basal Levels of VCAM-1 Gene Expression
Immunocytochemistry with RAM-11 showed that basal levels of VCAM-1
gene expression in the DTA and ABD of 6-week-old WHHL rabbits occurred
in the complete absence of macrophage-containing
atherosclerotic lesions. To investigate whether the unexpected
observation of a "basal" expression of VCAM-1 might be due to the
animal model used, eg, occurring as a result of a very early onset of
atherogenesis in WHHL rabbits and/or as a direct result of the very
high plasma cholesterol level in these animals, VCAM-1
expression was also determined in comparable aortic segments of
normocholesterolemic NZW rabbits. Four NZW rabbits fed
regular rabbit chow showed levels of aortic VCAM-1 expression very
similar to those seen in the lesion-free segments of the DTA of three
6-week-old WHHL rabbits (0.813±0.179 and 0.835±0.041 in NZW and WHHL
rabbits, respectively).
Comparison of VCAM-1 Gene Expression and VCAM-1 Protein
Immunocytochemical detection of rabbit VCAM-1 protein with the
monoclonal antibody Rb1/9 could not be carried out in the
formaldehyde-fixed and paraffin-embedded sections of the actual rabbits
used for our time-course experiment. To verify that the gene expression
of VCAM-1 results in the presence of VCAM-1 protein, frozen sections
were prepared from two additional NZW and two WHHL rabbits. These
sections were immunostained with Rb1/9 (Fig 4G through 4J).
Patches of VCAM-1 staining were consistently found in the
endothelial layer of normal (macrophage-free)
areas of both NZW (Fig 4G) and WHHL rabbits (data not shown). This
finding is consistent with our observation of a "basal"
level of gene expression in normal aortas. No VCAM-1 protein was
detected in the media of these aortas. VCAM-1 expression in early
lesions (comparable to those typically seen in the WHHL rabbits of our
time-course experiment) was also limited to the
endothelium. Most of the surface of lesions showed at
least some immunostaining for VCAM-1, whereas adjacent
areas of normal endothelium generally did not (Fig 4H).
The subintimal areas of these lesions occupied predominantly by
macrophage/foam cells did not stain for VCAM-1. Presence of
VCAM-1 in deeper layers of the intima and in the necrotic core was
observed only in lesions that were much more advanced than the ones
that typically occur in 18-week-old WHHL rabbits. Nevertheless, in more
advanced lesions, intimal cells other than endothelial
cells can clearly produce substantial amounts of VCAM-1. Fig 4I and 4J
show examples of such advanced atherosclerotic lesions from the ABD of
a 10-month-old WHHL rabbit. The presence of VCAM-1 in the vicinity of
the internal elastic lamina and in deeper areas of the intima (ie, in
areas staining for actin epitopes) suggests VCAM-1 expression by smooth
muscle cell–derived intimal cells. VCAM-1 gene expression on
corresponding aortic segments of the same rabbits was 1.004 in the DTA
of the normal NZW shown in Fig 4G, 2.697 in the DTA of the 10-month-old
WHHL rabbit shown in Fig 4H, and 4.476 in the ABD of the same rabbit
shown in Fig 4I and 4J. These results also suggest a strong correlation
between VCAM-1 gene and protein expression.
Correlation Between VCAM-1 Expression and Intimal Macrophage
Accumulation
To determine whether gene expression of VCAM-1 was correlated with
the extent of atherosclerosis, we compared the two
parameters, using data from all animals of the same
treatment group (irrespective of the age of the animals). For this
analysis, the level of gene expression in each animal
represented the mean of the levels in the DTA1 and DTA5
segments, and the intimal monocyte/macrophage presence was
expressed either as the cumulative macrophage score or as the
area of RAM-11–positive cells. In the 12 animals of the control group,
a linear relationship between VCAM-1 expression and the cumulative
macrophage score was apparent (Fig 5A). When VCAM-1 expression was
correlated with the area of RAM-11–positive intimal
macrophages determined by morphometry, a striking correlation
was found (R2=.89, P<.001; Fig 5B).
However, the influence of the two animals with the most extensive
atherosclerosis on the correlation of VCAM-1 expression
with the macrophage area was visibly greater than on the
comparison of gene expression with the macrophage score. The
correlation of gene expression with the size of atherosclerotic lesions
was also significant, but only when data from the animals with the most
advanced atherosclerosis were excluded, whereas
significance was lost when data from all animals were analyzed
(not shown). This was not surprising, because in most animals with
early stages of atherosclerosis, the overall lesion
size correlated very well with the macrophage area, whereas in
the animal with the most extensive atherosclerosis,
only 29% of the lesion area was taken up by macrophages.
|
), whereas the absence
of lesions and macrophages precluded a correlation in the 12
probucol-treated animals (
).In the probucol-treated group, no correlation between VCAM-1 gene expression and intimal macrophages could be established due to the almost complete absence of lesions. VCAM-1 expression in all 12 probucol-treated WHHL rabbits (0.605±0.080; mean±SEM) was 28% lower than that of the 6 untreated WHHL rabbits that had no or only minimal lesions (score <10) (0.840±0.095; P=.047). VCAM-1 expression in these 6 untreated WHHL rabbits was also similar to that of four lesion-free NZW rabbits (0.813±0.179).
Correlation Between MCP-1 Expression and Intimal Macrophage
Accumulation
Gene expression of MCP-1 in the 12 untreated animals also showed a
linear relationship with the macrophage score (Fig 6A). The correlation between MCP-1
expression and the macrophage area (Fig 6B) was also
significant (R2=.42, P=.022) but much
weaker than the correlation of VCAM-1 (Fig 5B). In the probucol-treated
group, 2 of 12 animals showed increased MCP-1 expression that was not
associated with the presence of intimal macrophages. This
increase also was not paralleled by an increase in VCAM-1
expression. However, no effect of probucol on MCP-1 expression was
noticeable when data from the 12 treated WHHL rabbits were compared
with those of 6 untreated WHHL rabbits with no or only minimal lesions
(0.740±0.093 versus 0.782±0.060; not significant). Expression of
MCP-1 in the aorta of the four lesion-free NZW aortas was lower (0.521)
and comparable to that in the ABD of WHHL rabbits.
|
Correlation Between M-CSF Expression and Intimal Macrophage
Accumulation
A different picture was seen for gene expression of M-CSF (Fig 7). Levels of expression were relatively
constant throughout the study period in both the treated group and the
untreated group, and M-CSF expression in the control group showed no
correlation with the macrophage score (left) or the
macrophage area (data not shown). However, the overall M-CSF
expression was 41.7% lower in the probucol-treated group (right). The
average value from the DTA1 and DTA5 segments was 0.271±0.058 in the 6
control rabbits with minimal or no lesions compared with 0.155±0.030
in the 12 probucol-treated rabbits (P=.0006). Similar
results were obtained when data from all 12 control rabbits were pooled
(average M-CSF expression, 0.266±0.006). The fact that an increase in
lesion size was not accompanied by an increase in M-CSF expression
suggests that macrophages are not the predominant source of
M-CSF.
|
Expression of Different Forms of VCAM-1
It has been previously reported that two alternately spliced forms
of VCAM-1 exist, but to date no functional difference between these
forms has been described.44 45 We therefore addressed the
question of whether the two forms might contribute differently to the
expression of VCAM-1 in aortic tissue. Since the pair of VCAM-1 primers
we had used for the studies described above did not differentiate
between the forms, an additional set of primers was developed (VCAM-1
"B"; see Table 1). The expected size of the PCR products
resulting from this new pair of primers was 511 bp for the unspliced
and 235 bp for the spliced product. Of all 72 WHHL tissue samples
analyzed, only 2 samples showed a weak band at a size
corresponding to the expected spliced product. The unspliced
product (511 bp) was by far the predominant form expressed in any
arterial segment tested. In contrast, the spliced
product only became detectable after 40 cycles of PCR
amplification. Under these conditions, quantification of the gene
products would not be reliable, because the amplification reaction
is no longer in the exponential phase.
Discussion
The penetration of monocytes into the intima, their phenotypic transformation into macrophages, and subsequent foam cell formation are key events in early atherogenesis. Substantial evidence suggests that oxidized lipoproteins, or oxidative processes in general, contribute to these events, and the fact that several different antioxidants, including probucol, reduce lesion formation in animal models strongly supports this. Some of the atherogenic mechanisms can be expected to be directly proportional to the amount of OxLDL (or its byproducts). Other atherogenic effects may stem from the modulation of gene expression of vascular cells by OxLDL and may be triggered by small amounts of OxLDL or mild degrees of oxidation.4 Genes that are upregulated by OxLDL alone or in conjunction with other factors (such as TNF) and that could have profound effects on the intimal accumulation of macrophages are VCAM-1, MCP-1, and M-CSF. In the present analysis, we studied the expression of these three genes in vivo, during the early stages of lesion formation.
It is likely that a broad variability exists in the expression of these genes in different animal models of atherosclerosis, in different parts of the vasculature, and in relation to the extent and rate of progression of the lesions. Thus, conclusions regarding the biological role of individual genes can only be drawn by comparing very similar arterial tissues under tightly controlled conditions. For example, one can follow gene expression over time or compare expression between antioxidant-treated and control animals. In the present study, we compared gene expression at defined arterial sites before lesion formation and during early atherogenesis in untreated WHHL rabbits with that of litter-matched probucol-treated animals. We hypothesize that this procedure would not only indicate the effects of probucol on the selected genes but also provide valuable information about the potential biological importance of the respective gene in atherogenesis. However, several caveats have to be addressed for such a study.
The first of these caveats regards the choice of the animal model. We felt that the availability of histological data on lesion composition would be essential for the interpretation of gene expression. Because it is not possible to perform both analyses on the same tissue, it is of the utmost importance that the degree of atherosclerosis in the tissue samples studied by PCR and immunocytochemistry be as similar as possible. This is clearly the case in the thoracic aorta of WHHL rabbits, where very consistent early lesions form distal to each pair of intercostal arteries. By choosing this model, we were able to perform quantitative PCR on two aortic segments flanking the segment used for immunocytochemistry. The comparison of the two PCR quantifications may also be used as additional evidence for the similarity of the lesions in each animal. We also chose the WHHL rabbit, because hypercholesterolemia and atherogenesis in this model occur spontaneously and because we have accumulated extensive data on atherogenesis in WHHL rabbits.20 21 48 49 50 51 Most importantly, several previous studies have demonstrated that probucol reduces lesions by 50% to 70% in this strain.11 12 21 We were of course aware of the possibility that the early onset of hypercholesterolemia in WHHL rabbits might affect the base levels of gene expression. This was controlled for by studying comparable aortic segments from NZW rabbits fed a nonatherogenic diet and by carefully documenting that virtually no atherosclerosis was present in the WHHL rabbits we studied at age 6 weeks.
The second caveat concerns the fact that we analyzed both the media and intima by PCR. This was imposed by our focus on early atherogenesis. A separate analysis of the intima would be desirable, because it would be more sensitive to changes in gene expressions in very small lesions, in particular if low-level expression of the same gene occurs in the media. However, a clean and consistent separation of the normal intima or that of very early lesions from the media is difficult, and the quantities of intact mRNA isolated from the normal intima were insufficient for multiple PCR analyses.
The third caveat regards the choice of the antioxidant. Although probucol is very effective in reducing the progression of atherosclerosis in rabbits11 12 21 and primates,15 its efficacy in humans has not been demonstrated, and concerns have been raised about the risk associated with lowering plasma HDL levels. Furthermore, probucol may have other cellular effects unrelated to its antioxidant properties.39 40 One effect of probucol that could influence our study is the lowering of the total plasma cholesterol level mentioned above. Although the studies of Carew et al11 demonstrated that the antiatherogenic effect of probucol is independent of its lipid-lowering effect in WHHL rabbits, we could not rule out a priori that small differences in plasma cholesterol levels might influence gene expression. However, this is very unlikely, because expression of each of the three genes was similar in probucol-treated WHHL rabbits throughout the study period, in control WHHL rabbits before the onset of measurable atherosclerosis, and in NZW rabbits (which had very low plasma cholesterol levels) fed a nonatherogenic diet.
The last caveat regards the quantification of gene expression.
Quantitative PCR technology is prone to errors, such as
inconsistent yields of the RNA isolation, the RT reaction, or
the PCR amplification itself. We have chosen an experimental approach
that greatly reduces the impact of such errors. A particular advantage
of the semiquantitative method used is the fact that the internal
standard and the target are processed in parallel throughout the
procedure. As described in "Methods," we have amply documented
the accuracy and reproducibility of the results. Because earlier
reports had raised the possibility that GAPDH may be
inconsistently expressed in tissues,52 53 54 a
thorough analysis of GAPDH expression in arterial
tissues was performed. Data shown in Fig 2 demonstrated that GAPDH
expression was sufficiently constant throughout the study period and
was not affected by probucol treatment. This is in agreement with
published reports showing that GAPDH expression was the most
consistent among several genes frequently used as internal
standards.55 56
VCAM-1 is considered to be a key molecule in the initiation and growth of atherosclerotic lesions. Its role in the adhesion of T cells has been demonstrated in in vitro flow chambers,57 and the presence of T cells in atherosclerotic lesions is well known.58 59 In contrast, the contribution of VCAM-1 to monocyte adhesion is less well established. VCAM-1 also seems to play a substantial role in cardiac transplantation and allograft rejection.60 61 VCAM-1 has previously been demonstrated in aortic endothelium overlaying atherosclerotic lesions of animal and human arteries.41 42 43 62 Two differentially spliced forms of VCAM-1 exist, but no functional difference has been described.44 45 Our data indicate that the long form of VCAM-1 is the predominant, if not exclusive, form expressed in vivo in the artery wall, a finding that supports earlier reports by Cybulsky et al.44
It was recently reported that expression of VCAM-1 preceded that of
intimal macrophages in rabbits fed an atherogenic
diet.41 The present study demonstrated that gene
expression of VCAM-1 is strongly correlated with the presence of
subendothelial macrophages, irrespective of
whether the macrophage score or the area of RAM-11–positive
cells was used as a parameter of intimal macrophage
accumulation (Fig 5A and 5B). Interestingly, "basal" levels of
VCAM-1 gene expression were found even in aortic segments of WHHL
rabbits free of intimal macrophages. Normal aortic segments of
NZW rabbits fed a nonatherogenic diet also showed very similar basal
expression levels. Furthermore, the finding of a basal level of VCAM-1
expression was supported by the demonstration of the presence of VCAM-1
protein in normal intima of NZW aortas (Fig 4G), consistent
with an earlier report of VCAM-1 in the endothelium of
normal rabbit coronary arteries.60 It is not clear
why these results differ from those of other immunocytochemical studies
that found VCAM-1 only at lesion sites.41 42 43
Probucol not only effectively inhibited atherogenesis and upregulation
of VCAM-1 expression during the first 18 weeks but also reduced VCAM-1
expression below the basal level. These results support the notion that
VCAM-1 plays an important role in early atherogenesis, possibly as a
"rate-limiting factor." However, the present studies were not
designed to provide evidence for a causal relationship between VCAM-1
expression and intimal monocyte accumulation. The correlation between
the two parameters could indicate increased monocyte
adhesion and penetration caused by increased VCAM-1 expression, as well
as an increased VCAM-1 expression by endothelial cells
overlying progressively bigger lesions (possibly induced by secretory
products of intimal macrophages). In more advanced lesions,
VCAM-1 expression by intimal smooth muscle cells is likely to
contribute significantly to the overall expression43 (Fig 4I and 4J). A gene knockout of VCAM-1 was lethal,63 and no
intervention inhibiting VCAM-1 in vivo, eg, by blocking antibodies, has
been reported to date.
Even in the absence of conclusive data establishing the biological role
of VCAM-1 in monocyte adhesion, the vascular expression of VCAM-1 and
its regulation during atherogenesis are of great interest, eg, because
of its role in T-cell adhesion. A number of molecules are known to
affect VCAM-1 expression. These include byproducts generated during
the oxidation of LDL, such as 13-HPODE35 and
lysophosphatidylcholine,64 as well as the
cytokines IL-4 and TNF-.65 Native LDL or LDL
oxidized to different degrees does not induce VCAM-1 expression in
cultured endothelial cells. However, OxLDL increases
VCAM-1 expression induced by TNF-.35 More recently, it
was also reported that advanced glycosylation end products promote
VCAM-1 expression and atheroma formation in normal
rabbits.66 This may contribute to enhanced
arteriosclerosis in diabetic subjects, as well as
in hypercholesterolemic subjects.50 Our
data clearly indicate that probucol downregulates VCAM-1 expression in
the vascular wall even before lesion formation is initiated (Fig 3).
The mechanism of this effect is unclear. If one assumes that VCAM-1 is
indeed a rate-limiting factor in early atherogenesis and that small
amounts of oxidized lipoproteins or their byproducts are sufficient
to trigger its upregulation, a high threshold of antioxidant protection
may be required to prevent this occurrence. This may explain why
antioxidants less powerful than probucol, eg, vitamin E, are not very
effective in reducing lesion progression.20 21 On the
other hand, direct or indirect cellular effects of probucol independent
of the formation of OxLDL, or even independent of its antioxidant
effect, may occur. Furthermore, De Catarina et al67 68
showed that fatty acids can exert direct endothelial
effects and reduce cytokine-induced VCAM-1 expression.
Future studies investigating the effect of other antioxidants (which do
not inhibit atherogenesis) on VCAM-1 expression in vivo and the effects
of probucol treatment in more advanced atherosclerosis
are necessary to address these questions.
Expression of MCP-1 mRNA and the presence of MCP-1 protein have been
shown by in situ hybridization and immunocytochemistry in
atherosclerotic lesions of rabbits, primates, and
humans.31 69 70 71 While there is a consensus that MCP-1
expression in normal arteries ranges from low to undetectable, the
cells responsible for increased expression in atherosclerotic arteries
are less well established. It appears that in early lesions,
endothelial cells and macrophages contribute
evenly, whereas in more advanced lesions, macrophages become
the predominant source of MCP-1.69 Contradictory
observations exist concerning the role of smooth muscle
cells.31 69 70 71 Data on the regulation of MCP-1 expression
include in vitro studies demonstrating that the contact of monocytes
with endothelial cells on transmigration stimulates
MCP-1 expression.72 Increased MCP-1 release by
endothelial cells and smooth muscle cells can also be
induced by incubation with minimally modified LDL.73 We
found that a significant correlation between macrophage
infiltration and levels of MCP-1 expression exists during early
atherogenesis in WHHL rabbits (Fig 6). These results support the view
that macrophages are an important source of MCP-1 in developing
lesions (either directly or indirectly, by stimulating MCP-1 expression
by other vascular cells) and that upregulation of MCP-1 occurs on
monocyte transmigration. However, macrophages cannot be the
sole source, because basal level expression was found before the onset
of lesion formation and was also present in normal aortas from NZW
rabbits. Furthermore, the increase in MCP-1 expression with progressive
atherogenesis was less dramatic than that seen for VCAM-1 (Fig 6). The
failure of probucol treatment to significantly affect MCP-1 expression
suggests that the antiatherogenic effect of probucol is not mediated by
changes in MCP-1.
M-CSF has also been demonstrated in atherosclerotic
lesions.32 33 In lesions of cholesterol-fed
rabbits, macrophages were the predominant source of M-CSF gene
expression,74 but endothelial cells and
smooth muscle cells can also express M-CSF. Recent in vitro studies
showed that minimally modified LDL but not extensively OxLDL induced
M-CSF expression in endothelial cells or smooth muscle
cells.27 33 The involvement of the oxygen
radical–sensitive transcription factor NF-B in this process has
been suggested to mediate this effect.28 Our studies
showed that M-CSF expression did not change during the 12-week study
period in either the untreated or probucol-treated group (Fig 7).
However, overall levels of M-CSF expression were decreased by probucol
treatment. Although lesion progression clearly is not dependent on
increased levels of M-CSF, it cannot be ruled out that a certain basal
level of M-CSF expression in the lesion is necessary for atherogenesis
and that probucol may in part act by reducing the basal expression.
This would be in agreement with the observation of reduced
atherosclerosis in apoE-deficient mice that are M-CSF
deficient.75 The fact that we did not find increased M-CSF
expression in arteries with increased numbers of
subendothelial macrophages indicates that
macrophages are not the predominant cells responsible for M-CSF
expression in the vascular wall. This conflicts with the results of an
earlier study.74 However, the type of lesion in this study
(induced by feeding NZW rabbits a high-cholesterol diet for
10 weeks) is quite different from that in WHHL rabbits. Our finding
that probucol lowered overall levels of M-CSF expression supports the
notion that oxidative processes may be involved in regulating M-CSF
expression, as suggested by the upregulation of M-CSF by minimally
modified LDL.
Selected Abbreviations and Acronyms
|
Acknowledgments
These studies were supported by National Heart, Lung, and Blood Institute grant HL14197 (La Jolla Specialized Center of Research in Arteriosclerosis). The probucol used in the experiments was a generous gift from Merrell Dow Pharmaceuticals; competitor cDNA was provided by Dr H. Lukhaup, University of Heidelberg; RAM-11 monoclonal antibody specific for rabbit macrophages was given by Dr A.M. Gown, University of Washington; and Rb1/9 monoclonal antibody against rabbit VCAM-1 was given by Drs M.I. Cybulsky and M.A. Gimbrone, Harvard Medical School. We thank Drs Daniel Steinberg, Joseph L. Witztum, Peter D. Reaven, and Christopher K. Glass for discussions and review of the manuscript. We also thank Jennifer Pattison, Florencia Casanada, Stephan Palmer, Suzan Butler, Joe Juliano, and Frank Peralta for technical assistance.
Received June 3, 1996; accepted September 24, 1996.
References
- Steinberg D, Parthasarathy S, Carew TE, Khoo JC, Witztum JL. Beyond cholesterol: modifications of low density lipoprotein that increase its atherogenicity. N Engl J Med. 1989;320:915-924.[Medline] [Order article via Infotrieve]
- Witztum JL, Steinberg D. Role of oxidized low density lipoprotein in atherogenesis. J Clin Invest. 1991;88:1785-1792.[Medline] [Order article via Infotrieve]
- Palinski W, Witztum JL. Oxidative stress and diabetes mellitus. In: Born GVR, Schwartz CD, eds. New Horizons in Diabetes Mellitus and Coronary Heart Disease. London, UK: Current Science; 1995:111-123.
- Berliner JA, Navab M, Fogelman AM, Frank JS, Demer LL,
Edwards PA, Watson AD, Lusis AJ.
Atherosclerosis: basic mechanisms: oxidation,
inflammation, and genetics. Circulation. 1995;91:2488-2496.
[Abstract/Free Full Text] - Quinn MT, Parthasarathy S, Fong LG, Steinberg
D. Oxidatively modified low density lipoproteins: a potential
role in recruitment and retention of monocyte/macrophages
during atherogenesis. Proc Natl Acad Sci U S A. 1987;84:2995-2998.
[Abstract/Free Full Text] - Quinn MT, Parthasarathy S, Steinberg D.
Lysophosphatidylcholine: a new chemotactic factor for human monocytes
and its potential role in atherogenesis. Proc Natl Acad
Sci U S A. 1988;85:2805-2809.
[Abstract/Free Full Text] - Goldstein JL, Ho YK, Basu SK, Brown MS. Binding
site on macrophages that mediates uptake and degradation of
acetylated low density lipoprotein, producing massive
cholesterol deposition. Proc Natl Acad Sci
U S A. 1979;76:333-337.
[Abstract/Free Full Text] - Kodama T, Freeman M, Rohrer L, Zabrecky T, Matsudaira P, Krieger M. Type I macrophage scavenger receptor contains alpha-helical and collagen-like coiled coils. Nature. 1990;343:531-535.[Medline] [Order article via Infotrieve]
- Endemann G, Stanton LW, Madden KS, Bryant CM, White
RT, Protter AA. CD36 is a receptor for oxidized low density
lipoprotein. J Biol Chem. 1993;268:11811-11816.
[Abstract/Free Full Text] - Ramprasad M, Fischer W, Witztum JL, Sambrano G,
Quehenberger O, Steinberg D. The 94-96 kDa mouse
macrophage membrane protein that recognizes oxidized low
density lipoprotein and phosphatidylserine-rich
liposomes is identical to macrosialin, the mouse homologue of human
CD68. Proc Natl Acad Sci U S A. 1995;92:9580-9584.
[Abstract/Free Full Text] - Carew TE, Schwenke DC, Steinberg D.
Antiatherogenic effect of probucol unrelated to its
hypocholesterolemic effect: evidence that antioxidants
in vivo can selectively inhibit low density lipoprotein degradation in
macrophage-rich fatty streaks slowing the progression of
atherosclerosis in the WHHL rabbit. Proc
Natl Acad Sci U S A. 1987;84:7725-7729.
[Abstract/Free Full Text] - Kita T, Nagano Y, Yokode M, Ishii K, Kume N, Ooshima A,
Yokida H, Kawai C. Probucol prevents the progression of
atherosclerosis in Watanabe heritable
hyperlipidemic rabbit, an animal model for familial
hypercholesterolemia. Proc Natl
Acad Sci U S A. 1987;84:5928-5931.
[Abstract/Free Full Text] - Björkhem I, Henriksson-Freyschuss A, Breuer O,
Diczfalusy U, Berglund L, Henriksson P. The antioxidant
butylated hydroxytoluene protects against
atherosclerosis. Arterioscler
Thromb. 1991;11:15-22.
[Abstract/Free Full Text] - Sparrow CP, Doebber TW, Olszewski J, Wu MS, Ventre J, Stevens KA, Chao YS. Low density lipoprotein is protected from oxidation and the progression of atherosclerosis is slowed in cholesterol-fed rabbits by the antioxidant N,N'-diphenyl-phenylenediamine. J Clin Invest. 1992;89:1885-1891.[Medline] [Order article via Infotrieve]
- Sasahara M, Raines EW, Chait A, Carew TE, Steinberg D, Wahl PW, Ross R. Inhibition of hypercholesterolemia-induced atherosclerosis in Macaca nemestrina by probucol, I: intimal lesion area correlates inversely with resistance of lipoproteins to oxidation. J Clin Invest. 1994;94:155-164.[Medline] [Order article via Infotrieve]
- Tangirala RK, Casanada F, Witztum JL, Steinberg D,
Palinski W. Effect of the antioxidant N,N'-diphenyl
1,4-phenylenediamine (DPPD) on atherogenesis in
apoE-deficient mice. Arterioscler Thromb Vasc Biol. 1995;15:1625-1630.
[Abstract/Free Full Text] - Williams RJ, Motteram JM, Sharp CH, Gallagher PJ. Dietary vitamin E and the attenuation of early lesion development in modified Watanabe rabbits. Atherosclerosis. 1992;94:153-159.[Medline] [Order article via Infotrieve]
- Prasad K, Kalra J. Oxygen free radicals and hypercholesterolemic atherosclerosis: effect of vitamin E. Am Heart J. 1993;125:958-973.[Medline] [Order article via Infotrieve]
- Willingham AK, Bolanos C, Bohannan E, Cenedella RJ. The effects of high levels of vitamin E on the progression of atherosclerosis in the Watanabe heritable hyperlipidemic rabbit. J Nutr Biochem. 1993;4:651-654.
- Fruebis J, Carew TE, Palinski W. Effect of vitamin E on atherogenesis in LDL receptor-deficient rabbits. Atherosclerosis. 1995;117:217-224.[Medline] [Order article via Infotrieve]
- Fruebis J, Steinberg D, Dresel HA, Carew TE. A comparison of the antiatherogenic effects of probucol and of a structural analog of probucol in LDL-receptor–deficient rabbits. J Clin Invest. 1994;94:392-398.[Medline] [Order article via Infotrieve]
- Cathcart MK, Morel DW, Chisolm GM. Monocytes and neutrophils oxidize low density lipoprotein making it cytotoxic. J Leukoc Biol. 1985;38:341-350.[Abstract]
- Kugiyama K, Kerns SA, Morrisett JD, Roberts R, Henry PD. Impairment of endothelium-dependent arterial relaxation by lysolecithin in modified low density lipoproteins. Nature. 1990;344:160-162.[Medline] [Order article via Infotrieve]
- Radomski MW, Moncada S. Regulation of vascular homeostasis by nitric oxide. Thromb Haemost. 1993;70:36-41.[Medline] [Order article via Infotrieve]
- Clinton SK, Libby P. Cytokines and growth factors in atherogenesis. Arch Pathol Lab Med. 1992;116:1292-1300.[Medline] [Order article via Infotrieve]
- Cushing SD, Berliner JA, Valente AJ, Territo MC, Navab
M, Parhami F, Gerrity R, Schwartz CJ, Fogelman AM. Minimally
modified low density lipoprotein induces monocyte chemotactic protein 1
in human endothelial cells and smooth muscle
cells. Proc Natl Acad Sci U S A. 1990;87:5134-5138.
[Abstract/Free Full Text] - Rajavashisth TB, Andalibi A, Territo MC, Berliner JA, Navab M, Fogelman AM, Lusis AJ. Induction of endothelial cell expression of granulocyte and macrophage colony stimulating factors by modified low density lipoproteins. Nature. 1990;334:254-257.
- Rajavashisth TB, Yamada H, Mishra NK. Transcriptional activation of the macrophage-colony stimulating factor gene by minimally modified LDL: involvement of nuclear factor-kappa B. Arterioscler Thromb Vasc Biol. 1995;15:1