Scx^Lin cells directly form a subset of chondrocytes in temporomandibular joint that are sharply increased in Dmp1-null mice

INTRODUCTION

The temporomandibular joint (TMJ) is a complex skeletal structure that is essential for jaw movement in mammals ¹. This joint is comprised of the mandibular condyle, glenoid fossa, an articular disc and numerous tendons and ligaments ^2,3 (two extremely similar connective tissues simply called tendon in this paper). The articular disc is dense fibrocartilaginous tissue rich in fibroblasts and type I collagen ⁴. The disc displacement is one of the most common causes for TMJ disorders which affects 5–10% of the population in the US ^5–7. Yet, the cell origin of the disc is still unclear, which further increases the challenge to develop appropriate treatment strategies.

The condyle is covered by the mandibular condyle cartilage, which is considered as a fibrocartilage with well-organized cellular zones of maturation. These zones include a distinct fibrous layer followed by prechondroblastic layer, and chondrocytic layers including chondrocytes and hypertrophic chondrocytes ^8–10. Similar to the disc, the fibrous zone contains fibroblasts as the primary cell type and an ECM (extracellular matrix) of dense, organized, type I collagen. The in vitro study suggested that the cells in the fibrous layer can give rise to mature chondrocytes ¹⁰. In contrast to large joints ^11,12, the TMJ condyle has a unique prechondroblastic layer (also rich in type I collagen), although the cell source and biological function of this layer is largely unknown.

Cell transdifferentiation plays a critical role in skeletal formation ¹³. Our recent publications showed that up to 80% of the TMJ-ramus bone cells are chondrocyte-derived controlled by Wnt/β-catenin signaling ^14–16. This finding is similar to studies in limb development ^13,17–19. Similarly, fish tendon forms a fibrocartilaginous pad, followed by sesamoid bone formation under comprehensive loading ²⁰. Furthermore, tendon calcifies in birds under tensile loadings ^21,22. In humans, the calcific tendonitis occurs either through degenerative or reactive calcification ^23–26. However, it is largely unclear if tendon lineage contributes to TMJ formation.

DMP1 (dentin matrix protein 1), highly expressed in osteocytes that account for more than 90% of bone cells ^27–29, is critical for normal postnatal chondrogenesis and subsequent osteogenesis ³⁰. DMP1 mutations in human ^28,30 or Dmp1 deletion in mice ³¹ or rabbits ³² result in hypophosphatemic rickets caused by an increase of fibroblast growth factor 23 ^33–35. Our recent studies demonstrated there is an acceleration of chondrogenesis and bone transdifferentiation from chondrocytes in hypophosphatemic rickets ³⁶. However, some increased chondrocytes remained as cartilage cells and do not transdifferentiate into bone cells, resulting in more accumulated cartilage tissue. This finding raises a new puzzle: is there a special subset of chondrocytes that has a different origin and a different gene expression pattern from the perichondrium-derived chondrocytes?

Scleraxis (Scx), a transcriptional factor highly expressed in tendon cells, is essential for tendon development ^26,37–39. Currently there are two lineage tracing lines reported in literatures: ScxCre^{ERT2 24} (widely used for studying the tendon fate during development) and Scx-Cre ⁴⁰, although these tracing lines are mainly used for studying tendon formation and trauma reactions in limbs.

In this study, we attempted to address whether Scx lineage directly contributes to chondrogenesis in normal and diseased TMJ growth. Our findings suggest that Scx lineage directly contributes to disc formation, and forms a new subset of chondrocytes for the TMJ condyle, which is highly sensitive to a change in phosphorus levels as demonstrated in the Dmp1 KO mice (a hypophosphatemic rickets model).

METHOD

Mouse strains

For tendon-derived tracing studies, Scx-Cre^{ERT2 41} was crossed to R26R^tdTomato (B6;129S6-Gt(ROSA)26Sor^{tm9(CAG-tdTomato)Hze}/J, stock number: 009705) and named Scx^Lin. For chondrocyte-derived tracing studies, Acan-Cre^{ERT2 42} and Col10a1-Cre ⁴³ were respectively crossed to R26R^tdTomato, and named Acan^Lin and Col10a1^Lin, respectively. For studying the impact of hypophosphatemia on cell fate, Scx^Lin was further crossed to Dmp1 KO mice ²⁸, with the Scx^Lin; Dmp1+/− as the controls. For studying the impact of DTA-mediated ablation in tendon cells, R26R^DTA mice (B6.129P2-Gt(ROSA) 26Sor^tm1(DTA)Lky/J, stock number: 009669) were crossed to Scx^Lin mice, with the Scx^Lin mice as the controls. Tamoxifen was injected to induce the Cre^ERT2 activity. The tamoxifen powder (Sigma T5648) was dissolved in 10% ethanol and 90% corn oil (Sigma C8267) at a concentration of 10 mg/ml. A low dosage of 75 mg/kg was injected into mice intraperitoneally ^14,15. Since the windows of time for bioactivity of tamoxifen and its metabolites are between 4–24 h ⁴⁴, two types of injection strategies were applied: single injection was started at P3 for tracing the fate of the early Scx^Lin cells; and multiple injections were given twice per week for late recruitments studies. The validation of tamoxifen-negative controls for Scx^Lin was shown in Fig. S1.

All protocols were reviewed and approved by the Institutional Animal Care and Use Committee (IACUC) at Texas A&M College of Dentistry.

Immunohistochemistry, H&E, Toluidine Blue, Safranin O Staining

The TMJs were fixed in 4% paraformaldehyde and decalcified in 10% EDTA at 4°C, followed by either CryoJane frozen sections as previously described⁴⁵ or embedded in paraffin, sectioned and stained with H&E, Safranin O (proteoglycans) or Toluidine blue stain⁴⁶. Immunostaining was proceeded as previously described⁴⁷ with the following antibodies: rabbit anti-aggrecan antibody (Abcam; 1:400), rabbit anti-DMP1 antibody (provided by Dr. Chunlin Qin at Texas A&M University, 1:400), rabbit anti-Collagen I antibody (Abcam; 1:100), rabbit anti-Sox9 antibody (Abcam; 1:100), rabbit anti-Collagen X antibody (Abcam; 1:200), rabbit anti-MEPE antibody (Kerafast, 1:100). The immunofluorescent signals were detected with the corresponding Alexa second antibody (Thermofisher; 1:200) at room temperature for 2 hours.

Cell proliferation

To evaluate cell proliferation, 5-ethynyl-2’-deoxyuridine (EdU) (Invitrogen A10044) dissolved in PBS was administered to mice at the indicated postnatal days 3 hours before sacrifice. Click-iT Imaging Kit with Alexa Flour 488-azide (Invitrogen, C10337) or Alexa Flour 647-azide (A10277) was used to detect EdU in cryosections^15,48.

TEM

TMJ condyles were excised and fixed for several days in 1.5% glutaraldehyde/1.5% formaldehyde, rinsed, then decalcified in 0.2 M EDTA with 50 mM TRIS in a laboratory microwave (Ted Pella, Inc) operated at 97.5 watts for fifteen, 99 min cycles. Samples were fixed again in 1.5% glutaraldehyde/1.5% formaldehyde with 0.05% tannic acid overnight, then rinsed and post-fixed overnight in 1% OsO4. Samples were dehydrated and extensively infiltrated in Spurr’s epoxy and polymerized at 70 °C⁴⁹. One-micron thick sections stained with an epoxy tissue stain (Electron Microscopy Sciences) were used to identify regions of interest. Ultrathin sections containing ROI were cut at 80 nm, contrasted with uranyl acetate and lead citrate, and imaged using a FEI G20 TEM operated at 120 kV with montages collected using an AMT XR-41 2 × 2K camera. The acquired images were stitched using ImageJ software⁵⁰.

Tissue clearing, Confocal microscopy and Image analysis

10-week mandibular condyles from Scx-Cre^ERT2; R26R^tdTomato with Dmp1 KO or Dmp1 +/− background were used for 3-D reconstruction images, and movies were collected. Tissue clearing was performed on these samples with the PEGASOS method as previously described ⁵¹. Fluorescence cell images were captured using SP5 Leica confocal microscope. All images were obtained at light ranging from 488 (green) to 647 (red) μm. Multiple stacked images were taken at 200Hz (1024×1024) and shot in 10X, 20X and 63X objectives ⁵². For tissue clearing samples, red Tomato reporter represented Scx^Lin cells. Image processing, 3-D reconstruction images, and movies were used with Image J software and Imaris 9.0 (Bitplane) ⁵². For IHC staining, red Tomato reporter represented the Cre event and the daughter cells they derived; the green color indicated the corresponding marker for immunostaining. Cells were counted using NIH ImageJ software.

Statistical Analysis

All data were reported as mean (SD). The Kruskal-Wallis test was used to detect significant differences among samples (n=4). The Mann-Whitney U test (post hoc test) was used to compare differences between the mutant and age-matched control groups (n=4). Significance level was defined as *p < 0.05; **p < 0.01.

RESULTS

The prechondroblastic layer has a unique feature of cell morphology and fiber distribution

One of the most unique features of the TMJ condyle is its prechondroblastic layer, which is between the fibrous and chondrocyte layers (Fig.S2A) and expresses both cartilage and bone markers ^53,54. The representative TEM image (Fig.1A) of the TMJ condyle from a 3-week old mouse exhibited a distinctive ultrastructural feature of the prechondroblast compared with the fibroblast (flat with dendrites) and chondrocyte (round and surrounded by lacunae) as below: 1) irregular cell shape with prominent dendrites the size of which gradually reduced during postnatal growth; 2) rich collagen fibers either in fibril bundles (yellow dashed circles) or loosely distributed format in ECMs, which gradually disappeared in the cartilage layer; 3) active cell proliferation as indicated by red dashed circles; and 4) lack of lacunae (white dashed circles), which were gradually formed along with chondrogenesis (Fig.1A–B). The non-decalcified Safranin O staining for a 2-week old mouse condyle cartilage displayed a transition from the fibrous layer to the prechondroblastic layer followed by the chondrocytic layer (Fig.S2B). The H&E images also showed a gradual reduction in the thickness of the prechondroblastic layer from P49 to 8-months (Fig.S2C). Similarly, the prechondroblastic layer was demonstrated in rabbit (Fig.S3A) and human cadaver TMJ condyles (Fig.S3B) in a corresponding style as shown in mice, indicating a progressive transition from fibroblasts to chondrocytes via a prechondroblast step is likely a universal event in other species.

Fig. 1. There are unique features of the TMJ prechondroblastic layer in cell morphology and fiber distribution.

A) The TEM image from a P21 WT mouse condyle displayed unique features of prechondroblasts, which includes dendritic cell shapes, a transition of ECM environment from fibril bundle (yellow dashed circle, FB) to collagen fibers (CF) and to lacuna (L, white dotted lines), and active cell proliferation (blue dotted lines), compared to the other two layers in condyle cartilage. B) A summarized cartoon indicates dual transition from flat fibrous cell to dendritic prechondroblasts and then round chondrocytes along with the changes in ECM environment. F: fibrous layer; P: prechondroblastic layer; C: chondrocytic layer; white solid lines: boundaries between each layer.

Scx^Lin directly contributes to growth and modeling of the TMJ disc and condyle during entire postnatal development

To trace the Scx^Lin cell fate, the Scx-Cre^ERT2 mice were crossed with R26R^tdTomato. Single tamoxifen injection was given at P3 and mice were harvested at P4, P9, P33 and P60, respectively. At P4, there were very few Scx^Lin cells in the fibrous and prechondroblastic layers, whereas more Scx^Lin cells were detected in the TMJ disc (Fig.2A). At P9, there were even more Scx^Lin cells identified in the disc, fibrous and prechondroblastic layers with few in the chondrocyte layer (Fig.2B). By P33 (i.e., 4-weeks after induction), there was an expansion of Scx^Lin cells into cell clusters, which spread from the condyle surface to a deeper portion of the cartilage. The representative immunofluorescent confocal image exhibited overlapping of Sox9 (a key transcription factor for chondrogenesis) and Scx^Lin cells (yellow color; Fig.2C, right panel), supporting a cell transdifferentiation from the initially labelled fibroblast cell (Sox9⁻) to the Sox9⁺ chondrocytes. At P60, more Scx^Lin clusters were documented with Sox9 expressed in their nuclei (Fig.2D). The quantitation data showed a gradual increase of the Scx^Lin cells number in both the disc (~50% of cell populations) and condyle cartilage (~18% of the cell populations) from P4 to P60 (Fig.2E). The EdU labelled Scx^Lin cells counted for 6.6 % of the cell population in P19 condyles (Fig.2F), suggesting a low proliferation rate.

Fig. 2. A set of early labeled Scx^Lin cells contribute to late condyle head expansion.

A) At P4 (one day after injection of tamoxifen) few Scx^Lin cells were observed in the fibrous layer with no Sox9 expression (right panel). B) At P9, more Scx^Lin cells were shown in the fibrous layer, in which there was no Sox9 expression (right panel). C) At P33, some Scx^Lin cells in the fibrous layer continuously differentiated into prechondroblasts and chondrocytes with Sox9 expressed in the nuclei (right panel). D) At P60, more Scx^Lin cells accumulated into clusters that extended from fibrous to chondrocytic layers with Sox9 expression in nuclei (right panel). E) The quantitation data showed that the percentage of Scx^Lin cells in disc and condyle cartilage gradually increased from P9 to P60. F) At P19, 6.6 % Scx^Lin cells in condyle cartilage were EdU positive (n=4). D, disc. * P<0.05; ** P<0.01; *** P<0.001

Next, we set out to investigate the functional importance of Scx^Lin in TMJ growth. A 30-day inducible cell ablation experiment (from P3 to P33 with single tamoxifen injection at P3) was performed using Scx^Lin as the control, and Scx^Lin; R26R^DTA (carrying a loxP-flanked stop cassette associated with an attenuated DT for the ablation of cells when Cre is active) as the ablation group. The tracing results revealed a great reduction in the number of Scx^Lin cells in the ablation group (Fig.3A, lower) compared to the control (Fig.3A, upper). The quantitation data showed an over 50% decrease of the red cells in the ablation group, which is significant between these two groups (Fig.3B). Together, the above data supports a notion that Scx^Lin cells are largely quiescent at the early developmental stage but rapidly expand during late development.

Fig. 3. There is a sharp reduction of Scx^Lin cells in the DTA group.

A) DTA-mediated ablation of Scx^Lin cells (induction at P3 and sacrifice at P33) led to fewer red cells in TMJ condyle compared to the control group. B) There was a more than 50% reduction in the percentage of Scx^Lin cells in the DTA condyles compared to the controls, which is significant between these two groups (P < 0.01).

Scx^Lin cells are continuously recruited during TMJ expansion

In contrast to the growth plate, TMJ condyle growth starts late and continues to expand from a carrot shape to a mushroom-like structure under the mechanical stimulation during postnatal growth ^55–57. Additionally, the condylar fibrous layer remains active in mouse (Fig.S2) and human (Fig.S3B). Importantly, one-time tracing data showed a gradual increase of Scx^Lin cells in condylar expansion when mice start to be more active in mastication (Fig.2). Next, we asked whether there is a continuous recruitment of Scx^Lin cells during TMJ condylar expansion. To test this hypothesis, we did multiple injections of tamoxifen at two age groups as below.

First, we did multiple injections of tamoxifen into Scx^Lin mice two times/week starting from P3 and harvested them at P21. The cell lineage tracing data showed a more and broad distribution of Scx^Lin cells in all three layers (Fig.4A) compared to a single injection experiment (Fig.2). The tracing data also showed a change in cell shape from flat in the fibrous layer to dendritic in the prechondroblastic layer, followed by a round shape in the chondrocytic layer. The quantitation data (the ratio of Scx^Lin cells/DAPI⁺ cells) revealed that 91% of cells in the disc and fibrous layers, 67% in the prechondroblastic layer, and 31% in the chondrocytic layer were Scx^Lin derived (Fig.4B). The strong expressions of Sox9 and Aggrecan in the Scx^Lin cells within the chondrocytic layer further confirmed the transdifferentiation into cartilage cells, whereas the Col 1 expression was limited in the fibrous and prechondroblastic layers (Fig.4C).

Fig. 4. Scx^Lin cells are continuously recruited during TMJ condylar expansion.

A) A confocal view of a 3-week-old TMJ (induced at P3 with two tamoxifen injections each week) revealed a strong Tomato signaling (low magnification, left panel; high magnification, right panel), indicating a transition from fibroblasts to prechondroblasts and chondrocytes. B) The quantitation data showed that over 90% cells in both disc and fibrous layers, 67% in prechondroblastic layer, and 31% in chondrocytic layer were from Scx^Lin. C) Co-immunostaining with Sox9 (left), aggrecan (middle) and Col 1 (right) showed a distinct feature of Scx^Lin cells in the TMJ condyle. D-E) Late activation of Scx^Lin (induced at P28 with two tamoxifen injections each week and harvested at P42) showed a continuous recruitment of Scx^Lin cells (D), some of which were Sox9⁺ chondrocytes (E, arrows). F) The quantitation results showed that 53% cells in disc and 22% in the condyle cartilage were Scx^Lin cells. F: fibrous layer; P: prechondroblastic layer; C: chondrocytic layer; white solid lines: boundaries between each layer.

Second, we activated the Scx-Cre^ERT2 event with two times of tamoxifen injections each week from P28 to P42. The Scx^Lin cells were found in the disc and all three layers of the condyle cartilage (Fig.4D). Some of the Scx^Lin cells expressed Sox9 in the chondrocyte zone, confirming the transdifferentiation from Scx⁺ fibroblasts into chondrocytes (Fig.4E). The quantitation data showed that 53% of cells in the disc and 22% in the condyle cartilage originated from Scx^Lin (Fig.4F), supporting the notion that Scx^Lin cells contribute to TMJ expansion during the entire postnatal development.

Scx^Lin chondrocytes do not transdifferentiate into bone cells

Previously, we reported over 70% of the bone cells in TMJ condyle ramus were transdifferentiated from Aggrecan⁺ (Acan^Lin) chondrocytes ¹⁴. Here we asked whether the Scx^Lin chondrocytes follow the suit as the Acan^Lin does by comparing these two tracing lines at two age groups (P30 and P60). Both Cre lines received single tamoxifen induction at P3 and co-immunostained with DMP1 (a bone marker). There was a progressive increase in the tomato signal from P30 to P60 with a much lower level in the Scx^Lin (Fig.5A–B, left panels) and a very high level in the Acan^Lin (Fig.5A–B, right panels). There were essentially no red bone cells in the Scx^Lin subchondral bone, whereas most osteoblasts (OBs) and osteocytes (OCYs) were derived from the Acan^Lin chondrocytes. This information supports the notion that the Scx^Lin chondrocyte does not transdifferentiate into bone cells as the Acan^Lin chondrocyte does.

Fig. 5. There is a lack of transdifferentiation from Scx^Lin chondrocytes into subchondral bone cells during TMJ condyle growth (one-time tamoxifen induced at P3 with animals sacrificed at P30 and P60, respectively).

A) There was a progressive increase in Scx^Lin chondrocytes from P33 to P60 but no Dmp1⁺/Scx⁺ bone cells in the subchondral bone, although there were some Scx⁺ fibroblasts in the bone marrow with no Dmp1⁺ ECM surrounding. B) Many Acan^Lin chondrocytes rapidly transdifferentiated into osteoblasts (Ob) and osteocytes (Ocy) with high level of Dmp1 in the surrounded ECM in both age groups.

A low phosphorus level accelerates Scx^Lin derived chondrogenesis

Our recent studies ³⁶ showed a disproportional increase of chondrogenesis and osteogenesis in Dmp1 KO condyles (i.e., some increased chondrocytes remained as cartilage cells and do not transdifferentiate into bone cells, resulting in more accumulated cartilage tissue). To address this puzzle, we first confirmed a greater increase in cartilage mass verse bone volume in the Dmp1 KO TMJ condyle using classical histology assays at different growth timelines (P10, P21, P49, and P60) (Fig.S4A–B). Next, we showed a great expansion in the fibrous layer and rapid transitions from prechondroblastic cells to much enlarged hypertrophic chondrocytes at TEM level (Fig.S4C). The chondrocyte tracing studies using Col10a1^Lin displayed a drastic increase of Aggrecan expression in both Col10a1^Lin and non-Col10al derived chondrocytes plus a moderate increase in osteogenesis compared to the age matched control (Fig.S4D). This disproportional increase of cartilage mass indicates that some of the increased chondrocytes are likely derived from Scx^Lin cells, which do not transdifferentiate into bone cells (Fig.5).

To test this hypothesis, Dmp1 KO and Scx^Lin mice were internally crossed, which were induced at P3 and harvested at P70. The snapshot of 3D images from a tissue clearance treated control TMJ condyle (with an approximate 100 μm thickness) showed that the Scx^Lin cells were mainly distributed in fibrous and prechondroblastic layers with few in the chondrocyte layer (Fig.6A, left panels). In the KO Scx^Lin mice, there was a sharp increase in the number of Scx^Lin cells in all three layers of the condyle cartilage (Fig.6A, right panels). The enlarged confocal images clearly displayed multiple expanded Scx^Lin chondrocyte columns (or clusters) in the KO (Fig.6B, right panel). The quantitative data confirmed that this 2.5-fold increase in Scx^Lin cells in the KO condyle cartilage versus the age-matched control is statistically significant (Fig.6C, n=4; left panel). Additional analyses of immunostainings displayed a much higher level of aggrecan expression in the KO Scx^Lin chondrocytes (Fig.6D) and many more Scx^Lin hypertrophic chondrocytes expressing type X collagen versus the control mice (Fig.S5). However, there was no co-localization of Scx^Lin cells and MEPE⁺ bone cells in either the control (left panel) or KO (right panel) mice (Fig.6E). Thus, we conclude that it is the Scx^Lin chondrocyte that is responsible for an increase in the KO cartilage volume with no contribution to further cell transdifferentiation into bone cells.

Fig. A low phosphorus level accelerates tendon-derived chondrogenesis with no bone cell transdifferentiation.

A) The confocal images (~ 100 mm thickness) of the compound mice Dmp1 KO; Scx-Cre^ERT2; R26R^tdTomato showed a drastic increase in the number of Scx^Lin in all layers of condyle cartilage (right panels). B) The enlarged confocal images revealed many round Scx^Lin-derived chondrocytes in columns originated from fibrous layer (right panels), whereas the tendon-derived cells remained low numbers in the controls (left panels). C) The quantitative data showed a significant increase in the number of Scx^Lin cells in condyle cartilage in the KO mice compared to the controls (n=4). D) A co-localization of Aggrecan and Scx^Lin signal images displayed a great increase in Aggrecan expression in the KO chondrocytes compared to the controls. E) MEPE immunostaining combined with Scx^Lin signal confirmed no bone cell transdifferentiation from the Scx^Lin derived chondrocyte in both the control and KO condyles. D: disc; white dotted line: the surface of condyle cartilage

DISCUSSION

The TMJ is a critical yet overlooked joint whose defects affect a large population, especially women ⁵⁸. Although it is increasingly clear that the TMJ is distinct from other large joints, our understanding on this mysterious joint is primitive. In this study, we investigated the features of cell morphologies, ECMs, and the cell origin of the TMJ during postnatal growth using multiple image techniques and the Scx^Lin mice with and without a Dmp1 KO (a hypophosphatemic rickets model) background. Our key findings (Fig.7) are: 1) Scx^Lin directly contributes to a subset of TMJ chondrocyte population via two steps: Scx^Lin cells in fibrous layer first form dendritic prechondroblasts (rich of type I collagen in its ECM) and subsequently transdifferentiate into Scx^Lin chondrocytes; 2) Scx^Lin chondrocytes do not transdifferentiate into bone cells as the non-Scx^Lin chondrocytes do ^14,59; and 3) the Scx^Lin derived chondrogenesis, similar to the non-Scx^Lin derived chondrocyte ³⁶, is highly sensitive to a change of phosphorus level as demonstrated in the Dmp1 KO mice.

Fig. 7. Working Hypothesis: Direct contribution of Scx^Lin cells to the development of TMJ.

There are two sets of progenitors in the fibrous layer (F) on the top of the condyle. One is Scx⁻, the commonly studied chondrocytes, which form prechondroblasts (P) followed by chondrocytes (C). Subsequently, these Scx⁻ chondrocytes transdifferentiate into endochondral bone for TMJ condyle growth. The other set is Scx^Lin cells that are responsible for a small portion of the progenitors in the fibrous layer. This type of progenitor cells also sequentially differentiates into prechondroblasts and chondrocytes for joint expansion, although they do not further transdifferentiate into bone cells. We name this process as tendon-derived chondrogenesis. Both types of chondrogenesis are highly sensitive to the change of phosphorus level and responsible for TMJ defects caused by hypophosphatemia. In addition, Scx^Lin cells are solely responsible for TMJ disc formation.

The cell origin of the prechondroblastic layer which is a unique feature of the TMJ condyle is largely unclear, although it is considered to be derived from the perichondrium based on early histological studies ^4,60,61. Apparently, our cell lineage study supports a different cell source: the Scx^Lin origin (Figs.2–4). In contrast to a one-step transition from chondrocytes to subchondral bone cells ^14,59, it takes multiple steps for cell transdifferentiation from a flat fibrous cell to a round or oval chondrocyte via the irregular and dendritic prechondroblasts in addition to complex remodeling in ECM fibers (Fig.1). The co-immunostaining with different cell markers (Fig.4) further showed a dual transition of Scx^Lin cells from flat fibrous cell to dendritic prechondroblasts (with Col 1 but no Sox9 expressions) and then round chondrocytes (with Sox9 and aggrecan but no Col 1).

The epithelial-mesenchymal cell transdifferentiation (EMT) or vice-versa occurs in gastrulation, neural crest and somite dissociation, craniofacial development, wound healing, organ fibrosis, and tumor metastasis ⁶². Similarly, our recent publication ^14,59,63 using multiple approaches (including cell lineage tracing), in agreement with studies in limb development ^13,17–19, demonstrated that the vast majority of endochondral bone cells are directly originated from mature chondrocytes via a one-step cell transdifferentiation. Interestingly, in this study we found the Scx^Lin derived chondrocytes do not undergo osteogenesis as the Acan^Lin derived chondrocytes do ^14,59 (Figs.5–6). Although we do not know exactly the impact of the Scx^Lin chondrocyte in TMJ condyle growth, the Scx^Lin derived chondrocytes may help to maintain a life-span cartilage mass in TMJ condyles in both mice (Fig. S2) and human (Fig. S3B).

Recently, we reported a disproportional increase in the condyle cartilage mass (a sharp increase) and the subchondral bone (a moderate increase) in the Dmp1 KO mice in response to hypophosphatemia ³⁶ (also see Fig.S4). This study suggests two separate chondrocyte pools: one, as a bone precursor, transdifferentiates into an endochondral bone cell; and the other remains as a “permanent chondrocyte” to maintain a stable cartilage mass in the TMJ condyle. The Scx^Lin chondrocytes appear to belong to the latter group with the following feature: 1) they do not undergo further transdifferentiation into bone cells during normal growth (Figs.2–5); and 2) they are highly sensitive to hypophosphatemia, which is responsible for the disproportional expansion between the condyle cartilage and bone (Fig.6; Figs.S4–5). To further enforce the contribution of Scx^Lin to the formation of a subset of TMJ chondrocytes that does not transdifferentiate to bone cells as the Acan^Lin derived chondrocytes do, we plan to use a powerful multicolor labeling mouse line ⁶⁴ in our future studies so that the tracking of differential fate of cells, including neighboring cells, can be assessed effectively.

The TMJ disc is a fibrous extension of the capsule in between the temporal bone and mandible. The anterior portion of the disc is coincident with the insertion of the superior head of the lateral pterygoid, and the posterior portion is continuous with the joint capsule. The origin of the TMJ disc is largely unclear. Here, our lineage tracing results demonstrate a key contribution of Scx^Lin to the formation of the TMJ disc.

The Scx lineage line is widely used in studies for the fate of tendon/ligament cells ^24,65. There were no reports about a direct contribution from Scx⁺ cells in tendon/ligament to the joint chondrogenesis during normal limb development, although it has been reported that Scx^Lin participated in ectopic endochondral bone formation in tendon trauma ^26,41. Scx⁺ cells were also found in other tissues, such as periodontal ligament (PDL) ⁶⁶ and heart valves ⁶⁷, with no evidence of transdifferentiation into other cell types. Similarly, in this study we observed sparse Scx^Lin-derived bone marrow fibroblasts in several developmental stages, but they appear not to transdifferentiate into bone cells (Figs.5–6). Currently, we do not know the impact of those sparse Scx^Lin-derived bone marrow fibroblasts in the skeletal development, and more studies are needed to uncover the mechanism for the unique roles of Scx^Lin cells in the chondrogenesis of TMJ condyle.

Taken together, the above comprehensive studies demonstrate the novel roles of Scx^Lin in TMJ chondrogenesis via the prechondroblastic layer (Fig.7). This new information provides an innovative idea that tendon cells may have a more extensive function beyond a simple mechanical support in body movement, and it will likely revise the current dogma and provide awareness in this understudied area. However, there are more challenging questions to be addressed in the future. First, there is an urgent need to identify more tendon specific tracing mouse lines, as a few Scx^Lin cells are observed in the bone marrow. Second, both Scx^Lin and Scx⁻ chondrocytes are observed in fibrous and prechondroblastic layers, which raises a question on how to distinguish their features from early progenitors to late daughter cell stages at the gene and protein levels. Particularly, it is unclear if there are even more specific transcriptional factors beyond the known ones for chondrogenesis or osteogenesis, which define the cell fate of these progenitors to transdifferentiate into an osteocyte or to remain as a “permanent chondrocyte”. Lastly, more studies are required to demonstrate the implication of tendon-derived chondrocytes in future diagnosis or treatment of TMJ disorders, which affect a large TMD population ⁶⁸.

Supplementary Material

1

Fig. S1 There is no tamoxifen-independent recombination in Scx-Cre^ERT2 line. A) No tomato signaling were found in the Scx-Cre^ERT2; R26R^tdTomato mice in different ages that did not receive the tamoxifen injection. B) There was no tomato signaling of P70 Scx-Cre^ERT2; R26R^tdTomato; Dmp1 KO mice and the controls that did not receive the tamoxifen injection.

2

Fig. S2 The prechondroblastic layer continuously exists in mice TMJ condyle from young to adult age. A) An illustration showed disc (D), fibroblast layer (F); prechondroblasts (P) and chondrocytes (C) in TMJ. B) a non-decalcified Safranin O stain for 2-week mouse condyle showed the histology of the three layers. C) H&E staining for 7-week TMJ condyle showed a substantial portion of the prechondroblast layer (upper panel), and a sharp reduction in all three layers at the age of 8 months (lower panel, red arrows indicating the remained cell clusters in fibrous and prechondroblastic layers). F: fibrous layer; P: prechondroblastic layer; C: chondrocytic layer; white solid lines: boundaries between each layer.

3

Fig. S3 The prechondroblastic layer is identified in both rabbit and human TMJ condyle samples. A) An identical 3-layer structure was documented in a 3-week-old rabbit condyle by Safranin O stain. B) A very similar F-P-C layer structure was also observed in newborn (left) and 40-year old (right) TMJ condyle from human cadaver. F: Fibrous layer; P: prechondroblastic layer; C: chondrocytic layer; white dotted lines: the boundaries between each layer.

4

Fig. S4 Accelerated chondrogenesis does not results in proportional increase in chondrocyte-derived osteogenesis in the Dmp1 KO condyle. A) Toluidine blue stain displayed an expanded hypertrophic chondrocyte area and disorganized subchondral bone in different age groups of Dmp1 KO condyle (right) compared to the WT (left). B) Safranin O stain further confirmed the increased chondrogenesis and malformed subchondral bone in 2-month Dmp1 KO condyle. C) TEM showed a great expansion in the fibrous layer but dramatic decreases in prechondroblastic and chondrocyte layers, and an accumulation of hypertrophic chondrocytes in the 3-week DMP1 KO condyle. D) The cell lineage tracing from Col10a1-Cre; R26R^tdTomato mice revealed more red chondrocytes and bone cells in the 3-week Dmp1 KO condyle, with dramatic increase of Aggrecan expression.

6

Fig. S5 Scx^Lin cells actively transdifferentiate into hypertrophic chondrocytes in low phosphorus environment. There were many more Scx^Lin cells expressing type X collagen (Col10) in the Dmp1 KO condyle cartilage (arrows) compared to the control mice.

Highlights.

• Scx^Lin cells are responsible for formation of TMJ disc postnatally;

• Scx^Lin cells contribute to formation of a subset of TMJ chondrocytes;

• Scx^Lin chondrocytes do not undergo osteogenesis as the Scx⁻ chondrocytes do;

• Scx^Lin-derived chondrocytes are sensitive to hypophosphatemia