The recycling of apolipoprotein E and its amino-terminal 22 kDa fragment: evidence for multiple redundant pathways

A portion of apolipoprotein E (apoE) internalized by hepatocytes is spared degradation and is recycled. To investigate the intracellular routing of recycling apoE, primary hepatocyte cultures from LDL receptor-deﬁcient mice and mice deﬁcient in receptor-associated protein [a model of depressed expression of LDL receptor-related protein (LRP)] were incubated with human VLDL containing 125 I-labeled human recombinant apoE3. Approximately 30% of the internalized intact apoE was recycled after 4 h. The N-terminal 22 kDa fragment of apoE was also resecreted, demonstrating that this apoE domain contains sufﬁcient sequence to recycle. The 22 kDa fragment has reduced afﬁnity for lipoproteins, suggesting that apoE recycling is linked to the ability of apoE to bind directly to a recycling receptor. Finally, apoE was found to recycle equally well in the presence of brefeldin A, a drug that blocks transport from the endoplasmic reticulum and leads to collapse of the Golgi stacks. Our studies demonstrate that apoE recycling occurs 1) in the absence of the LDL receptor or under conditions of markedly reduced LRP expression; 2) when apoE lacks the carboxyl-terminal

Apolipoprotein E (apoE) is a 34 kDa protein with functions in plasma lipoprotein metabolism and intracellular lipid disposal. Extracellularly, apoE is a key mediator of the internalization of lipoprotein particles, serving as a ligand for the LDL receptor (1,2) and the LDL receptorrelated protein (LRP) (3,4). ApoE is also internalized by binding heparan sulfate proteoglycans (HSPGs) alone (5,6) or via a mechanism involving both HSPGs and LRP (7,8). Intracellularly, apoE modulates lipid metabolism (9,10) and functions in the routing of internalized lipoprotein remnants (11,12). In addition, it is involved in both the assembly (13,14) and secretion (15,16) of VLDLs. ApoE also plays a critical role in cholesterol efflux from macrophages (17)(18)(19)(20), a role that could be due to both intracellular and extracellular effects. Because apoE can easily transfer between lipoproteins and binds tightly to its receptors, we and others hypothesized and proved that a substantial amount of internalized apoE escapes lysosomal degradation and is routed into the secretory pathway.
We have established that a portion of apoE internalized on triglyceride-rich lipoproteins, as well as on HDL, is spared degradation and recycled (21)(22)(23). We also demonstrated that internalized apoE is found in the media of hepatocyte cultures from apoE Ϫ / Ϫ mice transplanted with wild-type bone marrow (22), a proof that systemic apoE can be retained by the liver cell and is eventually regurgitated. In addition, we reported that apoE was resecreted from apoE Ϫ / Ϫ hepatocytes after incubation with apoEcontaining mouse VLDL (23). The fact that apoA-I stimulated apoE resecretion from hepatocytes suggests a role for recycling apoE in HDL metabolism (23). We have also demonstrated the presence of internalized apoE in Golgi apparatus-rich fractions isolated from mouse liver (21,22), suggesting that a portion of apoE recycles through the Golgi apparatus. ApoE has also been shown to recycle in macrophages, fibroblasts, and hepatoma cell lines (23)(24)(25)(26).
The studies in this paper were undertaken to investigate the intracellular routing of recycling apoE. We explored the roles of the LDL receptor and LRP in apoE recycling, and investigated the trafficking of apoE in the presence of the Golgi-disrupting agent brefeldin A (BFA). Our studies suggest that apoE recycling occurs in the absence of the LDL receptor as well as under conditions of markedly reduced LRP expression. We also show that recycling occurs even when apoE lacks its lipoprotein binding domain, an indication that receptor binding, and not lipoprotein association, directs the routing of internalized apoE to the escape pathway. Finally, we demonstrate that apoE recycles in the absence of an intact Golgi apparatus. The data show that there are multiple recycling pathways for apoE that allow a maximum impact on cellular functions.

Animals
ICR mice were purchased from Harlan (Indianapolis, IN). RAP Ϫ / Ϫ mice were obtained from Jackson Laboratories (Bar Harbor, ME). C57BL/6 mice were purchased from Harlan and Jackson Laboratories. A colony of LDL receptor-deficient (LDLR Ϫ / Ϫ ) mice is maintained in our animal facility. All mice were kept on a 12 h light/dark cycle and were fed a normal mouse chow diet (RP5015, PMI Feeds, Inc., St. Louis, MO). Food and water were available ad libitum. All animal procedures were carried out in accordance with institutional guidelines and approved by the Institutional Animal Care and Use Committee of Vanderbilt University.

Isolation and culture of primary mouse hepatocytes
Hepatocytes were isolated as described previously (22) and cultured in 6-well (4.5 ϫ 10 5 cells/well) 35 mm collagen IVcoated dishes (Becton-Dickinson, Franklin Lakes, NJ) in media containing 1% BSA, 0.8 mM oleate, 0.02 g/ml dexamethasone, 4 mU/ml insulin, and 100 U penicillin and 100 g streptomycin/ml. Cells were used for experiments after overnight culture (16-20 h). Experiments conducted in the presence of serum used media as described above, without BSA and oleate, and including 10% fetal bovine serum (Invitrogen Corporation, Carlsbad, CA).

Isolation of plasma lipoproteins
Human VLDL was isolated at d Ͻ 1.019 g/ml from a healthy, fasted donor as described previously (23). The protocol was approved by the Vanderbilt Institutional Review Board, and informed consent was obtained prior to collecting blood. Apolipoprotein composition of VLDL was determined by SDS-PAGE as described previously (23).

Preparation of [ 125 I]apoE-VLDL
Purified apoE was produced as previously described (27) and iodinated by the chloramine T method (28). Aliquots of labeled protein were analyzed by SDS-PAGE using the NuPAGE system (Invitrogen) followed by autoradiography. [ 125 I]apoE ( ‫ف‬ 4 ϫ 10 6 cpm; or 3.5 ϫ 10 5 cpm/well) was incubated with VLDL (240 g; or 20 g/well) for 60 min at 37 Њ C. Association of labeled apoE with VLDL was confirmed by fast-protein liquid chromatography (FPLC) analysis. Aliquots (100 l) of [ 125 I]apoE-VLDL were separated using a Biologic HR Workstation (BioRad Laboratories, Hercules, CA) with a Superose 6 HR 10/30 column (Amersham Biosciences, Piscataway, NJ). Fractions (0.5 ml) were collected and cpm/fraction determined using a Beckman 5500 ␥ counter. Analysis of triglyceride was performed using an enzymatic assay adapted to microtiter plates (Raichem, San Diego, CA). The presence of apoE in the fractions was demonstrated by SDS-PAGE, followed by transfer to nitrocellulose and exposure to a Cyclone SR Screen.

Recycling experiments
Hepatocytes were incubated (pulsed) with [ 125 I]apoE-VLDL for 2 h. Media was aspirated, and cells were washed five times with ice-cold phosphate-buffered saline (PBS), incubated with heparin (10 mg/ml in PBS) for 3 min at room temperature, and washed five additional times with PBS. Fresh media without lipoproteins was added, and hepatocytes were incubated (chased) for various time periods. Media was collected at the end of the chase period, and cells were treated with heparin as described above. The heparin wash was added to the media for analysis of resecreted apoE. Following heparin treatment, cells were collected in 0.1 N NaOH. Total cell protein was determined using the bicinchoninic assay method (Pierce, Rockford, IL) using BSA as standard and modified to eliminate lipid interference (29). Degradation was defined as TCA-soluble radioactivity in the media. Recycling apoE was defined as total TCA-precipitable label in the media at the end of the chase period. Counts per minute were normalized to cell protein and data expressed as percent of total cpm/mg at each time point. Resecretion of apoE was confirmed by treatment of media with Liposorb (PHM-l Liposorb, Calbiochem-Novabiochem, La Jolla, CA) followed by SDS-PAGE and Western transfer as described previously (22). Images from cells represent approximately one-thirtieth of total cell For labeled apoE, dried gels were exposed to film overnight and developed. Lanes 1 and 3: molecular weight markers. B: [ 125 I]apoE was incubated with human VLDL for 1 h at 37ЊC and fractionated by fast-protein liquid chromatography (FPLC). Total triglyceride was determined by enzymatic assay. C: FPLC fractions were separated by SDS-PAGE, transferred to nitrocellulose, and exposed to a Cyclone SR Screen. Fraction numbers are listed below each lane.

BFA studies
These experiments were performed as described above, except that BFA (Sigma, St. Louis, MO) at varying concentrations was added to cells for 15 min prior to the pulse and included in media throughout the experiment. BFA stock solution (5 mg/ml in ethanol) was stored at Ϫ 20 Њ C. Control cells contained the same percent ethanol as the samples with the highest concentration of BFA (0.2% ethanol). For analysis of secreted apolipoproteins, media was treated with Liposorb and adsorbed proteins were separated by SDS-PAGE and transferred to nitrocellulose membranes as described (22). Membranes were blocked in 5% nonfat milk, incubated with primary antibodies, washed extensively, and incubated with horseradish peroxidase-conjugated secondary antibodies. Bands were visualized using enhanced chemiluminescence (Amersham Pharmacia Biotech).

Thrombin treatment of apoE
Human thrombin was purchased from Sigma ( ‫ف‬ 1,000 U/mg), and a stock solution (20 g/ml in 0.1 M NH 4 HCO 3 ) was made and stored at Ϫ 20 Њ C. ApoE was treated with 0, 380, 760, and 950 U thrombin/mg apoE for 20 h at room temperature. Samples were solubilized in sample buffer and separated on SDS-PAGE gels, followed by transfer to nitrocellulose membranes as described above. ApoE was visualized using the Cyclone Storage Phosphor System as described above. For recycling experiments, [ 125 I] apoE was treated with 950 U thrombin/mg apoE for 20 h at room temperature. Thrombin-treated [ 125 I]apoE was incubated with VLDL for 60 min at 37 Њ C. An aliquot (100 l) was used for analysis by FPLC. Recycling experiments were performed as described above.

Statistical analyses
One-way ANOVA was performed using GraphPad InStat (v. 3.00). Data are expressed as mean Ϯ SD. Figure 1A shows an SDS polyacrylamide gel of the purified human apoE (lane 2) as well as an autoradiogram of 125 I-labeled apoE (lane 4). After reconstitution with human VLDL, ‫ف‬ 80% of the labeled apoE was associated with the VLDL (Fig. 1B, C). In addition, autoradiograms of SDS polyacrylamide gels of the individual fractions demonstrated that the radioactivity was indeed intact apoE ( Fig. 1C).

RESULTS
In the initial studies, hepatocytes from wild-type (C57BL/6) mice were incubated with the [ 125 I]apoE-VLDL for 2 h, washed, and the resecretion of labeled apoE monitored. The appearance of intact (TCA-precipitable), la- Primary hepatocytes from C57BL/6 mice were isolated and cultured in media containing 1% BSA and 0.8 mM oleic acid or 10% FBS. Cells were pulsed with [ 125 I]apoE-VLDL for 2 h, washed extensively, treated with heparin, and incubated in the appropriate media without lipoproteins for 0, 60, and 240 min. Radioactivity (cpm/mg cell protein) in cells, TCA-precipitable media fraction (Recycled), and TCA-soluble media fraction (Degraded) was determined and expressed as percent of total uptake at each time point. Data are shown as mean Ϯ SD; n ϭ 3 separate experiments.

Fig. 2. Recycling of [ 125 I]apoE-VLDL in primary hepatocytes
from wild-type mice. Hepatocytes from C57BL/6 mice were pulsed for 2 h with [ 125 I]apoE-VLDL, washed extensively, treated with heparin, and incubated in media without lipoproteins for the times indicated. A: Radioactivity (cpm/mg cell protein) in cells, TCA-precipitable media fraction (recycled), and TCA-soluble media fraction (degraded) was determined and is expressed as percent of total uptake at each time point. Data are shown as mean Ϯ SD; n ϭ 3 separate experiments. Inverted triangle, cellular apoE; square, recycled apoE; triangle, degraded apoE. B: Media lipoproteins were adsorbed using Liposorb, and cells were solubilized in 0.1 N NaOH. Proteins were separated on SDS-PAGE gels, transferred to nitrocellulose, and exposed to a Cyclone SR Screen. Aliquots of [ 125 I]apoE-VLDL were used as controls for apoE and show the composition of the starting material. Images from cells represent approximately one-thirtieth of total cell protein.
by guest, on  www.jlr.org Downloaded from beled apoE was time dependent, with ‫ف‬ 30% of the internalized apolipoprotein being recycled after 240 min ( Fig.  2A ). The levels of apoE within the cell seemed to stabilize after 60 min of chase (Fig. 2B). To determine whether the presence of serum altered apoE recycling, the experiments with wild-type hepatocytes were repeated using serum-containing media. Although the mass of apoE internalized and resecreted was greater in the presence of serum, there were no changes in the percent of internalized apoE that was recycled or degraded ( Table 1 ). Consequently, all subsequent studies were carried out in serumfree media.
To determine whether the LDL receptor is critical for the recycling of apoE, primary hepatocytes from LDLR Ϫ / Ϫ mice were incubated with 125 I-labeled apoE and resecre-tion of apoE monitored ( Fig. 3A ). The uptake of labeled apoE in hepatocytes from LDLR Ϫ / Ϫ mice was reduced by nearly 50% when compared with control cells (LDLR Ϫ / Ϫ , 7,699 Ϯ 2,154 cpm/mg cell protein; control, 16,403 Ϯ 4,482 cpm/mg cell protein); however, as observed with wild-type cells, ‫%03ف‬ of internalized apoE was recovered in the media after 240 min (Fig. 3A). To investigate the role of LRP in apoE recycling, similar experiments were performed using hepatocytes from mice lacking receptorassociated protein (RAP), an intracellular chaperone essential for proper LRP trafficking (30)(31)(32). Hepatic LRP expression is decreased by 75% in RAP Ϫ/Ϫ mice (33); however, the uptake of labeled apoE in the RAP Ϫ/Ϫ hepatocytes was not significantly different from that of control cells (RAP Ϫ/Ϫ , 6,716 Ϯ 925 cpm/mg cell protein; control, 9,777 Ϯ 5304 cpm/mg cell protein). The relative amount of labeled apoE resecreted was similar to that found with hepatocytes from both wild-type and LDLR Ϫ/Ϫ mice (Figs. 2, 3, Table 2). Furthermore, the cellular levels of apoE remained constant after 60 min in hepatocytes from all three genotypes (Figs. 2A, 3). Degradation of apoE was similar in hepatocytes from wild-type and RAP Ϫ/Ϫ mice; however, it was significantly increased in hepatocytes from LDLR Ϫ/Ϫ mice (Figs. 2, 3, Table 2).
To investigate the cellular routing of recycling apoE, primary hepatocytes from wild-type mice were incubated with [ 125 I]apoE-VLDL in the presence of increasing concentrations of BFA. BFA blocks trafficking from the endoplasmic reticulum to the Golgi apparatus, which ultimately leads to dissolution of the Golgi complex (34)(35)(36). ApoE recycled equally well in the presence of all concentrations of BFA (Fig. 4). To determine whether the concentrations of BFA used were effective in blocking protein secretion, we analyzed the accumulation of endogenous apolipoproteins in the media by means of SDS-PAGE and immunoblotting (Fig. 5). BFA (2.5 g/ml) decreased secretion of apoB and apoE by ‫%58ف‬ and that of apoA-I by 75%. Higher concentrations of BFA (100 g/ml) had no additional effect on the secretion of any of the apolipoproteins (data not shown).
In addition to the recycling of full-length apoE, we noted the presence of a radiolabeled 22 kDa apoE frag-    (Fig. 2B). This fragment represented ‫%01ف‬ of the radioactivity found with apoE in the starting material and 50% of radioactivity in the resecreted material (Fig. 2B). The increasing proportion of 22 kDa protein in the media as a function of chase time may be the result of proteolysis of the 34 kDa protein by a cellular or secreted protease. Thrombin digestion of intact apoE produces a 22 kDa fragment. This fragment contains the N-terminal globular domain of apoE (amino acids  and the heparin and receptor binding sites, but it lacks the major lipoprotein binding region (1). We designed experiments to address the pattern of recycling of the N-terminal fragment in the absence of intact apoE. ApoE was completely digested at thrombin concentrations of Ն380 U/mg apoE (Fig. 6A). When the thrombin-digested apoE was reconstituted with VLDL, only a small portion of the apoE fragments associated with the VLDL particles. Recycling studies utilizing the thrombin-treated [ 125 I] apoE-VLDL demonstrated that the fragments were internalized and recycled (Fig. 6B).

DISCUSSION
In this study, we explored the internalization and trafficking of recycling apoE from VLDL in primary hepatocytes. We found that apoE recycling is unaffected by the absence of the LDL receptor or under conditions of reduced LRP expression. ApoE also recycles equally well in the presence of BFA, indicating either that an intact Golgi apparatus is not required for apoE recycling, or that a redundant mechanism will maintain optimal recycling even when the Golgi route is interrupted. The 22 kDa thrombolytic fragment of apoE recycled efficiently, suggesting that the N-terminus contains sufficient sequence to recycle and that the association with lipoproteins is not a necessary element for the activation of this pathway.
Recycling was investigated using 125 I-labeled recombinant apoE that was reconstituted with human VLDL (Fig.  1). We used this approach as opposed to 125 I-labeling of intact VLDL so that apoE recycling could be easily monitored without interference from other labeled apolipoproteins. Pulse-chase experiments utilizing hepatocytes from wild-type mice demonstrated that the appearance of recycling apoE in the media was time dependent (Fig. 2). Cellular levels of labeled apoE stabilized after 60 min, providing evidence for retention of internalized apoE. Previous studies using primary hepatocytes from apoE Ϫ/Ϫ mice reconstituted with bone marrow from wild-type mice also suggested that internalized apoE remained within the cell for an extended period of time (22). The delay in clearance of apoE in these studies suggests an intracellular reserve of apoE that is resistant to degradation, or a pool of continually recycling apoE.
The present study suggests that apoE recycles effectively Primary hepatocytes were isolated from C57BL/6 mice and incubated with 0, 2.5, 5, and 10 g/ml BFA for 15 min. Hepatocytes were then pulsed for 2 h with [ 125 I]apoE-VLDL, washed extensively, treated with heparin, and incubated in media without lipoproteins for 0, 60, and 240 min in the continued presence or absence of BFA. Data were calculated as described in the legend to Fig. 2, and are shown as mean Ϯ SD; n ϭ 3 separate experiments. in the absence of the LDL receptor. Although uptake of [ 125 I]apoE-VLDL was reduced in hepatocytes from LDLR Ϫ/Ϫ mice, as compared with wild-type controls, the relative mass of apoE recycled was unchanged (Fig. 3A, Table 2). We previously recovered labeled apoE in hepatic Golgi-rich fractions from LDLR Ϫ/Ϫ mice after injection of 125 I-labeled mouse VLDL (21), suggesting that internalized apoE is spared degradation even in the absence of the LDL receptor. Similarly, Heeren et al. (25) demonstrated that although uptake of 125 I-labeled remnant lipoproteins by fibroblasts from patients with familial hypercholesterolemia was approximately half of uptake in normal fibroblasts, the relative amount of resecreted labeled apolipoproteins was the same. However, Rensen et al. (37) suggested that apoE recycles through an LDL receptor-mediated pathway. Interestingly, in our studies, degradation of apoE was increased in the hepatocytes isolated from LDLR Ϫ/Ϫ mice, as compared with both wildtype and RAP Ϫ/Ϫ mice ( Table 2), suggesting that alternate pathways may exist for routing of lipoproteins via the LDL receptor or LRP. Because recycling does not appear to be dependent on the LDL receptor, we focused on the role of LRP in inter-nalization and recycling of apoE. LRP binds at least 30 ligands and functions in many processes, including lipoprotein metabolism, regulation of proteolytic enzymes and inhibitors, entry of pathogens, and cell signaling (38). Deletion of LRP in mice is embryonic lethal (39). RAP is a 39 kDa protein that serves as a chaperone for LRP and escorts the receptor to the cell surface (30)(31)(32). Hepatic expression of LRP is reduced by 75% in RAP Ϫ/Ϫ mice (33). Our data indicate that a severe reduction in LRP expression has no effect on apoE recycling (Fig. 3B, Table 2). Even though it is possible that RAP Ϫ/Ϫ mice could have sufficient LRP expression for normal apoE recycling, it is more likely that the process of apoE recycling is influenced by both the LDL receptor and LRP, and that in the absence of one receptor, the other compensates.
BFA was used to investigate the effects of disruption of intracellular trafficking on apoE recycling. BFA is a fungal metabolite that blocks membrane export out of the endoplasmic reticulum (34,35) by preventing association of coatamer proteins on transport vesicles (40,41), which ultimately results in dissolution of the Golgi apparatus (36). Interestingly, incubation of hepatocytes with BFA concentrations up to 10 g/ml did not alter recycling (Fig. 4). Previous studies in our laboratory indicated that recycling apoE is routed to the Golgi apparatus (21,22). To explain this apparent discrepancy, we looked at the possibility that BFA treatment of hepatocytes did not completely disrupt the Golgi. Therefore we monitored secretion of apoB, apoE, and apoA-I from BFA-treated hepatocytes. Although secretion of apoB, apoE, and apoA-I was not completely blocked, it was dramatically reduced (Fig. 5), suggesting that there was a substantial effect of BFA treatment on the Golgi. Sparks et al. (42) reported that 10 g/ml BFA greatly reduced but did not completely inhibit secretion of apoB, albumin, and transferrin in primary rat hepatocytes. Thus, it is possible that in the presence of BFA, apoE continues to recycle in a quantitatively unmodified fashion through remnants of the Golgi apparatus. However, the results also suggest that there are multiple intracellular trafficking routes for recycling apoE. Indeed, studies by Stoorvogel and coworkers (43,44) demonstrated that BFA treatment resulted in rerouting of recycling transferrin from a perinuclear recycling compartment to a direct recycling route from early endosomes to the plasma membrane. There may be a rapid recycling pathway in which apoE is internalized, trafficked close to the cell periphery, and resecreted, similar to that described for transferrin (45)(46)(47). Evidence for a slower recycling pathway has been demonstrated by the retention of recycling apoE in the cell for an extended period of time (Figs. 2, 3) (22,23). It is possible that the Golgi apparatus is involved in this slower recycling pathway and that BFA treatment directs apoE recycling away from the Golgi apparatus and into the direct pathway from endosomes to the plasma membrane.
The appearance of a lower-molecular-weight form of labeled apoE in the media and cells suggests that the N-terminal domain of apoE contains sufficient sequence to recycle (Figs. 2B, 6B). ApoE has two distinct domains linked by a hinge region (48). Thrombin treatment results in cleavage of apoE into a 22 kDa fragment and a 10 kDa fragment, representing the N-and C-terminal domains of the protein, respectively (1). Although the C-terminal domain contains the lipid binding domain (48)(49)(50), the N-terminal domain has been shown to have some lipid binding affinity (1,51). After thrombin digestion of [ 125 I] apoE and reconstitution with VLDL, only a minor portion of apoE fragments associated with VLDL. However, the 22 kDa fragment was recovered in both the cells and media (Fig. 6). In vivo, the 22 kDa fragment of apoE is not found in plasma, probably because the cleavage sites are protected from proteases by the glycosylation site at threonine-194 (52,53). However, apoE fragments of ‫22ف‬ kDa have been found in brain and cerebrospinal fluid (54,55), and both synthetic fragments and the 22 kDa N-terminal domain produced by thrombin digestion of apoE have been shown to be neurotoxic (54,(56)(57)(58). In addition, the fact that apoE recycles in fetal human brain cultures (59) suggests a role for recycling of apoE fragments in neurological disease. At the very least, the efficient recycling of the N-terminal domain of apoE demonstrates that apoE does not need to be bound to lipoproteins in order to engage the escape route.
Our finding that ‫%03ف‬ of internalized [ 125 I]apoE is resecreted after 240 min of chase ( Fig. 2A) is different from earlier estimates that as much as 60% of internalized apoE may be recycled in hepatocytes from apoE Ϫ/Ϫ mice transplanted with bone marrow from wild-type mice (22). The models used to investigate recycling may explain these differences. In the bone marrow transplantation (BMT) model, the plasma apoE derives exclusively from macrophages. Macrophages secrete apoE on small, nascent HDL-like discoidal particles enriched in phospholipids (60)(61)(62); however, apoE is capable of exchanging between lipoprotein classes in the plasma. Therefore apoE may be presented to hepatocytes on VLDL, HDL, or lipoprotein remnants. The present study utilized [ 125 I]apoE reconstituted with VLDL isolated from a healthy, 16-h-fasted human donor at d Ͻ 1.019 g/ml. Consequently, there would be few remnants and no HDL in the experimental system. We have shown that apoE can recycle from HDL (23), and Rensen et al. (37) and Heeren et al. (24)(25)(26) have shown that apoE recycles from large, triglyceride-rich remnants. It is possible that the presentation of apoE to cells on VLDL in this study, as compared with HDL, remnants, and VLDL in the BMT studies, alters the relative proportion of apoE that is recycled. Furthermore, the relative proportion of internalized apoE could be different in the presence of hepatic apoE production.
In summary, we investigated the internalization and trafficking of recycling apoE from VLDL in primary hepatocytes. The observation that apoE recycling is not dependent on entry point, combined with previous data that apoE recycles from both VLDL and HDL, suggests that the apoE protein determines recycling, rather than how it is presented to the cell. However, the pathway of apoE recycling could be different depending on the lipoprotein particle or internalization mechanism. In addition, we provide evidence that apoE recycles in the absence of an intact Golgi apparatus and that binding of apoE to the lipoprotein is not necessary for recycling to occur. In light of the implications of apoE fragments as neurotoxins, our discovery of efficient recycling of the 22 kDa fragment of apoE allows us to speculate on a role for apoE recycling in neurological disease.