LRP-1-mediated endocytosis regulates extracellular activity of ADAMTS-5 in articular cartilage

Abstract

Aggrecan is a major matrix component of articular cartilage, and its degradation is a crucial event in the development of osteoarthritis (OA). Adamalysin-like metalloproteinase with thrombospondin motifs 5 (ADAMTS-5) is a major aggrecan-degrading enzyme in cartilage, but there is no clear correlation between ADAMTS-5 mRNA levels and OA progression. Here, we report that post-translational endocytosis of ADAMTS-5 by chondrocytes regulates its extracellular activity. We found 2- to 3-fold reduced aggrecanase activity when ADAMTS-5 was incubated with live porcine cartilage, resulting from its rapid endocytic clearance. Studies using receptor-associated protein (RAP), a ligand-binding antagonist for the low-density lipoprotein receptor-related proteins (LRPs), and siRNA-mediated gene silencing revealed that the receptor responsible for ADAMTS-5 clearance is LRP-1. Domain-deletion mutagenesis of ADAMTS-5 identified that the noncatalytic first thrombospondin and spacer domains mediate its endocytosis. The addition of RAP to porcine cartilage explants in culture increased the basal level of aggrecan degradation, as well as ADAMTS-5-induced aggrecan degradation. Notably, LRP-1-mediated endocytosis of ADAMTS-5 is impaired in chondrocytes of OA cartilage, with ∼90% reduction in protein levels of LRP-1 without changes in its mRNA levels. Thus, LRP-1 dictates physiological and pathological catabolism of aggrecan in cartilage as a key modulator of the extracellular activity of ADAMTS-5.—Yamamoto, K., Troeberg, L., Scilabra, S. D., Pelosi, M., Murphy, C. L., Strickland, D. K., Nagase, H. LRP-1-mediated endocytosis regulates extracellular activity of ADAMTS-5 in articular cartilage.

Articular cartilage consists of a sparse population of chondrocytes in an abundant extracellular matrix (ECM), whose major components are collagen fibrils and aggrecan proteoglycans (1). Collagen fibrils, mainly type II collagen, form a meshwork and provide the tissue with tensile strength. Aggrecan, present as large aggregated complexes interacting with hyaluronan and link proteins, forms a hydrated gel within the collagen meshwork and gives cartilage its ability to withstand mechanical compression. Chondrocytes, the only cell type present in articular cartilage, regulate tissue homeostasis by balancing the synthesis and degradation of the ECM macromolecules. A disruption in this balance results in the cartilage destruction seen in rheumatoid arthritis and osteoarthritis (OA), largely due to elevated proteolytic enzyme activities (2, 3). Because the presence of aggrecan prevents collagenolysis by collagenases of the matrix metalloproteinase (MMP) family, aggrecan loss is considered to be a crucial early event in the development of arthritis, particularly OA (4, 5).

The proteinases responsible for aggrecan degradation are MMPs and aggrecanases; the latter being members of the adamalysin with thrombospondin motifs (ADAMTS) family (3). Aggrecanases were defined by their ability to cleave the Glu³⁷³-Ala³⁷⁴ bond of the aggrecan core protein (6, 7). Elevated aggrecanase-generated aggrecan fragments were found in synovial fluids of patients with OA and inflammatory joint disease (8, 9). These fragments were also detected in normal synovial fluid and serum of animals (7), suggesting that aggrecanases function in both physiological and pathological catabolism of aggrecan.

The ADAMTSs are multidomain metalloproteinases consisting of a metalloproteinase domain, a disintegrin domain, a thrombospondin (TS) domain, a cysteine-rich (CysR) domain, a spacer (Sp) domain and a number of additional TS domains (10). Among the ADAMTSs that have aggrecanase activity, ADAMTS-4 and ADAMTS-5 have been considered as the major aggrecanases involved in cartilage matrix turnover because of their effective aggrecanase activity in vitro (11, 12). The expression of ADAMTS-4 at mRNA and protein levels correlate with the progression of OA in humans (13). In contrast, ADAMTS-5-null mice, but not ADAMTS-4-null mice, showed protection of their cartilage from destruction when challenged in an OA model induced by surgically induced joint destabilization (14, 15) or antigen-induced arthritis (16), indicating that ADAMTS-5 plays a key role in aggrecan degradation, at least in mice. ADAMTS-5 is ∼30 times more active on aggrecan than ADAMTS-4 (12). Nevertheless, mRNA levels for ADAMTS-5 in OA cartilage are not significantly elevated compared to that in normal cartilage (13, 17, 18). Treatment of human chondrocytes with the proinflammatory cytokine interleukin-1 (IL-1) increased ADAMTS-4 mRNA levels (17), but the levels of ADAMTS-5 mRNA were reported to be inconsistent and do not correlate with degradation of aggrecan in cartilage (see ref 19 for review). This led us to postulate that the aggrecanase activity of ADAMTS-5 in cartilage may be regulated at the protein level, and changes at the mRNA level may not be the major factor controlling its aggrecanase activity. The aggrecanase activity of ADAMTS-5 is inhibited by tissue inhibitor of metalloproteinases 3 (TIMP-3), which is expressed in cartilage (20). Furthermore, processing of the C-terminal ancillary domain of ADAMTS-5 reduces the aggrecanase activity (21).

Biochemical characterization of ADAMTS-5 has been carried out in vitro using purified monomeric aggrecan as a substrate. These studies do not reflect the complexity of the cartilage matrix where numerous minor ECM components, such as fibromodulin; decorin; biglycan; cartilage oligomeric matrix protein; type VI, IX, and XI collagens; matrillins; and cell surface proteoglycans assemble together with type II collagen fibrils and aggrecan (1). Furthermore, ADAMTS-5 binds to the negatively charged cell surface and ECM molecules (21), and sulfated polysaccharides, such as heparan sulfate, may regulate the aggrecanase activity of ADAMTS-5 (20). Therefore, we tested aggrecanase activity of ADAMTS-5 in the context of the cartilage matrix using dissected porcine articular cartilage, which presents a substrate close to physiological conditions. We confirmed that ADAMTS-5 has a greater aggrecan-degrading activity than ADAMTS-4, MMP-1, or MMP-13, and we found that the aggrecanase activity of ADAMTS-5 was much lower when live cartilage was used as a substrate. This was due to a rapid endocytic clearance and degradation of ADAMTS-5 by chondrocytes, which is mediated by low-density lipoprotein receptor-related protein (LRP)-1, but this endocytic pathway is dysregulated in human OA cartilage due to a loss of LRP-1.

MATERIALS AND METHODS

Reagents and antibodies

The sources of materials used were as follows: dimethylmethylene blue (DMMB), dynasore, β-cyclodextrin (β-CD), polymyxin B, and the anti-FLAG M2 mouse monoclonal antibody from Sigma-Aldrich (Dorset, UK); the anti-early endosome antigen 1 (EEA1) rabbit polyclonal antibody, and the anti-LRP-1 mouse monoclonal antibodies 5A6 and 8G1 from Abcam (Cambridge, UK) and Calbiochem (San Diego, CA, USA); the anti-actin antibody from Santa Cruz Biotechnology (Santa Cruz, CA, USA); the anti-tubulin antibody from Cell Signaling (Danvers, MA, USA); and BC-3 mouse monoclonal antibody that recognizes the N-terminal ³⁷⁴ARGSV generated by aggrecanase cleavage of aggrecan core protein from Abcam. The anti-human ADAMTS-5 catalytic domain rabbit polyclonal antibody was raised in rabbits and characterized (21). Recombinant human ADAMTS-5 and its domain-deletion mutants, ADAMTS-4 lacking the Sp domain, MMP-1, and MMP-13, were prepared as described previously (11, 21, 22). Recombinant human IL-1α was kindly provided by Prof. J. Saklatvala (Kennedy Institute of Rheumatology, London, UK). All other reagents used were of the highest analytical grade available.

Expression and purification of human receptor-associated protein (RAP)

Recombinant human C-terminally His-tagged RAP was expressed in Escherichia coli using a pET3a-based expression vector (Novagen/EMD Biosciences, Madison, WI, USA). The human RAP cDNA was isolated by polymerase chain reaction (PCR) using cDNA from HT1080 cells as a template with the sense primer 5′-TGGCATATGTACTCGCGGGAGAAGAACCAGCCCAAGCCGTCCCCGAAACGC-3′ containing an NdeI site (underscored) and RAP N-terminal sequence and the antisense primer 5′-CCAGGATCCCTAATGGTGATGGTGATGGTGGAGTTCGTTGTGCCGAGCTCT-3′ containing a BamHI site (underscored), a 6-histidine tag (in italics) and the C-terminal sequence of RAP. PCR fragments were ligated into CloneJET cloning vector (Fermentas, Glen Burnie, MD, USA), digested with NdeI and BamHI, and ligated into similarly cut pET3a. The sequence was confirmed by DNA sequencing. E. coli BL21(DE3) cells were transformed with the RAP expression plasmid, and cultures grown at 37°C in 1 L of Luria-Bertani broth with 50 μg/ml carbenicillin. Once the culture reached an OD₆₀₀ of 0.6, protein expression was induced by addition of 1 mM isopropyl-β-d-thiogalactoside for 3 h at 37°C. Cells were harvested by centrifugation and lysed in 40 ml of phosphate-buffered saline using a French press. The soluble fraction of the cell lysate was applied to nickel-nitrilotriacetic acid Superflow resin (Qiagen, Valencia, CA, USA) equilibrated in 20 mM HEPES (pH 7.5) and 150 mM NaCl. The column was washed in 20 mM HEPES (pH 7.5), 1 M NaCl, and 50 mM imidazole, and further washed with an excess amount of 20 mM HEPES (pH 7.5), 150 mM NaCl, and 60% isopropanol to remove endotoxin (23). Bound protein was eluted in 20 mM HEPES (pH 7.5), 150 M NaCl, and 500 mM imidazole. RAP-containing fractions were pooled and dialyzed against 20 mM HEPES (pH 7.5) and 150 mM NaCl. To ascertain the levels of LPS in the recombinant RAP preparation, the Limulus amoebocyte lysate assay (Cambrex, East Rutherford, NJ, USA) was used, according to the manufacturer's instructions. The recombinant RAP used in this study had an endotoxin level < 10 pg/ml.

Porcine articular cartilage culture

Porcine articular cartilage from the metacarpophalangeal joints of 3- to 9-mo-old pigs was dissected into small pieces (∼6 mm³: 10 mg wet vol). Each piece was placed in one well of a round-bottom 96-well plate and allowed to rest for 24 h in 200 μl of Dulbecco's modified Eagle's medium (DMEM) containing 10% fetal bovine serum (FBS) before use. The medium was then replaced, and the cartilage was rested for a further 24 h in 200 μl of DMEM at 37°C (for fresh live cartilage) before the aggrecan degradation and endocytosis assays. For freeze-thawed cartilage experiments, cartilage pieces were frozen at −80°C and thawed in the same medium and replaced with fresh medium before subjecting to subsequent assays.

Analysis of aggrecan degradation

Each piece of cartilage was incubated in 100 μl of DMEM with or without IL-1α (10 ng/ml) or various concentrations of MMPs and ADAMTSs. Three pieces of cartilage were subjected to each treatment. After incubation for various periods of time, the conditioned media were harvested, and glycosaminoglycan (GAG) released into the medium was measured using the DMMB assay (5). Aggrecan fragments in the medium were deglycosylated, as described previously (11), and analyzed by Western blotting using an anti-ARGSV aggrecan neoepitope antibody. Immunoreactive bands were quantified using ImageJ (U.S. National Institutes of Health, Bethesda, MD, USA) and results are presented as relative intensities.

Isolation of chondrocytes and cell culture

Human articular cartilage was obtained from patients after they provided informed consent and following approval by the Riverside Research Ethics Committee (Riverside Health Authority, London, UK). Non-OA cartilage was obtained from the knee following amputation due to soft tissue sarcoma and osteosarcoma with no involvement of the cartilage. Tissues were obtained from 8 patients (5 male, 3 female; aged 12–55 yr, mean age 31.6 yr). OA cartilage was obtained from patients undergoing joint replacement surgery. Tissues were obtained from 7 patients (2 male, 5 female; aged 53–66 yr, mean age 61 yr). Chondrocytes were isolated as described previously (20). Primary porcine cells and both primary and passaged human cells were used in the experiments. For the ADAMTS-5 endocytosis assay, cells were plated at a density of 2.5 × 10⁵ cells/well (24-well plate) in DMEM containing 10% FCS.

Analysis of ADAMTS-5 clearance

Cartilage explants or cells were incubated in 100 or 400 μl of DMEM with or without 80 μM dynasore, 10 mM β-CD, or 500 nM RAP. After incubation for 30 min, media were replaced with DMEM containing 10 nM ADAMTS-5 or its domain deletion mutants with or without 80 μM dynasore or 500 nM RAP. Four pieces of cartilage were subjected to each treatment. After incubation for various periods of time, media were collected; the protein was precipitated with TCA and dissolved in 50 μl of 1× SDS sample buffer [50 mM Tris-HCl (pH 6.8)/5% 2-mercaptoethanol, 2% SDS and 10% glycerol]. The cartilage explants and cells were washed with DMEM and lysed in 50 μl of 2× SDS-sample buffer. All samples were analyzed by SDS-PAGE and Western blotting using an anti-ADAMTS-5 catalytic domain antibody. Immunoreactive bands were quantified using ImageJ, and the amounts of ADAMTS-5 remaining in the medium and cell lysate at each time point were calculated as a percentage of the amount of ADAMTS-5 at 0 h.

Immunocytochemical localization of endocytosed ADAMTS-5

Cultured cells on 4-well Lab-Tek chamber slides (Nunc, Roskilde, Denmark) were incubated in DMEM with 10 nM ADAMTS-5 in the absence or presence of 500 nM RAP for 1 h. Cells were washed with DMEM, fixed with 3% paraformaldehyde in Tris-buffered saline (TBS; 10 min, room temperature) and permeabilized with TBS containing 10 mM CaCl₂ and 0.1% Triton X-100 (15 min, room temperature). To visualize ADAMTS-5 within the cartilage, porcine articular cartilage explants were incubated as described above for 3 h. Explants were washed with DMEM and fixed with 4% formalin (24 h, room temperature). Explants were then snap-frozen and sectioned (5-μm sections) using a CM1900 cryostat (Leica Microsystems, Wetzlar, Germany). Each sample was incubated with anti-FLAG M2 mouse monoclonal antibody and anti-EEA1 rabbit polyclonal antibody (3 h, room temperature). Alexa Fluor 488-conjugated anti-mouse IgG and Alexa Fluor 568-conjugated anti-rabbit IgG (Molecular Probes, Eugene, OR, USA) were used to visualize the antigen signals (1 h, room temperature). Actin was stained with actin stain 670 phalloidin (Cell Signaling), and nuclei were stained with DAPI. Samples were viewed using a Nikon Eclipse TE2000-U confocal laser scanning microscope (Nikon, Tokyo, Japan). The data were collated using Volocity software (Improvision, Coventry, UK).

siRNA-mediated knockdown of LRP-1 in human articular chondrocytes

siRNA oligonucleotides for LRP-1 (On-TargetPlus SMARTpool siRNA) and nontargeting oligonucleotide were purchased from Thermo Scientific Dharmacon (Lafayette, CO, USA). Human articular chondrocytes were plated at a density of 1.5 × 10⁶ cells/dish (6-cm dish) in DMEM containing 10% FCS and incubated until 50% confluent. Lipofectamine 2000 (Invitrogen, Carlsbad, CA, USA) was used to transfect cells with siRNA at a final concentration of 10 nM in Opti-MEM I. At 4 h after transfection, the Opti-MEM was removed and replaced with DMEM containing 10% FCS.

Quantitative reverse transcriptase–PCR

RNA was extracted and prepared using the RNeasy mini kit (Qiagen, Valencia, CA, USA) following the manufacturer's guidelines. cDNA was generated using a reverse-transcription kit (Applied Biosystems, Foster City, CA, USA) and random primers from 0.5 μg of total RNA. Newly synthesized cDNA was diluted 5-fold in DNase-free water. Four percent of this cDNA was then used for real-time PCR assays using TaqMan technology (Applied Biosystems). The ΔΔ threshold cycle (ΔΔC_t) method of relative quantitation was used to calculate relative mRNA levels for each transcript examined. The 60S acidic ribosomal protein P0 (RPLP0) gene was used to normalize the data. Predeveloped primer/probe sets for LRP-1 and RPLP0 were purchased from Applied Biosystems.

Western blotting analysis of LRP-1 protein

Cultured cells (2×10⁵) were lysed in 50 μl of 2× SDS-sample buffer without 2-mercaptoethanol, and a portion of the lysates was analyzed by SDS-PAGE and Western blotting using anti-LRP-1 heavy-chain and light-chain antibodies. Immunoreactive bands were quantified using ImageJ, and the amounts of LRP-1 heavy chain and light chain in cell lysates were calculated as a percentage of the amount of LRP-1 or shown as relative LRP-1 expression. Cellular tubulin was used to normalize the data.

RESULTS

Aggrecanase activity of ADAMTS-5 is greatly reduced in live compared to dead cartilage

To investigate the ability of ADAMTS-4 and ADAMTS-5 to cleave aggrecan in cartilage, recombinant ADAMTS-4 lacking the C-terminal Sp domain or ADAMTS-5 lacking the C-terminal TS2 domain was added to live or freeze-thawed porcine cartilage explants, and subsequent GAG release was monitored. These two forms of aggrecanases were used as they are readily available in a pure form (11, 21). Freeze-thawed cartilage was used to eliminate the involvement of metabolically active live chondrocytes. Freeze-thawing of cartilage-killed chondrocytes, and these cartilages did not respond to IL-1α stimulation, while live cartilage exhibited a 4.0-fold increase in GAG release into the medium on IL-1α treatment (Fig. 1A). Treatment of dead cartilage with 50 nM ADAMTS-4, 5 nM ADAMTS-5, 100 nM MMP-1, or 50 nM MMP-13 caused a 4.3-, 11.0-, 2.4-, or 2.2-fold increase in GAG release, respectively. When the rate of aggrecan cleavage was compared by measuring GAG release from dead cartilage, ADAMTS-5 cleaved aggrecan fastest, at a rate ∼19 times faster than that of ADAMTS-4 (Fig. 1B). When live cartilage was used, the amounts of GAG released by ADAMTS-4, MMP-1, or MMP-13 were similar to those released from dead cartilage (Fig. 1A). However, ADAMTS-5 showed much lower levels of GAG release when added to live cartilage. A time-course study indicated that ADAMTS-5 digested aggrecan in dead cartilage 1.6- to 2.0-fold more effectively than in live cartilage (Fig. 1C). Analysis of aggrecan breakdown products with an anti-ARGSV neoepitope antibody indicated that aggrecan was cleaved at the Glu³⁷³-Ala³⁷⁴ bond, characteristic of aggrecanase activity (Fig. 1D).

Figure 1.

Aggrecanase activity of ADAMTS-5 is greatly reduced in live compared to dead cartilage. A) Live and freeze-thawed (dead) porcine cartilage explants were incubated with 10 ng/ml IL-1α (IL-1), 50 nM ADAMTS-4 (TS-4), 5 nM ADAMTS-5 (TS-5), 100 nM MMP-1, or 50 nM MMP-13 for 24 h. GAG released into the medium was measured by DMMB assay. B) Time course of GAG released into the medium of dead cartilage explants incubated with 20 nM ADAMTS-4 or 5 nM ADAMTS-5. C) Time course of GAG released into the medium of live and dead cartilage explants incubated with 5 nM ADAMTS-5. Bars and points represent means ± sd for triplicate assays (n=3). *P < 0.05; unpaired t test. D) Western blotting analysis of the medium obtained after 24 h culture in A for aggrecan fragments using an anti-ARGSV neoepitope antibody. L, live cartilage; D, dead cartilage.

Rapid clearance of ADAMTS-5 by articular chondrocytes

To investigate the cause of reduced aggrecanase activity of ADAMTS-5 with live cartilage, we monitored the level of exogenously added ADAMTS-5 in the medium and in the cartilage explants by Western blot analysis. When added to live cartilage, ADAMTS-5 disappeared from the culture medium and was almost completely absent after 4 h without any detectable fragments (Fig. 2A, B). In contrast, relatively large amounts of ADAMTS-5 were detected in the medium of dead cartilage even after 4 h. ADAMTS-5 was also detected in extracts from dead cartilage after 2–4 h, but not in live cartilage. We could not detect endogenous ADAMTS-5, as its level was too low to detect by this method. When ADAMTS-5 was incubated with isolated chondrocytes, it was also depleted from the medium with similar kinetics to that seen in live cartilage, with a half-life of ∼80 min. We estimated ∼3.5 × 10⁶ ADAMTS-5 molecules were cleared by a single chondrocyte within 1 h (Fig. 2C).

Figure 2.

Rapid clearance of ADAMTS-5 by articular cartilage chondrocytes. A) Live (L) and dead (D) porcine cartilage explants were incubated with 10 nM ADAMTS-5 (TS-5) for 0–4 h, and ADAMTS-5 in the medium and cartilage explants was detected by Western blot analysis using an anti-ADAMTS-5 catalytic domain antibody. B) Densitometric analysis of immunoreactive ADAMTS-5 bands detected in the medium of A. Amount of ADAMTS-5 is expressed as a percentage of the amount of ADAMTS-5 at 0 h. C) Porcine chondrocytes were incubated as in A, and ADAMTS-5 remaining in the medium was measured as in B. Points represent means ± sd for triplicate assays (n=3).

LRP-mediated endocytic clearance of ADAMTS-5

To investigate the mechanism of ADAMTS-5 disappearance, we first examined the effect of proteinase inhibitors, but neither the broad-spectrum hydroxamate metalloproteinase inhibitor GM6001 nor a proteinase inhibitor cocktail (mixture of inhibitors for serine, cysteine, and aspartic proteinases) inhibited ADAMTS-5 disappearance. However, dynasore, an inhibitor of dynamin, which is required for most endocytic pathways, including clathrin- and caveolae-mediated endocytosis (24), almost completely inhibited ADAMTS-5 depletion from the medium of live cartilage (Fig. 3A). By contrast, β-CD, which inhibits caveolae-mediated endocytosis by depleting cholesterol from the cell membrane (25), did not inhibit it (Fig. 3A), indicating that ADAMTS-5 disappearance is due to a clathrin-dependent endocytic clearance. No significant cytotoxicity of dynasore on chondrocytes was found by 3-(4,5-dimethylthiazole-2-yl)-2,5diphenyl tertasolium bromide (MTT) assay (data not shown). We then found that RAP, an antagonist of ligand binding to LRP receptors, inhibited ADAMTS-5 uptake by live cartilage and chondrocytes (Fig. 3B, C), suggesting that ADAMTS-5 is endocytosed by a member of the LRPs. The LRP-dependent internalization of ADAMTS-5 was confirmed by immunofluorescent confocal microscopy. Punctate staining of ADAMTS-5 was detected within the cells, and the staining colocalized with EEA1, a marker for early endosomes (Fig. 3D). The fluorescent signal of ADAMTS-5 was absent in cells incubated without ADAMTS-5, indicating the specificity of the staining. Consistent with the data from Western blot analysis, the intracellular fluorescent signal for ADAMTS-5 was abolished in the presence of RAP. Heparin also blocked the internalization of ADAMTS-5, suggesting that basic residues of the enzyme are involved in interaction with LRP. When live cartilage was incubated with ADAMTS-5, fluorescent ADAMTS-5 signal was observed within cells located in the superficial zone of the cartilage, and this signal was abolished by RAP (Fig. 3E). ADAMTS-5 was internalized even in the presence of 50 μM GM6001 (data not shown), suggesting that the enzyme can penetrate into the cartilage without matrix degradation.

Figure 3.

LRP-mediated endocytic clearance of ADAMTS-5. A) Live porcine articular cartilage explants were first incubated with 80 μM dynasore or 10 mM β-CD for 30 min. Then, 10 nM ADAMTS-5 was added to the medium and further incubated for 0–4 h. ADAMTS-5 remaining in the medium was measured as in Fig. 2. B) Live porcine cartilage explants were incubated with 10 nM ADAMTS-5 in the presence of 500 nM RAP for 0–4 h. C) Porcine chondrocytes were incubated as in B, and ADAMTS-5 remaining in the medium was measured as in Fig. 2. Points represent means ± sd for triplicate assays (n=3). D) Confocal microscopy analysis of ADAMTS-5 endocytosis by porcine chondrocytes. Cells were incubated with 10 nM ADAMTS-5 (TS-5) in the presence or absence of 500 nM RAP or 100 μg/ml heparin for 1 h. Endocytosed ADAMTS-5, EEA1, cytoskeleton and nucleus were visualized as described in Materials and Methods. E) Confocal microscopy analysis of ADAMTS-5 endocytosis by porcine cartilage explants. Live cartilage explants were incubated with 10 nM ADAMTS-5 in a similar manner, and endocytosed ADAMTS-5 was detected as in D. Yellow arrowheads indicate the intact surface of cartilage explants.

siRNA-mediated knockdown of LRP-1 impairs ADAMTS-5 endocytosis in human chondrocytes

Among the members of the low-density lipoprotein receptor family, LRP-1 is known to mediate endocytosis of a number of proteinases and proteinase-inhibitor complexes (26). Therefore, we silenced LRP-1 in normal human chondrocytes with a gene-specific siRNA to investigate its possible involvement. As shown in Fig. 4A, LRP-1-targeting siRNA-depleted LRP-1 mRNA levels by 93% compared to the nontargeting siRNA control. Western blot analysis of the cell extracts confirmed that the levels of the 515-kDa extracellular heavy chain and the 85-kDa light chain containing the transmembrane domain were reduced by 95 and 80% by targeting siRNA compared to the nontargeting siRNA control (Fig. 4B). Cellular uptake of ADAMTS-5 was markedly reduced in LRP-1-depleted cells (Fig. 4C). Immunocytochemical analysis further confirmed that ADAMTS-5 endocytosis was impaired in the LRP-1-depleted human chondrocytes (Fig. 4D). From these results, we conclude that LRP-1 is the primary endocytic receptor for ADAMTS-5.

Figure 4.

siRNA-mediated knockdown of LRP-1 impairs ADAMTS-5 endocytosis in human chondrocytes. Human chondrocytes transfected with nontargeting siRNA (siCtrl) or LRP-1 targeting siRNA (siLRP-1) were cultured for 2 d in DMEM containing 10% FCS. A) Results of TaqMan real-time PCR showing relative levels of mRNA for LRP-1. B) Left panels: LRP-1 heavy-chain (515 kDa) and light-chain (85 kDa) in cell lysates were assessed by Western blot analysis using anti-LRP-1 heavy-chain and light-chain antibodies, respectively. Right panels: densitometric analysis of immunoreactive LRP-1 bands detected in the cell lysate. Amount of LRP-1 is expressed as a percentage of the amount of LRP-1 in untransfected cells (none). C) Cells were incubated with 10 nM ADAMTS-5 in the presence or absence of 500 nM RAP for 0–6 h, and ADAMTS-5 remaining in the medium was measured as in Fig. 2. Bars and points represent means ± sd for triplicate assays (n=3). D) Confocal microscopy analysis of ADAMTS-5 endocytosis by human chondrocytes. Cells were incubated with 10 nM ADAMTS-5 in a similar manner, and endocytosed ADAMTS-5 was detected as in Fig. 3.

TS1 and Sp domains mediate ADAMTS-5 endocytosis

ADAMTS-5 is a multidomain metalloproteinase. Therefore, we searched which domains are required for its endocytosis by testing a series of domain-deletion mutants (Fig. 5A). Since full-length ADAMTS-5 (ADAMTS-5-1) is not available in a high quantity, it was incubated with porcine chondrocytes only for 3 h. The results showed that ∼75% of ADAMTS-5-1 disappeared from the medium, and this disappearance was inhibited by RAP (Fig. 5B). Domain-deletion mutants were subjected to time-course studies (Fig. 5C). Compared to deletion of the TS2 domain alone (ADAMTS-5-2), deletion of both TS2 and Sp domains (ADAMTS-5-3) significantly reduced the rate of endocytosis. Further deletion of the CysR domain resulted in only a slight further reduction. However, the mutant without the TS1 domain (ADAMTS-5-5) was not internalized. When the mutants were tested for their aggrecanase activity using live and dead cartilage explants in culture, all except ADAMTS-5-5 showed a markedly reduced GAG release with live cartilage compared to dead cartilage (Fig. 5D). From these data, we conclude that the TS1 domain is essential for ADAMTS-5 endocytosis but that the Sp domain enhances the endocytic process.

Figure 5.

TS1 and spacer domains mediate ADAMTS-5 endocytosis. A) Schematic representation of ADAMTS-5 and its domain deletion mutants. Pro, prodomain; Cat, catalytic domain; Dis, disintegrin-like domain; TS1, thrombospondin 1 domain; Cys, cysteine-rich domain; Sp, spacer domain; TS2, thrombospondin 2 domain. B) Porcine primary chondrocytes were incubated with 10 nM ADAMTS-5-1 in the presence or absence of 500 nM RAP for 3 h, and ADAMTS-5 in the medium and cell lysate was detected as Fig. 2. C) Porcine chondrocytes were incubated with each of 10 nM ADAMTS-5-2, ADAMTS-5-3, ADAMTS-5-4, or ADAMTS-5–5 for 0–6 h, and each ADAMTS-5 mutant remaining in the medium was measured as in Fig. 2. D) Live and freeze-thawed (dead) porcine articular cartilage explants were incubated with 10 ng/ml IL-1α, 4 nM ADAMTS-5-1, 5 nM ADAMTS-5-2, 10 nM ADAMTS-5-3, 20 nM ADAMTS-5-4, 50 nM ADAMTS-5-5 for 24 h. GAG released into the medium was measured by DMMB assay. Bars and points represent means ± sd for triplicate assays (n=3). *P < 0.05; unpaired t test.

Inhibition of ADAMTS-5 endocytosis by RAP enhances aggrecan degradation in articular cartilage

We then examined the effect of RAP on normal turnover of aggrecan in cartilage. The addition of RAP to live porcine cartilage increased the constitutive GAG release an ∼2-fold, whereas GAG release from dead cartilage was not enhanced by RAP (Fig. 6A).This effect of RAP was not altered in the presence of polymyxin B, an inhibitor of endotoxin, indicating that the effect was not due to endotoxin contamination in the E. coli-expressed RAP preparation. Western blot analysis of the products with an anti-ARGSV antibody indicated that RAP-enhanced aggrecan degradation was due to an increase in aggrecanase activity (Fig. 6B). These results suggest that aggrecanases are constitutively produced by normal cartilage, but their activities are down-regulated by LRP-1-mediated endocytosis. The addition of 20 nM recombinant ADAMTS-5-4 to dead cartilage caused 1.7-fold higher GAG release over that with live cartilage after 4-h incubation, and this increased to 2.9-fold after 24 h. The addition of RAP to the ADAMTS-5-4-mediated aggrecan degradation system significantly enhanced GAG release from live cartilage but had no effect on GAG release from dead cartilage (Fig. 6C). Notably, the level of GAG released from live cartilage in the presence of RAP was equivalent to that from dead cartilage. Western blotting using an anti-ARGSV antibody confirmed that the 3.0-fold increase in aggrecan degradation observed in the presence of RAP is solely due to increased aggrecanase activity of ADAMTS-5 in live cartilage (Fig. 6D).

Figure 6.

Inhibition of ADAMTS-5 endocytosis by RAP enhances aggrecan degradation in articular cartilage. A) Time course of GAG released into the medium of live cartilage explants incubated in the presence or absence of 500 nM RAP. B) Densitometric analysis of immunoreactive bands of aggrecan fragments detected in the medium obtained after 48 h culture in A by Western blotting using an anti-ARGSV neoepitope antibody. Amount of aggrecan fragment in the medium of live cartilage explants incubated alone was taken as 1. As a control, live cartilage explants were incubated with RAP in the presence of 100 μg/ml polymyxin B (PolyB) for the indicated period of time. C) Time course of GAG released into the medium of live and freeze-thawed (dead) cartilage explants incubated with 20 nM ADAMTS-5-4 in the presence or absence of 500 nM RAP. D) Amount of aggrecan fragments in the medium obtained after 24 h culture in C were measured as in B. Amount of aggrecan fragment in the medium of live cartilage explants incubated with 20 nM ADAMTS-5-4 was taken as 1. Bars and points represent means ± sd for triplicate assays (n=3).

ADAMTS-5 endocytosis is impaired in OA cartilage due to reduced protein levels of LRP-1

Because the extracellular aggrecanase activity of ADAMTS-5 is regulated by LRP-1-mediated endocytosis, we considered whether the increased degradation of aggrecan observed in OA cartilage is due to a dysregulated endocytic pathway. Quantitative mRNA analysis of OA and normal human cartilage showed similar levels of LRP-1 mRNA (Fig. 7A). However, Western blot analysis of LRP-1 protein indicated that OA cartilage has lost most of the receptor protein (Fig. 7B). Therefore, we considered the possibility that ADAMTS-5 endocytosis is impaired in OA cartilage, making it more readily susceptible to degradation, while normal human cartilage fully retains this function and efficiently removes aggrecanases from the extracellular space, as demonstrated in Fig. 4. To test this possibility, we examined ADAMTS-5-induced aggrecan degradation in live and freeze-thawed (dead) OA cartilage. We found that the addition of ADAMTS-5 to OA cartilage resulted in a similar release of GAG from the cartilage regardless of whether the tissue was live or dead (Fig. 7C). Specific aggrecanase activity detected with these cartilage specimens and the similar susceptibility of live and dead cartilage to degradation was confirmed using an anti-ARGSV antibody (Fig. 7D).

Figure 7.

ADAMTS-5 endocytosis is impaired in OA cartilage due to reduced protein levels of LRP-1. A) Results of TaqMan real-time PCR showing relative levels of mRNA for LRP-1 in human OA and normal chondrocytes. B) Left panel: LRP-1 heavy chain was assessed by Western blotting using an anti-LRP-1 heavy-chain antibody. Right panel: densitometric analysis of immunoreactive LRP-1 bands. Each value was obtained from chondrocytes isolated from articular cartilage of patients with OA and normal patients (n=6 each). C) Time course of GAG released into the medium of live and freeze-thawed (dead) cartilage explants dissected from knee joints of patients with OA (n=5) incubated with 20 nM ADAMTS-5-4. D) Left panels: representative Western blot analysis of the media obtained from two patients after 24 h culture in C for aggrecan fragments using an anti-ARGSV neoepitope antibody. L, live cartilage; D, dead cartilage. Right panel: densitometric analysis of immunoreactive bands of aggrecan fragments (n=4). Amount of aggrecan fragment in the medium of live cartilage explants was taken as 1. Bars and points represent means ± sd. **P < 0.005; unpaired t test.

DISCUSSION

LRP-1 is a multifunctional endocytic type 1 transmembrane receptor consisting of an extracellular 515-kDa heavy chain and an associated 85-kDa light chain containing the transmembrane and cytoplasmic domains. It mediates internalization of >40 ligands, including lipoproteins, ECM proteins, cell surface receptors, proteinases, and proteinase-proteinase inhibitor complexes (26). The ablation of the LRP-1 gene in mice is embryonically lethal (27), but tissue specific deletion of the LRP-1 gene has indicated that it plays an important role in protecting the vasculature, β-amyloid precursor protein trafficking, and lipid metabolism in adipocytes and macrophage biology (see ref. 26 for review). The work presented here demonstrates that LRP-1 also plays a key role in regulating aggrecan turnover in cartilage by internalizing a major aggrecanase, ADAMTS-5.

Many proteinases are internalized by LRP-1 after forming a complex with specific inhibitors, as conformational changes of the inhibitor on complex formation give a signal for LRP-1 recognition (see ref. 26 for review). We considered the possibility that ADAMTS-5 might have been internalized after forming a complex with TIMP-3, as TIMP-3 is also endocytosed by an LRP-1-mediated pathway (ref. 20 and unpublished results). However, we concluded that TIMP-3 is not required for ADAMTS-5 internalization from the following observations: first, GM6001, a broad-spectrum hydroxamate metalloproteinase inhibitor that blocks metalloproteinase-TIMP complex formation, did not block ADAMTS-5 internalization; second, ADAMTS-5-5, consisting of the catalytic and the disintegrin domains, was not internalized, although it can form a complex with TIMP-3; and third, chondrocytes isolated from TIMP-3-null mice internalized ADAMTS-5 (data not shown). Thus, ADAMTS-5 and TIMP-3 can be endocytosed independently or as a complex. Domain deletion mutagenesis studies then indicated that both the TS1 domain and Sp domain are involved in effective binding of ADAMTS-5 to LRP-1, but the TS1 domain alone is sufficient. We postulate that basic residues in these domains are responsible for LRP-1 binding, as ADAMTS-5 internalization is inhibited by heparin. The importance of basic residues for interaction with LRP-1 has been shown for many protein ligands (26). When full-length ADAMTS-5 is secreted from the cell, however, it can bind to sulfated proteoglycans on the cell surface and in the ECM through basic residues in its CysR and Sp domains (21). The CysR and Sp domains are also necessary for the expression of full aggrecanase activity of ADAMTS-5. Therefore, the location of the enzyme and expression of aggrecanase activity is determined by competition between the ECM and LRP-1. However, ADAMTS-5 is internalized by cartilage explants and isolated chondrocytes with similar kinetics (Fig. 2). When live cartilage was incubated with ADAMTS-5, fluorescent ADAMTS-5 signal was observed within chondrocytes but not in the cartilage ECM (Fig. 3E). These results suggest that ADAMTS-5 binds loosely to the ECM and thus that the enzyme is readily removed from the tissue by LRP-1. When the spacer domain is removed from ADAMTS-5 (ADAMTS-5-3), its aggrecanase activity reduces only by 25%, but the extracellular half-life increases >3-fold. Deletion of both the CysR and Sp domains (ADAMTS-5-4) results in the loss of most of the aggrecanase activity (only 0.2-0.3% of the original activity remains; ref. 21). This form can diffuse through the interterritorial region of cartilage, as it no longer binds to the ECM. Interestingly, ADAMTS-5-4 retains ∼25% of full-length ADAMTS-5 proteolytic activity against biglycan, decorin, and fibromodulin (21). Such processing would have potential to dysregulate tissue matrix turnover and cellular environments. However, this form of the enzyme is still removed from the tissue by LRP-1, indicating that LRP-1 is therefore an important regulator of normal cartilage homeostasis.

A number of ADAMTS metalloproteinases have been reported to have an important role in development and organ morphogenesis. For example, ADAMTS-1 is essential for normal growth, structure, and function of the kidney, adrenal gland and female reproductive organ, myocardial trabeculation during heart development (see ref. 10 for review), and limb joint development (28). ADAMTS-5 plays a key role in cardiac valve development (29), regression of the interdigital web during mouse limb morphogenesis along with ADAMTS-9 and ADAMTS-20 (30), control of fibroblast-myofibroblast transition (31), and skin excision wound healing (32). In those processes, ADAMTSs function as versican- and aggrecan-cleaving enzymes. Although the role of LRP-1 in post-translational turnover of these ADAMTSs during organ morphogenesis or wound healing is yet to be investigated, our study implies that LRP-1 may regulate the activities of other members of the ADAMTS family, if we consider the importance of LRP-1 in development (27). During morphogenesis and tissue remodeling, timely ECM degradation is essential, and one possible mechanism is endocytosis of matrix-degrading enzymes. The recent study of Sorvillo et al. (33) showed that endocytosis of ultralarge von Willebrand factor-cleaving ADAMTS-13 by dendritic cells is mediated by a different scavenger receptor, the macrophage mannose receptor. This process is postulated to be involved in generation of the autoantibodies against the enzyme seen in patients with acquired thrombotic thrombocytopenic purpura (33).

ADAMTS-5 has been implicated in the development of OA, but this is controversial, because no correlation between ADAMTS-5 mRNA levels and OA progression have been reported (13, 17, 18). Our study has provided a new insight into this discrepancy by introducing a role of LRP-1 in the progression of OA. In normal healthy cartilage, the activity of ADAMTS-5 is regulated by endocytosis, as shown by an increased aggrecan turnover on addition of RAP. This pathway is, however, impaired in OA cartilage. This is evident from our observation that there were no significant differences in the ability of exogenously added ADAMTS-5 to digest aggrecan in live or freeze-thawed OA cartilage. A similar impairment of LRP-1-dependent endocytosis of MMP-13 has been reported in OA cartilage without alteration in the mRNA levels of LRP-1 (34), but its mechanism was not known. We have confirmed unaltered mRNA levels of LRP-1 in normal and OA human cartilage and found that levels of LRP-1 protein is substantially lower in OA cartilage. This leads to an increase in extracellular levels of ADAMTS-5, MMP-13, and possibly other metalloproteinase in OA cartilage. Although the mechanism by which OA cartilage loses LRP-1 from the cell surface has yet to be investigated, we postulate that it is due to increased shedding of LRP-1 from the cell surface as has been shown in malignant cells (35, 36) and in cycling human endometrium at menses (37). The shedding of LRP-1 increases under inflammatory conditions, and soluble LRP-1 is elevated in plasma of patients with rheumatoid arthritis and systemic lupus erythematosus (38). Proteinases that were reported to shed LRP-1 include β-amyloid precursor protein-cleaving enzyme (39), MMP-14 (35), ADAM-10 (40), ADAM-12 (36, 37), and ADAM-17 (38, 40). LRP-1 is reported to protect against the progression of atherosclerosis in mice by suppressing the expression of inflammatory mediators, such as monocyte chemoattractant protein 1, tumor necrosis factor α, and MMP-9, while deletion of macrophage LRP-1 increases MMP-9, which is associated with a high frequency to plaque rupture (41). Increased shedding of LRP-1 in chondrocytes will shift the homeostatic balance of ECM turnover to catabolic, resulting in a gradual loss of aggrecan and collagen fibrils. A slight shift in enzymatic balance due to loss of LRP-1 may be an important part of the dysregulation of aggrecanase activity. Identification of the mechanism by which LRP-1 is post-translationally processed will therefore provide further insights into our understanding of the pathogenesis of OA.