Degradation of Macrophage ApoE in a Nonlysosomal Compartment

Macrophage-derived apoE has been shown to play an important role in the susceptibility of the vessel wall to atherosclerosis. Previous studies have shown that macrophage sterol content modulates apoE synthesis and secretion, associated with a large transcriptional response of the apoE gene. The current studies were undertaken to evaluate the existence of additional post-transcriptional regulatory loci for the effect of sterols on apoE synthesis and secretion. Using a macrophage cell line transfected to constitutively express an apoE cDNA to facilitate detection of a post-transcriptional regulatory locus, we demonstrated that preincubations in 25-hydroxycholesterol and cholesterol lead to increased apoE secretion in pulse/chase experiments. Examination of cell lysates in these experiments showed that apoE not secreted by control cells was degraded and not detectable, suggesting that the preincubation in sterols increased secretion by decreasing degradation of newly synthesized apoE. The measurement of total protein and apoE degradation in cell fractions revealed an intermediate density fraction that degraded significant amounts of newly synthesized total protein and newly synthesized apoE. In this fraction, degradation of total protein and apoE was unaffected by chloroquine but was substantially reduced by N-acetyl-Leu-Leu-norleucinal plusN-acetyl-Leu-Leu-methioninal or by lactacystin, suggesting the involvement of proteasomes. Preincubation in sterol/oxysterol or acetylated low density lipoprotein did not modify total protein degradation by this fraction but inhibited apoE degradation. Similar results were obtained using intermediate density fractions isolated from human monocyte-derived macrophages. The results of our studies indicate that newly synthesized apoE in the macrophage can be degraded in an intermediate density nonlysosomal cellular compartment, which is sensitive to proteasomal inhibitors. Alteration of cellular lipid homeostasis by preincubation in sterol/oxysterol or acetylated low density lipoprotein inhibits apoE, but not total protein, degradation in this fraction. Inhibition of the degradation of apoE in this fraction likely contributes to the increased apoE secretion observed in sterol-enriched cells.

serves as a ligand for several cell membrane endocytic receptors found in liver and peripheral cells. In addition, apoE is a highly charged protein and has strong affinity for negatively charged proteoglycans, including those found on the cell surface. Consistent with this, apoE has been detected in a cell surface pool in hepatic cells and macrophages, where it may be involved in modulating cellular lipid and lipoprotein metabolism (3)(4)(5)(6)(7).
The function of apoE synthesized in multiple extrahepatic cell types remains incompletely defined. In macrophages, the regulation of apoE gene expression and apoE secretion by small increments in cellular cholesterol mass led to the hypothesis that apoE could serve a role in macrophage cholesterol homeostasis, as part of a regulatory loop to reduce cellular free cholesterol content (8). Subsequent studies confirmed such a role and indicated that expression of apoE in macrophages functions to enhance cholesterol efflux and reduce macrophage cholesterol stores in the presence and absence of extracellular acceptor particles (9,10). Studies using genetically engineered mice and bone marrow transplantation technology have confirmed an important role for macrophage-derived apoE in modulating susceptibility to atherosclerosis (11)(12)(13)(14). Mice with a specific defect in macrophage apoE expression displayed a 10fold increase in susceptibility to atherosclerosis while on a high fat diet (14). Conversely, macrophage-specific expression of apoE has been shown to protect mice from the effect of atherogenic hyperlipidemia (13). There are multiple properties of apoE which can be considered in evaluating the mechanism of such a protective effect (1,2). However, modulation of macrophage cholesterol content by endogenously synthesized apoE probably contributes to this important role in determining susceptibility of the vessel wall to atherogenic insult. In light of the apparent importance of macrophage-derived apoE in vessel wall homeostasis, it is imperative to fully understand the regulation of apoE expression in macrophages.
Sterol modulation of apoE synthesis and secretion has been demonstrated in multiple macrophage types, including human monocyte-derived macrophages and mouse peritoneal macrophages (8,(15)(16)(17)(18). Increased apoE expression has been demonstrated after cholesterol loading of cells with modified lipoproteins or exposure of cells to cholesterol and oxysterol (8,(15)(16)(17)(18). Previous studies have indicated that sterol modulation of apoE synthesis and secretion is associated with a substantial increase in apoE gene transcription, with reports showing up to a 10-fold increase in apoE mRNA levels with a similar increase in apoE gene transcription rates (15). In the current studies, we wished to address the issue of a potential post-transcriptional site for modulation of apoE secretion by sterols. A post-translational site was of particular interest based on previous observations that a significant portion of newly synthesized apoE in the macrophage is degraded prior to its release from the cell (19). To approach this issue, we reasoned that a post-transcriptional and post-translational effect would be more easily de-monstrable in cells in which the large transcriptional response of the apoE gene to sterols could not obscure other potential regulatory loci. We therefore utilized a J774 cell macrophage model, which was transfected to constitutively express a human apoE cDNA. Expression of the apoE cDNA in this cell line is mediated by the human metallothionine IIa promoter and has been previously characterized in detail (3,10,19,20). These cells secrete 0.9 -3.5 g/ml/mg of cell protein of apoE over 24 h depending on culture conditions, an amount similar to that reported for human monocyte-derived macrophages.
Cell Culture-J774 cells stably transfected to constitutively express a wild type human apoE cDNA have been characterized previously in detail (3,10,19,20). Cells were selected and maintained in G418 (400 g/dl) until 1 week prior to the initiation of experiments. Cells were grown in 10% fetal bovine serum and Dulbecco's modified Eagle's medium until the start of the described experimental incubations. Freshly isolated human monocytes were purified by elutriation. The elutriated cell population used for experiments was Ͼ95% monocytic, as determined by differential counts of Wright-stained smears. Human monocyte macrophages were plated and grown as described previously (21). Cells were incubated with sterols dissolved in ethanol vehicle (final concentration Ͻ 0.5%). Control cultures were incubated in equivalent amounts of vehicle only.
Subcellular Fractionation-All steps were carried out at 4°C. Cultures to be fractionated on continuous sucrose gradients were washed two times and scraped from the culture flasks with a rubber policeman in 5 mM sodium phosphate (pH 7.5), 0.1 M sucrose and incubated on ice for 20 min to swell the cells. Cell suspensions were homogenized using 100 strokes in a Dounce homogenizer, and the sucrose concentration of the cell homogenate was adjusted to 20%. A linear sucrose gradient was prepared by mixing 6 ml of 20% sucrose (mixed with the cell homogenate) and 5.5 ml of 56% sucrose and centrifuging at 35,000 rpm for 18 h in a Beckman SW41 rotor, as described previously (22). Eleven fractions of equal volume were collected from the bottom of the tube, and the density of the fractions was monitored by comparing the refractive index of each fraction with a standard curve of density versus refractive index. Fractions were assayed immediately for marker enzyme activity. Recovery of enzyme activity and total labeled protein approximated 70% on these gradients.
Isolation of intermediate density, nonlysosomal cell fractions was based on previously published methods (23,24) with minor modifications. Cultures to be fractionated were washed two times and scraped from the culture flasks with a rubber policeman in homogenizing medium (37.5 mM Tris-HCl, 0.5 M sucrose, 1% dextran, 5 mM MgC1 2 , 0.1 mM phenylmethylsulfonyl fluoride, 2 g/ml leupeptin, pH 6.5). Cells in suspension were homogenized using 100 strokes in a Dounce homogenizer. The homogenate was then centrifuged at 2,000 rpm in a Beckman SW41 rotor for 5 min, followed by 10,000 rpm for 30 min. The upper one-half of the pellet was resuspended in 2 ml of homogenizing medium and then layered onto 9 ml of ice-cold 1.2 M sucrose and centrifuged at 25,000 rpm for another 30 min in a SW41 rotor. Two closely spaced opaque bands at the interface were recovered with a Pasteur pipette and diluted 20-fold with 1 ϫ phosphate-buffered saline. Lysosomes were removed by centrifuging at 5,000 rpm for 20 min in a SW41 rotor and recovering the pellets. Pelleted membranes were assayed immediately for marker enzyme activity.
Quantitation of ApoE Secretion, Distribution in Subcellular Fractions, and Gene Expression-Immunoprecipitations and Western and Northern blot hybridizations were performed as described previously in detail (3, 10, 15, 19 -21). ApoE was biosynthetically labeled with [ 35 S]methonine (100 -200 Ci/ml), and pulse-chase incubations were carried out as described previously. ApoE bands in SDS-polyacrylamide gel electrophoresis of immunoprecipitates were localized using a radiofluorescent image scanner and quantitated using ImageQuant software. Northern and Western blot signals were measured using a Molecular Dynamics laser densitometer with ImageQuant software.

Measurement of ApoE and Total Protein Degradation in Subcellular
Fractions-The method used to measure the degradation of total proteins or apoE in subcellular fractions was based on previously published methods (25,26) with minor modifications. After isolation of fractions at 4°C, equal aliquots were mixed with half volumes of 0.4 M sodium potassium phosphate (pH 6.8) containing 8 mM dithiothreitol. Equal portions of each fraction were then incubated for an additional 3 h at 40°C or kept at 4°C. At the end of that time, total labeled protein was measured by trichloroacetic acid precipitation, and apoE was measured by Western blot hybridization or immunoprecipitation. The difference between the amount of protein present in a fraction after incubation at 40 versus 4°C was taken as the amount degraded during the 3-h incubation period. ApoE in immunoprecipitates and Western blots was quantitated, as detailed above, using ImageQuant software. The amount of apoE present at the end of the 4°C incubation was assigned a value of 100%, and the amount at the end of the 40°C incubation was expressed as a fraction of the 4°C value. In our experiments, the percentage of apoE degraded ranged from 68 to 73% in control cells and from 27 to 43% in sterol-treated cells.

Sterols Modulate ApoE Secretion at a Post-transcriptional
and Post-translational Locus-In this series of studies, we used a preincubation protocol that included sterol and oxysterol. We first wished to evaluate apoE mRNA levels in response to this experimental preincubation to confirm the absence of a transcriptional response and eliminate the possibility that posttranscriptional stabilization of apoE mRNA levels occurred in response to the sterol/oxysterol preincubation. The results of a Northern blot hybridization, shown in Fig. 1, indicate that 24-h FIG. 1. ApoE mRNA levels in control and sterol-treated cells. Cells were plated, grown, and incubated as described under "Experimental Procedures." After 24 h in 1 g/ml of 25-hydroxycholesterol (25-OH CH) and 10 g/ml of cholesterol, 20 g of total cellular RNA was utilized for Northern blot hybridization. Northern blots were probed for apoE and ␤-actin mRNA species. The level of mRNA was quantitated as described under "Experimental Procedures," and apoE/␤-actin ratios were calculated. The values shown are the mean Ϯ S.D. from triplicate measurements.
preincubations in sterol/oxysterol produced no change in apoE mRNA levels. We next conducted a series of pulse-chase experiments after a 24-h preincubation in sterol/oxysterol. The results of a representative experiment are shown in Fig. 2. Cells preincubated in sterol/oxysterol show a 1.7-fold increase and a 1.5-fold increase in medium apoE after 30 and 60 min of chase, respectively. Comparable increases were observed in multiple experiments conducted in a similar manner. Examination of the cellular content of apoE at 0, 30, and 60 min of chase showed no difference. These results indicate that the preincubation in sterol/oxysterol could enhance apoE secretion at a translational or post-translational locus. The apoE that was not secreted in control cells did not appear to be retained within the cell and, therefore, was probably degraded. Given previous observations regarding the significant rate of degradation of newly synthesized apoE (19), these results were consistent with the hypothesis that the sterol/oxysterol preincubation modified the degradation rate for newly synthesized apoE and thereby enhanced its secretion. We therefore directly investigated the effect of the sterol/oxysterol incubation on the degradation rate of apoE.

Sterol/Oxysterol Preincubation Does Not Modulate ApoE Distribution in Subcellular Fractions or Alter the Compartment
Responsible for ApoE Degradation-In addressing the potential for modification of apoE degradation by the sterol/oxysterol preincubation, we first wished to examine the subcellular distribution of apoE in control cells and cells preincubated in sterol/oxysterol and to determine which subcellular fractions were primarily responsible for degrading newly synthesized apoE. To study the subcellular distribution of newly synthesized apoE and the distribution of the activity for degrading it, macrophage cell homogenates were fractionated on 23-42% continuous sucrose gradients. On these gradients, the plasma membrane was contained in the three most buoyant fractions, the ER marker was found in the three densest fractions, and the lysosome and Golgi markers were found in five fractions of intermediate density and were incompletely resolved. A predominantly nonglycosylated isoform of apoE was recovered from the dense fractions containing ER, whereas larger glycosylated forms were easily detectable in Golgi/lysosome and plasma membrane fractions. Multiple experiments were performed utilizing cells harvested after a 30-min pulse label and no chase or after a 30-min pulse and 30-min chase period. We did not detect a consistently significant difference between the distribution of apoE in control cells and cells preincubated in sterol/oxysterol (not shown). We next assayed each of the fractions for its ability to degrade newly synthesized apoE. We found that 60 -70% of total apoE degradation occurred in intermediate density fractions in both control and sterol/oxysterol-treated cells (not shown). We therefore undertook a more focused study of apoE degradation and the effect of sterols in an intermediate density cell fraction.
Degradation of ApoE by a Nonlysosomal Pathway in Intermediate Density Fractions-An increasingly important role has been ascribed to nonlysosomal degradative pathways in the turnover of newly synthesized protein. We therefore undertook a set of studies to evaluate the role of a nonlysosomal intermediate density compartment in the macrophage for the degrada- tion of newly synthesized apoE and evaluate whether apoE degradation in this compartment could be modulated by the sterol/oxysterol preincubation. This intermediate density fraction, isolated by previously published techniques (23,24), had the characteristics shown in Table I. The fraction was enriched in the Golgi marker with an 8-fold increase in its specific activity and a 30% recovery. The specific activities for the ER, lysosome, and plasma membrane markers were essentially unchanged. Significantly, only 5.3% of the lysosomal marker activity was recovered in this fraction.
This intermediate density fraction degraded almost twothirds of newly synthesized total cellular protein, suggesting the presence of a major cellular degradative pathway. Major cellular pathways for degrading proteins are believed to primarily involve lysosomes or proteasomes (27)(28)(29). Our fractionation technique significantly excluded lysosomes; however, we wished to exclude the possibility that the small lysosomal contamination could still account for a significant portion of the degradation observed in our intermediate density fraction. We utilized chloroquine, a lysosomotropic protease inhibitor, and compared its ability to inhibit total protein and apoE degradation in this fraction compared with ALLN and ALLM. As shown in Table II, chloroquine at 100 M essentially was without effect on the degradation of total proteins or apoE in this fraction, confirming the lack of involvement of lysosomes. The inclusion of ALLN and ALLM, however, did significantly modulate total protein and apoE degradation. These latter agents are cysteine protease inhibitors; however, they have broad inhibitory activity and have been shown to also inhibit proteasomal hydrolases. We next examined the effect of a more specific proteasomal inhibitor on the degradation of total protein and apoE in this fraction. The results in Table III show that lactacystin, a specific inhibitor for proteasomal degradation (30,31), significantly interfered with total protein and apoE degradation.
Sterols Modulate ApoE Degradation in Intermediate Density Fractions-We next evaluated the effect of the sterol/oxysterol incubation on the degradation of total proteins and apoE in this intermediate density fraction. For the experiment shown in Fig. 3, degradation of total labeled proteins in the intermediate density fraction was high (63.6 Ϯ 4.2% degraded) and was not altered by the sterol/oxysterol preincubation (60.8 Ϯ 9.5% degraded). ApoE was also substantially degraded in this fraction (30% remaining in the 40°C incubation compared with the 4°C incubation, 70% degraded) from control cells (Fig. 3). However, the rate of apoE degradation in this fraction was significantly reduced by the sterol/oxysterol preincubation (73% remaining in the 40 versus 4°C incubation, 26.7% degraded). In Fig. 4, the results of an identical experiment conducted in human monocyte-derived macrophages is shown, with a similar result.
Sterol/oxysterol modulation of apoE degradation could reflect a direct regulatory action of these compounds on constituents of the degrading mechanism or could reflect changes in the cell (e.g. signaling pathways or membrane composition) engendered during the 24-h preincubation period. To evaluate the potential of a direct regulatory activity, we measured the effect of sterol/oxysterol added only during the 3-h incubation period at 4 or 40°C. A direct addition, without prior preincubation, did not modulate degradation rates of apoE (not shown), thereby suggesting that a secondary change in the cell produced during the preincubation was responsible for modulating apoE degradation. A likely effect of the 24-h sterol/ oxysterol preincubation is alteration of the lipid composition of cellular membranes. The data shown in Table IV confirm that 24 h in sterol/oxysterol altered the lipid composition of the intermediate density nonlysosomal fraction that degraded apoE with an increase in free cholesterol and phospholipid mass. Free cholesterol/phospholipid molar ratios remained unchanged.
To further evaluate the possibility that altered lipid composition could be related to modulation of apoE degradation, we tested the effect of another preincubation protocol known to modulate cellular membrane composition in a similar manner (32), preincubation in Ac-LDL. As shown in Fig. 5, preincubation in Ac-LDL also significantly suppressed degradation rates of apoE in the intermediate density nonlysosomal fraction.

DISCUSSION
The current studies indicate the existence of an important post-transcriptional locus for regulation of macrophage apoE secretion by sterols. While there may be other sites that could be involved in the degradation of newly synthesized apoE, our studies specifically establish the importance of an intermediate density nonlysosomal cellular compartment in which apoE degradation is modulated by sterols. Previous reports have shown that a substantial portion (over 50%) of newly synthesized apoE is degraded in macrophages prior to its secretion (19). Detailed examination of the kinetics of apoE synthesis and secretion in macrophages led us to postulate that intracellular degradation and secretion were competing fates for newly synthesized apoE and that newly synthesized apoE was not retained, to any substantial degree, within cells (19). We have previously shown that the presence of phospholipid vesicles or high density lipoprotein could enhance apoE secretion, primarily by directly interacting with a cell surface pool of apoE (3, 19,  4. The effect of sterols on apoE degradation in human monocyte-derived macrophages. Human monocytes were obtained by elutriation and plated and grown as described under "Experimental Procedures." Cells were incubated in growth medium alone or in this medium containing 1 g/ml of 25-hydroxycholesterol (25-OH CH) and 10 g/ml of cholesterol (CHOL) for 24 h. At that time, cells were harvested for isolation of an intermediate density nonlysosomal fraction, and the degradation of apoE was measured as described under "Experimental Procedures." At the end of the experimental incubations, apoE was quantitated in each fraction by Western blot hybridization, and the amount of apoE present after the 40°C incubation is expressed as a percentage of that present after the 4°C incubation. Values shown are the mean Ϯ S.D. from triplicate cultures. The difference between the fraction of apoE remaining at 40°C between control and sterol/oxysterol-treated cells is significant at p Ͻ 0.05. 20). This effect was rapid and could be observed within 1 h (3,19,20). In the current studies, we demonstrate the existence of an intracellular locus in which the degradation rate of apoE responds to sterol enrichment of cells. Sterols do not appear to act rapidly or directly on the degrading mechanism but, after preincubation with cells, they can suppress the degradation of apoE in a nonlysosomal compartment. The levels to which sterols suppress degradation (50 -70%) and stimulate secretion (1.7-2 fold) are in good agreement.
Additional characterization of the macrophage compartment responsible for degrading apoE in a sterol-responsive manner will require further investigation. We have already ruled out significant involvement of the plasma membrane, ER, and lysosomes and have shown that the enzyme marker for the Golgi was enriched approximately 8-fold in specific activity. However, the fraction was not likely purified Golgi, and our results do not exclude the presence of other cell constituents. The high rate of degradation of all newly synthesized proteins in this fraction suggested the presence of a major cellular degradative pathway. Degradation of the vast majority of cellular proteins is thought to be mediated by lysosomes or proteasomes (27)(28)(29). The likely involvement of proteasome-mediated degradation in the fractions we have isolated is suggested by the significant exclusion of the lysosomal marker, the lack of a chloroquine inhibitory effect, and the profound inhibitory effect of lactacystin. Lactacystin is the most specific inhibitor of proteasomal degradation reported to date (30,31).
As noted above, apoE can be substantially degraded prior to its secretion, and our data support a role for proteasomes in this degradation. In immunofluorescent studies of macrophages, apoE has been largely associated with the ER, Golgi, and secretory granules (33). ApoE has also been detected in surface-connected compartments of the plasma membrane (34). Detection of apoE within lysosomes has been inconsistent, and such distribution could result from receptor-mediated internalization of extracellular apoE (33,35). How apoE that is traversing the cellular secretory apparatus becomes a substrate for the proteasomal complex, which is thought to be predominantly cytosolic in distribution, is not completely clear. However, as investigation into the function of proteasomes has proceeded, it has been shown that a number of membrane-associated proteins are substrates for proteasomal degradation (27)(28)(29)36). ApoB has also been shown to be subject to proteasomal degradation, and this may occur prior to its translocation into the ER (37)(38)(39)(40). It is of interest that apoB, although much larger than apoE, is also a lipid-binding protein that is largely degraded prior to secretion; and that its degradation rate is modulated by the intracellular lipid environment.
How could preincubation in sterol/oxysterol or Ac-LDL modulate the susceptibility of apoE to degradation? One attractive explanation is that modulation of apoE conformation by the lipid composition of internal membranes is involved. Previous studies have demonstrated that Golgi membranes, as well as the membranes of secretory granules, are cholesterol-rich and that the cholesterol content of intracellular membranes is responsive to the lipid milieu of the cell (41)(42)(43)(44)(45)(46)(47). For example, cholesterol in the Golgi can be depleted by incubation of cells in lipoprotein-deficient serum or by inhibitors of cholesterol synthesis. Conversely, the cholesterol of these membranes can be replenished by incubation in serum containing lipoproteins or mevalonate. In addition, plasma membrane, which contains the largest portion of cell cholesterol, is recycled through the Golgi; labeling of plasma membrane lipids with fluorescent tags results in rapid labeling of the Golgi (43). It is also clear that association with different types of lipid can influence apoE conformation. Based on primary structure, apoE is predicted to contain a number of amphipathic helical segments with affinity for lipid, particularly at the carboxyl-terminal end (for a review, see Ref. 48). In the absence of lipid, apoE self-associates. In the presence of lipid, apoE conformation will depend on the composition of the lipid complex with which it is associated. Different conformations present different epitopes when mapped with monoclonal antibodies and manifest different receptor binding affinities. Further, apoE is secreted from  5. The effect of cholesterol loading with Ac-LDL on apoE degradation. Cells were plated and grown as described in the legend to Fig. 2. Where indicated, cultures were incubated in growth medium with 80 g/ml of Ac-LDL protein for 48 h prior to being harvested for measurement of apoE degradation in a nonlysosomal intermediate density fraction. The apoE present at the end of the 40°C incubation is expressed as a percentage of that present after the 4°C incubation. Values shown are the mean Ϯ S.D. from triplicate cultures. macrophages in a lipoprotein particle that contains apoE, phospholipid, and free cholesterol. The cellular source of this secreted lipid and where in the secretory pathway it becomes associated with apoE have not been determined. Consistent with this formulation, there has been a recent report of intracellular lipid acting as a non-protein molecular chaperone to maintain proper folding of an endogenously synthesized protein (49). Our measurements confirm a change in the lipid composition of the fraction responsible for degrading apoE after incubation in sterol/oxysterol. We therefore postulate that interaction with cellular membranes of differing lipid composition could modulate the folding of apoE, thereby modulating its susceptibility to proteasomal degradation. Such a hypothesis is also consistent with the dominant role, ascribed to proteasomes, for degrading misfolded proteins.
In summary, these studies show for the first time a posttranscriptional and post-translational locus for control of macrophage apoE expression by sterols. This regulatory control by sterols is associated with modulation of the degradation rate of newly synthesized apoE in a nonlysosomal cellular compartment. Our results rule out a direct effect of sterol on the degrading mechanism; sterol/oxysterol had no effect on apoE degradation when added directly to the degradation incubation. Further, preincubation in sterol/oxysterol selectively suppressed apoE degradation, while degradation of total proteins remained unchanged, suggesting that the effect of this preincubation on apoE degradation was due to a change in apoE susceptibility rather than a change in constituents of the degrading pathway. These results provide further insight into the mechanism for regulation of macrophage apoE expression in the vessel wall. They also provide a rationale for further investigating the post-translational processing of apoE in macrophages and its relationship to the intracellular lipid environment.