Cry2 prompts clock oscillation and temporomandibular joint homeostasis under mechanical loading

Abstract

Aims

Cartilage metabolism exhibits circadian rhythms (CR), and disruption of these often correlates with cartilage degeneration. Mechanical stimulation is a vital zeitgeber of CR in cartilage. However, the effects of mechanical loading on CR and cartilage homeostasis have not been explored. In this study, we aimed to explore the involvement of CR in mediating temporomandibular joint (TMJ) cartilage homeostasis under mechanical loading.

Methods

We introduced the mechanical loading models of both rat TMJ chondrocytes and condyle explants. The mechanical loading was applied through rhythmic compression (12 hrs on and 12 hrs off), while the mechanical unloading model was set as control through static culture of cells and explants.

Results

One-week static culture led to imbalanced cartilage metabolism and accelerated cartilage degeneration. This was accompanied by diminished levels of core clock genes (bmal1, clock, per1, and cry2), compared with oscillations of these genes in one-day static cultured chondrocytes and condyle explants. Rhythmic physiological mechanical loading for one week not only sustained the oscillations of CR genes, but also partially maintained homeostasis of cartilage. The mechanistic delineation confirmed that one week of mechanical loading promoted the nuclear-cytoplasmic shuttling of Per1 and Cry2, which were obvious in one-day static cultured chondrocytes but weakened in one-week static cultured cells. Furthermore, nuclear translocation of Cry2, but not Per1, was dependent upon mechanical loading-induced Rho-associated protein kinase (ROCK) activation and actin polymerization. Inhibition of ROCK caused actin depolymerization and partially blocked Cry2 nuclear-cytoplasmic trafficking. Finally, down-regulation of Cry2 in both chondrocytes and condyle explants attenuated mechanical loading-sustained circadian oscillations and cartilage homeostasis.

Conclusion

This study elucidated the involvement of CR in mediating mechanical loading-related cartilage homeostasis, as well as the critical role of ROCK-actin-Cry2 in regulating normal CR and homeostasis of TMJ cartilage under rhythmic mechanical loading.

Cite this article: Bone Joint Res 2026;15(1):25–41.

Article focus

• This study investigated the role of circadian gene cry2 in mediating the beneficial effect of mechanical stimuli on maintaining the cartilage homeostasis, and the underlying mechanism involving Rho-kinase (ROCK) pathway and cytoskeletal actin dynamics.

Key messages

• One-week static culture of both temporomandibular joint (TMJ) chondrocytes and cartilage explants led to disturbed cartilage matrix gene expressions and cartilage matrix degeneration, which could be compromised by rhythmic daily mechanical compression.

• Rhythmic mechanical compression-induced nuclear-cytoplasmic translocation of cry2 protein is critical for oscillations of other circadian genes and balanced expressions of cartilage matrix genes and proteases.

• The mechanical stimuli-induced rhythmic ROCK activation and actin polymerization resulted in Cry2 rhythmic nuclear-cytoplasmic translocation in chondrocytes.

Strengths and limitations

• These data provided a potential correlation among mechanical stimuli, circadian clock, and TMJ cartilage matrix homeostasis, and elucidated the mechanism underlying this correlation.

• The sampling rate is relatively low for the analysis of key circadian parameters such as magnitude, period, and phase. A higher sampling rate, or real-time analysis of circadian rhythm, would greatly improve this study.

• This study only employed cell model and explant model to investigate the effects of mechanical loading and unloading on TMJ cartilage circadian clock and homeostasis. Further animal model and clinical data are warranted to better explain the correlation among mechanical stimuli, circadian clock, and TMJ cartilage matrix homeostasis.

Introduction

Mechanical stimuli are vital in promoting proper cartilage development and maintaining its homeostasis. Conversely, absence of mechanical stimulations causes degeneration and loss of cartilage, which is commonly seen in people engaging in spaceflight or on bedrest.^1,2 Even though the development of small molecules and synthetic proteins are promising in inhibiting the progression of cartilage degeneration,^3,4 daily exercise is still a potentially efficient and cost-effective method to counteract the harmful effect caused by the long-term absence of mechanical loading.

In terms of temporomandibular joint (TMJ), the condyle cartilage is frequently subjected to various mechanical stimuli such as compression, fluid shear stress, stretch, and during daily activities (chewing, speech, etc). However, whether physiological loading of TMJ by these mechanical stimuli is necessary for TMJ cartilage homeostasis, as well as whether TMJ cartilage suffers degeneration in the absence of these mechanical stimuli, has not been thoroughly explored due to the difficulties in establishing relevant animal models. To overcome this obstacle, we cultured TMJ condyle explants and chondrocytes to investigate the effects of mechanical loading on TMJ cartilage homeostasis, as well as the underlying mechanisms.

It has been known for a decade that cartilage mitotic activity, metabolism, and mineralization exhibit circadian rhythms (CR), as reported in transcriptomic analyses in rodents.^5,6 The CR in various tissues of the body can be entrained by distinct external signals, such as light-dark cycle, feeding, sleeping, and exercise. Notably, the mechanoregulation of CR in cartilage was first reported in cyclic vibration-induced oscillation in Drosophila (fruit fly) chordotonal organ clocks, implying that mechanical stimulation is a vital zeitgeber of circadian clock in cartilage.⁷ Moreover, Heywood et al⁸ demonstrated for the first time that cyclic tensile strain could reset or shift the phase of chondrocyte CR in vitro. Recently, compressive loading has been confirmed to promote chondrogenic differentiation, through entraining the autonomous clock within chondrocytes.⁹ This evidence indicates that mechanical stimuli are critical in maintaining CR of cartilage. Accordingly, TMJ degeneration in response to mechanical unloading, which often occurs due to mandibular immobilization and function restriction, could result from CR disruption.

So far, the molecular networks regarding mechanical stress-regulated CR in chondrocytes has been addressed by several research groups.^10,11 However, to the best of our knowledge, there are no studies comparing the effects of mechanical loading and unloading on CR in chondrocytes, and the involvement of CR in these effects on TMJ cartilage homeostasis. In this study, we used both an in vitro cultured primary chondrocyte model and ex vivo cultured cartilage explant model to exclude the influences of system cues in modulating cartilage clock, and investigate the effects of mechanical loading and unloading on CR of chondrocytes.

Methods

Alpha-modified Eagle’s minimal essential medium (α-MEM), Dulbecco’s-modified Eagle’s medium (DMEM), and fetal bovine serum (FBS) were purchased from Invitrogen (Thermo Fisher Scientific, USA). Antibodies against Bmal1 (ab3350), Clock (ab3517), Aggrecan (ab313636 and ab150079), matrix metalloproteinase 3 (MMP3) (ab52915), and MMP13 (ab39012) were purchased from Abcam (UK). Antibodies against Per1 (PA5-119798), Cry2 (PA5-89135), and Col2 (PA5-99159) were purchased from Thermo Fisher Scientific. Antibody against β-actin was purchased from Beyotime (China). All the other chemicals used in this study are described in the subsequent sections.

Six-week-old male Sprague-Dawley rats were used in this study. Male and female mice were used in all experiments, as we have not observed sex differences in any of the measured experiments. We have adhered to the ARRIVE guidelines and included the ARRIVE checklist in the Supplementary Material. Organotypic rat TMJ tissue explants were prepared by dissecting the whole TMJ condyle with cartilage, subchondral bone, and synovium. The explants were randomly cultured in six-well culture plates at 37°C, 5% CO₂ in DMEM medium with 5% serum (culturing medium) for one day or one week without changing the medium. For synchronization of the explants, dexamethasone (100 nM) was added into medium for one hour and the medium was changed back to culturing medium. The zeitgeber time (ZT) was set as 0 hours after synchronization. Mechanical loading of tissue explants was performed by mechanical compression stimulator (NK-P40; Naturethink, China) as described by Xiao et al.¹² The compression system comprised a culture chamber, electrothermostatic water bath, and controllable gas cylinder (filled with 76% N₂, 19% O₂, and 5% CO₂) (Supplementary Figure a). Rat explants were subjected to 0 KPa (static control, defined as mechanical unloading), or 50 Kpa, 100 Kpa, 200 Kpa, and 400 Kpa (defined as mechanical loading) compression (1 Hz; 12 hrs on and 12 hrs off) for one week. The TMJ explants were randomly assigned to control and experimental groups, and randomization sequence was created using Stata 9.0 (StataCorp, USA) statistical software. For the sample size determination, PASS.15 software (NCSS, USA) was used by inputting the following parameters: type I error (α) was set as 0.05, type I error (β) was set as 0.1, and the means and SDs of each group were calculated in GraphPad Prism 9 (GraphPad Software, USA) based on our pilot experiments. The sample size was calculated using the PASS.15 software as follows:

$${\displaystyle {\displaystyle N=\frac{\lambda }{\mathrm{\Delta }};\mathrm{\Delta }=\frac{1}{{\sigma }^2}\sum _{i=1}^k({\mu }_i-\overline{\mu }{)}^2;\overline{\mu }=\frac{1}{k}\sum _{j=1}^k{\mu }_j;\lambda =N\frac{{\sigma }_m^2}{{\sigma }^2}}}$$

where N is the sample size, σ is the SD, k is the group number, and μi is the mean of each group. λ does not need to be inputted in PASS.15 software.

Rat TMJ cartilage tissues were scraped from the condyle surface, and were treated with 0.5% type II collagenase for four hours at 37°C. Isolated chondrocytes were centrifuged and resuspended with α-MEM. The condyle chondrocytes were cultured in α-MEM containing 15% FBS. After passage for three times, the chondrocytes were incubated in the mechanical compression stimulator (NK-P40, Naturethink), and subjected to the cyclic compression protocol (50, 100, 200, and 400 kPa; 1 Hz; 12 hrs on and 12 hrs off) for one day or for one week. For inhibition of ROCK, Y27632 was added into culture medium in a final concentration of 10 μM right after one hour of synchronization with dexamethasone during the required mechanical loading period. DMSO was added in the control group. Cell samples were collected at six, 12, 18, and 24 hours ZT of the first day and seventh day of both static and compressive conditions.

The rat TMJ explant samples were dissected and fixed with 4% paraformaldehyde, decalcified with 4% EDTA for four weeks, dehydrated in ethanol, embedded in paraffin, and sagittally cut into 5 μm-thick serial sections. The central sagittal sections of each condyle were selected randomly for haematoxylin and eosin (H&E) staining and Safranin-O-Fast Green staining. The H&E staining kit (C0105S) was purchased from Beyotime, and Safranin-O-Fast Green staining kit (S0335) was purchased from Bioss Antibodies (USA).

The sections for IHC and IF staining were prepared similarly to those for H&E staining. For IHC staining, the sections were incubated with the primary antibody against MMP3 (1:200) and MMP13 (1:200) overnight at 4°C. Following this, sections were incubated with HRP-conjugated goat antirabbit IgG polyclonal antibody at 37°C for one hour. The signal developed as a brown reaction product using the peroxidase substrate 3,3’-diaminobenzidine (Bioss Bioscience, China). For quantification of MMP3 and MMP13 positive area in IHC staining, the region of interest (ROI) fields (300 μm × 200 μm) were randomly chosen in each sample, and ImageJ software (National Institutes of Health, USA) was used to calculate optical density (OD) of each ROI.

For IF staining of ex vivo cultured rat TMJ explants, the sections were incubated with primary antibody against Bmal1 (1:200), Clock (1:200), Per1 (1:250), Cry2 (1:100), Aggrecan (1:500), and Col2 (1:500), before incubation overnight at 4°C. DAPI staining was performed after incubation with fluorescent-labelled secondary antibodies. Digital images were obtained using a laser scanning confocal microscope (Nikon, Japan). For quantification of fluorescent positive area in IF staining, the ROI fields in explant samples (300 μm × 200 μm) were randomly chosen in each sample, and ImageJ software was used to calculate fluorescence intensity (FI) of each ROI.

For IF staining of in vitro cultured chondrocytes, cells were firstly cultured on 0.1% Gelatin (Sigma-Aldrich) pre-coated glass coverslips. After mechanical compression experiments, culturing medium was removed, and the cells were fixed with 4% formaldehyde containing 0.1% glutaraldehyde for 15 minutes at 37°C. After being rinsed with cold PBS (pH 7.4), the cells were permeabilized using 0.1% Triton X-100 (Thermo Fisher Scientific) for ten minutes and then incubated with 1% bovine serum albumin for one hour at room temperature. Antibodies against Bmal1 (1:200), Clock (1:200), Per1 (1:250), Cry2 (1:100), Aggrecan (1:500), MMP3 (1:200), MMP13 (1:200), and Col2 (1:500) were added, and the fixed cells were incubated with antibodies at 4°C overnight followed by incubation with anti-immunoglobulin G (IgG)-FITC or IgG-Texas Red (1:150 dilution) for one hour. After removal of antibodies, cells were rinsed with PBS and mounted with 90% glycerol. DAPI staining was conducted right after washing with PBS. Fluorescence was immediately observed using a Nikon ECLIPSE E400 microscope.

We have reported the methods in our previously published study.¹³ Briefly, total messenger RNA (mRNA) was extracted from primary chondrocytes or cartilage layer of TMJ condyle explant, and reverse transcription (RT) and real-time polymerase chain reaction (PCR) were carried out. Then, quantitative PCR was performed using LightCycler 96 System (Roche, Switzerland) with the primers shown in Table I. The results were presented as the calculated comparative expression ratios of the target sample to control group for each sample using the CT method (2^−ΔΔCT).

Table I.

Primers used for reverse transcription quantitative polymerase chain reaction of endogenous reference genes and target genes.

Genes	Forward	Reverse
β-actin	5'-TGCTATGTTGCCCTAGACTTCG-3'	5'-GTTGGCATAGAGGTCTTTACGG-3'
bmal1	5'-CCGATGACGAACTGAAACACC-3'	5'-TCTTCCCTCGGTCACATCCT-3'
clock	5'-AGAACTTGGCGTTGAGGAGTCT-3'	5'-GTGATCGAACCTTTCCAGTGCT-3'
per1	5'-CTGCCTCAGGCCCTCGA-3'	5'-GTCCGAGTGGCCAGGATCTT-3'
cry2	5'-TCAGCGTGAATGCAGGCA-3'	5'-AGGGCAGTAGCAGTGGAAGAAC-3'
collagen2	5'-GACCTGCCGGTGAACAAG-3'	5'-GGTACCAGGTTCTCCATCTCT-3'
aggrecan	5'-GCAGACATTGATGAGTGCCTC-3'	5'-CTCACACAGGTCCCCTCTGT-3'
mmp3	5'-CAATCCCTCTATGGACCTCCC-3	5′-CCCTCCATGAAAAGACTCAGAGG-3′
mmp13	5'-ACCCAGCCCTATCCCTTGAT-3′	5′-GGCCCAGAATTTTCTCCCTCT-3′
Tbp	5'-TGGGATTGTACCACAGCTCCA-3'	5'-CTCATGATGACTGCAGCAAACC-3'

Chondrocytes were collected after experiments and prepared using G-Actin/F-Actin In vivo Assay Biochem Kit according to the manufacturer’s instructions (Cytoskeleton, Inc, USA). Briefly, cells were scraped and cell lysate were homogenized at 37°C. Then, the cells were centrifuged at 100,000 g, 37°C for one hour to pellet F-actin and leave G-actin in the supernatant. Add F-actin depolymerization buffer to allow actin depolymerization. Finally, add sodium dodecyl sulphate (SDS) buffer to each of the pellet and supernatant samples and mix for actin quantitation by SDS polyacrylamide gel electrophoresis (SDS-PAGE) and immunoblotting analysis.

We have reported the method in an earlier study.¹⁴ Chondrocytes were collected after experiments and prepared using nuclear and cytoplasmic protein extraction kit (Beyotime). The cytoplasmic and nuclear proteins were sequentially extracted according to the manufacturer’s instructions. Finally, the cytoplasmic and nuclear components were subjected to immunoblotting.

The compressed chondrocytes were washed with PBS pre-cooled at 4°C and lysed using radioimmunoprecipitation assay (RIPA) buffer. The cell lysate was centrifuged and the supernatant was separated at 14,000× g and 4°C for ten minutes. Proteins were measured using the bicinchoninic acid (BCA) protein assay. SDS-PAGE electrophoresis, transferring, fixation, incubation, and development were performed. The concentrations of primary antibodies for Col 2, Aggrecan, Bmal1, Clock, Per1, and Cry2 were 1:1,000. β-actin was used as control, and concentration of antibody for β-actin was 1:3,000. Finally, the expressions of the specific proteins were quantified as relative expressions of target proteins against expression of β-actin in each protein sample.

The methods have previously been reported by Yadav et al.¹⁵ Briefly, chondrocytes were grown in six-well culture plates containing coverslips and subjected to mechanical loading. After that, cells were fixed, blocked, and permeabilized as mentioned in the immunofluorescence (IF) preparation procedure. The cells were then stained with FITC-Phalloidin (Sigma-Aldrich) for 30 minutes in the dark room. After Hoechst staining for five minutes, the cells were mounted on slides and images were obtained at 495 nm in Olympus FluoView FV1000.

To genetically downregulate Cry2 level in rat TMJ cartilage and chondrocytes, we used adeno-associated virus (AAV) vectors carrying green fluorescent protein (GFP) and the short hairpin RNA (shRNA) against cry2 mRNA (AAV-GFP-shCry2). AAV vector carrying both scrambled shRNA and GFP (AAV-GFP-scrambled) was used as control. Delivery of AAV vectors into the TMJ condyle explants and chondrocytes was accomplished as described previously.¹² Briefly, a preliminary experiment was conducted to obtain optimal value of multiplicity of infection (MOI) in explants and chondrocytes. After that, the appropriate amounts of AAV were added into culturing medium of explants and chondrocytes according to the appropriate MOI values (MOI = 50 for chondrocytes and MOI = 100 for explants). After three days of AAV transfection, the efficiency of transfections was evaluated by observing green fluorescence and Cry2 expressions at both mRNA and protein levels using IF, immunoblotting, and qRT-PCR.

Statistical analysis

All measurements were performed at least three times. Shapiro-Wilk tests were used to confirm normal distribution. Statistical analysis was performed using GraphPad Prism 6.0 and the Statistical Product and Service Solutions (SPSS) 11.0 software (USA). One-way analysis of variance (ANOVA) was used for multiple comparisons among the groups. All data are presented as the mean (SD). A p-value of < 0.05 was considered statistically significant.

Results

To test the effect of mechanical unloading on the TMJ cartilage homeostasis, the TMJ condyles of rats were cultured for up to one week ex vivo (Figure 1a, Supplementary Figure b). H&E and Safranin-O-Fast Green staining, as well as IF and immunohistochemistry (IHC) staining of key cartilage matrix components and proteases, was performed on one-day and one-week cultured TMJ explants (Figures 1b to 1j). A minor decrease in Safranin-O staining could be seen in the one-week sample compared with the one-day sample, which is indicative of cartilage degeneration after one week of ex vivo culture (Figures 1c and 1d). The mRNA and protein levels of collagen II (Col2) and aggrecan (Agcn) were lower, while those of MMP3 and MMP13 were higher, in one-week samples than one-day samples (Figures 1e to 1j).

Fig. 1.

Mechanical unloading disrupted temporomandibular joint (TMJ) homeostasis ex vivo. a) Diagram of the experimental design to compare the effects of one-day and one-week static culturing on matrix proteins and proteases expressions of TMJ condyle explants and chondrocytes. b) and c) Haematoxylin and eosin (H&E) and Safranin-O-Fast Green staining of one-day and one-week static cultured TMJ condyle explants. Scale bar: 200 μm. d) Quantification of Safranin O staining density in one-day and one-week static cultured TMJ condyle explants. e) Immunohistochemistry (IHC) staining of matrix metalloproteinase 3 (MMP3) and MMP13 in one-day and one-week static cultured TMJ condyle explants (n = 3 rats/group). Scale bar: 50 μm. f) Quantification of mean optical density of MMP3 and MMP13 staining in one-day and one-week static cultured TMJ condyle explants. g) Quantification of relative messenger RNA (mRNA) expressions of mmp3 and mmp13 in cartilage of one-day and one-week static cultured TMJ condyle explants, with β-actin as the normalizing gene. h) Immunofluorescence (IF) staining of Aggrecan and Collagen 2 in one-day and one-week static cultured TMJ condyle explants (n = 3 rats/group). Scale bar: 50 μm. i) Quantification of mean fluorescence intensity of Aggrecan and Collagen 2 staining in one-day and one-week static cultured TMJ condyle explants. j) Quantification of relative mRNA expressions of aggrecan and collagen 2 in one-day and one-week static cultured TMJ condyle explants, with β-actin as the normalizing gene. Values are expressed as the mean (SD). *p < 0.05. **p < 0.01, one-way analysis of variance.

Primary chondrocytes isolated from rat TMJ cartilage were cultured for one day or one week without passaging (Supplementary Figure c). The protein levels of MMP3 and MMP13 increased in the one-week group compared to the one-day group, while those of Aggrecan and Col2 decreased (Figures 2a and 2b). The mRNA expressions of agcn decreased, while those of mmp3 and mmp13 increased in the one-week group compared to the one-day group (Figure 2c). IF staining of MMP3, MMP13, Agcn, and Col2 in primary chondrocytes cultured for one day or one week showed similar results (Figures 2d and 2e).

Fig. 2.

Mechanical unloading disrupted temporomandibular joint (TMJ) chondrocyte homeostasis in vitro. a) Representative immunoblotting results of core clock proteins expressions at different timepoints of one-day and one-week static cultured chondrocytes. b) Quantification of immunoblotting results of core clock protein expressions at different timepoints of one-day and one-week static cultured chondrocytes. c) Quantification of relative messenger RNA (mRNA) expressions of core clock genes at different timepoints of one-day and one-week static cultured chondrocytes, with β-actin as the normalizing gene. d) Immunofluorescence (IF) staining of Aggrecan and Collagen 2 in one-day and one-week static cultured TMJ chondrocytes. e) IF staining of matrix metalloproteinase 3 (MMP3) and MMP13 in one-day and one-week static cultured TMJ chondrocytes. Values are expressed as the mean (SD). *p < 0.05. **p < 0.01 (one-day groups vs one-week groups), one-way analysis of variance.

For evaluation of circadian clock in both TMJ explants and chondrocytes, explants and cells were first synchronized by dexamethasone (Dex), followed by culturing for one day or one week, and samples were collected every six hours in the one-day and one-week groups (Figure 3a). Both mRNA and protein levels of brain and muscle Arnt-like protein 1 (Bmal1), circadian locomotor output cycles kaput (Clock), Period 1 (Per1), and Cryptochrome 2 (Cry2) in cultured chondrocytes exhibited rhythmic expressions among the four timepoints in the one-day group (Figures 3b to 3d, Supplementary Figure d). The expression of Per2 did not show an oscillated pattern, while that of Cry1 was not observed, across all timepoints in the one-day group (Supplementary Figure e). Similar results were obtained in one-day cultured TMJ explant (Supplementary Figure f). In contrast, static culturing of both explants and chondrocytes for one week resulted in dampening of periodic levels of these core clock genes and proteins (Supplementary Figure f). Thus, both the ex vivo and in vitro results demonstrated that imbalanced expressions of cartilage matrix components (Col2 and Agcn) and matrix degrading proteinase (MMP3 and MMP13) correlated with disrupted circadian rhythm of core clock genes after one week of mechanical unloading.

Fig. 3.

Disrupted temporomandibular joint (TMJ) homeostasis correlated with weakened oscillations of key clock genes. a) Diagram of the experimental design to compare the effects of one-day and one-week static culturing on core clock gene expressions in TMJ condyle explants and chondrocytes. b) Quantification of relative messenger RNA (mRNA) expressions of core clock genes at different timepoints of one-day and one-week static cultured chondrocytes, with β-actin as the normalizing gene. c) Representative immunoblotting results of core clock proteins expressions at different timepoints of one-day and one-week static cultured chondrocytes. d) Quantification of immunoblotting results of core clock proteins expressions at different timepoints of one-day and one-week static cultured chondrocytes. Values are expressed as the mean (SD). *p < 0.05. **p < 0.01 (one-day groups vs one-week groups), one-way analysis of variance.

Cyclic compressive loading (1 Hz, 12 hours on and 12 hours off) with a range of magnitudes (50 Kpa, 100 Kpa, 200 Kpa, and 400 Kpa) were applied to TMJ explants and chondrocytes for one week (Figure 4a). H&E and Safranin-O-Fast Green staining results of these explants are shown in Figures 4b and 4c. Lower magnitudes of mechanical loading (100 Kpa and 200 Kpa) were demonstrated to protect cartilage matrix from degradation, while higher magnitudes of mechanical loading (400 KPa) failed to have the same effect (Figure 4c). IHC staining of 100 Kpa- and 200 Kpa-compression groups displayed lower levels of MMP3 and MMP13, and higher levels of Col2 and Aggrecan, compared with unloaded one-week group and the other two loaded groups (Figures 4d to 4g).

Fig. 4.

Cyclic mechanical loading partially rescued cartilage from unloading-induced degeneration. a) Diagram of the experimental design to compare the effects of one-week mechanical loading on matrix protein and protease expressions, as well as oscillations of core clock genes, of temporomandibular joint (TMJ) condyle explants and chondrocytes. b) Haematoxylin and eosin (HE) staining of one-week static cultured TMJ condyle explants, as well as one-week mechanical loaded TMJ condyle explants with magnitudes of 50 Kpa,100 Kpa, 200 Kpa, and 400 Kpa. c) Safranin-O-Fast Green staining and quantification of Safranin O staining density of one-week static cultured TMJ condyle explants, as well as one-week mechanically loaded TMJ condyle explants with magnitudes of 50 Kpa, 100 Kpa, 200 Kpa, and 400 Kpa. d) and e) Immunohistochemical staining and quantification of mean optical density of matrix metalloproteinase 3 (MMP3) and MMP13 in one-week static cultured and one-week mechanically loaded TMJ condyle explants (n = 5 rats/group). Values are expressed as the mean (SD). *p < 0.05. **p < 0.01 (mechanical unloading vs mechanical loading). f) Immunofluorescence (IF) staining of Aggrecan and Collagen 2 in one-week static cultured and one-week mechanically loaded TMJ condyle explants (n = 5 rats/group). Scale bar: 50 μm. g) Quantification of mean fluorescence intensity of Aggrecan and Collagen 2 staining in one-week static cultured and one-week mechanically loaded TMJ condyle explants. Values are expressed as the mean (SD). *p < 0.05. **p < 0.01 (unloading groups vs loading groups), one-way analysis of variance.

For mechanical loading of cultured chondrocytes for one week, results proved that 50 Kpa- and 100 Kpa-compression groups had anabolic effects, while 200 Kpa- and 400 Kpa-compression groups had catabolic effects, as revealed by distinct levels of matrix components and proteases (Figures 5a to 5c). We chose the protocol (100 Kpa, 1 Hz, 12 hrs on and 12 hrs off) to demonstrate the effect of one-week mechanical loading on circadian rhythm of primary chondrocytes or TMJ explant, and samples were collected at four timepoints (6 hrs, 12 hrs, 18 hrs, and 24 hrs) on the seventh day (Figure 4a). Compared with the samples from one-week unloaded explants, the one-week loaded samples revealed rhythmic expression patterns of Bmal1, Clock, Per1, and Cry2 (Figures 5d to 5h), although the amplitude and phase were different from the one-day unloaded samples in Figures 3b to 3d. Similarly, in one-week mechanically loaded chondrocytes, oscillated mRNA and protein expressions of Bmal1, Clock, Per1, and Cry2 were also observed (Supplementary Figure g). Therefore, these data uncovered that mechanical loading acted as a timing cue that could entrain the rhythmicity of circadian clock in chondrocytes, and maintain the homeostasis of one-week cultured cartilage explants.

Fig. 5.

Cyclic mechanical loading maintained the circadian rhythm of chondrocytes in vitro and ex vivo. a) Quantification of relative messenger RNA (mRNA) expressions of matrix metalloproteinase 3 (mmp3), mmp13, aggrecan, and collagen 2 in one-week static cultured and one-week mechanically loaded chondrocytes, with β-actin as the normalizing gene. b) Representative immunoblotting results of MMP3, MMP13, Aggrecan, and Collagen 2 expressions in one-week static cultured and one-week mechanically loaded chondrocytes. c) Quantification of immunoblotting results of MMP3, MMP13, Aggrecan, and Collagen 2 expressions in one-week static cultured and one-week mechanically loaded chondrocytes. d) Immunofluorescence (IF) staining of Bmal1 at different timepoints of one-week static cultured and one-week mechanically loaded TMJ condyle explants. e) IF of Clock at different timepoints of one-week static cultured and one-week mechanically loaded TMJ condyle explants. f) IF staining of Per1 at different timepoints of one-week static cultured and one-week mechanically loaded TMJ condyle explants. g) IF staining of Cry2 at different timepoints of one-week static cultured and one-week mechanically loaded TMJ condyle explants (n = 3 rats/group). h) Quantification of mean fluorescence intensity of Bmal1, Clock, Per1, and Cry2 staining in one-week static cultured and one-week mechanically loaded TMJ condyle explants. Values are expressed as the mean (SD). *p < 0.05. **p < 0.01. #p < 0.001 (unloading groups vs loading groups), one-way analysis of variance. N.S., not significant.

The core clock proteins (Bmal1, Clock, Per1, and Cry2) have been demonstrated to exhibit variable tissue-specific patterns of nuclear-cytoplasmic cycling.¹⁶ We first compared the subcellular localizations of these proteins in static cultured chondrocytes (Figure 6a). Both Bmal1 and Clock proteins stayed in nuclei at all timepoints of one-day static cultured chondrocytes (Figures 6b to 6d), and almost disappeared in one-week static cultured chondrocytes (Supplementary Figure h). Interestingly, the subcellular localizations of Per1 and Cry2 displayed circadian patterns of nuclear-cytoplasmic localizations in one-day static cultured chondrocytes (Figures 6b to 6d). However, the nuclear-cytoplasmic cycling of Per1 and Cry2 weakened in one-week static cultured chondrocytes, and only a minor portion could be detected in the cytoplasm (Supplementary Figure h).

Fig. 6.

The spatiotemporal subcellular localizations of circadian rhythm genes. a) Diagram of the experimental design. b) Representative immunoblotting results of nuclear-cytoplasmic localizations of Bmal1, Clock, Per1, and Cry2 at different timepoints of one-day static cultured chondrocytes. c) Quantification of nuclear and cytoplasmic portions of Bmal1, Clock, Per1, and Cry2 proteins at different timepoints of one-day static cultured chondrocytes. Values are expressed as the mean (SD). *p < 0.05, **p < 0.01 (comparing between different timepoints). #p <0.05, ##p < 0.01, ###p < 0.001 (comparing between nuclear portion and cytoplasmic portion), one-way analysis of variance. d) Immunofluorescence staining of Bmal1, Clock, Per1, and Cry2 at different timepoints of one-day static cultured chondrocytes. Scale bar: 20 μm.

When chondrocytes were subjected to compression (100 Kpa, 1 Hz, 12 hrs on and 12 hrs off) for one week, the nuclear-cytoplasmic shuttling of Per1 and Cry2 proteins were sustained compared with the one-week static control group. Specifically, Per1 and Cry2 translocated to nuclei during 12-hour loading time (12-hour and 18-hour timepoints), and accumulated in cytoplasm during 12-hour unloading time (six-hour and 24-hour timepoints) (Figure 7). The subcellular localizations of Bmal1 and Clock proteins were not affected by mechanical loading. Thus, the effects of cyclic mechanical loading on nuclear-cytoplasmic shuttling of core Clock proteins were more potent on Per1 and Cry2.

Fig. 7.

Cyclic mechanical loading affected the spatiotemporal subcellular localizations of PER1 and CRY2. a) Representative immunoblotting results of nuclear-cytoplasmic localizations of Bmal1, Clock, Per1, and Cry2 at different timepoints of one-week mechanical loaded chondrocytes. b) Quantification of nuclear and cytoplasmic portions of Bmal1, Clock, Per1, and Cry2 proteins at different timepoints of one-week mechanical loaded chondrocytes. c) Immunofluorescence staining of Bmal1, Clock, Per1, and Cry2 at different timepoints of one-week mechanical loaded chondrocytes. Scale bar: 20 μm. Values are expressed as the mean (SD). *p < 0.05, **p < 0.01 (comparing between different timepoints). #p < 0.05, ##p < 0.01, ###p < 0.001 (comparing between nuclear portion and cytoplasmic portion), one-way analysis of variance.

The actin of cytoskeleton is known to be dynamic, switching between filamentous actin (F-actin) and monomeric globular actin (G-actin). Rho-associated kinases (ROCK) are upstream signalling molecules transmitting mechanical stimuli to modulate actin dynamics.¹⁷ We assumed that mechanical loading-induced ROCK activation promoted the formation of F-actin, which was required for Cry2 and Per1 proteins to translocate from cytoplasm into nuclei. The TMJ primary chondrocytes mainly expressed ROCK1 but not ROCK2 (Supplementary Figure i). Phospho-ROCK1 (p-ROCK) level was shown to be oscillated upon cyclic mechanical loading at different timepoints of the one-week group, but remained unchanged in the unloaded group (Figures 8a to 8c). In addition, there were dynamic changes in the ratio of F-actin/G-actin in the one-week loading group, but not in the one-week static control group (Figures 8b to 8d). The timepoints when phospho-ROCK1 level and F-actin/G-actin ratio increased correlated with nuclear localization of Cry2 and Per1 (Figure 7), suggestive of ROCK-mediated actin polymerization in regulating nuclear shuttling of Cry2 and Per1.

Fig. 8.

Cyclic mechanical loading promoted ROCK1 activation-mediated actin polymerization. a) Diagram of the experimental design. b) Representative immunoblotting results of ROCK, p-ROCK, G-actin, and F-actin at different timepoints of one-week static and mechanical loaded chondrocytes. c) Quantification of relative protein levels of p-ROCK, G-actin, and F-actin, as well as ratios of G-actin/F-actin, at different timepoints of one-week static and mechanical loaded chondrocytes. d) FITC-Phalloidin staining of chondrocytes at different timepoints of one-week static and mechanical loading groups. e) FITC-Phalloidin staining of chondrocytes with addition of Y27632 at different timepoints ofone-week mechanical loading groups. Scale bar: 20 μm. f) Representative immunoblotting results of ROCK, p-ROCK, G-actin, and F-actin at different timepoints of one-week mechanical loaded chondrocytes with the addition of DMSO or Y27632. g) Quantification of relative protein levels of p-ROCK, G-actin, and F-actin, as well as ratios of G-actin/F-actin, at different timepoints of one-week mechanical loaded chondrocytes with the addition of dimethyl sulfoxide. Values are expressed as the mean (SD). *p < 0.05,**p < 0.01 (comparing between different timepoints). #p < 0.05, ##p < 0.01, one-way analysis of variance. N.S., not significant.

To further elucidate the role of ROCK1 in circadian rhythm of chondrocytes, the specific inhibitor Y27632 was applied along with mechanical loading on chondrocytes, and actin fibres showed increased depolymerization as demonstrated by decreased F-actin/G-actin ratio (Figures 8e to 8g). In addition, the mechanical loading-induced oscillations of core clock proteins (Bmal1, Clock, Per1, and Cry2) were weakened by the addition of Y27632 (Figures 9a and 9b). Notably, the rhythmic nuclear translocation of Cry2, but not Per1, was partially inhibited by the addition of Y27632 in the one-week loading group (Figures 9c to 9f), compared with one-week loaded chondrocytes without Y27632 addition (Figure 7). These data confirmed that Cry2, but not Per1, acted as a mechanosensory circadian factor downstream of ROCK1-mediated actin polymerization in chondrocytes.

Fig. 9.

Cyclic mechanical loading promoted CRY2 nuclear translocation through ROCK1 activation. a) Representative immunoblotting results of Bmal1, Clock, Per1, and Cry2 at different timepoints of one-week mechanical loaded chondrocytes with the addition of DMSO or Y27632. b) Quantification of relative protein levels of Bmal1, Clock, Per1, and Cry2 at different timepoints of one-week mechanical loaded chondrocytes with the addition of DMSO or Y27632. Values are expressed as the mean (SD). *p < 0.05, **p < 0.01. c) Representative immunoblotting results of nuclear-cytoplasmic localizations of Per1 and Cry2 at different timepoints of one-week mechanical loaded chondrocytes with the addition of DMSO or Y27632. d) Quantification of relative nuclear and cytoplasmic portions of Per1 and Cry2 at different timepoints of one-week mechanical loaded chondrocytes with the addition of Y27632. e) Immunofluorescence (IF) staining of Cry2 at different timepoints of one-week mechanical loaded chondrocytes with the addition of Y27632. f) IF staining of Per1 at different timepoints of one-week mechanical loaded chondrocytes with the addition of Y27632. Scale bar: 20 μm. Values are expressed as the mean (SD). *p < 0.05, **p < 0.01 (comparing between different timepoints). #p < 0.05, ##p < 0.01, one-way analysis of variance. N.S., not significant (comparing between nuclear and cytoplasmic proteins).

Finally, we aimed to investigate whether Cry2 was responsible for the effects of mechanical loading on circadian rhythm and cartilage homeostasis. Down-regulation of Cry2 expression in chondrocytes and cartilages was performed by infection with AAV vectors containing short hairpin RNA (shRNA) against Cry2 mRNA (Figure 10a), and the efficiency in Cry2 down-regulation was confirmed at both mRNA and protein levels (Supplementary Figure i). The mRNA and protein levels of core Clock proteins (Bmal1, Clock, and Per1) from Cry2 down-regulated chondrocytes showed weakened oscillations in the one-week loaded group, compared with the wild type (WT) chondrocytes (Figures 10b to 10d). Similar results were also obtained from Cry2 down-regulated TMJ condyle explants that had been cultured for one week under cyclic mechanical loading (Supplementary Figure k).

Fig. 10.

CRY2 was involved in cyclic mechanical loading-sustained circadian rhythm of chondrocytes. a) Diagram of theexperimental design. b) Quantification of the effect of CRY2 down-regulation on one-week mechanical loading-mediated oscillations of bmal1, clock,and per1 at relative messenger RNA (mRNA) levels. c) Representative immunoblotting results showing the effect of CRY2 down-regulation on one-week mechanicalloading-mediated oscillations of Bmal1, Clock, and Per1 in chondrocytes. d) Quantification of immunoblotting results showing the effect of CRY2down-regulation on one-week mechanical loading-mediated oscillations of Bmal1, Clock, and Per1 in chondrocytes. Values are expressed as the mean (SD). *p < 0.05. **p < 0.01, one-way analysis of variance. N.S., not significant.

H&E and Safranin-O-Fast Green staining of explants demonstrated that Cry2 down-regulation partially impaired the effect of one-week mechanical loading on cartilage integration (Figures 11a and 11b). Furthermore, the expressions of cartilage matrix proteins and proteases in one-week mechanically loaded chondrocytes and explants were also compared between the Cry2 down-regulated group and WT group. Mechanical loading-induced increased expressions of matrix proteins (Col2 and Agcn) and reduction of proteases (MMP3 and MMP13) in WT cartilage explants were partially blunted by Cry2 down-regulation (Figures 11c to 11f). Similarly, Cry2 down-regulated chondrocytes also had reduced levels of Agcn and elevated levels of MMP3 and MMP13 compared with WT cells, after being subjected to one week of mechanical loading (Figures 11g to 11i). Altogether, these results provide strong evidence that Cry2 was involved in periodic mechanical loading-stimulated cycling of circadian oscillation in the chondrocytes and homeostasis of cartilage.

Fig. 11.

CRY2 was involved in cyclic mechanical loading-sustained homeostasis of cartilage. a) Haematoxylin & eosin (HE) and Safranin-O-Fast Green staining of one-week unloading temporomandibular joint (TMJ) condyle explants, one-week loading TMJ condyle explants, and one-week loading TMJ condyle explants with Cry2 down-regulation. Scale bar: 200 μm. b) Quantification of Safranin O staining density in one-week unloading TMJ condyle explants, one-week loading TMJ condyle explants, and one-week loading TMJ condyle explants with Cry2 down-regulation. c) Immunohistochemical (IHC) staining of matrix metalloproteinase 3 (MMP3) and MMP13 in one-week unloading TMJ condyle explants, one-week loading TMJ condyle explants, and one-week loading TMJ condyle explants with Cry2 down-regulation (n = 5 rats/group). Scale bar: 50 μm. d) Quantification of mean optical density of MMP3 and MMP13 staining in one-week unloading TMJ condyle explants, one-week loading TMJ condyle explants, and one-week loading TMJ condyle explants with Cry2 down-regulation. e) Immunofluorescence staining of Aggrecan and Collagen 2 in one-week unloading TMJ condyle explants, one-week loading TMJ condyle explants, and one-week loading TMJ condyle explants with Cry2 down-regulation (n = 5 rats/group). Scale bar: 50 μm. f) Quantification of mean fluorescence intensity of Aggrecan and Collagen 2 staining in one-week unloading TMJ condyle explants, one-week loading TMJ condyle explants, and one-week loading TMJ condyle explants with Cry2 down-regulation. g) Quantification of relative mRNA levels of mmp3, mmp13, aggrecan, and collagen 2 in one-week unloading chondrocytes, one-week loading chondrocytes, and one-week loading chondrocytes with Cry2 down-regulation, with β-actin as the normalizing gene. h) Representative immunoblotting results of MMP3, MMP13, Aggrecan, and Collagen 2 in one-week unloading chondrocytes, one-week loading chondrocytes, and one-week loading chondrocytes with Cry2 down-regulation. i) Quantification of immunoblotting results of MMP3, MMP13, Aggrecan, and Collagen 2 in one-week unloading chondrocytes, one-week loading chondrocytes, and one-week loading chondrocytes with Cry2 down-regulation. Values are expressed as the mean (SD). *p < 0.05. **p < 0.01, one-way analysis of variance. N.S., not significant.

Discussion

To our knowledge, this is the first study to uncover the relationship between mechanical loading and circadian clock in TMJ cartilage.

Recently, global transcriptomic studies have confirmed the rhythmic expressions of core clock genes in cartilage.^5,6,18 Moreover, many of the genes involved in cartilage homeostasis, such as ECM structural components, remodelling enzymes, and growth factors, are targets of clock genes and exhibited rhythmic expressions.^5,6 In our study, one-week ex vivo culturing of TMJ explants resulted in cartilage degeneration accompanied by weakened oscillations of Bmal1, Clock, Per1, and Cry2, providing direct evidence of the relationship between circadian clock and cartilage metabolism.

Circadian clock can be entrained by various timing cues, such as light, temperature, hormones, sleeping, feeding, and exercise, to modulate tissue homeostasis and functions.¹⁹ Several studies have revealed that exercise acts as an important timing cue to maintain the normal cycling of circadian clock in skeletal tissues (bone, cartilage, and muscle).²⁰ For example, the bones’ responses to mechanical loading are dependent upon loading time of day.²¹ Another study revealed that mechanical loading could promote muscle cell proliferation through regulating core clock genes, while mechanical unloading alleviated the oscillation of the circadian cycle.^22-24 With regard to cartilage, it has been shown that chondrogenesis caused by mechanical loading was partially attributed to the rhythmic expressions of the core clock genes, and that chondrogenesis was blocked by clock modulator longdaysin.⁹ One study by Heywood et al⁸ reported that cyclic mechanical stretch acted as a timing cue to reset chondrocyte clock in vitro. Dudek et al¹⁰ provided similar evidence that 12-hour compression followed by 12-hour relaxation could reset the circadian clock phase and amplitude in cartilage and intervertebral disc tissues in tissue explant cultures. Consistent with the results of these previous studies, our study uncovered an interesting phenomenon that daily periodic mechanical compression could maintain TMJ cartilage homeostasis for at least one week. Thus, all these studies together with ours illustrate that mechanical loading is a critical and specific zeitgeber to correlate circadian clock with normal function of these skeletal tissues.

Our study compared the effects of compressive load with a wide range (from 50 Kpa to 400 KPa) on TMJ chondrocytes and TMJ explants, and found that a lower magnitude (100 to 200 KPa) of compression was beneficial to sustain circadian clock and normal metabolism of chondrocytes and TMJ cartilage, while a higher magnitude of compression (400 KPa) disrupted the circadian clock and degraded the cartilage matrix. Thus, we assumed that mechanical stimuli within physiological range are essential to exert a beneficial effect on circadian rhythm. This was supported by studies showing a dampened circadian clock in chondrocytes or nucleus pulposus cells, as well as degraded intervertebral disc, by excessive mechanical stimuli.^11,25,26

How does a mechanical signal transmit into the cell and act as a resetting cue for the circadian clock system? The RhoA/ROCK pathway is a classical mechanosensitive pathway, and was proposed to correlate with circadian clock in various cell types.^11,23,26 In addition, Dudek et al¹⁰ uncovered a novel idea that compressive loading resets the circadian clock of cartilage through compression-generated hyperosmolarity and activation of the PLD2-mTORC2-AKT-GSK3β pathway in the tissue niche. Our study proposed another novel mechanism that mechanical loading modulates the circadian clock of chondrocytes through ROCK-mediated actin polymerizing-depolymerizing dynamics. The involvement of actin dynamics in modulating the circadian clock could be explained by the nuclear translocation of myocardin-related transcription factor-B (MRTF-B), which is sequestered to the cytoplasm by G-actin and released into nuclei upon F-actin formation. Nuclear translocation of MRTF stimulated the downstream target gene expression, including circadian clock genes, such as clock, per1, per2, Nr1d1, and Nfil3.^27-29 In addition to actin-MRTF signalling, a study by Abenza et al³⁰ focused on another important mechanosensory pathway, YAP/TAZ-TEAD, which affected the expressions of clock genes bmal1, cry1, per2, and rev-erb in matrix stiffening-induced actin polymerization. Overall, all these studies and ours provide a network of mechanosensors that could entrain the circadian clock in various cell types under distinct mechanical conditions.

So far, the intracellular translocations of core clock proteins, as well as how they are regulated under various conditions, have not been extensively explored. Our study revealed a novel phenomenon of nuclear-cytoplasmic shuttling of Per1 and Cry2 proteins in mechanically loaded chondrocytes. Specifically, Per1 and Cry2 entered nuclei during 12 hours of mechanical loading, while they transferred to cytoplasm during 12 hours of mechanical unloading. Furthermore, our data attest that mechanical loading-induced nuclear entry of Cry2, but nor Per1, was dependent on ROCK activation and actin polymerization. Thus, we proposed a hypothesis that mechanical stimuli promoted ROCK-mediated actin polymerization, subsequently released Cry2 from being sequestered by depolymerized actin, and guided Cry2 into nuclei. This was consistent with previous studies that there was a close relationship between Cry cytoplasmic localization and cytoskeletal dynamics.^31,32 Moreover, we proposed that Cry2 nuclei-cytoplasmic shuttling resulted in oscillations of circadian rhythms in chondrocytes in response to rhythmic mechanical compression. This notion was supported by the fact that oscillated abundance of Bmal1 and/or Clock was not necessary to generate rhythmic expressions of per and cry,³³ but the rhythmic availability of nuclear Per:Cry protein complexes was sufficient to determine the rhythmic gene expressions.^34-36

In the last part of our study, we clarified the potential role of Cry2 in mechanical loading-sustained circadian rhythm and cartilage homeostasis by AAV-mediated Cry2 downregulation in chondrocytes and cartilage explants. This was partially consistent with other studies showing that Cry2 knock-down or knock-out resulted in disrupted circadian rhythm and cartilage degeneration.^37,38 Although other core clock genes, such as bmal1^18,39,40 and clock,^41,42 have also been shown to be involved in regulating cartilage homeostasis, the exact functions of these clock genes were still not conclusive. Whether these core clock genes also play important roles in maintaining cartilage homeostasis in a proper mechanical environment requires further investigation.

This study has several limitations. First, the circadian clock is a complex system consisting of many clock-related genes, and our study only investigated the core clock genes (bmal1, clock, per1, and cry2). A transcriptomic result would provide more information on mechanical loading-regulated circadian rhythm in TMJ chondrocytes. Moreover, an appropriate in vivo animal model of TMJ immobilization and functional restriction should be developed to more deeply understand the role of mechanical loading and circadian clock in modulating TMJ cartilage homeostasis in future.

In conclusion, the study provided direct evidence that appropriate mechanical loading was efficient in sustaining circadian oscillations of core clock genes in chondrocytes and cartilage homeostasis in TMJ condyle, suggesting that mechanical loading is a vital zeitgeber for TMJ cartilage. In addition, we uncovered a novel role of mechanical compression-induced ROCK activation and actin polymerization in regulating Cry2 nuclear-cytoplasmic translocations, which partially contributed to mechanical loading-generated oscillations of core clock genes and homeostasis of TMJ cartilage.

Data Availability

Data are available from the corresponding author upon reasonable request.