Rapid Endocytosis of the Low Density Lipoprotein Receptor-related Protein Modulates Cell Surface Distribution and Processing of the β-Amyloid Precursor Protein*

The low density lipoprotein receptor-related protein (LRP) is a ∼600-kDa multifunctional endocytic receptor that is highly expressed in the brain. LRP and its ligands apolipoprotein E, α2-macroglobulin, and β-amyloid precursor protein (APP), are genetically linked to Alzheimer disease and are found in characteristic plaque deposits in brains of patients with Alzheimer disease. To identify which extracellular domains of LRP interact with APP, we used minireceptors of each of the individual LRP ligand binding domains and assessed their ability to bind and degrade a soluble APP fragment. LRP minireceptors containing ligand binding domains II and IV, but not I or III, interacted with APP. To test whether APP trafficking is directly related to the rapid endocytosis of LRP, we generated stable Chinese hamster ovary cell lines expressing either a wild-type LRP minireceptor or its endocytosis mutants. Chinese hamster ovary cells stably expressing wild-type LRP minireceptor had less cell surface APP than pcDNA3 vector-transfected cells, whereas those stably expressing endocytosis-defective LRP minireceptors accumulated APP at the cell surface. We also found that the steady-state levels of the amyloid β-peptides (Aβ) is dictated by the relative expression levels of APP and LRP, probably reflecting the dual roles of LRP in both Aβ production and clearance. Together, these data establish a relationship between LRP rapid endocytosis and APP trafficking and proteolytic processing to generate Aβ.

Alzheimer disease (AD) 1 is one of the leading causes of dementia in elderly persons. Although the cause of AD dementia is unknown, it is probably related to the deposition of characteristic extracellular amyloid plaques and intracellular neurofibrillary tangles found in the brains of patients with AD. Amyloid plaques are composed of fibrillar aggregates of amy-loid-␤ peptides (A␤), which are derived from the processing of a ϳ120-kDa transmembrane protein known as ␤-amyloid precursor protein (APP) (1). APP, which exists in three main isoforms (APP695, APP751, and APP770), can follow two alternate posttranslational processing pathways. In the amyloidogenic pathway, APP is cleaved first at a ␤-secretase site by the enzyme ␤-site APP-cleaving enzyme and subsequently by a ␥-secretase within its intramembrane region to release the A␤ peptide (2). In the non-amyloidogenic pathway, APP is processed by an ␣-secretase that clips within the A␤ region, resulting in the release of a soluble APP fragment (sAPP␣), which may function in blood coagulation and neurite outgrowth (3).
Several studies have shown that APP internalization through the endocytic pathway can lead to A␤ generation (4 -7). The APP tail has a YENPTY motif similar to tyrosine-based NPXY and YXXØ endocytic motifs (where X can be any amino acid and Ø is an amino acid with a bulky hydrophobic group) found in other endocytic receptors (8). A decrease in APP endocytosis caused either by mutations within its endocytosis motif or by potassium depletion, which inhibits the formation of clathrin-coated pits, results in an accumulation of APP at the cell surface and favors the non-amyloidogenic processing of APP (5).
The low density lipoprotein receptor-related protein (LRP) interacts with APP, and its expression facilitates APP processing via the amyloidogenic pathway. Cells overexpressing APP that are transiently transfected with LRP produce higher levels of A␤ and lower levels of sAPP␣ than mock-transfected cells (9). LRP may influence the processing of APP through both extracellular and intracellular interactions (10 -12). Expression of the C-terminal fragment of LRP increases A␤ production and decreases the release of sAPP␣, similar to the expression of full-length receptor (13). Blocking the extracellular interaction between APP and LRP with receptor-associated protein (RAP) results in a decrease in A␤ levels and an increase in cell surface APP and sAPP␣ levels (9). Consistent with these in vitro studies, overexpression of a functional LRP minireceptor in the PDAPP amyloid mouse model increases soluble brain A␤ in aged mice (14). We postulate that LRP may regulate APP processing and A␤ levels by its very rapid rate of endocytosis (t 1/2 Ͻ 0.5 min) (15). Through extracellular and intracellular interactions with APP (9,10,12,16), LRP could facilitate the trafficking of APP through the endocytic pathway such that its availability to ␤and ␥-secretases is favored over its accessibility to ␣-secretases at the cell surface (17).
In this study, we found that minireceptors containing domains II and IV of LRP can bind and degrade a soluble APP fragment. Using endocytosis-deficient LRP minireceptors containing domain IV, we also found that the endocytosis rate of LRP influences the cell surface level of APP and A␤ secretion in the media. Our data establish a relationship between the rapid endocytosis rate of LRP and cellular distribution and trafficking of APP.

MATERIALS AND METHODS
Antibodies and Reagents-The C-terminal APP antibody BC1 (recognizes amino acids 704 -730) was the kind gift of Barbara Cordell (Scios). The monoclonal APP antibody 6E10 raised against residues 1-17 of A␤ was purchased from Signet. A previously described monoclonal anti-HA antibody was used for FACS analysis of LRP minireceptors (15). 6E10, 22C11 (RDI, Inc.), and 8E5 (gift from Kelly Bales, Lilly Research Laboratories) were used for FACS analysis of APP. Fluorescein isothiocyanate-conjugated goat anti-mouse IgG was obtained from BD Biosciences. Soluble APP770␣ (sAPP770␣) was purified from the media of human embryonic kidney cells stably overexpressing APP770 as described previously (18). Soluble APP695␣ (sAPP695␣) and holotransferrin were purchased from Sigma. For ELISA assays, antibodies 266 (recognizes A␤ 13-28), 21F12 (recognizes A␤1-42), and 2G3 (recognizes A␤1-40) were the kind gifts of Kelly Bales (Lilly Research Laboratories). Carrier-free Na 125 I was purchased from PerkinElmer Life and Analytical Sciences. Proteins were iodinated using the IODO-GEN method as described previously (15).
cDNA Construction-Generation of LRP minireceptors has been described previously (15). The LRP minireceptors were designated as mLRP1, mLRP2, mLRP3, or mLPR4, with "m" representing "membrane-containing" and the number representing the cluster of cysteinerich ligand-binding repeats (beginning from the amino terminus). A chimeric mLRP4-APPtail minireceptor was generated by excising the cytoplasmic tail of mLRP4 plasmid and ligating a PCR fragment of the APP tail. Constructs were sequenced and their expression tested by in vitro translation using the TNT coupled reticulocyte lysate system in vitro translation kit (Promega) according to the manufacturer's instructions. Generation of mLRP4 endocytosis mutants has been described previously (15).
Cell Culture and Transfection-CHO LRP-null cells were transfected using the calcium phosphate precipitation method or Lipofectamine 2000 (Invitrogen). Stably transfected cells were selected in 700 g/ml geneticin (and 750 g/ml zeocin for APP770/minireceptor cell lines), and stable clones were maintained in 350 g/ml geneticin (and 350 g/ml zeocin for APP770/LRP minireceptor doubly transfected cell lines).
Kinetic Analysis of Endocytosis-Kinetic analysis of receptor-mediated endocytosis was carried out as described previously (15). In brief, stably transfected cells were plated at a density of 2 ϫ 10 5 cells/well in a 12-well plate and used after overnight culture. Cells were rinsed twice with cold PBS and then incubated in 0.5 ml of ice-cold ligand binding buffer (minimal Eagle's medium containing 0.6% bovine serum albumin) with 2-5 nM 125 I-RAP, 125 I-anti-HA IgG, or 125 I-1G7 anti-APP IgG. The binding of iodinated proteins was carried out at 4°C for 60 min with gentle rocking. Unbound ligand was removed by washing cell monolayers three times with cold binding buffer. Ice-cold stop/strip solution (0.2 M acetic acid, pH 2.6, and 0.1 M NaCl or PBS, pH 2.0) was added to one set of plates without warming up and kept on ice. The remaining plates were then placed in a 37°C water bath, and 0.5 ml of prewarmed ligand binding buffer was quickly added to the cell monolayers to initiate internalization. After each time point, the plates were placed on ice, and the ligand binding buffer was replaced with ice-cold strip/stop solution. Ligand that remained on the cell surface was stripped by incubation of cell monolayers with ice-cold stop/strip solution for a total of 10 min and counted. The sum of internalized ligand plus those on the surface after each assay was used as the maximum potential internalization. The fraction of internalized ligand at each time point was calculated and plotted. For degradation assay, cells were plated as described and incubated with 1 nM 125 I-sAPP770␣ in serumfree media for 6 h at 37°C in the absence or presence of 1 M RAP. Media was then collected and subjected to trichloroacetic acid precipitation. Cells were washed and lysed for protein determination.
Co-immunoprecipitation-Cells were lysed in phosphate-buffered saline, pH 7.4, containing 1% Triton X-100 and protease inhibitors. Cell extracts were precleared for at least 2 h with Protein-A agarose beads and then incubated overnight with antibody at 4°C. Immunocomplexes were precipitated with Protein-A agarose beads for 45 min, washed three times with phosphate-buffered saline, and boiled in SDS sample buffer containing ␤-mercaptoethanol. The supernatants were subjected to SDS-PAGE and Western blotting.
Immunoblotting-Proteins were separated on 6% or 7.5% SDS-PAGE gels under reducing conditions and transferred to Immobilon-P membrane (Millipore). Membranes were blocked in Tris-buffered saline containing 0.1% Tween 20 and 5% nonfat dried milk and subjected to incubation with primary and secondary antibodies. For detection of APP, 6E10 antibody was used. For detection of LRP minireceptors, membranes were probed with anti-85kD tail fragment of LRP or anti-HA antibodies. Goat anti-rabbit or anti-mouse IgGs conjugated to horseradish peroxidase were developed with ECL Plus reagent (Amersham Biosciences) and exposed to film. For quantitation of minireceptor expression in each cell line, equal amounts of protein were separated and transferred as above. After detection with ECL Plus reagent (Amersham Biosciences), blots were analyzed using the variable mode imager Typhoon 9410 (Amersham Biosciences).
Flow Cytometry Measurements-For cell surface APP analysis, CHO cells were first detached by incubation with non-enzymatic cell dissociation solution (Sigma). To measure total levels of APP, cells were treated with 0.1% saponin. Successive incubations with anti-APP antibody 6E10 (Signet) or 22C11 (RDI, Inc.) (50 g/ml) and goat anti-mouse Ig-fluorescein isothiocyanate were carried out at 4°C for 1 h. As a control, background fluorescence intensity was assessed in the absence of primary antibody. All measurements were performed on a FACSCalibur (BD Biosciences) equipped with an argon ion laser. Laser excitation of 488 nm for fluorescein isothiocyanate was used. Ten thousand cells from each sample were analyzed. Histograms were generated using the CellQuest software; median values, after subtraction of controls, were compared among samples.
A␤ Enzyme-linked Immunosorbent Assay of Conditioned Media-Cells were plated into 6-well plates at a density of 1 ϫ 10 5 cells/well. The next day, cells were washed two times with serum-free media and incubated in low serum media (containing 1% fetal bovine serum). Media were collected after 24 -48 h of conditioning with the addition of proteinase inhibitors and then spun at 20,200 ϫ g for 5 min to remove cellular debris. A␤ in the media was analyzed by ELISA adapted from previous studies (20). A␤ in the conditioned media was captured with antibody 2G3 or 21F12 for A␤40 and A␤42, respectively, and subsequently detected with biotinylated 266 or 3D6 antibody. Endogenous A␤ values were normalized to the amount of total protein in cell lysates.
Human A␤ values were normalized to the amount of APP expression determined by Western blot.
Human A␤ Clearance-Conditioned media from primary neurons derived from PDAPP mice were used as source of human A␤. PDAPP mice overexpress mutated human APP V717F under the platelet-derived growth factor promoter (21). Diluted media were added to CHO-pcDNA3 and CHO-mLRP4 cells (ϳ500 pg of total A␤), or blank wells without cells. After 24 h, media were collected with the addition of proteinase inhibitors and then spun at 20,200 ϫ g for 5 min to remove cellular debris. Human A␤ remaining in the conditioned media was analyzed by ELISA using capturing antibody 2G3 for A␤40 or 21F12 for A␤42, and subsequent detection with biotinylated 3D6 antibody, specific for human A␤. Remaining A␤ in the media incubated with cells was subtracted from the amount of A␤ present in the media from empty wells and normalized to the amount of total protein in cell lysates. Values were normalized to that of CHO-pcDNA3 cells.

RESULTS
To determine which extracellular domains of LRP are important for an interaction with APP, we assessed the degradation of a soluble purified APP770 fragment (sAPP770␣) by cells expressing minireceptors of LRP containing each individual ligand-binding domain (CHO-mLRP1, mLRP2, mLRP3, or mLRP4) (22). To determine receptor-specific degradation, 125 I-sAPP770␣ degradation was assessed in the absence or presence of excess amounts of the LRP antagonist RAP. RAP binds to domains II, III, and IV of LRP with high affinity (23). We found that 125 I-sAPP770␣ was specifically degraded by CHO-mLRP2 and CHO-mLRP4 cells but not CHO-mLRP1 or CHO-mLRP3 cells (Fig. 1A). We detected similar levels of the 85-kDa subunit of the mature minireceptors among these cell lines (Fig. 1B) suggesting that the differences in the ability to degrade 125 I-sAPP770␣ were not caused by differences in the expression levels of these minireceptors. In addition, mLRP2 and mLRP4, two LRP domain minireceptors that degraded 125 I-sAPP770␣, exhibited similar rates of endocytosis (Fig. 1C). None of the cell lines tested significantly degraded 125 I-sAPP695␣, which lacks a KPI domain (data not shown). These findings are in agree-ment with previous reports that full-length LRP degrades only KPI-containing sAPP␣ (18). These results also suggest that an extracellular interaction between LRP and APP is mediated through the second and fourth ligand-binding domains of LRP and the KPI domain of APP.
We hypothesize that LRP, with its fast rate of endocytosis, influences the trafficking of APP through the endocytic pathway. To directly compare the endocytosis rates of LRP and APP, we generated a stable CHO cell line expressing a chimeric receptor containing the fourth ligand binding domain and transmembrane region of LRP and the APP tail (mLRP4-APPtail) (Fig. 2A). The APP tail has a YENP tetrapeptide that is important for its endocytosis and can mediate endocytosis when fused to the human transferrin receptor (7, 24). Using 125 I-RAP to measure the endocytosis of the mLRP4-APPtail chimeric receptor compared with the wild-type LRP minireceptor (mLRP4), we found mLRP4-APPtail had a much slower rate of endocytosis (ϳ10%/min) than mLRP4 (Ͼ70%/min) (Fig. 2B). These results support our hypothesis that LRP endocytosis could facilitate APP internalization and processing to the amyloidogenic pathway.
Because APP and LRP interact through extracellular interactions mediated by the KPI domain of APP and also intracellular interactions via the adaptor protein FE65 (10 -13,16), we tested whether endogenous APP within CHO LRP-null cells could interact with overexpressed LRP domain IV minireceptors. In a previous study from our laboratory, mLRP4 minireceptors containing site-directed mutations within putative cytoplasmic endocytosis motifs were generated and stably transfected into CHO LRP-null cells (15). Minireceptors bearing mutations within the YXXL motif (Y63A and L66A) and distal di-leucine repeat (L86A/L87A) (Fig. 3A) have a reduced rate of endocytosis compared with wild-type minireceptor (15). To determine whether an extracellular interaction could occur between the minireceptors and endogenous APP, we first examined whether CHO cell APP is KPI-containing APP similar to APP751 or non KPI-containing APP similar to APP695. Lysates from cells transiently transfected with hAPP751 or hAPP695 were compared with CHO LRP-null lysates. Immunoblot of lysates for APP revealed that CHO LRP-null cells contained an APP immunoreactive band that migrated at the same molecular size as APP751 and was larger than the molecular size of APP695, suggesting that endogenous CHO cell APP was primarily KPI-containing APP (Fig. 3B).
Next, we determined whether endogenous APP could interact with mLRP4 and its endocytosis mutants by co-immunoprecipitation. Immunoblot of APP immunoprecipitates with anti-HA antibody revealed immunoreactive bands that migrate at the same molecular sizes as the furin-cleaved processed form (ϳ120 kDa) and the unprocessed form (ϳ210 kDa) of LRP minireceptors (Fig. 3C). No anti-APP or anti-HA signals were detected in lysates immunoprecipitated with control IgG antibody (data not shown). We also detected an interaction between APP and mLRP4Tailess. Because the major form of APP in CHO cells contained the KPI domain that can bridge an extracellular interaction with LRP, it is probably that the association between LRP minireceptors and APP are predominantly extracellular and independent of cytoplasmic interactions.
After confirming an interaction between endogenous APP and the LRP minireceptors, we tested whether the rate of LRP endocytosis could influence the trafficking of APP by comparing cell surface APP between cells expressing mLRP4 and mLRP4 endocytosis mutants by FACS analysis. We detected significantly less cell surface APP in CHO-mLRP4 cells compared with CHO-pcDNA3 cells (Fig. 4A). In support of our hypothesis, cells expressing endocytosis mutants of mLRP4, mLRP4(Y63A), mLRP4(L66A), mLRP4(Y63A, L86A/L87A), and mLRP4Tailess exhibited Ն3-fold more cell surface APP compared with CHO-mLRP4 wild-type cells (Fig. 4B). To confirm that the effect of LRP overexpression on APP distribution was specific to APP, we assessed for changes in the amount of another cell surface receptor, the transferrin receptor, within these cell lines. The amount of bound iodinated transferrin to the transferrin receptor did not change because of LRP minireceptor expression, indicating that alterations in the cell surface distribution of APP were specific to an interaction between APP and LRP minireceptors (data not shown). To determine whether LRP minireceptor expression mediated changes in the total cellular level of APP, we measured total levels of APP in CHO stable cell lines by treatment with saponin (a detergent) to permeabilize cell membranes, followed by FACS analysis of APP, and by quantitative Western blotting for APP. We found that expression of mLRP4 or its endocytosis mutants did not alter total cellular APP. By both FACS analysis (Fig. 4C, top) and immunoblot (Fig. 4C, bottom), total APP levels were similar in pcDNA3, mLRP4, and endocytosis mutants of mLRP4. Taken together, these results suggest that changes in cell sur-face APP were caused solely by altered trafficking of APPmediated by the expression of LRP minireceptors.
Mutations in the LRP tail that do not influence the endocytosis rate of LRP may also change APP trafficking because of potential structural alterations that may interfere with their ability to interact with APP. Therefore, we measured levels of cell surface APP in CHO cells stably transfected with LRP minireceptors containing mutations that had no change in endocytosis rate compared with wild-type mLRP4. CHO cells expressing LRP minireceptors with mutations within the proximal NPXY (N26A and Y29A), proximal di-leucine motifs (L43A/L44A), or the distal NPXY motif (N60A) had no significant effects on cell-surface distribution of APP compared with cells expressing wild-type LRP minireceptor (Fig. 5). These data indicate that the fast endocytosis rate of LRP is the major contributor to alterations in the cellular distribution of APP.
The above results suggest that LRP endocytosis decreases cell surface APP levels by increasing its endocytosis rate. Be- cause A␤ has been shown to be produced within endocytic compartments, we assessed the effect of LRP minireceptor expression on APP processing to A␤ by measuring endogenous levels of A␤40 and A␤42 in conditioned media of cells expressing pcDNA3 or mLRP4 by ELISA. We found a trend toward less A␤40 and significantly less A␤42 in 48-h conditioned media from CHO-mLRP4 cells compared with CHO-pcDNA3 cells (Fig. 6A).
Based on our hypothesis and previous published studies (9, 13), we expected to find increased secreted A␤ levels in CHO-mLRP4 cells. Because CHO-mLRP4 cell media had less A␤ than that of the vector control cells, we considered the possibility that the A␤ clearance pathway mediated by the LRP minireceptor was dominant over the production pathway. To test whether A␤ could be cleared by CHO-mLRP4 cells, we incubated CHO-pcDNA3 and CHO-mLRP4 cells with human A␤ obtained from conditioned media of PDAPP mouse primary cultured neurons. To measure the disappearance of human A␤ without the interference of endogenously produced CHO cell hamster A␤, we used an ELISA that specifically detects human A␤ but not the endogenously produced hamster A␤. We found that cells expressing mLRP4 cleared significantly more human A␤40 and A␤42 than CHO-pcDNA3 cell lines (Fig. 6B). These findings suggest that the lower levels of endogenous A␤ detected in CHO-mLRP4 cells could be caused by increased clearance mediated by the overexpressed LRP minireceptor.
Because previous studies reporting increased A␤ levels with LRP expression used cells overexpressing human APP (9, 13), we generated CHO LRP-null stable cell lines that overexpressed human APP770 together with pcDNA3 or mLRP4. After selecting stable clones with equal expression of APP as detected by Western blot analysis, we measured steady state levels of human A␤40 and A␤42 in conditioned media of CHO-pcDNA3 cells compared with CHO-mLRP4 cells. Opposite to our findings with endogenous A␤, but in agreement with previously published reports, we found ϳ3-fold more A␤40 and A␤42 levels in CHO-mLRP4 cells compared with CHO-pcDNA3 with overexpression of APP770 (Fig. 6C). DISCUSSION Several lines of evidence suggest that LRP has a role in AD pathogenesis either by interacting with APP and influencing its processing (9,13,16), or as a clearance receptor for A␤ (25)(26)(27). The objective of the current study was to investigate how LRP expression and rate of endocytosis influence APP cellular distribution and processing to A␤. Using LRP-deficient CHO cells overexpressing functional LRP minireceptors, we found that the extracellular domains II and IV of LRP can both individually bind and degrade a soluble form of the KPI-containing APP (sAPP770␣), but not a soluble KPI-lacking APP (sAPP695␣). In addition, the cellular degradation of soluble APP770 by CHO-mLRP2 and CHO-mLRP4 was blocked by excess RAP, an LRP antagonist. These results are consistent with previous findings that a KPI domain is necessary for the interaction between APP and LRP (18). Similar to our findings, it has been shown that another KPI-containing ligand of LRP, tissue factor pathway inhibitor, specifically interacts with ligand-binding domains II and IV of LRP but not domains I or III (28). Most LRP ligands have been shown to bind domains II and IV of LRP, and no ligands have been reported to bind to domain I (29). Although the predominant APP isoform in neurons does not contain a KPI domain (APP695), the proportion of KPIcontaining forms of APP is significantly elevated in AD brains (30). Therefore, it is possible that increased extracellular interactions between LRP and KPI-containing isoforms of APP could contribute to the A␤ accumulation seen in AD.
We also established that the LRP minireceptor containing the ligand binding domain IV of LRP (mLRP4) interacts with endogenous CHO cell APP and that this interaction leads to decreased cell surface APP levels compared with pcDNA3 vector-transfected cells. In addition, cells expressing endocytosisdeficient mLRP4, which were able to interact with endogenous CHO cell APP, showed an accumulation of APP at the cell surface. These results are consistent with our previous findings showing that overexpression of a minireceptor of LRP1B, a member of the LDLR family that has high homology to LRP but significantly slower rate of endocytosis, interacts with APP and leads to its cell surface accumulation (31). More importantly, mLRP1B4-expressing CHO cells also produce less A␤ and secrete more soluble APP (31).
Previous studies have shown that retention of APP at the cell surface decreases A␤ production and increases sAPP␣ (5)(6)(7)32). Overexpression of an LRP C-terminal fragment in MEF LRPϪ/Ϫ cells bearing a point mutation within its distal NPXY motif increased sAPP␣ levels compared with a wild-type LRP fragment (13). This mutation is equivalent to the mLRP4(Y63A) endocytosis mutant from which we detected an increase in cell surface APP. Our findings that a chimeric mLRP4/APPtail receptor had markedly decreased rate of endocytosis compared with wild-type mLRP4 further suggest that, upon interaction of LRP and APP at the cell surface, the fast rate of endocytosis of LRP could indeed facilitate the trafficking of APP within the endocytic compartments.
It has been postulated that APP endocytosis facilitates A␤ generation because it brings both APP and the ␤-secretase ␤-site APP-cleaving enzyme, which is concentrated in lipid rafts, into close proximity (33,34). Furthermore, ␤-site APPcleaving enzyme optimum activity would be favored at the slightly acidic pH of endosomes (19). If LRP association with APP at the cell surface increases APP trafficking within the endocytic compartments where A␤ can be generated, we expected A␤ secretion in the media to be higher in CHO-mLRP4 cells than in CHO-pcDNA3 control cells. Past reports in the literature have indeed shown that LRP expression is associated with increased A␤ levels in the media (9, 13). Ulery et al. reported a 3-fold increase in A␤ levels in CHO cells stably transfected with APP751 and transiently transfected with fulllength LRP. Pietzrik et al. (13) found that A␤ levels were increased in LRPϩ/Ϫ mouse embryonic fibroblasts stably transfected with APP751 compared with LRPϪ/Ϫ mouse embryonic fibroblasts. However, contrary to our expectations and to these previous studies, we found decreased endogenous A␤ in the media of CHO-mLRP4 cells compared with CHO-pcDNA3 cells and no alterations in A␤ levels with mLRP4 endocytosis mutant receptor expression. Although we were initially surprised with these findings, a detailed comparison between these studies and ours revealed a major difference between them. Whereas we analyzed the effect of LRP overexpression on endogenous A␤ secretion, the other studies used cells overexpressing human APP to evaluate the effect of LRP on A␤ secretion. In view of these differences, we considered the possibility that when cells did not overexpress APP, the A␤ clearance pathway mediated by the LRP minireceptor could be dominant over its effect on increasing A␤ production, leading to the decreased steady-state levels of A␤ in the media.
Supporting this hypothesis, we found that the disappearance of exogenously added human A␤ was significantly greater in CHO-mLRP4 than that of CHO-pcDNA3 control cells. These results were consistent with the findings that LRP can bind and clear A␤ either directly (25) or indirectly via LRP ligands (e.g. apolipoprotein E and ␣2-macroglobulin) (26,27). The ability of CHO-mLRP4 cells to clear A␤ more effectively could therefore explain why lower levels of A␤ were detected in CHO-mLRP4 than in CHO-pcDNA3 control cells. This hypothesis was further confirmed by the generation of CHO-mLRP4 stably transfected with human APP770, which showed increased A␤ levels over pcDNA3 control cells, similarly to previous reports (9,13).
Taken together, our data support a link between LRP endocytosis, APP trafficking, and APP proteolytic processing. Previous in vivo studies from our laboratory have shown that overexpression of mutated human APP V717F and LRP minireceptor domain II in central nervous system neurons results in a net increase of soluble A␤ levels in the brain (14). This current study demonstrated that in CHO cells, LRP minireceptor functions in both the cellular clearance and production of A␤, and that the net result of these opposing pathways seems to depend on the ratio LRP/APP that is expressed in the cell. Therefore, under physiological and pathological conditions such as AD, the stoichiometry of APP to LRP within cells may dictate the overall levels of A␤ in the brain and the deposition of A␤ into plaques.