(Received for publication, July 9, 1996, and in revised form, November 27, 1996)
From the Departments of Medicine and Biochemistry,
Rush Medical College, Chicago, Illinois 60612, the
§ Veteran's Affairs Medical Center and Department of
Internal Medicine, University of Utah, Salt Lake City, Utah 84117, and
the ¶ Biochemistry Laboratory, American Red Cross,
Rockville, Maryland 20855
Macrophages are a significant source of
lipoprotein lipase (LPL) and apolipoprotein E (apo E) in the developing
arterial wall lesion, and each of these proteins can importantly
modulate lipid and lipoprotein metabolism by arterial wall cells. LPL
and apo E share a number of cell surface binding sites, including
proteoglycans, and we have previously shown that proteoglycans are
important for modulating net secretion of apoprotein E from
macrophages. We therefore evaluated a potential role for LPL in
modulating net secretion of macrophage-derived apo E. In pulse-chase
experiments, addition of LPL during the chase period produced a
decrease in secretion of apoprotein E from human monocyte-derived
macrophages, from the human monocytic THP1 cell line, and from J774
cells transfected to constitutively express a human apo E cDNA. LPL
similarly reduced apo E secretion when it was prebound to the
macrophage cell surface at 4 °C. A native LPL particle was required
to modulate apo E secretion; addition of monomers and aggregates did
not produce the same effect. Depletion of cell surface proteoglycans by
a 72-h incubation in 4-methylumbelliferyl--D-xyloside
did not attenuate the ability of LPL to reduce apo E secretion.
However, addition of receptor-associated protein attenuated the effect
of LPL on apo E secretion. Although LPL could mediate removal of
exogenously added apo E from the culture medium, detailed pulse-chase
analysis suggested that it primarily prevented release of newly
synthesized apo E from the cell layer. Cholesterol loading of cells or
antibodies to the low density lipoprotein receptor attenuated LPL
effects on apo E secretion. We postulate that LPL sequesters
endogenously synthesized apo E at the cell surface by a low density
lipoprotein receptor-dependent mechanism. Such
post-translational regulation of macrophage apo E secretion by LPL
could significantly influence apo E accumulation in arterial vessel
wall lesions.
Macrophages synthesize and secrete both apo E1 and LPL (1, 2). These cells, therefore, can be a significant source of each of these proteins in the arterial wall. Both proteins have important roles in systemic lipoprotein metabolism (3, 4) as well as in lipoprotein metabolism at the level of the individual cell (5-14). Regional accumulation of each of these proteins has been detected at sites of atherosclerotic lesion development (15); apo E has been found predominantly on the surface of macrophages and in the matrix surrounding macrophages in atherosclerotic lesions, whereas LPL is also associated with arterial smooth muscle cells. apo E expression in the vessel wall has been shown to be important for modulating vessel wall cholesterol homeostasis (16, 17). In fact, macrophage-specific expression of apo E in apo E-null mice has been shown to protect against atherosclerotic lesion development even in the presence of high levels of circulating atherogenic lipoproteins (17). As noted above, LPL also accumulates in diseased vessel wall. A number of cell surface binding sites for LPL have been identified (11, 18, 19), including the LRP, plasma membrane-associated proteoglycans, and a 116-kDa nonproteoglycan binding site on endothelial cells. LPL accumulates in the extracellular matrix and has been shown to directly bind apoprotein B (6). LPL bound to cell surface or extracellular matrix sites, therefore, has been implicated in enhancing cellular uptake of lipoproteins, including LDL and triglyceride-rich lipoproteins (4).
apo E also binds to proteoglycans and to the LRP (13, 14). In a recent report, it was found that preincubation of endothelium-derived matrix with apoprotein E did not reduce the binding of subsequently added LPL to this matrix (10). This observation suggested that apoprotein E and LPL may bind to distinct proteoglycan species in the subendothelial matrix. In addition, apo E reduced LPL-mediated retention of LDL in the subendothelial matrix in a dose-dependent manner.
Expression of macrophage apo E is modulated transcriptionally by cholesterol and by cytokines (20, 21). There is also significant post-transcriptional regulation of macrophage apo E secretion, and we have recently shown that one important site for post-translational regulation of macrophage apo E secretion is at the pericellular proteoglycan matrix (22). A substantial portion of the apo E synthesized by the macrophage is retained in the pericellular proteoglycan matrix and is rapidly returned to the cell for degradation (22). In view of this observation and previous reports regarding the interaction between pericellular proteoglycans and LPL, we formally evaluated a role for LPL in modulating the secretion/metabolism of endogenously produced apo E in the macrophage. For these studies, we utilized a macrophage cell model in which apo E is constitutively synthesized and evaluated the effect of exogenously added LPL.
4-Methylumbelliferyl--D-xyloside
was obtained from Sigma. Lipoprotein lipase was isolated from bovine
milk as described previously (23). Human recombinant RAP was prepared
as described (24). Acetylated LDL was prepared by previously described
methods (20). Human apo E isolated from plasma was purchased from Alpha
Biomedical (Bellevue, WA) and iodinated using Iodobeads (Pierce)
according to the manufacturer's instructions. Unbound iodine was
removed by chromatography on a PD-10 column followed by dialysis.
Polyclonal rabbit anti-LDLR antiserum was prepared against LDLR protein
isolated from bovine adrenal glands as described previously (25).
Monoclonal antibody IgGC7 was prepared from mouse ascites by standard
techniques. This antibody has been previously characterized and is
directed against the first cysteine-rich repeat of the LDLR. All other materials were from previously described sources (5, 22, 26).
J774 cells were stably transfected to express a wild type human apo E cDNA as described previously in detail (26). This cell line constitutively expresses a human apo E cDNA (E 3 isoform) and secretes 900 ng of apo E/mg of cell protein in 24 h, an amount similar to that produced by mature cholesterol-loaded human monocyte-derived macrophages in culture. Cells were maintained in 400 µg/ml selection agent G418 (Geneticin, Sigma) until 1 week prior to the initiation of experiments. Cells were grown in 10% fetal bovine serum in Dulbecco's modified Eagle's medium until the start of the described experimental incubations. apo E-expressing J774 cells were used for all experiments unless indicated otherwise. For studies using J774 cells not secreting apo E, the same parental J774 cell line transfected only with a neomycin-resistance construct was used. These cells were maintained in G418 exactly as for apo E-secreting cells. Human monocytes were purified by elutriation. Cells were >90% monocytic (as determined by differential counts of Wright-stained smears) and grown as described previously (21). The THP1 human monocytic cell line was obtained from the American Type Culture Collection and grown as described previously (21).
Isolation and Characterization of Monomeric/Aggregated LPLBovine LPL (1 mg), isolated as described previously (23), was
diluted in 300 ml of 0.15 M NaCl, 1 mM EDTA, 10 mM sodium phosphate buffer (pH 7.5) containing 0.1 mg/ml
ovalbumin and incubated at 37 °C for 30 min in this dilute solution.
After the incubation, the sample was chilled on ice and passed through
a 1-ml heparin-agarose column (HiTrap Heparin, Pharmacia Biotech Inc.)
using a peristaltic pump at flow rate of 2.5 ml/min. The column,
submerged in an ice bath and attached to a Pharmacia fast protein
liquid chromatography system, was washed with 20 ml of 0.15 M NaCl, 1 mM EDTA, 10 mM sodium
phosphate buffer (pH 7.5) at a flow rate of 1 ml/min and then eluted
with a 15-ml gradient of 0.15-1.5 M NaCl in the same buffer at a flow rate of 0.25 ml/min while 0.25-ml fractions were collected and immediately chilled on ice. The effluent was monitored by
absorbance at 280 nm. As described previously for human LPL (27), the
protein emerged in a low affinity inactive and a high affinity active
peak with material of heterogeneous heparin affinity in between.
Fractions containing the inactive monomeric and aggregated material
preceding the high affinity active peak were pooled, dialyzed against
50% (v/v) glycerol, 10 mM sodium phosphate (pH 7.5), and
stored at 20 °C. The yield of inactive monomeric and aggregated
enzyme (0.36 mg) was determined by absorbance at 280 nm.
2-3 × 106 cells were plated into 35-mm wells and grown for 48-72 h. All media used during the following procedures were warmed to 37 °C before use. Pulse medium contained 100 µCi/ml [35S]methionine with 1-2 µM unlabeled methionine added to methionine-free Dulbecco's modified Eagle's medium. Pulse and chase times were as indicated in the figure legends. Chase medium contained 500 µM unlabeled methionine. At the end of the chase period, apo E secreted into the medium and apo E that remained associated with the cell was quantitatively immunoprecipitated and analyzed by SDS-polyacrylamide gel electrophoresis as described previously in detail (26). All immunoprecipitations from cell lysates and media were performed starting with equal numbers of total trichloroacetic acid-precipitable counts. Therefore, the disintegrations per minute in secreted or cell-associated apo E are normalized for total labeled secreted protein or total labeled cell-associated protein, respectively, in each experiment. For most experiments, the apo E bands were excised from the gel after they were localized by autoradiography. The dried gel piece from each sample was cut out and, after the gel was rehydrated with 3% glycerol, the backing paper and glycerol solution were removed and the gel slices were digested in 0.5 ml of 30% hydrogen peroxide. After 48-72 h at 60 °C, scintillant was added and the samples were counted against an external standard so that counting efficiency could be used to calculate the disintegrations per minute incorporated into apo E for each sample. For some experiments, apo E radioactivity on SDS-polyacrylamide gel electrophoresis gels was quantitated using the Storm PhosphorImager (Molecular Dynamics), and results are expressed as -fold change over control. For our analyses, "secreted" apo E is considered to be that recovered from the medium and "cell" apo E is considered to be that recovered from lysis of the cell layer.
The effect of LPL on the net medium accumulation of apo E from
J774 macrophages transfected to constitutively express a human apo E
cDNA, as a function of LPL concentration, is shown in Fig. 1. Cells were incubated for 45 min with labeled
methionine and chased for an additional 45 min with vehicle or LPL at
the indicated concentrations. Addition of LPL at 10 µg/ml led to a
39% reduction of apo E content, and 20 µg/ml produced a 49%
reduction. The LPL effect was maximal at 20 µg/ml, and higher doses
of LPL, up to 80 µg/ml, had no additional effect on apo E levels. A
final LPL concentration of 40 µg/ml was used for all subsequent
experiments.
Similar results were obtained in experiments using human monocyte-derived macrophages. Elutriated human monocytes (>90% pure) were allowed to differentiate to a macrophage phenotype for 4 days in culture before being placed in 0.2% bovine serum albumin for an overnight incubation. Cultures were then pulse-labeled with methionine for 45 min followed by a 45-min chase at 37 °C with vehicle or LPL (40 µg/ml). apo E immunoprecipitated from the medium (quantitated by PhosphorImager) showed a 43% reduction in the presence of LPL (10.6 ± 0.6 versus 6.1 ± 0.4 phosphorimaging units in control versus LPL-treated, respectively).
Because LPL can bind to several cell surface sites, the observed effect
on apo E when LPL was present in the medium during the 37 °C chase
could represent the effect of LPL bound at the cell surface or LPL in
the culture medium. The next experiment was performed to address this
issue. For the 4 °C bars in Fig. 2, LPL
was prebound to cells at 4 °C for 60 min immediately following a
45-min pulse labeling period. Following this 4 °C incubation, cells
were extensively washed and incubated for an additional 45 min at
37 °C with no added LPL. For the 37 °C bars, LPL was added only during a 45-min chase at 37 °C. apo E was then
immunoprecipitated from chase medium for each experimental incubation.
As shown, apo E medium content was reduced to similar levels by LPL
whether it was present in the chase medium or prebound to the cell
surface at 4 °C.
Both apo E and LPL bind to cell surface proteoglycans (14, 18, 22). The
results above suggest that these ligands do not compete for the same
saturable proteoglycan binding sites, because LPL does not displace apo
E from the cell layer. LPL binding of the newly synthesized apo E
particle could, therefore, increase the binding capacity for apo E in
the pericellular proteoglycan matrix. We investigated the potential
involvement of cellular proteoglycans by depleting cell surface
proteoglycans using a 72-h preincubation in DX (Table
I).
DX substitutes for the core protein moiety of
proteoglycans during their synthesis and substantially reduces their
appearance at the cell surface (22). We have previously shown (22) that
cell surface proteoglycans retain a large portion of newly synthesized
apo E in the pericellular matrix so that preincubation in
DX alone
led to a large increase in the release of apo E into the medium
(26,159 ± 162 versus 12,512 ± 112 dpm). In cells
not preincubated in
DX, LPL addition reduced medium apo E content as
previously observed. In proteoglycan-depleted cells, LPL retained its
ability to reduce apo E levels (from 26,159 ± 162 to 6,164 ± 79 dpm). This result indicated that proteoglycan binding sites are
not necessary for LPL modulation of apo E expression.
|
LPL can bind to additional cell surface nonproteoglycan sites (11, 18,
19, 28), and we utilized RAP to gain further insight into important
cell surface sites. This protein has been shown to bind to multiple
members of the LDL receptor family including the LRP, LDLR, VLDL
receptor, and gp 330/megalin and to inhibit their ligand binding
properties. As shown in Table II, addition of RAP
enhanced apo E release from the cell layer into the medium. Furthermore, RAP significantly interfered with LPL-mediated
sequestration of apo E. Contrary to what was observed in the experiment
using preincubations in DX, in which LPL addition reduced apo E in the medium to similar levels whether cells were depleted of
proteoglycans or not, RAP at a 2-fold molar excess (relative to LPL)
attenuated LPL-mediated sequestration of apo E compared with addition
of LPL alone. This result, in conjunction with the results shown in
Table I, suggested that a nonproteoglycan, RAP-sensitive binding site
was involved in LPL sequestration of apo E.
|
In Fig. 3, we investigated the requirement for the
addition of native LPL for its effect on medium apo E accumulation.
Prior to addition to cells, inactive monomeric and aggregated LPL was prepared as described under "Experimental Procedures"; native LPL
was excluded from the cultures labeled "aggregrated/monomeric" LPL.
Addition of native LPL produced the expected result on medium apo E
content. Monomeric/aggregated LPL, however, not only failed to decrease
medium apo E level but actually increased it by 40%.
The reduction we observed in medium apo E content in the presence of LPL could be related to LPL-mediated reduction of apo E release from cells or LPL-mediated enhancement of apo E reuptake from the medium. Therefore, we next evaluated the effect of LPL on the removal of exogenously added apo E (Table III). Delipidated apo E, isolated from human plasma lipoproteins, was iodinated and added to apo E-secreting and nonsecreting J774 cells with or without LPL. After 120 min at 4 °C, labeled apo E remaining in the medium was recovered by trichloroacetic acid precipitation. Total apo E radioactivity added to the cell cultures equaled 53,360 cpm. In the absence of LPL, apo E-secreting J774 macrophages sequestered very little exogenous apo E (52,839 ± 734 dpm remaining), whereas nonsecreting cells may have sequestered a small amount (44,223 ± 6,917 remaining). Addition of LPL stimulated removal of apo E only from apo E-secreting cells (23,489 ± 2,428 remaining). This observation suggested that synthesis of endogenous apo E was necessary for LPL-mediated uptake of exogenously added apo E. This could implicate a cell surface pool of apo E or the lipid component of the endogenously synthesized apo E particle as important for LPL-mediated uptake (see "Discussion"). These results indicated that LPL could reduce apo E in the medium by enhancing its reuptake.
|
To further investigate this issue, we studied cells after a 20-min
pulse with labeled methionine followed by multiple chase times. For the
experiment shown in Fig. 4, LPL was added at time 0 (immediately following a 20-min pulse period). By 15 min, apo E in the
medium from LPL-treated cells was already lower than that in control
medium (6,288 ± 667 versus 9,588 ± 2,432 dpm). The difference in medium apo E between LPL-treated and control cultures
also continues to increase through the 45-min chase period. Fig.
4B shows the results of immunoprecipitating apo E from the cell layer. apo E in the cell layer tends to be higher in LPL-treated cells; however, the magnitude of the difference between LPL-treated and
control cultures appears to change very little after 15 min. The above
results suggested that LPL prevents release of apo E from the cell
layer and that the apo E retained by the cells is subject to
degradation.
To gain further insight into the physiologic parameters modulating LPL-mediated sequestration of macrophage-derived apo E, we evaluated the effect of cellular cholesterol loading on this process using the transfected J774 cell model. Cholesterol loading was accomplished by a 48-h preincubation in acetylated LDL. This preincubation was followed by an 18-h equilibration period in serum-free medium to allow for the internalization of remaining acetylated LDL and equilibration of cell cholesterol. Cells preincubated with acetylated LDL contained 56.1 and 12.8 µg/mg free and esterified cholesterol, respectively (average of duplicate determinations), compared with 23.3 µg/mg free cholesterol and no detectable cholesterol ester in cells grown without acetylated LDL. As shown in Table IV, cholesterol loading of J774 cells completely abolished the effect of LPL on apo E secretion.
|
We next investigated the potential involvement of specific cell surface binding sites for the LPL effect. The results shown in Table I indicate that proteoglycans on the cell surface are not likely involved. The results in Table II suggest the potential involvement of a member of the LDL receptor gene family (e.g. LRP, VLDL receptor, or LDLR). Of these, only the LDLR would be expected to behave in a sterol-suppressible fashion, as is demonstrated in Table IV. Therefore, we focused on this receptor as important in mediating LPL effects in our experiments. As shown in Tables V and VI, polyclonal or monoclonal antibodies to the LDLR abrogated LPL-mediated reduction of apo E secretion by THP1 cells and transfected J774 cells.
|
|
The data presented indicate that native LPL reduces the secretion of newly synthesized apo E in the macrophage in an LDLR-dependent manner. We postulate as an explanation for our findings that native LPL can directly interact with the endogenously synthesized apo E-lipid particle and act as a bridge to sequester it at the cell surface. apo E that is not secreted in the presence of LPL is thereby returned to the cell for degradation. LPL and apo E may bind to unique cell surface proteoglycan species, since LPL does not displace apo E from the cell layer. Binding to different proteoglycan species would allow for alternative fates in the extracellular matrix for these two proteins. In addition, binding of LPL and apo E to unique proteoglycan species is consistent with previous observations that apo E and LPL act in a complementary fashion to enhance binding of proteoglycans to triglyceride-rich lipoproteins (29). This would not likely occur if LPL and apo E were competing for binding to the same proteoglycan sites.
The data shown in Fig. 3 indicate that native LPL is required for LPL-mediated reduction of macrophage apo E secretion. Addition of LPL monomers and aggregates prepared by column chromatography actually increased macrophage apo E secretion. We believe that the altered properties of aggregated/monomeric LPL in our experiments are related primarily to loss of critical secondary structure for a bridging function and not to loss of enzyme catalytic activity. Our experiments conducted with native LPL were performed in serum-free medium without a source of apolipoprotein CII. The latter is a critical cofactor for LPL activity that can be supplied by serum or triglyceride-rich lipoproteins. Because of the absence of apolipoprotein CII in our experimental system, there was likely little LPL enzymatic activity in any of our experiments. Additional insight into the contribution of enzymatic activity will require extensive analysis of LPL structural mutants. Using such an analysis, the bridge and catalytic functions for LPL-induced metabolism of VLDL have been separated (30). In these studies it was determined that although catalytic function was not important for the "bridging" function of LPL for VLDL, intact folding of the catalytic loop must be maintained for such function. Studies using inhibitors of LPL activity that produce covalent modification of the enzyme active site may also produce changes in critical secondary structure for bridging and therefore cannot substitute for mutational analysis.
The data shown in Table III suggest that the lipid portion of the apo E-lipid particle may, in fact, be responsible for LPL-mediated effects on apo E secretion in the macrophage. The uptake of exogenously added delipidated apo E to nonexpressing J774 cells was not enhanced by the addition of LPL. However, LPL did enhance the uptake of exogenous apo E by apo E-secreting J774 cells. As an explanation for this observation, we speculate that the exogenously added apo E is able to rapidly exchange onto an apo E-lipid particle (produced by the apo E-secreting J774 cells) and thereby be removed by LPL binding to the lipid portion of this particle. Uptake of exogenously added apo E cannot proceed in nonsecreting J774 cells because these cells do not produce the apo E-lipid particles required for this exchange. Thus, our data predict that an intact lipid binding domain of native LPL is required for its reduction of apo E secretion by the macrophage. This domain may be irreversibly altered during preparation of aggregrated/monomeric LPL.
Reduction of medium apo E by LPL could be related to reduced release of apo E from the cell layer, enhanced reuptake of apo E after secretion, or both. LPL can mediate the removal of exogenously added apo E (Table III). The turnover study presented in Fig. 4, however, does not appear to support the enhanced reuptake mechanism as an explanation for the LPL effect on endogenous apo E expression. apo E levels are already reduced as early as 15 min after the addition of LPL and do not change substantially over the next 30 min, whereas apo E levels in control cultures continue to rise. These results appear to be more consistent with a reduction of apo E release from the cell layer by LPL. This type of result, however, cannot exclude an enhanced reuptake mechanism that functions in the immediate pericellular space and rapidly comes to equilibrium with apo E secretion rate after the addition of LPL. A more complete understanding of these issues will require more detailed characterization of the apo E-lipid particle produced by the macrophage and extensive evaluation of its potential interactions at the macrophage cell surface.
The data shown in Tables IV, V, VI indicate that the LDLR is involved in the LPL-mediated reduction of apo E secretion. Recently, Medh et al. (31) have shown that LPL binds to the LDLR. The binding and degradation of VLDL induced by LPL was found to significantly depend on LDL receptor expression, and the LPL-dependent contribution of the LRP to VLDL catabolism was small when LDL receptors were up-regulated. Furthermore, in these studies, the LDLR interaction appeared to be associated with the carboxyl-terminal noncatalytic domain of LPL.
In conclusion, increased amounts of pericellular LPL will post-translationally suppress the net secretion, and therefore the local accumulation, of apo E at any tissue site where macrophages are the major source of apo E, e.g. the atherosclerotic vessel wall lesion (15, 32). Because macrophage expression of apo E has been shown to be anti-atherogenic (16, 17), increased LPL could have deleterious effects on the cholesterol homeostasis of the vessel wall. This is consistent with observations in intact mice indicating that high macrophage lipoprotein lipase expression and secretion are associated with enhanced susceptibility to atherosclerosis (33). It is also of interest that the regulation of macrophage apo E and LPL expression are different. For example, it has been recently demonstrated that oxidant stress (e.g. by addition of hydrogen peroxide) enhances macrophage LPL mRNA levels and LPL production (34). We have been unable to detect any effect of hydrogen peroxide on macrophage apo E mRNA levels.2 Platelet-derived growth factor appears to have a major role in modulating the expression of macrophage LPL (35), but such a regulatory effect for apo E has not been demonstrated.
Cholesterol loading of macrophages enhances macrophage apo E synthesis and secretion predominantly by stimulating apo E gene transcription (20). The data in this report indicate that cholesterol loading will also suppress the LPL-mediated degradation of endogenously synthesized apo E in the macrophage. Cholesterol loading will, therefore, function synergistically at transcriptional and post-translational loci to enhance the net secretion of apo E by macrophage foam cells in the arterial wall.
The authors thank Beverly Burge for typing the manuscript.