Low-Density Lipoprotein Receptor-Related Protein 1 (LRP1) Mediates Neuronal Aβ42 Uptake and Lysosomal Trafficking

Background Alzheimer's disease (AD) is characterized by the presence of early intraneuronal deposits of amyloid-β 42 (Aβ42) that precede extracellular amyloid deposition in vulnerable brain regions. It has been hypothesized that endosomal/lysosomal dysfunction might be associated with the pathological accumulation of intracellular Aβ42 in the brain. Our previous findings suggest that the LDL receptor-related protein 1 (LRP1), a major receptor for apolipoprotein E, facilitates intraneuronal Aβ42 accumulation in mouse brain. However, direct evidence of neuronal endocytosis of Aβ42 through LRP1 is lacking. Methodology/Principal Findings Here we show that LRP1 endocytic function is required for neuronal Aβ42 uptake. Overexpression of a functional LRP1 minireceptor, mLRP4, increases Aβ42 uptake and accumulation in neuronal lysosomes. Conversely, knockdown of LRP1 expression significantly decreases neuronal Aβ42 uptake. Disruptions of LRP1 endocytic function by either clathrin knockdown or by removal of its cytoplasmic tail decreased both uptake and accumulation of Aβ42 in neurons. Finally, we show that LRP1-mediated neuronal accumulation of Aβ42 is associated with increased cellular toxicity. Conclusions/Significance These results demonstrate that LRP1 endocytic function plays an important role in the uptake and accumulation of Aβ42 in neuronal lysosomes. These findings emphasize the central function of LRP1 in neuronal Aβ metabolism.


Introduction
Alzheimer's disease (AD) is a neurodegenerative disorder of the central nervous system characterized by a progressive decline in cognitive functions and neuronal loss. AD was originally attributed to the accumulation of extracellular amyloid and intraneuronal tau deposition [1]; however, mounting evidence indicates that intracellular accumulation of amyloid-b 42 (Ab42) is an early pathological marker that precedes neuronal cell death and correlates with cognitive decline. Significant intraneuronal Ab42 is found in neuronal cell bodies of both AD patients and patients with mild cognitive impairment, a preclinical condition that precedes AD [2,3]. Animal models for AD corroborate the presence of intraneuronal Ab42 prior to the appearance of amyloid plaques. Functionally, intraneuronal Ab42 load is associated with decreased long-term potentiation and increased synaptotoxicity in Ab42-containing neurons [4,5,6]. Since the endosomal/lysosomal pathway plays an important role in mediating intracellular degradation of Ab, an impairment of this pathway has been suggested as a possible mechanism for increased neuronal Ab42 load within multivesicular bodies, especially in the postsynaptic compartments [3,7,8]. The molecular mechanisms that lead to Ab42 accumulation during AD are not clear. Recent evidence indicates that autophagy, an upstream branch of the endosomal/lysosomal cascade, is impaired in AD brains and might be responsible for intraneuronal Ab42 accumulation [9]. In addition to intraneuronal Ab production, Ab42 internalized from extracellular space might represent a possible mechanism that contributes to intracellular Ab42 accumulation [10,11,12,13,14]. Therefore, understanding the mechanisms of Ab42 internalization is crucial for defining the neuropathological process of AD and for designing novel methods of AD therapy.
Apolipoprotein E (apoE) secreted by glia in the brain is required for the internalization of lipoproteins by neuronal apoE receptors and plays important roles in brain lipid transport and neuronal repair [15]. It has been suggested that apoE regulates Ab42 accumulation in neurons during AD. Strong co-immunostaining of intracellular apoE and Ab42 was evident in APP transgenic mice [16]. The e4 allele of the APOE gene represents the most prevalent genetic risk factor for late-onset AD [17,18,19], and apoE4 causes lysosomal leakage when applied exogenously to cultured cells [20,21]. Furthermore, apoE4 targeted-replacement mice induced to accumulate endogenous Ab42 show increased neurodegeneration compared to apoE3 mice [22]. Notably, metabolism of both apoE and Ab42 are regulated by the low-density lipoprotein (LDL) receptor-related protein 1 (LRP1), a large endocytic receptor belonging to the LDL receptor family [15].
LRP1 is synthesized as a 600-kDa precursor protein that interacts with the ER chaperone receptor-associated protein (RAP) [23,24] and matures in the biosynthetic pathway into an extracellular, ligand-binding subunit of 515-kDa (LRP1-515) and a transmembrane 85-kDa subunit (LRP1-85) that binds several adaptor proteins for efficient endocytic trafficking and signaling [25]. Several lines of evidence suggest a role for LRP1 in AD pathogenesis and in the metabolism of neuronal Ab42. 1. LRP1 is abundantly expressed in the cell body and in proximal processes of cortical and hippocampal neurons in the brain [26,27,28]. 2. LRP1 binds to Ab42 either directly or via Ab chaperones such as apoE to mediate brain Ab clearance [29,30,31]. 3. Several genetic risk factors for sporadic AD are ligands of LRP1 which, together with LRP1, are found in senile plaques in AD brains (for reviews see [31,32]. 4. In AD patients and in elderly people, brain LRP1 levels are significantly decreased and inversely correlate to the age of onset of AD, suggesting that a decrease in LRP1 function might contribute to the cognitive decline [33]. 5. Increased extracellular deposition of Ab42 was detected in RAP-deficient mice, which display reduced levels of LRP1 [29]. 6. Increased intraneuronal Ab42 was found in AD model mice overexpressing a functional LRP1 minireceptor, mLRP2 [34]. Together, this evidence indicates a role for LRP1 in the etiology of AD and suggests that LRP1 likely mediates neuronal Ab42 uptake via receptor-mediated endocytosis. Despite these clues, direct evidence supporting a role for LRP1 in neuronal Ab42 uptake and accumulation is lacking. Here, using both gain-of-function and loss-of-function approaches, we demonstrate a direct role for LRP1mediated endocytosis in neuronal Ab42 uptake and accumulation. These studies define a novel receptor-mediated pathway for Ab42 entry to neuronal lysosomes, where Ab42 accumulation may contribute to both Ab aggregation and neuronal toxicity.

Overexpression of LRP1 increases Ab42 uptake and accumulation in lysosomes in N2a cells
To test whether the endocytic activity of LRP1 mediates the uptake of Ab42 and its accumulation within neurons, we analyzed the uptake of fluorescently labeled Ab42 in control N2a cells and in N2a cells that overexpress a LRP1 minireceptor, mLRP4. This minireceptor contains the fourth ligand-binding domain of LRP1 and is capable of binding several LRP1 ligands with endocytic properties identical to the full-length LRP1 [35]. We generated stable N2a-mLRP4 and N2a-pcDNA3 cell lines and characterized them for total mLRP4 expression levels by Western blot and FACS (Fig. 1). Our results indicate that functional mLRP4 is properly folded and matured in N2a cells, as evidenced by its furin-induced cleavage into 120-and 85-kDa bands (Fig. 1A). To determine if mLRP4 reached the plasma membrane after synthesis, we analyzed cell surface levels of mLRP4 by FACS in non-permeabilized cells. Approximately 80% of the cells showed expression of cell surface mLRP4 (Fig. 1B). To assess whether cell surface mLRP4 could interact and internalize ligands, we performed uptake experiments with fluorescently labeled RAP. RAP is an LRP1 chaperone of 39-kDa that binds LRP1 with high affinity and is commonly used as a model extracellular ligand for LRP1 [23,36]. Cell-associated RAP was primarily detected intracellularly, as demonstrated by confocal microscopy ( Figure 1C) and increased in N2a-mLRP4 cells compared to N2a-pcDNA3 cells. To quantify the amount of internalized RAP, we performed FACS analysis of similarly treated cultures. Cells were detached and additionally treated with pronase to remove any cell-surface labeled RAP. N2a-mLRP4 cells have an 8-fold increase in the uptake of fluorescently labeled RAP compared to N2a-pcDNA3 cells (Fig. 1D), indicating that mLRP4 expressed in N2a cells is fully functional in mediating ligand internalization. Based on these results, we analyzed the uptake of Ab42 in N2a-pcDNA3 cells and N2a-mLRP4 cells by performing binding/ endocytosis assays and fluorescence microcopy analysis. We choose FAM-Ab42 as our amyloid-peptide probe since the addition of the carboxyfluorescein moiety at the N-terminal region does not change the properties of Ab42 aggregation and internalization into neurons [37]. For binding-endocytosis experiments, N2a-pcDNA3 and N2a-mLRP4 cells were first incubated with 5 mM Ab42 at 4uC for 1 h to allow the binding of Ab42. Cells were then warmed to 37uC for different periods of time to allow ligand internalization, and analyzed by confocal microscopy as described in the Experimental Procedures section. N2a-pcDNA3 cells accumulated a small amount of Ab42 after 8 h of warming; this accumulated Ab42 disappeared by 24 h (Fig 1E). In contrast, Ab42 uptake was faster and accumulation greater in N2a-mLRP4 cells, peaking at 4 h and decreasing over time. These results indicate that increased FAM-Ab42 uptake occurred in N2a cells overexpressing LRP1 minireceptor and demonstrate that the lysosomal function is unaffected by LRP1 overexpression, given the similar clearance rate of FAM-Ab42 in the two cell types analyzed.
To determine whether internalized Ab42 is accumulated over time, we performed continuous feeding experiments of FAM-Ab42 in N2a-pcDNA3 and N2a-mLRP4 cells. The steady-state levels and localization of Ab42 were determined by flow cytometry and confocal microscopy, respectively (Fig. 2). Cells were incubated with 500 nM FAM-Ab42 at 37uC for 24, 48 or 72 h and the content of fluorescence per cell was quantified by FACS and presented as a percentage of the levels in N2a-pcDNA3 at each time point ( Fig. 2A). Only a slight increase in the levels of FAM-Ab42 was detected in N2a-mLRP4 cells after 24 h incubation. However, twice the levels were detected at 48 h and 72 h in N2a-mLRP4 cells compared to N2a-pcDNA3 cells, indicating that LRP1 minireceptor overexpression increases the accumulation of FAM-Ab42 in N2a cells. It has been previously demonstrated that endocytosed Ab enters the lysosomal pathway in both neuronal and non-neuronal cells, and it has been suggested that both exogenously added Ab42 and de novo synthesized Ab42 accumulate and aggregate in these organelles [10,12,13]. Therefore, we analyzed the subcellular localization of internalized FAM-Ab42 by confocal microscopy in N2a cells treated as before and analyzed co-localization with lysosomes. For the visualization of acidic organelles, a brief incubation with LysoTracker Red at the end of the experiment was included (Fig. 2B). After 24 h of incubation, a small amount of cell-associated Ab42 can be detected in N2a-mLRP4 cells. However, this FAM-Ab42 does not localize into acidic organelles, as evidenced by low co-localization with LysoTracker. However, increased localization of Ab42 into LysoTracker-positive organelles can be detected in both N2a-pcDNA3 and N2a-mLRP4 cells after 48 and 72 h, and an increased number of double-positive cells were detected in N2a-mLRP4 cells compared to N2a-pcDNA3 cells after 48 h of treatment. Given that pulse-chase experiments showed normal lysosomal activity and increased FAM-Ab42 internalization in LRP1-expressing cells (Fig. 1E), and because increased steady-state levels of FAM-Ab42 were found in LRP1-expressing cells at longer incubation times (Fig. 2), these results collectively indicate that LRP1 minireceptor overexpression increases the uptake, delivery and accumulation of Ab42 into lysosomes in N2a cells.

LRP1 increases Ab42 accumulation in GT1-7 and MEF cells
To rule out cell type-specific effects and artifacts due to the expression of LRP1 minireceptor, we performed similar FAM-Ab42 uptake experiments in GT1-7 neuronal cells with reduced endogenous LRP1 levels by siRNA knockdown (Fig. 3A), and in mouse embryonic fibroblasts from both wild-type (MEF1) and LRP1 knockout mice (MEF2) [38] (Fig. 3B). GT1-7 cells were plated and treated as indicated in the Experimental Procedures and GT1-7 and MEF cells were incubated with 500 nM of FAM-Ab42 at 37uC for 4 h. Steady-state levels of intracellular Ab42 were determined by flow cytometric analysis of pronase-treated cells. A decrease in the internalization of FAM-Ab42 was observed in GT1-7 cells with decreased LRP1 levels, and in MEF2 cells compared to MEF1. Depletion of LRP1 levels in both knockdown and knockout cells were confirmed by Western blot (Fig. 3C). Collectively, these results indicate that LRP1 expression directly impacts the internalization of FAM-Ab42 in both neuronal and non-neuronal cells.

LRP1 endocytosis is required for Ab42 accumulation in neuronal cells
To determine whether the endocytic function of LRP1 is required for Ab42 internalization and accumulation, we first analyzed the effect of clathrin knockdown on the uptake of FAM-Ab42 in N2a-mLRP4 cells. We decreased the expression of clathrin heavy chain by lentiviral-mediated delivery of small hairpin interfering RNA (shRNA) in N2a-mLRP4 cells and analyzed cell surface levels of mLRP4 by FACS. Lentiviralmediated knockdown of clathrin heavy chain substantially decreases the protein levels of clathrin heavy chain and evokes a twofold-increase in the levels of cell surface mLRP4 in N2a cells (Fig. 4A), consistent with an impairment of clathrin-mediated endocytosis of this LRP1 minireceptor. We therefore tested the uptake of FAM-Ab42 in similarly infected cultures. Approximately a twofold decrease in FAM-Ab42 uptake was observed in clathrin knocked down cells compared to cells infected with a control, and N2a-mLRP4 cells were incubated with 500 nM Alexa488-RAP at 37uC for 30 min, then fixed and analyzed by fluorescence microscopy. D, flow cytometric analysis from experiments as in C. N2a-pcDNA3 and N2a-mLRP4 cells were detached and additionally treated with pronase to remove cell surface-associated RAP. Quantification of FACS experiments indicates that N2a-mLRP4 cells internalize approximately 8 times more RAP than N2a-pcDNA3 cells. ** p,0.01, t-test. n = 3. E, LRP1 minireceptor expression increases the endocytosis of Ab42 in N2a cells. N2a-pcDNA3 and N2a-mLRP4 cells were incubated with 5 mM FAM-Ab42 for 1 h at 4uC and warmed to 37uC for 0.5, 4, 8 or 24 h, then fixed and examined by confocal microscopy. In N2a-mLRP4 cells, FAM-Ab42 accumulation peaked 4 h after warming and decreased over time, suggesting that intracellular Ab42 is eventually delivered for degradation. doi:10.1371/journal.pone.0011884.g001 empty lentivirus (Fig.4B). As a control for specificity, we analyzed the uptake of a FAM-Ab42 peptide that contained the same amino acid composition but in a scrambled sequence. Very little uptake of fluorescently labeled scrambled FAM-Ab42 was detected in both control and clathrin knockdown N2a-mLRP4 cells, indicating that Ab42 internalization requires the intact sequence of Ab42. Together, these results indicate that clathrin-mediated endocytosis is a critical pathway for Ab42 uptake and strongly suggest that neuronal Ab42 uptake is a receptor-mediated process that requires the endocytic properties of LRP1. To further confirm this, we specifically targeted the endocytic function of mLRP4 and analyzed the internalization of FAM-Ab42 in LRP1 endocytosisdeficient cells. The molecular determinants of LRP1 endocytosis reside in its cytoplasmic tail [39,40]. We have previously demonstrated that, despite reaching the plasma membrane, a deletion of the complete cytoplasmic tail of mLRP4 (mLRP4-Tless) impairs its endocytic function [34]. We therefore prepared N2a-mLRP4-Tless stable cell lines and first characterized the endocytic properties of the LRP1 minireceptor in this neuronal cell type. Flow cytometric analysis indicates that deletion of the mLRP4 cytoplasmic tail increases its distribution at the cell surface (Fig. 4C), consistent with impaired endocytosis. Additionally, the endocytosis rate for 125 I-RAP was decreased in these cells (Fig. 4D). We compared the uptake of FAM-Ab42 in N2a-pcDNA3, N2a-mLRP4 and N2a-mLRP4-Tless by FACS in both pronase-treated and untreated cells. Again, cell-associated FAM-Ab42 increases in N2a-mLRP4 cells compared to N2a-pcDNA3 cells. Notably, FAM-Ab42 uptake was similar in N2a-mLRP4-Tless and N2a-pcDNA3 cells, consistent with a role for LRP1 endocytosis in Ab42 uptake (Fig. 4E). Therefore, impairing the endocytic function of mLRP4 by clathrin knockdown or by the removal of its cytoplasmic tail decreased FAM-Ab42 uptake, demonstrating that LRP1 endocytosis is required for Ab42 internalization in N2a cells.

Increased cellular toxicity in N2a-mLRP4 cells incubated with Ab42
It has been previously demonstrated that increased delivery of Ab42 into lysosomes increases leakage of the lysosomal contents into the cytosol and hence induces cell toxicity [20,21], an effect that is increased by apoE4 and might depend on LRP1 function. We hypothesized that Ab42 treatment might increase cell death in mLRP4-overexpressing N2a cells because we detected increased co-localization of Ab42 with lysosomes in these cells. We treated N2a-pcDNA3 and N2a-mLRP4 cells with Ab42 for 24, 48 and  Ab42, suggesting that LRP1-mediated delivery of Ab42 into lysosomes might not be sufficient to trigger cell death in N2a cells at early time points or at lower Ab42 concentrations ( Figure 5). Together, our results demonstrate that LRP1 mediates neuronal endocytosis of Ab42, which leads to eventual lysosomal accumulation at high Ab42 concentrations and slightly decreases neuronal viability.

Discussion
The importance of LRP1 in Ab metabolism has been recognized because of its ability to mediate brain efflux at the blood-brain barrier and because of its role in modulating APP processing to Ab [31]. Here, we show that LRP1 endocytosis plays an additional role in Ab42 uptake and delivery into intracellular compartments of neuronal cells. A faster transient accumulation of intracellular Ab42 was detected in N2a-mLRP4 cells compared to N2a-pcDNA3 cells after 4 h of endocytosis in binding-endocytosis experiments (Fig. 1E). However, continuous uptake experiments suggest that, over time, Ab42 might escape the degradation pathway and start to accumulate, which is evident after 48 h of incubation (Fig. 2B). It is interesting that N2a-pcDNA3 cells showed a delayed accumulation of Ab42 in binding-endocytosis experiments assays, with a peak at 8 h after warming. Because N2a cells express very low levels of endogenous LRP1 (Fig. 1A, anti-LRP1 blot), and given that endogenous LRP1 and the mLRP4 minireceptor possess similar kinetics of endocytosis, these results suggest that an LRP1-independent pathway is responsible for Ab42 internalization in N2a-pcDNA3 cells, which mediate slower kinetic delivery into lysosomes. The LRP1 minireceptor mLRP4 might compete for these Ab42 cell surface binding sites and give rise to the rapid internalization observed in N2a-mLRP4 cells. Notably, at 8 h after warming up, N2a-mLRP4 cells exhibited decreased levels of Ab42 in our imaging experiments, indicating that the degradation of Ab42 by the LRP1-dependent pathway is very efficient. The mechanisms that are responsible for Ab42 accumulation in N2a-mLRP4 cells in the continuous endocytosis experiments are unknown. However, it is tempting to speculate that the constant presence of relatively high concentration of Ab42 leads to high Ab42 concentrations in the lysosomes that might exceed its degradation capacity. Our data demonstrate that the endocytic activity of LRP1 is required for Ab42 internalization and accumulation because both clathrin heavy chain knockdown and deletion of the LRP1 tail decreased the levels of intracellular Ab42.
Previous reports addressed the question of whether LRP1 mediates the uptake of Ab into neurons. Experiments performed in cortical neurons from Tg2576 APP transgenic mice suggested that LRP1 might endocytose and degrade endogenously produced Ab complexed with LRP1 ligands a2M and lactoferrin [41]. However, these experiments relied primarily on the use of RAP to define the LRP1 pathway. Because RAP also inhibits binding of ligands to several other members of the LDLR family, the specific function of LRP1 in these studies was not clear. Endogenously produced Ab adds a second level of complexity to that study [41] as Ab might be either produced in the endocytic compartments or trafficked there directly without secretion. Similarly, the uptake of Ab40 complexed with apoE has been previously characterized by FACS into purified rat brain synaptosomes, but again only RAP was used to define the LRP1 pathway [42]. To our knowledge, our study is the first to report changes in Ab trafficking by direct targeting the expression and endocyitic function of LRP1 in neural cells. We show that in neural cells, overexpressing LRP1 minireceptor increases Ab42 internalization and accumulation, whereas knockdown of endogenous LRP1 decreases these processes. These findings establish a direct role for LRP1 in neuronal Ab42 uptake, degradation and/or accumulation. It is worth mentioning that induction of Ab accumulation in the brains of apoE-targeted replacement mice through inhibition of the extracellular Ab protease neprilysin [22] showed that Ab accumulates intracellularly along with apoE in CA1 and septal neurons of apoE4 targeted-replacement mice, and that this effect correlates with increased LRP1 levels in these neurons [22,43]. This in vivo evidence is in line with our current findings in cultured neurons and our previous studies in mouse brain [34] which demonstrated that LRP1 function is critical for the uptake and accumulation of Ab42. Since apoE3 and apoE4 bind similarly to LRP1 [44], additional factors might account for the apoE isoformspecific effects on neuronal Ab42 accumulation. We are currently evaluating the contribution of differential trafficking of apoE isoforms in neuronal Ab42 metabolism and accumulation. Kim et al. have recently demonstrated that LDLR, the founding member of the LDLR gene family to which LRP1 belongs, negatively regulates apoE levels in the brain and promotes Ab clearance in vivo [45]. They elegantly showed that as little as a twofold increase in brain LDLR decreases apoE levels and enhances the elimination of extracellular Ab from the brain interstitial fluid, with a concomitant decrease in amyloid plaque formation. Although it remains to be established whether the effect of LDLR on Ab metabolism is apoE dependent, this finding identifies another apoE endocytic receptor that modulates Ab metabolism in the brain. It is interesting that LDLR and LRP1 have overlapping binding properties for apoE3 and apoE4; however, LRP1's high expression level in the brain and fast endocytosis rate likely make this receptor an important candidate for Ab metabolism in the brain. Supporting this hypothesis, we have previously shown that LRP1 knockout in forebrain neurons leads to increased brain apoE levels in mice [46]. On the other hand, Ab42 immunoreactivity was significantly decreased in PDAPP mice lacking apoE, supporting a role for apoE in mediating Ab uptake in vivo [34]. More detailed characterization of APP, apoE, LDLR and LRP1 is required to define the precise interplay among these lipid metabolic proteins. The increased accumulation of Ab42 in lysosomes of LRP1expressing cells (Fig. 2B) and its correlation with decreased cell viability (Fig.5) is intriguing. Imaging experiments of endocytosed Ab42 suggest that a large portion of LRP1-delivered Ab42 is degraded. However, when the amount of Ab42 exceeds lysosomal degradation capacity, Ab42 likely accumulates and aggregates. It has been demonstrated that localization of Ab42 into lysosomes is toxic to neurons, in particular in the presence of apoE4 [20,21,22]. The acidic environment of the lysosomal compartment might be key for the potentiating effect of apoE4 on Ab42-induced lysosomal leakage and apoptosis, as demonstrated by Ji et al.
Notably, this group found that reducing LRP1 levels by siRNA knockdown decreases the stimulatory effects of apoE4 [21], which is consistent with our finding that LRP1 mediates endocytosis and lysosomal trafficking of Ab42 in neuronal cells. In addition, it has been demonstrated that LRP1 plays a role in the TGFb-induced, Ab42-mediated lysosomal leakage [47,48]. Our neural viability experiments showed that cell viability was slightly decreased after 72 h of incubation with a high concentration of Ab42 in LRP1expressing cells (Fig. 5). It will be of great interest to evaluate whether increased co-localization of Ab42 with lysosomes in apoE4-treated N2a-mLRP4 cells leads to more severe lysosomeassociated neuronal toxicity in our cell-based system.
Increased accumulation of intracellular Ab has various pathological effects on cell and organelle function, including proteasome inhibition, mitochondrial abnormalities, tau hyperphosphorylation and presumably, the seeding of amyloid plaques after neuronal cell death [49]. Immunohistochemical analyses of AD brains strongly suggests that amyloid plaques arise when an Ab-accumulating neuron undergoes a single lysis event [50,51]. In support of this, 3xTg-AD mice treated with anti-Ab antibodies show a parallel decrease of extracellular amyloid deposits and intraneuronal Ab in pyramidal cells, which suggests there is a dynamic relationship between the intracellular and extracellular pools of Ab [52]. Therefore, it is conceivable that the endocytic activity of LRP1 might contribute to the appearance of plaques by increasing the internalization of Ab, with a consequent decrease in neuronal function and viability and release of amyloid seeds into the brain parenchyma. Our LRP1-overexpressing mouse model, however, did not show an increase in the amyloid plaque burden when crossed to PDAPP amyloid model mice [53]. Additional compensatory mechanisms mediated by LRP1, including the activation of survival signaling pathways, might prevent the adverse effects of intraneuronal Ab accumulation in vivo and its subsequent release. It is worth mentioning that LRP1 is indeed a signaling receptor capable of activating survival pathways including ERK, Akt and b-catenin [54,55,56,57,58]. Among these pathways, Akt is regulated by LRP1 in differentiated neurons and targeted by intraneuronal Ab [57,59]. We propose that dysregulated neuronal functions of LRP1 in Ab production, clearance, as well as cell survival might contribute to the neuronal pathology in AD.
In conclusion, we show that modulation of LRP1 expression or its endocytic function directly impacts neuronal Ab42 uptake, degradation and/or accumulation. These results provide strong evidence that an LRP1-mediated pathway is crucial for Ab42 trafficking and metabolism in neurons.

FACS analysis of surface and total ratio and purity of clones
Sub-confluent N2a cells in 100 mm Petri dishes were washed twice in PBS and non-enzymatically detached by incubating 5 min at 37uC in cell dissociation solution. After washing in growth media, samples were divided in two parts and half of the samples were kept in a minimal volume of PFN buffer (PBS supplemented with 1.5% FBS and 0.1% sodium azide) for the non-permeabilized group. The remaining cells were resuspended in PFN containing 0.05% saponin and gently rocked for 30 min at 4uC. After washing with growth media, permeabilized and non-permeabilized cells were incubated with 50 mg/mL anti-HA antibody for 2 h at 4uC. After two washes with PFN, cells were incubated with 50 mg/mL goat anti-mouse-Ig fluorescein isothiocyanate as a secondary antibody and similarly incubated for 45 min. Cells were washed twice, fixed in PFN-1% PFA and analyzed in a FACSCalibur cytometer (BD Bioscience) equipped with an argon ion laser. Laser excitation of 488 nm for fluorescein isothiocyanate was used. Twenty thousand cells from each sample were analyzed and histograms were generated using CellQuest software. After the subtraction of blank controls lacking primary antibody, the geometric means of fluorescence intensity were obtained for non-permeabilized and permeabilized samples, representing surface and total minireceptor levels, respectively. The surfaceto-total ratio was calculated and utilized for comparison purposes. Unpermeabilized conditions were used to determine the purity of stable cell lines.

Immunoblot Analysis
After indicated treatments, cells were washed twice in PBS and cell lysates were prepared in PBS containing 1% TX-100, supplemented with 1 mM PMSF and a protease inhibitor cocktail. After 30 min on ice, cells were briefly vortexed and cell debris discarded by centrifugation at 10,0006g for 10 min at 4uC. 20 mg of protein was resolved by reducing 10% SDS-PAGE and transferred to PVDF membranes overnight. After the membranes were blocked, proteins were detected by using 1:1,000 dilution of primary antibodies for commercial antibodies or 1 mg/mL of purified anti-LRP polyclonal IgG. The immunoblotting was followed by detection with a horseradish peroxidase-conjugated secondary antibody and enhanced chemiluminiscence substrate (GE Healthcare, Uppsala, Sweden). Bands on films were scanned using Quantity-One software (Bio-Rad, Hercules, CA).

Preparation of Ab42 and RAP
Synthetic FAM-Ab42 was resuspended essentially as described by Stine et al. to yield an essentially assembly-free/monomeric preparation [61]. Briefly, 0.1 mg vial of FAM-Ab42 was resuspended in HFIP, divided into aliquots and dried in a current of nitrogen gas. FAM-Ab42 and Ab42 protein films were thoroughly resuspended in DMSO at 5 mM and immediately diluted with water to form a 100 mM Ab42 stock solution. To avoid aggregation of the peptides during treatment of cells, resuspended aliquots were used within 30 min after resuspension. RAP protein was routinely produced in our lab and was labeled with an Alexa-488 labeling kit (Millipore, Bedford, MA) and stored at 280uC until use [23].
In vitro FAM-Ab42 uptake and accumulation assays by laser scanning confocal microscopy For the imaging experiments, stable N2a cells were plated into an 8-well chambered coverglass at 16,250 cells/cm 2 with one independent chambered coverglass for each time point, and cells were grown overnight. For uptake experiments, cells were washed three times with cold DMEM and binding of 5 mM FAM-Ab42 was performed in DMEM at 4uC for 1 h. After binding, unbound FAM-Ab42 was removed by three washes with cold PBS. To initiate the uptake, cells were switched to 37uC for different time periods by adding pre-warmed internalization media (DMEM containing 10% FBS and 2 mM L-glutamine) and bringing the chambered coverglass back to the incubator until completion of the kinetic time point. For steady-state accumulation experiments, cells were washed twice with PBS and incubated for the indicated periods of time with pre-warmed internalization media containing 500 nM FAM-Ab42. In some experiments, 100 nM LysoTracker Red probe was added 30 min prior to the end of the experiment. Cells were then fixed and analyzed by a confocal laser scanning microscope (Fluoview 500, Olympus) using a 40x objective lens. The signal corresponding to Ab42 was visualized in the fluorescein channel and LysoTracker was visualized in the Alexa-569 channel.

Quantification of Ab42 accumulation by FACS analysis
For quantification of Ab uptake experiments, stable N2a cells were plated into 12-well plates at 16,250 cells/cm 2 and grown overnight. Similar to imaging assays, cells were washed twice with PBS and incubated for the indicated periods of time with prewarmed internalization media containing 500 nM FAM -Ab42. After incubation, cells were washed twice with PBS and nonenzymatically detached as before in 500 mL cell dissociation solution. Cells were sequentially washed once with 1 mL of cold internalization media and twice with 2 mL cold PFN. Cells were finally resuspended in PFN-1% PFA and analyzed by FACS as before. For MEF-1 and MEF-2 experiments, 5,000 cells/cm 2 were plated in 12-well plates the day before. Cells were then washed twice with PBS and incubated 4 h with pre-warmed internalization media containing 500 nM FAM-Ab42. After completion of the incubation time, cells were processed by FACS analysis as described above for N2a cells.

Kinetic Analysis of Endocytosis
Kinetic analysis of receptor-mediated endocytosis was carried out as described previously [40]. Briefly, stably transfected N2a cells were plated at a density of 2610 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 ice-cold ligand binding buffer (DMEM-5 mM CaCl 2 containing 0.6% bovine serum albumin) with 5 nM 125 I-RAP. The binding of iodinated proteins was carried out at 4uC 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, and 0.1 M NaCl, pH 2.6) was added to one set of plates without warming and kept on ice. The remaining plates were then placed in a 37uC water bath and 0.5 ml of pre-warmed 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 remaining on the cell surface was stripped by incubation of cell monolayers with ice-cold stop/strip solution for 10 min and counted. The sum of internalized and cell-surface ligand after each assay was used as the maximum potential internalization. The fraction of internalized ligand at each time point was calculated and plotted.

LRP1 and clathrin heavy chain knockdown
LRP1 knockdown was carried out in GT1-7 cells by siRNA transfection using Lipofectamine 2000 (Invitrogen, Carlslab, CA) and 120 nM of dsRNA duplex (Ambion, Austin, TX). Briefly, cells were plated in 6-well plates the day before transfection and grown overnight. Scrambled and LRP1 siRNA duplexes were generated according to the manufacturer's instruction and complexed with Lipofectamine 2000 as previously described [62]. FAM-Ab42 experiments were performed 72 h after siRNA transfection and analyzed as previously described for N2a cells. For clathrin heavy chain knockdown, shRNA lentiviruses were produced in the Viral Vectors Core facility at Washington University School of Medicine. In brief, 293T cells were transfected with pLKO.1derived constructs together with the pHR'8.2DR and pCMV-VSV-G packaging systems as previously described [63]. Conditioned media were concentrated by ultracentrifugation, titrated against HT1080 cells and kept frozen in 20 mL aliquots at 280uC until use. N2a cells were plated at 16,250 cells/cm 2 in 100 mm Petri dishes and grown overnight. N2a cells were infected two times by incubating the cultures with 1610 7 transforming units of lentivirus overnight. After the second infection, the cells were split into 12-well plates for Western blot and FAM-Ab42 accumulation experiments or were kept an additional day without being split up for cell surface and total mLRP4 levels determination by FACS.

MTS redox assay in Ab42 treated cells
N2a-pcDNA3 and N2a-mLRP4 cells were plated at 16,250 cells/cm 2 in 48-well plates and grown overnight. Serial dilutions of freshly prepared synthetic Ab42 were prepared by diluting the Ab42 stock with internalization media. Culture media was then replaced with media containing 0, 1 or 3 mM Ab42 and cells were incubated for an additional 24, 48 or 72 h. Cell viability was measured by using one-solution cell titer kit at each corresponding time point (Promega, Madison, WI). After 4 h labeling, the soluble reduced formazan product was quantified by directly reading the plate in a Bio-Tek plate reader. After background subtraction from media-only containing wells, results were represented as % of the corresponding control and untreated N2a cells.