LRP6 Protein Regulates Low Density Lipoprotein (LDL) Receptor-mediated LDL Uptake

Background: Elevated serum LDL cholesterol is a major risk factor for atherosclerosis. Mechanisms that regulate LDL homeostasis are not well understood.

Results: LRP6 forms a complex with LDLR and other endocytic proteins, and its knockdown or mutation impairs LDLR endocytosis.

Conclusion: LRP6 regulates LDLR-dependent LDL uptake.

Significance: LRP6 is a potential target for development of novel lipid-lowering drugs.

Abstract

Genetic variations in LRP6 gene are associated with high serum LDL cholesterol levels. We have previously shown that LDL clearance in peripheral B-lymphocytes of the LRP6_R611C mutation carriers is significantly impaired. In this study we have examined the role of wild type LRP6 (LRP6_WT) and LRP6_R611C in LDL receptor (LDLR)-mediated LDL uptake. LDL binding and uptake were increased when LRP6_WT was overexpressed and modestly reduced when it was knocked down in LDLR-deficient CHO (ldlA7) cells. These findings implicated LRP6 in LDLR-independent cellular LDL binding and uptake. However, LRP6 knockdown in wild type CHO cells resulted in a much greater decline in LDL binding and uptake compared with CHO-ldlA7 cells, suggesting impaired function of the LDLR. LDLR internalization was severely diminished when LRP6 was knocked down and was restored after LRP6 was reintroduced. Further analysis revealed that LRP6_WT forms a complex with LDLR, clathrin, and ARH and undergoes a clathrin-mediated internalization after stimulation with LDL. LDLR and LRP6 internalizations as well as LDL uptake were all impaired in CHO-k1 cells expressing LRP6_R611C. These studies identify LRP6 as a critical modulator of receptor-mediated LDL endocytosis and introduce a mechanism by which variation in LRP6 may contribute to high serum LDL levels.

Introduction

Elevated serum LDL cholesterol is a major risk factor for atherosclerosis and myocardial infarction (1). Despite great advances in development of effective lipid-lowering drugs, an adequate control of serum lipids in patients with very high serum LDL levels is seldom achieved (2). The major determinant of plasma LDL cholesterol levels is the rate of LDL clearance from the plasma. Much of our knowledge about the LDL clearance and trafficking comes from rare Mendelian disorders that impair its endocytosis (3–7). However, the identified genetic variants account for only a fraction of inherited lipid abnormalities in the general population. Accordingly, our knowledge about mechanisms that regulate LDL clearance is far from complete.

We recently reported that LDL receptor-related protein 6 (LRP6)² regulates LDL cholesterol clearance (8). Individuals with a rare nonconservative LRP6_R611C mutation have in their third or fourth decades of life LDL cholesterol levels that are comparable with values observed in patients with heterozygote familial hypercholesterolemia (9). Furthermore, common variations within LRP6 gene have been associated with a modest elevation in serum LDL in the general population (10). We have previously demonstrated that an intact LRP6 function is necessary for normal LDL uptake (8). In the same study we showed that the splenic macrophages of LDLR^+/− mice display reduced LDL uptake compared with wild type mice. We also demonstrated that the peripheral B-lymphocytes of LRP6_R611C mutation carriers exhibit impaired LDL internalization compared with their non-carrier relatives (8). Conversely, in vitro overexpression of LRP6 in NIH3T3 cells increased cellular cholesterol uptake (same reference). Because these studies were all carried out in cells that express LDL receptor, it remained to be determined as to whether and to what extent the function of LRP6 in LDL clearance is LDLR-dependent. Furthermore, the extent of apoB binding of LRP6 was not sufficiently strong to explain the severe degree of hyperlipidemia in LRP6 mutation carriers. In this study we examined the effect of LRP6 on LDLR function and LDLR-dependent LDL uptake. In addition, the interaction between LRP6 and key proteins involved in vesicular cholesterol transport was investigated. Finally, the effect of LRP6_R611C on LDLR function and LDLR-mediated LDL uptake in CHO-k1 cells was examined.

EXPERIMENTAL PROCEDURES

Antibodies, Cell Lines, and Human Skin Fibroblasts

Antibodies for LRP6, HA tag, clathrin, caveolin-1, CD44, and β-actin were purchased from Cell Signaling Technology. Antibody for Na⁺,K⁺-ATPase 1 was from Santa Cruz Biotechnology. Dil-LDL (LDL labeled with 1,1′-dioctadecyl-3,3,3′,3′-tetramethylindo-carbocynanine perchlorate; (BT-904)), and human ¹²⁵I-LDL (BT-913R, specific activity 0.20 μCi/μg) was purchased from Biomedical Technologies Inc. Antibodies for ARH and LDLR were purchased from Novus Biologicals. Clathrin-specific shRNAs were purchased from Santa Cruz Biotechnology. CHO-ldlA7 cells were a gift from Dr. Monty Krieger. CHO-k1 cells and CHO-ldlA7 cells were maintained in F12 medium supplemented with 10% FBS and 1% penicillin-streptomycin. Wild type and Cav-1 knock-out mouse embryonic fibroblasts (MEFs) were kindly provided by Dr. Martin Schwartz at Yale. Human skin fibroblasts were obtained from LRP6_R611C mutation carriers and four unaffected relatives by routine skin biopsies. MEF, HepG2, HEK293 cells, and human skin fibroblasts were maintained in DMEM medium supplemented with 10% FBS and 1% penicillin-streptomycin.

Plasmids, Point-mutant Generation, and Cell Transfection

Vectors expressing HA-tagged LRP6_WT or HA-tagged LRP6_R611C were generated as previously described (9). Plasmid LRP6-EYFP was a gift from Dr. Christof Niehrs. LRP6_R611C-EYFP was generated by point mutation. Briefly, a C for T mutation at the nucleotide position 1831 in human LRP6 gene was introduced using QuikChange site-directed mutagenesis kit (Stratagene) as instructed by the manufacturer. A thermal cycling reaction was carried out using high fidelity DNA polymerase and complementary mutagenic primers. The forward and reverse primer sequences were 5′-CTATAGACCTCAGGGCCTTTGCTGTGGCTTGCCCTATTG-3′ and 5′-CAATAGGGCAAGCCACAGCAAAGGCCCTGAGGTCTATAG-3′ respectively. Cell transfections were carried out with FuGENE 6 (Roche Applied Science) as the per manufacturer's instructions.

Immunocoprecipitation

Cell lysates (2 mg) were subjected to overnight immunoprecipitation with HA antibody in the presence of protein A/G-Sepharose beads at 4 °C. The following day the immunoprecipitants were washed with lysis buffer three times, applied to SDS-gel electrophoresis, and subsequently analyzed by Western blotting using antibody against LDLR, clathrin, and ARH. Immunoprecipitation with LDLR antibody and immunoblotting with HA antibody was carried out similarly.

Metabolic Labeling

CHO cells were plated in 6-well plates and grew to 70% confluent. Cells were washed multiple times and incubated in 0.2 mCi/ml [³⁵S]cysteine in minimum Eagle's medium with and without Cys + 1× Gln for 30 min for pulse-chase labeling or 4 h for continuous labeling at room temperature. After several washes with cold PBS, 2 ml/well of Chase media (37 °C) were added. At the end of each time point, cells were washed, lysed in cold lysis buffer in the presence of PMSF, and placed on a rocker for 30 min at 4 °C. Lysates were centrifuged for 1 min at 14,000 rpm in a 4 °C, and the supernatant was transferred to a new set of tubes where 500 μl/tube HA monoclonal antibodies were added for immunoprecipitation. Subsequently 50 μl/tube Protein A beads were added to the samples, placed on a rocker at 4 °C overnight, and subsequently washed with cold PBS. Beads were spun down for 1 min at 4000 rpm and boiled for 5 min, and the supernatant was loaded onto gels, fixed, and exposed to films.

Stable shRNA Knockdown of LRP6 and Clathrin

The lentivirus vectors expressing LRP6 targeting shRNA (5′-CGGCGAATTGAAAGCAGTGAT-3′) were constructed as described (11). Briefly 0.5 million CHO-k1 or CHO-ldlA7 cells were plated in 6-well plates 1 day before infection. Polybrene was added into medium to the final concentration of 5 μg/ml. Lentivirus particles were added into the medium for 24 h. On the next day medium was replaced, and transduced cells were transferred to medium supplemented with 5 μg/ml puromycin.

Binding and Uptake of LDL

CHO-k1 and CHO-ldlA7 cells either overexpressing wild type LRP6 or LRP6_R611C or transfected with shGFP or shLRP6 were cultured in F-12 medium supplemented with 5% human lipoprotein-deficient serum for 24 h followed by treatment with ¹²⁵I-LDL for binding assay or with dil-LDL for uptake study. For binding assays, cells were prechilled for 30 min at 4 °C followed by adding ¹²⁵I-LDL (10 μg/ml) in F-12 culture medium supplemented with lipoprotein-deficient serum for 2 h at 4 °C. After several washes at 4 °C with DPBS, cells were incubated with 2 ml of sodium dextran sulfate (4 mg/ml) in DPBS for 1 h at 4 °C. An aliquot was placed in liquid scintillation counter to determine the total amount of ¹²⁵I-LDL bound to the cell surface. Cells were harvested, and the lysate was used to measure protein concentration. For LDL uptake, cells were incubated in lipoprotein-deficient serum with dil-LDL (10 μg/ml) for 2 h at 37 °C. Cells were harvested and washed twice with ice-cold PBS and analyzed by FACS. For treatment with lipoprotein lipase (LPL), cells were incubated in medium B supplemented with 1 μg/ml LPL inactivated with tetrahydrolipostatin (1:2,000).

Confocal Imaging

Cells were placed in 24-well plates containing poly-d-lysine-coated coverslips. After lipoprotein starvation, cells were treated with human LDL (20 μg/ml). Cells were fixed by 4% paraformaldehyde for 10 min, permeabilized with 0.05% Triton X-100 in PBS for 5 min, and blocked by 3% BSA in PBS for 1 h. Cells were then washed and incubated with 1:100 diluted antibodies for LRP6 (Abgent), clathrin (BD Biosciences), caveolin (BD Biosciences), or LDLR (Abcam) overnight at 4 °C. Subsequently cells were washed and incubated with 1:100 diluted Alexa Fluor 488 and Alexa Fluor 568 fluorescence-conjugated secondary antibodies (Invitrogen) at room temperature for 1 h. Coverslips were mounted with ProLong Gold Antifade with DAPI (Invitrogen). Specimens were examined by Nikon Ti-E Eclipse inverted microscope equipped with Perfect Focus using excitation and emission filters at 488 and 561 nm, respectively.

Sucrose Density Gradient

Cultured CHO-k1 cells 3 × 10⁸ were harvested in cold PBS on ice and lysed in 700 μl of extraction buffer (25 mm HEPES (pH 6.5), 150 mm NaCl, 1 mm EDTA, 1% Triton X-100, 1 mm PMSF, and protease inhibitor mixture). The lysate was mixed with 1 ml of 80% sucrose cushion to yield a mixture of 40% sucrose gradient and then transferred into a 12-ml ultracentrifuge tube for SW41 rotor. At the top of the sample-sucrose mixture (2 ml), 6.5 ml of 30% sucrose and 3.5 ml of 5% sucrose cushion were overlaid respectfully. Ultracentrifugation was done at 34,500 rpm for 20 h at 4 °C using a Beckman SW 41 rotor. After centrifugation, fractions were collected from the bottom of the tube with 20-gauge needles and analyzed by immunoblotting with the indicated antibodies.

Isolation of Cell Surface Proteins

Endocytosis of the LDLR or LRP6 was monitored using a protocol described previously (12). HA-tagged LRP6_WT- or LRP6_R611C-transfected CHO cells were used for LDLR endocytosis, and wild type or caveolin1 knock-out MEFs were used for LRP6 endocytosis. After 24 h of lipoprotein starvation, monensin (25 μm) and human LDL (20 μg/μl) were added and incubated for 0, 5, 20, or 60 min. Cells were plunged into ice-cold PBS to inhibit further endocytosis, washed with cold PBS, resuspended in 1 ml of PBS plus sulfosuccinimidyl-6-biotinamido hexanoate (1 mg/ml), and incubated for 30 min at 4 °C with end-over-end mixing. Samples were then processed for surface expression of LDLR using neutravidin-agarose. Cell surface proteins were eluted from the beads by adding 1× SDS loading buffer and immunoblotted using indicated antibodies.

Statistical Analysis

All experimental data represent the results from four independent experiments. Protein expression levels were quantified by densitometry of Western blots. Statistical analysis was carried out with two-factor analysis of variance. Statistical comparisons were done using Analyze-it® statistic software. A probability value of p < 0.05 was considered as statistically significant for all experiments.

RESULTS

LRP6 Regulates LDLR-independent Cellular LDL Uptake

To study the LDLR-independent effect of LRP6 on LDL uptake, CHO-ldlA7 were transfected with plasmids containing HA tagged LRP6_WT or empty vectors. These cells lack LDLR (Fig. 1A). Binding of ¹²⁵I-LDL to the cell surface was analyzed using a liquid scintillation counter, and cellular uptake of dil-LDL was examined using FACS analysis. CHO-ldlA7 cells expressing LRP6_WT showed significantly higher LDL binding at 4 °C compared with vehicle alone (p < 0.001) (Fig. 1B). Accordingly, LDL uptake at 37 °C was increased in cells expressing LRP6_WT by 25% compared with cells transfected with vehicle alone (p < 0.01) (Fig. 1C).

FIGURE 1.

LDLR-independent binding and internalization of LDL by LRP6. CHO-ldlA7 cells were transfected with plasmids encoding HA-tagged LRP6_WT, LRP6_R611C, or with vector alone (CTL). CHO-LdlA7 cells lack LDLR (A). Cells were cultured in 5% human lipoprotein-deficient serum for 24 h followed by the addition of ¹²⁵I-LDL (10 μg/ml) for 2 h at 4 °C. For LDL uptake dil-LDL (10 μg/μml) was added to the medium at 37 °C for 2 h and analyzed by FACS. Cells expressing LRP6_WT had significantly greater LDL binding, and those expressing LRP6_R611C significantly lower LDL binding compared with controls (B). Similarly, cells expressing LRP6_WT had significantly higher LDL uptake compared with vector alone (CTL) and cells expressing LRP6_R611C (C). LRP6-specific shRNA knocked down LRP6 in CHO-ldlA7 cells by >80% (D). LRP6 knockdown by RNA interference modestly reduced LDL binding (E) and uptake (F) compared with GFP shRNA (mean ± S.E.; *, p < 0.05; **,p < 0.01; ***, p < 0.001 by analysis of variance)

We next knocked down LRP6 by RNA interference in CHO-ldlA7 cells and examined its effect on LDL clearance. The shRNA reduced LRP6 protein levels by greater than 80% (Fig. 1D). RNA interference (shLRP6) resulted in a modest but significant reduction in LDL binding and internalization of these cells compared with cells transfected with GFP-shRNA (p < 0.05) (Fig. 1, E and F). These findings implicated LRP6 in LDLR independent LDL uptake.

LRP6 Modulates LDLR-mediated LDL Uptake

The modest effect of LRP6 on LDL uptake could not explain the severity of hyperlipidemia in patients with loss of function LRP6 mutation. To study the effect of LRP6 on LDLR-mediated LDL uptake, wild type CHO cells (CHO-k1) were transfected with plasmids containing HA-tagged LRP6_WT or vehicle alone. LDL binding significantly increased in cells overexpressing LRP6_WT (p < 0.01) compared with vehicle alone (Fig. 2A). Base-line LDL uptake was significantly higher in CHO-k1 compared with CHO-ldlA7 cells. Analogous to CHO-ldlA7 cells, CHO-k1 cells expressing LRP6_WT exhibited increased LDL internalization compared with vector alone (Fig. 2B) (p < 0.05). Strikingly, there was a much greater increase in LDL uptake caused by LRP6_WT overexpression in CHO-k1 versus CHO-ldlA7 cells. Conversely, LRP6 knockdown with LRP6-specific shRNA reduced LDL binding (p < 0.05) and internalization (p < 0.01) in CHO-k1 cells compared with cells transfected with a GFP shRNA(Fig. 2, C and D). In these cells LDLR was expected to compensate for the impaired function of LRP6 in LDL uptake. On the contrary, LRP6 knockdown resulted in a much greater decline in LDL binding and internalization in CHO-k1cells compared with CHO-ldlA7 cells. Furthermore, opposite to CHO-ldlA7 cells, CHO-k1 cells exhibited a steady decline in LDL clearance after LRP6 knockdown. These findings indicated impaired function of LDLR in the absence of LRP6 and strongly implicated LRP6 in LDLR-dependent LDL uptake. The LPL-facilitated LDL uptake has shown to be a LDLR-dependent process (13). The effect of LRP6 on LPL-facilitated LDL uptake was examined in CHO-k1 cells. Cells overexpressing LRP6_WT exhibited a >1.8-fold (p < 0.01) increase in LDL uptake in response to LPL compared with vector alone (Fig. 2E). LPL did not change LDL uptake in CHO-ldlA7 cells before and after overexpression of LRP6_WT (data not shown). This finding underscores the critical role of LRP6 in LDLR-mediated LDL uptake.

FIGURE 2.

Binding and internalization of LDL by LRP6 in the presence of LDLR. CHO-k1 cells were transfected with plasmids encoding HA-tagged LRP6_WT, LRP6_R611C, or with vector (CTL) alone. LDL binding and internalization assays were carried out as described. Cells expressing LRP6_WT had significantly higher LDL binding, and those expressing LRP6_R611C showed significantly lower LDL binding compared with the vector alone (A). LRP6_WT caused higher LDL uptake compared with LRP6_R611C or vector alone (B). LRP6 knockdown by RNA interference significantly reduced LDL binding (C) and uptake (D). The decrement in LDL uptake of CHO-k1 cells was more dramatic compared with CHO-ldlA7 cells, suggesting impairment of the LDLR activity. Interaction between LRP6 and LDLR was further examined by examining the effect of the LPL on LDL uptake. LPL-induced increase in uptake of LDL in CHO-k1 cells expressing LRP6_WT was twice as high compared with vector alone (E) (mean ± S.E.; *, p < 0.05; **, p < 0.01 by analysis of variance).

LRP6 Is Necessary for LDLR Internalization

The effect of LRP6 on LDLR function was assessed by the LDLR internalization in response to LDL. The rate of LDLR disappearance from the cell surface in response to LDL (10 μg/ml) was examined in CHO-k1 cells after LRP6 was knocked down by RNA interference. Cells were infected with lentivirus vectors expressing LRP6 targeting shRNA and treated with monensin to block LDLR recycling. LDLR on the cell surface was biotinylated and was isolated using neutravidin-agarose. The isolated protein was immunoblotted to assess for surface LDLR 30 and 60 min after treatment with LDL. In cells expressing sham shRNA, LDLR internalization was detectable 30 min after LDL was added to the medium. The amount of LDLR in the cell lysates before isolation of the biotinylated protein was unchanged, indicating that the loss of surface LDLR was not due to the reduction in total LDLR content. The membrane LDLR in response to LDL remained relatively unchanged when LRP6 was knocked down, indicative of its impaired internalization (Fig. 3A). The impaired function of the LDLR was rescued when cells were transfected with plasmid containing LRP6_WT (Fig. 3B). These results indicated that LRP6 is indispensable for proper LDLR internalization.

FIGURE 3.

LRP6 mediates LDLR internalization. LRP6 knockdown by RNA interference significantly impaired LDLR internalization in CHO-k1 cells treated with LDL and recycling inhibitor monensin (A). This effect was rescued with transfection of LRP6_WT (B). Sucrose density gradient centrifugation in CHO-k1 cells shows that LRP6 resides predominantly in the membrane fractions containing clathrin and LDLR compared with caveoline-1 (C). LRP6_WT and LRP6_R611C form complexes with LDLR, clathrin and ARH but not IgG (used as control) M-DRM, detergent-resistant plasma membrane (D). CHO-k1 cells were transfected with plasmids containing either LRP6_WT or LRP6_R611C. Proteins from cell lysates were immunoprecipitated (IP) with either anti-HA or anti-LDLR followed by Western blotting (IB) with either anti-LDLR or anti-HA antibodies, respectively. In addition, after immunoprecipitation with anti-HA, Western blotting was carried out with anti-ARH and anti-clathrin. Immunohistochemical studies in skin fibroblasts of R611C mutation carriers and noncarriers showed colocalization of LDLR and wild type and mutant LRP6 (E). However, LDLR/LRP6 internalization was defective in the skin fibroblasts of R611C mutation carriers. In the upper corner of the right panels higher magnification of the cell surface area, shown by the arrows, are depicted for better visualization).

LRP6 Colocalizes and Forms Complex with LDLR and Clathrin

The conventional wisdom is that LRP6 is a lipid raft receptor protein that is internalized after Wnt stimulation in a Cav-1-mediated process (14). The intriguing effect of LRP6 on LDLR internalization prompted us to readdress this issue. The subcellular localization of LRP6 in relationship to LDLR and clathrin was examined by sucrose gradient fractionation in CHO-k1 cells. The analysis showed that the majority of LRP6 resides in membrane fractions that contain clathrin and LDLR (Fig. 3C), and only a small quantity of LRP6 was present in Cav-1-containing fraction. These findings prompted further studies to examine the interaction between LRP6 and LDLR.

To determine whether LRP6 and LDLR form a complex, immunocoprecipitation studies in CHO-k1 cells transfected with plasmids either expressing HA-tagged LRP6_WT or vectors alone were carried out. Proteins from cell lysates were immunoprecipitated with anti-HA or anti-LDLR antibodies followed by Western blotting with anti-LDLR or anti-HA antibodies, respectively. The analysis showed that LRP6 forms a complex with LDLR (Fig. 3D).

Associations between LRP6, clathrin, and ARH were also examined by immunocoprecipitation studies (Fig. 3D). There was only a weak association between LRP6 and clathrin at base line that dramatically increased in presence of LDL. In contrast, the maximum physical association between ARH and LRP6 was at base line (see below). The specificity of these reactions was demonstrated by the absence of coimmunoprecipitation of LRP6 with IgG.

We examined cellular localization of the LDLR and LRP6 in normal cultured human skin fibroblasts by immunocytochemistry and confocal microscopy (Fig. 3E). LRP6 and LDLR colocalized both on the cell surface and within the cytoplasm. After stimulation with LDL, both membrane proteins translocated to the same juxtanuclear region.

LDL-mediated LRP6 Internalization Is Clathrin-dependent

We examined membrane expression of LRP6_WT in CHO-k1 cells over a time course of 60 min after treatment with LDL and in the presence of monensin. There was no change in total expression of LRP6_WT (data not shown). Membrane expression of the LRP6_WT decreased steadily during this time period, indicating its internalization in response to LDL (Fig. 4A).

FIGURE 4.

Clathrin-dependent internalization of LRP6 and its impairment by R611C mutation. LRP6 on the cell surface was biotinylated and were precipitated using neutravidin-agarose. The immunoprecipitated complex was immunoblotted to assess for surface LDLR 30 and 60 min after treatment with LDL. Wild type LRP6 started to internalize 30 min after LDL was added to the medium (A). In contrast, internalization of LRP6_R611C was significantly impaired. Clathrin-specific shRNA knocked down clathrin in MEF cells by more than 90% (B). LRP6 internalization in Cav1^−/− MEFs was comparable to those of the wild type MEFs (C) but was significantly impaired in MEFs after clathrin was knocked down. Immunohistochemical studies in normal human skin fibroblasts showed significant colocalization of LRP6 with clathrin but not with caveolin 1 (D). Clathrin and LRP6 but not Cav-1 internalized in response to LDL stimulation. Immunocoprecipitation (IP) of LRP6, clathrin, and ARH over a time course of 60 min after LDL exposure were carried out. IB, immunoblot. LRP6/clathrin immunocoprecipitation peaked 30 min after stimulation with LDL (E). ARH andLRP6_WT immunocoprecipitated, but their association decreased over time after exposure to LDL (F). LRP6_R611C immunocoprecipitated with ARH, but its association with LRP6 remained unchanged over a time course of 60 min after LDL exposure, suggesting impaired endocytosis.

Non-vesicular LDL uptake is a clathrin-independent process (15) that is associated with up-regulation of large network of invaginations originating from the plasma membrane (13). Cav-1 promotes formation of tubular invaginations (16, 17) and has been implicated in LDL trafficking within cells (18, 19). In addition, Cav-1 is required for Wnt3a-mediated LRP6 internalization and activation (14). We investigated if Cav-1 is required for LDL-mediated LRP6 internalization. LRP6 internalization in Cav1^−/− MEFs was compared with control MEFs in the presence of monensin. There was no significant difference in LRP6 internalization between the two cell types (Fig. 4C). This finding suggested that the LDL-mediated internalization of LRP6 is caveolin-independent.

The process of LDLR-dependent LDL clearance starts from binding of LDL by LDLR and its endocytosis mediated by clathrin-coated vesicles. LRP6 is a member of LDLR family with structural domains similar to LDLR. To examine the role of clathrin in LRP6 endocytosis during LDL uptake, clathrin was knocked down by RNA interference in MEFs. The RNA interference reduced expression of clathrin by more than 90% (Fig. 4B). LRP6 internalization in response to LDL was significantly impaired when clathrin was knocked down by RNA interference (Fig. 4C). Immunohistochemical studies of normal human skin fibroblasts showed colocalization of LRP6 with clathrin and their translocation to the juxtanuclear region 30 min after exposure to LDL (Fig. 4D). Surprisingly, there was no significant colocalization between LRP6 and Cav-1 in these cells before or after stimulation with LDL. These findings suggest that clathrin mediates internalization of both LDLR and LRP6.

We next examined the time course of physical association between LRP6 and clathrin in response to LDL by immunocoprecipitation studies. The immunocoprecipitation of LRP6 and clathrin was examined in CHO-k1 cells for 60 min after LDL (10 μg/dl) stimulation. Complex formation between LRP6 and clathrin was weak at base line but peaked at 30 min post-stimulation with LDL (Fig. 4E). In contrast, ARH and LRP6 formed a complex in absence of LDL. This complex is barely detectable after 30 min and is undetectable after 60 min stimulation with LDL (Fig. 4F). These findings are consistent with the roles of clathrin and ARH as endocytic and adaptor proteins, respectively (20). Taken together, these results suggest that LRP6 serves as a critical protein for vesicular LDL uptake.

LRP6_R611C Impairs LDL Uptake and LDLR Internalization

LRP6_R611C mutation carriers have such dramatically elevated serum LDL cholesterol levels that could not be solely explained by isolated impairment of the LDLR-independent LDL uptake. This raised the question as to whether R611C mutation impairs LDLR function. We expressed LRP6_R611C in CHO-k1 and CHO-ldlA7 cells and compared its effect on cell surface LDL binding and LDL internalization with those of the LRP6_WT or empty vector. There was no significant change in total expression levels of LRP6_WT and LRP6_R611C (Fig. 5A). Similarly, a pulse-chase study carried out for 120 min showed no decline in total LRP6 at the given time points, suggesting the absence of protein degradation (Fig. 5B). The membrane expression of LRP6_R611C was slightly lower than LRP6_WT (Fig. 5C). This we had previously shown to be caused by impaired recycling of the mutant protein. Accordingly, the amount of LRP6_R611C protein immunocoprecipitated with LDLR and clathrin was significantly lower compared with LRP6_WT (Fig. 3D). Expression of LRP6_R611C in both CHO cell types, however, completely failed to increase LDL binding and internalization (Fig. 1, A and B, and 2A, and B). The rate of LDLR disappearance from the cell surface in response to LDL (10 μg/ml) was examined between CHO-k1 cells expressing LRP6_WT and LRP6_R611C. The base-line expression levels of the membrane LDLR were slightly higher in cells expressing LRP6_R611C compared with LRP6_WT, likely due to a feedback mechanism triggered by the impaired LRP6-dependent LDL uptake. However, the LDLR internalization in response to LDL (10 μg/ml) was markedly reduced in CHO-k1 cells expressing LRP6_R611C compared with LRP6_WT and vector alone (Fig. 5D). Immunostaining of the LDLR and clathrin in skin fibroblasts of noncarriers of the LRP6 mutation showed redistribution of the LDLR-clathrin complex from the cell surface to a juxtanuclear region 30 min after stimulation with LDL (Fig. 5E). However, the LDLR-clathrin complex was at this time point still significantly present on the surface of the skin fibroblasts of LRP6_R611C mutation carriers. Taken together, these findings strongly suggested impaired function of the mutant receptor protein in promoting LDLR internalization and LDL uptake.

FIGURE 5.

R611C mutation impairs vesicular LDL uptake. CHO-k1 cells were transfected with vectors containing HA-tagged LRP6_R611C, LRP6_WT, or empty vectors. The total expression of LRP6_R611C and LRP6_WT was not significantly different (A). A pulse-chase study carried out to assess decay of the LRP6_R611C protein showed no change in its expression at specified time interval (B). There was slight reduction in membrane expression levels of LRP6_R611C compared with LRP6_WT (C). Membrane expression levels of LDLR in response to LDL in CHO-k1cells expressing LRP6_R611C and LRP6_WT were compared. LRP6_R611C significantly impaired LDLR internalization in (D). HMGCR, the key enzyme of the LDL synthesis, was expressed at significantly higher levels in cells expressing LRP6_R611C compared with LRP6_WT (E). Immunofluorescent staining of the skin fibroblasts from R611C mutation carriers and noncarriers using antibodies against clathrin and LDLR was carried out (F). LDLR and clathrin internalized after LDL stimulation in the fibroblasts of mutation noncarriers. In contrast, LDLR and, clathrin in the skin fibroblasts of the mutation carriers remained largely on the cell surface after stimulation with LDL.

We next compared internalization of LRP6_R611C and LRP6_WT in CHO-k1 cells treated with monensin to block recycling. The internalization of the LRP6_R611C compared with LRP6_WT was dramatically reduced (Fig. 4A). At the base line_, there was a modest association between LRP6_R611Cand ARH demonstrated by immunocoprecipitation (Figs. 3E and 4E). This association, however, remained unchanged over a time course of 60 min after LDL exposure (Fig. 4E). Taken together these findings suggested impaired vesicular endocytosis caused by LRP6 mutation. Diminished LDL uptake results in increased LDL synthesis through a feedback mechanism (21). Accordingly, the key enzyme of the cholesterol biosynthesis, HMGCR, was expressed at significantly higher levels in cells expressing LRP6_R611C compared with LRP6_WT (Fig. 5E). These studies imply the critical role of LRP6 in vesicular trafficking and suggest that the impairment of this function is an important contributor to the elevated LDL cholesterol levels in LRP6_R611C mutation carriers.

DISCUSSION

LRP6 is a member of the LDL receptor-related family, which are transmembrane cell surface proteins involved in receptor-mediated endocytosis. LRP6, however, is widely known for its role as a Wnt coreceptor in the canonical signaling pathway during embryonic development (22). Although this protein is ubiquitously expressed post-embryonically (23), its function in adult tissues has remained largely elusive. We have previously shown that individuals with rare nonconservative LRP6 mutations have LDL levels that resemble those of the individuals with heterozygote familial hypercholesterolemia (9, 10). Common variants of this gene have been associated with modest elevation of the LDL in independent populations (10). These studies have implicated the emerging role of LRP6 in regulation of the circulating LDL and as a potential target for lipid-lowering therapy.

We have previously shown that LRP6 promotes LDL uptake, and this function is impaired in the lymphocytes of LRP6_R611C mutation carriers. However, whether LRP6-mediated LDL uptake is independent of LDLR function was not examined. The most critical finding of this study is the identification of LRP6 as a regulator for LDLR-mediated LDL uptake. The evidences came from much greater alterations in LDL binding and clearance in wild type CHO compared with LDLR-deficient CHO-ldlA7 cells when LRP6 was overexpressed or knocked down. Most notably, the LDLR endocytosis was significantly impaired in CHO-k1 cells when LRP6 was knocked down or mutated. LRP6 colocalized and immunocoprecipitated with LDLR. A sucrose gradient centrifugation demonstrated that LRP6 resides in the same membrane fraction with LDLR and clathrin. Immunohistochemical studies confirmed the colocalization of LRP6, clathrin, and LDLR. Further analysis showed that LRP6 forms a complex with LDLR, ARH, and clathrin and undergoes a clathrin-dependent endocytosis after exposure to LDL. These findings suggested that LRP6 functions as a critical scaffolding protein that promotes formation of a complex between LDR and endocytic machinery and triggers their endocytosis. Dissection of the intracellular trafficking of LRP6 requires additional investigations that were beyond the scope of this study.

LRP6_R611C expression in CHO-k1 significantly impaired LDLR internalization and failed to enhance LDL binding or uptake. Because LDLR is the major regulator of the cellular LDL clearance, its altered function has possibly a major impact on the circulating LDL levels of LRP6 mutation carriers. Impaired LDL uptake in cells expressing LRP6_R611C should also trigger LDL biosynthesis. Such evidence comes from higher HMGR expression in cells expressing LRP6_R611C versus LRP6_WT. Whether and to what extent HMGCR inhibitors can reduce serum LDL in LRP6 mutation carriers remains to be determined.

In this study we also demonstrate a modest but significant role of LRP6 in LDLR-independent LDL uptake. We have previously shown that LRP6 binds apoB and is present in early endosomes. Whether LRP6 is recognized by the LDLR-related adaptor proteins is unknown at this point. LRP6 contains several putative YXXϕ motifs for potential binding to AP-2 (where X stands for any amino acid, and ϕ is for a bulky hydrophobic residue) (24–26) (27, 28). One of these motifs is in close proximity to the R611C mutation site. Interestingly, the internalization of LRP6_R611C was also significantly impaired. Additional experiments are necessary to determine which motif(s) of LRP6 is required for its recognition by adaptor proteins and for its internalization.

Cav-1 is a protein required for LRP6 internalization upon Wnt stimulation (14) and has been shown to be involved in LDL trafficking (18, 19). Cav-1 was, however, not necessary for LRP6 internalization in response to LDL. Strikingly, sucrose gradient centrifugation demonstrated that LRP6 resides mainly in fractions containing clathrin as opposed to caveolin. Taken together, these results suggest that LRP6-mediated LDL uptake is a clathrin but not a caveolin-dependent process.

In sum, our results strongly indicate that LRP6 is a critical modulator of vesicular LDL uptake. This conclusion comes from collective data ranging from identification of a rare LRP6 mutation in a large kindred with dramatically elevated LDL cholesterols, demonstration of defective LDL clearance in cells from mutation carries, and strong in vitro findings of impaired LDLR function in cells expressing mutant LRP6 or deficient for it. Thus, LRP6 should be regarded as a potential target for development of novel therapeutics in hyperlipidemia.