This study was presented in part at the 25th Annual Meeting of the American Society for Bone and Mineral Research, Minneapolis, Minnesota, September 19–23, 2003.
Article first published online: 1 NOV 2004
Copyright © 2005 ASBMR
Journal of Bone and Mineral Research
Volume 20, Issue 2, pages 283–293, February 2005
How to Cite
Niemeier, A., Kassem, M., Toedter, K., Wendt, D., Ruether, W., Beisiegel, U. and Heeren, J. (2005), Expression of LRP1 by Human Osteoblasts: A Mechanism for the Delivery of Lipoproteins and Vitamin K1 to Bone. J Bone Miner Res, 20: 283–293. doi: 10.1359/JBMR.041102
The authors have no conflict of interest.
- Issue published online: 4 DEC 2009
- Article first published online: 1 NOV 2004
- Manuscript Accepted: 31 AUG 2004
- Manuscript Revised: 27 JUL 2004
- Manuscript Received: 14 MAR 2004
- low-density lipoprotein receptor-related protein;
- vitamin K;
- apolipoprotein E
Accumulating clinical and experimental data show the importance of dietary lipids and lipophilic vitamins, such as vitamin K1, for bone formation. The molecular mechanism of how they enter the osteoblast is unknown. Here we describe the expression of the multifunctional LRP1 by human osteoblasts in vitro and in vivo. We provide evidence that LRP1 plays an important role in the uptake of postprandial lipoproteins and vitamin K1 by human osteoblasts.
Introduction: Chylomicrons (CM) and their remnants (CR) represent the postprandial plasma carriers of dietary lipids. Dietary vitamin K1 is known to be transported in the circulation as part of CM/CR and is required by osteoblasts as an essential co-factor for the γ-carboxylation of bone matrix proteins. The molecular mechanisms underlying the delivery of lipophilic substances to bone are not understood. In this study, the expression and function of CM/CR receptors was examined in human osteoblasts.
Materials and Methods: Four human osteoblast-like cell lines were analyzed: two osteosarcoma lines (MG63, SaOS-2) and two telomerase-immortalized human bone marrow stromal cell lines (hMSC-TERT ‘4’ and ‘20’) after 1,25(OH)2vitamin D3 induction of osteoblastic differentiation (hMSC-TERT-OB). Receptor expression was examined by Western blotting and immunohistochemistry of normal human bone sections. Endocytotic receptor function was analyzed by cellular uptake assays using fluorescent and radiolabeled human CR. Vitamin K1-enriched CR (CR-K1) were generated in vivo after oral vitamin administration and vitamin K1 uptake by osteoblasts was measured by HPLC. The effect of CR-K1 uptake on osteocalcin carboxylation was measured by ELISA.
Results: Osteoblasts exhibit high levels of protein expression of the CR receptors LRP1 and LDLR. VLDLR is expressed to a lower degree. Immunohistochemistry of normal human bone sections showed strong LRP1 expression by osteoblasts and marrow stromal cells. Uptake of fluorescent CR by osteoblasts resulted in the typical pattern of receptor-mediated endocytosis. CR uptake was stimulated by the exogenous addition of the lipoprotein receptor ligands apolipoprotein E and lipoprotein lipase. Uptake was reduced by the known LRP1 inhibitors RAP, lactoferrin, and suramin, but not by LDL, which exclusively binds to the LDLR. Vitamin K1 uptake by hMSC-TERT-OB after incubation with CR-K1 was also shown to be sensitive to LPL stimulation and the LRP1 specific inhibitor lactoferrin. CR-K1 uptake into osteoblasts stimulated the γ-carboxylation of osteocalcin.
Conclusion: Human osteoblasts express receptors of the LDLR family with a capacity for vitamin K1 uptake through CR endocytosis, a novel mechanism for the delivery of dietary lipids and lipophilic vitamins to human bone. The current data suggest that, among the expressed receptors, LRP1 plays a predominant role.
CHYLOMICRONS (CM) ARE triglyceride-rich lipoproteins (TRL), which are synthesized in the intestine and function as carriers of dietary lipids and lipophilic vitamins in the postprandial phase. In the circulation, they are hydrolyzed by lipoprotein lipase (LPL), an enzyme residing at the luminal site of endothelial cells of the vascular wall. The majority of resulting chylomicron remnants (CR), relatively poor in triglycerides (TGs), are cleared by the liver in a process that is known to be mediated by LDLR and LRP1.(1, 2) Extrahepatic tissues, depending on the metabolic need of the respective organs, also have a capacity for CR uptake.(3) The VLDLR, expressed by endothelial cells, has been discussed to function as a peripheral CR receptor,(4, 5) but in general, the molecular uptake mechanism in extrahepatic organs is far less well characterized than in the liver. Binding of CR to all known CR receptors is mediated by apolipoprotein E (apoE), a major protein constituent of CR. Three different major apoE alleles (e2, e3, e4) exist, and the resulting protein isoforms bind to lipoprotein receptors with different affinities, with apoE4 having the highest and apoE2 having the lowest binding capacity.(6–9) The apoE genotype contributes considerably to the plasma lipoprotein profile and thereby to the risk of atherosclerosis.(7) In addition, some recent epidemiological studies have found an association of the e4 genotype with reduced BMD(10, 11) and an increased risk of fracture,(12) whereas others have not.(13, 14) Next to apoE, LPL is another key player in the postprandial metabolism of CR that enhances cellular CR-binding and uptake through CR receptors, independent of its catalytic activity.(15, 16)
Accumulating experimental and epidemiological data indicate that various dietary lipids, which are known to be part of postprandial CR, play an important role in bone formation. These include essential fatty acids, polyunsaturated fatty acids, and lipophilic vitamins.(17) In this context, vitamin K is of particular interest, because unlike vitamin A and D, no other plasma-carrier protein next to lipoproteins has been described. Whereas vitamin A and vitamin D are known to be transported by CR in the early postprandial phase, they also bind to the retinol-binding protein (RBP) and vitamin D-binding protein (DBP) for later distribution and target organ uptake in the fasting period.(18, 19) Vitamin K, in contrast, remains associated with CR to a high percentage in the fasting state and is not transferred to other plasma carriers than lipoproteins.(20) The primary dietary form of vitamin K is phylloquinone (vitamin K1). Next to CR, which carry 70–90% of plasma vitamin K1, minor proportions are bound to LDL and high-density lipoproteins (HDL).(20) Osteoblasts require vitamin K as a co-factor for the enzyme γ-carboxylase that accomplishes the post-translational conversion of glutamic acid to γ-carboxyglutamic acid (Gla) residues.(21) The regular γ-carboxylation of osteocalcin (bone Gla protein [BGP]) seems to have important clinical implications.(22) Undercarboxylation of osteocalcin has been reported to be associated with an increased risk of fracture in some studies,(23, 24) although this association was not confirmed by others.(25) However, oral supplementation with vitamin K1 as well as K2, can correct undercarboxylation of osteocalcin.(26, 27) Taken together, these data suggest that an impaired delivery of vitamin K to osteoblasts may affect bone quality. This leads to the question of how vitamin K is taken up by the osteoblast.
Although CR are known carriers of the above-mentioned dietary fatty acids and vitamins that are known to affect bone metabolism, few studies have investigated CM or CR uptake into bone. Three reports exist in the literature about clearance of intravenous-injected radiolabeled CM by the bone marrow in several species.(28–30) These studies focused on the overall organ distribution and relative quantification of uptake. They describe the bone marrow as an organ of major quantitative relevance in CM uptake in rodents and mammals. However, the question of which bone cells participate in the uptake process and which molecular uptake mechanisms are at work have not been addressed in detail. One recent paper provides evidence that human osteoblast-like cells are able to take up lipoproteins, but the underlying molecular mechanisms have not been identified.(31)
Receptor-mediated endocytosis of CR may represent the mechanism for delivery of dietary lipids and lipophilic vitamins to bone, in particular to the anabolic osteoblast population and their stromal cell precursors. Thus, this study aimed at investigating the expression of known CR receptors in human osteoblasts and their potential function in CR and vitamin K1 metabolism.
MATERIALS AND METHODS
Antibodies and reagents
The mouse monoclonal antibodies 8G1 and 5A6 against LRP1 and the chicken polyclonal against the LDLR were obtained from Progen. The hybridoma cell line producing the mouse monoclonal antibody 6A6 against the VLDLR was bought from ATTC. The polyclonal rabbit anti-apoE antibody was from DAKO. 1,25(OH)2vitamin D3 was from Sigma. The bovine LPL was a kind gift from G Olivecrona (Umea, Sweden), recombinant RAP was from M Merkel (Hamburg, Germany), and the polyclonal anti-RAP antibody was from J. Gliemann (Aarhus, Denmark). Oral administration of vitamin K1 was with 10 mg Konakion tablets (Roche, Basel, Switzerland). Heparin and tetrahydrolipstatin (THL) were purchased from Roche, and bovine lactoferrin and suramin were from Sigma.
All cells were grown in DMEM supplemented with 10% FCS and penicillin/streptomycin at 37°C in 5% CO2. Derivation and characterization of hMSC-TERT has been previously described.(32) Two sublines derived from the parental cells were used in this study: hMSC-TERT(4) and hMSC-TERT(20). These sublines were established based on differences in their morphology and growth rates (M Kassem, unpublished data, 2004). To induce osteoblast differentiation, hMSC-TERTs were cultured in the presence of 10−8 M 1,25(OH)2vitamin D3 for 6 days. The osteoblastic phenotype was confirmed by upregulation of osteocalcin by RT-PCR as described previously.(32)
Plasma and lipoprotein isolation, hydrolysis, and labeling
CM were obtained from plasma of a nonfasted patient with an apoC-II deficiency by density ultracentrifugation (density ≤ 1.006 g/ml) and hydrolyzed in vitro to obtain CR as described before.(16) For uptake studies, CR were labeled with125I by the iodine monochloride method(16) or with the fluorochrome Cy3 from Molecular Probes according to the manufacturer's recommendations.
For CR-K1 generation, healthy male individuals (age, 23–38 years) carrying the apoE3/3 genotype underwent a standardized oral fat load in the form of a heavy breakfast (1950 kcal, 46% fat, 37% carbohydrates, and 17% protein), supplemented with or without an oral dose of 10 mg vitamin K1. In contrast to the apoC-II-deficient patient, the lipoprotein fraction with a density ≤1.006 g/ml from healthy individuals is always partially hydrolyzed by normal LPL activity. These particles were isolated 4 h postprandially by ultracentrifugation without further in vitro hydrolysis and designated CR (without supplementation) and CR-K1 (with supplementation of vitamin K1).
CM and CR uptake experiments and confocal laser microscopy
For the uptake experiments, osteosarcoma and hepatoma cells were used as confluent monolayers at day 2 after plating, whereas hMSC-TERT were used after 6 days of osteogenic differentiation. CR were added in a concentration of 4 μg protein/ml incubation media in radioactive and 1 μg/ml in fluorescence uptake assays. All data-points were obtained in at least duplicate. Uptake was performed as previously described.(4) Immunodetection of apoE was performed, and confocal laser microscope images were taken as described.(33)
Vitamin K1 uptake assays
For vitamin K1 uptake assays, CR-K1 were used in a concentration of 20 μg protein/ml, corresponding to 40 ng vitamin K1/ml. Uptake was allowed for 60 minutes at 37°C as in the radioactive uptake assays. To extract the lipophilic vitamin K1, cells were treated with ethanol and 100 ng vitamin K2 as an internal standard for 30 minutes at room temperature in the dark. Hexane was added in a ratio of 5:1 to ethanol (v/v), followed by another 30-minute incubation in the dark. The hexane supernatant was removed and further processed for HPLC analysis as described below. The ethanol phase was allowed to evaporate overnight, cell proteins were solubilized with 0.1 N NaOH, and the protein concentration was measured as described.(16)
Vitamin K1 HPLC analysis
For plasma, vitamin K1 was analyzed by HPLC in modification of methods described earlier.(34, 35) Briefly, 0.5 ml plasma was incubated with 1 ml ethanol, 50 μl hexane, and 2 μg/ml vitamin K2 (Sigma) as an internal standard for 30 minutes and was extracted with 5 ml hexane. Hexane was evaporated, and the residue was dissolved in 2 ml hexane. Samples were passed through a SepPak silica cartridge (500 mg silica; Waters, Eschborn, Germany) conditioned with 2 ml hexane, followed by a wash with 2 ml hexane, and elution with 5 ml hexane/ether 95:5. The eluate was evaporated and dissolved in 300 μl acetonitrile/methanol 3:1. For vitamin K1 extraction from osteoblasts, 0.75 ml 0.9% NaCl, 1.5 ml ethanol, and 50 μl hexane with 2 μg/ml vitamin K2 were added to each well of a 6-well plate. After 30 minutes, 7.5 ml hexane was added per well. The liquid supernatant was centrifuged 5 minutes at 2000g in a screw-cap vial, and the hexane layer was further processed as described for plasma. All glassware was rinsed with hexane before use. Incubations and centrifugations took place in the dark. HPLC analysis was performed in variation to the method of Haroon et al.(34) with an Agilent LC 1100 (Agilent) and an external Shimadzu RF-10axl fluorescence detector (Shimadzu) linked to the LC 1100 through a Hewlett-Packard 35900 interface (Agilent). Stationary phase was a 150 × 4.5-mm Prevail C18 5-μm column with a 4-mm precolumn (Alltech). Reduction was accomplished with a 50 × 2-mm Peek column filled with analytical grade zinc, grid size <45 μm. The mobile phase was composed of acetonitrile/methanol 3:1 with a 5-ml/liter additive of a solution containing 2.0 M ZnCl2, 1.0 M Na-acetate, and 1.0 M acetic acid in methanol. For quantification, calibration curves with vitamin K1 (Sigma) and vitamin K2 as internal standard were established. Data were processed with Agilent ChemStation. The analytical procedure was validated by adding 1.0 ng of vitamin K1 to hMSC-TERT-OBs. Recovery was 0.98 ± 0.06 ng (n = 6; two injections per sample).
Immunohistochemistry and immunoblotting
Bone and liver samples were obtained from the tissue bank of the Institute of Pathology, University Hospital Odense, Odense, Denmark. Tissues were derived from patients undergoing various surgical procedures after informed consent and acceptance of tissue sampling by the local scientific-ethical committee. Paraffin-embedded sections were deparaffinized and boiled in TRIS-buffer in the microwave for 15 minutes for antigen retrieval. Incubation with the monoclonal anti-LPR1 antibody (8G1; Progen) was performed in a dilution of 1:20 for 60 minutes at room temperature. Detection was carried out with the EnVision system (DAKO). For Western blot analysis, cells were grown under standard conditions, and cell membrane proteins were isolated by homogenization in 20 mM Tris-HCl, pH 7.4, 2 mM MgCl2, 0.25 M sucrose, 10 μg/ml leupeptin, 20 μg/ml aprotinin, and 5 μg/ml pepstatin A, followed by a sequential centrifugation at 4°C for 15 minutes at 800g and 60 minutes at 100,000g. Pellets were then resuspended in 50 mM Tris-HCl, pH 8.0, 2 mM CaCl2, 80 mM NaCl, and 1% TritonX 100 with the above proteinase inhibitors. A last centrifugation was for 30 minutes at 100,000g, protein concentrations in the supernatant were determined, and 20 μg of membrane protein was loaded per lane and separated by 8% SDS PAGE. Blotting to nitrocellulose was followed by immunodetection with ELC (Amersham).
Measurement of intracellular γ-carboxylated osteocalcin
For the determination of intracellular γ-carboxylated osteocalcin, CR-K1 were isolated, and uptake assays were performed with hMSC-TERT-OBs (4) as described above. CR-K1 were used in a concentration of 40 μg protein/ml, corresponding to 20 ng vitamin K1/ml in this particular CR-K1 preparation. LPL was added in a concentration of 1 μg/ml and recombinant human apoE3 at 10 μg/ml. Total cell protein was extracted by solubilizing cells in 50 mM Tris, pH 8.0, 2 mM CaCl2, 80 mM NaCl, and 1% Triton X100, followed by a centrifugation for 20 minutes at 13.000 rpm. The amount of intracellular γ-carboxylated osteocalcin was quantified by a commercially available ELISA (TaKaRa) that specifically recognizes the carboxylated glutamic acid residues of osteocalcin. γ-Carboxylated osteocalcin concentrations were calculated in picograms per milligram total cell protein.
Endocytosis of chylomicrons and their remnants by human osteoblasts
The first set of experiments was carried out to investigate whether human osteoblasts are able to take up TRL and whether the process resembles uptake by liver cells, for which this process has been extensively characterized and is known to involve receptor-mediated endocytosis through the LDLR and LRP1. For this purpose, MG63, SaOS-2, and hMSC-TERT-OB (4) cells were incubated with Cy3-labeled CM and CR for 20 minutes, and confocal images were taken under identical laser and detection values (Fig. 1). The resulting intracellular Cy3-signal, representing CM and CR proteins, displayed a distribution pattern comparable with that of HUH7 hepatoma cells and was typical of endocytosis.(33, 36) Because of the hydrolysis-induced enrichment with apoE and LPL, the binding of CR to lipoprotein receptors is known to be more efficient than CM binding.(15, 16) Here, the direct comparison revealed that CR were actually taken up much more effectively than CM by all cell lines. Of note, the relative difference in HUH7 cells (Figs. 1A versus 1E) was markedly less pronounced than in all osteoblast-like cell lines (Figs. 1B-1D versus 1F-1H).
To confirm the impression that CR are internalized through lipoprotein receptor-mediated endocytosis, the next experiment was performed with the addition of exogenous apoE and LPL. For this experiment, only the nontumor hMSC-TERT-OB (4) cells are shown in Fig. 2. Because of their origin from normal cells, they were considered to represent a better model for human physiology than the osteosarcoma cells. The laser was adjusted to basal CR uptake signal intensity. The addition of exogenous apoE to the incubation media (Figs. 2D-2F) clearly stimulated Cy3-CR uptake compared with uptake without the addition of extra apoE (Figs. 2A-2C). Furthermore, the addition of inactive LPL resulted in a similar increase of Cy3-CR uptake as did apoE (Figs. 2G-2I). Compared with Cy3-CR uptake without the addition of exogenous stimulators of ligand binding (compare Fig. 1H), the subcellular signal distribution within the peripheral endosomal compartment was not altered by the addition of either apoE or LPL. The experiment was performed with HUH7, MG63, and SAOS-2 cells as well, yielding comparable results (data not shown), indicating that lipoprotein receptor-mediated endocytosis of CR by osteoblasts is not a phenomenon specific to hMSC-TERT-OB but a common feature of human osteoblast-like cells in general.
To quantify the CR uptake and to discriminate between the potentially involved CR receptors, internalization of125I-CR was measured. These assays were performed with the addition of LPL. Among all known LDLR family members, LPL preferentially binds to LRP1(15) and the VLDLR.(4, 37) Figure 3 shows the stimulatory effect of inactive LPL (hatched bars) on the uptake of125I-CR uptake (white bars) in the different osteoblast-like cells in comparison with HUH7 liver cells. LPL significantly increased CR uptake by all cell lines. Total uptake in absolute terms was lower on the nontumor osteoblast-like cells compared with the hepatoma and osteosarcoma cells. The fluorescence and radioactive CR uptake experiments displayed in Figs. 1, 2, and 3 strongly indicate that osteoblasts internalize CR through lipoprotein receptor-mediated endocytosis.
Expression of lipoprotein receptors by human osteoblasts
The cellular receptor protein expression of all lines was studied by immunofluorescence (data not shown) and Western blot analysis (Fig. 4). LRP1, LDLR, VLDLR, and their common intracellular chaperone, the 39-kDa receptor-associated protein (RAP), were detectable on SAOS-2, MG63, and hMSC-TERT-OB (4 and 20) cells. LRP1 consists of two subunits: one intracellular unit of 85 kDa and one extracellular domain of 515 kDa.(38) Two forms of VLDLR transcripts, with or without an O-linked sugar domain, are generated by alternative splicing.(39) Both are recognized by the antibody we used, yielding a double band in the gel. HUH7 cells served as a positive control for the expression of LRP1 and LDLR and as a negative control for VLDLR protein expression. By comparison of equal amounts of membrane proteins, SAOS-2 (Fig. 4, lane 2) seemed to differ from the other cell lines in that SAOS-2 cells displayed a relative high amount of VLDLR protein and less LRP1 than the other cell lines. Western blots were run repeatedly under identical conditions, reproducing the pattern shown in Fig. 4 with high VLDLR expression by SAOS-2 and relative high LRP1 and LDLR expression by MG63 and hMSC-TERT-OB (lanes 3–5) compared with HUH7 (lane 1). Expression of megalin/LRP2 was not detectable by Western blotting (data not shown). These results suggest that LRP1 and the LDLR may be the predominant CR receptors on human osteoblasts.
Inhibition of osteoblast CR uptake by LRP1 inhibitors but not by LDL
A series of experiments with established inhibitors of ligand binding to CR receptors was performed to further discriminate between the relative functional importance of the LDLR, LRP1, and VLDLR. Figure 5A shows co-incubation with RAP, which represents a universal inhibitor of ligand binding to all members of the LDLR gene family.125I-CR uptake was reduced about 20–50% (hatched bars) compared with125I-CR uptake without RAP (white bars). The inhibitory effect of RAP on the nontumor cell line hMSC-TERT-OB was most pronounced (50%) and displayed the smallest SD compared with the three tumor-derived cell lines. For all subsequent experiments in this study, hMSC-TERT-OB (4 and 20) cells were used.
To narrow the number of candidate receptors on osteoblasts for CR uptake, the next experiments were performed with inhibitors of ligand binding that, unlike RAP, are specific either for the LDLR or LRP1. In contrast to apoE, the apoB100 moiety of LDL does not bind to LRP1 or the VLDLR, but exclusively to the LDLR. In the experiment in Fig. 5B, the inhibition of basic125I-CR uptake by a 30-fold excess of unlabeled CR or LDL protein was measured. Whereas CR reduced uptake by about 50% in both cell lines, LDL did not have an inhibitory effect on either cell line. In analogy, the stimulation of125I-CR uptake by LPL (about 2-fold), was reduced by CR to the same residual 50% of basic uptake, whereas LDL did not have an effect in either cell line. These data strongly indicate that the LDLR is involved in osteoblast125I-CR uptake only to a minor degree. Figure 5C shows the inhibitory effect of LRP1 specific inhibitors, namely lactoferrin and suramin. Basic125I-CR uptake by hMSC-TERT-OB (4 and 20) cells was inhibited by lactoferrin to about 40–50%. After stimulation of uptake with LPL, inhibition by co-incubation with lactoferrin and suramin reduced the LPL-stimulated125I-CR uptake by about 40–60%, with some minor variation between cell lines. Taken together, the results in Figs. 5A-5C suggest that LRP1 is the predominant CR receptor on human osteoblasts.
Co-localization of LRP1 and Cy3-CR after endocytosis and expression of LRP1 by human osteoblasts in vivo
The uptake inhibition experiments with LRP1-specific inhibitors pointed to a relatively important role of LRP1 in osteoblast CR uptake. Therefore, co-localization studies of LRP1 with Cy3-CR were performed by confocal laser scanning microscopy (Fig. 6). After uptake of Cy3-CR into hMSC-TERT-OB (4) cells, anti-LRP1 immunofluorescence stain (green) yielded a prominent co-localization of Cy3-CR (red) and LRP1 signals (Figs. 6A-6C). To determine whether the in vitro observations represent a physiologically relevant mechanism in vivo, LRP1 protein expression was analyzed by immunohistology on normal human bone sections. Strong LRP1 expression was detected on osteoblasts (arrows) and marrow stromal precursors, but not by osteocytes (Fig. 7A). Figure 7B represents the hepatocytic LRP1 expression, serving as the positive control.
Incorporation of dietary vitamin K1 into CR in vivo and osteoblast vitamin K1 uptake
To examine whether dietary vitamin K1 is taken up by osteoblasts bound to lipoproteins or through a selective uptake pathway, the following experiments were designed to directly trace vitamin K1 in osteoblast uptake assays. To be able to mimic the physiological situation as closely as possible, we isolated vitamin K1 in its naturally occurring constitution from the postprandial plasma of healthy male probands. The individuals underwent a standardized oral fat load supplemented with vitamin K1. By intraindividual comparison of fat loads with and without vitamin K1 supplementation, the oral administration of 10 mg vitamin K1 together with the fat load resulted in a >50-fold relative enrichment of CR vitamin K1 content, resulting in concentrations of 0.5-2 ng vitamin K1/μg CR protein (data not shown). Figure 8A shows a representative vitamin K1 distribution profile among the plasma lipoprotein classes from one individual. Recovery of vitamin K1 in the sum of the respective lipoprotein classes was about 90% of plasma concentrations, with 70–90% of lipoprotein-associated vitamin K1 being found in the TRL fraction. In the postprandial phase, the TRL fraction essentially consists of CR. These particles, termed CR-K1 in the following, were used to directly trace vitamin K1 in osteoblast uptake assays with hMSC-TERT-OB (4) cells. The uptake of vitamin K1, shown as picograms of intracellular vitamin K1 per milligram cell protein in Fig. 8B, was sensitive to stimulation by LPL and inhibition by lactoferrin (Fig. 8B) in a similar manner as shown for125I-CR in Fig. 5C. These data indicate that LRP1-mediated CR internalization substantially contributes to osteoblast vitamin K1 uptake.
CR-K1 induced γ-carboxylation of osteocalcin
To test whether the vitamin K1 carried by CR-K1 particles is bioactive and yields a higher degree of osteoblast γ-carboxylase activity, we performed a further series of CR-K1 uptake experiments. hMSC-TERT-OB (4) cells were incubated with CR-K1 in analogy to the uptake experiments described in Fig. 8. The consequences on γ-carboxylase activity were measured by analyzing the amount of intracellular γ-carboxylated osteocalcin after isolation of total osteoblast protein. When incubated with CR-K1 for 60 and 240 minutes, we observed a time-dependent increase in the amount of intracellular γ-carboxylated osteocalcin compared with control cells, which were treated identically except for the addition of CR-K1. Figure 9 displays the relative increase of intracellular γ-carboxylated osteocalcin for each time-point compared with controls (white bars, set to 100% for each assay). By the addition of inactive LPL or apoE, a further increase was observed, more pronounced after 240 than after 60 minutes. This increase of γ-carboxylated osteocalcin was also measured in the assay media (data not shown), indicating that the entire physiological pathway from vitamin K1 uptake to γ-carboxylated osteocalcin secretion is functionally intact. These data clearly show that the apoE- and LPL-dependent CR-K1 uptake mediated by lipoprotein receptors in human osteoblasts has functional consequences for the degree of osteocalcin γ-carboxylation.
The molecular uptake mechanism for lipids and lipophilic vitamins into bone has not been elucidated up to date. This study shows the expression of lipoprotein receptors by human osteoblasts in vitro and in vivo. We describe the expression of the LDLR, VLDLR, and LRP1 by different human osteoblast cell lines. LRP1 seems to play a predominant role in the uptake of CR by osteoblasts and is involved in the uptake of vitamin K1 in its physiological conformation (i.e., after in vivo incorporation into postprandial lipoproteins). Intracellularly, the internalized CR-K1s lead to an increase in the γ-carboxylation of osteocalcin by osteoblasts.
Hussain et al.(28–30) reported that CM are taken up by the bone marrow of several rodent and mammalian species without further in-depth characterization of the uptake mechanism. It is reasonable to assume that, among the diverse bone cell populations, the anabolic osteoblast would have a particular need for the lipophilic nutrients carried by CM/CR in the blood stream. Newman et al.(31) have published a report that addressed the question as to whether osteoblasts are able to take up lipoproteins in an apoE-dependent manner. Their data were suggestive of a lipoprotein receptor-mediated uptake process, but the question of receptor expression was neither looked at nor examined in any functional detail.
In this study, several osteoblast-like cell lines were investigated. MG63 and SAOS-2 represent widely used and well-accepted models for human osteoblasts in vitro. One disadvantage, however, is that they are tumor-derived, which may alter metabolic needs as well as cellular uptake mechanisms. Therefore, we made use of another cell line, hMSC-TERT, after osteoblastic differentiation, which is likely to be close to the naturally occurring osteoblast in bone.(32) HUH7 cells were included as a positive control because they have been well characterized for receptor-mediated lipoprotein endocytosis.(16) After initial experiments (Figs. 1, 2, 3, 4, and 5A) revealed that there were no major qualitative differences between tumor-derived and nontumor cell lines, we continued to perform experiments with the hMSC-TERT-OB cells only, because they were considered to be the most physiological cell line. The quantitative difference in CM/CR uptake (Figs. 1 and 3) between the hMSC-TERT-OB on the one side and all tumor cells on the other side may be caused by a higher need for metabolic energy of tumor versus nontumor cells. Figures 1, 2 and 6 display an endosomal pattern of ligand uptake, implying that CM and CR are internalized by human osteoblasts through endocytosis. The relative difference between the amount of CM and CR uptake is caused by better exposure of apoE on the CR surface and an enrichment of CR with LPL, thereby facilitating binding to lipoprotein receptors.(15, 16) The fact that this relative difference in uptake is higher in all osteoblasts than in liver cells gives rise to the thought that the relative quantitative importance of lipoprotein receptor-mediated endocytosis may be higher on osteoblasts than on liver cells. This interpretation is supported by the fact that the known inhibitors and stimulators of ligand-receptor interaction, namely RAP, LPL, and apoE, also have a stronger relative impact on CR uptake by osteoblasts than by hepatocytes (Figs. 3 and 5). The notion that lipoprotein receptors are important molecules on osteoblasts is further strengthened by the observation that the amount of LRP1 and LDLR expressed by osteoblasts seemed to be higher than the expression by hepatoma cells (Fig. 4). However, such an interpretation must be made with caution, because Fig. 4 displays relative amounts of the respective lipoprotein receptors normalized to total cell membrane proteins (20 μg of membrane protein/lane). Therefore, Fig. 4 gives the impression of the relative importance of the respective receptors within the single cell lines rather than the absolute amounts of receptors expressed. Consequently, high expression levels in Fig. 4 do not necessarily imply high absolute values of CR uptake. Furthermore, it should be taken into account that tumors display a higher rate of cell division and consequently require higher amount of lipids for membrane synthesis than nontumor cells. Therefore, alternative pathways for lipoprotein uptake, mediated by proteoglycans or scavenger receptors, may contribute to CR uptake by the tumor cells in this study. Interestingly, the relative proportion of RAP-sensitive (i.e., lipoprotein receptor-mediated) CR uptake in the nontumor hMSC-TERT OB cells is in fact higher than in tumor cells (Fig. 5A). Osteoblast expression of VLDLR is an interesting finding in light of previous publications that have described its expression almost exclusively by endothelial cells in organs of known abundant VLDLR detection (i.e., heart, skeletal muscle, and fat).(40) Unlike the LDLR, which is ubiquitously expressed, the VLDLR and LRP1 are known to be expressed in a very distinct organ distribution pattern and with high cell-type specificity.(41, 42) Therefore, osteoblast expression of these receptors was not to be expected a priori and suggests that they have a distinct osteoblast-specific role that goes beyond the universal and sole function of the LDLR, which is to deliver blood-borne cholesterol to any single cell in the organism.(43) The immunohistochemical localization of LRP1 on normal human bone sections proves the high cellular specificity of expression in vivo (Fig. 7). The importance of LRP1 for CR uptake is underlined by the sensitivity to inhibitors of ligand binding to LRP1 (Figs. 5A and 5C). Whereas RAP represents a universal inhibitor of ligand binding to the LDLR gene family, lactoferrin and suramin are considered to be specific for LRP1.(44) Because LDL did not inhibit CR uptake at all (Fig. 5B), the LDLR is likely to play a subordinate role in the observed CR uptake. However, it is known that there exists a certain degree of functional redundancy within the receptor family in other organs, which is partly explained by the overlapping ligand spectrum of the family members.(2, 45) To finally delineate the definitive role of single receptors on osteoblasts, double or triple gene inactivations will be needed in future experiments.
Given these limitations, Fig. 8 provides clear evidence that LRP1 participates in vitamin K1 uptake by osteoblasts. This is a significant finding in view of recent epidemiological studies that had revealed the importance of vitamin K1 as well as the K1-dependent γ-carboxylation of osteocalcin for BMD and fracture rates.(22–24, 46, 47) ApoE-dependent lipoprotein receptor-mediated vitamin K1 delivery to the osteoblast might represent the physiological pathway that results in normal intracellular γ-carboxylation of osteocalcin and secondarily contributes to normal serum levels of carboxylated osteocalcin. The data shown in Fig. 9 support such a scenario. The vitamin K1 concentrations of internalized CR-K1 are sufficient to induce an increased activity of osteoblast γ-carboxylase. The probability that this is a relevant mechanism in vivo is underlined by the notion that cells in these experiments were incubated with 20 ng vitamin K1/ml, which corresponds to physiological concentrations in human plasma between the fasted and postprandial states.(20) The increase in intracellular concentrations of γ-carboxylated osteocalcin after 240 minutes may indicate that the intracellularly available γ-carboxylase is not saturated yet with the amount of CR-K1 taken up after 60 minutes. Also, the lysosomal degradation of CR-K1 and subsequent delivery of the vitamin to the endoplasmic reticulum (ER) resident γ-carboxylation system (consisting of γ-carboxylase, vitamin K1 2,3-epoxide reductase, and further regulatory proteins) as well as activation of the system may need more time than 60 minutes. In light of these findings, one would expect a higher degree of undercarboxylated serum osteocalcin in cases of apoE deficiency or reduced apoE binding capacity. In accordance, Schilling et al.(48) describe serum osteocalcin undercarboxylation in apoE knockout mice in an accompanying manuscript. They also show that the bone phenotype of apoE-deficient mice is similar to that of mice lacking functional osteocalcin,(49) which is in full agreement with the above hypothesis.
Another interesting aspect, recently brought up in the literature, is that vitamin K1 stimulates osteogenic(50) and inhibits adipogenic(51) differentiation of marrow stromal cells. Therefore, an advanced understanding of the underlying cellular vitamin K1 uptake mechanisms both in osteoblasts and adipocytes and their respective progenitor cells is of fundamental interest to bone physiology.
One major strength of this study is the fact that vitamin K1 is incorporated into CR through the physiological route in vivo. Furthermore, native CR-K1 were used for uptake assays, and vitamin K1 was directly measured by HPLC. Therefore, potential secondary label-induced artifacts were ruled out. We believe that by choosing this experimental system, the human physiological situation was mimicked as closely as possible in vitro.
In addition to CR and vitamin K1 uptake, LRP1 may well have further functions on osteoblasts, for example, by involvement in signal transduction pathways. Both the VLDLR and LRP1 have been described to be involved in cellular signaling in other organs.(52, 53) It is particularly tempting to speculate about signaling through LRP1 on osteoblasts. Two distant relatives of LRP1, namely LRP5 and LRP6, are Wnt co-receptors,(54) and LRP5 is known to play a central role in the Wnt/β-catenin signaling pathway in the osteoblast.(55) Clearly, in vivo studies will be needed to prove the functional importance of LRP1 on osteoblasts. Because LRP1 knockout animals are not viable and die perinatally,(56) we are currently establishing an osteoblast-specific LRP1 knockout mouse model to learn more about the above-mentioned open questions.
In conclusion, aspects of lipid and lipoprotein metabolism of osteoblasts have traditionally been largely neglected in bone biology research. In recent years, however, a growing body of evidence has emerged that lipid constituents of postprandial lipoproteins are crucial to normal bone formation. With LRP1, as well as the VLDLR and LDLR to a minor degree, this study describes a molecular mechanism for the uptake of dietary lipids, in particular, the lipophilic vitamin K1, into human osteoblasts and its consequences for the carboxylation status of osteocalcin.
The authors thank Dr Henrik Schroeder (Institute of Pathology, University Hospital Odense) for providing bone samples for immunohistochemistry, Magdalini Tozakidou (Hamburg) for measuring vitamin K, and Walter Tauscher (Hamburg) for excellent technical assistance. This work was supported by the Deutsche Forschungsgemeinschaft Grants Ni 637/1-1 to AN and GRK336 to UB, and grants from the Danish Medical Research Council, Danish Stem Cell Center, and Novo Nordisk Foundation (MK).
- 11995 Receptors for triglyceride-rich lipoproteins and their role in lipoprotein metabolism. Curr Opin Lipidol 6: 117–122.
- 21998 Inducible inactivation of hepatic LRP gene by cre-mediated recombination confirms role of LRP in clearance of chylomicron remnants. J Clin Invest 101: 689–695., , ,
- 31997 Clearance of lipoprotein remnant particles in adipose tissue and muscle in humans. J Lipid Res 38: 2335–2343., , , ,
- 41996 VLDL receptor mediates the uptake of human chylomicron remnants in vitro. J Lipid Res 37: 1733–1742., , , , ,
- 52000 LDL receptor deficiency unmasks altered VLDL triglyceride metabolism in VLDL receptor transgenic and knockout mice. J Lipid Res 41: 2055–2062., , , , , ,
- 61999 Comparison of the LDL-receptor binding of VLDL and LDL from apoE4 and apoE3 homozygotes. Am J Physiol 276: E553–E557., , , ,
- 71999 Apolipoprotein E: From atherosclerosis to Alzheimer's disease and beyond. Curr Opin Lipidol 10: 207–217.,
- 81996 Apolipoprotein (apo) E genotype and apoE concentration determine binding of normal very low density lipoproteins to HepG2 cell surface receptors. J Lipid Res 37: 1316–1324., , , ,
- 91987 Dietary fat clearance in normal subjects is regulated by genetic variation in apolipoprotein E. J Clin Invest 80: 1571–1577., ,
- 101997 Association of bone mineral density with apolipoprotein E phenotype. J Bone Miner Res 12: 1438–1445., , , , , ,
- 112002 Apolipoprotein E4 is associated with reduced calcaneal quantitative ultrasound measurements and bone mineral density in elderly women. Bone 31: 497–502., , , , , ,
- 121998 Bone fracture history and prospective bone fracture risk of hemodialysis patients are related to apolipoprotein E genotype. Calcif Tissue Int 62: 278–281., , ,
- 132001 Osteoporosis and apolipoprotein E genotype in older adults: The Rancho Bernardo study. Osteoporos Int 12: 332–335., , , ,
- 142003 Association between polymorphisms of apolipoprotein E, bone mineral density of the lower forearm, quantitative ultrasound of the calcaneus and osteoporotic fractures in postmenopausal women with hip or lower forearm fracture. Scand J Clin Lab Invest 63: 247–258., , , , ,
- 151991 Lipoprotein lipase enhances the binding of chylomicrons to low density lipoprotein receptor-related protein. Proc Natl Acad Sci USA 88: 8342–8346., ,
- 162002 Endothelial-derived lipoprotein lipase is bound to postprandial triglyceride-rich lipoproteins and mediates their hepatic clearance in vivo. J Mol Med 80: 576–584., , ,
- 172001 Lipids as modulators of bone remodelling. Curr Opin Clin Nutr Metab Care 4: 105–110., ,
- 181993 Human plasma transport of vitamin D after its endogenous synthesis. J Clin Invest 91: 2552–2555., , , ,
- 191990 Transport and storage of vitamin A. Science 250: 399–404., , ,
- 201998 Plasma lipoproteins as carriers of phylloquinone (vitamin K1) in humans. Am J Clin Nutr 67: 1226–1231., , , , ,
- 211985 The propeptide of rat bone gamma-carboxyglutamic acid protein shares homology with other vitamin K-dependent protein precursors. Proc Natl Acad Sci USA 82: 6109–6113.,
- 222000 Role of vitamin K and Gla proteins in the pathophysiology of osteoporosis and vascular calcification. Curr Opin Clin Nutr Metab Care 3: 433–438.
- 231993 Serum undercarboxylated osteocalcin is a marker of the risk of hip fracture in elderly women. J Clin Invest 91: 1769–1774., , ,
- 241996 Serum undercarboxylated osteocalcin is a marker of the risk of hip fracture: A three year follow-up study. Bone 18: 487–488., , ,
- 251998 Age-related changes in serum undercarboxylated osteocalcin and its relationships with bone density, bone quality, and hip fracture. Calcif Tissue Int 62: 286–289.,
- 262002 A high phylloquinone intake is required to achieve maximal osteocalcin gamma-carboxylation. Am J Clin Nutr 76: 1055–1060., , , , ,
- 272000 Vitamin K supplementation reduces serum concentrations of under-gamma-carboxylated osteocalcin in healthy young and elderly adults. Am J Clin Nutr 72: 1523–1528., , , ,
- 281997 Uptake of chylomicrons by the liver, but not by the bone marrow, is modulated by lipoprotein lipase activity. Arterioscler Thromb Vasc Biol 17: 1407–1413., , , ,
- 291989 Chylomicron metabolism. Chylomicron uptake by bone marrow in different animal species. J Biol Chem 264: 17931–17938., , , , ,
- 301989 Chylomicron-chylomicron remnant clearance by liver and bone marrow in rabbits. Factors that modify tissue-specific uptake. J Biol Chem 264: 9571–9582., , , , ,
- 312002 The uptake of lipoprotein-borne phylloquinone (vitamin K1) by osteoblasts and osteoblast-like cells: Role of heparan sulfate proteoglycans and apolipoprotein E. J Bone Miner Res 17: 426–433., , , , ,
- 322002 Telomerase expression extends the proliferative life-span and maintains the osteogenic potential of human bone marrow stromal cells. Nat Biotechnol 20: 592–596., , , , , , ,
- 332003 Recycling of apoprotein E is associated with cholesterol efflux and high density lipoprotein internalization. J Biol Chem 278: 14370–14378., , , , , ,
- 341986 Liquid-chromatographic determination of vitamin K1 in plasma, with fluorometric detection. Clin Chem 32: 1925–1929., ,
- 351993 Serum vitamin K concentration in pediatric patients receiving total parenteral nutrition. JPEN J Parenter Enteral Nutr 17: 465–467., ,
- 361999 Intracellular processing of endocytosed triglyceride-rich lipoproteins comprises both recycling and degradation. J Cell Sci 112: 349–359., ,
- 371995 Enhancement of the binding of triglyceride-rich lipoproteins to the very low density lipoprotein receptor by apolipoprotein E and lipoprotein lipase. J Biol Chem 270: 15747–15754., , , , , , ,
- 381990 Proteolytic processing of the 600 kd low density lipoprotein receptor-related protein (LRP) occurs in a trans-Golgi compartment. EMBO J 9: 1769–1776., , ,
- 391997 Breast carcinoma epithelial cells express a very low-density lipoprotein receptor variant lacking the O-linked glycosylation domain encoded by exon 16, but with full binding activity for serine proteinase/serpin complexes and Mr-40,000 receptor-associated protein. Eur J Biochem 248: 583–591., , , , , , , ,
- 401996 Expression of the VLDL receptor in endothelial cells. Arterioscler Thromb Vasc Biol 16: 407–415., , ,
- 411994 Mouse very-low-density-lipoprotein receptor (VLDLR) cDNA cloning, tissue-specific expression and evolutionary relationship with the low-density-lipoprotein receptor. Eur J Biochem 224: 975–982., , , , , ,
- 421992 Distribution of the alpha 2-macroglobulin receptor/low density lipoprotein receptor-related protein in human tissues. Cell Tissue Res 269: 375–382., ,
- 431981 Regulation of plasma cholesterol by lipoprotein receptors. Science 212: 628–635., ,
- 442000 Relative roles of LDLr and LRP in the metabolism of chylomicron remnants in genetically manipulated mice. J Lipid Res 41: 205–213., , , , ,
- 451999 Reeler/Disabled-like disruption of neuronal migration in knockout mice lacking the VLDL receptor and ApoE receptor 2. Cell 97: 689–701., , , , , , , ,
- 461997 Bone health of adult hemodialysis patients is related to vitamin K status. Kidney Int 51: 1218–1221., , , ,
- 471994 Serum undercarboxylated osteocalcin correlates with hip bone mineral density in elderly women. J Bone Miner Res 9: 1591–1595., , , , ,
- 482004 Increased bone formation in mice lacking apolipoprotein E. J Bone Miner Res 20: 274–282., , , , , , , ,
- 49Ducy P and others 1996 Increased bone formation in osteocalcin-deficient mice. Nature 382: 448–452.
- 502003 Vitamin K stimulates osteoblastogenesis and inhibits osteoclastogenesis in human bone marrow cell culture. J Endocrinol 176: 339–348., , , ,
- 512000 Vitamin K(2) inhibits adipogenesis, osteoclastogenesis, and ODF/RANK ligand expression in murine bone marrow cell cultures. Bone 27: 769–776., , ,
- 522003 Diverse role of LDL receptor-related protein in the clearance of proteases and in signaling. J Thromb Haemost 1: 1663–1670.,
- 532003 Integration of endocytosis and signal transduction by lipoprotein receptors. Sci STKE 2003: E12., ,
- 542000 LDL-receptor-related proteins in Wnt signal transduction. Nature 407: 530–535., , , , , , , ,
- 55Osteoporosis-Pseudoglioma Syndrome Collaborative Group 2001 LDL receptor-related protein 5 (LRP5) affects bone accrual and eye development. Cell 107: 513–523., , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , ,
- 561992 LDL receptor-related protein internalizes and degrades uPA-PAI-1 complexes and is essential for embryo implantation. Cell 71: 411–421., ,