Post-traumatic osteoarthritis (OA) is the most prevalent form of arthritis in young adults, and often results in long-term detrimental influence on quality of life (1). One hallmark of posttraumatic OA is the irreversible degradation of articular cartilage following traumatic injuries and/or aberrant joint loading, leading to joint dysfunction, pain, and limited locomotion (2). In posttraumatic OA, elevated chondrocyte catabolism results in aggravated proteolysis of the cartilage extracellular matrix (ECM) (3). Aggrecan, the major proteoglycan, is one of the first ECM constituents to undergo fragmentation due to enzymatic cleavage by aggrecanases and matrix metalloproteinases (MMPs). This leads to the disassembly of the aggrecan-hyaluronan (HA) supramolecular network and the loss of aggrecan from the ECM (4). In turn, the loss of aggrecan impairs cartilage biomechanical functions (5), disrupts chondrocyte mechanotransduction (6), and accelerates the damage of collagen fibrils (7) and formation of fibrocartilage (8), and so contributes to the vicious loop of irreversible cartilage breakdown. Inhibition of aggrecan depletion from degenerative tissue has the potential to delay cartilage degradation, attenuate OA progression, and prolong joint use.
Decorin, a small leucine-rich proteoglycan (SLRP), could play such a role in regulating cartilage degradation in OA (9). Decorin is a class I SLRP characterized by an ~36-kd leucine-rich protein core harboring one chondroitin sulfate or dermatan sulfate glycosaminoglycan chain at its N-terminus (10). One canonical structural function of decorin is to regulate collagen fibril diameter and inter-
| 1267 fibrillar spacing during fibril assembly in tension-bearing tissues such as tendon, cornea, and skin (11). In cartilage, decorin is actively expressed from the newborn period to adulthood (12), and is one of the most abundant small proteoglycans in the ECM, with a molar concentration (~15 nmoles/ml) similar to that of aggrecan (~20 nmoles/ml) (13). In early human OA, decorin is significantly up-regulated (14,15). Despite this up-regulation, decorin is not released into the synovial fluid at higher levels (16), suggesting that decorin may participate in stabilizing cartilage matrix. It is postulated that this up-regulation may be a compensatory response by chondrocytes to ameliorate cartilage damage (17). This hypothesis is supported by our recent study showing that decorin increases the retention of aggrecan in healthy cartilage ECM (9). However, one recent study showed that decorin-null (Dcn -/-) mice develop higher resistance to forced exercise-induced OA, which was attributed to enhanced transforming growth factor β (TGFβ) signaling in the absence of decorin, indicating a detrimental role of decorin (18). Given these findings, the role of decorin in OA remains inconclusive.
The objective of this study was to elucidate the role of decorin in cartilage degradation and post-traumatic OA progression in vivo. Mild-to-moderate post-traumatic OA was induced in young adult mice via surgical destabilization of the medial meniscus (DMM) (19). Given the crucial role of decorin in cartilage development (9), we first investigated whether the absence of decorin increased susceptibility to OA in Dcn -/-mice (20). Next, to separate decorin activity in OA from that in normal joint growth, we tested the recently established inducible decorin-knockout (Dcn iKO ) mouse model (21). We allowed for normal joint growth in these mice, and then induced the knockout of decorin expression at the time of surgical DMM. Thus, the resulting phenotype represents the impact of decorin loss during OA progression, with developmental defects being minimized. In these models, we assessed cartilage damage and sulfated glycosaminoglycan (sGAG) loss, cartilage surface fibrillation and tissue modulus changes, as well as alterations in subchondral bone structure. Furthermore, we tested whether decorin alters chondrocyte catabolism, or the retention of fragmented aggrecan in the matrix, or both, using cartilage explants exposed to inflammatory stimuli. Taken together, our findings pointed to a crucial protective role of decorin in increasing aggrecan retention and inhibiting cartilage surface fibrillation in post-traumatic OA.
Animal models. Dcn -/-mice (20) and Dcn iKO mice (Dcn flox/flox / Rosa26Cre ER ) (21) in the C57BL/6 strain were generated as previously described, and were housed in the Calhoun animal facility at Drexel University. To induce the homozygous knockout of the Dcn gene in 3-month-old mice, tamoxifen was injected intraperitoneally on 3 consecutive days beginning 1 week prior to surgical DMM at a dose of 3 mg/40 gm body weight in the form of 20 mg/ml suspended in sesame oil (catalog no. S3547; Sigma) with 1% volume/volume benzyl alcohol (catalog no. 305197; Sigma). Quantitative polymerase chain reaction (qPCR) was performed on day 5 to confirm that tamoxifen-induced gene excision reduced the expression of Dcn to the baseline level (Supplementary Figure 1, available on the Arthritis & Rheumatology web site at http://onlin elibr ary.wiley.com/doi/10.1002/art.41254/ abstract). For Dcn -/-mice, age-matched wild-type (WT) littermates from the breeding of decorin heterozygous (Dcn +/-) mice were used as controls. For Dcn iKO mice, 2 control groups were used, including Dcn iKO mice injected with vehicle (the same amount of sesame oil and benzyl alcohol but without tamoxifen), and WT mice injected with tamoxifen at the same dose and frequency. All mice used in this study were genotyped according to standard procedures (20,21). Animal experiments were approved by the Institutional Animal Care and Use Committee at Drexel University.
Surgical DMM was performed on the right hind knees of 3-month-old male mice for all genotypes, according to an established procedure (19), with Sham surgery performed on the contralateral left knees. Briefly, after anesthesia, the joint capsule was opened and the medial meniscotibial ligament was cut to de stabilize the medial meniscus. The Sham surgery was performed by opening the joint capsule in the same manner to expose the ligament, but without further damage. Dcn -/-and WT mice were euthanized 2 or 8 weeks after surgery, and Dcn iKO mice and their controls were euthanized 8 weeks after surgery, for further analyses.
Histological analysis and immunofluorescence imaging. Whole murine hind knee joints (n = 6 per group) were harvested and fixed in 4% paraformaldehyde, first used for microcomputed tomography (micro-CT) analysis, and then decalcified in 10% EDTA for 4 weeks before being embedded in paraffin for histological analysis. Serial 6-μm-thick sagittal sections were prepared, and 2 sections out of every consecutive 6 sections of the medial side of the murine knees subjected to Sham operation or DMM were stained with Safranin O-Fast Green. For each joint, ~15 sections were obtained and scored in a blinded manner by 2 observers (QL and CW) using a modified Mankin scale (22). Uncalcified cartilage thickness (t uncalcified ) and total cartilage thickness (t total ) were determined by averaging 6 thickness values evenly distributed across the entire cartilage, according to an established procedure (23). For immunofluorescence imaging of decorin, additional paraffin sections were treated with 0.1% pepsin (catalog no. P7000; Sigma) for antigen retrieval, and blocked with 5% bovine serum albumin (BSA) in phosphate buffered saline (PBS) for 1 hour at room temperature. Sections were first incubated with primary antibody (LF-114; a gift from Dr. Larry W. Fisher, National Institute of Dental and Craniofacial Research, Bethesda, MD) (1:100 dilution) overnight at 4°C, and then with secondary antibody (Alexa Fluor 594; ThermoFisher) (1:500) for 2 hours at room temperature. The sections were washed with PBS, counterstained, mounted with DAPI (Fluoromount-G) (catalog no. 0100-20; SouthernBiotech), and imaged with a Zeiss Axio Observer microscope (Carl Zeiss).
Additional immunofluorescence imaging was performed on healthy and OA human cartilage specimens obtained from deidentified donors who had undergone total arthroplasty (n = 3). Similar to murine specimens, serial paraffin sections were treated with 0.1% pepsin, blocked with 5% BSA in PBS, and incubated with primary antibody (20 μg/ml) (AF-143; R&D Systems) and then secondary antibody (Alexa Fluor 594; ThermoFisher). The specificity of decorin antibodies was confirmed by the staining of isotype controls (1:100) (for mouse specimens, AB37415 [Abcam]; for human specimens, AB-108-C [Novus Biologicals]) (Supplementary Figure 2, available on the Arthritis & Rheumatology web site at http://onlin e libr ary.wiley.com/doi/10.1002/art.41254/ abstract).
Atomic force microscopy (AFM)-based nanoindentation. Atomic force microscopy (AFM)-based nanoindentation was applied to freshly dissected femoral condyle cartilage (n = 5 or more specimens per group), according to an established procedure (23). The indentation tests were performed using borosilicate microspherical colloidal tips (R ~5 μm; nominal k ~8.9 N/m; HQ:NSC35/tipless/Cr-Au; cantilever A; NanoAndMore) and a Dimension Icon AFM (Bruker Nano) at a rate of 10 μm/second up to ~1 μN maximum load in PBS with protease inhibitors (Pierce 88266; ThermoFisher). For each joint, at least 10-15 locations were tested on the load-bearing region of the medial condyle to account for spatial heterogeneity. The effective indentation modulus (E ind ) was calculated by fitting each force-indentation depth loading curve with the Hertz model.
Scanning electron microscopy was used to quantify the fibril nanostructure on condyle cartilage surfaces, according to an established procedure (24). Immediately after the AFM tests, joints were treated with 0.1% trypsin (catalog no. T7409; Sigma) and 20 units/ml hyaluronidase (catalog no. H3506; Sigma) at 37°C for 24 hours each to remove proteoglycans, fixed with Karnovsky's fixative at room temperature for 3 hours, sequentially dehydrated in graded water-ethanol and ethanolhexamethyldisilazane mixtures, and air dried overnight (n = 4 samples per group). Samples were then coated with ~6-nm-thick platinum and imaged using a Supra 50VP scanning electron microscope (Carl Zeiss) (Supplementary Figure 3, available on the Arthritis & Rheumatology web site at http://onlin elibr ary.wiley.com/ doi/10.1002/art.41254/ abstract). Collagen fibril alignment angles, θ, were measured using ImageJ, and fitted with von Mises probability density function to calculate the von Mises concentration parameter, κ, a quantitative measure of the degree of fibril alignment (25).
Micro-CT scanning. Micro-CT scanning was performed to assess concurrent changes in subchondral bone 8 weeks after DMM. For mice used for histological analysis, prior to demineralization, knee joints (n = 5 per group) were scanned ex vivo using MicroCT 35 (Scanco Medical) at 6 μm isotropic voxel size and smoothed by a Gaussian filter (sigma = 1.2, support = 2.0). Each region of interest (ROI) for subchondral bone plate, subchondral trabecular bone (STB), and medial meniscal ossicles was contoured at a threshold corresponding to 30% of the maximum image gray scale. The subchondral bone plate of the tibia plateau on the medial side central loading region was contoured to calculate the subchondral bone plate thickness (SBP.Th), according to an established procedure (26). STB was contoured on the entire load-bearing ROI on the medial side (27) to calculate structural parameters, including bone volume/total volume fraction, trabecular number, and trabecular thickness via Scanco software for trabecular bone 3-dimensional standard microstructural analysis. In addition, meniscal ossicle bone volume was directly measured via Scanco software for standard microstructural analysis.
In vitro cartilage explant model. To assess the impact of decorin loss on sGAG release from degenerative cartilage, femoral head cartilage explants were isolated from 3-week-old WT and Dcn -/-mice (n = 6 per group), according to an established procedure (28). For live explants, immediately after harvesting, explants were sterilized and pre-cultured at 37°C with 5% CO 2 for 2 days in Dulbecco's Modified Eagle's Medium (catalog no. 11960; Ther-moFisher) mixed with 10% fetal bovine serum, 1× insulin-transferrin-selenium-sodium pyruvate (catalog no. 51300; ThermoFisher), 2 mM l-glutamine (catalog no. 25030; ThermoFisher), 250 μM lascorbic acid 2-phosphate (catalog no. A8960; Sigma), and 2× penicillin-streptomycin (catalog no. 15140; ThermoFisher). The explants were then cultured in the same media but with 1× penicillin-streptomycin, supplemented with 10 ng/ml recombinant murine interleukin-1β (IL-1β) (catalog no. 211-11B; PeproTech) for 3 days, with the control group cultured in the same manner but without IL-1β. On day 3, the amounts of sGAGs released to the media and retained in cartilage were assessed via dimethylmethylene blue dye binding assay after papain digestion, and qPCR was performed on additional explants to measure the expression of anabolic and catabolic genes. Total RNA (250 ng per well) was subjected to reverse transcription using a TaqMan reverse transcription kit (catalog no. N8080234; ThermoFisher) with amplification carried out via Pow-erUp SYBR Green Master Mix (catalog no. A25742; ThermoFisher) on a RealPlex 4S Mastercycler (Eppendorf). The genes tested and their associated primer sequences are listed in Supplementary Table 1, available on the Arthritis & Rheumatology web site at http://onlin elibr ary.wiley.com/doi/10.1002/art.41254/ abstract.
For devitalized explants, after extraction and pre-culture, explants underwent 3 freeze-thaw cycles between -80°C for 2 hours and 37°C for 45 minutes (29), and were cultured in the same medium supplemented with 20 nM recombinant human ADAMTS-5 (a disintegrin and metalloproteinase with thrombospondin motifs 5) (30) (catalog no. 2198-AD; R&D Systems) or MMP-13 (31) (catalog no. 511-MM-010; R&D Systems) for 4
| 1269 days. The amounts of sGAGs released to media and retained in explant were assessed according to the same procedure as described above. For both live and devitalized explants, cell viability was assessed via fluorescein diacetate and propidium iodide staining (Supplementary Figure 4, Arthritis & Rheumatology web site at http://onlin elibr ary.wiley.com/doi/10.1002/ art.41254/ abstract) (29).
To avoid the assumption of normal distribution of the data, nonparametric statistical tests were applied. To test the significance between genotypes within each surgery type or treatment condition, Mann-Whitney U test was applied to compare t total , t uncalcified , E ind , modified Mankin score, micro-CT outcomes, gene expression, and ratio of sGAG release. Wilcoxon's signed rank test was applied to compare these parameters between surgery types within each genotype. To compare the degree of fibril alignment, θ, the Mardia and Jupp test of concentration equality (32) was used to compare the von Mises concentration parameter κ between genotypes and surgery types. All quantitative outcomes and statistical analysis results are summarized in Supplementary Tables 2345, available on the Arthritis & Rheumatology web site at http://onlin e libr ary.wiley.com/doi/10.1002/art.41254/ abstract. For all tests, the significance level was set at α = 0.05.
Accelerated cartilage degradation in Dcn -/-mice after surgical DMM. In healthy human cartilage matrix, decorin was present in both the pericellular and further-removed territorial/interterritorial domains of the ECM (Figure 1A). In human OA specimens, decorin was significantly up-regulated and present throughout the damaged cartilage matrix. This observation was consistent with the findings of previous studies showing an increase in decorin in human OA cartilage (14,15). In WT mice subjected to Sham surgery, decorin was also distributed throughout the ECM. By 8 weeks after surgical DMM, we detected increased staining of decorin in WT mouse cartilage (Figure 1B), which validated DMM as an appropriate model for investigating the role of decorin and its up-regulation in post-traumatic OA.
In the DMM model, Dcn -/-mice developed accelerated cartilage degradation compared to WT mice, as shown by histological analysis (Figure 1C). Two weeks after surgery, while WT mouse cartilage did not show appreciable damage, Dcn -/-mouse cartilage started to develop loss of sGAG staining on the surface, contributing to higher modified Mankin scores (Figure 1D). By 8 weeks after surgery, both genotypes exhibited salient OA signs, which were more pronounced in Dcn -/-mice. Specifically, Dcn -/-mouse cartilage was characterized by the formation of surface fissures, a further reduction in sGAG staining, more substantial cartilage thinning (lower t uncalcified ) (Figure 1E), and thus, significantly higher Mankin scores (Figure 1D).
the cartilage surface in Dcn -/-mice after surgical DMM. One distinctive OA phenotype of Dcn -/-mouse cartilage was the nanoscale surface fibrillation, as observed by scanning electron microscopy (Figure 2A). In healthy joints, the cartilage surface is covered by a transversely random mesh of collagen fibrils (33), present in both WT and Dcn -/-adult mouse cartilage (Supplementary Figure 3, available on the Arthritis & Rheumatology web site at http://onlin elibr ary.wiley.com/doi/10.1002/art.41254/ abstract) (9). In the Sham-operated groups, both ge notypes retained this random fibrillar architecture. In the surgical DMM groups, this feature was retained in WT mice up to 8 weeks after surgery. In contrast, Dcn -/-mouse cartilage surface started to develop aligned fibrils as early as 2 weeks after surgery, and were dominated by highly aligned, densely packed collagen fibrils 8 weeks after surgery. These changes were signified by the much higher von Mises concentration, κ, in Dcn -/-mouse cartilage (Figure 2B). These fibrils were aligned along the mediolateral orientation (Supplementary Figure 3), suggesting that the surface fibrillation could be induced by extensive shearing of the destabilized medial meniscus during joint loading. In both genotypes, cartilage damage was associated with marked reduction in modulus (E ind ) (Figure 2C). This reduction was attributed to the loss of cartilage ECM structural integrity and illustrated the impaired cartilage load-bearing function in OA (23). In ROLE OF DECORIN IN POST-TRAUMATIC OA | 1273
both the Sham-operated group and the group subjected to surgical DMM, Dcn -/-mouse cartilage showed lower modulus compared to WT mouse cartilage, suggesting that loss of decorin impairs cartilage load-bearing function both during normal skeletal growth and in DMM-induced cartilage degradation.
Lack of an appreciable subchondral bone phenotype in Dcn -/-mice after surgical DMM. Given the crucial interplay between cartilage and subchondral bone in OA development (2), we investigated whether loss of decorin also impacted subchondral bone after DMM. Eight weeks after surgery, for both mouse genotypes, the DMM group showed no significant structural changes in either the subchondral bone plate or STB relative to the Sham-operated group (Figures 3A-C). This observation is consistent with the findings of previous studies showing that in the DMM model, subchondral bone changes occur only after the erosion of cartilage in late OA (34). No differences were detected between the WT and Dcn -/-mice with regard to subchondral bone plate or STB structure in either the group subjected to DMM or the Sham-operated group, indicating that loss of decorin does not impact the remodeling of subchondral bone in the DMM model. In contrast, 8 weeks after surgical DMM, we detected the formation of osteophytes in Dcn -/-, but not WT, mouse joints (Figure 3D), which indicated more advanced OA (35). DMM increased meniscal ossification at the horns in both genotypes, and Dcn -/-mouse joints showed a marginally higher degree of ossification at the posterior, but not the anterior, end compared to WT mouse joints (Figure 3E). Taken together, micro-CT findings suggest that the accelerated OA progression in Dcn -/-mice is more likely to be a direct impact of decorin loss in cartilage, rather than a secondary effect arising from changes in underlying subchondral bone.
The more severe OA phenotype in Dcn -/-mice could be attributed to both altered OA pathology and impaired joint growth prior to surgery in the absence of decorin. In the Sham-operated group, Dcn -/-mouse cartilage showed moderate Mankin scores (Figure 1D) and lower modulus relative to WT mice (Figure 2C), which was due to the lower sGAG content in Dcn -/-mouse cartilage even without surgery (9). To separate the role of decorin in OA from its role in joint growth, we studied OA progression in Dcn iKO mice 8 weeks after surgery. The expression of decorin was maintained up to maturity, and only ablated at the time of DMM. In these mice, we also detected accelerated OA, with features comparable to those seen in Dcn -/-mice. Compared to control mice, Dcn iKO mice developed higher Mankin scores, more pronounced reduction in sGAG staining, and salient surface fibrillation (Figures 4A-E). Dcn iKO mice also developed osteophytes (Supplementary Figure 5, available on the Arthritis & Rheumatology web site at http://onlin elibr ary.wiley.com/doi/10.1002/ art.41254/ abstract) but did not show appreciable subchondral bone phenotype (Figure 4G). Thus, the accelerated OA phenotype in Dcn -/-mice and Dcn iKO mice was primarily due to the loss of the protective role of decorin in DMM-induced cartilage degradation.
In comparison to Dcn -/-mice (Figures 1C-E), Dcn iKO mice did not demonstrate more severe cartilage thinning (Figure 4C) and had moderately lower Mankin scores (Figure 4B). This confirmed that the OA phenotype in Dcn -/-mice was a result of the combined effects of both altered OA pathology and impaired joint growth. Further, while the modulus of Dcn iKO cartilage was also reduced by DMM, it was similar to that of control mouse cartilage (Figure 4F) and higher than that of Dcn -/-mouse cartilage. The modulus of Dcn iKO cartilage could reflect the properties of newly formed fibrous tissues on the surface, which lack aggrecan and associated sGAGs, and thus are incapable of dissipating energy through poroelasticity as normal hyaline cartilage.
To determine if loss of decorin directly impacts chondrocyte catabolism, we analyzed chondrocyte gene expression in mouse femoral head cartilage explants (Figure 5A). Upon stimulation with IL-1β, there was a decrease in the expression of the anabolic genes Acan and Col2a1 in WT but not Dcn -/-mouse cartilage explants (Figure 5A). In contrast, major catabolic genes, including aggrecanases and MMPs, were significantly up-regulated by 10-100-fold after IL-1β stimulation in both mouse genotypes. We did not detect significant differences between the 2 genotypes in any of the genes tested except for decorin. This suggests that decorin does not directly regulate the anabolic or catabolic activities of chondrocytes, either with or without IL-1β stimulation.
Finally, we compared the amount of sGAGs released from explants in both live and devitalized conditions. Dcn -/-mouse cartilage has lower amounts of aggrecan and sGAGs than WT mouse cartilage, which indicates a lower concentration gradient for diffusion-driven release of sGAGs. Despite this, in live explants, when IL-1β aggravated chondrocyte catabolism, a greater proportion of sGAGs was released from Dcn -/-mouse explants than from WT mouse explants (Figure 5B). Further, in devitalized explants, when chondrocyte metabolism was abolished, Dcn -/-mouse explants still experienced a higher degree of sGAG release upon exogenous proteolysis of aggrecan by ADAMTS-5 or MMP-13 (Figure 5C). Therefore, in the explant model, the absence of decorin accelerates the loss of fragmented aggrecan from cartilage matrix, but does not directly alter chondrocyte metabolism when stimulated with IL-1β.
This study highlights the crucial role of decorin in mediating cartilage degradation in post-traumatic OA, as evidenced by the accelerated OA phenotype in both Dcn -/-and Dcn iKO mice (Figures 1234). We hypothesize that in degenerative cartilage, decorin increases the retention of fragmented aggrecan within degrading matrix, thereby delaying aggrecan loss and cartilage damage (Figure 6A). This hypothesis extends findings from our recent study on the regulatory roles of decorin in postnatal cartilage growth (9). Specifically, we showed that in healthy cartilage, decorin functions as a "physical linker" (Figure 6B) to increase molecular adhesion of aggrecan-aggrecan and aggrecan-type II collagen fibrils. The canonical assembly mechanism of the aggrecan network in the ECM is the aggregation of aggrecan-HA via the G1 domain (36), but this mechanism does not fully explain the integrity of aggrecan across development and disease states (37). While aggrecan can also form networks through interacting with tenascins (38) and fibulin -2 (39) via its G3 domain, the presence of G3 domain decreases markedly with age (40). To this end, the decorin-mediated aggrecan network assembly could be a crucial mechanism in maintaining the integrity of the aggrecan network in the ECM; this notion is supported by the markedly impaired biomechanical functions of Dcn -/-mouse cartilage (9). In OA, when aggrecan molecules become increasingly fragmented and dissociated from the aggrecan-HA aggregates (4), the upregulation of decorin could be a reparative attempt to increase the retention of fragmented aggrecan, and so, attenuate aggrecan depletion (Figure 6A). Indeed, this hypothesis is supported by outcomes from explant models, in which loss of decorin accelerates the release of fragmented aggrecan from both live and devitalized cartilage (Figures 5B and C). In the DMM model, both Dcn -/-and Dcn iKO mouse cartilage develop extensive surface fibrillation (Figures 2A and B and Figures 4D and E). We attribute this effect to the loss of the protective effect of aggrecan against collagen remodeling. In the ECM, the densely packed, aggrecan-HA aggregates occupy the ~100-nm-sized interfibrillar spacing within the porous type II/IX/XI collagen network (41), limiting aberrant collagen fibril lateral fusion and overgrowth. When loss of decorin leads to accelerated aggrecan depletion and impairs aggrecan's protection of the collagen fibrillar network following DMM, the extensive shear and frictional forces from the destabilized meniscus can induce the alignment of collagen fibrils along the shear direction (Figures 2A and 4D). In addition, decorin could also directly inhibit fibril lateral fusion given its capability of binding to type II collagen (42), but such contribution may be less important given its low mass concentration relative to aggrecan in cartilage (13). The presence of aligned fibrils suggests that hyaline cartilage loses its integrity and transforms into fibrocartilage, which does not possess the energy dissipation function endowed by aggrecan (8). These results thus support the notion of a critical protective role of decorin in inhibiting cartilage fibrillation, an irreversible degenerative step in OA progression (8).
Integrating this study with our recent work on young adult Dcn -/-mouse cartilage (9), we show that in both healthy and degenerative cartilage, the primary role of decorin is to mediate ECM assembly, rather than to influence chondrocyte metabolism. The absence of the impact of decorin on chondrocyte response to IL-1β (Figure 5A) is similar to its lack of impact on the response to growth factor TGFβ1 (9). In both scenarios, although loss of decorin does not alter chondrocyte metabolism, it significantly reduces the retention of aggrecan in cartilage matrix, both during degeneration in situ (Figures 5B and C) and during regeneration in alginate culture (9). In vivo, the role of decorin in increasing aggrecan retention could be more crucial, as the extensive physiologic joint loading and associated interstitial fluid flow can further promote aggrecan depletion when its assembly is impaired. This finding is different from previous studies showing higher resistance of Dcn -/-mice to forced exercise-induced OA (18). We attribute this contrast mainly to differences in the OA models used. The forced exercise model represents joint overuse (43), while DMM induces joint instability and inflammation (19), and so the two models could have different disease etiology. For example, in WT mouse cartilage, the expression of decorin is suppressed by forced exercise (18), but elevated by DMM (Figure 1B). This increase in decorin expression corroborates observations in both murine (Figure 5A) and bovine (44) cartilage explants stimulated with IL-1β, as well as cartilage specimens from OA patients (14,15). To this end, the protective role of decorin in post-traumatic OA is confirmed by the consistency of an accelerated OA phenotype in Dcn -/-and Dcn iKO mice (Figures 1234), which were established in different manners (20,21).
This study has several limitations. First, while we showed that decorin increases molecular adhesion of aggrecan (9), we did not identify how decorin specifically interacts with aggrecan core protein or its sGAGs, and how such interaction is altered when decorin becomes increasingly fragmented as OA advances (45)(46)(47). To address this limitation, our ongoing studies sought to test the interactions between purified decorin, aggrecan, and type II collagen at the single-molecule level. Second, the half-life of decorin in adult murine cartilage is unknown. The lack of phenotype in the Sham-operated Dcn iKO mice (Figure 4) is in stark contrast with the pronounced phenotype that resulted from induced decorin knockout at 1 month of age (9). This could be due to either the presence of residual decorin, or the fact that decorin plays a less essential role in adulthood (≥3 months of age). However, while we cannot delineate the influence of residual decorin in DMM, the accelerated OA phenotype in Dcn iKO mice (Figure 4) clearly demonstrated the impact of ablating the decorin up-regulation on DMM-induced cartilage degradation. Third, although we showed that decorin does not significantly impact chondrocyte catabolism under stimulation with IL-1β, given its versatile interactomes (10), it is possible that decorin may affect chondrocyte signaling through its bindings with cell surface receptors and cytokines in other OA models. For example, spontaneous OA is associated with reduced autophagy in cartilage (48), and decorin may also ameliorate OA through evoking autophagy (49), which will be a topic of our future work.
In summary, this study identifies decorin as a central player in cartilage degradation in post-traumatic OA. Loss of decorin aggravates aggrecan depletion and surface fibrillation, leading to accelerated cartilage damage. We hypothesize that upregulation of decorin in early OA acts as a reparative attempt to increase the retention of fragmented aggrecan in the ECM, thereby delaying aggrecan loss, cartilage fibrillation, and irreversible cartilage breakdown (Figure 6A). This role is mediated by the interactions of decorin with aggrecan molecules and type II collagen fibrils. Therefore, modulating decorin activities through decorin-targeting gene therapies or decorin-based biomaterials has the potential to attenuate OA progression and prolong joint use.
Forward Primer Reverse Primer Dcn 5'-TGAGCTTCAACAGCATCACC-3' 5'-AAGTCATTTTGCCCAACTGC-3' Acan 5'-GACTGTGTGGTGATGATCTG-3' 5'-CTCGTAGCGATCTTTCTTCTG-3' Col2a1 5'-GCTGGTGCACAAGGTCCTAT-3' 5'-ACCCTGCAGTCCAGTGAAAC-3' Bgn 5'-CTACGCCCTGGTCTTGGTAA-3' 5'-ACTTTGCGGATACGGTTGTC-3' Adamts4 5'-AGGTAAGGCCTCAGTCAGCA-3' 5'-AAGGAGCACAGACGATGCTT-3' Adamts5 5'-TTGCTCTCCTCGAAGTGGTT-3' 5'-ATGGGTCTGGAGATCGAGTG-3' Mmp3 5'-GATCTCTTCATTTTGGCCATCTCTTC-3' 5'-CTCCAGTATTTGTCCTCTACAAAGAA-3' Mmp9 5'-GGAACTCACACGACATCTTCCA-3' 5'-GAAACTCACACGCCAGAAGAATTT-3' Mmp13 5'-CCTTCTGGTCTTCTGGCACAC-3' 5'-GGCTGGGTCACACTTCTCTG-3' Actb 5'-AGATGACCCAGATCATGTTTGAGA-3' 5'-CACAGCCTGGATGGCTACGT-3' Gapdh 5'-TCAACAGCAACTCCCACTCTTCCA-3' 5'-ACCCTGTTGCTGTAGCCGTATTCA-3'
Table S2. Summary of structural and biomechanical analysis outcomes of wild-type (WT) and decorin-null (Dcn -/-) joints at 2 and 8 weeks post-surgery, shown as mean ± 95% CI from values averaged by each animal, except for κ as mean [95%CI] 2 weeks Sham DMM p-value (Sham vs DMM) n Unit WT Dcn -/- p WT Dcn -/- p WT Dcn -/- WT Dcn -/- tuncalcified, femur 40 ± 7 43 ± 6 0.589 41 ± 7 41 ± 6 0.818 0.844 0.688 6 6 μm ttotal, femur 110 ± 17 110 ± 15 0.937 107 ± 15 108 ± 15 0.818 0.844 1.000 6 6 tuncalcified, tibia 53 ± 4 57 ± 5 0.093 53 ± 4 51 ± 5 0.589 1.000 0.026 6 6 ttotal, tibia 124 ± 8 126 ± 9 0.818 128 ± 13 122 ± 12 0.240 0.394 0.485 6 6 Mankin Score 0.33 ± 0.54 2.33 ± 0.86 0.006 0.67 ± 0.54 4.17 ± 0.54 0.002 0.317 0.034 6 6 a.u. Eind 1.54 ± 0.20 0.47 ± 0.12 0.004 0.58 ± 0.19 0.41 ± 0.20 0.177 0.031 0.625 6 5 MPa κ 0.37 [0.23 0.53] 0.40 [0.26 0.56] 0.963 0.38 [0.24 0.54] 1.39 [1.10 1.67] < 0.001 0.869 < 0.001 4 4 --8 weeks Sham DMM p-value (Sham vs DMM) N Unit WT Dcn -/- p WT Dcn -/- p WT Dcn -/- WT Dcn -/- tuncalcified, femur 42 ± 7 40 ± 6 0.132 27 ± 2 20 ± 3 0.002 0.031 0.031 6 6 μm ttotal, femur 116 ± 10 107 ± 16 0.240 100 ± 16 82 ± 5 0.002 0.031 0.031 6 6 tuncalcified, tibia 55 ± 5 53 ± 5 0.310 34 ± 3 27 ± 7 0.041 0.002 0.002 6 6 ttotal, tibia 125 ± 11 120 ± 7 0.485 113 ± 13 91 ± 9 0.180 0.002 0.002 6 6 Mankin Score 1.22 ± 0.42 2.72 ± 0.92 0.006 5.86 ± 1.02 8.10 ± 0.55 0.002 0.031 0.031 6 6 a.u. Eind 1.67 ± 0.46 0.44 ± 0.21 < 0.001 0.29 ± 0.08 0.18 ± 0.08 0.029 0.008 0.031 8 6 MPa κ 0.35 [0.22 0.50] 0.42 [0.27 0.58] 0.690 0.40 [0.25 0.56] 3.30 [2.73 3.93] < 0.001 0.793 < 0.001 4 4 --SBP.Th 124 ± 41 108 ± 18 0.690 124 ± 23 128 ± 13 0.841 0.813 0.125 5 5 μm STB BV/TV 50 ± 18 40 ± 11 0.190 54 ± 16 48 ± 8 0.421 0.138 0.043 5 5 % STB Tb.N 6.8 ± 1.4 6.2 ± 0.9 0.310 6.8 ± 1.3 6.5 ± 0.9 0.421 0.225 0.043 5 5 mm -1 STB Tb.Th 89 ± 24 67 ± 19 0.095 103 ± 28 90 ± 8 0.222 0.138 0.043 5 5 μm Men. OVant. 96 ± 6 85 ± 5 0.016 204 ± 67 158 ± 48 0.310 0.043 0.043 5 5 × 10 -3 mm 3 Men. OVpost. 13 ± 10 22 ± 10 0.151 42 ± 24 80 ± 30 0.056 0.043 0.043 5 5 Dcn 1.00 ± 0.43 0.06 ± 0.04 0.002 1.98 ± 0.53 0.06 ± 0.03 0.002 0.004 0.485 6 6 relative mRNA (to WT without IL-1β) Acan 1.00 ± 0.15 1.04 ± 0.30 0.937 0.57 ± 0.16 0.80 ± 0.38 0.394 0.002 0.310 6 6 Col2a1 1.00 ± 0.43 0.98 ± 0.56 0.818 0.54 ± 0.26 0.71 ± 0.21 0.310 0.026 0.589 6 6 Bgn 1.00 ± 0.24 0.83 ± 0.38 0.310 1.05 ± 0.36 0.92 ± 0.25 0.310 0.937 0.699 6 6 Adamts4 1.00 ± 0.79 1.14 ± 0.34 0.394 13.9 ± 10.1 16.3 ± 7.9 0.394 0.002 0.002 6 6 Adamts5 1.00 ± 0.48 1.18 ± 0.59 0.699 75.8 ± 23.3 73.6 ± 20.4 0.818 0.002 0.002 6 6 Mmp3 1.00 ± 0.48 1.00 ± 0.56 0.965 109 ± 43 98 ± 30 1.000 0.002 0.002 6 6 Mmp9 1.00 ± 0.41 1.06 ± 0.45 0.818 7.06 ± 2.09 6.64 ± 3.15 0.699 0.002 0.002 6 6 Mmp13 1.00 ± 0.49 0.90 ± 0.41 0.699 34.3 ± 10.9 36.8 ± 10.3 0.589 0.002 0.002 6 6 sGAG media 12.6 ± 2.8 10.5 ± 2.8 0.180 23.2 ± 5.8 18.4 ± 4.2 0.093 0.002 0.002 6 6 μg/mg wet wt. sGAG expl. 58.7 ± 6.8 28.2 ± 7.8 0.002 48.8 ± 10.9 18.6 ± 3.0 0.002 0.093 0.015 6 6 sGAG total 71.3 ± 4.9 38.8 ± 6.4 0.002 72.0 ± 8.2 37.1 ± 3.1 0.002 0.937 0.818 6 6 Loss ratio 17.9 ± 4.5 28.0 ± 9.4 0.132 32.6 ± 9.4 49.5 ± 8.6 0.015 0.004 0.002 6 6 %