The low-density lipoprotein receptor-related protein 1 and amyloid-β clearance in Alzheimer’s disease

1The low-density lipoprotein receptor (LDLR) family consists of structurally related single transmembrane receptors, including LDLR, LDLR-related protein 1 (LRP1), LRP1B, megalin/LRP2, very-LDLR (VLDLR), apolipoprotein E receptor 2 (ApoER2)/LRP8, sortilin-related receptor (SorLA/LR11), LRP5, and LRP6. These cell surface receptors recognize extracellular ligands for subsequent signaling and/or trafficking to either degradation or recycling pathways. While the LDLR family members often recognize the same ligands, they regulate distinct physiological or pathophysiological pathways due to unique tissue expression patterns. In particular, apolipoprotein E (apoE), which transports cholesterol, is a critical ligand for several receptors of the LDLR family. Since the APOE ∊4 allele increases the risk for late-onset Alzheimer’s disease (AD) compared with the APOE ∊2 and ∊3 alleles, LDLR family has been vigorously studied as a target to explore the complex pathogenesis of AD.

2Amyloid-β (Aβ) peptides cleaved from amyloid precursor protein (APP) are the key molecules involved in AD pathogenesis; deposition of Aβ in the brain as senile plaques and cerebral amyloid angiopathy (CAA) likely triggers a cascade of events leading to disease onset. Recent evidence has also shown that soluble Aβ oligomers injure synapses resulting in cognitive impairment prior to Aβ deposition. While familial AD, which accounts for ~1% of AD cases, is likely caused by genetic mutations in APP, PSEN1, and PSEN2 leading to enhanced Aβ production, a positive correlation between Aβ levels and APP processing is not evident in sporadic late-onset AD, which represents the bulk of all AD cases. In fact, the disturbance of Aβ clearance machinery appears to be a leading cause of Aβ accumulation in the brain. Thus, the dysregulation of Aβ clearance pathways may be a central disease event in the majority of AD cases. Improved understanding of such pathways should help to both understand the complex pathogenesis of AD and allow for rationale design for AD therapy.

3Among the LDLR family members, LRP1 is the most studied receptor due to its involvement in multiple pathways in AD pathogenesis. LDLR also mediates Aβ metabolism and SorLA/LR11, which controls APP trafficking/processing, is genetically associated with AD. LRP1 is a large multi-functional receptor that regulates the endocytosis of diverse ligands and transduces several cell signaling pathways by coupling with other cell surface receptors. LRP1 is detected in most tissues and is highly expressed in liver, brain and lung. In the central nervous system, LRP1 is abundantly expressed in neurons, glial cells and vascular cells, and plays a critical role in maintaining brain homeostasis. In this review, we discuss how LRP1 regulates AD pathogenic pathways in different cell types with particular focus on Aβ clearance pathways.

4Low-density lipoprotein receptor-related protein 1 was initially identified in liver cells as LDLR homology. It is composed of two subunits including an 85-kDa C-terminal transmembrane domain and a 515-kDa N-terminal extracellular domain. LRP1 is synthesized as a glycosylated precursor protein and then cleaved into two subunits in the Golgi complex. After proteolytic processing, the extracellular domain of LRP1 is non-covalently connected to the transmembrane domain as it matures to the cell surface. The extracellular domain of LRP1 contains four ligand-binding domains I-IV with 2, 8, 10, and 11 cysteine-rich complement-type repeats, respectively. These motifs contain a net negative charge, which allows the bindings of a variety of positively charged ligands. The domains II and IV of LRP1 are the major binding regions. The cytoplasmic tail of LRP1 contains two copies of NPXY motifs, which commonly present in most members of the LDLR family and serve as the endocytosis signal for the LDLR. In addition to the two NPXY motifs, the LRP1 cytoplasmic tail has a YXXL motif, which along with two di-leucine motifs serve as the dominant endocytosis signals for its rapid endocytosis. When the endocytosis rates of several LDLR family members were compared using in vitro cellular models, the LRP1 tail showed faster endocytosis with t1/2 of ~0.5 min compared with those of the LDLR tail (t1/2 = 4.8 min) and megalin/LRP2 tail (t1/2 = 3.1 min), whereas VLDLR and ApoER2 exhibit relatively slower endocytosis rates (t1/2 = ~8 min; ). Thus, the main function of LRP1 is to capture its ligands through extracellular ligand-binding domains, rapidly internalize them through its unique cytoplasmic tail and deliver them to the endosomal/lysosomal compartments. After dissociation of ligands in the early endosome, LRP1 is known to efficiently recycle back to the cell surface by coupling with sorting nexin 17.

5Low-density lipoprotein receptor-related protein 1 also regulates signaling pathways in response to extracellular ligands by several mechanisms. Binding of tissue-type plasminogen activator (tPA) or α2-macroglobulin (α2M) to LRP1 induces Src family kinase (SFK) activation and SFK-dependent Trk receptor transactivation in neuronal cells. LRP1 also controls cell signaling by mediating the endocytosis of preformed receptor-ligand complexes into endosomes as a co-receptor. For example, LRP1 couples with the platelet-derived growth factor (PDGF) receptor and traffics into endosomes, where the phosphorylation of the PDGF receptor is induced in the presence of PDGF. In addition, LRP1 modifies the distribution of several membrane proteins between the cell surface and intracellular pools, which impacts their signaling strength. In some cases, LRP1 deletion appears to increase total, or cell surface, levels of PDGF receptor and urokinase plasminogen activator (uPA) receptor, and activate their downstream signaling pathways. In neurons, LRP1 interacts with the N-methyl-D-aspartate (NMDA) receptor through PSD-95, and regulates its trafficking to recycling compartments or degradation pathway. When LRP1 is deleted in neurons, degradation of the NMDA receptor is accelerated, resulting in decreased NMDA receptor levels and disturbed signaling pathways. Furthermore, the C-terminal intracellular domain of LRP1 (LRP1-ICD), cleaved from the transmembrane domain by γ-secretase, likely regulates the transcription of target genes. In fact, it has been demonstrated that LRP1-ICD suppresses LPS-induced inflammatory responses by binding to the interferon-γ promoter.

6Taken together, LRP1 serves as a multifunctional receptor that controls the endocytosis of a variety of ligands, influences signaling pathways by coupling with other cell surface receptors or proteins, and directly regulates gene expression through its intracellular domain.

7Low-density lipoprotein receptor-related protein 1 ligands, specifically apoE, α2M, tPA, uPA, plasminogen activator inhibitor-1, lipoprotein lipase, and lactoferrin co-deposit with Aβ in senile plaques in AD brains. In fact, the soluble form of LRP1 is also likely a prominent component of senile plaques and has been found to co-localize exclusively with these ligands. Although it is not clear whether LRP1 and its ligands are independently associated with senile plaques, these observations suggest that they do interact with Aβ in AD brains. Immunohistochemical analysis has revealed that LRP1 is up-regulated in neurons and in GFAP-positive activated astrocytes, in particular in astrocytic processes surrounding senile plaques in AD. Consistent with this finding, LRP1 mRNA levels are shown to be increased in temporal neocortex in AD patients. Since the mRNA levels of both LRP1 and GFAP are up-regulated in AD brains with a positive correlation, LRP1 expression is likely enhanced in activated astrocytes. On the other hand, LRP1 levels have been shown to be decreased in midfrontal cortex in AD cases (n = 39) as compared to age-matched controls (n = 39; ). Interestingly, higher LRP1 levels significantly correlate with later ages at onset of AD, while age and LRP1 expression in normal individuals appears inversely correlated. In addition, when the regional correlations between LRP1 and synaptic markers (synaptophysin and PSD95) or GFAP were assessed, a moderate-to-strong positive regional correlation was observed between LRP1 and postsynaptic marker PSD95, but not presynaptic marker synaptophysin and GFAP in the brains from non-demented individuals. Thus, LRP1 expression may be either up-regulated in glial cells due to neuroinflammation or suppressed in neurons due to postsynaptic damages in AD. Further studies are needed to clarify the temporal and spatial regulation of LRP1 expression in AD brains.

8Several lines of evidence support a role of LRP1 in regulating APP endocytosis. LRP1 can bind to the Kunitz-type protease inhibitor (KPI) domain of APP and mediate its degradation. The association of APP with LRP1 leads to increased trafficking of APP through the endosomal compartments, resulting in accelerated Aβ production and APP processing. Consistent with these results, the regional correlation between LRP1, APP, and Aβ showed positive correlations in non-demented individuals. Therefore, LRP1 may function to increase Aβ levels from the perspective of APP processing and Aβ production. On the other hand, LRP1 plays a critical role as an endocytic receptor to eliminate Aβ from the brain. In fact, after intracerebral microinjections of [125I]-Aβ40 in young mice, Aβ was rapidly removed from the brain (t1/2 < 25 min). Aβ40 clearance was significantly inhibited by LRP1 antagonist, RAP, or antibodies against LRP1. Thus, the ability of LRP1 to regulate both Aβ production and clearance suggests a critical role of this receptor in AD pathogenesis. In Aβ-independent pathways, conditional deletion of the Lrp1 gene in forebrain neurons in mice leads to age-dependent dendritic spine degeneration, synapse loss, neuroinflammation, memory loss, and neurodegeneration, which are all common features of AD.

9Low-density lipoprotein receptor-related protein 1 is a major apoE metabolic receptor in the brain. As APOE4 dramatically increases AD risk and accelerates disease onset compared with APOE2 and APOE3, understanding how apoE is involved in AD pathogenesis has been both an opportunity and a challenge. Given that apoE4 is related to increased Aβ aggregation and deposition in the brain, apoE-Aβ interaction has been actively studied to understand the specific roles of apoE isoforms. While the main function of apoE is to transfer lipid from cell to cell through cell surface LRP1 and other apoE receptors, apoE can also bind to Aβ through a region overlapping with its receptor-binding site or lipid-binding site in an isoform-dependent manner. Although growing evidence supports apoE-Aβ binding, a recent report showed that only a small portion of soluble, cell-derived Aβ interacted with astrocyte-secreted or artificially reconstituted apoE particles in solution. Thus, the effects of apoE on Aβ cellular uptake are likely complex. While recombinant apoE accelerates neuronal Aβ uptake in an isoform-dependent manner (apoE3 > apoE4; ), apoE particles inhibited the cellular uptake of soluble Aβ in astrocytes. Of note, the suppressive effect of apoE particles on Aβ uptake was not detected in LRP1-deficient cells. The LRP1-blocking antibody also significantly decreased the effect of apoE on Aβ uptake in astrocytes. Thus, apoE may either facilitate or inhibit LRP1-dependent or independent Aβ endocytosis depending on its concentration, Aβ aggregation state, apoE isoform, apoE lipidation and the expression pattern of the receptors on the cell surface.

10The activated form of α2M (α2M*) is also a well validated LRP1 ligand. While α2M* associates with Aβ and prevents fibril formation, α2M* enhances the clearance of soluble Aβ via LRP1 in neurons. RAP, which is an LRP1 chaperone and antagonist, can also interact with Aβ and facilitate its cellular uptake through heparan sulphate proteoglycan (HSPG), rather than LRP1. Thus, it is interesting to note that several major LRP1 ligands can bind to Aβ, suggesting the existence of common mechanisms by which Aβ and other ligand interact with LRP1 and/or HSPG. Further biochemical and structural studies are needed to determine the binding properties among Aβ, LRP1 and its ligands, which may provide insights as to the differential effects of LRP1 ligands on cellular Aβ metabolism.

11Cellular Aβ clearance through lysosomal degradation in brain parenchyma cells (microglia, astrocytes, neurons) and in cerebrovascular system constitutes a major pathway (Figure 1), while Aβ is also efficiently degraded by a large set of proteases including neprilysin and insulin-degrading enzyme in extracellular space (Figure 1; ). Neurons not only produce Aβ from APP proteolytic processing but also eliminate it through cellular uptake and lysosomal degradation. If neuronal Aβ endocytosis is disturbed, the accumulation and aggregation of Aβ may lead to synaptic injury and eventual neuronal death. In neurons, LRP1 is expressed predominantly in the postsynaptic region and the cell body, where LRP1 mediates Aβ uptake and subsequent degradation (Figure 1;, ). When LRP1 is deleted in neurons in adult mice, the half-life of interstitial fluid (ISF) Aβ in cortex increases, resulting in exacerbated amyloid pathology. In addition, it is interesting to note that upon internalization from distal axons, Aβ can also be transported to neighboring neurons after secretion through exosomes. Pharmacological inhibition of dynamin-mediated endocytosis leads to accumulation of Aβ on the cell surface and further prevents the transneuronal transmission of Aβ. Although potential involvement of LRP1 in the Aβ transcytosis pathway is not clear, it is tempting to speculate that the cellular uptake of Aβ through LRP1-dependent pathway might be an important step. In this regard, it might be interesting to test the effect of neuronal LRP1 deletion on Aβ propagation pathway. While lysosome has a strong ability to degrade Aβ, any disturbances of this pathway or when the accumulation of Aβ exceeds its degradation capacity could lead to Aβ aggregation in the lysosomes, thus accelerating intraneuronal Aβ aggregation and deposition. In fact, lysosomal enzymes, cathepsins B and D, seem to be reduced when Aβ accumulates in the lysosomes of amyloid model mice. In addition, if these lysosome-initiated Aβ aggregates are spread through neuronal connections, it may contribute to the propagation of Aβ aggregation as well as neuronal toxicity.

12Cellular Aβ uptake by glial cells (i.e., astrocytes and microglia) is likely to represent alternative Aβ clearance pathways (Figure 1). When adult mouse astrocytes are co-cultured with brain sections from amyloid model mice containing Aβ deposition, Aβ levels in these sections are reduced. Excessive astrocyte activation is a common pathological feature of AD; while activated astrocytes promote neurodegeneration, they might also have protective functions by facilitating Aβ clearance. Interestingly, exogenous adult astrocytes can efficiently eliminate Aβ in an apoE-dependent manner, perhaps in a manner that depends on LRP1 function. Several in vitro experiments have also shown that LRP1 controls Aβ uptake in astrocytes and further mediates Aβ-induced astrocyte activation. In vivo studies using astrocyte-specific Lrp1 knockout mice should address the specific role of LRP1 in astrocyte-mediated Aβ clearance.

13In microglia, soluble Aβ is likely internalized by fluid-phase macropinocytosis into lysosomes. On the other hand, microglia takes up larger Aβ fibrils through phagocytosis via a multi-component cell surface receptor complex. Importantly, the uptake of Aβ-coated yeast particles in microglia was suppressed by the presence of LRP1 ligands including lactoferrin, α2M* or RAP, indicating that LRP1 might regulate Aβ phagocytosis (Figure 1). LRP1 is also shown to mediate phagocytosis of apoptotic cells by binding to cell surface calreticulin in macrophages, suggesting that LRP1 may play a role in eliminating apoptotic cells containing Aβ. In addition, ABCA7, another membrane protein that is implicated in AD risk, co-localizes with LRP1 on cell surface and enhances the phagocytosis of apoptotic cells through LRP1 in macrophages. Whereas CR1 is a receptor for the complement fragments C3 and C4b, LRP1 directly binds to C1q which triggers a complement activation cascade, affecting phagocytic function. Of note, several inflammation-related genes expressed in macrophage/microglia, including TREM2, CD33, CR1, and ABCA7, have been shown to be related to the risk of late-onset AD and are generating great interests in the AD research community. Further studies may demonstrate cooperative roles among these GWAS-identified molecules and LRP1 for Aβ clearance in microglia. Interestingly, LRP1-deletion exacerbates inflammation by activating the NFκB pathway in peripheral macrophages. Thus, it is possible that LRP1 also regulates Aβ uptake in glial cells by controlling inflammatory responses and phagocytic machinery.

14The cerebrovascular system, which is composed of endothelial cells, vascular mural cells (i.e., vascular smooth muscle cells and pericytes) and astrocytes, plays critical roles in maintaining brain homeostasis and perturbations of this system lead to neuronal loss and cognitive decline. Interestingly, epidemiological studies have clearly shown that cerebrovascular disturbances, including diabetes mellitus, atherosclerosis, stroke, hypertension, transient ischemic attacks, microvessel pathology, and smoking, are risk factors for AD. Vascular smooth muscle cells, which are attached to endothelial cells and covered by astrocytes, are a major component of intracerebral arteries. At the brain capillary level, vascular endothelial cells and pericytes attached to the basement membrane form the blood-brain barrier (BBB) together with astrocyte end-foot processes. Importantly, the cerebrovascular system is yet another major pathway that mediates brain Aβ clearance by either transporting Aβ out of the brain via BBB or degrading it in vascular mural cells, which include vascular smooth muscle cells in cerebral arteries and pericytes in the capillaries (Figure 1;; ).

15Low-density lipoprotein receptor-related protein 1 is highly expressed in cerebrovasculature. In vascular smooth muscle cells, serum response factor and myocardin suppress Aβ clearance by down-regulating LRP1. Our group has also directly demonstrated that conditional deletion of LRP1 in vascular smooth muscle cells in amyloid model mice exacerbated Aβ deposition as amyloid plaques and CAA (Figure 1; ). While LRP1 mediates lysosomal Aβ degradation in vascular smooth muscle cells, LRP1 may be involved in controlling the cerebrovascular function as a signal transducing receptor through coupling with other receptors as described in previous sections. Since the cerebroarterial pulsations provide the driving force for drainage of ISF along the cerebrovasculature, deletion of LRP1 in vascular smooth muscle cells may cause Aβ accumulation by disturbing overall cerebroarterial functions as well as cellular Aβ clearance. In addition, pericyte loss caused by haploinsufficiency of Pdgfrβ gene suppresses Aβ clearance and accelerates Aβ deposition in amyloid model mice. Given that LRP1 mediates Aβ uptake and lysosomal degradation in cultured pericytes; LRP1 is likely a critical player in Aβ clearance within the cerebrovascular system (Figure 1).

16Low-density lipoprotein receptor-related protein 1 is expressed in mouse brain capillaries and mediates Aβ binding, internalization, and clearance (Figure 1; ), although the LRP1-independent pathway is also shown. Vascular endothelial cells are major components of brain capillaries and BBB, which critically regulates the influx and efflux of components between cerebral ISF and blood flow. Thus, BBB breakdown may lead to the disturbance of endothelial cell-mediated Aβ clearance across BBB. In vitro BBB model using primary cultures of mouse endothelial cells harboring mutated LRP1 endocytosis signal has shown that radiolabeled Aβ is transcytosed through LRP1 rather than degraded. On the other hand, other groups using gene overexpressing or knockdown method have demonstrated that endothelial cell lines can internalize and degrade Aβ through LRP1. Thus, further studies using endothelial cell-specific Lrp1 knockout mice are needed to assess how much endothelial cell-internalized Aβ undergoes transcytosis through LRP1 at the BBB. LRP1 antisense treatment reduced Aβ clearance at the BBB, resulting in exacerbated brain Aβ accumulation and cognitive impairment. Interestingly, LRP1-mediated Aβ clearance at the BBB was reduced by apoE in an isoform-dependent manner (apoE4 > apoE3 > apoE2; ). Furthermore, apoE4 activates the cyclophilin A-MMP9 pathway through LRP1 in pericytes, leading to damages of BBB integrity. Thus, apoE isoforms differentially regulate Aβ clearance from the brain through several LRP1-regulated pathways in the cerebrovascular system.

17In leptomeningeal arteries from AD patients with CAA, Western blot analysis revealed elevated Aβ levels and lower levels of LRP1 compared with age-matched, non-demented controls. Immunohistochemical analysis also showed that LRP1-positive vessels were reduced in patients with AD and cerebrovascular β-amyloidosis Dutch-type compared to controls, although there are also conflicting reports. While Aβ exposure decreased LRP1 levels in endothelial cells in a dose-dependent manner, hypoxia or reactive oxygen species, conditions often detected in AD, reduced LRP1 expression in vascular smooth muscle cells. In fact, LRP1 expression in brain microvasculature is known to decline in an age-dependent manner in mice. Therefore, aging or vascular damage may impair Aβ clearance by reducing LRP1 expression in the cerebrovascular system, which leads to eventual Aβ accumulation and aggregation as senile plaques and CAA. In addition, accumulation of copper in brain capillaries was associated with LRP1 reduction in mice. When mice were chronically treated with low levels of copper in their drinking water, copper disrupted brain Aβ clearance by decreasing LRP1 at BBB in a mouse model of AD. Thus, toxic chemical elements also appear to be involved in AD pathogenesis by influencing LRP1, although further studies are needed to clarify the pathways.

18In addition, LRP1 likely mediates Aβ clearance at the blood-cerebrospinal fluid (CSF) barrier as well as BBB. After intracerebroventricular administration, radiolabeled Aβ was eliminated from the CSF with a half-life of 17.3 min, which was significantly suppressed in the presence of RAP or anti-LRP1 antibody. These findings suggest that LRP1 is involved in the elimination of Aβ from CSF in epithelial cells at choroid plexus.

19The β-secretase BACE1 cleaves LRP1 on the cell surface, resulting in the release of the LRP1 extracellular domain termed soluble LRP1. In addition, α-secretases ADAM10 and ADAM17 are also likely involved in LRP1 shedding. In MEF cells lacking ADAM10 and/or ADAM17, LRP1 shedding was significantly decreased, while overexpression of ADAM10 increased the release of soluble LRP1. While the biological functions of soluble LRP1, which normally circulates in plasma, are not fully understood, plasma soluble LRP1 level appears to be decreased in AD patients compared with control individuals. Zlokovic’s group has shown that 70–90% of Aβ is bound to soluble LRP1 in plasma from cognitively normal individuals. Furthermore, the remaining soluble LRP1 in AD patients appears to be highly oxidized, which results in a lower binding affinity for Aβ. Thus, soluble LRP1 may be a novel target as a plasma biomarker for AD, although the reliability of blood biomarkers has been questioned. When recombinant soluble LRP1 domain IV (LRPIV) or LRPIV-D3674G mutant were administrated intraperitoneally into amyloid model mice for 3 months, brain Aβ levels were significantly decreased. Since these recombinant soluble LRP1 proteins do not cross BBB, they are predicted to eliminate brain Aβ through a peripheral Aβ sink mechanism. Decreasing peripheral Aβ levels likely drives Aβ transport across the BBB, although there are conflicting reports on such a hypothesis. Pharmacological approaches to increase circulating soluble LRP1 also appears to ameliorate amyloid pathology in amyloid model mice. In addition, it is interesting to note that LRPIV-D3674G increased cerebral blood flow responses to whisker stimulation in amyloid model mice. Further studies are again needed to investigate both the mechanism and the therapeutic value of soluble LRP1 in circulation.

20Internalized Aβ has been shown to predominantly traffic through the endosome/lysosome pathways for degradation (Figure 2;;; ). Overexpression of small GTPases Rab5 and Rab7, which regulates vesicle fusion for early and late endosomes, respectively, facilities the trafficking of Aβ into lysosomes. Blocking the late endocytic pathway by Rab7 knockdown induced the enlargement of early endosomes and amyloid fibril formation. In addition, a small portion of endocytosed Aβ likely traffics through the recycling vesicles (Figure 2) as a blockage of this pathway by a constitutively active Rab11 mutant significantly accelerated cellular Aβ accumulation in the recycling endosomes.

21Recent GWAS studies have also identified several endocytosis-related genes, including BIN1, PICALM, and CD2AP, as novel AD risk genes. These genes are likely involved in clathrin-mediated endocytosis and vesicular trafficking to the lysosome. Thus, the altered endocytosis pathways likely contribute to modifying AD pathology, although it is unclear whether LRP1 function is related to these genes. LRP1 is a major clathrin-dependent endocytic receptor. We have shown that the endocytic function of LRP1 is required for neuronal Aβ uptake. Overexpression of a functional LRP1 minireceptor, mLRP4, increased Aβ uptake in neuronal cells, where the effect is reversed when LRP1 endocytic function was disturbed by either clathrin knockdown or by removal of its cytoplasmic tail. However, the endocytosis rate of Aβ was slower than another LRP1 ligand RAP. These results suggest that LRP1 plays an important role in regulating Aβ endocytosis, although it is possible that other mechanisms are also involved in the process (Figure 2). In fact, it is controversial whether LRP1 directly binds to Aβ for its endocytosis. Surface plasmon resonance (SPR) analysis showed the high binding affinity of Aβ40 to immobilized recombinant LRP1 fragments of its ligand-binding domains II and IV with Kd values of 0.57 ± 0.12 and 1.24 ± 0.01 nM, respectively. In case of Aβ42, the binding affinity for LRP1 was reduced compared with Aβ40, where Kd values for LRP II and IV domains were 3.00 ± 0.11 and 10.10 ± 0.03 nM, respectively. In contrast, reported that Aβ did not show any significant binding to these LRP1 domains immobilized to microtiter plates in a solid phase binding assay.

22It is important to note that HSPG also serves as major Aβ binding receptor on the cell surface. HSPG mediates the entry of diverse molecules such as exosomes, cell penetrating peptides, polycation-nucleic acid complexes, viruses, lipoproteins, growth factors, and morphogens into cells. The HSPG endocytosis pathway likely varies depending on the cellular context and type of extracellular ligands. In the case of Aβ, HSPG appears to provide an initial binding site for Aβ on the cell surface, where LRP1 then mediates its endocytosis (Figure 2; ) by forming LRP1-HSPG complexes. It is also possible that LRP1 constitutively endocytoses Aβ that is bound to cell surface HSPG. GPI-anchored cellular prion protein (PrPc), which is localized in cholesterol-rich lipid raft microdomains of the plasma membrane, has also been demonstrated to mediate Aβ oligomer binding on the cell surface. Interestingly, LRP1 interacts with PrPc and facilitates its clathrin-mediated endocytosis. Thus, LRP1 is required for Aβ oligomer-PrPc interaction and internalization. While LRP1 is mainly localized in non-raft regions of the plasma membrane, it is also known to interact transiently with lipid rafts under specific conditions. In addition, cell surface HSPG glypican-1 is co-localized with PrPc and recruits it to lipid rafts. Therefore, LRP1 may form a complex with PrPc and HSPG, and regulate Aβ endocytosis in either clathrin- or caveolin-dependent manner (Figure 2).

23In addition to functioning as an endocytic receptor, LRP1 may control cellular Aβ uptake by modulating signaling pathways. LRP1 is known to regulate Rac1 and RhoA activities in Schwann cell, which influences cell migration and adhesion. In fact, a part of Aβ42 oligomers is likely internalized through a dynamin-dependent and RhoA-mediated endocytic pathway in neuronal cells. While dynamin regulates both clathrin or caveolin-dependent and -independent pathways, RhoA-mediated endocytosis does not require clathrin or caveolin. Although it is not fully understood how LRP1 mediates the activity of these Rho family GTPases in different cell types, LRP1 may also be involved in RhoA-dependent endocytic mechanisms of Aβ (Figure 2). In addition, larger size Aβ aggregates are thought be taken up by cells through macropinocytosis or phagocytosis, where actin polymerization is a critical step. Since LRP1 is predicted to control cytoskeleton architectures by modifying phosphoinositide 3-kinase (PI3K)/extracellular signal-regulated kinase (ERK) and/or focal adhesion kinase (FAK)/paxillin pathways, LRP1 may also affect macropinocytosis or phagocytosis of Aβ (Figure 2).

24Low-density lipoprotein receptor-related protein 1 regulates cellular Aβ uptake and degradation in neurons, astrocytes, and microglia in brain parenchyma, and in vascular smooth muscle cells and pericytes in cerebrovasculature. It also mediates Aβ clearance at the BBB by facilitating Aβ transport from brain to blood (Figure 1). LRP1-mediated cellular Aβ uptake likely depends on diverse molecular mechanisms including: (1) endocytosis of Aβ through direct binding; (2) regulation of trafficking for other Aβ receptors such as HSPG and PrPc; (3) regulation of RhoA-dependent endocytosis pathway by controlling Rho family GTPase activity; and (4) micropinocytosis/phagocytosis of Aβ by affecting actin polymerization (Figure 2). Thus, LRP1 likely mediates cellular Aβ clearance through several endocytic pathways depending on each brain cell type.

25Apolipoprotein E is a major ligand for LRP1 and influences AD risk by affecting Aβ aggregation, cellular uptake and degradation. While decreased apoE levels reduce Aβ deposition, the pharmacological approaches to increase lipidated apoE by liver X receptor (LXR) and retinoid X receptor (RXR) agonists facilitate Aβ clearance and restore cognitive function in amyloid model mice. While apoE and Aβ can interact with each other, they also share common receptors including LRP1, LDLR, and HSPG on cell surface. ApoE likely competes with Aβ for their receptor binding but can also facilitate cellular Aβ uptake by forming apoE/Aβ complexes depending on their concentrations, apoE isoform involved, lipidation status, Aβ aggregation status and receptor distribution patterns. Dissecting how LRP1 participates in apoE-mediated Aβ clearance will be critical to develop apoE-targeted therapy for AD.

26There have been several studies investigating the effects of altered LRP1 expression on Aβ metabolism. It was shown that treatment with a hydroxymethylglutaryl-CoA reductase inhibitor, fluvastatin, increases LRP1 in mouse cerebral vessels, which results in reduced brain Aβ level. Rifampicin and caffeine also upregulate LRP1 levels at the BBB and enhance Aβ elimination from the mouse brain. In peripheral tissues, insulin facilitates the hepatic clearance of plasma Aβ by increasing cell surface LRP1 distribution in hepatocytes, which in turn affects brain Aβ clearance. Given that LRP1 can control Aβ elimination from the brain in a variety of cell types, it might be important to define potentially different LRP1-mediated Aβ clearance mechanisms in each cell type to develop novel AD therapeutic methods, which target LRP1 and its ligands.

27Taken together, it is clear that LRP1 plays a critical role in cellular Aβ uptake and brain Aβ clearance. It remains to be elucidated how much of LRP1 function depends on interplay with other mechanisms. Future studies are also needed to address how LRP1 in each cell type participates in AD pathogenesis through Aβ-dependent and/or independent pathways using both in vitro and in vivo models.

28The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.