In vivo effect of opticin deficiency in cartilage in a surgically induced mouse model of osteoarthritis

The SLRP opticin (OPTC) has been demonstrated to be produced and degraded in osteoarthritic (OA) human cartilage. Here, we investigated the in vivo effect of OPTC deficiency in OA cartilage. OA was induced in 10-week-old Optc −/− and Optc +/+ mice. Ten weeks post-surgery, cartilage was processed for histology and immunohistochemistry. SLRP expression was determined in non-operated mouse cartilage. OA Optc −/− demonstrated significant protection against cartilage degradation. Data revealed that in non-operated Optc −/− cartilage, expression of SLRPs lumican and epiphycan was up-regulated at day 3 and in 10-week-olds (p ≤ 0.039), and fibromodulin down-regulated in 10-week-olds (p = 0.001). Immunohistochemistry of OA mice showed a similar pattern. In OA Optc −/− cartilage, markers of degradation and complement factors were all down-regulated (p ≤ 0.038). In OA Optc −/− cartilage, collagen fibers were thinner and better organized (p = 0.038) than in OA Optc +/+ cartilage. The protective effect of OPTC deficiency during OA results from an overexpression of lumican and epiphycan, known to bind and protect collagen fibers, and a decrease in fibromodulin, contributing to a reduction in the complement activation/inflammatory process. This work suggests that the evaluation of the composition of the different SLRPs in OA cartilage could be applied as a new tool for OA prognosis classification.

Sample collection and preparation. DMM and sham operated mice were euthanized at 10 weeks after surgery and the non-operated mice at 20 weeks old. Right knee joints were dissected free of tissue.
The severity of the OA cartilage lesions was determined using the Osteoarthritis Research Society International (OARSI) scoring method 25 .
Subchondral bone plate histomorphometry was done on the medial compartment as described 23,26 using a Leitz Diaplan microscope coupled to a personal computer as above using three Safranin-O/fast green stained sections. To measure the subchondral bone plate thickness, the bone volume/total volume ratio (%BV/TV) and trabecular thickness, a box with a fixed width (1,000 μm) and variable length was created with the upper limit at the calcified cartilage-subchondral bone junction and the lower limit at the subchondral bone-trabecular bone junction. The mean distance between the upper and lower limit was calculated automatically by the software as well as the trabecular thickness.
Histomorphometric quantitative analysis of the anterior synovial membrane thickness was performed by capturing images at 63X with a Leitz Diaplan microscope (Leica Microsystems, Wetzlar, Germany) coupled to a personal computer and data determined with Bioquant OSTEO Image Analysis Software version 12.5.60 MIR (Bioquant, Nashville, TN); data are expressed as micrometers.
Collagen organization. Cartilage collagen organization was evaluated on 5 µm paraffin sections following picrosirius red staining, as described 27 . In brief, each slide was evaluated under polarized light microscopy using a scale of 0-2, where 0 = no birefringence, completely disorganized collagen; 1 = birefringence on 25%-50% of the total cartilage thickness, collagen partially disorganized; and 2 = bright birefringence on more than 50% of total cartilage thickness, organized collagen network. Three areas for each slide were evaluated and the scores summed (maximum score 6).
Control procedures were performed according to the same experimental protocol as follows: (1) omission of the primary antibody, (2) substitution of the primary antibody with a non-specific IgG from the same host as the primary antibody (Santa Cruz Biotechnology), and (3) a third control for type X collagen was performed by adsorption with the peptide YNRQQHYDPRSGIFTCKIPGIYYFSYGGC (provided by Dr. E. Lee, Shriners Hospital for Children, McGill University Hospital Centre, Montreal, Quebec, Canada) at 10-fold. Controls only showed background staining.
For each specimen, positive cells were quantified as previously described 23 . In brief, for each section, 1 microscopic field at 250X captured with a Leitz Diaplan microscope (Leica Microsystems) was assigned and chondrocytes staining positive were quantified following the determination of the total number of cells and of those staining positive for the antigen. Final results were expressed as the percentage of cells staining positive for the antigen with the maximum score being 100% for the entire cartilage, except for lumican which was expressed according to the upper (first half) or deep (second half) cartilage zone.
For type II collagen (Col2-3/4Cshort antibody) and VDIPEN, the matrix staining was assessed as follows. For type II collagen staining, images were captured at 250X with a Leitz Diaplan microscope connected to BIOQUANT OSTEO II Image Analysis software 23 . Surface area of positively stained extracellular cartilage matrix was measured and data expressed as % of positive stained area over total area. VDIPEN staining was graded on a scale of 0-3, where 0 = no staining; 1 = minor staining; 2 = marked staining; and 3 = maximal staining, as described previously 23 . Each slide was examined and scored by 2 independent observers who were blinded to group allocation.
Transmission electron microscopy (TEM). Right knees from 10-week-old Optc +/+ and Optc −/− mice were used. For further details see Supplementary Material, Materials and Methods. Ultra-thin (100 nm) sections from the zone of interest were obtained using a microtome (Leica Microsystems UCT ultramicrotome, Vienna, Austria) and coated with 1-2 nm platinum. From each specimen, images were taken at magnification X1200 and X30000.
Statistical analysis. Statistical significance was assessed by the Mann-Whitney and one sample t test where P values < 0.050 were considered significant. Data availability. The datasets generated and/or analysed during the current study are available from the corresponding author on reasonable request. Data generated or analysed during this study are included in this published article and its Supplementary Material files or are noted as data not shown. Optc −/− present different SLRP expression/production in articular cartilage. Analysis of the expression (mRNA) of SLRP members from each class revealed that biglycan (class I), osteomodulin (class II), osteoglycin (class III), tsukushi and nyctalopin (class IV), and podocan (class V) expression showed no significant difference between non-operated Optc +/+ and Optc −/− mice at either day 3 (P03) or at 10 weeks of age in articular cartilage (data not shown). Interestingly, lumican (class II) (Fig. 3A) and epiphycan (Fig. 3B), a member of the same class as OPTC (class III), were significantly overexpressed at both P03 (p ≤ 0.039) and at 10 weeks of age (p ≤ 0.010) in Optc −/− compared to Optc +/+ mice. Moreover, compared to Optc +/+ , in Optc −/− mice fibromodulin (class II) (Fig. 3C) and PRELP (class II) (Fig. 3D) had similar values at P03, but were significantly down-regulated at 10 weeks of age (p < 0.001) (Fig. 3C,D).
As fibromodulin contributes to an increase in the complement activation/inflammatory process, we further looked at the levels of two factors of the complement activation, C5b-9 and CCL2. Data revealed that for both factors, DMM-Optc −/− mice demonstrated a significantly decreased level of positive chondrocytes (p = 0.004) compared to -Optc +/+ mice (Fig. 4C,D).  showed similar findings for both zones studied and only data of the deep zone are illustrated (Fig. 5B,C). Findings revealed that Optc −/− mice had a significantly higher frequency of cartilage collagen fibers of smaller diameter compared to Optc +/+ mice (Fig. 5D). TEM images also showed that in Optc −/− mice, type II collagen fibers were more tightly packed than in Optc +/+ mice. This was further substantiated by the use of the collagen birefringence methodology (Fig. 5E), in which collagen fibers in the DMM-Optc +/+ mouse cartilage had a significantly lower birefringence intensity (p = 0.038) than in the -Optc −/− mice, indicating a higher disorganization of the collagen fibers in the Optc +/+ mice.

Collagen fibers of Optc
Optc −/− effect on two other articular tissues. As OPTC was also found present in both osteoblasts and synoviocytes 19 , we further looked at the effect of Optc deficiency in both subchondral bone and synovial membrane. Histomophometric evaluations of the subchondral bone: plate thickness, percentage of bone volume/ total volume and trabecular thickness, demonstrated similar values for the DMM-Optc +/+ and -Optc −/− mice ( Supplementary Material Fig. S6). However, synovial membrane thickness comparison of DMM-Optc +/+ with -Optc −/− revealed a significant increase (p = 0.036) (Supplementary Material Fig. S7).

Discussion
Our group previously reported that OPTC is expressed and produced in articular cartilage and degraded during human OA 19,20 . Because SLRPs contribute to the maintenance and integrity of articular cartilage, OPTC reduction/cleavage in OA was suggested to predispose cartilage to degeneration. Here, we further investigated the in vivo effect of the lack of OPTC during OA, using an Optc-null mouse model.
Although mice with a deficiency of one or more SLRPs have been shown to develop premature OA 11,13 , unexpectedly and in contrast to our hypothesis, the present study showed that OA Optc-null mice exhibited a cartilage protective effect. The lack of OPTC in OA mice resulted in significantly better cartilage histology features and lower expression markers of catabolic processes. We then further explored the relevant in vivo mechanisms implicated in OA cartilage associated with the OPTC absence. Data revealed that this ensues from an overexpression of two other members of the SRLP super-family, lumican and epiphycan, combined with a down-regulation of fibromodulin, a factor involved in activation of the complement/inflammatory process. These changes are associated with a better preservation of the collagen fibril organization observed in the Optc-null mice.
Our observation of the cartilage protection in OA Optc −/− mice, together with the literature demonstrating that the lack of some members of the SLRP super-family is compensated by an overexpression of other members 12,16,28,29 , led us to investigate the expression of SLRPs from each class on articular mouse cartilage at birth and in adults. Data revealed that in Optc −/− mice, lumican and epiphycan are overexpressed at birth (P03) and in adults (10 weeks old), and fibromodulin down-regulated in adult mice. Such overexpression of lumican in adult mice has also been reported in the tendons when fibromodulin is absent 15 , and agrees with our data on the significant decrease in fibromodulin in adult Optc −/− mice. The epiphycan overexpression (SLRP belonging to the same class as OPTC) also concurs with cartilage protection during OA, as this SLRP was shown to play an important role in maintaining joint integrity and its absence (knockout mice) resulted in OA development with aging 13 . In addition, interaction between SLRP members has also been documented; for example, epiphycan/biglycan double-deficient mice present a more severe OA than only epiphycan knock-out mice 13 .
Findings of the OA Optc −/− cartilage protective effect also concur with those of the markers of the cartilage degradation, including products of type II collagen and aggrecan, as well as chondrocyte hypertrophy, MMP-3 and MMP-13, which were all significantly and markedly reduced in the DMM-Optc −/− compared to DMM-Optc +/+ mice. SLRPs are encoded in 18 different genes clustered on 7 chromosomes, suggesting duplication to generate functional redundancy during evolution 2 . Interestingly, OPTC and fibromodulin share the same gene cluster, suggesting that OPTC could have a regulatory role in fibromodulin transcription 18 . This could explain, at least in part, the decrease in fibromodulin expression in Optc −/− mouse cartilage. In addition, it could be that fibromodulin degradation by proteases is high in OA cartilage or at least higher than other SLRPs. Indeed, previous data reported that although MMP-13, a major MMP involved in collagen degradation, is able to cleave both lumican and fibromodulin, fibromodulin appears to be the preferred substrate 30 .
Collagen fibrillogenesis during normal cartilage development is finely regulated and involves multiple steps. Type II collagen fibrils impart a strength and compressibility to the cartilage matrix and permit this tissue to resist large deformations in shape, allowing joints to absorb shocks. During the first steps of fibrillogenesis, a population of collagen molecules is assembled into quarter staggered arrays forming short small-diameter fibril intermediates 31 (Supplementary Material Fig. S8). After the molecular assembly phase, fibrils start growing and become longer and larger in diameter and the stabilization of the assembled fibrils is mediated by interactions with fibril-associated molecules such as SLRPs 29 , more specifically by lumican and fibromodulin, two SLRPs differently regulated in the Optc −/− mice. Lumican and fibromodulin belong to the same SLRP subfamily (class II) and show a 47% sequence identity. Additionally, these SLRPs are similar in their post-translational modifications. Both carry tyrosine sulfate residues in their N-terminal domains and are substituted with carbohydrates in a similar manner 32 . They bind to the same binding site of type I collagen fibrils, although fibromodulin has a higher affinity than lumican 33 . Even though their binding characterization to type II collagen has not been studied, one could speculate that these two SLRPs could also compete for binding in this collagen type. Moreover, data from  Fig. S8). At the beginning of the collagen fibril growth phase, lumican promotes the lateral fibril growth by fusion of intermediate subunits and, as fibrillogenesis progresses, the lumican level decreases to a barely detectable level, while the fibromodulin level increases significantly and is essential for diameter growth of the mature fiber at later stages 29 .
These data, together with our findings on the SLRP expression/production and the cartilage collagen analysis, and further experiments in which the Optc +/+ lumican, epiphycan and fibromodulin fold change expression (mRNA) in 10-week-old mice over P03 mice was investigated and found down-regulated at 10 weeks old (Supplementary Material Fig. S9), are suggestive of a collagen fibrillogenesis model in the absence of OPTC (Supplementary Material Fig. S8).The decreased level of fibromodulin is translated by a compensatory overexpression of lumican over time. As a result, the collagen fibers of the Optc −/− mice are over-coated by lumican and epiphycan molecules during the fibrillogenesis process, preventing the final steps of the fibril process from developing and the fibers to fuse laterally, thus remaining thinner and tightly packed, and therefore less susceptible to degradation. This scheme is strengthened by the picrosirius red data that showed that OA-Optc −/− mice maintained a better organization of the cartilage collagen network than OA-Optc +/+ mice.
In turn, the decrease in fibromodulin in Optc −/− mice could be responsible for less catabolic activity in the cartilage. Indeed, fibromodulin, in addition to its structural role in the cartilage matrix, is known to activate the classical complement pathway 9 . Unbalanced complement activation has been shown to play a central role in OA cartilage 34 . The complement system consists of pathways that converge at the formation of the C3 and C5 convertases, enzymes that mediate activation, among others, of the formation of membrane attack complex (MAC), comprising the complement effector C5b-9 35 . The levels of C5b-9 have been found elevated in OA synovial fluid when compared to healthy individuals 34,36 . Interestingly, MAC co-localized with MMP-13 around human OA chondrocytes and was shown to play a critical role in the pathophysiology of this disease, in which in addition to inducing cell lysis, it increases the chondrocytes' expression of multiple genes including MMPs, inflammatory cytokines and complement effectors 34 . We then looked at the effect of the lack of OPTC on two representative complement factors, C5b-9 and CCL2 (also referred to as monocyte chemoattractant protein 1 [MCP-1]), which have been found significantly increased in serum and synovial fluid of human OA patients 34,36,37 . Data revealed that the levels of these two factors were significantly decreased in OA-Optc −/− compared to -Optc +/+ mice contributing to a reduction in the inflammatory/catabolic processes during the disease process. Although the hallmark of OA is the progressive degeneration of articular cartilage, it is acknowledged that the subchondral bone and synovial membrane alterations are also components of the disease process. An interdependence between subchondral bone alterations and cartilage degeneration, as well as inflammatory factors released from the synovial membrane, is gaining strong support. As data showed that OPTC deficiency imparts a protective effect on the synovial membrane thickness, but not on the subchondral bone, the question arises as to whether a protective effect on the synovial membrane alterations could also be beneficial to the cartilage in this OA mouse model. Although more research should be performed to that effect, in this OA model it is noteworthy that inflammation of the synovial membrane is not the main feature 38 and that OA alterations in this tissue are more related to non-inflammatory progressive fibrosis. However, it could be that the synovial membrane preservation occurs via a direct effect of the cartilage preservation, as less degradation products will be released from the cartilage.
In summary, this is the first report showing that the in vivo lack of OPTC destabilizes the natural balance of SLRP members in the cartilage, which is translated into a protective environment in OA cartilage. Such effect is a result of different events comprising an overexpression of lumican and epiphycan expression/production at birth and over time and therefore a protection of the collagen fibers by their surface deposition, enabling the fibrils to better resist catabolic factors. In addition, the protective effect also reflects the decrease in fibromodulin, which will result in a reduction of the classical complement pathway activation during the progression of the disease, favoring production of less catabolic factors, thus less degradation.
In conclusion, although OPTC, a SLRP associated with collagen fibrils, contributes to the structural stability of cartilage, its lack leads to a protection of cartilage due to an overexpression of other SLRP members throughout adult life, thus preserving the collagen fibril from degradation, and by a reduction in the complement activation/ inflammatory process in this tissue during OA through the reduction of another member. This work suggests that the evaluation of the composition of the different SLRPs in human OA cartilage could be applied as a new tool for OA prognosis classification.