Annexin A6 Interacts With p65 and Stimulates NF-κB Activity and Catabolic Events in Articular Chondrocytes

Authors


Abstract

Objective

ANXA6, the gene for annexin A6, is highly expressed in osteoarthritic (OA) articular chondrocytes but not in healthy articular chondrocytes. This study was undertaken to determine whether annexin A6 affects catabolic events in these cells.

Methods

Articular chondrocytes were isolated from Anxa6-knockout mice, wild-type (WT) mice, and human articular cartilage in which ANXA6 was overexpressed. Cells were treated with interleukin-1β (IL-1β) or tumor necrosis factor α (TNFα), and expression of catabolic genes and activation of NF-κB were determined by real-time polymerase chain reaction and luciferase reporter assay. Anxa6−/− and WT mouse knee joints were injected with IL-1β or the medial collateral ligament was transected and partial resection of the medial meniscus was performed to determine the role of Anxa6 in IL-1β–mediated cartilage destruction and OA progression. The mechanism by which Anxa6 stimulates NF-κB activity was determined by coimmunoprecipitation and immunoblot analysis of nuclear and cytoplasmic fractions of IL-1β–treated Anxa6−/− and WT mouse chondrocytes for p65 and Anxa6.

Results

Loss of Anxa6 resulted in decreased NF-κB activation and catabolic marker messenger RNA (mRNA) levels in IL-1β– or TNFα-treated articular chondrocytes, whereas overexpression of ANXA6 resulted in increased NF-κB activity and catabolic marker mRNA levels. Annexin A6 interacted with p65, and loss of Anxa6 caused decreased nuclear translocation and retention of the active p50/p65 NF-κB complex. Cartilage destruction in Anxa6−/− mouse knee joints after IL-1β injection or partial medial meniscectomy was reduced as compared to that in WT mouse joints.

Conclusion

Our data define a role of annexin A6 in the modulation of NF-κB activity and in the stimulation of catabolic events in articular chondrocytes.

Annexins, cytoplasmic proteins that, in the presence of Ca2+, translocate and bind to membranes, are involved in a diverse range of cellular functions, including membrane-related events and membrane-trafficking events, ion channel activity, inflammation, and fibrinolysis ([1, 2]). The role of annexins in disease pathology is an emerging area of interest, and many studies have highlighted their active involvement in diseases, such as Alzheimer's disease, cancer, diabetes, and cardiovascular and autoimmune diseases, suggesting that annexins as disease modifiers could be novel therapeutic targets ([3-5]). Annexins, including annexin A6, are expressed by articular chondrocytes, and expression of some annexins is up-regulated during the progression of osteoarthritis (OA) ([6-10]). In a recent study, Anxa5 overexpression in articular chondrocytes resulted in increased crystal- or tumor necrosis factor α (TNFα)–induced apoptosis ([11]), suggesting that high Anxa5 expression in articular chondrocytes modulates the apoptotic response of these cells to these factors. Other studies have demonstrated that annexins modulate major signaling pathways involved in disease progression but do not modulate these same pathways during physiologic processes; these studies have provided novel insights into the potential mechanisms by which annexins may affect disease pathology ([12-14]). For example, ANXA1 and ANXA4 modulate NF-κB signaling activity in cancer cells ([13, 14]).

NF-κB signaling is one of the major catabolic signaling pathways involved in OA pathogenesis ([15]). The canonical NF-κB signaling pathway occurs through the activation of the IKK complex. Activation of this complex results in the phosphorylation and degradation of IκBα, and the release, phosphorylation, and nuclear translocation of the p50/p65 (RelA) heterodimeric complex, which then activates expression of various catabolic genes, including ADAMTS-5, interleukin-6 (IL-6), inducible nitric oxide synthase (iNOS), and matrix metalloproteinase 13 (MMP-13) in articular chondrocytes ([15, 16]). Since NF-κB signaling is highly activated in OA, and given the high expression of ANXA6 in human OA cartilage and the finding that annexins can act as modulators of NF-κB signaling, we questioned whether high ANXA6 expression in OA articular cartilage stimulates NF-κB signaling and ultimately accelerates catabolic events in articular chondrocytes. Using in vitro mouse and human chondrocyte cultures, a mouse model of IL-1β–mediated cartilage destruction, an in vivo mouse model of surgically induced OA, and Anxa6-knockout mice, we analyzed whether and how annexin A6 affects NF-κB signaling in human and mouse articular chondrocytes and how the modulatory role of annexin A6 affects cartilage degradation.

MATERIALS AND METHODS

Mice

Anxa6−/− mice with a C57BL/6 genetic background were provided by Dr. S. E. Moss (University College London, London, UK). Mice heterozygous for the mutation in Anxa6 were used for breeding ([17]). All protocols involving mice were approved by the New York University School of Medicine Institutional Animal Care and Use Committee.

Experimental OA

Experimental OA in 12-week-old male Anxa6−/− mice and wild-type (WT) littermates was induced using medial collateral ligament transection and partial medial meniscectomy as previously described ([18]). Briefly, the animals were placed under general anesthesia, the medial collateral ligament was transected, and the medial meniscus was partially removed using a surgical microscope. A sham operation was performed on the contralateral knee joint using the same approach, with no ligament transection or meniscectomy. The animals were then allowed unrestricted activity and food and water ad libitum.

IL-1β injections

Left knee joint cavities of 12-week-old male Anxa6−/− mice and WT littermates were injected with 6 μl of phosphate buffered saline (PBS) containing 25 ng/ml of recombinant mouse IL-1β (R&D Systems). This was performed on day 0, day 3, and day 6. The left knee joints of control mice were injected with PBS alone. The dose, frequency, and number of IL-1β injections were based on previous studies that have shown that this pattern of injection causes a mild arthritic insult accompanied by suppression of proteoglycan synthesis and a concomitant accelerated breakdown ([19-21]).

Histologic assessment and immunohistochemical analysis

Mice were killed 8 weeks after partial medial meniscectomy or on day 7 (1 day after the last IL-1β injection) for histologic analyses. The knee joints were harvested and fixed in 4% paraformaldehyde, decalcified in 0.2 moles/liter of EDTA, embedded in paraffin, and 4-μm–thick sections were cut perpendicular to the cartilage surface. Cartilage destruction in mice with partial medial meniscectomy was examined using Safranin O staining and scored based on the Osteoarthritis Research Society International (OARSI) histologic scoring system for murine OA ([22]).

Osteophyte formation was evaluated based on osteophyte formation scores, size, and maturity as previously described ([22-24]). For evaluation of osteophyte size, the thickness of the osteophyte was compared to that of the adjacent articular cartilage using a scoring system based on a scale of 0–3 (0 = none, 1 = small [the same thickness as the adjacent articular cartilage], 2 = medium [1–3 times the thickness of the adjacent articular cartilage], and 3 = large [>3 times the thickness of the adjacent articular cartilage]). Osteophyte maturity was scored using the following scale: 0 = none, 1 = precartilaginous lesion, 2 = predominantly cartilage, 3 = mixed cartilage and bone, and 4 = predominantly bone. Cartilage damage in IL-1β–injected knee joints was measured using the following arbitrary scale as previously described ([25]): 0 = no changes in cartilage, 1 = destaining of cartilage in either the deep or the superficial zone, 2 = destaining of cartilage in both the deep and the superficial zones, 3 = fibrillations of the cartilage surface, 4 = deep fissures in the cartilage, 5 = partial erosion of the cartilage up to the tidemark, and 6 = full-thickness erosion of the noncalcified and calcified layers. Five sections from each knee joint (spaced ∼70 μm apart) were scored; the score for an individual joint consisted of the summed values for these sections.

For immunohistochemical analysis, sections were incubated with sheep testicular hyaluronidase (2 mg/ml; Sigma) in PBS, pH 7.5, for 30 minutes at 37°C. Immunostaining was performed using a Histostain SP Kit according to the instructions of the manufacturer (Zymed Laboratories). Briefly, after incubation with a blocking solution for 10 minutes at room temperature, sections were incubated overnight at 4°C with primary antibodies specific for MMP-13 (Abcam), followed by incubation for 10 minutes at room temperature with biotinylated secondary antibodies. After washing, sections were incubated for 10 minutes at room temperature with a streptavidin–peroxidase conjugate, followed by a solution containing diaminobenzidine (chromogen) and 0.03% hydrogen peroxide for 5 minutes at room temperature. Control sections were incubated with nonimmune rabbit serum.

Cell cultures

Chondrocytes were isolated from the articular cartilage of 5-day-old Anxa6−/− mice and WT littermates as previously described ([26]). Cells were plated at a density of 1 × 105 cells/well into 6-well tissue culture plates and grown in monolayer cultures in Dulbecco's modified Eagle's medium (DMEM; Life Technologies) containing 10% fetal calf serum (HyClone), 2 mM of L-glutamine (Invitrogen), and 50 units/ml of penicillin and streptomycin (Invitrogen) (complete medium).

Human articular chondrocytes were isolated from articular cartilage samples obtained from patients (age range 48–67 years) undergoing total knee replacement surgery at New York University Hospital for Joint Diseases. Knee cartilage was harvested from regions with no macroscopically evident degeneration. The collection of tissue from patients undergoing knee replacement surgery was approved by the New York University School of Medicine Institutional Review Board. Human chondrocytes were isolated from these cartilage samples as previously described ([27]).

Cells were plated at a density of 5 × 105 cells/well into 12-well tissue culture plates and cultured in complete medium as described above. Semiconfluent human chondrocytes were transfected with empty pcDNA expression vector or pcDNA expression vector containing ANXA6 complementary DNA (cDNA) using FuGene 6 transfection reagent according to the manufacturer's protocol (Roche). We used a pcDNA vector that contained a c-Myc tag. We obtained a transfection rate of ∼40–50% as determined by immunostaining of transfected cells with fluorescein isothiocyanate–labeled antibodies specific for c-Myc (Abcam) and counterstaining of cell nuclei with DAPI (results not shown). Twenty-four hours after transfection, human chondrocyte cultures were switched to serum-free medium for 24 hours, followed by treatment with 10 ng/ml of recombinant human IL-1β or 10 ng/ml of recombinant human TNFα (both from R&D Systems) in PBS/0.1% bovine serum albumin (BSA) for up to 6 hours. Control cultures were treated with PBS/0.1% BSA (vehicle).

Anxa6−/− and WT mouse articular chondrocytes were serum-starved for 24 hours after cells reached confluence and then treated with 10 ng/ml of recombinant mouse IL-1β (R&D Systems), 10 ng/ml of recombinant mouse TNFα (aa 80–235; R&D Systems), or vehicle for up to 6 hours. To determine glycosaminoglycan (GAG) release into the culture medium, Anxa6−/− and WT femoral heads were cultured in DMEM in the absence or presence of 10 ng/ml of recombinant mouse IL-1β for 4 days.

Luciferase reporter assays

For luciferase assays, cells were cotransfected with a firefly NF-κB–specific luciferase reporter vector (pNFκB-Met-Luc2 reporter; Clontech). Luciferase activity in the medium from the secreted Metridia luciferase reporter gene was monitored using a Ready-To-Glow Secreted Luciferase Reporter Assay (Clontech) and a Tristar LB 941 luminometer (Berthold). Transfection efficiency was monitored by cotransfection with pSEAP vector (Clontech), which provides constitutive expression of the human placental form of secreted enhanced alkaline phosphatase (SEAP). Secreted SEAP activity was monitored with the chemiluminescent substrate CSPD (Clontech). All experiments were performed in triplicate and repeated 3–5 times.

Reverse transcription–polymerase chain reaction (PCR) and real-time PCR analysis

Total RNA was isolated from chondrocyte cultures using an RNeasy Mini kit (Qiagen). Levels of messenger RNA (mRNA) for ADAMTS-5, aggrecan, IL-6, iNOS, MMP-13, SOX9, and type II collagen were quantified by real-time PCR as previously described ([28]). Briefly, 1 μg of total RNA was reverse transcribed with an Omniscript RT kit (Qiagen). A 1:100 dilution of the resulting cDNA was used as the template to quantify the relative content of mRNA by real-time PCR (ABI Prism 7300 sequence detection system; Applied Biosystems), using the appropriate primers and SYBR Green. PCRs were performed with a SYBR Green PCR Master Mix kit (Applied Biosystems) at 95°C for 10 minutes, followed by 40 cycles of 95°C for 15 seconds and 60°C for 1 minute, and 1 cycle of 95°C for 15 seconds and 60°C for 1 minute. The 18S RNA was amplified at the same time and used as an internal control. The cycle threshold values for 18S RNA and the samples were measured and calculated. Transcript levels were calculated according to the equation x = 2–ΔCq, where ΔCq = Cqexp – Cq18S.

Coimmunoprecipitation

Mouse articular chondrocyte cultures were extracted with a buffer containing 50 mM of Tris HCl, pH 7.4, 1% Nonidet P40, 0.1% Triton X-100, 150 mM of NaCl, 5 mM of EDTA, and a proteinase inhibitor mixture. For the coimmunoprecipitation experiments, we used an immunoprecipitation kit, Dynabeads Protein G, according to the instructions of the manufacturer (Invitrogen). Briefly, the Dynabeads were incubated with 2 μg of a monoclonal mouse anti-p65 antibody or a mouse control IgG (Cell Signaling Technology) for 10 minutes with rotation, followed by incubation with the cell extracts (1 mg of total protein) for 30 minutes at room temperature with rotation. After washing, the antibody–antigen complex was eluted from the beads by incubation with 20 μl of elution buffer and 10 μl of 4× NuPAGE sodium dodecyl sulfate (SDS) sample buffer containing a reducing agent (Invitrogen) at 70°C for 10 minutes, and analyzed by electrophoresis in 10% Bis-Tris polyacrylamide gels. Samples were electroblotted onto nitrocellulose filters after electrophoresis. After blocking with a solution of low-fat milk protein, blotted proteins were immunostained with primary polyclonal rabbit antibodies specific for annexin A6 or monoclonal mouse anti-human p65 antibody (Cell Signaling Technology) and then peroxidase-conjugated secondary antibodies. The signal was detected by enhanced chemiluminescence (Pierce Protein Biology Products) as previously described ([28]). Production and specificity of the polyclonal rabbit anti-human Anxa6 antibody have also been previously described ([8]).

Subcellular fractionation

To extract the cytoplasmic, nuclear, and plasma membrane fractions from IL-1β–treated and untreated Anxa6−/− and WT mouse articular chondrocytes, we used NE-PER, according to the instructions of the manufacturer (Pierce Biology Products). Total cytoplasmic, nuclear, or plasma membrane protein fractions (30 μg) were analyzed by SDS–polyacrylamide gel electrophoresis and immunoblotting with antibodies specific for annexin A6 or NF-κB p65 as described above. For normalization of the protein expression levels, the membranes were immunostained with antibodies specific for β-actin (cytoplasmic fraction; Cell Signaling Technology), lamin B (nuclear fraction; Abcam), or α1 ATP1A1 (plasma membrane fraction; Abcam).

Quantification of GAG release

The amount of GAGs released into the culture media from IL-1β– and vehicle-treated Anxa6−/− and WT femoral head explants was measured with the metachromatic dye 1,9-dimethylmethylene blue (DMMB) using a proteoglycan detection kit (Astarte Biologics). Briefly, 100 μl of DMMB reagent was added to 100 μl of culture medium. The GAG–dye complex resulted in an absorption spectrum shift, which was measured at 525 nm. Values were derived from a standard curve using different concentrations of chondroitin sulfate, and the results are expressed as micrograms of GAG released per grams of cartilage.

Statistical analysis

Student's t-tests were performed to evaluate differences between 2 groups; analysis of variance was performed to evaluate differences among ≥3 groups. Tukey's multiple comparison test was applied as a post hoc test. P values less than 0.05 were considered significant.

RESULTS

Effect of loss of annexin A6 function on the expression of catabolic markers and proteoglycan loss after IL-1β treatment

Levels of mRNA for catabolic markers, including ADAMTS-5, IL-6, iNOS, and MMP-13, increased in articular chondrocytes isolated from 5-day-old WT and Anxa6−/− mice after IL-1β or TNFα treatment for 6 hours compared to the levels in vehicle-treated WT or Anxa6−/− mouse chondrocytes, while the levels of mRNA for cartilage markers, including aggrecan, type II collagen, and SOX9, decreased (Figures 1A and C). However, increases in levels of mRNA for these catabolic markers in Anxa6−/− mouse chondrocytes treated with IL-1β or TNFα were reduced by 40–60% compared to WT mouse cells treated with IL-1β or TNFα (Figure 1A), while cartilage marker mRNA levels were increased by 30–50% (Figure 1C). IL-1 receptor mRNA levels were similarly increased in IL-1β–treated WT and Anxa6−/− mouse chondrocytes (Figure 1B). Furthermore, the amount of GAGs released into the medium was decreased in femoral head explants of IL-1β–treated Anxa6−/− mice compared to IL-1β–treated WT mice (Figure 1D).

Figure 1.

A–C, Levels of mRNA for catabolic markers (ADAMTS-5, interleukin-6 [IL-6], inducible nitric oxide synthase [iNOS], and matrix metalloproteinase 13 [MMP-13]) (A), IL-1 receptor (B), and articular cartilage markers (aggrecan, SOX9, and type II collagen) (C) in vehicle-treated and IL-1β–treated or tumor necrosis factor α (TNFα)–treated wild-type (WT) and annexin A6–knockout mouse chondrocytes. Levels of mRNA were determined by real-time polymerase chain reaction (PCR) using SYBR Green and normalized to the level of 18S RNA. Data were obtained from triplicate PCRs using RNA from 3 different cultures. Values are the mean ± SEM. ∗ = P < 0.01 versus IL-1β– or TNFα-treated WT mouse cells. D, Glycosaminoglycan (GAG) release into culture medium of vehicle-treated and IL-1β–treated Anxa6−/− and WT femoral head explants. Values are the mean ± SEM of 4 different explant cultures. ∗ = P < 0.01 versus IL-1β–treated WT femoral head explants.

To determine the role of annexin A6 in IL-1β–mediated articular cartilage degradation in vivo, recombinant mouse IL-1β was injected intraarticularly into the left hind limb knee joints of WT and Anxa6−/− mice. IL-1β–injected WT mouse knee joints showed markedly increased loss of proteoglycans as indicated by reduced Safranin O staining, increased cartilage degradation, and increased MMP-13 immunostaining compared to WT mouse knee joints injected with PBS (Figure 2A). Proteoglycan loss, cartilage degradation, and MMP-13 immunostaining were markedly reduced in IL-1β–injected Anxa6−/− mouse knee joints compared to IL-1β–injected WT mouse knee joints (Figure 2A). Quantification of cartilage degradation after IL-1β injection revealed a reduction of cartilage degradation with the loss of Anxa6 (Figure 2B). PBS-injected WT and Anxa6−/− mouse knee joints showed no loss of Safranin O staining and no MMP-13 immunopositive cells (Figure 2A). These findings demonstrate that loss of Anxa6 results in reduced catabolic marker expression and loss of cartilage degradation after IL-1β treatment.

Figure 2.

Cartilage destruction in the medial portion of tibial cartilage of wild-type (WT) and annexin A6–knockout mice after interleukin-1β (IL-1β) injection. Twelve-week-old male mice received IL-1β injections in the knee joints on days 0, 3, and 6. Control mice were injected with phosphate buffered saline (PBS). Tissue samples were harvested on day 7. A, Representative histologic features of sections in the medial portion of tibial cartilage from WT and Anxa6−/− mice stained with Safranin O and immunostained with antibodies specific for matrix metalloproteinase 13 (MMP-13). Bar = 200 μm. B, Histologic scoring of cartilage degradation according to the grading system described in Materials and Methods. Values are the mean ± SEM of 10 samples per genotype for WT and Anxa6−/− mice. ∗ = P < 0.01 versus IL-1β–injected WT mice.

Loss of annexin A6 function results in reduced NF-κB signaling

Since NF-κB signaling plays a major role in the IL-1β– and TNFα-mediated stimulation of the expression of catabolic markers in articular chondrocytes, we determined whether and how annexin A6 affects NF-κB activities in chondrocytes. Luciferase activity from the NF-κB luciferase reporter was increased in WT and Anxa6−/− mouse chondrocytes after IL-1β or TNFα treatment compared to vehicle-treated cells (Figure 3). The increase, however, was markedly less in IL-1β– or TNFα-treated Anxa6−/− mouse chondrocytes compared to IL-1β– or TNFα-treated WT cells (Figure 3).

Figure 3.

NF-κB activity in vehicle-treated, interleukin-1β (IL-1β)–treated, and tumor necrosis factor α (TNFα)–treated wild-type (WT) and annexin A6–knockout mouse articular chondrocytes. WT and Anxa6−/− mouse articular chondrocytes were transfected with pNFκB-Met-Luc2 luciferase reporter. After serum starvation, the cells were treated for 6 hours with IL-1β, TNFα, or vehicle. The samples were then analyzed for luciferase activity. Transfection efficiency was monitored by cotransfection with pSEAP vector, which provides constitutive expression of a secreted form of human placental alkaline phosphatase (SEAP). Secreted SEAP activity was monitored with the chemiluminescent substrate CSPD. Values are the mean ± SEM of 4 different experiments. ∗ = P < 0.01 versus IL-1β– or TNFα-treated WT mouse cells.

The release of the p50/p65 (RelA) heterodimeric complex from IκBα allows this heterodimeric complex to translocate to the nucleus, where it binds to specific DNA promoter regions and activates the expression of catabolic genes ([15, 16]). Coimmunoprecipitation experiments with antibodies specific for p65 and annexin A6 revealed that annexin A6 coimmunoprecipitated with p65, indicating an interaction between these 2 proteins (Figure 4A). To obtain insights into how the annexin A6–p65 interaction affects NF-κB activity, we determined the amount of p65 in the nuclear and cytoplasmic fractions isolated from Anxa6−/− and WT mouse chondrocytes at various time points after IL-1β treatment. The amount of nuclear p65 was markedly decreased in Anxa6−/− mouse chondrocytes at 15, 30, and 60 minutes posttreatment with IL-1β compared to WT mouse chondrocytes, whereas cytoplasmic p65 amounts were increased (Figure 4B). Immunoblot analysis of the nuclear fraction isolated from WT mouse chondrocytes with antibodies specific for annexin A6 revealed the presence of annexin A6. The amount of nuclear annexin A6 increased with IL-1β treatment, while the amount of cytoplasmic annexin A6 decreased (Figure 4C). The amounts of annexin A6 in the total cell extract and plasma membrane fraction were not affected by IL-1β (Figure 4C). These findings suggest that annexin A6 stimulates the nuclear translocation and prolonged nuclear localization of the active p50/p65 heterodimeric complex, leading to decreased NF-κB activity in Anxa6−/− mouse chondrocytes.

Figure 4.

A, Interaction between annexin A6 and p65 as revealed by coimmunoprecipitation of both proteins from cell lysates of mouse chondrocytes. Immunoprecipitation (IP) was performed with p65-specific antibodies or control IgG; immunoblot (IB) analysis of the immunoprecipitates was performed with annexin A6–specific antibodies or p65-specific antibodies. L = lysate; B = beads; HC = heavy chain of IgG; LC = light chain of IgG. B, Immunoblot analysis of the cytoplasmic (C) and nuclear (N) fractions isolated from wild-type (WT) mouse chondrocytes (left) and Anxa6-knockout mouse chondrocytes (right) treated with IL-1β for the time periods indicated and analyzed with antibodies specific for p65. Blots were also analyzed with antibodies specific for β-actin (a cytoplasmic protein) and lamin B (a nuclear protein) to control for equal loading in each fraction. C, Immunoblot analysis of the cytoplasmic, nuclear, and plasma membrane (PM) fractions and total cell extract isolated from WT mouse articular chondrocytes treated with IL-1β for the time periods indicated and analyzed with antibodies specific for annexin A6. Blots were also analyzed with antibodies specific for ATP1A1 (a plasma membrane protein), β-actin, and lamin B to control for equal loading. Blots in A–C are representative of 3 separate experiments with similar results.

Altered ANXA6 expression affects NF-κB activity and MMP-13 expression in human articular chondrocytes

To extend our findings to human cells, we overexpressed ANXA6 in human articular chondrocytes to mimic the high expression of ANXA6 in severe OA cartilage ([8]). Transfection of human articular chondrocytes with pcDNA expression vector containing ANXA6 cDNA resulted in a 2–3-fold increase in annexin A6 protein levels compared with empty vector–transfected cells (data not shown). In empty vector–transfected human chondrocytes treated with IL-1β or TNFα, an ∼8-fold increase in the luciferase activity from the NF-κB luciferase reporter was noted as compared to vehicle-treated cells (Figure 5A). ANXA6 overexpression in IL-1β– or TNFα-treated human chondrocytes resulted in a further increase of NF-κB activity as indicated by the increased luciferase activity from the NF-κB luciferase reporter compared to IL-1β– or TNFα-treated, empty vector–transfected cells (Figure 5A). Overexpression of ANXA6 alone in human articular chondrocytes was sufficient to increase luciferase activity from the NF-κB luciferase reporter by ∼1.5-fold compared to empty vector–transfected cells (Figure 5A). Concurrently, MMP-13 mRNA levels were increased in human chondrocytes overexpressing ANXA6 compared to empty vector–transfected chondrocytes, and MMP-13 mRNA levels were further increased in ANXA6-overexpressing cells treated with IL-1β or TNFα compared to IL-1β–treated, empty vector–transfected cells (Figure 5B). These findings demonstrate that annexin A6 stimulates NF-κB activity and MMP-13 expression in human articular chondrocytes in the absence or presence of IL-1β or TNFα.

Figure 5.

NF-κB activity as determined by luciferase activity from an NF-κB luciferase reporter (A), and matrix metalloproteinase 13 (MMP-13) mRNA levels (B), in empty pcDNA expression vector or pcDNA expression vector containing annexin A6–transfected human articular chondrocytes in the absence or presence of interleukin-1β (IL-1β) or tumor necrosis factor α (TNFα). For determining NF-κB activity, human articular chondrocytes were cotransfected with pNFκB-Met-Luc2 luciferase reporter and an ANXA6 overexpression vector or an empty control vector. Transfection cells were serum-starved for 24 hours and then cultured in the presence of vehicle, IL-1β, or TNFα for 6 hours. Levels of mRNA were determined by real-time polymerase chain reaction (PCR) using SYBR Green and normalized to 18S RNA levels. Each PCR was run in triplicate. Transfection efficiency was monitored by cotransfection with pSEAP vector as described in Figure 3. Values are the mean ± SEM of 3 different cultures. ∗ = P < 0.01 versus vehicle-treated, empty vector–transfected cells; ∗∗ = P < 0.01 versus IL-1β– or TNFα-treated, empty vector–transfected cells.

Loss of annexin A6 protects against cartilage degradation after partial medial meniscectomy

In the final set of experiments, we determined whether loss of annexin A6 protects against cartilage destruction in a mouse model of surgically induced OA. OA was induced in the left knee joints of 12-week-old male Anxa6−/− mice and WT littermates using partial medial meniscectomy as previously described ([18]). The right knee joints of these mice underwent sham surgery as controls (Figures 6A and B). Eight weeks after partial medial meniscectomy, the medial tibial plateau and femoral condyles of WT mouse knee joints showed major losses of articular cartilage and loss of Safranin O staining in the remaining cartilage (Figure 6C). In addition, osteophyte formation, a characteristic feature of OA, was detected at the medial edge of WT mouse joints (Figure 6E). Conversely, cartilage destruction was much milder in the knee joints of Anxa6−/− mice (Figure 6D). These observations were corroborated by the OARSI scores, which were significantly lower (P < 0.0001) in the partial medial meniscectomy–operated Anxa6−/− mouse knee joints compared to the partial medial meniscectomy–operated WT mouse knee joints at 8 weeks after surgery (Figure 6G). In addition, osteophyte formation was reduced in Anxa6−/− mouse knee joints 8 weeks after partial medial meniscectomy compared to osteophyte formation in WT mouse knee joints as indicated by reduced osteophyte size and maturity (Figures 6F and H).

Figure 6.

A–D, Safranin O staining of the medial portion of femoral and tibial cartilage in representative samples of knee joints of sham-operated wild-type (WT) mice (A) and sham-operated annexin A6–knockout mice (B), as well as knee joints 8 weeks after partial medial meniscectomy (PMX) in WT mice (C) and Anxa6−/− mice (D). E and F, Representative Safranin O–stained sections of knee joints from WT (E) and Anxa6−/− (F) mice 8 weeks after partial medial meniscectomy, demonstrating osteophyte development on the anterior tibial plateau. Osteophyte size is greatly reduced in the joints of Anxa6−/− mice compared with WT mice (arrows). Bars = 200 μm. G, Quantification of osteoarthritis development using the Osteoarthritis Research Society International (OARSI) scoring system. Ti = tibia; Fe = femur. H, Osteophyte size and maturity in knee joints from WT and Anxa6−/− mice 8 weeks after partial medial meniscectomy, using the scoring system described in Materials and Methods. Values in G and H are the mean ± SEM of 8 mice per group.

DISCUSSION

Intensified and sustained NF-κB signaling activity has been shown in OA cartilage, and it is one of the key signaling pathways to stimulate catabolic events in articular chondrocytes during OA progression ([15]). We found that upon stimulation of articular chondrocytes with NF-κB activation signals, such as IL-1β or TNFα, annexin A6 increased NF-κB activity. Consequently, articular chondrocytes lacking Anxa6 exhibited decreased NF-κB activity after IL-1β or TNFα treatment and reduced catabolic events. Conversely, NF-κB activity and MMP-13 expression were increased in human articular chondrocytes with overexpressed ANXA6 after treatment with IL-1β or TNFα. We also demonstrated that lack of Anxa6 resulted in decreased cartilage degradation in a mouse model of surgically induced OA. This suggests that increased annexin A6 expression in OA cartilage may lead to exacerbated and sustained NF-κB activation in OA cartilage, causing increased catabolic events in articular chondrocytes and accelerated cartilage destruction.

How does annexin A6 stimulate NF-κB activity? We have demonstrated that annexin A6 interacts with the p65 subunit of the NF-κB complex and that the lack of Anxa6 results in decreased amounts of nuclear p65 over a prolonged period of time in chondrocytes treated with IL-1β. We can only speculate about how annexin A6 enhances the nuclear translocation and retention of the p50/p65 NF-κB complex. Previous studies have shown that the translocation of the p50/p65 complex is mediated by importins and the recognition of nuclear localization sequence in p65 by these importins ([29]). Since we also detected increasing amounts of annexin A6 in the nucleus of articular chondrocytes after IL-1β treatment, it is possible that a complex containing annexin A6, p50, and p65 translocates to the nucleus and that binding of annexin A6 to p65 leads to enhanced exposure of the nuclear localization sequence for recognition by importins. Accelerated nuclear transport of the complexed annexin A6 and p50/p65 into the nucleus could then result.

In addition, our finding that nuclear annexin A6 increased the amount of nuclear p65 after 60 minutes of IL-1β treatment suggests that nuclear annexin A6 may inhibit the nuclear export of NF-κB. Newly resynthesized nuclear IκBα has been shown to bind to the p50/p65 heterodimeric NF-κB complex; it dissociates the NF-κB complex from DNA and promotes its nuclear export, thereby providing a negative regulatory feedback mechanism that critically influences the duration of the NF-κB response ([30]). It is possible that nuclear annexin A6 binding to p65 prevents binding of nuclear IκBα to the active NF-κB complex, and as a consequence the nuclear export of NF-κB may be inhibited. Future experiments are needed to establish the exact mechanism by which annexin A6 interacts with p65 and mediates the stimulation of the activity of the p50/p65 NF-κB complex in articular chondrocytes.

ANXA6 is expressed at low levels in healthy articular cartilage, and its expression markedly increases with OA progression ([8]). In this study, we demonstrated that overexpression of ANXA6 in human articular chondrocytes, which mimics high ANXA6 expression in OA cartilage, resulted in increased NF-κB activity and increased MMP-13 expression after IL-1β or TNFα treatment. These findings suggest that the stimulation of NF-κB activity, and ultimately the expression of catabolic markers, depends on ANXA6 expression. Recently studies have shown that expression of several annexins, including ANXA6, changes during various diseases, including cancer, cardiovascular diseases, brain ischemia, and Alzheimer's disease, and that the interaction of several annexins with components of signaling pathways and the modulation of these signaling pathway activities depends on annexin expression ([3, 13, 14, 31, 32]). For example, constitutive activation of NF-κB is seen in breast cancer cells with highly expressed ANXA1, and as a consequence, high metastasis occurs. Low expression of ANXA1 in breast cancer cells, however, does not lead to the constitutive activation of NF-κB, and therefore low metastasis is seen ([13]). In addition, it is likely that factors other than annexin expression, including Ca2+, regulate interactions between annexins and components of the signaling pathways. Further studies are necessary to identify these factors and elucidate how ANXA6 expression, together with these factors, modulates NF-κB activity and catabolic events in articular chondrocytes.

In summary, we have identified annexin A6 as a stimulator of catabolic events in articular chondrocytes, and ultimately, of cartilage degradation, in a model of IL-1β–mediated cartilage destruction and in a model of surgically induced OA. Specifically, our evidence shows that high expression of ANXA6 in articular chondrocytes from OA cartilage resulted in increased NF-κB activity and stimulation of the expression of catabolic genes. However, lack of ANXA6 expression in articular chondrocytes led to reduced NF-κB activity, reduced expression of catabolic genes, and eventually, reduced cartilage degradation. Our findings suggest that targeting annexin A6 functions or its high expression in OA cartilage may be a more effective and specific way to control NF-κB signaling in OA than targeting this pathway directly. Therapeutic targeting of the NF-κB pathway in OA carries the high risk of affecting physiologic NF-κB signaling. Our results have provided a novel and solid foundation for future studies of the molecular mechanisms governing the role of annexin A6 and other annexins in OA, as well as the interplay of these annexins with other proteins involved in OA initiation and progression. Annexin A6 appears to have an important role in OA pathogenesis, and it may prove to be an effective therapeutic target for this disease.

AUTHOR CONTRIBUTIONS

All authors were involved in drafting the article or revising it critically for important intellectual content, and all authors approved the final version to be published. Dr. Kirsch had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.

Study conception and design. Campbell, Minashima, Hadley, Lee, Giovinazzo, Kirsch.

Acquisition of data. Campbell, Minashima, Zhang, Hadley, Giovinazzo, Lee, Quirno, Kirsch.

Analysis and interpretation of data. Campbell, Minashima, Hadley, Lee, Quirno, Kirsch.

Ancillary