Chondrocytes, the sole cell type in articular cartilage, maintain the extracellular matrix (ECM) through a homeostatic balance of anabolic and catabolic activities that are influenced by genetic factors, soluble mediators, and biophysical factors such as mechanical stress. Chondrocytes are encapsulated by a narrow tissue region termed the “pericellular matrix” (PCM), which in normal cartilage is defined by the exclusive presence of type VI collagen. Because the PCM completely surrounds each cell, it has been hypothesized that it serves as a filter or transducer for biochemical and/or biomechanical signals from the cartilage ECM. The present study was undertaken to investigate whether lack of type VI collagen may affect the development and biomechanical function of the PCM and alter the mechanical environment of chondrocytes during joint loading.
Col6a1−/− mice, which lack type VI collagen in their organs, were generated for use in these studies. At ages 1, 3, 6, and 11 months, bone mineral density (BMD) was measured, and osteoarthritic (OA) and developmental changes in the femoral head were evaluated histomorphometrically. Mechanical properties of articular cartilage from the hip joints of 1-month-old Col6a1−/−, Col6a1+/−, and Col6a1+/+ mice were assessed using an electromechanical test system, and mechanical properties of the PCM were measured using the micropipette aspiration technique.
In Col6a1−/− and Col6a1+/− mice the PCM was structurally intact, but exhibited significantly reduced mechanical properties as compared with wild-type controls. With age, Col6a1−/− mice showed accelerated development of OA joint degeneration, as well as other musculoskeletal abnormalities such as delayed secondary ossification and reduced BMD.
These findings suggest that type VI collagen has an important role in regulating the physiology of the synovial joint and provide indirect evidence that alterations in the mechanical environment of chondrocytes, due to either loss of PCM properties or Col6a1−/−-derived joint laxity, can lead to progression of OA.
Articular cartilage is the tissue that lines the surfaces of diarthrodial joints and serves as the resilient, low-friction, load-bearing material for joint motion. A sparse population of cells—chondrocytes—maintains the extracellular matrix (ECM) of this tissue through a balance of anabolic and catabolic activities. The micromechanical environment of chondrocytes, in conjunction with biochemical factors (e.g., growth factors, cytokines) and genetic factors, plays an important role in cartilage homeostasis and, as a consequence, the health of the joint (1–3). Chondrocytes in articular cartilage are enclosed by a narrow region of tissue termed the “pericellular matrix” (PCM), which, together with the enclosed chondrocyte, has been termed the “chondron” (4–7). The PCM is characterized primarily as being the exclusive location of type VI collagen in normal cartilage, but proteoglycans, fibronectin, and types II and IX collagen are also present in high concentrations in the PCM (4, 8).
The functional role of the PCM in articular cartilage is still unknown, although the fact that it completely surrounds the cell suggests that it regulates the biomechanical, biophysical, and biochemical signals that the chondrocyte perceives (9). For example, interactions between cell surface receptors and the ECM significantly influence matrix metabolism, gene expression, and response to growth factors (10–12). Furthermore, cytokines and growth factors that interact with the chondrocyte surface traverse the pericellular environment, where they may be retained and modified (13, 14). From a biomechanical standpoint, there has been considerable speculation that the PCM plays a critical role in either protecting the cells or serving as a “filter” or transducer of physical signals in the ECM (4, 6, 9, 15, 16), potentially through an interaction of type VI collagen with integrins or other cell surface receptors (17–20). Indirect evidence in support of these hypotheses is provided by experimental data showing that a newly formed PCM augments the cellular metabolic response to biomechanical loading (21).
Type VI collagen serves as the defining boundary of the PCM in articular cartilage, but it is also found in the ECM in many connective tissues (22). It has a characteristic beaded filamentous structure of tetrameric units consisting of 3 different α-chains, α1(VI), α2(VI), and α3(VI). Type VI collagen has high affinity with numerous ECM components (i.e., biglycan, decorin, hyaluronan, fibronectin, perlecan, and heparin) as well as with the cell membrane (21, 23–26). Thus, it has been hypothesized that type VI collagen plays important roles in mediating cell–matrix interactions as well as intermolecular interactions in various tissues and cell cultures (27–31). In articular cartilage, type VI collagen forms a network that anchors the chondrocyte to the PCM (32–35) through its interaction with hyaluronan (21, 36), decorin (24), and fibronectin (37).
The goal of this study was to examine the hypothesis that lack of type VI collagen alters the biomechanical properties of the PCM and ECM of articular cartilage. Type VI collagen–deficient mice were generated by targeted disruption of the Col6a1 gene (38). Histologic analysis and dual x-ray absorptiometry (DXA) were used to examine differences in skeletal development, bone mineral density (BMD), and progression of osteoarthritic (OA) joint degeneration in wild-type and type VI collagen–deficient mice. In addition, micromechanical testing was performed on the articular cartilage and on isolated chondrons, using microindentation and micropipette aspiration techniques, respectively, to determine the role of type VI collagen in the elastic properties of the articular cartilage ECM and PCM.
MATERIALS AND METHODS
Type VI collagen–knockout mice.
All procedures were approved by the Duke University Institutional Animal Care and Use Committee. Type VI collagen–knockout mice were generated on a CD1 genetic background by targeted disruption of the Col6a1 gene, which is responsible for the production of the α1(VI)-chain (38). The elimination of the α1(VI)-chain resulted in the absence of triple-helical type VI collagen molecules in the ECM (38). Mice were killed at age 1, 3, 6, or 11 months.
Type VI collagen immunostaining was performed using a polyclonal anti–type VI collagen antibody raised against a peptide mapping near the amino-terminus of the murine α1(VI)-chain (Santa Cruz Biotechnology, Santa Cruz, CA). Cryostat sections from the coronal plane were obtained from decalcified femoral heads, using standard histologic methods.
One-month-old mice were killed, skinned, and eviscerated. Alcian blue and alizarin red were used to stain the cartilage and bone, respectively, by standard skeletal staining techniques.
Histologic analysis and morphometric grading.
Fixed cryostat sections were stained with toluidine blue and hematoxylin and eosin, using standard histologic techniques. Osteoarthritic and developmental changes in the femoral head of 60 mice (ages 1, 3, 6, and 11 months) were assessed using a quantitative histomorphometric grading system (39, 40). OA was graded based on the sum score of surface fibrillation (0–4), intensity of toluidine blue staining (0–3), and fibrocartilage presence (0–2) (maximum possible total sum score 9) and was reported as a percentage of the 0–9 scale, with 0% corresponding to no histologic signs of OA and 100% corresponding to the most severe changes (grade 9). Specimens were classified as non-OA (score 0–1), mild OA (>1–5), or severe OA (>5–9).
Developmental changes were assessed histologically based on the presence of the growth plate and the extension of the secondary ossification center (40), which was graded from 0 to 5 (0 corresponding to findings in normal 1-month-old mice, and 5 corresponding to findings in normal 11-month-old mice). The percentage of ossified area relative to the total area was graded as follows: 0 = cartilage with growth plate and no area of secondary ossification present, 1 = area of secondary ossification <10%, 2 = area of secondary ossification >10% and <50%, 3 = area of secondary ossification >50% and <90%, 4 = area of secondary ossification >90%, and 5 = no growth plate present. Secondary ossification was reported as a percentage of the 0–5 scale, with 0% corresponding to grade 0 (no secondary ossification) and 100% corresponding to the fully developed femoral head (grade 5).
BMD was measured by DXA (PIXImus; Lunar, Madison, WI) (41). The mice were weighed and placed in the DXA unit in a supine position, and the whole body except the skull was measured. A total of 82 mice were analyzed, at age 1, 3, 6, or 11 months.
Mechanical testing of articular cartilage.
A total of 18 right hip joints from 1-month-old mice (7 Col6a1−/−, 7 Col6a1+/+, 4 Col6a1+/−) were tested in indentation using an electromechanical test system (ELF 3200; EnduraTEC, Minnetonka, MN) instrumented with a low-capacity load cell (250 gm; Sensotec, Columbus, OH) and extensometer (1 mm; Epsilon, Jackson, WY) (42). Plane-ended microindenters were machined from glass fibers (110 μm diameter; Thorlabs, Newton, NJ). A dual-angle camera system was used to optically align the indenter tip perpendicular to the cartilage surface. After applying a tare load of 0.3 gm force and allowing it to equilibrate, 4 consecutive indentation displacements (5 μm per step with a ramping speed of 1 μm per second) were applied to the cartilage surface and allowed to equilibrate for 200 seconds per step. The time, reaction force, and displacement data were collected at 1 Hz throughout the test. The equilibrium force versus displacement curve was obtained from the linear region of the curve. After mechanical testing, the thickness of cartilage from the tissue surface to the calcified cartilage was measured at a site adjacent to the test site, using routine histologic procedures (5-μm sections labeled with Safranin O and fast green). The Young's modulus of mouse cartilage was calculated using an elastic indentation model (43) with an assumed Poisson's ratio of 0.25 (42).
Mechanical testing of the PCM.
Chondrons were mechanically isolated from the femoral articular cartilage of 1-month-old mice (93 chondrons from 26 mice) with a custom-built “microaspirator,” which applies suction pressure to the cartilage surface with a modified syringe, as described previously (44). The micropipette aspiration technique (45–47) was used to measure the mechanical properties of the PCM, as described previously (44, 48, 49). With this technique, the surface of the PCM is aspirated into a glass micropipette (12 μm diameter) by the application of a series of controlled pressures up to 18 kPa, and the ensuing equilibrated aspiration length is measured using video microscopy. The Young's modulus of the PCM was determined using a theoretical model that represents the chondron as an elastic, compressible layer (i.e., PCM) overlying an elastic half-space (i.e., chondrocyte) (44).
Statistical analysis was performed using a multifactorial analysis of variance setup in Statistica (StatSoft; Tulsa, OK). The only categorical predictors considered were age (1, 3, 6, 9, or 11 months) and genotype (+/+, +/−, or −/−). We assumed that these variables would be able to predict 4 dependent measurements, i.e., weight, OA score, BMD, and ossification score. Full factorial design revealed that both age, genotype, and the age × genotype effects were significant contributors; thus, post hoc comparison was performed using Fisher's least significant difference method. Age effects were significant for all 4 measurements, as expected. Genotype effects between Col6a1+/+ and Col6a1−/− mice were also significant for all 4 measurements. Genotype effects between Col6a1+/+, Col6a1+/−, and Col6a1−/− mice were significant only for OA and ossification scores. Age × genotype effects within the same age groups were not significant for OA. For ossification scores, a significant difference between Col6a1+/+ and Col6a1−/− mice was observed at 3 months only. For BMD, a significant difference between Col6a1+/+ and Col6a1−/− mice was observed at all ages except 1 month.
Mice lacking type VI collagen exhibited no apparent abnormalities, and all animals survived the full 11 months of the study unless killed earlier per the protocol. Articular cartilage of wild-type and heterozygous mice showed extensive pericellular labeling for type VI collagen, whereas type VI collagen was absent in the knockout mice (Figure 1). Intense labeling for type VI collagen was also observed in the growth plate of 1-month-old wild-type mice (Figure 1A).
Staining of skeletal bone and cartilage from 1-month-old mice (Figure 2) indicated that Col6a1−/− mice were smaller and exhibited a slower ossification process in the upper extremities (Figures 2C and D) and lower extremities (Figures 2E and F) than their wild-type counterparts. The smaller size was also consistent with a trend toward lower body weight in 1-month-old Col6a1−/− mice compared with Col6a1+/+ mice (mean ± SD 16.5 ± 2.4 gm versus 18.3 ± 1.99 gm; P = 0.11 by 2-tailed t-test).
To better evaluate the developmental process, we measured the secondary ossification process in the femoral head (Figures 3A–C). Col6a1−/− mice showed significantly delayed ossification at 3 months, compared with Col6a1+/+ mice (mean ± SD grade 2.2 ± 1.8 versus 4.1 ± 0.2; P < 0.02) (Figure 3D). Among 3-month-old animals, only 1 of 6 Col6a1−/− mice had an ossification grade of > 4, whereas the ossification grade was ≥4 in 3 of 3 Col6a1+/+ mice. In all mice, the ossification process was complete after 6 months.
Semiquantitative histologic analysis of cartilage degeneration revealed significant age-dependent osteoarthritic changes in the Col6a1−/− mice. OA changes depended on age (P < 0.001) and genotype (P < 0.05) (Figure 4). Among 6–11-month-old animals, only 2 of 12 Col6a1+/+ mice, 3 of 7 Col6a1+/− mice, and 11 of 16 Col6a1−/− mice scored >1 and were thus characterized as having OA (either mild or severe; see Materials and Methods). However, scores were >5, i.e., representing severe OA, in 0 of 12, 0 of 7, and 2 of 16 of the Col6a1+/+, Col6a1+/−, and Col6a1−/− mice, respectively.
BMD, as measured by DXA, was significantly higher in wild-type mice than their type VI collagen–knockout counterparts at ages 3–6 months (P < 0.001). These differences were no longer evident at 11 months (Figure 5).
Mechanical properties of articular cartilage and PCM.
The PCM exhibited linear elastic behavior, and the Young's modulus of the PCM of chondrons isolated from Col6a1+/+ mice was significantly higher than that of the PCM of chondrons from Col6a1+/− mice. This was further reduced in the knockout Col6a1−/− mice (Figure 6C). Microindentation tests revealed no significant differences between Col6a1+/+, Col6a1+/−, and Col6a1−/− mice in terms of mechanical properties of femoral head articular cartilage (Figure 6D).
This report presents new evidence of significant musculoskeletal changes in Col6a1−/− mice. Primarily, our findings show that mice lacking type VI collagen exhibit accelerated development of hip osteoarthritis, as well as a delayed secondary ossification process and reduced BMD. Lack of type VI collagen resulted in a loss of the stiffness of the articular cartilage PCM (decreased modulus), prior to the occurrence of any detectable histologic changes. However, no differences in ECM properties were observed. These findings provide indirect evidence of a role of type VI collagen in regulating the physiology of the articular cartilage chondrocyte, potentially via alterations in the biologic and mechanical environment of chondrocytes due to changes in biomechanical properties of the PCM or increased joint laxity associated with a type VI collagen deficiency.
The mechanical environment of the chondrocytes is one of several environmental factors that influence the normal balance between the synthesis and breakdown of articular cartilage, and it is an important participant in the etiopathogenesis of OA (1–3, 50, 51). Thus, changes in the mechanical interactions between the cell and the ECM may have a significant influence on the regulatory response of the chondrocyte.
A biomechanical function of the PCM has long been hypothesized (5, 6, 9), and there is growing evidence from both theoretical modeling and experimental studies that the PCM plays a significant role in regulating the biomechanical signals perceived by the chondrocyte (16, 52, 53). The mechanical properties of the PCM are significantly altered in OA, exhibiting reduced stiffness and increased fluid permeability (44, 49). These changes occur throughout the thickness of the articular cartilage, affecting the properties of the PCM in the superficial and middle/deep zones in a similar manner (48). The PCM appears to function by providing a relatively uniform cellular microenvironment despite a great lack of homogeneity in local tissue strain (16, 54). Thus, a compromised PCM could significantly affect the mechanical environment of the chondrocytes in articular cartilage, leading to increased strain at the cellular level (53), which may affect catabolic responses at the level of single cells (55). In other tissues such as bone, however, the pericellular region (i.e., the glycocalyx) can serve as a strain amplifier by coupling fluid drag forces to the actin cytoskeleton within the processes of osteocytes (56–58). In the present study, Col6a1−/− mice showed significantly reduced PCM stiffness at 1 month of age, prior to the appearance of any histologic or biomechanical changes in the overall articular cartilage. With age, these mice exhibited accelerated development of OA. These findings provide indirect evidence that early alterations in the mechanical properties of the PCM are associated with the progression of OA.
In normal articular cartilage, type VI collagen is present exclusively in the PCM, and it has been characterized as being a discrete marker of chondron anatomy (34). For this reason, we hypothesized that type VI collagen is necessary for providing the structural integrity and mechanical properties of the PCM. Contrary to our hypotheses, though, Col6a1−/− mice exhibited intact chondrons that could be isolated despite the lack of type VI collagen. This finding suggests that proteins other than type VI collagen provide some of the structural integrity of cartilage PCM. Nonetheless, the Young's modulus (stiffness) of the PCM in Col6a1−/− mice was dramatically decreased (approximately one-third that of the PCM in wild-type controls), illustrating the important role of type VI collagen in the properties of the PCM.
An important issue that must be considered is the link between type VI collagen deficiency and changes in muscle physiology displayed by Col6a1−/− mice (38). Such a link has also been observed in humans, with studies showing that mutations of type VI collagen genes play a causal role in two inherited disorders of muscle: Bethlem myopathy and Ullrich congenital muscular dystrophy (UCMD) (59, 60). It is possible that some features of UCMD, particularly joint laxity or predisposition to hip dislocation, may also contribute to the accelerated hip degeneration observed in Col6a1−/− mice. Since joint laxity and mechanical alterations of the PCM are heritably coupled in Col6a1−/− mice and both lead to an altered mechanical environment in chondrocytes, both factors could contribute to the development of OA.
While the present findings clearly demonstrate an association between Col6a1 deficiency and OA, presumably via mechanical alterations caused by joint laxity or altered PCM properties, further studies aimed at developing and characterizing conditional or tissue-specific knockout animals may be needed to fully understand the mechanisms by which Col6a1 deficiency leads to OA. Nonetheless, our results are consistent with the hypothesized role of type VI collagen as an integrating molecule in the structure of cells and tissues; down-regulation of type VI collagen is associated with tissue laxity and wasting (e.g., Bethlem myopathy, UCMD, joint hyperlaxity), whereas up-regulation of type VI collagen results in increased fibrosis and tissue stiffness (e.g., bullous keratopathy, scleroderma) (38, 61–68).
Our studies revealed no gross morphologic differences between wild-type and type VI collagen–knockout mouse chondrons, other than reduced skeletal size of Col6a1−/− mice. Skeletal changes were apparent as a retardation of the developmental process until 11 months of age. During development, histogenesis of long bones occurs via endochondral ossification of cartilage tissue. During this process, chondrocytes in the epiphyseal plate differentiate into mature hypertrophic cells and finally are eliminated from the growth plate (69). The hypertrophic cell lacunae are invaded by vessels carrying mesenchymal and osteogenic cells that differentiate into osteoclasts and synthesize a bony matrix. A similar procedure, known as secondary ossification, takes place at the end of the bone, where the formation of the bony epiphysis occurs.
The present results provide evidence of a slowing of secondary ossification changes and decreased BMD in Col6a1−/− mice. While there is no known mechanism directly linking type VI collagen deficiency with endochondral ossification, type VI collagen may provide a scaffold for osteoblasts, preosteoblasts, and chondrocytes to proceed to osteochondral ossification (70). In addition, type VI collagen has been linked to the early events of chondrocyte differentiation (71), the regulation of mesenchymal cell proliferation in vitro (72), and ECM stabilization during development (34). It has also been hypothesized that type VI collagen is important for chondrocyte proliferation and hypertrophy in cartilage. Taken together with the findings of these previous studies (34, 73, 74), our observation of the ubiquitous presence of type VI collagen in the growth plate (Figure 1A), supports the notion that type VI collagen deficiency may delay cell differentiation and proliferation, resulting in delayed development and reduced bone formation.
Interestingly, the COL6A1 gene was recently identified as the locus for ossification of the posterior longitudinal ligament of the spine (70) and has also been associated with increased systemic BMD and diffuse idiopathic skeletal hyperostosis (75). These findings provide evidence of a role of type VI collagen in diseases associated with high bone-mass, consistent with our observation of decreased BMD in Col6a1−/− mice. While it was beyond the scope of the present study to analyze the mechanisms resulting in altered BMD, these changes may also be biomechanical in origin, since type VI collagen deficiency causes muscular dystrophy (38, 63), which can lead to abnormal mechanical loading of the musculoskeletal system.
In conclusion, results of the present study, in which the effect of an abnormal mechanical environment on chondrocytes was investigated using type VI collagen–knockout mice, suggest that type VI collagen plays a major role in the mechanical properties of the PCM, and thus, in the mechanical environment of chondrocytes. Col6a1−/− mice showed accelerated development of osteoarthritis that may be “biomechanical” in nature, via either altered properties of the PCM or heritable joint laxity. In addition, our findings provide direct evidence that type VI collagen might have a significant role in the osteochondral ossification process, by modulating chondrocyte and mesenchymal cell differentiation and proliferation activities. This model may provide a valuable tool for better understanding of the way changes in the mechanical environment of chondrocytes may lead to abnormal skeletal development and development of OA.
Dr. Guilak 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. Alexopoulos, Bonaldo, Guilak.
Acquisition of data. Alexopoulos, Youn.
Analysis and interpretation of data. Alexopoulos, Youn, Bonaldo, Guilak.