Uptake of High Density Lipoprotein Cholesterol Ester by HepG2 Cells Involves Apolipoprotein E Localized on the Cell Surface*

High density lipoprotein (HDL) cholesterol ester (CE) is taken up by many cells without simultaneous uptake of HDL apoprotein. The studies described herein demonstrate that the selective uptake of cholesterol ester by HepGZ cells is reduced by antibody di- rected against the receptor-binding domain of apoE (monoclonal antibody (mAb) 1D7) but not by antibody directed against the NHz-terminal portion of the molecule. The reduction, by 1D7, of HDL cholesteryl ester uptake is not due to apoE acquisition by the labeled HDL preparation or by the transfer of r3H]CE of HDL to apoE-containing lipoproteins and uptake by the apoB/E or apoE receptors. Rather, it appears that mAb 1D7 recognizes apoE localized at the cell surface of HepGZ cells. This conclusion is supported by the fact that:


Uptake of High Density Lipoprotein Cholesterol Ester by HepG2 Cells Involves Apolipoprotein E Localized on the Cell Surface*
(Received for publication, July 13, 1992) Lorraine Leblond  High density lipoprotein (HDL) cholesterol ester (CE) is taken up by many cells without simultaneous uptake of HDL apoprotein. The studies described herein demonstrate that the selective uptake of cholesterol ester by HepGZ cells is reduced by antibody directed against the receptor-binding domain of apoE (monoclonal antibody (mAb) 1D7) but not by antibody directed against the NHz-terminal portion of the molecule. The reduction, by 1D7, of HDL cholesteryl ester uptake is not due to apoE acquisition by the labeled HDL preparation or by the transfer of r3H]CE of HDL to apoE-containing lipoproteins and uptake by the apoB/E or apoE receptors. Rather, it appears that mAb 1D7 recognizes apoE localized at the cell surface of HepGZ cells. This conclusion is supported by the fact that: 1) reduction of HDL-CE uptake by HepG2 cells is observed within 15 min after the addition of the antibody-ligand mixture; 2 ) 1D7 is similarly effective in reducing the selective uptake of HDL-CE when added to the ligand or to the cells; 3) three different anti-apoE mAbs (1D7, 3B7, and 3H1) bind specifically to the surface of the cells. We have also demonstrated that heparin (5 mg/ml) does not reduce the amount of apoEimmunoreactive material bound at the cell surface when added before or after the binding period. 1D7, but not 3B7 or 3H1, binds less in the presence of heparin. The observations are consistent with a localization of apoE on the cell membrane rather than on lipoproteins bound to apoB/E or apoE receptors.
HDL' plays a central role in the return of excess cholesterol from extrahepatic tissues to the liver for excretion or reutilization (1). HDL induces cholesterol efflux from the peripheral cells by a nonspecific exchange process with possible involvement of specific recognition site for HDL (2). Free cholesterol is esterified by 1ecithin:cholesterol acyltransferase, and cholesteryl esters, transported in HDL, may be cleared from the circulation by at least three processes. In species with CETP, HDL-CE can be transferred to very low density lipoprotein and chylomicrons remnants which are largely cleared by the liver (3). Second, HDL can acquire apoE and be targeted for hepatic uptake via the apoB/E receptor (4). Finally, HDL-CE may be directly taken up without a parallel uptake of the HDL particle, a process that has been termed selective uptake by Pittman and colleagues (5, 6).
The selective uptake of HDL cholesteryl ester has been observed in a number of rat tissues ( 5 , 6), in a variety of cultured cells from other species (6)(7)(8), and in some organ culture systems (9)(10)(11). Selective uptake of HDL-CE does not correlate with high affinity binding of HDL (8,12), does not depend upon any specific HDL apoprotein component (7,13), and may be enhanced by CETP (14). In vitro studies have shown that selective uptake is a saturable process involving a net mass uptake of CE from HDL and regulated by cell cholesterol status (8,15), this regulation being apparently at the level of the plasma membrane pool (16).
The importance of a specific interaction between HDL and the cell membrane in the uptake of cholesteryl ester remains unclear. It is possible that HDL binds to the cell surface and that cholesteryl esters are preferentially transferred into the cell. Some studies suggest that dissociation of HDL components may occur at the hepatocyte surface (11, 17) and may be followed by a specific cholesteryl ester membrane transport (18) and hydrolysis in a non-lysosomal compartment (19). Another possibility is that the entire HDL particle is internalized with the apoprotein component being released from the cell. There is evidence that at least part of the HDL apoprotein may be degraded by the liver (20). Resecreted HDL are depleted of cholesteryl ester and have an altered size distribution (21).
In the present paper, we demonstrate that a monoclonal antibody directed against the receptor-binding domain of apoE (mAb 1D7) reduces the selective uptake of cholesterol ester from HDL to HepG2 cells. This reduction is attributed to the interaction of the antibody with the apoE molecule localized on the cell surface of HepGZ cells. The apoE was detected by three different anti-apoE mABs and was not associated with lipoproteins bound to apoB/E or apoE receptors. These studies suggest that the amino acid sequence 139-169 of apoE plays a role in the selective uptake of HDL cholesterol ester. Lipoprotein Preparation and Labeling-ApoE-free HDL were isolated by density gradient ultracentrifugation from serum of normal fasted subjects as described earlier (22). The HDL fraction was passed over two immunoadsorbers containing anti-apoE mAbs 1D7 and 3B7. HDL fractions were checked for the presence of apoE by SDS-polyacrylamide gel electrophoresis, followed by Coomassie Blue staining, or by immunoblotting with specific monoclonal antibodies (3H1 and 6C5). A faint contamination by apoE was detectable only by immunoblot analysis.
HDL was labeled with lZ5I by the iodine monochloride method of McFarlane (23) with the appropriate modifications for lipoproteins (24). Specific activities of preparations varied between 300-500 cpm/ ng of protein. More than 95% of the radioactivity in T -H D L preparations was associated with protein.
[3H]cholesteryl linoleyl ether was incorporated into HDL essentially as described by Roberts et al. (25). The procedure involved incubation of HDL in human lipoprotein-deficient serum (d > 1.215) that had been prelabeled with [3H] cholesteryl linoleyl ether, followed by reisolation of the labeled lipoprotein by ultracentrifugation. The specific activities of the labeled HDL were between 20,000 and 30,000 cmp/pg of esterified cholesterol. The labeled HDL were analyzed for cholesterol, cholesteryl esters, and protein. No significant compositional differences were observed between the labeled and unlabeled HDL.
When the cell monolayer was almost confluent, the medium was removed, cells were rinsed twice with sterile phosphate-buffered saline, and incubated overnight in serum-free medium containing 2 mg/ml fatty acid-free albumin. The cells were washed again (phosphate-buffered saline 2 X), and the HDL association assays or Fab fragment binding were carried out. Cell protein was about 1.0-1.2 mg/35-mm well.
HDL Cell Association-Assays for the cell association of labeled '"I or [3H]CE-HDL were carried out as previously described (22). Cells were incubated (37 "C, 5% CO,) for the indicated times in MEM containing BSA (2 mg/ml) and labeled HDL (usually 5-10 pg of HDL protein/ml) with or without antibodies. At the termination of the cell association assay, the medium was removed and the cells were washed as previously described (22). To determine the amount of '251-cell associated radioactivity, 2 ml of 0.1 N NaOH was added directly to the washed monolayer. Part of the NaOH digest was used for protein determination and part to determine the radioactivity. Cell lipids were extracted directly from HepG2 cell monolayers in the plastic culture dish. The washed monolayers were incubated with hexane/ isopropyl alcohol (3:2, v/v) as described (22). The protein remaining on the dish after the lipid extraction was allowed to dry and digested in 2 ml of 0.1 N NaOH. Aliquots were taken and protein was determined by the method of Lowry et al. (26) using bovine serum albumin as a standard. Total radioactivity was determined directly by counting an aliquot of the hexane-isopropyl alcohol extract in a liquid scintillation counter.
Production of Monoclonal Antibodies, Preparation of IgG and Fab Fragments, Iodination, and Binding Studies-Production and characterization of mAbs against human apoE and CETP have been described previously (27)(28)(29). The mAbs were isolated from a mixture of plasma and ascitic fluid from hybridoma-bearing mice by affinity chromatography on Protein A-Sepharose 4B (Pharmacia Inc., Uppsala, Sweden) (30) and stored at 4 "C. Fab fragments were prepared from IgG fractions using papain as described previously (31) and were separated from undigested IgG and Fc fragments by affinity chromatography on Protein-A Sepharose (32). The purity of the Fab fragments was checked by SDS-gel electrophoresis.
The four anti-human apoE mAbs used in this study were 1D7, 6C5, 3B7, and 3H1; mAbs 6C5 and 3B7 compete for either the same or closely situated antigenic determinants, which are located near the amino-terminal of apoE, probably between amino acid 1 and 11 (33). The epitope of 3H1 is situated in the COOH-terminal part of apoE, probably between amino acids 243 and 272. The epitope of 1D7 is in the center of the molecule, between amino acids 139 and 169, and this mAb can inhibit apoE-mediated binding of lipoproteins to the apoB/E (34) and apoE receptors (64). method (35).
Antibodies were labeled by a modification of the chloramine T The method used for the binding assay of antibodies to cells is essentially the same as described for the cell association of HDL except that all the procedures were done at 4 "C. Briefly, nearly confluent cells exposed for 18 h to MEM medium supplemented with 2 mg/ml BSA (fatty acid-free) were prechilled at 4 "C for 30 min and then incubated at 4 "C for 3 h with a range of radiolabeled antibody concentration, in the presence or absence of an excess of unlabeled antibody. The '251-Fab-specific activities were preadjusted to 1500 cpm/ng protein by the addition of unlabeled Fab. The cells were then washed and treated as described in the lipoprotein association studies. Nonspecific binding of Fab was also determined in the presence of Fab prepared from non-immune IgG mice preparation or from nonrelated IgG preparation.
ApoE Radioimmunoassay-The sandwich assays for the measurement of apoE were carried out using 6C5 as the immobilized capture mAb and the 3H1 mAb as the labeled probe. The amount of apoE in the sample was determined using a calibration curve with known amounts of apoE in the range of 0.1-1 pglassay. To measure the release of apoE by HepG2 cells, cultures grown for 4 or 5 days were used. The monolayers were about 90% confluent at this time. The cell were incubated in the desired medium for a specified period and a small aliquot of medium, usually 10-100 pl, was taken for the assay. A similar aliquot of medium not incubated with the cells was used as blank and, wherever necessary, this value was subtracted from the assay values. All results were calculated in terms of micrograms of apoprotein secreted per milligrams of cellular protein per specified time.
Polyacrylamide Gel Electrophoresis and Immunoblotting-Electrophoresis of proteins and lipoproteins in the presence of SDS-polyacrylamide gel electrophoresis was performed according to the method of Schagger (36). The transfer from SDS gels was performed in 20 mM Tris, 150 mM glycine, pH 8.3, 30% methanol (v/v). The nitrocellulose replicas were used for immunochemical detection of proteins as described earlier (37). Incubation with anti-apoE antibodies were made at 37 "C in 10 mM Tris, 150 mM NaC1, 0.01% NaN3, pH 7.4, containing 1% BSA. '9-Labeled second antibody (rabbit anti-mouse IgG) was added subsequently and the nitrocellulose paper was autoradiographed on XAR-5 Kodak films with an intensifier screen (Cronex, Du Pont).

RESULTS
The Anti-apoE mAb 107 Reduces Selective Uptake of HDL Cholestelyl Ester by HepG2 Cells-In previous studies on the HDL-binding site specificity (22), intriguing results were obtained with two anti-apoE mAbs suggesting that an apoEcontaining lipoprotein, which binds to the apoE and/or apoB/ E receptors, participates in the selective uptake of cholesteryl ether. We proceeded to examine this phenomenon in more detail. In parallel experiments, lZ5I or [3H]cholesteryl etherlabeled HDL were treated with anti-apoE mAbs prior to their incubation with cells, and the cell association of HDL tracers was studied during a 4-h incubation. As shown in Fig. 1, mAb 3B7, which does not interfere with the binding of apoE to its receptor (27), had no effect on the cellular association of CE-HDL, while 1D7 which prevents binding of apoE to its receptors did reduce the cellular uptake of labeled cholesteryl ether by 40% or more. The addition of either 1D7 or 3B7 minimal or no effect on the cell association of Iz5I-HDL. ApoE Secretion by the HepG2 Cells-The previous finding suggested that apoE is involved in the uptake of HDL cholesteryl ether by the apoB/E and/or apoE receptor. The origin of the apoE that mediated the uptake of the labeled CE-HDL which had previously been depleted of endogenous apoE was unclear. Only traces of apoE remained associated with the HDL particle after their passage on the anti-apoE columns, an amount insufficient to mediate the cell association of the labeled lipoproteins. The other possibility was that the apoE was secreted by the HepG2 cells. HepG2 cells are known to secrete cholesterol ester-poor, apoE-rich particles (38,39). The apoE secretion by the cells was quantified under our experimental conditions (Table I). The amount of apoE detected in the media by immunoassay (0.38 pg apoE/mg cell protein/4 h) is in agreement with previous observations (39, 40). On the other hand, calculated per milligram of cell protein, apoE secreted by the cells in the presence of HDL was 1.6 times higher than in the absence of HDL. This amount does not take into account endocytosis or degradation of secreted protein during incubation (4 h).
Association of the newly secreted apoE with HDL can lead to the formation of HDL with apoE which can target these lipoproteins for uptake via the apoB/E or apoE receptors. However, these receptors are known to internalize the entire particle (3,41,42), and the results in Fig. 1 do not support the participation of such a pathway, which requires the whole HDL particle to be internalized and mAb 1D7 should be as effective on the cellular uptake of '251-labeled HDL as on the cellular uptake of [3H]CE-labeled HDL. We have in fact consistently observed that addition of 1D7 or 3B7 had minimal or no effect on the cell association of 1251-HDL while 1D7 was consistently more effective than 3B7 in reducing [3H]CE-HDL cell association. The reduction of the cellular uptake of [3H]CE-HDL observed with 1D7 is therefore specific to the lipid component of HDL. The small reduction observed with T -H D L in the presence of 1D7 is probably due to the residual trace of apoE in our apoE-depleted HDL preparations.
CETP Is Not Involved in the Selective Uptake of HDL Cholesteryl Ester by HepG2 Cells-The significant reduction in the cell association of HDL-CE observed with 1D7 might be due to the transfer of [3]cholesteryl ether from HDL to other apoE-containing lipoproteins secreted by HepG2 cells in the medium which are then removed by the apoE and the apoB/E receptors. We examined the possibility that CETP secreted by the HepG2 cells (43) might be involved. CETP is known to mediate the transfer of cholesteryl esters and triglycerides between lipoproteins (44). We explored this possibility by adding anti-CETP mAbs in the culture medium during cellular HDL association assay. The antibody, TP2, has been found to neutralize the cholesteryl ester and triglyceride transfers catalyzed by CETP, whereas antibody TP6, directed to another epitope, has no effect on these transfer activities (29). As shown in Fig. 2, neither TP2 nor TP6 had any significant effect on the cellular association of CE-HDL during a 4-h incubation. Others have shown that secreted CETP can play a role in CE uptake by HepG2 cells but only during long incubation periods (20)(21)(22) and in the situation where there is accumulation of secreted lipoproteins containing apoB or apoE in the culture medium (45,46). ApoE Is Localized on the Cell Surface of the HepG2 Cells-We proceeded to evaluate, in the most general way possible, the relation between the cell association of HDL cholesterol ester and the appearance of apoE-immunoreactive products secreted in the medium by HepG2 cells. As with any secretary product, the effect should depend on the quantity of substances accumulating in the medium and thus on the time of incubation. In the experiment illustrated in Fig. 3, [3H]CE-HDL cell association was compared from 15 min to 4 h in the presence and absence of 1D7 mAb in the incubation medium. The decrease in HDL-CE association to cells caused by mAb 1D7 was clearly very rapid. Within 15 min after addition of 1D7, cellular uptake of HDL-CE dropped by 25%. When incubation was continued, there was a further decrease to 60% of control after 4 h. The incubation was not prolonged beyond this period to avoid too large an accumulation of apoE-containing lipoproteins in the medium.
This suggested that some of the cell association of CE may  be mediated by the accumulation of secreted products in the medium. However, the most interesting conclusion of this experiment was that apoE is secreted very rapidly into the culture medium and/or that apoE is already present and accessible on the surface of HepG2 cells before addition of the HDL preparation.
In order to verify the presence of apoE at the surface of the HepG2 cells and the role of this apoE in the cellular association of [3H]CE-HDL, the cells were preincubated with 1D7 mAb for 2 h and washed prior to the addition of labeled HDL. In another set of experiments, the cells were treated in the same way but without 1D7 in the preincubation medium, and 1D7 was added late in the experiment, concomitantly with the HDL preparation. As shown in Table 11, the inhibition of CE association to the cells was as effective when 1D7 was added prior to the incubation with labeled HDL as it was when added together with the labeled HDL. The 3B7 mAb used as control was completely ineffective under both conditions. This experiment demonstrated that a direct interaction of the Fab preparation with an apoE molecule on the cell surface caused the inhibition of selective uptake of HDL cholesteryl ester.
Fab Fragments of Anti-apoE mAbs Bind Specifically to the Surface of HepGP Cells-As a more direct test to demonstrate the presence of apoE on the surface of the cells, the binding of three different anti-apoE Fab fragments to HepG2 cells was studied at 4 "C. The three mAbs selected recognize one of the NH'-terminal domain (3B7), the COOH-terminal domain (3H1), or a middle area of the molecule (1D7). Nonspecific binding was assessed by using Fab fragment prepared from IgG of non-immune mouse or from IgG of unrelated antibody. As shown in Fig. 4, all three antibodies bound specifically and with saturability although different amounts of each antibody bound to the surface of HepG2 cells. The saturation of binding sites occurs at antibody concentrations over 6 pg of Fab fragment protein/ml. The difference in the absolute values of antibody bound may indicate that the antigenic determinants of 1D7 or 3B7 are either partially inaccessible to antibody on the cell surface or that these determinants are poorly recognized by the antibodies. However, the results obtained with very low density lipoprotein, reconstituted apoE-containing liposomes, and free apoE attribute decreased recognition by 1D7 and 3B7 to the lower affinity of these antibodies for apoE rather than to inaccessibility of the antigenic site.' Scatchard analysis of the binding data give a linear curve indicating the presence of an unique binding site for each of them (results not shown). We have also repeated the same experiments with and without a 20-fold excess of cold ligand. For all mAbs, binding results are essentially the same as those L. Lehlond and Y. L. Marcel, unpublished data. reported earlier with the non immune preparation (not shown).
Effect of Heparin on the Binding of mAbs to HepG2 Cells-HepG2 cells express both LDL (47) and LDL receptor-related protein receptors (48) and secrete a virtually full complement of plasma proteins including apoB and apoE (49). The apoE is largely associated (98%) with material of density greater than 1.063 g/ml and of particle size greater than 12.2 nm (39). One can speculate that some of these large apoE-rich discoidal particles secreted by the HepG2 cells are still at the cell surface or in association with an apoE receptor. To rule out the possibility that apoE detected at the cell surface is present on lipoprotein particles bound to a receptor, the cells were treated with heparin (5 mg/ml). At this concentration, heparin blocks apoB/E receptor-mediated lipoprotein binding in many cells (50). Therefore, cells were preincubated with or without heparin for 1 h at 4 "C before the binding assay of antibody. As shown in Fig. 5A, the binding of ID7 to cells was essentially not affected by the heparin treatment. We also demonstrated the lack of effect of washing the cells with heparin after binding of the anti-apoE mAbs (Fig. 5 B ) . The small amount of radioactivity released in the heparin medium is essentially the same as with the control medium. Similar results were obtained with 3B7 and 3Hl mAbs (not shown).
We also tested the binding of Fab fragments to cells in presence of heparin at 4 "C. As shown in Table 111, binding of 3B7 and 3H1 was unaffected while binding of 1D7 was reduced by 31%. Since only 1D7 binds less in the presence of heparin, this suggests that we have a competition between heparin and antibody 1D7 for the same binding site rather than a dissociation of apoE from the cell membrane. Indeed the epitope of 1D7 overlaps with a heparin binding site on apoE, i.e. residues 142-147 (33). In addition, since only one heparin-binding site is expressed, the apoE detected is possibly in association with phospholipids as reported by Weisgraber and colleagues (32) for apoE-dimyristoylphosphatidylcholine complexes.
Effect of HDL on the Binding of Anti-apoE mAbs to HepG2 Cells-It has been previously suggested that HDLs can induce apoE secretion in macrophages and that this effect is dependent on an interaction between HDL, and a plasma membrane receptor (51). As reported above (Table I), we also noted that apoE secretion by HepG2 cells is significantly increased in the presence of HDL. In light of this observation, we inves- tigated the possibility that the increase in apoE secretion in presence of HDL may be related to a loss of membrane associated apoE. Table IV shows that the addition of HDL (50 pg/ml) to the HepG2 cells before or after the addition of labeled mAbs does not significantly modify the amount of apoE detected on the cell surface by the antibodies. These results suggested that the increased amount of apoE detected in the culture media did not result from a loss of apoE anchored at the plasma membrane and that the apoE associated with the plasma membrane represents a pool distinct from that of the apoE secreted in the medium.

DISCUSSION
The interaction of HDL with hepatocytes results in a selective uptake of cholesterol esters such that the amount of HDL cholesterol ester transferred to the cell is disproportionately higher than that of HDL protein (6,11,22). This transfer of cholesterol esters to cells is partially inhibited by mAb 1D7, an antibody that recognizes the receptor-binding domain of apoE but not by another anti-apoE mAb, 3B7, directed against the NHp-terminal part of the molecule. The present study Values shown are mean f S.D. triplicate wells. Antibody concentration used in this experiment was 3.0 pg/ml. The binding studies were carried out as described under "Experimental Procedures," except that in column A cells were first incubated with HDL for 3 h a t 37 "C, washed three times with PBS and then incubated for 3 h a t 4 "C with the indicated Fab; and, column B cells were first incubated for 3 h at 4 "C with the indicated Fab, washed three times with phosphate-buffered saline and then incubated to 3 h a t 37 "C with HDL. The binding of labeled Fab relative to control (incubation in absence of HDL) is indicated in parenthesis. demonstrates that this inhibition is neither the result of association of newly secreted apoE with the HDL particle nor the transfer by CETP of cholesteryl ester from HDL to an apoE-secreted lipoprotein followed by the uptake of this apoEcontaining particle by the apoB/E or E receptors. These possibilities were excluded even though a greater amount of apoE can be detected in the culture medium after addition of HDL. We eliminated the possibility of a direct action of CETP secreted by the cells on the HDL-CE cell association as reported by Granot et al. (14). The presence of mAb TP2 or mAb T P 6 in the incubation medium did not affect at all the amount of HDL cholesterol ester transferred to the cell.
The inhibition of the selective uptake of HDL cholesteryl ester observed in the presence 1D7 was, however, associated with an action of the antibody with an apoE molecule localized on the cell surface. Indeed, mAb 1D7 was equally effective in reducing the cell association of CE-HDL when added before or with the HDL preparation.
Moreover, the presence of apoE on the cell membrane was confirmed by the specific and saturable binding of three anti-apoE mAbs directed against epitopes localized on different parts of the molecule. The presence of this apolipoprotein at the cell surface of hepatocytes has also been reported by others (52). The present study provides evidence that this apoE is not associated with lipoproteins bound to apoB/E or apoE receptors as suggested by Hamilton et al. (52) since heparin treatment does not reduce the amount of labeled antibodies bound to the cell surface. In addition, it seems unlikely that apoE is immobilized on the cell surface of HepG2 by the mechanism that anchors hepatic lipase, as indeed the heparin sulfate proteoglycan that retains hepatic lipase is also released by the addition of heparin (53).
Our demonstration of the participation of apoE in the selective uptake of CE-HDL is not in full accord with previous studies (6,7,19,45), but some of the apparent discrepancies can be explained. In two studies (6, 71, exogenous apoE was used either as reconstituted apoE-containing liposomes or added to HDL present in culture medium. The results presented here clearly show that it is the apoE localized on the plasma membrane that is important in this uptake not that present in the medium or on the ligand. Delamatre and colleagues (19) reported that the incubation of [3H]CE-HDL with polyclonal antibodies against apoE did not inhibit total ["ICE uptake. Since not all anti-apoE mAbs reduce uptake of CE but only mAb 1D7, it is quite possible that the polyclonal antiserum used lacked antibodies with the specificity of 1D7. On the other hand, Rinninger and Pittman (451, who also used our antibody 1D7, failed to observe an effect of this sions and a critique of the manuscript, Carole Poudrier and Claude mAb on the selective uptake of HDL-CE. This clearly contra-Godbout for technical assistance in cell culture, and Christine Lemire dicts our findings reported here for the effect of 1D7 on the cell association of '*'I-HDL and I3H1CE-HDL. Although REFERENCES for preparing the manuscript. methods used to prepare apoE-free HDL, to label it, and to measure selective uptake may have differed, we cannot fully explain this discrepancy. However, complete details on how the authors conducted this particular experiment are not available although we were surprised by the large amount of apoE associated with their HDL preparations. In our preparations, almost no apoE was initially associated with HDL even if a greater amount of this apolipoprotein were secreted in the medium when cells were incubated with HDL. Selective uptake of cholesterol esters has also been reported in 5774 macrophages which fail to synthesize apoE (54). However, these cells show an alteration in the mechanism of esterified cholesterol clearance in the presence of HDL (55). LOSS or acquisition of apoE expression by cells have recently been associated with a number of failures in cellular response to external agents (55,56).
In the light of recent studies, it seems very possible that apoE may participate in the selective uptake of CE but the exact mechanism by which apoE is involved in this process remains to be established. HDL associates with the cell membrane by a mechanism that may or may not involve a specific receptor and that results in the insertion of the HDL cholesterol ester in the plasma membrane (16,17). The extent to which hepatic lipase may participate in this process is unclear. Close associations between HDL and cell membrane have been reported (57,58), and the insertion may be followed by interaction between the HDL lipid and apoE in such a way that a complex of cholesteryl ester and apoE, possibly including phospholipids, segregates and is taken up by a mechanism that involves the positively charged region of apoE in which the 1D7 epitope is localized. This may involve binding to the LDL receptor-related protein receptor (48) or by another receptor or by association to a proteoglycan-like "perlecan," reported to contain a domain with high homology to the binding site of apoB or apoE in the LDL receptor (59). This remains to be established. An apoE-binding macromolecule like the mitochondrial F1-ATPase or the 59-kDa protein described by Bieseigel and colleagues (60) could be also considered. Antibody 1D7 has been shown to inhibit the binding of apoE to F1-ATPase (60). This hypothesis implies that apoE localized on/at the cell surface is in close association with a protein that serves in the intracellular transport of apoEcontaining complexes. It has been also suggested that apoE itself can act as a carrier for cholesterol transport between subcellular components within the hepatocytes (52). The earlier studies of Gwynne and Mahaffee (18) and Pittman and colleagues (15,16,61) suggest the existence of specific transmembrane and intracellular transfer of cholesteryl ester followed by hydrolysis.
Another possibility may be the interaction of HDL with apoE at the cell surface followed by internalization, delivery to the trans-Golgi apparatus, and resecretion out of the cells as described recently for macrophages (62, 63) and rat hepatoma cells (21). Resecreted HDL are depleted of CE and have an altered size distribution (19), and some evidence of apoE enrichment of retroendocytosed HDL was recently presented (63). Association with membraneous apoE may target HDL differently than does association with secreted and circulating apoE. The presence of a regulated, recycling receptor-like molecule involved in the binding and intracellular routing of HDL has already been discussed (20). Is apoE this molecule?