Reduced cartilage matrix stiffness in temporomandibular joint osteoarthritis impairs the functions of superficial zone chondrocytes via downregulation of CREB5-PLPP3 signaling

Abstract

Temporomandibular joint (TMJ) plays critical roles in the movement of mandible. TMJ osteoarthritis (TMJOA) leads to pain and limited jaw function. Histologically, TMJOA causes cartilage degradation and a reduction in extracellular matrix (ECM) stiffness. The superficial zone chondrocytes (SZC) contribute in the regeneration of the condylar fibrocartilage in TMJ, while their responses to the softened ECM remains unclear. Here, we showed that the ECM stiffness was decreased in the superficial zone cartilage of TMJOA patients and rat models. Single-cell RNA sequencing demonstrated the diminished phospholipid phosphatase 3 (Plpp3) expression, impaired migration, and ECM secretion, as well as the down-regulated PI3K-AKT pathway of SZC in rat TMJOA. Such alternations were also revealed by mRNA sequencing of SZC cultured on the softened ECM. Further studies disclosed that reduced ECM stiffness induced decreased PLPP3 on the endoplasmic reticulum (ER), which inhibited mitochondrial fission and respiration via increasing phosphatidic acid (PA) in the mitochondria. Meanwhile, deactivated PI3K-AKT pathway reduced the intra-nuclear translocation of transcriptional factor Cyclic AMP responsive element binding protein 5 (CREB5), which limited PLPP3 expression. Overexpression of PLPP3 alleviated the functional damage of SZC in vitro and the cartilage ECM degradation in vivo. This work displayed the functional impairment of SZC on the softened ECM and the underlying mechanism, as well as suggested PLPP3 as a potential target in the regenerative treatments for TMJOA.

Supplementary Information

The online version contains supplementary material available at 10.1186/s43556-026-00435-2.

Introduction

The temporomandibular joint (TMJ) is the only movable joint in the mature orofacial area, playing critical roles in phonation, chewing, and breathing. Multiple pathologies, including malocclusion and trauma, can impact the intricate anatomy of the joint and lead to TMJ disorders (TMD), which has a prevalence of 29.5% in the global population [1], and an annual economic burden of more than $4 billion worldwide [2–4]. This condition can progress to TMJ osteoarthritis (TMJOA) because of factors such as inflammation and mechanical overload [5], which affects 1.9%–3.2% of adult/older population [6], predominantly women [7]. TMJOA is characterized by synovitis, degeneration of the mandibular condyle cartilage, and loss of subchondral bone, which cause phonic disturbances, pain, and limited jaw movement [8].

The onset of OA is associated with the degradation of cartilage extracellular matrix (ECM) starting from the superficial area [9, 10]. The cartilage ECM is composed mainly of collagen fibrils and proteoglycans, which are enmeshed together to form a stable load-bearing network [11, 12]. Highly charged proteoglycans bind to water to form high osmotic pressure, while the collagen network provides the restraining forces to maintain such pressure, thus contributing to the hydration and biomechanical properties of the cartilage [13]. In OA pathogenesis, the accumulation of matrix metalloproteins (MMPs) and proteoglycanases leads to disorganization of the collagen network and the of proteoglycans, further resulting in impaired cartilage integrity and decreased ECM stiffness [10, 14, 15]. Through multiple mechano-sensing and mechano-transducing pathways, ECM stiffness regulates biological processes such as proliferation, migration, secretion, and differentiation [16, 17]. In TMJOA, softening of the ECM promotes the adipogenic differentiation of SOX9 + and COL10 + chondrocytes, however, the underlying mechanisms remain unclear [18].

The surface of the mandibular condyle is covered by fibrocartilage, which adapts to functional movements and loading while possessing life-long regenerative capacity [19, 20]. Many studies have demonstrated the regenerative potential of Prg4-expressing surface layer chondrocytes in knee cartilage [21, 22]. In the condyle fibrocartilage, superficial zone chondrocytes (SZC) contain Prg4 + fibrocartilage stem cells (FCSC), which possess multipotential differentiation, cartilage-repairing, and bone-regenerative abilities [23–26]. They are partially responsible for TMJOA formation in the prenatal stage, as well as maintaining the homeostatic microenvironment in the fibrocartilage after birth [23, 27]. Multiple signaling pathways, including β-catenin, WNT, NOTCH, and TNF-α/NFκB pathways, can regulate the phenotypes and fates of SZC, hence participating in TMJOA pathogenesis and progression [27–30]. However, few studies have investigated the effects of biomechanical cues on SZC. Therefore, our work aimed to elucidate the responses of these cells to decreased ECM stiffness in TMJOA and the underlying mechanism, to identify potential mechanically sensitive targets for TMJOA management.

Phospholipid phosphatase 3 (PLPP3) belongs to the phosphatase/phosphotransferase family, whose function is to dephosphorylate various lipid phosphates and thus participate in phospholipid metabolism and signal transduction [31]. As an integral membrane protein, PLPP3 can locate on plasma membranes and/or internal membranes, depending on the cell type. On the cell membrane, PLPP3 can degrade extracellular lysophosphatidate (LPA) and sphigosine 1-phosphate (S1P), whereas it dephosphorylates phosphatidic acid (PA) and ceramide 1-phosphate (C1P) on inner organelle membranes [32–34]. Intriguingly, PLPP3 is mechanosensitive and contributes to the mechano-transduction mechanism of endothelial cells under hemodynamic forces [35]. Nevertheless, the role of PLPP3 in SZC remains unexplored, which encouraged us to further elucidate its role in the altered mechano-environment of TMJOA.

In the present study, decreased ECM stiffness was detected in the superficial cartilage of patients and animal models of TMJOA. Single-cell RNA sequencing of the condylar cartilage revealed enriched expression of PLPP3 in SZC. Combined with mRNA sequencing of SZC cultured on stiffness-modified ECM, we demonstrated that softening ECM impaired the PI3K-AKT pathway, and limited the intra-nuclear translocation of Cyclic AMP-Responsive Element-Binding Protein 5 (CREB5), resulting in the down-regulation of PLPP3 expression on the ER of SZC. Such decreased PLPP3 expression induced increased PA levels in the ER and mitochondria, which further led to mitochondrial elongation and dysfunction, resulting in reduced migration and ECM secretion of SZC. PLPP3 overexpression effectively ameliorated TMJOA cartilage degradation. These findings provide new insights into the pathogenesis and treatment targets of TMJOA.

Results

ECM stiffness is reduced in the condylar cartilage superficial zone of TMJOA

To assess the change in ECM stiffness in TMJOA, we procured condylar cartilage tissue from patients diagnosed with OA, osteochondroma (OC), condyle fracture (FR), or benign condylar hypertrophy (CH). Macroscopic views revealed uneven cartilages from the OA patients, with rough, eroded OA parts and relatively smooth normal control (NC) parts (Fig. 1a). By contrast, cartilages from the OC, FR and CH patients were continuous and glossy, similar to those from the NC parts. Hematoxylin–eosin (HE) staining revealed the irregular alignment of cells and ECM in OA cartilage, compared with that in the NC parts and from OC, FR and CH samples.

Fig. 1.

Tissue stiffness is reduced in the condylar cartilage superficial zone of TMJOA. a Macroscopic photos and HE staining of the condylar cartilage of patients. The age, gender and diagnosis of the patients were provided above the photos. Number, age in years, F, female, M, male; OA, osteoarthritis, OC, osteochondroma, FR, condylar fracture, CH, benign condylar hypertrophy. Scale bar, black, 2 mm, white, 100 µm. Dashed lines indicate the boarders between SZ and deeper layers. b Young’s modulus of condylar cartilage SZ of patients. c Young’s modulus of condylar cartilage SZ of control and UAC rats

A nano-indentor was used to detect ECM stiffness in the surface layer of the condylar cartilage samples. In the OA samples, NC parts of cartilage possessed the Young’s modulus of 193.49 ± 35.42 kPa to 317.32 ± 112.57 kPa, higher than the modulus of OA parts that ranged from 82.16 ± 9.42 kPa to 128.89 ± 6.11 kPa (Fig. 1b). The stiffness of OC and FR cartilages were close to that of NC parts, with the modulus of 241.97 ± 27.17 kPa to 259.70 ± 42.04 kPa. The CH cartilage showed the highest modulus of 398.03 ± 22.50 kPa. Moreover, we established the unilateral anterior crossbite (UAC) model to mimic TMJOA in 8-week-old rats. After 4 weeks, ECM stiffness in the surface layer of rat condyle cartilage from both the control (CTL) and UAC groups was measured through nanoindentation. The SZ of the CTL group maintained a modulus of 203.67 ± 35.81 kPa, which was higher than the 128.98 ± 17.38 kPa of the UAC group (Fig. 1c). Hence, our data demonstrates decreased ECM stiffness in the surface layer of condylar cartilage in TMJOA patients and in rat UAC model.

Single-cell RNA sequencing reveals the cell cluster composition of the condylar cartilage and the functional impairment of SZC in TMJOA

To explore the distinct features of SZC among other chondrocytes, as well as their altered functions in TMJOA, we performed single-cell RNA sequencing (scRNA-seq) on isolated condylar chondrocytes from rats in the CTL and UAC groups (Fig. 2a). In normal condylar cartilage, chondrocytes fell in 6 clusters according to their marker genes, as shown in the 2-dimensional and 3-dimensional UMAPs (Fig. 2b and Fig. S1) and feature plots (Fig. S2a). Histologically, the condyle cartilage was divided into superficial zone (SZ), polymorphic zone (PZ), flattened chondrocyte zone (FZ), and hypertrophic chondrocyte zone (HZ) (Fig. 2a) [23, 27]. As shown in the expression heatmap (Fig. 2c), Cluster C0 and C1 displayed similar marker genes such as Col1a1, Col2a1, Bgn and Clec3a, which matched the ECM components in the FZ [36–38]; thus, these clusters were named flattened chondrocyte zone chondrocyte 1 and 2 (FC1, FC2), respectively. Cluster C2 cells were regarded as polymorphic zone chondrocytes (PMC) since they expressed genes such as Dkk3, Tnc and Postn, which were enriched in the PZ [27, 39]. Produced in cluster C3, Prg4 and fibronectin (Fn1) are ECM components located especially in the SZ of articular cartilage, as displayed by immunohistological staining (Fig. S2b) [22, 40]. Therefore, cluster C3 cells were designated SZC. Intriguingly, mechanosensitive Plpp3 [35] was shown to be expressed especially by SZC (Fig. 2c and S2). The hypertrophic marker Col10a1 was expressed especially in cluster C4, whereas the proliferative markers Mki67 and Top2a were produced in C5 cells. As a result, these two clusters were named hypertrophic chondrocytes (HC) and proliferative chondrocytes (PC), while locating in the HZ and PZ, respectively [41].

Fig. 2.

ScRNA-seq reveals limited functions of SZC in the condylar cartilage of TMJOA. a Schematic diagram of rat condyle cartilage. CTL: control group; UAC: unilateral crossbite group. The shapes of chondrocytes were adopted from BioRender. https://BioRender.com/8cm30e1b 2-dimensional UMAP plot showing the CTL chondrocyte clusters of condyle cartilage. c Heatmap of marker genes in each CTL chondrocyte cluster. d Radar plot of functional enrichment in CTL chondrocyte clusters. e Pseudotime trajectory of CTL chondrocyte clusters. f UMAP plot showing the matched chondrocyte clusters in CTL and UAC group. g Cell cluster composition in CTL and UAC group. h GO and KEGG enrichment of down-regulated genes of SZC in UAC group

Functional enrichment elucidated that SZC featured in ECM organization, migration, regeneration, growth, and growth factor binding, whereas FC1, PMC, HC and PC were associated with higher chondrocyte differentiation, glycolysis, endochondral bone morphogenesis and proliferation, respectively (Fig. 2d). Peudotime analysis unveiled that SZC were located at the start point of the developmental trajectory, with the potential to differentiate into PMC and then into two branches: PC and FC2-FC1-HC (Fig. 2e). These results highlighted the contribution of SZC to the regenerative potential of the condyle cartilage.

In the UAC group, condylar chondrocytes remained in the 6 clusters, yet the proportions of the clusters were altered (Fig. 2f, g). The proportions of FC2 and PC shrank significantly, whereas those of HC and PMC increased, indicating the potentially limited transition of PMC to PC and FC2, as well as the boosted FC-HC differentiation, which aggravated OA [42]. As the proportion of SZC changed insignificantly, we focused on the differentially expressed genes (DEGs) of SZC in the two groups. The significantly enriched DEGs in the control group SZC included Plpp3, Aoc3, Gsta1, Ccn3 and Abi3bp, as displayed in the volcano plot and heatmap (Fig. S3a-b). According to the results of the Gene Ontology (GO) and Kyoto Encyclopedia of Genes and Genomes (KEGG) enrichment analysis of the DEGs, SZC displayed compromised functions, including ECM organization, response to mechanical stimulus, and cell migration (Fig. 2h). Therefore, TMJOA causes functional impairment of SZC.

Softened ECM leads to impaired migration, ECM secretion and PLPP3 expression of SZC

To explore the underlying mechanisms of the functional alternation in SZC caused by decreased ECM stiffness, we first isolated primary rat SZC and verified their surface markers and trilineage differentiation potential following published methods [23, 25] (Fig. S4). We subsequently cultured the SZC on stiff and soft polyacrylamide (PAM) hydrogels, with stiffness of 412.27 ± 2.42 kPa and 129.90 ± 1.93 kPa respectively, to simulate the alteration of ECM stiffness (Fig. S5). Phalloidin staining showed that cells on the soft ECM had larger spreading area, with extended F-actin organized along the longer longitudinal axis, than on the stiff one (Fig. 3a). Moreover, the migration of SZC was impaired by the soft ECM, and the secretion of fibronectin and PRG4 decreased (Fig. 3b-d). Furthermore, mRNA sequencing revealed the DEGs associated with the effects of SZCs on stiffness-altered ECM. As demonstrated in the heatmap, most of the DEGs were related to ECM components and organization in condylar cartilage, including Fbn1, Col1a1, Fndc1, and Fmod [26, 36, 38, 39, 43] (Fig. 3e). The SZC genes, whose expression was downregulated both under UAC, as determined by scRNA-seq, and on the softened ECM, as determined by mRNA-seq, were cross-verified (Fig. 3f). Among these genes, Plpp3 expression significantly decreased with decreasing ECM stiffness, both in vivo and in vitro (Fig. 3g-i, Fig. S6). Intriguingly, the scRNA-seq data revealed that Plpp3 was expressed especially in SZC (Figs. 2c, 3g, S2a, S2b), suggesting its potential key regulatory role in their functions. Moreover, the diminished expression of PLPP3 in the SZ was confirmed with immunofluorescence (IF) staining of clinical TMJOA samples, as well as with fluorescence in situ hybridization (FISH) in UAC rat samples, comparing with control tissues (Fig. 3 j-m). Collectively, these results revealed that the impairment of migration, ECM secretion and PLPP3 expression in the SZC was caused by the softening of the ECM in TMJOA.

Fig. 3.

Softened ECM leads to impaired function and PLPP3 expression in SZC. a Phalloidin staining shows the organization of F-actin in SZC. b The trajectory and c speed of SZC during 12 h. d Western blot showing ECM secretion of SZC. e Heatmap of differentially expressed genes of SZC on stiff and soft ECM. f Venn plot of down-regulated genes of SZC in UAC and on soft ECM. g Violin plot of Plpp3 expression in chondrocyte clusters of scRNA-seq. h-i mRNA and protein expression of PLPP3 in SZC. j Immunofluorescence staining of PLPP3 in the condyle cartilage of patient samples and (k) the proportion of positive cells (n = 3). l FISH assay of Plpp3 in the condyle cartilage of rats and the (m) proportion of positive cells (n = 6). Scale bars, 50 µm. White dashed lines indicate the boarders between SZ and deeper layers. **, p < 0.01; ***, p < 0.001

Down-regulated PLPP3 inhibits mitochondrial fission and respiration in SZC

Next, we aimed to elucidate the role of PLPP3 in regulating SZC’s cellular responses to a softened ECM. Immunofluorescence staining showed the co-localization of PLPP3 and the endoplasmic reticulum (ER) marker Glucose-Regulated Protein 78 (GRP78), suggesting that PLPP3 is located in the ER of SZC, instead of in the Golgi apparatus or mitochondrion, with the markers of cis-Golgi matrix protein (GM130) and ATP synthase F1 subunit alpha (ATP5A1), respectively (Fig. 4a). We then overexpressed PLPP3 in SZC with lentiviruses, and confirmed its elevated expression in the ER, under both stiff and soft ECM conditions (Fig. 4b-d). As reported, PLPP3 degrades PA on the ER, which interchanges with mitochondrial PA and subsequently mediates cellular oxidative phosphorylation [34]. Therefore, concentrations of PA in the ER, mitochondria and whole cells were evaluated by enzyme-linked immunosorbent assay (ELISA). On the softened ECM compared with the stiff one, SZC presented higher PA in both the ER and mitochondria, which matched the decreased PLPP3 expression. PLPP3 overexpression significantly decreased the PA in the two organelles, whereas the total cellular PA remained unchanged (Fig. 4e). The accumulation of PA promotes mitochondrial fusion and inhibits mitochondrial fission, leading to mitochondrial elongation and impaired respiration [44, 45]. Accordingly, we assessed the morphology of the mitochondria in SZC. Mito-Tracker illustrated the elongated mitochondria of SZC on the soft ECM, with a relatively large area, perimeter, aspect ratio and branch length. By contrast, overexpression of PLPP3 helped maintain mitochondrial morphology on the softened ECM. (Fig. 4f, g). Transmission electron microscope (TEM) revealed similar changes in mitochondrial morphology among the groups (Fig. 4h). Fluorometric analysis revealed an increase in the oxygen consumption rate (OCR) in PLPP3-high cells (Fig. 4i, j). Moreover, we measured the adenosine triphosphate (ATP) concentration in the SZC. The cells cultured on the soft ECM exhibited less ATP concentrations, whereas the PLPP3-overexpresing ones displayed higher ATP concentrations (Fig. 4k). These results indicated that PLPP3 contributed to the downregulation of PA-induced mitochondrial elongation and hence increased mitochondrial respiration in the SZC. Consequently, the constrained SZC cell migration (Fig. 4l, m), beside with the restricted secretion of fibronectin and PRG4 on the soft ECM (Fig. 4n), were restored by the PLPP3 overexpression.

Fig. 4.

PLPP3 regulates mitochondrial fission and respiration in SZC. a Immunofluorescence staining of PLPP3 and ER marker GRP78, Golgi apparatus marker GM130 and mitochondria marker ATP5A1 in SZC (Scale bars, 20 µm). b-c The mRNA and protein expression of PLPP3 in control and PLPP3-overexpressed SZC. d PLPP3 expression in ER, mitochondria, and whole cells. GAPDH, GRP78 and ATP5A1 served as loading control for whole-cell, ER and mitochondria protein, respectively. e ELISA assay of PA in ER, mitochondria, and whole cells. f Mito-tracker showing the shapes of mitochondria (Scale bar, 2 µm) and (g) the according morphological analysis (n = 200). h TEM displaying the morphology of mitochondria (Scale bar, 0.5 µm). i Fluorometric assay detecting oxygen concentration in extracellular space of SZC and (j) the accordingly calculated OCR. k ATP concentration in SZC. l The trajectory and (m) speed of SZC during 12 h. n Western blot showing ECM secretion of SZC. NC, Normal Control; OE, Plpp3 overexpression *, p < 0.05; **, p < 0.01; ***, p < 0.001

PI3K-AKT-CREB5 axis regulates PLPP3 expression in SZC

Furthermore, we explored the mechanism through which ECM stiffness affectes the expression of PLPP3 in SZC. With respect to the genes downregulated in the SZCs on the softened ECM as well as in the UAC cartilage, we highlighted the Phosphatidylinositol 3-kinase (PI3K)- RAC serine/threonine-protein kinase (AKT) pathway, which exhibited significant functional enrichment (Figs. 5a, 2h). Cells cultured on the soft ECM displayed deactivation of the PI3K-AKT pathway, as indicated by decreased phosphorylation of PI3K and AKT (Fig. 5b). The specific AKT activator, SC-79 [46], with the concentration of 4 µg/mL (MCE, CAT: HY-18749, LOT: 517,547), restored the expression of p-AKT and promoted the fibronectin and PRG4 expression in SZC in 48 h (Fig. 5c). The arrested cell migration on the soft ECM was also rescued by SC-79 (Fig. 5d-e). Interestingly, single-cell regulatory network inference and clustering (SCENIC) analysis demonstrated the exclusive regulatory role of the transcription factor CREB5 in SZC (Fig. 5f). Moreover, according to the GSEA database (GSEA, map04151), CREB5 appeared to be a member of the PI3K-AKT pathway, with its intranuclear translocation being activated by phosphorylated AKT [47]. Hence, we speculated that activation of the PI3K-AKT pathway could improve the regulatory role of CREB5. Immunofluorescence staining (Fig. 5g) and the extraction of cytosolic and nuclear proteins (Fig. 5h), revealed that the soft ECM could cause the extranuclear localization of CREB5 in the SZC, whereas SC-79 could enhance its intranuclear transportation. Moreover, we aimed to elucidate the role of CREB5 in PLPP3 expression. Chromatin immunoprecipitation followed by quantitative PCR (ChIP-qPCR) unveiled the binding of CREB5 to the Plpp3 promoter DNA sequence (Fig. 5i). Using a dual-luciferase assay, we detected that CREB5 could increase the promoter activity of Plpp3 (Fig. 5j). Furthermore, CREB5 overexpression increased the mRNA and protein levels of PLPP3 in SZC (Fig. 5k-m). Therefore, our work displayed that the softened ECM deactivated the PI3K-AKT pathway, and inhibited the positive regulatatory effect of CREB5 on the Plpp3 promoter, consequently causing a decrease in PLPP3 expression in SZC.

Fig. 5.

PI3K-AKT-CREB5 axis regulates PLPP3 expression in SZC. a GO and KEGG enrichment of down-regulated genes of SZC on soft ECM. b Western blot of PI3K-AKT pathway in SZC. c Western blot of AKT phosphorylation and ECM secretion in SZC after SC-79 treatment. d The trajectory and (e) speed of SZC during 12 h. f SCENIC analysis showing TF regulon activities in the condylar chondrocyte clusters. g Immunofluorescence staining of CREB5 in SZC (Scale bar, 10 µm). h Western blot of cytosolic and nuclear CREB5 in SZC. i ChIP-qPCR showing the binding of CREB5 to Plpp3 promoter. j Dual luciferase assay displaying the regulation of CREB5 on Plpp3 promoter activity. k The mRNA expression of Creb5 and (l) Plpp3 after Creb5 overexpression. m The protein expression of CREB5 and PLPP3 after Creb5 overexpression. NC, Normal Control; OE, Creb5 overexpression. **, p < 0.01; ***, p < 0.001

Overexpression of PLPP3 alleviates the cartilage degradation in TMJOA

Finally, we investigated the role of PLPP3 overexpression in the condylar cartilage of rats with TMJOA. After one week of UAC establishment, adeno-associated virus (AAV) 9 with Plpp3-overexpressing vectors (AAV9_Plpp3) or control vectors (AAV9_NC) was injected into the TMJ articular cavity. Three weeks later, the joints were collected for histological analysis. The adequate effectiveness of viral transfection was assessed using the ZsGreen signal on the AAV (Fig. S7). Compared with those in the CTL group, HE staining showed the decreased total and SZ thicknesses of condylar cartilages in the UAC and UAC + AAV9_NC groups (Fig. 6a, f, g), while safranin O-fast green (SO) staining displayed the loss of proteoglycans, and elevated Modified Mankin Scores in these two groups (Fig. 6b, h, i). Injection of AAV9_Plpp3 effectively rescued Plpp3 expression in the SZ, which was inhibited by UAC, and alleviated the histological changes in the cartilage (Fig. 6c, j). According to the immunohistochemical staining, fibronectin and lubricin (encoded by Prg4 gene) were down-regulated from normal in the UAC and UAC + AAV9_NC groups, which could be partially restored through the overexpression of Plpp3 (Fig. 6d, e, k, l). Collectively, the results of this work demonstrated that the softened ECM in TMJOA impaired PLPP3 expression in the SZC via deactivating the PI3K-AKT-CREB5 pathway. It was highlighted that PLPP3 boosted the migration and ECM secretion by reducing PA-induced mitochondrial elongation, which contributed in alleviating cartilage damage in TMJOA (Fig. 6m).

Fig. 6.

Overexpression of PLPP3 alleviates the cartilage degradation in TMJOA. a HE and (b) SO staining showing the histological changes of rat condylar cartilage under UAC and the treatment of AAVs. c FISH assay displaying the expression of Plpp3 in the condylar cartilage. d Immunohistology staining of fibronectin and (e) lubricin in the condylar cartilage of different groups. f The thickness of whole condylar cartilage and (g) the SZ. (h) The percentage of SO-positive area and (i) modified Mankin score from SO staining of condylar cartilage in different groups. j The percentage of Plpp3 positive cells in condylar cartilage. k The positive area of fibronectin and (l) lubricin in condylar cartilage of different groups. n = 6. Scale bars, 50 µm. Dashed lines indicate the boarders between SZ and deeper layers. *, p < 0.05; **, p < 0.01; ***, p < 0.001 (m) Schematic graphics demonstrating the role of PLPP3 in alleviating the impairment of SZC from softened ECM in TMJOA. The shapes of DNA, ER, and mitochondria were adopted from BioRender. https://BioRender.com/8cm30e1

Discussion

TMJOA induces pain and functional limitations in the jaw, greatly impairing the quality of life of patients. Regular conservative treatments such as the application of nonsteroidal Anti-inflammatory drugs (NSAIDs) can temporarily relieve discomfort, but contribute little in promoting TMJ tissue regeneration [5]. The condyle fibrocartilage possesses life-long reconstructive potential, which is partially maintained by the regenerative SZC [19, 23]. However, the functional impairments of SZC in the pathological mechanical environment of TMJOA remains to be elucidated. Here, we showed that the mechanosensitive PLPP3 was expressed in SZC. The softening of the ECM in TMJOA deactivated the PI3K-AKT-CREB5 pathway in SZC, causing decreased expression of PLPP3 in the ER and compromising mitochondrial fission. The overexpression of PLPP3 could enhance the cell migration and ECM secretion of SZC and alleviate cartilage damage in TMJOA. These findings indicated the role of PLPP3 in maintaining the regenerative potential of SZC in TMJOA.

The onset of OA features degradation of cartilage ECM. The damage to the collagen-proteoglycan meshwork caused by MMPs and proteoglycanases reduces the water-constraining capacity of the ECM, leading to altered mechanical properties of the cartilage [10, 14]. Decreased ECM stiffness was revealed in the knee cartilages of both humans and mice OA [15, 48, 49]. In TMJOA, the loss of proteoglycans also results in the declined elastic modulus of condylar fibrocartilage [18], which shows the similar trend as our data. The altered pressure, stretch and shear stress during the changes of ECM stiffness are sensed by cells through surface proteins such as integrins, ion channels, G-protein-coupled receptors (GPCRs) and adhesion proteins, which regulate cell morphology and function [50]. In fibrosis, the stiffened ECM promotes the proliferation and ECM generation of fibroblasts through pathways such as Hippo, WNT/β-catenin and FAK, causing a self-amplifying loop for fibroblast activation and boosts fibrogenesis [17, 51]. In tumors, cancer cells can sense changes in ECM stiffness, and adapt their cytoskeleton arrangement to migration or metastasis [52]. Importantly, the mechano-signaling pathways in chondrocytes involve PI3K-AKT, MAPK, SMAD and JNK, which mediate the regulation of cell survival, proliferation, migration, as well as the anabolic or catabolic processes through the mechanical properties of the ECM [53]. In our work, we highlighted the downregulation of the PI3K-AKT pathway, as well as the compromised migration and ECM generation of SZC on soft ECM. These results elucidated that mechanical properties of the ECM could affect the regenerative potential of condylar cartilage in TMJOA pathogenesis.

Histologically, the condylar fibrocartilage consists of four layers: superficial zone (SZ), polymorphic zone (PZ), flattened chondrocyte zone (FZ), and hypertrophic chondrocyte zone (HZ) [23, 27]. Researchers have displayed the layer-specific alignment of ECM niche markers: 1) PRG4 in the SZ [23]; 2) TNC, POSTN and DKK3 in the PZ [27, 39]; 3) COL1A1, COL2A1 and BGN in the FZ [36–38], and 4) COL10A1 in the HZ [41]. The PZ was also designated the proliferative layer [41], suggesting the presence of proliferative cells in this zone. In this study, scRNA-seq revealed the composition of chondrocyte clusters in the condyle cartilage, including SZC in the SZ, PMC, and PC in the PZ, FC1 and FC2 in the FZ, and HC in the HZ. The pseudotime analysis illustrated a differentiation trajectory of SZC to PMC and then to PZ/FC-HC. Moreover, functional enrichment displayed that SZC possessed the migration and regeneration capabilities, FC1 and FC2 ranked high in chondrocyte differentiation, while HC featuring in endochondral bone morphogenesis. These results matched the top-to-deep-cartilage and then hypertrophic-cartilage-to-bone trans-differentiation potential of the mandibular condyle [28, 54]. Nonetheless, beside with the established marker PRG4 [23], we found the mechano-sensitive PLPP3 to be especially expressed in SZC, suggesting its important role in regulating SZC’s function.

Moreover, mitochondrial fission is vital for maintaining the quality of the mitochondrial network, and the excessive elongation of mitochondria can lead to impaired cellular respiratory function [55–57]. Dynamin-related protein 1 (DRP-1) is a central mediator of mitochondrial fission with a GTP-hydrolyzing ability [57]. Mitochondrial PA binds to DRP-1, inhibits its oligomerization‐stimulated GTP hydrolysis and consequently restricts the constriction of the mitochondrial membrane [44]. Interactions between the ER and mitochondria contribute in the regulation of mitochondrial fission, with their contact site (mitochondria-associated membrane, MAM) participating in the coordination of phospholipids and calcium signaling between the two organelles [44]. In renal cell carcinoma, PLPP3 on the ER membrane dephosphorylates PA to generate diacylglycerol, which leads to declined PA in the mitochondria through MAM, and subsequently boosted mitochondria bioenergetics [34]. Our work pinpointed PLPP3 to be especially expressed in the ER of SZC, whose expression was significantly decreased in TMJOA. Meanwhile, the elongated mitochondria as well as increased PA in both the ER and mitochondria were detected on soft ECM. The overexpression of PLPP3 rescued the impaired mitochondrial fission and respiration, promoted the functions of SZC and alleviated the cartilage damage in TMJOA. Therefore, our data suggested PLPP3 to be a potential energy metabolism regulator in promoting TMJ regeneration.

In addition, we found a specific regulatory role of CREB5 in SZC. CREB5 is expressed in the perichondrium of the developing knee joint as well as the SZ of adult knee cartilage [58]. It is crucial for the TGF-β and EGFR signaling to maintain the expression of PRG4 in superficial cartilage [58]. In the synovial joint development, CREB5 is responsible for initiating the expression of signaling molecules to guide perichondral tissue to form articular cartilage instead of bone, partially through inhibiting WNT5A [59]. In postnatal joints, the expression of PRG4 and WIF1 is sustained by CREB5 to promote the functional development of chondrocytes [58]. Our work pointed out that CREB5 activation contributed in promoting the PLPP3 expression and boosting the functions of SZC.

However, several limitations of this work remain for further exploration. First, we used the mechanically-detached condylar superficial zone cartilage to isolate the SZC as reported by Embree et al. [23–25]. This method inevitably maintained the potential heterogeneity of SZC, which contained both FCSC and other non-stem cells. The cell surface markers of mesenchymal stem cells such as CD90, CD44 and Gli1, have been reported as markers of FCSC, yet their expression were also found in deeper-layer condylar chondrocytes [30, 60–62]. Advances in the establishment of FCSC-specific markers could help in the isolation and purification of the cells through fluorescence or magnetically-activated cell sorting, leading to a more precise investigation on the features of FCSC and other cells in the SZ. Second, given that the long-range migration of articular mesenchymal stem cells (MSC) could play a critical role in joint regeneration [63], and that ECM organization contributes to maintaining the integrity of the cartilage niche [27], we focused on these two functions of SZC, and investigated the underlying mechanism of their impairment in TMJOA. Future studies determining how the pathological microenvironment influences the differentiation trajectory of SZC by utilizing lineage tracing methods could provide new insights into TMJ regenerative medicine. Finally, our work revealed the role of CREB5 in regulating PLPP3 expression. Further exploration of the spatial and temporal expression of CREB5, as well as its other target genes and pathways, could shed light on the developmental process and regenerative potential of the TMJ.

In summary, this work revealed the softening of ECM in the condylar cartilage SZ of TMJOA, which deactivated the PI3K-AKT-CREB5 pathway in SZC, leading to the decrease of PLPP3-regulated mitochondrial fission and respiration, and finally the restriction of migration and ECM secretion of the cells. Overexpression of PLPP3 alleviated the cartilage degradation in TMJOA. These findings highlighted the role of PLPP3 in boosting SZC reparative functions, and its potential in regenerative treatments for TMJOA.

Materials and methods

Tissue samples

The OA condyle tissues were acquired from patients who underwent temporomandibular joint replacement. The osteochondroma (OC) and benign condylar hypertrophy (CH) condyle tissues were acquired from patients who underwent condylectomy. The tissues from patients with condylar fracture were pieces not able to be rigidly fixed during surgery. The age, gender and diagnosis of the patients were provided above the macroscopic photos in Fig. 1a. Patients with OA, OC, FR or CH in the TMJ were included in the study. Patients with systemic diseases affecting the TMJ, such as systemic lupus erythematosus and rheumatoid arthritis were excluded from the study. All the sample collection was processed in Shanghai Ninth People’s Hospital (Shanghai, China).

Nano-indentation

The tissue samples were gently cleaned with gauze wet by normal saline. Without fixation, the samples were embedded in optimal cutting temperature compound (OCT, Biosharp, CAT: BL557A, LOT:03830153). Frozen-sections were made with a thickness of 50 µm. The sections were washed and immerged in double-distilled water and detected under Chiaro Nanoindentor (Optics11 Life). The k of the probe was 54.2 (N/m), and the tip radius was 27.5 µm (Optics11 Life, P230440). The Hertzian contact model was adopted to calculate the Young’s modulus in DataViewer V2.6.0 (Optics11 Life). The stiffness of the hydrogels was detected using the same probe in the medium of distilled water, and was calculated with the same model.

Animals and Unilateral Anterior Crossbite (UAC) model establishment

Female Sprague–Dawley (SD) rats were obtained from Central Laboratory of Shanghai Ninth People’s Hospital (Shanghai, China), where rat housing and welfare procedures were performed with the approval from the Institutional Animal Care and Use Committee (IACUC) of Institute of Health Sciences, and the National Research Council's Guide for the Care and Use of Laboratory Animals.

The UAC model was established on 6-week-old female rats, following the protocol of previous reports [64]. Briefly, 20# teat cannulas were cut and shaped to form the lower-jaw crowns while 25# cannulas cut to form the upper-jaw ones. Under 1% pentobarbital anesthesia, the crowns were attached to the rats’ left upper and lower incisor with zinc polyacrylate cement (Rong Xiang Dental Material, CAT: 20,153,630,981, LOT: 200,201). The rats were euthanized for tissue collection and analysis after 4 weeks. Rats in the control (CTL) group were female littermates raised in the same condition as the UAC group. They underwent anesthesia without the crowns attached, and were euthanized for tissue collection and analysis at the same time as the UAC group.

Condylar chondrocyte isolation and culture

For scRNA-seq, full-layer condyle cartilage was isolated from 8 rats in CTL or UAC group. The cartilage was cut into 1 mm pieces. The pieces were digested in dispase II (Roche, CAT: 65,558,200, LOT: 04942078001)/collagenase I (Gibco, CAT: 17,100, LOT:2,556,489)/collagenase II (Gibco, CAT: 17,101, LOT:2,556,491) with the concentration of 2 mg/ml each, in 37 ℃ for 4 h.

The isolation and culturing of SZC was according to the published researches [23]. Briefly, the surface layer of condyle cartilage was tear from 4-week-old female rats. The tissue was cut into 0.5*1 mm pieces, and digested in dispase II (Roche, CAT: 65,558,200, LOT: 04942078001)/collagenase I (Gibco, CAT: 17,100, LOT:2,556,489) with the concentration of 2 mg/ml each, in 37 ℃ for 4 h. Cells were incubated in culture medium containing Dulbecco's Modified Eagle Medium/Nutrient Mixture F-12 Ham (DMEM/F12) (Biosharp, CAT: BL305A, LOT:26625678CJ) supplemented with 10% fetal bovine serum (Biosharp, CAT: BL205B, LOT: 01426164EE), penicillin–streptomycin (Biosharp, CAT: BL505A, LOT: 28325009CJ), in 5% CO2, 37 °C. Passage 2–4 of cells were used in the in vitro experiments.

Single-cell RNA sequencing (ScRNA-seq) and bioinformatical analysis

BD Rhapsody system (BD Biosciences) was used to capture the transcriptomic information of the condylar chondrocytes. Sequencing was performed by illumina sequencer (Illumina) on a 150 bp paired-end run. R package Seurat V4 was utilized for cell clustering and visualization. Functional enrichment in Gene Ontology (GO) and Kyoto Encyclopedia of Genes and Genomes (KEGG) datasets was performed with R package ClusterProfiler. R package Monocle 2 was used for pseudo-time analysis and cell developmental trajectory building. Regulon activities of transcriptional factors (TF) were analyzed with R package SCENIC.

Preparation of polyacrylamide (PAM) hydrogels

The preparation of hydrogels was according to the published protocols [65]. Briefly, the 40% acrylamide (MACKLIN, CAT: A6299, LOT: C16236650) and 2% bis-acrylamide (MACKLIN, CAT: B885202, LOT: C17400209) solution were mixed, then combined with 10% ammonium persulfate (MACKLIN, CAT: A6295, LOT: C1577783) and tetramethyl ethylenediamine (MACKLIN, CAT: T6023, LOT: C14819585), and added into 0.75 mm parallel glass plates for polymerization. For the stiff hydrogels, the volume of acrylamide and bis-acrylamide was 36% and 58% respectively. For the soft hydrogels, the volume of acrylamide and bis-acrylamide was 30% and 6% respectively. The volume of ammonium persulfate and tetramethyl ethylenediamine was 0.6% and 0.4%, respectively. After 30 min at room temperature, the plates were removed and hydrogels detached. The gels were washed with 50 mM N-2-hydroxyethylpiperazineN’−2-ethanesulfonic acid (HEPES, Biosharp, CAT: BL1061A, LOT: 24135315H), cut to circles and transferred into 6-well plates. 5 mg/mL sulfo-SANPAH (MCE, CAT: HY-137383, LOT: 320,577) was added on the surface of gels and activated under ultraviolet light for 20 min. The hydrogels were coated with 100 µg/mL rat tail collagen I (Corning, CAT: 354,236, LOT: 27,724,003–1) at 4 ℃ overnight.

mRNA sequencing (mRNA-seq) and bioinformatical analysis

RNA was extracted using Trizol reagent (Solarbio, CAT: R1100, LOT: 24,007,003) with standard procedure. Poly-A-tailed RNA libraries were constructed, and quality control was conducted with the Bioptic Qseq100 platform. Illumina NovaSeq 6000 was utilized for sequencing. R package Deseq2 was utilized for analyzing differentially-expressed genes. GO and KEGG enrichment was performed with R package ClusterProfiler.

Cell migration tracking

To track the migration of cultured cells, quantitative phase image microscopy was adopted with the Livecyte microscope (Phase Focus Limited) according to the manufacturer’s indications. The cells were seeded on stiff or soft hydrogels in 12-well plates and cultured in the Livecyte system at 37 °C and 5% CO2. Phase images were captured every 2 h for 12 h. The data were analyzed with the Cell Analysis Toolbox software (Phase Focus Limited), and visualized with the Chemotaxis software.

Real-time quantitative PCR (qPCR)

Total RNA of cells was extracted using Trizol (Solarbio, CAT: R1100, LOT: 24,007,003) reagent with standard procedure. The complementary DNA (cDNA) was synthesized with Advance Fast One-step RT-gDNA Digestion SuperMix (YEASEN, CAT: 11,141-C, LOT: H2107011). qPCR was conducted using Universal Blue qPCR SYBR Green Master Mix (Biosharp, CAT: BL697A, LOT: 23,352,313). Relative gene expression was determined with the glyceraldehyde-3-phosphate dehydrogenase (GAPDH) gene as reference for normalization. The primers used are listed in Table S1.

Western blotting

Cells were lysed with RIPA lysis buffer with protease inhibitors (Biosharp, CAT: BL504A, LOT: 09125008 M). The protein concentration was quantified with BCA assay kit (Biosharp, CAT: BL521A, LOT: 12025030ZA). After incubation with SDS loading buffer (Biosharp, CAT: BL502A, LOT: 14125005AZ), protein fractions were then separated on 4–20% gradient SDS-PAGE gels, followed by electrotransfer onto nitrocellulose membranes (LABSELECT, CAT: TM-NC-R-45; LOT: 20024296AV). Membranes were then blocked with 5% BSA (Biosharp, CAT: BS114, LOT: 24,144,509) for one hour. Antibody incubation was performed under 4 ℃ overnight. The adopted antibodies are listed in Table S2. Chemiluminescence was visualized after incubation with exposure solution (NCM, CAT:10,300, LOT: ZH2101).

Lentivirus transduction

The vectors adopted to overexpress Plpp3 and Creb5 were constructed by Genomeditech. Briefly, rat Plpp3 sequence was inserted into the PGMLV-CMV-3 × Flag-eGFP-PGK-Blasticidin vector through double-restriction enzyme digestion to construct PGMLV-CMV-3 × Flag-Rat_Plpp3-eGFP-PGK-Blasticidin vector. Rat Creb5 sequence was inserted into the PGMLV-CMV-MCS-3 × Flag-EF1-ZsGreen1-T2A-Blasticidin vector to construct PGMLV-CMV-Rat_Creb5−3 × Flag-EF1-ZsGreen1-T2A-Blasticidin vector. The GM easyTM Lentiviral Packaging Kit (Genomeditech, CAT: GMLCP-20/HGTG-12, LOT: 143,356/143355) was used to generate vector-embedded lentiviruses. For transduction, the cells were infected with the lentiviruses and polybrene (10 µg/mL) (Genomeditech, CAT: GM-040901A, LOT: 517,547). After 72 h, the infected cells were selected using 5 µg/mL blastimicin (Selleckchem, CAT: S7419, LOT: S741905).

Mitochondria and ER isolation

The Mitochondria Isolation Kit (Biosharp, CAT: BL168A, LOT: 34524976D) and BBproExtra ER Isolation Kit (Bestbio, CAT: BB-314541, LOT: BB25011) were adopted respectively to extract the mitochondria and ER from cells, according to the manufacturer’s indications. Briefly, the cells were homogenized on ice with a Dounce homogenizer. For mitochondria isolation, the supernatant was centrifuged at 600 g at 4 ℃ for 10 min for two times, followed by 11000 g centrifuging at 4 ℃ for 10 min. For ER isolation, the supernatant was centrifuged at 4 ℃, 1000 g for 5 min, 11000 g for 10 min, and 15000 g for 10 min. After incubation overnight, the ER was harvested from centrifuging at 10000 g, 4 ℃ for 45 min.

Enzyme-Linked Immunosorbent Assay (ELISA)

The ELISA for PA was performed with a Rat Phosphatidic ELISA Kit (Baiyi Biology, CAT: BY-C460H6, LOT: 202,501) following the manufacturer’s indication. In brief, the cytosol, ER and mitochondria of the cells, as well as the standard solution and HRP-conjugated reagent was added into the testing well. After incubating for 60 min at 37 ℃, the chromogen was added and incubated for 15 min at 37 ℃. Finally, the absorbance of the wells was red at 450 nm using Gen5 Photometer (BioTek, Synergy H1). The concentration of PA was calculated according to the standard curve and standardized with the protein concentration of each sample.

Mitochondrial staining

For Mitochondrial staining, the cells were cultured on stiff or soft hydrogels for 48 h. Then the cells were incubated with MitoTracker® Orange CMTMRos (ThermoFisher, CAT: A66442, LOT: 20,240,417) for 20 min at 37 ℃. The images were captured under ZWISS Axio Observer 7 Confocal Microscope (ZEISS). The morphology of mitochondria was analyzed with the FIJI plugin Mitochondrial Network Analysis (MiNA) (https://github.com/StuartLab/MiNA).

Transmission Electron Microscopy (TEM)

The cells were cultured on stiff or soft hydrogels for 48 h. Then the cells were fixed with 2.5% glutaraldehyde, and detached with cell scrapers, followed by dehydration, embedding, sectioning, and staining for TEM observation. The images were captured by SU8100 TEM (Hitachi).

Adenosine Triphosphate (ATP) content assay

The ATP Assay Kit (Biosharp, CAT: BL852B, LOT: 06325847SG) was used to measure the intracellular ATP concentration. In brief, the cells were lysed in the buffer with ultrasonic disruptor (Fisher, FB120). After centrifuged at 4 ℃, 12000 rpm for 10 min, the sample and reagents were added into the microplate. The 340 nm absorbance was detected respectively after 5 min and 15 min incubation at 25 ℃. The results were calculated according to the formula provided by the manufacturer.

Oxygen Consumption Rate (OCR) assay

OCR of the cells was measured with the OCR Fluorometric Assay Kit (Elabscience, CAT: E-BC-F068). Briefly, the cells were culture in the 96-well detect plate at 37 ℃ overnight to reach 100% convergence. The oxygen probe was added to the culture medium and incubated for 30 min at 37 ℃. After adding the coating oil, the fluorescence (Ex = 405 nm; Em = 675 nm) of the wells were measured every 4 min for 60 min with Gen5 Photometer (BioTek). The results were analyzed according to the manufacturer’s indications.

Chromatin immunoprecipitation followed by quantitative PCR (ChIP-qPCR)

The ChIP assay was conducted with Sonication ChIP Kit (ABclonal, CAT: RK2025B, LOT: NK2358L20W12). After fixation in 1% formaldehyde, the DNA of the cells was sheared to 500–1000 bp by sonication (30 W, 15 min, on ice). The chromatin was immunoprecipitated with anti-CREB5 antibody (ABclonal, A14635, 3 μg for 15 μg chromatin) and IgG negative control (3 μg for 15 μg chromatin). After purification with the DNA purification kit (ABclonal, CAT: RK30100, LOT: 962400C14W01), the chromatin was quantified by qPCR. The adopted primer of rat Plpp3 was shown in Table S1.

Immunostaining

For immunofluorescence (IF), the condyle samples were fixed with 4% PFA, decalcified with EDTA solution, and then embedded in paraffin. The samples were dissected at the thickness of 6 μm. The sections underwent antigen retrieval, blocked with 5% BSA, and then were incubated with the primary antibody at 4 ℃ overnight. Then the slices were incubated with fluorescence-labeled secondary antibodies, and subsequently stained with DAPI (Biosharp, CAT: BL105A, LOT: 33224122AZ). The fluorescence signal was then detected using ZEISS Axio Imager M2 Microscope (ZEISS), with the channels chosen according to the secondary antibodies. For immunocytochemistry (ICC), the cells were fixed with 4% PFA, then blocked with 5% BSA. The incubation with primary and secondary antibodies, as well as the imaging was the same as IF.

For immunohistochemistry (IHC), the sections underwent antigen retrieval, endogenous peroxidase inactivation, blocking, primary antibodies incubation overnight, secondary antibody conjugation with anti-rabbit/mouse IHC kit (Biosharp, CAT: BL1659, LOT: 26725085DU) and DAB reaction (ZSGB-BIO, CAT: ZLI-9018, LOT: 240,010,303). The stained slices were photographed under ZEISS Axio Lab5 Microscope (ZEISS). The antibodies used in immunostaining were listed in Table S3.

Histologicochemistry

The sections were stained by hematoxylin–eosin (HE, Biosharp, CAT: BL735, LOT: 22,102,581), or 0.4% Safranin-O solution and 0.1% fast green solution (SO, Servicebio, CAT: G1053, LOT: CR2206140). Condylar cartilage destruction was scored according to the modified Mankin score [66], by three observers blinded to the experimental information.

Fluorescence in situ hybridization (FISH)

The RNASweAMI FISH Kit (Servicebio, CAT: GF007, LOT: CR2412138) was used to perform the FISH assay. The rat-Plpp3 probe mix was designed and synthesized by Servicebio. In short, after heat retrieval and proteinase K digestion, the sections were sequentially hybridized with probe 1, probe 2 and IF-550 probe. After DAPI staining, the images were captured with ZEISS Axio Imager M2 Microscope (ZEISS).

Dual‑luciferase reporter assay

The pGL3-basic-Rat_Creb5 vector, pGL3-basic-Rat_Plpp3 promoter(−2000to + 50) vector, and negative control vectors were synthesized by Genomeditech, and transfected into HEK293T cells. The luciferase reporter and Renilla luciferase vector were also transfected with Lipofectamine 3000 (Invitrogen, CAT: L3000015, LOT: 3,017,492). After 48 h, the luciferase activity was measured with Luciferase Assay Kit (Genomeditech, CAT: GM-040502A, LOT: 205,779) following the manufacturer’s instructions.

Adeno-associated virus (AAV) injection

The AAVs embedding rat Plpp3-overexpressing vectors GPAAV-CMV-R_Plpp3-EF1-ZsGreen1-WPRE and negative control vectors GPAAV-CMV-MCS-EF1-ZsGreen1-WPRE were synthesized by Genomeditech. After 1 week of UAC establishment, 5 × 10¹¹ AAV in 20 µL PBS was injected into the cavity between the mandibular disc and condylar cartilage of each rat’s left TMJ. The left TMJ condyle of the rats were used for histological analysis 3 weeks later.

Statistical analysis

The statistical analysis was performed with Graphpad Prism software (GraphPad Software Inc.,). All data was displayed as mean ± SD, with statistical difference considered significant when p < 0.05. The paired two-tailed Student’s t-test was used to compare the difference between two groups. The two-way ANOVA followed by Turkey’s multiple comparison test was adopted to compare the difference between four groups.

Authors’ contributions

Conceptualization: Z.Z., D.H., Y.Z., and Y.Y.; Data Curation: Y.Y., D.W., and Y.Z.; Coding and debugging: Y.Y., and D.W; Formal Analysis: Y.Y., D.W., L.Z., H.Z., and C.W; Investigation: Y.Y., D.W., X.Z., C.L., and J.Z.; Funding acquisition: Z.Z, D.H., and Y.Z.; Supervision: Z.Z, D.H., and Y.Z.; Writing-Original Draft Preparation: Y.Y. and D.W.; Writing-Review and Editing: Y.Y., D.W., C.W., Y.Z., D.H., and Z.Z. All authors have read and approved the final paper.

Funding

This work was supported by: the National Natural Science Foundation of China (82270996, 32071313); the Science and Technology Commission of Shanghai Municipality Science Research Project (20S31902500, 20Y11903900); Shanghai's Top Priority Research Center, Grant/Award Number: 2022ZZ01017; the National Natural Science Foundation of China (No. 82472554).

Data availability

The scRNA-seq data and mRNA-seq data analyzed in this work were created in-house and were uploaded to the NCBI BioProject (scRNA-seq, ID: PRJNA951365; mRNA-seq, ID: PRJNA1433058).

Declarations

Ethics approval and consent to participate

This study was performed in line with the principles of the Declaration of Helsinki. The collection of all human samples was authorized by the patients. Written informed consent was obtained from all participants. The sample collection was approved by the Ethics Committee of Shanghai Ninth People’s Hospital (Shanghai, China), with the approval number SH9H-2021-T141-1. The animal experiment was approved by the Ethics Committee of Shanghai Ninth People’s Hospital (Shanghai, China), with the approval number SH9H-2021-A96-1.

Consent for publication

The authors affirm that human research participants provided informed consent for publication of the images in Fig. 1a.

Competing interests

The authors have no relevant financial or non-financial interests to disclose.