To determine expression patterns of apoptotic and matrix-degrading genes during aging and development of osteoarthritis (OA), using a rabbit model of induced OA.
To determine expression patterns of apoptotic and matrix-degrading genes during aging and development of osteoarthritis (OA), using a rabbit model of induced OA.
Six mature and 6 aged rabbits underwent anterior cruciate ligament transection and were killed 4 and 8 weeks after surgery, respectively, to create early-grade and advanced-grade OA. RNA from articular cartilage and menisci was examined for expression of the genes caspase 8, Fas, Fas ligand, p53, aggrecanase, matrix metalloproteinase 1 (MMP-1), and MMP-3. A second cohort of animals that had undergone no intervention in the joint was also killed. Parametric data were analyzed with analysis of variance and Student's t-tests, while nonparametric data were assessed with the Mann-Whitney U test.
Expression levels of Fas, caspase 8, FasL, and MMP-1 were significantly higher (>100%) in aged cartilage compared with mature cartilage (P < 0.05). After induction of OA, expression of apoptotic genes in aged rabbits remained high, while significant up-regulation of Fas and caspase 8 (nearly 150% increase) was observed in mature rabbits (P < 0.05). No significant up-regulation of these genes was observed in the menisci of aged or mature rabbits prior to or after induction of OA. Development of OA occurred more rapidly in aged cartilage compared with mature cartilage (P < 0.05).
Differential expression of apoptotic and matrix-degrading genes occurs in aged compared with mature cartilage, both at baseline and during development of OA. This may be responsible for faster degradation of aged cartilage and its predisposition for developing OA.
Articular cartilage in adult animals and humans undergoes several age-related changes, including a decrease in the number of chondrocytes and a reduction in the extracellular matrix (ECM) (1–10). Recent animal studies have shown increased levels of apoptosis in aged cartilage compared with mature cartilage, suggesting that apoptosis, or programmed cell death, may be responsible for chondrocyte hypocellularity during the aging process (11, 12). Because chondrocytes maintain the dynamic equilibrium between production of the ECM and its enzymatic degradation, loss of cell viability may predispose aged individuals to matrix degradation and the development of OA.
Osteoarthritis (OA) is a progressive degradative joint disease that is characterized by articular cartilage breakdown, and, much like aging, it is associated with reductions in the ECM and chondrocyte hypocellularity. The exact etiology of the hypocellularity in OA is unclear, but most investigators currently believe that chondrocyte apoptosis plays a central role (13–18). In situ TUNEL and electron microscopy studies have shown increased levels of chondrocyte apoptosis in OA tissue, and further studies have correlated OA severity with higher rates of apoptosis (19–22). Several mediators that induce chondrocyte apoptosis have been identified, including nitric oxide, Fas/FasL, p53, and tumor necrosis factor α (17, 23–27). Moreover, previous studies have demonstrated that synovial fluid from patients with OA has elevated levels of nitric oxide and FasL (27–29).
Interest in chondrocyte apoptosis has grown in the past decade; however, the role of matrix-degrading enzymes in OA cannot be overstated. In OA, an imbalance exists between locally synthesized matrix metalloproteinases (MMPs) and their inhibitors, favoring a catabolic state that leads to proteolysis and cartilage destruction (30, 31). Furthermore, several MMPs, including MMP-1 and MMP-3, as well as aggrecanase have been identified in situ in OA cartilage. The level of expression of these enzymes was also shown to correlate with the histologic severity of OA (32–35). Also of note is a recent study showing that matrix degradation, especially collagen breakdown, leads to chondrocyte apoptosis through a Fas-dependent pathway (21).
Although these studies illustrate several shared characteristics between aging and OA, the exact relationship between these 2 entities remains largely unexplained. Therefore, the purpose of the present study was to determine whether gene expression differences exist in aged cartilage or menisci that predispose aged rabbits to the development of OA, and then to induce OA using an anterior cruciate ligament (ACL) transection model to identify differential gene expression patterns in aged rabbits compared with mature rabbits over the spectrum of early to advanced OA.
This study was approved by the University of California San Diego Institutional Animal Care and Use Committee. Two groups of NZW rabbits (mature and aged) underwent ACL transection. Mature rabbits were defined as 10–18 months of age with closed epiphyses, and aged rabbits were defined as older than 4 years of age. For the ACL transection, animals were first appropriately sedated and anesthetized. The left knee was then shaved, prepped, and draped in a sterile manner. A 3-cm medial parapatellar incision was made, and soft tissue was carefully dissected until the joint capsule could be opened. The patella was then dislocated, and the ACL was isolated and then sharply divided. ACL transection was confirmed both visually and with Lachman testing by both the surgeon and an observer. The wounds were then closed and antiseptically treated. Rabbits were given appropriate postoperative care and analgesia. All animals were allowed normal cage activity and were killed at either 4 weeks or 8 weeks following ACL transection. Another group of rabbits was maintained as normal, unoperated controls.
After the rabbits were killed, gross morphologic grading was performed, using a modified Outerbridge classification system, where grade I = normal cartilage, grade II = partial thickness defect with fissures on the surface that did not reach subchondral bone, grade III = fissuring to the level of subchondral bone, and grade IV = exposed subchondral bone (36). Additionally, any changes in the menisci were noted. Femoral and tibial articular cartilage specimens from both the medial and lateral compartments as well as the medial and lateral menisci (n = 6 per group) were then immediately harvested from the knee joints.
After being fixed in 10% buffered formalin, the specimens were placed in 5% hydrochloric acid for 17 hours. The decalcified tissue samples were then sequentially dehydrated in alcohol and embedded into paraffin blocks before being sectioned at 6 μm. The sections were stained with hematoxylin and eosin (H&E).
The samples of cartilage and menisci were pulverized in liquid nitrogen, and total RNA was isolated using the acid guanidinium thiocyanate–phenol extraction procedure (37). Starting with 1 μg of total RNA, first-strand complementary DNA was synthesized using oligo(dT)15 primers. Based on published sequences (BLAST), PCR primer sets specific to selected coding regions of Fas, FasL, caspase 8, p53, aggrecanase, MMP-1 and MMP-3, and GAPDH were constructed (12, 37). Cycle studies were undertaken for all genes to determine the linear range of amplification for each gene, using 34 cycles for all genes except GAPDH (for which 30 cycles were used). NIH Image software version 1.61 (National Institutes of Health, Bethesda, MD) was used to quantitatively scan RT-PCR profiles following agarose gel electrophoresis and ethidium bromide visualization. The NIH software measures relative mean density over a fixed gray scale range, after correction for background. All values were normalized to GAPDH.
Total protein was extracted from lyophylized articular cartilage in 3.5M urea, phosphate buffered saline (PBS) overnight at 4°C. Collagenous proteins were precipitated out with 0.02M disodium phosphate, and the remaining supernatant containing caspase 8 and other soluble proteins was dialyzed against water and lyophylized. The residue was dissolved in PBS, the total protein content was determined using the Bradford method, and immunoblot analysis was performed as follows. Equal amounts of total protein in the extracts were applied to nitrocellulose membranes and blocked for 24 hours with 3% nonfat dried milk, 1% bovine serum albumin in Tris buffered saline (TBS) at 4°C. A primary antibody (chicken anti-rabbit caspase 8; US Biological, Swampscott, MA) was applied to the membrane, followed by incubation at room temperature for 2 hours with gentle rotation. After washing in TBS–Tween (TBST), the secondary antibody (biotin-labeled anti-chicken IgG; US Biological) was applied, followed by incubation for 2 hours at room temperature. The membranes were then washed in TBST, and the blots were visualized by chemoluminescence (luminol reagent; Santa Cruz Biotechnology, Santa Cruz, CA). Negative controls consisted of blots without primary antibody.
Pilot data from preliminary experiments demonstrated a 25% coefficient of variation. For these experiments, we chose a level of significance (α) of 0.05 and desired a statistical power (1 − β) of 80%. On the basis of these criteria and an expected treatment effect of 50%, we calculated that the number of independent specimens required per group was 6. In this study, all parametric data were evaluated using unpaired Student's t-tests and analysis of variance followed by post hoc Bonferroni testing, with a level of significance (α) set at 0.05 (StatView software; SAS Institute, Cary, NC). Because the macroscopic grading of the cartilage was nonparametric, differences between groups were analyzed using the Mann-Whitney U test; P values less than 0.05 were considered significant.
All control rabbits, both aged and mature, had normal grade 1 cartilage with normal-appearing menisci. Four weeks after ACL transection, the articular cartilage overall had begun to degrade (mean ± SD grade 1.8 ± 0.3) relative to the control specimens (P < 0.05), and the aged cartilage (mean ± SD grade 2.0 ± 0.2) degraded faster than the mature cartilage (1.7 ± 0.2; P < 0.05). Gross examination of the meniscus showed no evidence of thinning, fraying, or tearing in either aged or mature animals. Eight weeks after ACL transection, the articular cartilage overall had further degraded (mean ± SD grade 2.4 ± 0.4) relative to the week 4 specimens (P < 0.05), and, again, the aged rabbits appeared to have higher-grade OA (mean ± SD grade 2.5 ± 0.3) compared with the mature rabbits (2.2 ± 0.3; P = 0.08) (Table 1). The morphologic appearance of the menisci at this time had begun to show evidence of thinning in both the aged and mature rabbits, especially in the medial compartment, but no fraying, tearing, or resorption was noted.
|Normal||1.0 ± 0.0||1.0 ± 0.0||1.0 ± 0.0|
|Early OA (4 weeks after ACL transection)||1.7 ± 0.2||2.0 ± 0.2‡||1.8 ± 0.3|
|Advanced OA (8 weeks after ACL transection)||2.2 ± 0.3||2.5 ± 0.3§||2.4 ± 0.4|
H&E staining demonstrated hypocellularity and open lacunae of apoptized chondrocytes in aged cartilage, which are associated with a more fibrillated surface and striated articular matrix. In the deep radial zone of normal mature cartilage, the characteristic columnar orientation of chondrocytes was prevalent, while this structure was deteriorating in aged cartilage, particularly below the tide mark (Figure 1).
Analysis of articular cartilage from normal non–ACL-transected mature and aged rabbits revealed significantly higher expression of several apoptotic genes in aged rabbits compared with mature rabbits. These included Fas, FasL, and caspase 8 (P < 0.05) (Figure 2). Additionally, expression of the catabolic gene MMP-1 was significantly higher in aged rabbits compared with mature rabbits (P < 0.05) (Figure 2). When comparing aged meniscus and mature meniscus prior to the induction of OA, no significant differences were observed in any of the apoptotic or matrix-degrading genes.
In mature articular cartilage, apoptotic gene expression was relatively low at baseline but was up-regulated during the development of OA (early stage, 4 weeks after ACL transection), as seen by the significant increases in Fas and caspase 8 expression levels (P < 0.05) (Figure 3A). Eight weeks after ACL transection, expression of FasL showed a trend (P = 0.09) for increased expression relative to control, while expression of the other apoptotic genes was similar to control levels. In aged cartilage, apoptotic gene expression was relatively high at baseline and remained unchanged during the development of OA, with the exception of caspase 8, which demonstrated a near-significant (P = 0.08) increase 4 weeks after ACL transection. At 8 weeks, caspase 8 showed decreased expression compared with control (Figure 3A).
Expression of matrix-degradation gene mRNA in mature articular cartilage demonstrated changes relative to control during the early stages of OA (i.e., 4 weeks after ACL transection) (Figure 3B). Specifically, levels of MMP-1 and MMP-3 were increased (P = 0.07 and P = 0.09, respectively). Aged rabbits showed increases in aggrecanase (P = 0.07) and MMP-3 (P < 0.05) at this time point. Eight weeks after ACL transection, expression of both aggrecanase and MMP-1 was reduced in both mature and aged cartilage, while MMP-3 expression in aged cartilage remained higher than control (P < 0.05).
In mature menisci, there were no significant differences in expression of apoptotic or matrix-degrading genes during the development of OA. In aged meniscus, no differences in gene expression were observed between the control and the early OA groups. In advanced OA, however, expression of several genes, including caspase 8, p53, MMP-1, and MMP-3, decreased significantly.
Immunoblot analysis revealed the presence of caspase 8 protein in articular cartilage. Figure 4 shows that 4 weeks after ACL transection, expression of caspase 8 protein in mature cartilage showed greater intensity than its contralateral control. Similar results were observed in aged articular cartilage at 4 weeks. Eight weeks after ACL transection, cartilage from mature control and mature ACL-transected animals showed similar intensities, while cartilage from aged ACL-transected rabbits showed decreased caspase 8 intensity compared with control.
To date, the genetic and molecular changes leading to the distinctive alterations of aged cartilage and its propensity for developing OA are unknown. In an attempt to address this complex issue, we designed a study with 3 specific aims, as follows: 1) to compare apoptotic and matrix-degrading gene expression in mature and aged rabbits to determine molecular alterations that may predispose aged cartilage to the development of OA, 2) to induce OA using an ACL transection model to elucidate molecular mediators of OA, and 3) to determine whether differential gene expression patterns occur in aged rabbits compared with mature rabbits over the spectrum of early to advanced OA.
Toward the first aim, we demonstrated that aged rabbits with macroscopically normal cartilage have higher rates of expression of apoptotic genes, including Fas, FasL, and caspase 8, compared with mature rabbits with a similar grade of cartilage. These results compare favorably with the results described by Allen et al (12), who showed that increased expression of Fas, FasL, caspase 8, and p53 in aged rabbit cartilage positively correlated with decreased chondrocyte density. Additionally, we observed increased expression of MMP-1 in normal aged cartilage compared with normal mature cartilage. This finding may provide a mechanism for the observations of other investigators in both humans and animals, that aging results in a loss of proteoglycans within the ECM and thinning of the articular cartilage (2, 8, 38, 39). Overall, these results suggest that although cartilage in aged rabbits may appear grossly normal, microscopic and molecular changes have occurred, including increased expression of apoptotic and matrix-degrading genes with loss of viable chondrocytes and ECM, which may predispose the tissue to OA.
Toward the second aim, we used an ACL transection model to induce OA in rabbits. Gross analysis of the articular cartilage surface 4 weeks after ACL transection revealed that the cartilage had degraded to early-stage OA (mean ± SD Outerbridge grade 1.8 ± 0.3); 8 weeks after ACL transection, the cartilage had further degraded to a more advanced stage of OA (Outerbridge grade 2.4 ± 0.4). These findings are consistent with the finding of other studies that used ACL transection models of OA, including those by Hashimoto et al (25), Batiste et al (40), and Le Graverand et al (41), who noted superficial cartilage irregularities at 4 weeks and more extensive lesions by 8 weeks, both grossly and histologically. Of note, aged cartilage degraded grossly faster than mature cartilage. This is an important finding, because we are aware of no study that has shown a more rapid progression of OA in aged tissue compared with mature tissue, using an animal model. We attribute this faster progression of cartilage degradation to the predisposition of aged cartilage to OA, as described above.
Finally, toward the third aim, we examined the differential expression of apoptotic and matrix-degrading genes in mature and aged rabbits over the spectrum of early to advanced OA. Interestingly, there were several key differences between the aged and mature cohorts. In mature rabbits, the key step in the development of OA appeared to be the increased expression of caspase 8 and Fas in the early stages of disease. Based on this finding, we hypothesize that increased expression leads to chondrocyte apoptosis and decreased density of viable chondrocytes. This, in turn, impairs the ability of the tissue to maintain the ECM, contributing to cartilage degradation. In aged rabbits, a statistically significant up-regulation of apoptotic genes was not observed, with the exception of caspase 8, which showed a strong trend for up-regulation (P = 0.08). Importantly, these results were mirrored by the protein expression levels of caspase 8, as demonstrated in the immunoblots.
We also examined the meniscus as a possible contributor to the processes of aging and the development of OA. The meniscus acts both to increase joint congruity and stability and to share load across the knee. Loss of meniscus, whether by injury or surgical removal, can lead to the development of OA. Also, with age, the meniscus desiccates, has diminished repair ability, and is more susceptible to degenerative tears. We hypothesized that in our model of ACL transection, the meniscus would also undergo injury, inducing alterations in gene expression that could prove detrimental to cartilage. However, contrary to our hypothesis, we observed no differences in gene expression or in gross examination at baseline, and after the induction of OA, no genes were up-regulated in either the aged or mature cohorts.
These findings suggest that in our ACL transection model, the role of the meniscus in the development of OA is likely minimal, at least with respect to the genes tested. These results slightly differ from those described by Hellio Le Graverand et al (42, 43), who assessed meniscal gene expression in rabbits after ACL transection. In that study, no changes in MMP expression were observed in the lateral meniscus during the development of OA, but several MMPs were up-regulated in the medial meniscus, particularly 8 weeks after surgery. The difference in results between the 2 studies likely stems from our pooling of the medial and lateral meniscus samples. This was done to yield more RNA, so that additional genes could be examined.
Based on these data, we believe the 2 key features in the development of OA are the health and density of viable chondrocytes and the status of the ECM. With chondrocyte apoptosis, through either aging or altered biomechanics (as seen in the ACL transection model), cartilage loses its ability to generate and maintain the ECM. As the matrix degrades, it becomes increasingly sensitive to alterations or imbalances in joint mechanics or matrix-degrading enzymes. Beyond a critical threshold, the joint loses its ability to further maintain the ECM, leading to reduced structural integrity of the cartilage and ultimate breakdown of the joint. The final stages of this process result in a self-propagating cycle that is independent of apoptosis and matrix-degrading gene expression, as was seen in this study. The principal role that aging plays in this model is expedition of the loss of ECM by the hypocellularity and open lacunae from apoptized chondrocytes in aged cartilage (Figure 1). This is effected by the increased baseline apoptosis and increased enzymatic degradation of the matrix, which ultimately lead to more rapid progression of OA. A schematic depiction of this hypothesized model is shown in Figure 5.
Two weaknesses of this study stemmed largely from tissue-related limitations. First, the amount of available tissue from a rabbit condyle and tibial plateau enabled only a limited number of genes to be assessed. Therefore, we selected representative apoptotic and matrix-degrading genes that have been shown to play important roles in the development of OA in previous studies (12, 44). Second, the results of this study are based on RNA extracted from the entire cartilage surface of the affected joint and not on focused regions of cartilage degradation. Therefore, although our study permits an assessment of the entire articular surface, it is difficult to assess what the specific gene expression is as a function of regional variations in degradation. Future studies focusing on regional gene expression within specific cartilage lesions would further advance our understanding of the molecular alterations in OA.
In conclusion, in this limited study we examined the expression of proapoptotic and matrix-degrading genes in a rabbit model of aging and induced OA. Although this is a complex interaction, several genes were identified that appear to play key roles in both the aging process and development of OA and likely explain the differential rates of degradation in aged and mature cartilage. It is our goal and hope that improved understanding of the molecular biology of aging and OA may lead to clinical solutions for a prevalent and debilitating disease.
Dr. Amiel 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 design. Amiel.
Acquisition of data. Pennock, Robertson, Emmerson.
Analysis and interpretation of data. Pennock, Amiel.
Manuscript preparation. Pennock, Harwood, Amiel.
Statistical analysis. Pennock, Harwood, Amiel.