SEARCH

SEARCH BY CITATION

Abstract

  1. Top of page
  2. Abstract
  3. MATERIALS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. AUTHOR CONTRIBUTIONS
  7. REFERENCES

Objective

The incidence of low back pain is extremely high and is often linked to intervertebral disc (IVD) degeneration. The mechanism of this disease is currently unknown. This study was undertaken to investigate the role of β-catenin signaling in IVD tissue function.

Methods

β-catenin protein levels were measured by immunohistochemical analysis of disc samples obtained from patients with disc degeneration and from normal subjects. To generate β-catenin conditional activation (cAct) mice, Col2a1-CreERT2–transgenic mice were bred with β-cateninfx(Ex3)/fx(Ex3) mice. Changes in disc tissue morphology and function were examined by micro–computed tomography, histologic analysis, and real-time polymerase chain reaction assays.

Results

β-catenin protein was up-regulated in disc tissue samples from patients with disc degeneration. To assess the effects of increased β-catenin levels on disc tissue, we generated β-catenin cAct mice. Overexpression of β-catenin in disc cells led to extensive osteophyte formation in 3- and 6-month-old β-catenin cAct mice, which were associated with significant changes in the cells and extracellular matrix of disc tissue and growth plate. Gene expression analysis demonstrated that activation of β-catenin enhanced runt-related transcription factor 2–dependent Mmp13 and Adamts5 expression. Moreover, genetic ablation of Mmp13 or Adamts5 on the β-catenin cAct background, or treatment of β-catenin cAct mice with a specific matrix metalloproteinase 13 inhibitor, ameliorated the mutant phenotype.

Conclusion

Our findings indicate that the β-catenin signaling pathway plays a critical role in disc tissue function.

Low back pain, which has an extremely high incidence, is believed to be linked to degenerative changes in the intervertebral disc (IVD) (1). This degenerative disc disease can occur in any of the 23 IVDs that span the cervical, thoracic, and lumbar regions of the spine. Spine diseases and low back pain are the leading cause of disability in people younger than 45 years of age, and result in national economic losses of >$90 billion per year in the US (2). Approximately 1% of the US population is chronically disabled because of back pain, and an additional 1% experiences temporary disability. Numerous surgical approaches have been developed in recent years to deal with damaged or traumatized discs. These procedures are aimed at symptom relief, and none of the current therapies can restore the normal function of a degenerated IVD.

The major cartilaginous joint of the vertebral column is the IVD. Each disc consists of an inner zone, the nucleus pulposus, surrounded by a peripheral outer region, the anulus fibrosus. On the inner border of the anulus, the fibers form a tissue composed predominantly of type II collagen. The inner anulus fibrosus encloses the nucleus pulposus, a highly viscous gel-like tissue that generates aggrecan and other proteoglycans, which provide the disc with its critical water-retaining characteristics. The superior and inferior boundaries of the IVD are formed by the growth plate and cartilage end plate. It has been suggested that mechanical factors produce growth plate and end plate damage, the antecedent to disc degeneration (3). The metabolism of the nucleus pulposus is dependent on diffusion of fluid either from the marrow of the vertebral bodies across the subchondral bone, growth plate, and cartilage end plate or through the anulus fibrosus from the surrounding blood vessels. Morphologic changes in the vertebral bone, growth plate, and cartilage end plate, which occur with advancing age or trauma-related degeneration, can interfere with normal disc nutrition and further the degenerative process. These changes could eventually lead to abnormal disc cell metabolism and alter the integrity of the proteoglycans and water concentration, reducing the number of viable cells, with subsequent alteration in the movement of solutes into and out of the disc (4).

Wnt proteins play critical roles in bone and cartilage development (5). Wnt proteins form a dual-receptor complex with Frizzled and low-density lipoprotein receptor–related protein 5 (LRP-5) or LRP-6 on cell surfaces. This triggers signaling through a large protein complex in the Wnt canonical pathway, including glycogen synthase kinase 3β (GSK-3β), casein kinase 1, and the scaffolding proteins, adenomatous polyposis coli (APC), disheveled, and Axins. This complex has multiple effects on β-catenin: it promotes β-catenin phosphorylation by GSK-3β at specific amino terminal residues and creates docking sites for F-box protein/E3 ligase complexes (5–7) and enables β-catenin to be detected and destroyed by the 26S proteasome in the absence of Wnt signaling (8). The activation of Wnt signaling inhibits the stimulatory effect of Axins on β-catenin phosphorylation and allows β-catenin to move to the nucleus (9). Nuclear β-catenin combines with the transcription factors T cell factors and lymphoid enhancer factor 1 to activate expression of target genes. β-catenin is a key molecule in the canonical Wnt signaling pathway and plays a critical role in multiple steps of osteoblast and chondrocyte differentiation. However, the role of β-catenin signaling in IVD tissue has not been fully investigated.

In the present study, we performed immunohistochemical analysis to analyze β-catenin expression in disc tissue obtained from patients with disc degeneration. We found that β-catenin protein was significantly up-regulated in the disc tissue of these samples. To create a mouse model that mimics the condition of β-catenin up-regulation observed in patients with disc degeneration, we generated β-catenin conditional activation (cAct) mice. Severe defects in disc tissue were found in these mice, including up-regulation of the expression of Mmp13, Adamts4, and Adamts5 genes, significant loss of growth plate cartilage, and severe osteophyte formation. Deletion of the Mmp13 or Adamts5 gene on the β-catenin cAct background or treatment of β-catenin cAct mice with a matrix metalloproteinase 13 (MMP-13) inhibitor significantly reversed the defective phenotype observed in the disc tissue of β-catenin cAct mice. These findings demonstrate that β-catenin signaling plays a critical role in disc tissue function and may be involved in the development of disc degeneration.

MATERIALS AND METHODS

  1. Top of page
  2. Abstract
  3. MATERIALS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. AUTHOR CONTRIBUTIONS
  7. REFERENCES

Immunohistochemical analysis.

Paraffin sections (3 μm) were heated at 95°C in citrate buffer (pH 6.0) for 2 hours, and then treated with a dual endogenous enzyme-blocking reagent for 10 minutes (S2003; Dako). After blocking with normal goat serum (S-1000; Vector) (1:20 dilution) for 30 minutes, sections were treated with rabbit anti–β-catenin antibody (9562; Cell Signaling Technology) (1:30 dilution) overnight at 4°C and incubated with secondary biotinylated goat anti-rabbit antibody (BA-1000; Vector) (1:200 dilution) for 30 minutes, followed by streptavidin–peroxidase (21130; Pierce) (1:250 dilution) for 30 minutes at room temperature. Peroxidase activity was revealed by staining with Romulin AEC Chromogen (RAEC810L; BioCare Medical).

Animals.

We obtained Rosa26 reporter mice and Mmp13fx/fx mice from The Jackson Laboratory (10, 11). We used Col2a1-CreERT2–transgenic mice as previously described (12, 13). The β-cateninfx(Ex3)/fx(Ex3) mice were originally described by Harada et al (14) and were used in our previous studies (13). Tamoxifen (Sigma) was administered to 2-week-old mice by intraperitoneal (IP) injection (1 mg/10 gm body weight once a day for 5 days). All protocols were approved by the University Committee on Animal Resources of the University of Rochester.

Micro–computed tomography (micro-CT).

Prior to histologic processing, we evaluated formalin-fixed mouse spines by micro-CT using a Scanco VivaCT40 cone-beam scanner (Scanco Medical) with a 55 kVp source and a 142 μAmp current. We scanned the mouse spines at a resolution of 10.5 μm. The scanned images from each group were evaluated at the same thresholds to allow 3-dimensional structural rendering of each sample.

Histologic and histomorphometric analysis.

We dissected lumbar spines from Col2a1-CreERT2;R26R mice, Col2a1-CreERT2;β-cateninfx(Ex3)/wt mice, Col2a1-CreERT2;Mmp13fx/fx mice, Col2a1-CreERT2;β-cateninfx(Ex3)/wt;Mmp13fx/fx mice, and their corresponding Cre-negative control mice. Samples were fixed in 10% formalin, decalcified, and embedded in paraffin. Serial sections (3 μm thick) were taken from 3 levels spaced 15 μm apart within the midsagittal region of the intervertebral bodies. The sections were stained with Alcian blue/hematoxylin and eosin (H&E) and Safranin O–fast green. We quantified growth plate cartilage area using ImagePro 4.5 (Leeds Precision Instruments) by tracing the Alcian blue–positive area. To quantify lumbar spine length, we measured the distance from the C1 vertebra to the S4 vertebra on micro-CT images using ImagePro 4.5. To quantify disc space, we measured the distance between L4 and L5 on micro-CT images using ImagePro 4.5. Statistical analysis was conducted using one-way analysis of variance followed by Dunnett's test or Student's unpaired t-test.

Treatment with MMP-13 inhibitor.

An MMP-13 inhibitor, CL82198 (Tocris), was used for the in vivo experiment. The inhibitory effect of CL82198 on MMP-13 enzymatic activity was confirmed with an MMP-13 fluorometric drug discovery kit (Enzo Life Sciences). Both β-catenin cAct mice and Cre-negative control mice were injected IP with tamoxifen (1 mg/10 gm body weight once a day for 5 days) at 2 weeks of age and then injected IP with MMP-13 inhibitor (5 mg/kg body weight per day) or vehicle (phosphate buffered saline) every other day. At 3 months of age, mice were killed, and the whole lumbar vertebrae were harvested for micro-CT and histologic analyses.

RESULTS

  1. Top of page
  2. Abstract
  3. MATERIALS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. AUTHOR CONTRIBUTIONS
  7. REFERENCES

Our recent studies demonstrated that β-catenin expression is highly up-regulated in knee joint samples from patients with osteoarthritis (OA) (13). Since disc degeneration is often correlated with OA development in patients (15, 16), in the present study we examined β-catenin expression in patients with disc degeneration by immunohistochemical analysis. Expression of β-catenin protein in human disc samples was analyzed semiquantitatively. Each sample was scored twice, once for the percentage of labeled disc cells (0 = absence of labeling of disc cells, 1 = <30% of disc cells labeled, 2 = 30–60% of disc cells labeled, and 3 = >60% of disc cells labeled) and once for the intensity of the immunohistochemical staining (0 = no staining, 1 = weak staining, 2 = mild staining, and 3 = strong staining). The disc labeling score and the staining score were multiplied to determine the final score (range 0–9). A double-blind analysis was performed by 3 independent graders (MW, BS, and JS). Final scores of 1–4 were considered to indicate weakly positive β-catenin protein expression, and final scores of 6–9 were considered to indicate strongly positive β-catenin protein expression.

We found that β-catenin protein was up-regulated in most samples obtained from patients with disc degeneration. Of 30 patient samples analyzed, 6 were found to be strongly positive and 22 were found to be weakly positive for β-catenin protein expression. In contrast, β-catenin was detected in only 1 sample, with weak expression, in disc tissue obtained from normal subjects (n = 10) (Figure 1). In most cases, β-catenin expression was detected in the anulus fibrosus and cartilage end plate cells. Formation of chondrocyte clusters was seen in degenerative disc samples with highly up-regulated β-catenin expression (Figures 1A and F). These results suggest that β-catenin expression may be activated in different types of IVD cells through different mechanisms during the development of disc degeneration.

thumbnail image

Figure 1. Up-regulation of β-catenin protein expression in disc tissue from patients with disc degeneration. A–F, Immunohistochemical (IHC) analysis of β-catenin protein expression (arrows) in samples obtained from patients with disc degeneration, as determined by magnetic resonance imaging and histologic examination (n = 30). The β-catenin protein was highly expressed in 6 of the samples from patients with disc degeneration, weakly expressed in 22 of the samples from patients with disc degeneration, but weakly expressed in only 1 of the samples from normal subjects. In some samples, disc cells formed clusters where β-catenin was highly up-regulated (A and F). In most samples, β-catenin expression was found in localized areas of disc tissue. Lower panels show higher-magnification views (original magnification × 20) of the boxed areas in the upper panels. Bars = 120 μm. G, Semiquantitative analysis of β-catenin protein expression in normal subjects and in patients with degenerative disc disease (DDD). Each sample was scored twice, once for the percentage of labeled disc cells (0 = absence of labeling of disc cells, 1 = <30% of disc cells labeled, 2 = 30–60% labeled, and 3 = >60% labeled) and once for the intensity of the immunohistochemical staining (0 = no staining, 1 = weak staining, 2 = mild staining, and 3 = strong staining). The 2 scores were multiplied to obtain the final score (range 0–9). Final scores of 1–4 were considered weakly positive, and final scores of 6–9 were considered strongly positive. The percentages of weakly positive and strongly positive β-catenin staining were much higher in samples from patients with disc degeneration than in those from normal subjects. Bars show the mean ± SEM. ∗∗ = P < 0.01 versus normal subjects, by Student's unpaired t-test.

Download figure to PowerPoint

To develop a mouse model to mimic human disc degeneration, we generated inducible β-catenin cAct mice by breeding β-cateninfx(Ex3)/fx(Ex3) mice with Col2a1-CreERT2–transgenic mice. The advantage of this approach is that it overcomes the embryonic death caused by chondrocyte-specific β-catenin activation (17, 18) and it permits normal spine development up to the time of Cre recombination. Tamoxifen (1 mg/10 gm body weight) was administered by IP injection to 2-week-old Col2a1-CreERT2;β-cateninfx(Ex3)/wt (β-catenin cAct) and Cre-negative control mice once a day for 5 consecutive days. These mice were killed at 3 and 6 months of age, and changes in the structure and morphology of IVD tissue were analyzed.

In order to determine Cre-mediated recombination efficiency, we bred Col2a1-CreERT2–transgenic mice with Rosa26 reporter mice. Tamoxifen was administered to 2-week-old mice, and X-Gal staining was performed when mice were 2 months of age. High Cre recombination efficiency in growth plate chondrocytes (75%) and inner anulus fibrosus cells (81%) was observed in Col2a1-CreERT2;R26R mice (data not shown) (19). This is consistent with the expression pattern of endogenous type II collagen in the disc tissue. We performed immunohistochemical analysis to examine β-catenin expression in 4-week-old β-catenin cAct mice that had received tamoxifen at 2 weeks of age. β-catenin protein was overexpressed in the disc cells of β-catenin cAct mice (Figure 2E). Due to the loss of growth plate chondrocytes in 4-week-old β-catenin cAct mice, β-catenin expression was mainly detected in anulus fibrosus cells in β-catenin cAct mice (Figure 2E).

thumbnail image

Figure 2. Characterization of β-catenin conditional activation (cAct) mice. The β-catenin cAct mice were generated by breeding β-cateninfx(Ex3)/fx(Ex3) mice with Col2a1-CreERT2 mice and treating them with tamoxifen. A, Growth plate cartilage (demarcated by the dotted line) in an Alcian blue/hematoxylin and eosin (H&E)–stained section from a Cre-negative control mouse. Loss of growth plate cartilage was seen in 1-month-old β-catenin cAct mice. B–D, Formation of chondrocyte clusters (black arrows) (B), osteophytes (C), and new blood vessels (black arrows) and new woven bone (blue arrows) (D) in Alcian blue/H&E–stained sections from 1-month-old β-catenin cAct mice. E, Overexpression of β-catenin protein, detected by immunohistochemical analysis, in β-catenin cAct mice (blue arrows). NP = nucleus pulposus; AF = anulus fibrosus. F and I, Micro–computed tomography images showing that the length of the spine was reduced 18% in 3-month-old (F) and 25% in 6-month-old (I) β-catenin cAct mice. Additional pathologic features in the mutant mice included extensive osteophyte formation (red arrows and green arrows in F and green arrows in I) and disc space narrowing (orange arrows in F and red arrows in I). G, H, and J, Alcian blue/H&E–stained spine sections from 3-month-old β-catenin cAct mice (G), Safranin O–fast green–stained spine sections from 3-month-old β-catenin cAct mice (H), and Alcian blue/H&E–stained spine sections from 6-month-old β-catenin cAct mice (J) showing loss of end plate cartilage and growth plate cartilage and dramatic intervertebral disc tissue disorganization.

Download figure to PowerPoint

The loss of growth plate chondrocytes was detected as early as 2 weeks after tamoxifen induction in 1-month-old β-catenin cAct mice. Formation of chondrocyte clusters was observed in 1-month-old β-catenin cAct mice (Figure 2B), and similar chondrocyte clusters were seen in disc samples from patients with disc degeneration (Figures 1A and F). Early osteophyte formation was detected in 1-month-old β-catenin cAct mice (Figure 2C). New blood vessels and new woven bone formation were also observed at the original location of the growth plate in these mice (Figure 2D).

We next examined phenotypic changes in the disc tissue of 3- and 6-month-old β-catenin cAct mice using micro-CT and histologic analysis. The spine length of 3-month-old β-catenin cAct mice was reduced 18% compared to Cre-negative control mice (Figure 2F). The mean ± SEM length of the spine, from C1 through S4 was 60 ± 0.58 mm (n = 3) in the Cre-negative control mice and 49 ± 2.19 mm (n = 6) in the β-catenin cAct mice (Figure 2F). The average body weight of β-catenin cAct mice was reduced 23% (19.7 ± 2.18 gm in Cre-negative mice [n = 3] and 15.0 ± 1.24 gm in β-catenin cAct mice [n = 6]). At 6 months of age, the mean ± SEM spine length (C1–S4) in the β-catenin cAct mice was reduced 25% compared to that in the Cre-negative control mice (62.5 ± 1.08 mm in control mice [n = 7] and 46.5 ± 1.26 mm in β-catenin cAct mice [n = 4]) (Figure 2I). The average body weight of β-catenin cAct mice at 6 months of age was reduced 39% (26.3 ± 1.95 gm in Cre-negative mice [n = 7] and 16 ± 0.71 gm in β-catenin cAct mice [n = 4]). Extensive osteophyte formation, disc space narrowing, and fusion of adjacent vertebrae were evident throughout the entire spine of β-catenin cAct mice, as determined by micro-CT analysis (Figures 2F and I). The disc space narrowing could be related to the loss of growth plate cartilage and severe osteophyte formation, which blocks the disc space.

Histologic analysis was performed on paraffin-embedded coronal sections, and samples were stained with Alcian blue/H&E and Safranin O–fast green. Histologic results demonstrated a severe loss of end plate and growth plate cartilage, a reduced number of growth plate chondrocytes, and disorganized anulus fibrosus and nucleus pulposus tissue in 3-month-old β-catenin cAct mice (Figure 2G and H). The defects in the disc tissue were worse in 6-month-old mice. Alcian blue/H&E staining showed that the end plate and growth plate cartilage had almost completely disappeared, many fewer chondrocytes remained, and the anulus fibrosus and nucleus pulposus structures were completely damaged (Figure 2J).

To investigate gene expression changes in β-catenin cAct mice, we isolated primary disc cells from 3-week-old β-catenin cAct and Cre-negative control mice for real-time polymerase chain reaction assay. Two-fold and 4-fold increases in the expression of Adamts4 and Adamts5, respectively, were detected in disc cells from β-catenin cAct mice compared to Cre-negative mice (Figures 3A and B). Mmp13 expression was also significantly increased (4-fold) (Figure 3C) and Mmp2 expression was not changed (data not shown) in β-catenin cAct mice. ColX expression was significantly increased (Figure 3D). Alp and osteocalcin expression were also up-regulated (each 2-fold) in β-catenin cAct mice (Figures 3E and F). In contrast, Col2a1 and Col9a1 expression were significantly reduced in the disc cells of β-catenin cAct mice compared to Cre-negative mice (Figures 3G and H). Consistent with the findings of increased Mmp13 messenger RNA (mRNA) expression, MMP-13 protein levels were significantly increased in the disc tissue of 4-week-old β-catenin cAct mice (Figure 3I).

thumbnail image

Figure 3. Changes in expression of disc tissue–specific genes and proteins in mice. A–H, Significant increase in the expression of Adamts4 (2-fold) (A), Adamts5 (4-fold) (B), Mmp13 (4-fold) (C), ColX (2.5-fold) (D), Alp (2-fold) (E), and osteocalcin (2-fold) (F) and significant reduction in Col2a1 (G) and Col9a1 (H) mRNA levels in disc cells obtained from 3-week-old β-catenin conditional activation (cAct) mice as compared to Cre-negative control mice. Total RNA was extracted from primary disc cells isolated from the mice and subjected to real-time polymerase chain reaction analysis. Bars show the mean ± SEM. ∗ = P < 0.05 by Student's unpaired t-test. I–K, Immunohistochemical analysis, demonstrating a notable increase in matrix metalloproteinase 13 protein expression (green arrows) (I), a slight reduction in type II collagen protein levels (black arrow) (J), and a significant reduction in type IX collagen protein levels (blue arrow) (K) in 1-month-old β-catenin cAct mice as compared to Cre-negative control mice.

Download figure to PowerPoint

To further determine whether activation of β-catenin signaling leads to the up-regulation of Mmp13 in human chondrocytes, we treated human articular chondrocytes with BIO (a GSK-3β inhibitor) for 48 hours. BIO stimulated Mmp13 expression in human articular chondrocytes (data not shown). To determine changes in protein levels of type II and type IX collagen, we performed immunohistochemical analysis using mouse monoclonal antibodies against Col2a1 and Col9a1 proteins (Developmental Studies Hybridoma Bank) and found that protein levels of type II collagen were slightly reduced, but protein levels of type IX collagen were markedly reduced, in β-catenin cAct mice compared to Cre-negative control mice (Figures 3J and K).

We next performed in vitro studies using the rat chondrosarcoma chondrogenic cell line, in order to determine the signaling mechanism by which β-catenin regulates Mmp13 expression, and to provide initial insight into the mechanism underlying the disc phenotype seen in β-catenin cAct mice. Treatment of rat chondrosarcoma cells with BIO (a GSK-3β inhibitor which induces β-catenin activation) had no effect on Mmp13 expression within 12 hours, but significantly up-regulated Mmp13 mRNA expression at the 24- and 48-hour time points (Figure 4A). Interestingly, treatment with BIO (for 48 hours) also significantly up- regulated runt-related transcription factor 2 (RUNX-2) protein expression in rat chondrosarcoma cells (Figure 4B), which is consistent with the up-regulation of Runx2 mRNA in the disc cells of β-catenin cAct mice (Figure 4C). The regulation of the Mmp13 gene by RUNX-2 has been reported in recent years (20–22).

thumbnail image

Figure 4. Activation of β-catenin signaling stimulates Mmp13 gene transcription in a runt-related transcription factor 2 (RUNX-2)–dependent manner. A, Effects of BIO, a glycogen synthase kinase 3β (GSK-3β) inhibitor, on Mmp13 mRNA expression in rat chondrosarcoma chondrogenic cells. Mmp13 expression was significantly up-regulated in rat chondrosarcoma cells after incubation with BIO (1 μM) for 24 or 48 hours. B, Significant up-regulation of RUNX-2 protein expression in BIO-treated rat chondrosarcoma cells (48-hour incubation). C, Immunohistochemical analysis showing dramatic up-regulation of RUNX-2 protein levels (arrows) in the disc cells of 3-week-old β-catenin conditional activation (cAct) mice. D, Mmp13 promoter activity. The 3.4-kb human Mmp13 promoter was cloned into the pGL3 vector, and a conserved RUNX-2 binding site was found in the proximal region of the human Mmp13 promoter (−138 to −132). Transfection of RUNX-2 or treatment with BIO (1 μM) significantly enhanced luciferase activity of the Mmp13 promoter. Mutation of the RUNX-2 binding site completely abolished the stimulatory effects of RUNX-2 as well as BIO on Mmp13 promoter activity in rat chondrosarcoma cells. WT = wild type. E, Significant inhibition of BIO-induced Mmp13 promoter activity by transfection of RUNX-2 small interfering RNA (siRNA). F, Results of chromatin immunoprecipitation assay, showing that RUNX-2 specifically bound to the proximal promoter region of the Mmp13 promoter, but not the upstream or downstream distant region of the Mmp13 gene. In A, D, and E, bars show the mean ± SEM (n = 4 samples per group). ∗ = P < 0.05 versus controls, by one-way analysis of variance followed by Dunnett's test.

Download figure to PowerPoint

To directly assess the effect of RUNX-2 on Mmp13 gene transcription and to determine if it might mediate the overall effect of β-catenin on Mmp13, we cloned the 3.4-kb human Mmp13 promoter into the pGL3 luciferase reporter plasmid. DNA sequencing of this Mmp13 promoter identified a RUNX-2 binding site in the proximal region (−138 to −132; AACCACA), which is conserved among human, mouse, and rat species. We found that transfection of RUNX-2 as well as treatment with BIO significantly stimulated Mmp13 promoter activity (Figure 4D). Mutation of the RUNX-2 binding site (mutant sequence ACTAACA) completely abolished the stimulatory effect of RUNX-2 as well as BIO-induced Mmp13 promoter activity (Figure 4D), suggesting that BIO (or activation of β-catenin signaling) may stimulate Mmp13 gene transcription through up-regulation of the transcription factor RUNX-2.

To further determine the role of RUNX-2 in BIO-induced Mmp13 promoter activity, we treated rat chondrosarcoma cells with BIO with or without transfection of Runx2 small interfering RNA (siRNA) and found that transfection of Runx2 siRNA completely inhibited BIO-induced Mmp13 promoter activity (Figure 4E). Results of chromatin immunoprecipitation assays further demonstrated that RUNX-2 specifically bound to the proximal region of the Mmp13 promoter, which contains the RUNX-2 binding site but not the upstream or downstream distant region of the Mmp13 gene (Figure 4F). These results demonstrate that activation of β-catenin signaling could stimulate RUNX-2 expression, which subsequently promotes Mmp13 gene transcription.

MMP-13 is a collagenase that degrades type II and type IX collagen, and ADAMTS-5 is an aggrecanase that degrades aggrecan. Type II and type IX collagen and aggrecan are the principal components of disc tissue. We observed significant up-regulation of Mmp13 and Adamts5 expression in β-catenin cAct mice. Because both MMP-13 and ADAMTS-5 play critical roles in the development of OA (23–25), we reasoned that Mmp13 and Adamts5 may be the key downstream target genes of β-catenin signaling in disc cells during the development of the defective disc phenotype. To test this hypothesis, we bred β-catenin cAct mice with Mmp13fx/fx mice (11) or Adamts5−/− mice (26) to produce Col2a1-CreERT2; β-cateninfx(Ex3)/wt;Mmp13fx/fx or Col2a1-CreERT2;β-cateninfx(Ex3)/wt;Adamts5−/− double-mutant mice.

Deletion of the Mmp13 gene in β-catenin–induced disc cells was confirmed by immunohistochemistry (data not shown). Micro-CT data showed that deletion of the Mmp13 or the Adamts5 gene in β-catenin cAct mice significantly ameliorated the phenotype of disc defects in 3-month-old mice, including the loss of growth plate cartilage and disc space narrowing (Figure 5A). In 6-month-old β-catenin cAct mice, deletion of the Mmp13 gene significantly decelerated the disc defects (Figure 5B). Although deletion of the Adamts5 gene in β-catenin cAct mice ameliorated disc defects to a certain degree, it was less potent than deletion of the Mmp13 gene (Figures 5A and B).

thumbnail image

Figure 5. Amelioration of the disc defect phenotype by deletion of the Mmp13 or Adamts5 gene in β-catenin conditional activation (cAct) mice. A and B, Results of micro–computed tomography (micro-CT) of the L3/4 disc in 3-month-old (A) and 6-month-old (B) Cre-negative, β-catenin cAct, β-catenin cAct;Mmp13 conditional knockout (cKO), and β-catenin cAct;Adamts5-knockout mice. Deletion of the Mmp13 or Adamts5 gene significantly ameliorated the disc space narrowing (orange arrows) and osteophyte formation (red arrows) observed in 3-month-old and 6-month-old β-catenin cAct mice. In the β-catenin cAct;Adamts5–knockout group, osteophyte formation was reduced compared to that in the β-catenin cAct group. C and D, Alcian blue/hematoxylin and eosin (H&E) staining, showing dramatic loss of growth plate cartilage (yellow arrows) in 3-month-old (C) and 6-month-old (D) β-catenin cAct mice. The structures of anulus fibrosus and nucleus pulposus tissues were severely disorganized in β-catenin cAct mice (green arrows). Deletion of the Mmp13 or Adamts5 gene significantly prevented the loss of growth plate cartilage tissue (yellow arrows) and preserved the structure disorganization of the anulus fibrosus and nucleus pulposus tissues (green arrows) observed in β-catenin cAct mice. E and F, Growth plate cartilage areas in 3-month-old (E) and 6-month-old (F) mice, as measured by histomorphometry. Bars show the mean ± SEM (n = 5 mice per group). ∗ = P < 0.05; # = P < 0.05, by one-way analysis of variance followed by Dunnett's test.

Download figure to PowerPoint

Results of histologic and histomorphometric analysis further demonstrated maintenance of normal disc tissue morphology, with a significant increase in proteoglycan levels and restoration of the normal growth plate cartilage, when the Mmp13 gene or the Adamts5 gene was deleted on the β-catenin cAct background in 3-month-old double-mutant mice (Figure 5C). In 6-month-old β-catenin cAct mice, deletion of the Mmp13 gene was more effective than deletion of the Adamts5 gene in reducing disc defects (Figure 5D). These results indicate that Mmp13 and Adamts5 are the critical downstream target genes of β-catenin signaling in disc cells and may play key roles in mediating β-catenin–induced disc defects. Since Mmp13 conditional knockout mice and Adamts5-knockout mice had relatively normal disc phenotypes, the results observed in β-catenin cAct;Mmp13 conditional knockout and β-catenin cAct;Adamts5−/− double-mutant mice were mainly caused by rescuing the β-catenin activation phenotype in disc cells.

Since β-catenin–dependent up-regulation of Mmp13 is a potential mechanism underlying disc defects, we investigated the efficacy of the MMP-13 inhibitor CL82198 in reversing the disc phenotype in β-catenin cAct mice. CL82198 is a specific MMP-13 inhibitor that binds to the S1′ pocket of MMP-13 rather than via metal chelation. It has no effect on MMP-1, MMP-9, or TACE (27–29) and so is relatively specific for inhibition of MMP-13 activity. CL82198 (5 mg/kg) was injected IP every other day into β-catenin cAct mice immediately after tamoxifen induction. Mice were killed at 3 months of age, and the effects of CL82198 on the disc phenotype were analyzed by micro-CT and histology. The disc space narrowing seen on micro-CT and the loss of proteoglycan and disorganized anulus fibrosus and nucleus pulposus tissue found in histologic and histomorphometric analyses were notably ameliorated by MMP-13 inhibition (Figures 6A–D). These results suggest that the use of an MMP-13 inhibitor may represent a potential therapy for the treatment of diseases such as disc degeneration.

thumbnail image

Figure 6. Inhibition of matrix metalloproteinase 13 (MMP-13) enzyme activity in β-catenin conditional activation (cAct) mice ameliorates the disc defect phenotype. A, Results of micro–computed tomography of the L1 to L6 discs in Cre-negative mice treated with phosphate buffered saline (PBS), β-catenin cAct mice treated with PBS, Cre-negative mice treated with CL82198, and β-catenin cAct mice treated with CL82198. Treatment with the MMP-13 inhibitor CL82198 significantly ameliorated the disc space narrowing (orange arrows) and osteophyte formation (red arrows) observed in β-catenin cAct mice. B, Alcian blue/hematoxylin and eosin staining showing dramatic loss of growth plate cartilage and proteoglycan (yellow arrows) in β-catenin cAct mice. Treatment with the MMP-13 inhibitor CL82198 significantly reversed the cartilage loss and disc structure disorganization phenotype observed in β-catenin cAct mice. C and D, Change in disc space (C) and growth plate cartilage area (D), as measured by histomorphometry. Bars show the mean ± SEM (n = 5 mice per group). ∗ = P < 0.05; # = P < 0.05, by one-way analysis of variance followed by Dunnett's test.

Download figure to PowerPoint

DISCUSSION

  1. Top of page
  2. Abstract
  3. MATERIALS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. AUTHOR CONTRIBUTIONS
  7. REFERENCES

The present study is the first to demonstrate that β-catenin protein is up-regulated and activated in disc tissue from patients with disc degeneration. We created a β-catenin cAct mouse model with a phenotype that resembles some of the features of human disc degeneration. Similar findings have been reported in recent studies using transgenic mice that overexpress the β-catenin gene (30). The β-catenin cAct mice have reduced spine length, which is likely related to the general growth retardation seen in these mice. We also observed severe loss of proteoglycan and growth plate cartilage, severely disorganized anulus fibrosus and nucleus pulposus tissue, and prevalence of osteophyte formation in β-catenin cAct mice. Although β-catenin cAct mice have multiple features that resemble human disc degeneration, this mouse model may only partially mimic the human disc diseases. One reason is that nucleus pulposus cells were not targeted by Col2a1-CreERT2–transgenic mice. The change in the shape of the nucleus pulposus may be caused by the osteophyte formation and structural changes in anulus fibrosus tissue. It is also possible that the defect seen in nucleus pulposus tissue is related to the disruption of nutrient and solute supplies to this region after the loss of the growth plate cartilage in β-catenin cAct mice.

We found that Mmp13 mRNA and protein expression were significantly up-regulated in β-catenin cAct mice, and our in vitro experiments demonstrated that activation of β-catenin signaling stimulates Mmp13 gene transcription in a RUNX-2–dependent manner. Determining the role of RUNX-2 in β-catenin–induced disc defects in vivo will require genetically targeting the Runx2 gene, and this important aspect needs to be further investigated. Deletion of the Mmp13 gene or the Adamts5 gene on the β-catenin cAct background significantly reversed the disc destruction phenotype observed in β-catenin cAct mice. It seems that deletion of the Mmp13 gene is more efficient than deletion of the Adamts5 gene in protecting against the loss of growth plate cartilage and preserving the disc tissue structure.

These results indicate that Mmp13 may act as a major downstream target of β-catenin signaling in disc cells, although both Mmp13 and Adamts5 were up-regulated in disc cells in β-catenin cAct mice. Notably, treatment of β-catenin cAct mice with an MMP-13 inhibitor also significantly inhibited disc defects. This study demonstrates that β-catenin is a key mediator causing defects in disc tissue. A recent study showed that CBP/p300–interacting transactivator with ED-rich tail 2 (CITED2) is a negative regulator of MMP-1 in articular chondrocytes and has a chondroprotective effect (31). Taken together, these findings suggest that either inhibition of MMP-13 or activation of CITED2 could serve as potential treatments of disc degeneration in patients.

Histologic and histomorphometric analysis showed that treatment with MMP-13 inhibitor was more effective than deletion of the Mmp13 gene in protecting against the disc defects observed in β-catenin cAct mice. There are two possible reasons for this. Although the MMP-13 inhibitor used in this study is relatively specific for MMP-13, it does have some additional nonspecific effects on inhibition of other MMPs that may contribute to the protection against the disc defects observed in β-catenin cAct mice. Alternatively, the Cre recombination efficiency mediated by Col2a1-CreERT2 is ∼80% in inner anulus fibrosus cells and ∼75% in growth plate chondrocytes (19); thus, MMP-13 activity in the other 20–25% of inner anulus fibrosus cells and growth plate chondrocytes as well as the entire population of nucleus pulposus cells remains normal in β-catenin cAct;Mmp13 conditional knockout double-mutant mice.

Although deletion of the Mmp13 gene significantly reversed the disc degeneration phenotype observed in β-catenin cAct mice, the phenotype was not completely protected by deletion of the Mmp13 gene. There was still significant loss of proteoglycans and growth plate cartilage tissue in these double-mutant mice compared to Cre-negative controls. It is known that deletion of Adamts5 protects against the development of OA (25, 32, 33). Since mRNA expression for both of these genes was significantly up-regulated in the disc tissue in β-catenin cAct mice, the impact of aggrecanase enzymatic activity likely contributed to the development of disc defects observed in β-catenin cAct mice. We found that deletion of the Adamts5 gene on the β-catenin cAct background reversed the disc destruction phenotype in β-catenin cAct mice as well, although it was less effective than deletion of the Mmp13 gene on the β-catenin cAct background. These results indicate that Mmp13 may play a major role in the β-catenin–mediated disc destruction observed in β-catenin cAct mice.

The present study shows that β-catenin plays a key role in activating downstream target genes such as Mmp13 and Adamts5 in disc cells, leading to severe defects in disc tissue. However, how β-catenin signaling is activated during disc degeneration remains unknown. Potential causes of β-catenin activation in disc cells include an activation mutation of the β-catenin gene or mutations of other genes that regulate and inhibit β-catenin signaling, such as Axin-1 and APC, injury or mechanical loading–induced β-catenin activation, and inflammation-induced β-catenin activation. The activation of mutations of the β-catenin gene has been linked to the development of multiple cancers, including colorectal cancer, acute and chronic myeloid leukemia, and multiple myeloma (34–37). Several lines of evidence have demonstrated that mechanical loading leads to activation of β-catenin signaling in bone cells (38–40). Finally, TNFα and IL-1β are two important inflammatory cytokines involved in the development of OA. Recent studies have demonstrated that these cytokines activate β-catenin signaling in other cell types (41, 42). Thus, they may also play a role in the activation of β-catenin signaling in disc cells and contribute to the development of disc diseases.

AUTHOR CONTRIBUTIONS

  1. Top of page
  2. Abstract
  3. MATERIALS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. AUTHOR CONTRIBUTIONS
  7. REFERENCES

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. Drs. Y. Wang and Chen had full access to all of the data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis.

Study conception and design. M. Wang, Tang, Shu, B. Wang, Jin, Hao, Dresser, Shen, Im, Sampson, Rubery, Zuscik, Schwarz, O'Keefe, Y. Wang, Chen.

Acquisition of data. M. Wang, Tang, Shu, B. Wang, Jin, Hao, Dresser, Shen, Im, Sampson, Rubery, Zuscik, Schwarz, O'Keefe, Y. Wang, Chen.

Analysis and interpretation of data. M. Wang, Tang, Shu, B. Wang, Jin, Hao, Dresser, Shen, Im, Sampson, Rubery, Zuscik, Schwarz, O'Keefe, Y. Wang, Chen.

REFERENCES

  1. Top of page
  2. Abstract
  3. MATERIALS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. AUTHOR CONTRIBUTIONS
  7. REFERENCES