Cell surface proteoglycans modulate net synthesis and secretion of macrophage apolipoprotein E.

Using a macrophage cell line that constitutively expresses a human apolipoprotein E (apoE) cDNA, we have investigated the post-translational metabolism of endogenously produced apoE. Inhibition of lysosomal or cysteine proteases led to significant inhibition of apoE degradation but did not increase apoE secretion, indicating that cellular degradation is not limiting for apoE secretion in macrophages. Treatment of macrophages with inhibitors of proteoglycan synthesis (4-methylumbelliferyl-β-D-xyloside) or sulfation (sodium chlorate) enhanced the release of apoE from cells and significantly attenuated the increase in secretion produced by incubation with phosphatidylcholine vesicles (PV). These observations suggested that a significant fraction of the apoE retained by cells (and released by incubation with PV) was associated with proteoglycans. Treatment of cells with exogenous heparinase led to a greater than 4-fold increase in apoE secretion and similarly attenuated the response to PV, suggesting that apoE was trapped in an extracellular proteoglycan matrix. This conclusion was confirmed in studies showing that PV could enhance the release of apoE from cells during an incubation at 4°C, but this enhanced release was abolished in proteoglycan-depleted cells. Incubation with lactoferrin at 4 or 37°C produced a similar decrement in cellular apoE, again indicating the existence of a cell surface pool of apoE. Pulse-chase studies showed that the apoE trapped in the proteoglycan matrix was susceptible to rapid cellular degradation such that net synthesis of apoE (secreted plus cell-associated) was increased significantly in proteoglycan-depleted cells compared with control cells as early as 45 min during a chase period.

Previous work has shown that a substantial portion of apolipoprotein E (apoE) 1 which is synthesized by macrophages is not secreted but is rapidly degraded (1). We have shown that the addition of HDL 3 or phosphatidylcholine vesicles (PV) can protect a portion of this apoE from degradation and promote its secretion (1). Cellular sterol balance, which modulates apoE gene transcription in macrophages, does not appear to modulate post-translational processing of apoE or to modulate its sorting between a degradative or secretory pathway (1). Cysteine and lysosomal proteases have been postulated to be important for cellular degradation of endogenously produced apoE based on studies using inhibitor molecules for these enzyme classes (2,3).
The observation that a substantial portion of newly synthesized apoE is never secreted from macrophages raises the question of its potential function. ApoE has been postulated to modulate signaling of second messenger pathways in steroidogenic cells (4). Alternatively, intracellular apoE could be involved in the subcellular transport of lipid. Insight into potential functions could be obtained by identifying cellular compartments that retain the apoE that is not secreted. In hepatocytes, intracellular apoE has been localized to peroxisomes, cytosol, as well as elements of the endocytic and secretory apparatus (5,6). These cells, however, have a unique role in assembling complex lipoprotein particles. Leblond and Marcel (7) have identified a plasma membrane pool of apoE in hepatocytes which appears to be important for selective uptake of cholesterol ester from HDL. Studies in mouse peritoneal macrophages have identified apoE in an endosomal/lysosomal compartment (3). Other studies, using human monocyte-derived macrophages, have identified focal accumulations of apoE associated with the plasma membrane within spaces formed by deep invaginations of this membrane (8).
Studies by Mahley and colleagues (9 -11) have demonstrated that exogenously added apoE is avidly bound by the extracellular matrix of hepatocytes, especially by proteoglycans in the extracellular matrix. It has been postulated that these proteoglycans may serve as a high capacity reservoir for presenting apoE to specific cellular receptors (9 -11). Studies by Stauderman et al. (12) have also documented an important role for extracellular proteoglycans in modulating hepatocyte apoE metabolism. In view of the above considerations and our previous observations regarding the effect of PV or HDL 3 on macrophage apoE secretion and degradation, we investigated a potential role for extracellular matrix in modulating the secretion and degradation of endogenously produced apoE in macrophages. The results of our investigations indicate that newly synthesized macrophage apoE is sequestered in a pericellular proteoglycan matrix and that the addition of PV stimulates release of this sequestered pool into the medium. Experimental interventions that inhibit the formation, or produce hydrolysis, of this proteoglycan matrix interfere with the sequestration of apoE and enhance its release into the medium, thereby attenuating the enhanced release of apoE in response to PV. Further, our studies demonstrate that apoE sequestered in the pericellular proteoglycan matrix is susceptible to rapid cellular degradation.

EXPERIMENTAL PROCEDURES
Materials-4-Methylumbelliferyl-␤-D-xyloside (␤DX), sodium chlorate, ALLN, monensin, chloroquine, ammonium chloride, and lactoferrin were obtained from Sigma. All other materials were from sources * This work was supported by National Institutes of Health Grant HL 39653. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
Cell Culture-J774 cells were stably transfected to express a wild type human apoE cDNA as described previously in detail (1,13). Cells were selected and maintained in G418 (400 g/ml) until 1 week prior to 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. Phosphatidylcholine vesicles were prepared as described by Batzri and Korn (14). Vesicles were dialyzed overnight against methoninine-free Dulbecco's modified Eagle's medium plus 1% penicillin/streptomycin and filtered through a 0.45-m membrane prior to use. The THP1 human monocytic cell line was obtained from the American Type Culture Collection and grown in RPMI 1640 containing 10% fetal bovine serum. THP1 cells were incubated in 100 ng/ml 12-Otetradecanoylphorbol-13-acetate for 72 h to induce differentiation to the macrophage phenotype before starting experimental incubations. For some experiments, cells were preincubated in 1 mM ␤DX or 30 mM sodium chlorate. In separate studies we showed that inclusion of these agents inhibited the incorporation of radiolabeled sulfate into cellassociated protein by 20 -80%.
Quantitation of ApoE Synthesis and Secretion-2-3 ϫ 10 6 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 unless otherwise indicated in figure or table legends. Pulse medium contained 100 Ci/ml [ 35 S]methoninine with 1-2 M unlabeled methoninine added to methoninine-free Dulbecco's modified Eagle's medium. Pulse times varied from 45 to 60 min. Chase medium contained 500 M unlabeled methoninine. Chase time varied as indicated in the legends. At the end of the chase period apoE secreted into the medium, and that remaining associated with the cell, were quantitatively immunoprecipitated and analyzed by sodium dodecyl sulfate-polyacrylamide gel electrophoresis as described previously in detail (1). All immunoprecipitations were performed using equal numbers of total trichloroacetic acid-precipitable counts. The disintegrations/min (dpm) in secreted or cell-associated apoE are therefore corrected for total labeled secreted protein or total labeled cell-associated protein, respectively, for each experiment. The apoE bands were localized by autoradiography and excised. The dried gel piece from each sample was cut out and placed in a glass 20-ml scintillation vial, and the gel was rehydrated with 3% glycerol. The backing paper and glycerol solution were removed, and 0.5 ml of 30% hydrogen peroxide was added to each sample. The vials were capped and heated at 60°C for 48 -72 h until the gel pieces dissolved. 10 ml of scintillant was added, and the samples were counted against an external standard so that counting efficiency could be used to calculate dpm incorporated into apoE for each sample. For some experiments, autoradiographs of sodium dodecyl sulfate-polyacrylamide electrophoresis gels were scanned using an LKB XL Ultrascan laser densitometer and results expressed as fold change over control. For the purposes of our analyses, "secreted" apoE is considered to be that recovered from the medium, and "cell" apoE is considered to be that recovered from lysis of the cell layer.

RESULTS
The human apoE-producing J774 cell line used for these studies constitutively expresses a human apoE cDNA and has been characterized in detail previously (1). We have reported previously that this cell line, when incubated under standard growth conditions, produces 900 ng of apoE/mg of cell protein/24 h, an amount similar to that produced by mature cholesterol-loaded human monocyte-derived macrophages in culture. Our previous observation that addition of PV or HDL 3 could preserve a substantial amount of newly synthesized apoE from degradation and enhance its secretion (1) led us to ask which of these effects was primary; i.e. was apoE preserved from degradation because its secretion was enhanced, or was secretion limited by competing cellular degradative pathways? The data in Tables I-III address this question. We used two different classes of protease inhibitors that have been shown to prevent the degradation of newly synthesized apoE: chloroquine and ammonium chloride as inhibitors of lysosomal proteases, and ALLN as a cysteine protease inhibitor. As shown in Tables I and II, these agents prevent degradation of newly synthesized apoE in macrophages that constitutively express a human apoE cDNA, leading to its accumulation within the cell. However, this intracellular stabilization and subsequent accu-mulation do not lead to a substantial increase in the secretion of apoE, especially when compared with the effect of PV on apoE secretion. Therefore, intracellular degradation is not rate-limiting for apoE secretion. This result could be explained if, after completion of its synthesis, apoE is committed to separate cellular pools destined for secretion or degradation which are not in equilibrium. We assessed this by evaluating the effect of monensin (to block movement of apoE out of the trans-cisternae of the Golgi) on PV-induced secretion of apoE. After a labeling period of 45 min, cells were incubated during a 45-min chase with monensin, PV, or both (Table III). As shown, the addition of PV effectively mobilized apoE from a post-Golgi compartment for secretion. The data in Tables I-III therefore indicate the presence of a post-Golgi pool of apoE in macrophages which is available for secretion or degradation. PV cause translocation of apoE out of this post-Golgi compartment and thereby protect it from rapid degradation. These observations led us to focus on enhanced secretion as the mechanism by which PV modulate net accumulation of apoE. When analyzing post-Golgi compartments from which apoE secretion could be enhanced by PV, we considered the observations of Mahley and colleagues (9 -11) and Harmony and colleagues (12) that extracellular matrix could serve as a cellassociated reservoir of apoE in hepatocytes. The data in Fig. 1 and Table IV address the importance of macrophage proteoglycans in modulating apoE secretion. ␤DX substitutes for the core protein moiety of proteoglycans and significantly reduces their synthesis and appearance at the cell surface (15)(16)(17). Sodium chlorate has been shown to inhibit the sulfation of proteoglycans (18,19). ␤DX or sodium chlorate alone each enhanced the release of apoE into the medium (3.2-fold for ␤DX and 1.7-fold for sodium chlorate), suggesting that proteoglycans are a significant cellular reservoir of newly synthesized apoE. Further, in proteoglycan-depleted cells the fold increase in apoE secretion after addition of PV was either abolished (␤DX) or significantly attenuated (sodium chlorate). This result indicates that PV were releasing apoE from a cellular proteoglycan reservoir. Similar effects were observed using differentiated THP1 macrophages (not shown).
Lactoferrin has been shown to bind to cell proteoglycans and the LRP in hepatocytes (11). We therefore evaluated its effect on macrophage apoE secretion. As shown in Table V, lactoferrin at 100 g/ml increased apoE secretion into the medium by 2.5-fold. However, reduction of cell proteoglycans by prior incubation in ␤DX did not appear to attenuate the effect of lactoferrin on apoE release. In fact, lactoferrin displaced an additional 3,192 dpm of apoE from ␤DX-treated cells compared with 906 dpm in control cells. This observation not only indicates the existence of an alternate binding site shared by endogenously produced apoE and lactoferrin, but also suggests a synergistic interaction between alternate site(s) and proteoglycans. As would be expected, lactoferrin caused a reduction in cell-associated apoE. The lowest cell-associated apoE is observed in cells grown in ␤DX and chased in the presence of lactoferrin. The next series of experiments were designed to more precisely define an extracellular versus intracellular location for the newly synthesized apoE that is released from cells by lactoferrin or by depletion of proteoglycans.
For the experiments shown in Table VI, cells were chased at 37 or 4°C, with or without lactoferrin. As shown, lactoferrin is equally effective at 4 or 37°C in reducing cell-associated apoE. Because cellular secretion is arrested at 4°C, this result indicates that lactoferrin is displacing apoE from an extracellular pool. We next assessed the accessibility of the proteoglycans responsible for sequestering apoE to exogenously added heparinase. As shown in Fig. 2, treatment with exogenous heparinase during a 60-min pulse period led to a 4.8-fold increase of apoE secretion during a subsequent 45-min chase in which heparinase was not present. Further, treatment with heparinase during the pulse period significantly blunted the effect of PV added only during the subsequent chase period. Therefore,  predigestion of extracellular proteoglycans during the pulse period led to enhanced apoE secretion during the subsequent chase period and attenuated the ability of PV to enhance apoE secretion further. We next considered whether the observed effects of ␤DX or sodium chlorate could be ascribed to effects on an extracellular proteoglycan pool, or if it was necessary to consider a role for these agents in altering the intracellular processing of apoE. Cells were preincubated with no addition, or with ␤DX or sodium chlorate to inhibit the synthesis and/or sulfation of proteoglycans respectively (Fig. 3). During the pulse period, cells were treated with exogenous heparinase for 45 min. The results in Fig. 3 are expressed as fold change with heparinase treatment compared with the same preincubation condition without heparinase treatment. Cells not preincubated in ␤DX or sodium chlorate released substantially more apoE during the chase period after treatment with exogenous heparinase during the preceding 45 min. However, cells preincubated in ␤DX or sodium chlorate did not respond to the heparinase treatment during the pulse with increased secretion of apoE during the chase. This result suggests that the effect of these agents on apoE secretion derives from inhibition of the formation of an extracellular pool of proteoglycans which is responsible for sequestering apoE. Next we more directly studied the relationship between the pool of apoE released by PV and that sequestered in extracellular proteoglycans (Fig. 4). Cells were grown with no addition, or with ␤DX or sodium chlorate for 72 h. After a 1-h labeling period, cells were washed with chase medium at 4°C and then chased at 4°C in the presence or absence of PV. The results of Fig. 4 are expressed as fold change with PV compared with the same preincubation condition without PV. First, it can be seen that the addition of PV at 4°C to cells not preincubated in ␤DX or sodium chlorate enhances the release of apoE, confirming the extracellular location of the PV-releasable pool. Second, cells preincubated in ␤DX or sodium chlorate show no increase in apoE release after the addition of PV because production of this pool has been inhibited.
Next we directly assessed the effect of proteoglycans on macrophage apoE turnover in more detail. In cells depleted of proteoglycans (Fig. 5) by a 3-day incubation in 1 mM ␤DX the amount of apoE secreted into the medium is higher at every point beyond 20 min compared with control cells. Further, the accumulation of apoE in the medium in ␤DX-treated cells appears to be linear for 120 min, whereas in control cells the amount of apoE in the medium appears to plateau and changes very little between 20 and 120 min. ApoE associated with the cell is also similar in the two groups at 20 min, but in proteoglycan-depleted cells it falls rapidly over the next 25 min to a level that is approached more slowly by control cells over the next 100 min. At 120 min, control and proteoglycan-depleted cells have similar levels of cell-associated apoE, but secreted apoE is 2.5-fold higher in the proteoglycan-depleted cells. These data indicate that newly synthesized apoE trapped in the extracellular proteoglycan matrix is susceptible to rapid cellular degradation. Cultures were preincubated for 72 h in growth medium alone or with 1 mM ␤DX or 30 mM sodium chlorate (NaChl). Where indicated, freshly made heparinase (Hep), 3 units/ml, was added only during a 60min labeling period. Cells were then washed and chased for 45 min with no additions. Autoradiographs were quantitated by laser densitometry and results expressed as fold change with heparinase compared with the same preincubation condition without heparinase treatment. Values shown are the mean Ϯ S.D. from triplicate cultures. DISCUSSION The data in this manuscript indicate that a substantial portion of the apoE that is newly synthesized by macrophages is sequestered in an extracellular proteoglycan pool. The extracellular location of this pool is confirmed by its accessibility to exogenously added heparinase (Figs. 2 and 3) and by the results of experiments utilizing chase incubations at 4°C (Table  VI and Fig. 4). The rapidity with which this pool turns over is indicated by the data in Fig. 5; after 45 min of chase time there is already a significant difference in the net accumulation of FIG. 4. Effect of PV on apoE secretion in control and proteoglycan-depleted cells at 4°C. Cells were seeded and grown as described under "Experimental Procedures." Cells were preincubated for 72 h in growth medium alone or with 1 mM ␤DX or 30 mM sodium chlorate (NaChl). Cells were pulse labeled for 1 h at 37°C followed by a 1-h chase period at 4°C. Where indicated, 1 mg/ml PV was added during the chase period. Results are expressed as fold change with PV compared with the same preincubation condition without PV. Values shown are the mean Ϯ S.D. from triplicate wells. apoE between proteoglycan-depleted and control cells. Sequestration of apoE within this extracellular proteoglycan matrix, therefore, renders it susceptible to rapid cellular degradation.
Ye and colleagues have studied the effect of protease inhibitors on apoE degradation in hepatocytes and Chinese hamster ovary cells (2). Both ALLN and lysosomal inhibitors increased apoE accumulation within cells. In our studies, we show a similar effect of ALLN and lysosomal inhibitors on macrophage apoE accumulation, and we demonstrate that such cellular accumulation is not reflected in increased apoE secretion. Degradation of apoE by cysteine proteases or lysosomal proteases therefore is not rate-limiting for apoE secretion.
Others have previously studied the role of proteoglycans on hepatocyte secretion of apoE (12). However, cellular apoE degradation and turnover data were not reported. In these studies, heparan sulfate or heparin increased apoE secretion into the medium. Preincubation of cells in ␤DX attenuated the effect of heparin, but ␤DX alone did not increase the amount of apoE secreted into the medium. On the basis of this observation, it was suggested that apoE and proteoglycans are transported to the plasma membrane as a complex (12). This result is different from what we have observed in macrophages in which preincubation with ␤DX or sodium chlorate or treatment with heparinase directly increase the amount of apoE which is secreted into the medium. In macrophages, therefore, our data suggest that newly synthesized apoE becomes trapped in a preexisting pool of pericellular proteoglycans.
Mahley and co-workers (9 -11) have reported extensively on the importance of extracellular proteoglycans for modulating cellular metabolism of exogenously added apoE, or lipoprotein ligands containing apoE, by hepatocytes. They have shown that cell surface proteoglycans act as an important intermediary serving to present apoE to specific cell surface receptors, for example the LRP. In one series of studies, it was demonstrated that lactoferrin inhibited remnant lipoprotein (and apoE) binding to hepatocytes and Chinese hamster ovary cells by interacting with both proteoglycans and the LRP (11). These observations are of specific interest for the interpretation of the data shown in Table V. Lactoferrin alone increased apoE released into the medium and decreased cell-associated apoE. This is consistent with its interaction with proteoglycans, or the LRP, leading to displacement of apoE from these sites. However, depletion of cellular proteoglycans does not attenuate the effect of lactoferrin on the release on apoE. Further, the displacement of apoE from proteoglycan-depleted cells appears to be enhanced compared with control cells (based on the apoE dpm displaced). This suggests that displacement from the non-proteoglycan site may in fact be facilitated in the absence of proteoglycans. This is consistent with the hypothesis put forth by Mahley and co-workers (9 -11) that proteoglycans and cell surface receptors interact synergistically to bind and degrade apoE.
The data in Fig. 5 indicate that the pool of extracellular proteoglycans that sequester apoE are involved in facilitating its rapid cellular degradation. This would suggest that these proteoglycans are intimately associated with, or perhaps even anchored in, the plasma membrane. The rapidity of apoE turnover in this pool may also indicate the involvement of one of the cell surface endocytic receptors for apoE (e.g. LRP, low density lipoprotein receptor, very low density lipoprotein receptor) acting in concert with cell surface proteoglycans. Receptors exhibit rapid lateral mobility in the plasma membrane (for review, see Ref. 20). Because of this, apoE bound to abundant proteoglycan binding sites will have frequent encounters with higher affinity endocytic receptors that can rapidly internalize and degrade apoE.
The model we propose for post-translational regulation of apoE by pericellular proteoglycans is presented in Fig. 6. In the macrophage, newly synthesized apoE may be directed to the cell surface (b) or may remain intracellular (a). However, additional investigation is required to confirm the existence of a pool of apoE that remains (and functions?) intracellular without ever reaching the cell surface (as indicated by the question mark). ApoE directed to the cell surface may be secreted into the medium (b3) or trapped at the cell surface (b1, b2). Our data indicate that a preexisting pool of pericellular proteoglycans is important for pericellular trapping of apoE (b1); however, there also appears to be evidence for alternate cell surface binding site(s) that function to sequester newly synthesized apoE at the cell surface (b2). The addition of PV or lactoferrin stimulates release of apoE from cell surface binding sites into the medium. ApoE that is sequestered at the cell surface is susceptible to rapid cellular degradation (c, d). Whether apoE bound to pericellular proteoglycans must first be transferred to endocytic cell surface receptors (e.g. the LRP) prior to cellular degradation (c) or can be internalized directly by plasma membrane-bound proteoglycans (d) requires additional investigation. Questions that also require further investigation include how apoE is sorted at the membrane between secreted and sequestered pools and whether PV and lactoferrin release apoE from b1 and b2 sites with equal efficacy. Further, the identity of the proteoglycan(s) that bind apoE and whether these are integral plasma membrane proteins will require additional study. This model, supported by the data from our studies, proposes that the pericellular proteoglycan matrix represents an important locus for post-translational control of macrophage apoE expression. Changes in extracellular matrix composition which accompany macrophage differentiation or activation (21)(22)(23) can thereby significantly modulate the net accumulation of apoE at any tissue site where macrophages are a significant source of apoE, for example in the atherosclerotic vessel wall (24). Further, binding of macrophage-derived apoE in the matrix could alter the biological activity of cytokines or FIG. 6. Proposed model for modulation of macrophage apoE production by cell surface proteoglycans. The figure presents the salient points of the model we propose for the modulation of macrophage apoE production by pericellular proteoglycans. Important aspects of post-translational regulation of macrophage apoE production which require additional investigation are designated by question marks. The a pathway represents a potential pool of apoE which is synthesized but never reaches the cell surface. The b pathway is apoE directed to the cell surface which may be (b1) attached to a preexisting pool of pericellular proteoglycans, (b2) attached to alternate cell surface binding site(s) (e.g. the LRP), or (b3) secreted into the medium. PV or lactoferrin (LF) enhances the release of apoE from the b1 and b2 sites. Pathway c represents endocytosis and degradation of apoE via cell surface endocytic receptors; pathway d represents potential internalization and degradation via membrane-bound proteoglycans. The model is explained further under "Discussion." growth factors by displacing them from matrix binding sites and altering their interaction with their respective cell surface receptors. In addition, apoE in pericellular matrix could influence macrophage interaction with other lipoprotein and nonlipoprotein ligands that can directly interact with apoE.