Accelerated, aging-dependent development of osteoarthritis in α1 integrin–deficient mice

Authors


Abstract

Objective

Cell–matrix interactions regulate chondrocyte differentiation and survival. The α1β1 integrin is a major collagen receptor that is expressed on chondrocytes. Mice with targeted inactivation of the integrin α1 gene (α1-KO mice) provide a model that can be used to address the role of cell–matrix interactions in cartilage homeostasis and osteoarthritis (OA) pathogenesis.

Methods

Knee joints from α1-KO and wild-type (WT) BALB/c mice were harvested at ages 4–15 months. Knee joint sections were examined for inflammation, cartilage degradation, and loss of glycosaminoglycans (by Safranin O staining). Immunohistochemistry was performed to detect the distribution of α1 integrin, matrix metalloproteinases (MMPs), and chondrocyte apoptosis.

Results

In WT mice, the α1 integrin subunit was detected in hypertrophic chondrocytes in the growth plate and in a subpopulation of cells in the deep zone of articular cartilage. There was a marked increase in α1-positive chondrocytes in the superficial and upper mid-zones in OA-affected areas in joints from old WT mice. The α1-KO mice showed more severe cartilage degradation, glycosaminoglycan depletion, and synovial hyperplasia as compared with the WT mice. MMP-2 and MMP-3 expression was increased in the OA-affected areas. In cartilage from α1-KO mice, the cellularity was reduced and the frequency of apoptotic cells was increased. These results suggest that the α1 integrin subunit is involved in the early remodeling process in OA cartilage.

Conclusion

Deficiency in the α1 integrin subunit is associated with an earlier deregulation of cartilage homeostasis and an accelerated, aging-dependent development of OA.

Cartilage formation, homeostasis, and remodeling are regulated in part by chondrocyte interactions with extracellular matrix components. Collagens represent the major structural components of the articular cartilage. The most abundant is type II collagen, which accounts for 30–60% of the cartilage dry weight. Other collagens in articular cartilage are types VI, IX, X, and XI (1, 2). Collagen degradation is a central event in the pathogenesis of cartilage destruction in arthritis (3). Integrins are α/β-heterodimeric receptors that bind extracellular matrix components and modulate cell differentiation, activation, proliferation, and survival (4–9). Chondrocyte integrins are also involved in transducing mechanical stress (10–13). The integrins α1, α2, α10, and α11 contain a unique collagen-binding domain and pair only with the β1 subunit (14–19). Of those integrins, α1β1 and α2β1 are considered major collagen receptors (20, 21). These integrins can also bind to laminin (α1β1 and α2β1) and tenascin (α2β1). Differences in collagen-binding specificity have also been observed: the α1β1 integrin prefers type IV collagen to type I collagen (8, 22). The α1 subunit can bind and transduce signals from type II collagen, although adhesion to type II collagen that is of structural importance is mediated by α2β1 (23, 24). Integrin α1β1 also binds to type VI collagen (25–27) and cartilage matrix protein (28). The collagen-binding integrins expressed on articular chondrocytes are α1β1, α2β1, and α10β1 (29–34).

While several studies have addressed the role of integrins in chondrocyte function in vitro, there is no direct analysis of integrins in cartilage in vivo. Mice with targeted disruption of genes that encode integrins provide models that can be used to study the role of these receptors in cartilage homeostasis in vivo. Mice deficient in α1 integrin (α1-KO) develop normally (35). The dermis of α1-KO animals is characterized by hypocellularity and an increased rate of collagen synthesis. Embryonic fibroblasts from α1-KO mice do not spread, migrate, or proliferate when plated on type IV collagen or laminin. Interestingly, embryonic fibroblasts from α1-KO mice die when plated on collagen in the absence of serum, while cells derived from normal mice survive (19). Fibroblasts from α1-KO mice also fail to activate the mitogen-activated protein (MAP) kinase pathway in response to collagen and show impaired regulation of collagen expression and an increase in collagenase synthesis compared with wild-type (WT) mice.

In the present study, we analyzed the impact of α1-KO deficiency on the maintenance of articular cartilage integrity. We also determined cellular changes and extracellular matrix changes associated with this gene deletion.

MATERIALS AND METHODS

Mice.

The α1-KO mice were generated on a BALB/c background as described previously (35). Breeding pairs of α1-KO mice were kindly provided by Dr. Humphrey Gardner (The Scripps Research Institute, La Jolla, CA). Mutant gene transmission was confirmed by genomic Southern blotting of mouse tail DNA. Only homozygous α1-KO mice were used in the experiments. Age- and sex-matched WT mice of the same parental BALB/c lineage were included in each experiment as controls.

During breeding and experiments, mice were housed in sterilized microbarrier units under germ-free conditions. Mice received autoclaved chow and acidified water ad libitum. Experiments were performed when the mice were ages 4–15 months.

Histologic assessment.

Mice were euthanized at specified time points. Knee joints from both hind legs were harvested, fixed in 10% zinc-buffered formalin (Z-Fix; Anatech, Battle Creek, MI) for 2–3 days, and decalcified in decalcifier (TBD-2; Shandon, Pittsburgh, PA) for 24 hours. For each animal, the right knee was sectioned coronally, the left knee was sectioned sagittally, and both knees were embedded in paraffin. Serial slices (3 μm each) were cut, deparaffinized, stained with Safranin O–fast green, and examined for histopathologic changes.

Grading of histopathologic changes.

Histopathologic changes in the synovium were graded 0–3, where 0 = normal, 1 = thickening of the synovial lining cells, with hypertrophy, 2 = villous changes of the synovium, and 3 = villous changes of the synovium and proliferation of the synovial sublining cells.

Glycosaminoglycan depletion from the articular cartilage was graded on Safranin O–stained sections using a linear scale of 0–3, where 0 = normal, 1 = reduced staining in the superficial zone, 2 = reduced staining in the superficial and upper mid-zones, and 3 = complete loss of staining in the noncalcified layer of articular cartilage.

Cartilage degradation was scored on a scale of 0–4, where 0 = normal, 1 = fibrillation in the superficial layer of cartilage but no loss of cartilage, 2 = fissuring and cracking of the matrix and significant loss of tissue, 3 = calcified cartilage exposed, forming the articular surface at the level of the tidemark, and 4 = deep lesion extending into the epiphyseal bone and at least two-thirds of the articular surface consisting of exposed bone.

Tissue preparation for immunohistochemistry.

Selected sections from the knee joints were deparaffinized in 3 changes of PRO-PAR clearant (Anatech) and rehydrated in graded ethanol and water. For horseradish peroxidase (HRP)–conjugated antibody, endogenous peroxidase was blocked by incubating the sections with 3% hydrogen peroxide for 5 minutes at room temperature. Endogenous biotin- or avidin-binding sites were blocked by sequential incubation for 15 minutes with avidin and biotin (Vector, Burlingame, CA).

Nonspecific staining was blocked by incubating the sections with 10% normal serum or bovine serum albumin (BSA) in phosphate buffered saline (PBS). Sections were digested in 2 mg/ml of hyaluronidase for 30 minutes and permeabilized in 0.2% Triton X-100/PBS for 5 minutes at room temperature.

Immunohistochemistry.

After blocking with 10% normal serum or BSA for 30 minutes, the sections were incubated with primary antibody against the intracellular part of α1 integrin or with antibody to matrix metalloproteinase 2 (MMP-2) or MMP-3 (Santa Cruz, Santa Cruz, CA) at 2 μg/ml for 1–2 hours at room temperature or overnight at 4°C. After washing the sections 3 times for 5 minutes in PBS, a second blocking was performed for 10 minutes. The sections were then incubated for 30 minutes with diluted biotinylated secondary antibody. The slides were washed 3 times in PBS and incubated for 30 minutes with Vectastain ABC-AP reagent (Vector), or the peroxidase-based Elite ABC system. The slides were washed, and the sections incubated for 4–20 minutes with alkaline phosphatase substrate solution or with 3,3′-diaminobenzidine (DAB) substrate. Slides were rinsed with water, counterstained with diluted hematoxylin or methyl green, rehydrated in 3 changes of 1-butanol and 3 changes of PRO-PAR clearant, and mounted with Refrax mounting medium (Anatech).

Detection of poly(ADP-ribose) polymerase (PARP) cleavage.

Polyclonal antibody specific for the p85 fragment of PARP (Promega, Madison, WI) was used. Sections were microwaved in PBS–citrate buffer, pH 6.0, then digested with hyaluronidase, washed, and blocked in blocking buffer (PBS containing 0.1% Tween 20 and 10% normal serum) for 30 minutes. Primary antibody was applied at a dilution 1:100 and incubated overnight at 4°C. The negative control was rabbit IgG (1 μg/ml). The following day, the sections were washed, blocked with 3% hydrogen peroxide for 5 minutes, washed again, and incubated with diluted secondary antibody (HRP-conjugated anti-rabbit IgG) for 1 hour. Slides were washed, and sections were incubated for 4–10 minutes in DAB substrate. The slides were then rinsed in tap water, counterstained with diluted hematoxylin or methyl green, rehydrated in 3 changes of 1-butanol and 3 changes of PRO-PAR clearant, and mounted with Refrax mounting medium.

Cartilage cellularity.

To determine whether the OA changes in α1-KO mice were associated with loss of chondrocytes, cartilage cellularity was quantified by counting the chondrocytes in a microscopic field. In normal cartilage from 9-month-old mice, 4 pictures were taken under 400× magnification, representing the entire patellofemoral joint and the medial tibial plateau. The total number of chondrocytes was evaluated in each picture.

Statistical analysis.

Statistical comparison between groups was performed with Student's t-test. Histologic scores were analyzed using Mann-Whitney rank sum test. P values less than 0.05 were considered significant.

RESULTS

Expression of α1 integrin in cartilage.

In normal joints of 4-month-old WT mice, α1 integrin was detected only in cells in the deep zone of the articular cartilage (Figure 1A). Joints from 9-month-old or 15-month-old mice with mild OA-like changes also showed strong expression of α1 in the superficial and upper mid-zones of nonlesional, normal-appearing cartilage (Figures 1B and C). These findings indicate a profound up-regulation of the number of α1 integrin–positive cells in joints that are undergoing matrix remodeling associated with early OA changes (Table 1).

Figure 1.

Expression of α1 integrin in wild-type (WT) BALB/c mice. Knee joint sections were stained for the α1 subunit using goat polyclonal antibody to the intracellular part of α1 integrin. Shown are sections of articular cartilage from A, a 4-month-old mouse without osteoarthritis (OA)–like changes, B, a 9-month-old mouse with OA-like changes, C, a 15-month-old mouse with mild OA-like changes, D, a 4-month-old mouse with mild OA-like changes, E, a 9-month-old mouse, showing a section of growth plate, and F, a 4-month-old mouse, stained with normal goat IgG as a negative control. (Original magnification × 100.)

Table 1. Percentage of α1 integrin–positive cells in articular cartilage from wild-type BALB/c mice
Age of miceCartilage zone
SuperficialMidDeep
4 months16.7 ± 10.28.6 ± 4.350.3 ± 5.2
9 months66.7 ± 0.861.0 ± 8.515.3 ± 2.8
15 months75.7 ± 4.551.3 ± 6.116.0 ± 2.5

The α1 integrin subunit was also detected in growth plates from 9-month-old mice. However, it was present only in the hypertrophic zone adjacent to the zone of cell death and vascular invasion (Figure 1E).

Absence of the α1 subunit in the articular cartilage and growth plate of α1-KO mice was confirmed by immunohistochemistry (Figure 2). The tissue section from a 9-month-old α1-KO mouse shown in Figure 2B demonstrates normal growth plate architecture. Normal growth plate architecture was also observed in all other age groups of α1-KO mice (results not shown).

Figure 2.

Lack of immunoreactivity in α1 integrin–knockout (α1-KO) mice. Shown are sections of A, articular cartilage and B, growth plate from the knee joint of a 9-month-old α1-KO mouse. Sections were stained with antibody against the α1 subunit. (Original magnification × 100.)

Increased severity of OA changes in α1-KO mice.

Joints from α1-KO mice did not show developmental abnormalities. All joint structures were developed as in the WT mice. At age 9 months, the joints of the WT mice were normal or showed only minimal changes (Figures 3A and B). In contrast, knee joints from α1-KO mice (Figures 3C and D) showed loss of cartilage in some areas, with exposure of subchondral bone. Glycosaminoglycan in the noncalcified articular cartilage was almost completely lost. This was associated with marked synovial hyperplasia and attachment to and invasion of the synovium into the articular cartilage surface (Figure 4). At the age of 9 months, the scores for synovial changes, glycosaminoglycan depletion, and cartilage degeneration were significantly higher in the α1-KO mice than in the WT mice (Figures 5A–C). In addition, at 9 months of age, no osteophytes were seen in any of the 9 WT mice examined, but were present in 3 of the 11 α1-KO mice (results not shown).

Figure 3.

Histopathologic features of the knee joints of 9-month-old wild-type (WT) and α1 integrin–knockout (α1-KO) mice. A, Coronal and B, sagittal sections of knee joints from WT mice, showing a smooth joint surface and normal Safranin O staining pattern. There are no osteoarthritic (OA) changes in these WT mice. C, Coronal section of a knee joint from an α1-KO mouse, showing grade 4 OA changes in the tibia. Glycosaminoglycan is almost completely depleted from the noncalcified part of the articular cartilage. D, Sagittal section of a knee joint from an α1-KO mouse, showing a cartilage defect in the anterior aspect of the joint surface. Subchondral bone is almost exposed. (Safranin O stained; original magnification × 40 in A and B; × 100 in C and D.)

Figure 4.

Synovial hyperplasia in 9-month-old α1 integrin–knockout (α1-KO) mice. Sections of synovium from the insertion of the medial meniscus from 9-month-old α1-KO and wild-type (WT) mice were stained with Safranin O. Shown are sections from A, a WT mouse, demonstrating mild hyperplasia, B, an α1-KO mouse, demonstrating grade 2 changes, and C, an α1-KO mouse, demonstrating grade 3 changes. Synovial hyperplasia and proliferation onto the cartilage surface are prominent. (Original magnification × 100.)

Figure 5.

Histologic scores in 9-month-old α1 integrin–knockout (α1-KO) and wild-type (WT) mice. Histopathologic changes in A, the synovium and C, the cartilage were graded as described in Materials and Methods. B, Loss of glycosaminoglycans from the articular cartilage was determined by Safranin O staining. Values are the mean ± SEM of 6 mice. There were significant differences between WT and α1-KO mice for each variable.

To determine the temporal pattern of OA changes in the knee joints of α1-KO mice, animals were evaluated at ages 4, 7, 9, 10, 12, and 15 months. At 4 months of age, cartilage degeneration was more severe in α1-KO mice (Figure 6A). This difference was statistically significant at 7, 9, and 10 months. The severity of changes was similar in the 2 strains at 12 and 15 months. Glycosaminoglycan depletion was also enhanced in α1-KO mice, as evidenced by reduced Safranin O staining. The difference between the 2 strains was significant only at age 9 months (Figure 6B). Histopathologic changes of the synovium were significantly more severe in the α1-KO animals at ages 9 and 10 months (Figure 6C). These results suggest that the α1-KO mice spontaneously develop more severe OA-like changes at a younger age than do WT mice.

Figure 6.

Kinetics of the histopathologic changes in α1 integrin–knockout (α1-KO) and wild-type (WT) mice. Histologic scores in both knee joints from α1-KO (solid line) and WT (broken line) mice were assessed at the ages indicated (n = 7 mice per group). Significant differences between the groups were evident with respect to A, cartilage degeneration, B, glycosaminoglycan loss (by Safranin O staining), and C, synovial changes. = P < 0.05; ∗∗ = P < 0.01.

Cartilage matrix degradation in α1-KO mice, as evidenced by MMP expression.

The results of Safranin O staining suggested an early and profound loss of glycosaminoglycans prior to the development of cartilage fibrillations (Figure 3). Immunohistochemical analysis showed increased MMP-2 expression in cartilage and synovium from the knee joints of 9-month-old α1-KO mice (Figure 7), while there was no detectable MMP-2 in the normal joints of 9-month-old WT mice. MMP-3 was also increased in OA-affected cartilage, with a pattern similar to that of MMP-2 (results not shown).

Figure 7.

Matrix metalloproteinase 2 (MMP-2) expression in 9-month-old α1 integrin–knockout (α1-KO) and wild-type (WT) mice. A, Normal cartilage from a WT mouse stained with antibody to MMP-2. Chondrocytes in normal cartilage are negative for MMP-2. B, Knee joint from an α1-KO mouse stained with antibody to MMP-2. Many positive cells are detected in the osteoarthritic cartilage as well as in the synovium. C, Normal cartilage from a WT mouse stained with negative control antibody (normal rabbit IgG). (Original magnification × 100.)

Cartilage cellularity and cell death.

Cartilage cellularity was quantified to determine whether the OA changes in α1-KO mice were associated with a loss of chondrocytes (Figure 8). The mean number of chondrocytes per defined cartilage area was significantly lower in α1-KO mice as compared with WT mice (92.8 cells/view versus 112.8 cells/view; P = 0.031).

Figure 8.

Cartilage cellularity in 9-month-old α1 integrin–knockout (α1-KO) and wild-type (WT) mice, quantified by counting the chondrocytes in a microscopic field. In macroscopically normal-appearing cartilage from WT or α1-KO mice, 4 pictures were taken under 400× magnification, representing the entire patellofemoral joint and the medial tibial plateau. The total number of chondrocytes was evaluated in each picture.

To characterize the occurrence and type of cell death, knee joints were subjected to immunohistochemical analysis for the p85 fragment of PARP. In normal cartilage from 9-month-old WT mice, the p85 fragment was not detected. Cartilage from 9-month-old α1-KO mice contained many chondrocytes positive for PARP p85, predominantly in the OA-affected areas, with lower frequencies in partially degraded or macroscopically normal-appearing areas (Figure 9).

Figure 9.

Poly(ADP-ribose) polymerase (PARP) cleavage as a marker of apoptosis in 9-month-old α1 integrin–knockout (α1-KO) and wild-type (WT) mice. To detect apoptotic chondrocytes, polyclonal antibody specific for the p85 fragment of PARP was used. A, Normal cartilage from a 9-month-old WT mouse stained with antibody to PARP p85. No signals for PARP p85 were found. B, Cartilage from an α1-KO mouse stained with antibody to PARP p85. Many chondrocytes in the osteoarthritic (OA) area are positive for PARP p85. C, Margin of OA lesion shown in B. D, Negative control (normal rabbit IgG) cartilage from an α1-KO mouse. (Original magnification × 100.)

DISCUSSION

Osteoarthritis is a common age-related joint disorder. Key pathogenetic features of OA are degradation of cartilage extracellular matrix and a reduction in cartilage cellularity. Chondrocyte function is controlled in part by cellular contacts with the pericellular matrix. Since this interaction is disturbed in OA as a consequence of extracellular matrix degradation, this may cause secondary changes in chondrocyte function and viability. The present study focused on the role of the α1 integrin in cell–matrix interactions in cartilage.

The α1β1 integrin is a major collagen receptor (22). It binds not only to type II collagen, the most abundant protein in cartilage, but also to type VI collagen, which is enriched in the pericellular matrix. The 15–amino acid cytoplasmic tail of the α1 subunit is the shortest of all integrins. Upon ligand binding, caveolin 1 links the α1 subunit to the tyrosine kinase Fyn. This results in Shc activation and recruitment of Grb2 (36). These events are necessary for activation of the Ras–MAP kinase signaling pathway, a feature that integrin α1β1 shares with only 3 other integrins (19). Adhesion mediated by Shc-linked integrins promotes cell survival and progression through the G1 phase of the cell cycle in response to mitogenic growth factors, whereas adhesion mediated by other integrins results in exit from the cell cycle and, in certain cases, cell death (37, 38). These in vitro studies suggest that α1β1 integrin may regulate cell survival and growth in response to extracellular matrix components.

In the present study, we first established the distribution of the α1 integrin. The α1 integrin was detected in growth plate only in the hypertrophic zone adjacent to the zone of cell death and vascular invasion. It is of interest to note that the structural organization and size of the growth plates in α1-KO mice were normal; these mice also had normal development of the skeleton and of all synovial joint structures. This indicates that the α1 integrin subunit may not exert an essential or exclusive function in development, partly because other integrins can recognize the same extracellular matrix components. In normal joints, α1 integrin was expressed only on a subset of chondrocytes in the deep zone of the articular cartilage. In joints from old WT mice with OA-like changes, α1 integrin was expressed in the superficial and upper mid-zones of nonlesional, normal-appearing cartilage. These findings indicate a profound up-regulation of α1 integrin in joints that are undergoing matrix remodeling. This is consistent with an up-regulation of the α1 subunit in the process of chondrogenesis in the periosteum (39). Expression patterns of α1 integrin suggest a role in specific stages of chondrocyte differentiation and matrix remodeling.

In α1-KO mice, cartilage degradation developed at a younger age. This was detectable at the age of 4 months and was significantly different from that in WT mice at ages 7–12 months. At 12 and 15 months of age, the severity of changes in the WT BALB/c mice approached that of the α1-KO mice. Increased severity of cartilage degradation is associated with more severe synovial inflammation and with the formation of osteophytes. The accelerated development of cartilage degradation in the α1-KO mice was associated with increased MMP-2 expression, reduced cellularity, and increased frequency of apoptosis. Cartilage matrix degradation and chondrocyte apoptosis are linked in OA (40–42), and chondrocyte death is correlated with age and disease severity (43–46). This linkage is supported by the present study. Areas of cartilage degeneration and even normal-appearing cartilage in OA-affected joints showed increased MMP-2 and MMP-3 expression and higher frequencies of PARP p85–positive apoptotic chondrocytes.

Apoptosis in the α1-KO mice could be partly due to a deficiency in the survival-promoting effect mediated by this integrin. Chondroprotective effects of extracellular matrix proteins have been documented (7). These were abrogated by anti–β1 integrin antibodies, indicating a role for β1 integrins in the protection of chondrocytes from cell death.

In conclusion, α1-KO mice show an accelerated development of OA-like lesions. This is characterized by more severe glycosaminoglycan loss, synovial proliferation, and the presence of osteophytes. Cartilage from α1-KO mice contains increased numbers of apoptotic chondrocytes and cells expressing MMP-2 and MMP-3. The earlier development of OA in the α1-KO mice and the up-regulation of α1 integrin in OA-affected joints suggest that this integrin supports cartilage remodeling. Deletion of α1 integrin compromises this response and leads to an earlier and more severe OA. This finding is consistent with the findings of a recent study demonstrating reduced callus formation and diminished collagen synthesis in α1-KO mice (47).

Ancillary