Chondrocyte-Specific Loss of the Branched Actin Mediator ARP2/3 Results in Growth Plate Fusion and Proteoglycan Loss in Articular Cartilage

Abstract

Purpose: The actin cytoskeleton regulates cell shape changes that are associated with chondrogenic differentiation in the developing growth plate and with phenotypic changes to articular chondrocytes during monolayer culture. However, it has been challenging to determine the role of the actin cytoskeleton during cartilage homeostasis in vivo. The Arp2/3 complex generates branched actin and is involved in many cellular processes including cell motility in response to extra-cellular matrix cues. We hypothesized that chondrocyte-specific deletion of Arp2/3 in vivo would alter cell-matrix interactions and result in both growth plate disorganization and premature osteoarthritis (OA). Methods: C57BL/6 mice harboring flanking loxP sites around a critical exon of an Arp2/3 subunit (p34) were crossed to two chondrocyte-specific Cre driver strains. Type II collagen (Col2)-Cre resulted in constitutive loss of Arp2/3, whereas Aggrecan-CreERT2 was used to induce Arp2/3 loss by intraperitoneal injection of tamoxifen. The cellular organization of the tibial growth plate was analyzed using hematoxylin and eosin stained histological sections. Mid-coronal sections stained with Safranin-O were evaluated for disrupted articular cartilage structure and proteoglycan loss using established semiquantitative scales of 0–12 for medial and lateral sides of the tibial and femoral cartilage. Statistical analysis was performed by ANOVA and Tukey's post-hoc test with greater than 6 mice per group. Live cell imaging was used to quantify cell spreading in isolated cells after induction of Arp2/3 loss in vivo. Pure populations of growth plate and articular chondrocytes were sorted based on fluorescence in cells successfully recombined by Aggrecan-CreERT2 using a lox-stop-lox zsgreen reporter allele. Results: Mice with constitutive loss of Arp2/3 in chondrocytes showed reduced viability (5/90 pups; expected ratio 1/4). Mutant mice that did survive were smaller than littermate controls and showed disorganization of the growth plate with secondary ossification failure (Fig. 1). When Arp2/3 loss was induced at 4 months, growth plates showed partial closure at 6 months and complete fusion at 12 months (Fig. 2). The articular cartilage structure at 6 and 12 months was not affected by Arp2/3 loss (Fig. 3a). However, the proteoglycan content was significantly reduced at 12 months in mice with Arp2/3 loss as compared to other groups (Fig. 3b). Osteophytes were noted in 4 out of 8 mice with Arp2/3 loss at 12 months as compared to 1 out of 6 controls. Using an in vitro assay of cell-matrix interaction, growth plate and articular chondrocytes directly isolated from 3 week old mice with induced Arp2/3 loss showed less spread cell area on fibronectin and type I collagen coated dishes as compared to heterozygous controls (all combinations p<0.01). Conclusions: The ability of chondrocytes to sense and respond to extra-cellular matrix cues plays important roles during both development and homeostasis. Arp2/3 appears to mediate chondrocyte-matrix interactions, as cells with Arp2/3 loss were unable to spread normally on collagen and fibronectin surfaces. Growth plate chondrocytes with constitutive Arp2/3 loss failed to organize the highly ordered columns that result in longitudinal bone growth. When Arp2/3 loss was induced after skeletal maturity, growth plates progressively fused in a manner that does not typically occur in mice. This suggests that cell-matrix interactions or other Arp2/3 functions contribute to the active maintenance of the growth plate throughout adulthood. Loss of Arp2/3 in articular chondrocytes did not result in extensive erosion of cartilage matrix at 12 months, but did cause reduced staining for proteoglycans and more osteophytes. These observations suggest a role for the actin cytoskeleton in maintaining cartilage homeostasis over the long-term, although longer follow-up may be necessary to determine whether loss of Arp2/3 would result in more severe OA. This work has implications for understanding developmental phenomena such as chondrodysplasias and catch-up growth, as well as providing insight into the mechanisms that govern how cell-matrix interactions in articular cartilage may alter the risk for OA progression.Figure 2: H&E stained sections of the tibia of mice at 6 months (top) and 12 months (bottom) of age. All mice had Aggrecan-CreERT2 and received tamoxifen at 4 months of age, but controls retained one copy of functional Arp2/3. Scale bars are 100 μm.View Large Image Figure ViewerDownload Hi-res image Download (PPT)Figure 3. Safranin-O stained sections were graded for articular cartilage structure (A) and proteoglycan loss (B). A semiquantitative score of 0–12 was summed across four quadrants of the joint (high scores consistent with OA, max = 48). # indicates p<0.05 to all other groups.View Large Image Figure ViewerDownload Hi-res image Download (PPT)