Abstract Changes in the mechanical homeostasis of the temporomandibular joint (TMJ) can lead to the initiation and progression of degenerative arthropathies such as osteoarthritis (OA). Cells sense and engage with their mechanical microenvironment through interactions with the extracellular matrix. In the mandibular condylar cartilage, the pericellular microenvironment is composed of type VI collagen. NG2/CSPG4 is a transmembrane proteoglycan that binds with type VI collagen, and has been implicated in the cell stress response through mechanical loading-sensitive signaling networks including ERK 1/2. The objective of this study is to define the role of NG2/CSPG4 in the initiation and progression of TMJ OA and to determine if NG2/CSPG4 engages ERK 1/2 in a mechanical loading dependent manner. In vivo, we induced TMJ OA in control and NG2/CSPG4 knockout mice using a surgical destabilization approach. In control mice, NG2/CSPG4 is depleted during the early stages of TMJ OA and NG2/CSPG4 knockout mice have more severe cartilage degeneration, elevated expression of key OA proteases, and suppression of OA matrix synthesis genes. In vitro, we characterized the transcriptome and protein from control and NG2/CSPG4 knockout cells and found significant dysregulation of the ERK 1/2 signaling axis. To characterize the mechanobiological response of NG2/CSPG4, we applied mechanical loads on cell-agarose-collagen scaffolds using a compression bioreactor and illustrate that NG2/CSPG4 knockout cells fail to mechanically activate ERK 1/2 and are associated with changes in the expression of the same key OA biomarkers measured in vivo. Together, these findings implicate NG2/CSPG4 in the mechanical homeostasis of TMJ cartilage and in the progression of degenerative arthropathies including OA. Keywords: temporomandibuar joint, mechanotranduction, cartilage, cell-matrix interactions, NG2/CSPG4, ERK signaling, chondrocyte, arthopathy Introduction Osteoarthritis (OA) is the most common pathology affecting the temporomandibular joint (TMJ) and is characterized by extracellular matrix and cartilage degradation that is associated with limited joint mobility, pain, and diminished quality of life ([37]1). While the etiopathogenesis of TMJ OA has yet to be fully resolved, there is consensus that a change in mechanical homeostasis is a strong initiating condition of cartilage degeneration. Mechanical forces on a tissue are transduced to the cell through the extracellular matrix ([38]2). The pericellular matrix is a division of the extra cellular matrix that encompasses the cell. In mandibular cartilage, the pericellular matrix is composed of collagen IV, laminin, perlecan, and collagen VI ([39]3). Collagen VI is a micro- fibrillar collagen that regulates the material properties of the pericellular matrix ([40]4). During TMJ OA, there are changes in the distribution, organization, and quantity of collagen VI, indicating that the mechanical microenvironment of the cell may be affected during the progression of TMJ OA ([41]5–[42]7). Neuron-Glial antigen 2 (AN2 in mice; Chondroitin sulfate proteoglycan 4 in humans, NG2/CSPG4) is an N-linked type I transmembrane glycoprotein with chondroitin sulfate proteoglycan components on the ectodomain. NG2/CSPG4 binds with collagen VI ([43]8, [44]9). The full-length core protein without chondroitin sulfate chains is represented by a 300 kDa band. Ectodomain shedding occurs through proteolytic processing by MMP13 ([45]10), MMP14 ([46]11), MMP9 ([47]12), and/or ADAM10 ([48]13), generating a truncated, cell-associated/membrane-tethered fragment between 260 and 275 kDa ([49]11, [50]13, [51]14). MMP13 is significantly elevated after mechanical loading on mandibular fibrochondrocytes ([52]15, [53]16), suggesting that NG2/CSPG4 ectodomain shedding is contextually linked with changes in mechanical homeostasis. NG2/CSPG4 ectodomain shedding is believed to initiate the release of the intercellular domain through an α-secretase mediated process of regulated intramembrane proteolysis ([54]17). This extracellular surface shedding and intracellular processing potentiates both inside-out and outside- in signaling ([55]18, [56]19). In mandibular condylar cartilage, NG2/CSPG4 binds with pericellular type VI collagen in articular/superficial layer cells. During the progression of TMJ OA, NG2/CSPG4 translocates from the cell membrane to the cytosol ([57]20). In other cell types, NG2/CSPG4 internalization occurs at focal adhesions and is associated with cell-matrix dynamics and motility ([58]19, [59]21). The mechanism and role of NG2/CSPG4 internalization and proteolytic processing in cartilage homeostasis has not been mechanistically resolved. However, NG2/CSPG4 has been implicated in collagen binding in limb OA chondrocytes ([60]22) and chondrosarcomas ([61]23). Together, these data indicate that NG2/CSPG4 signaling may be important for regulating the health and disease of cartilage in the TMJ and elsewhere. NG2/CSPG4 regulates a wide variety of cellular processes including migration, survival, apoptosis, and differentiation, due in part to multivalent nature of the inside-out and outside- in signaling modalities of the molecule ([62]18, [63]19). The intracellular C-terminal domain has acceptor sites for the extracellular signal-regulated kinases (ERK) 1/2 and protein kinase C-alpha (PKC-α) ([64]24), binding domains for multi-PZD domain protein 1 (MUPP-1) ([65]25), interactions with integrins ([66]9), and regulates mTORC1 ([67]18, [68]19). NG2/CSPG4 mediated regulation of the ERK 1/2 signaling cascade is particularly important for studies on mechanotransduction. The ERK 1/2 signaling axis is a critical regulator of the mechanobiological response in chondrocytes ([69]26–[70]31). Impact injury to cartilage activates ERK 1/2 signaling axis through MAPK, promoting key OA proteases including MMP13 and ADAMTS5 ([71]32). ERK 1/2 signaling is activated by the release of FGF2 from the pericellular matrix ([72]33), a growth factor that binds to NG2/CSPG4 ([73]34). ERK 1/2 is further implicated in the progression of OA by regulating the differentiation cascade of chondrocytes ([74]35, [75]36), hypertrophy ([76]37), osteoblast differentiation ([77]38), and regulating the fate of mesenchymal stem cells in response to mechanical loading ([78]39). Together, these studies illustrate that NG2/CSPG4 may be a key mechanotransductive signaling hub in mandibular fibrochondrocytes during TMJ health and disease. NG2/CSPG4 influences ERK 1/2 signaling in a wide range of cells, and chondrocyte differentiation pathways in particular. ERK 1/2 can modulate key OA proteases in a mechanical loading dependent manner in mandibular and limb chondrocytes. Here we address a key gap in knowledge regarding if NG2/CSPG4 is a significant regulator of TMJ health and disease in vivo and if mechanical activation of ERK 1/2 signaling occurs in an NG2/CSPG4 dependent manner. We hypothesize that NG2/CSPG4 regulates cartilage degeneration during degenerative arthropathy and the mechanobiological activation ERK 1/2. Here we mechanistically interrogate the role of NG2/CSPG4 in the homeostasis of mandibular condylar cartilage using an in vivo preclinical surgical instability murine model of TMJ OA and an in vitro compression bioreactor mechanical loading model. Results TMJ OA is associated with loss of the NG2/CSGP4 ectodomain and an internalized intracellular domain To characterize how NG2/CSPG4 is altered in during the progression of TMJ OA in vivo, we analyzed mandibular condyles from wild type (c57 BL/6J) mice at 2- and 4-weeks after discectomy, and compared them with age matched non-surgical controls with immunofluorescence using an antibody against the NG2/CSPG4 ectodomain. These data illustrate that NG2/CSPG4 levels are decreased in the mandibular condylar cartilage following discectomy ([79]Figures 1A–[80]C). The NG2/CSGP4 ectodomain can be present as a 300 kDa full-length protein, or a 260/275 kDa shed, membrane tethered fragment that lacks the intracellular domain. To confirm that the 260/275 kDa fragment lacks the intracellular domain, we performed western blot analysis on protein samples from primary mandibular fibrochondrocytes immunolabeled using a polyclonal antibody labeled against the ectodomain and a monoclonal antibody labeled against the intracellular domain. These data illustrate that the monoclonal antibody only labeled the 300 kDa full-length fragment, confirming that the 260/275 kDa fragment lacks the intracellular domain ([81]Figures 1D,[82]E). To characterize NG2/CSPG4 during the progression of TMJ OA, we analyzed isolated mandibular condyles by western blot from 2-, 4-, and 8-week TMJ OA samples and compared them with non-surgical and sham controls. Western blot analysis illustrates that the sham controls have significantly more full length and shed, membrane-tethered NG2/CSPG4 fragments than the non-surgical control. In the TMJ OA samples, there is a non-significant reduction in full length NG2/CSPG4 at all stages and a significant reduction in the shed, membrane tethered fragment at 4-weeks after discectomy ([83]Figures 1F–[84]H). RT-qPCR illustrates that NG2/CSPG4 gene expression is significantly increased at 2-weeks after discectomy but returns to the level of the non-surgical control shortly after ([85]Figure 1I). We had previously illustrated TMJ OA is associated with elevated levels of cytosolic NG2/CSPG4 in articular layer cells of the mandibular condylar cartilage using a polyclonal antibody raised against the full length protein ([86]20). Using immunohistochemistry with an antibody specific to the NG2/CSPG4 intracellular domain, we illustrate that TMJ OA is associated with high levels of the NG2/CSPG4 intracellular domain in the cytosol in both a preclinical murine model ([87]Figures 1J,[88]K) and in a sample from a human TMJ OA patient undergoing total joint replacement ([89]Figure 1L). Together, these data illustrate that in vivo TMJ OA is associated with loss of NG2/CSPG4 ectodomain fragments and an increase in the presence of the NG2/CSPG4 intracellular domain in the cytosol. These data also indicate that NG2/CSPG4 localized to the cytosol is characteristic of clinical TMJ dysfunction. FIGURE 1. FIGURE 1 [90]Open in a new tab Early stage TMJ OA is associated with a loss of the NG2/CSPG4 ectodomain and internalization of the intercellular domain. (A) NG2/CSPG4 immunofluorescent staining from TMJ tissue in a non-surgical control mouse. (B) NG2/CSPG4 immunofluorescent staining from TMJ OA tissue 2-weeks after discectomy. (C) NG2/CSPG4 immunofluorescent staining from TMJ OA tissue 4-weeks after discectomy. (D) Western blot analysis from protein lysate extracted from primary mandibular fibrochondrocytes, separated on a 12% SDS PAGE gel, and immunolabeled using a polyclonal antibody raised against the NG2/CSPG4 ectodomain. (E) Western blot analysis from the same protein sample, run on the same gel, and immunolabeled using a monoclonal antibody raised against the NG2/CSPG4 intracellular domain. Note two distinct bands at 300–260 kDa with the polyclonal antibody and only one band representing the full-length protein on the monoclonal intracellular domain antibody. (F) Western blot from mouse tissue 2-, 4-, and 8-weeks after discectomy compared to non-surgical and sham controls immunolabeled with antibodies against NG2/CSPG4. (G,H) Quantification of the western blots from the full-length 300 kDa (G) and 260/275 kDa shed, membrane tethered (H) NG2/CSPG4 fragments. Note a significant reduction in the shed, membrane-tethered fragment a 4-weeks post-operative compared to the non-surgical control and a significant increase in full length and shed, membrane-tethered NG2 in the sham. N = 4/experimental group. (I) RT-qPCR on mRNA isolated from murine mandibular condylar cartilage illustrating that the expression level of NG2/CSPG4 is significantly elevated 2-weeks post-operative but is not significantly different from the non-surgical control through the rest of the progression of TMJ OA. N = 6/experimental group. (J) The spatial distribution of the NG2/CSPG4 intracellular domain (NG2icd) in an articular/superficial layer mandibular fibrochondrocyte from non-surgical control TMJ tissue. (K) The spatial distribution of NG2icd in articular/superficial layer mandibular fibrochondrocytes 8-weeks after discectomy. (L) The spatial distribution of the NG2icd in an articular/superficial layer mandibular fibrochondrocyte from a human patient with TMJ OA. *p < 0.05. NG2/CSPG4 knockout mice have higher levels of cartilage degeneration during early stage TMJ OA and have increased levels of key OA biomarkers This study uses NG2/CSPG4 global knockout mice backcrossed to a c57 BL6/J line ([91]Figure 2A). Global expression of NG2/CSPG4 was assessed by whole mount staining for a LacZ reporter inserted in a promoter driven cassette. This staining illustrates that NG2/CSPG4 is present in both primary and secondary cartilages in the craniofacial skeleton, including Meckel’s and mandibular condylar cartilage respectively ([92]Figures 2B,[93]C). NG2/CSGP4 knockout mice are viable through skeletal maturity. Loss of NG2/CSPG4 protein in LacZ positive tissues was confirmed in both TMJ tissue and primary cells at the gene, mRNA, and protein level ([94]Figures 2D–[95]I). To determine if NG2/CSPG4 regulates the rate of cartilage degeneration, TMJ OA was induced in wild-type control and NG2/CSPG4 knockout mice and the tissue was stained by safranin-o/fast green. Cartilage degeneration was graded using a modified Mankin Scoring matrix [see references ([96]15,