Theobroma cacao improves bone growth by modulating defective ciliogenesis in a mouse model of achondroplasia

A gain-of-function mutation in the fibroblast growth factor receptor 3 gene (FGFR3) results in achondroplasia (ACH), the most frequent form of dwarfism. Constitutive activation of FGFR3 impairs bone formation and elongation and many signal transduction pathways. Identification of new and relevant compounds targeting the FGFR3 signaling pathway is of broad importance for the treatment of ACH, and natural plant compounds are prime drug candidate sources. Here, we found that the phenolic compound (-)-epicatechin, isolated from Theobroma cacao, effectively inhibited FGFR3’s downstream signaling pathways. Transcriptomic analysis in an Fgfr3 mouse model showed that ciliary mRNA expression was modified and influenced significantly by the Indian hedgehog and PKA pathways. (-)-Epicatechin is able to rescue mRNA expression impairments that control both the structural organization of the primary cilium and ciliogenesis-related genes. In femurs isolated from a mouse model (Fgfr3Y367C/+) of ACH, we showed that (-)-epicatechin eliminated bone growth impairment during 6 days of ex vivo culture. In vivo, we confirmed that daily subcutaneous injections of (-)-epicatechin to Fgfr3Y367C/+ mice increased bone elongation and rescued the primary cilium defects observed in chondrocytes. This modification to the primary cilia promoted the typical columnar arrangement of flat proliferative chondrocytes and thus enhanced bone elongation. The results of the present proof-of-principle study support (-)-epicatechin as a potential drug for the treatment of ACH.


INTRODUCTION
The most common form of dwarfism, achondroplasia (ACH), is caused by a gain-of-function mutation in the FGFR3 gene encoding fibroblast growth factor receptor 3. 1 FGFR3 gain-offunction mutations are also associated with hypochondroplasia, mild dwarfism, thanatophoric dysplasia (TD), and severe and lethal dwarfism. 2 Defective FGFR3 signal transduction impairs intracellular downstream signaling pathways, including extracellular signalregulated kinases 1 and 2 (ERK1/2), p38, phosphoinositide 3kinase/protein kinase B (PI3K/AKT), phospholipase Cγ (PLCγ), and signal transducers and activators of transcription (STATs), thus affecting chondrocyte proliferation, differentiation, and bone elongation. 2 FGFR3 plays a significant role in growth plate development, acting to inhibit both the rate of chondrocyte proliferation and differentiation and the interactions with the Indian hedgehog (Ihh) signaling pathway to control chondrocyte formation. 3 Ihh signaling, which plays an essential role in chondrocyte differentiation, is fully dependent on primary cilia. 4 FGF signaling also regulates the length of the primary cilia in many tissues. 5,6 FGFR3 signaling interacts with Hedgehog and the serine/threonine kinase intestinal cell kinase (Ick), which is involved in ciliogenesis and participates in the control of ciliary length. 7,8 For ACH, we demonstrated defective primary cilium elongation in mouse and human chondrocytes. [7][8][9] The regular alignment of primary cilia is responsible for columnar-stacked chondrocytes in growth plate cartilage. 10 The primary cilium is essential for the regulation of chondrocyte rotation, as demonstrated by the presence of defective primary cilium biosynthesis and/or function in many skeletal ciliopathies characterized by short ribs, short limbs and polydactyly. [11][12][13] Interestingly, some clinical features of skeletal ciliopathies (short ribs and short limbs) are shared with achondroplasia. 9 In recent years, a significant body of work has focused on treatments for ACH. Nonsurgical therapeutic strategies have been developed from insights gained from preclinical studies using Ach mouse models. 2,14 There are many therapeutic approaches with various mechanisms, including: (i) targeting the FGF ligand (recifercept) and aptamer (APT-F2P/RBM 007), (ii) targeting FGFR (anti-FGFR3 antibody-B701/vofatamab), (iii) inhibiting tyrosine kinase activity (BGJ398/infigratinib), or (iv) using a C-type natriuretic peptide (CNP) (TransCon CNP, BMN111/ vosoritide) analog 15 to antagonize the mitogen-associated protein kinase (MAPK) pathway. The vosoritide approach is currently the furthest along in the clinical development pathway. 16,17 Identification of new and relevant compounds targeting the FGFR3 signaling pathway is of broad importance for the treatment of FGFR3-related chondrodysplasia. Natural plant compounds are prime sources of drug candidates. 18 Plant polyphenols such as Theobroma cacao contain flavon-3-ols and polyphenols that have long been considered to have relevant biological activities for the treatment of a number of diseases and are known to act in various ways on MAPK signaling pathways. For example, these compounds can: (i) reduce reactive oxygen species production and ERK1/2 and p38 phosphorylation in neurons, 19 (ii) inhibit adipocyte differentiation through AMP-activated protein kinase (AMPK) and the ERK1/2 signaling pathways, 20 and (iii) modulating antioxidant enzyme activities through ERK1/2 and regulating glucose production through AMPK modulation in liver cells. 21 In this study, we identified (-)-epicatechin from a Theobroma cacao extract as a drug candidate. (-)-Epicatechin is able to (i) induce primary cilium elongation in chondrocytes and modify the differentially expressed genes in the growth plate cartilage of Fgfr3 Y367C/+ mice, (ii) decrease FGFR3's downstream signaling pathways in cell and organ cultures, and (iii) promote femur growth elongation in both femur explants and live Fgfr3 Y367C/+ mice. Our results highlight this molecule's specific action on the Ihh pathway related to primary cilia and suggest that (-)-epicatechin could be developed as a means to correct bone growth defects in patients with achondroplasia.

RESULTS
Theobroma cacao extract restores abnormal activation of the FGFR3 pathway and primary cilium defects in mutant Fgfr3 chondrocytes To investigate the therapeutic efficacy of Theobroma cacao on the abnormal activation of the FGFR3 pathway, we fractionated a Theobroma cacao extract by combining solid-phase extraction with semipreparative HPLC 18 and collected a total of 11 fractions (fractions 1 to 11) ( Supplementary Fig. 1). The activities of the fractions were determined according to two relevant criteria considered a hallmark of FGFR3-related disorders, namely, (i) the ability to inhibit the MAPK pathway, a downstream effector of the activated FGFR3 pathway, 22 and (ii) the ability to promote primary cilia elongation. 7,9 We found by western blot that Theobroma cacao extract fraction 5 significantly lowered ERK1/2 phosphorylation in human ACH and TD chondrocytes ( Supplementary Fig. 2a, b). Moreover, we found by immunolabeling chondrocytes isolated from Fgfr3 Y367C/+ mice exhibiting a dwarf phenotype 23 that fraction 5 treatment increased the initial primary cilium length defects in Fgfr3 Y367C/+ chondrocytes (4.06 ± 2.15 µm) compared to nontreated Fgfr3 Y367C/+ chondrocytes (2.71 ± 0.75 µm) (Fig. 1a, b).
(-)-Epicatechin counteracts the abnormal activation of the FGFR3 signaling pathway (-)-Epicatechin has shown beneficial effects on primary cilia formation in Fgfr3 Y367C/+ chondrocytes. Therefore, we transiently transfected human embryonic kidney 293 (HEK293) cells with  Table 2). The defective bone elongation observed in FGFR3-related disorders is due, at least in part, to dysregulation of chondrocyte differentiation. The successive differentiation steps are regulated by the ERK-MAPK pathway. 31 Here, we evaluated the propensity of (-)-epicatechin to interact with FGFR3 downstream signaling pathways (several distinct MAPKs, including ERK1, ERK2, and p38 kinases). We found that (-)-epicatechin remained bound to ERK1, ERK2, and p38 with only small structural variations from the starting dynamic state (static state) (  Table 1). To validate the interactions of (-)-epicatechin with FGFR3's downstream signaling pathways in cartilage cells, we quantified the levels of ERK1/2 and p38 phosphorylation in primary chondrocytes isolated from Fgfr3 Y367C/+ mice 23 treated with (-)-epicatechin. Compared with untreated Fgfr3 Y367C/+ chondrocytes, the levels of ERK1/2 phosphorylation (Fig. 2h, i) and p38 phosphorylation (Fig. 2j, k) were significantly lower in mutant chondrocytes treated with (-)-epicatechin. To definitively eliminate procyanidin C1 as a drug candidate, in silico analyses demonstrated that procyanidin C1 failed to interact with ERK1, ERK2, and p38 (Supplementary Fig. 6 and Supplementary Table 2) and did not reduce the level of ERK1/2 phosphorylation in either primary Fgfr3 Y367C/+ murine or human FGFR3-Y373C chondrocytes (data not shown). These findings indicate that (-)-epicatechin might be a promising compound for repressing the constitutive activation of FGFR3's downstream signaling pathways in cartilage cells.
We then hypothesized that (-)-epicatechin treatment mainly stimulated proliferative chondrocytes in growth plate cartilage. To investigate chondrocyte proliferation, we immunolabeled the proliferating cell nuclear antigen (PCNA) protein in ex vivo femur cultures (Fig. 4e). Elevated expression of PCNA in mutant growth Epicatechin restores cilium defects and promotes bone growth L Martin et al.
No interation allowed  Fig. 4) and p38 (g). The complexes obtained after molecular docking (i.e., static states) are shown on the left, and those obtained after the molecular dynamics and hyperdynamics simulations (i.e., dynamic states) are shown on the right. The ATP-binding sites are presented as molecular surfaces in ice blue. h Representative western blots of the expression levels of p-ERK1/2 and ERK1/2 in primary chondrocytes isolated from the femoral cartilage of Fgfr3 Y367C/+ mice. i Changes in the ratio of p-ERK1/2 to ERK1/2 (n = 5). j Representative western blots of the expression levels of p-p38 and p38 in primary chondrocytes isolated from the femoral cartilage of Fgfr3 Y367C/+ mice. k Changes in the ratio of p-p38 to p38 (n = 5). Data are presented as the mean ± SD. ns, not significant; *P < 0.05; **P < 0.01; ***P < 0.001; ****P < 0.000 1 by two-tailed, unpaired t test. The gray density analysis for the western blot data was performed by using ImageJ software Epicatechin restores cilium defects and promotes bone growth L Martin et al.
Altogether, these data show that (-)-epicatechin has a beneficial effect on the proliferation/differentiation balance in Fgfr3 Y367C/+ chondrocytes.
These findings confirmed that in vivo, systemically administered (-)-epicatechin penetrated the growth plate cartilage, strongly downregulated the expression of key regulators of chondrocyte proliferation and differentiation and modified the length of the primary cilia of chondrocytes. These results demonstrate that (-)-epicatechin can be used to control long bone elongation in vivo in FGFR3-related diseases.

DISCUSSION
A growing body of research has identified a number of therapeutic approaches for the treatment of defective bone growth in achondroplasia, the most common form of dwarfism. 2 Our present results provide strong evidence for a potential therapeutic effect from (-)-epicatechin on the cartilage growth plate and ciliogenesis in FGFR3-related disorders. Over the last few decades, the beneficial effects of polyphenols on human health have been increasingly well documented. Several preclinical studies have demonstrated that (-)-epicatechin is effective against sarcopenia, 36 alleviates inflammation in lipopolysaccharideinduced lung injury, 37 prevents the development of dilated cardiomyopathy, 38 and improves vascular function. 39 Here, we provide physiological evidence (in cell-based and murine models of ACH) and molecular evidence to show that (-)-epicatechin treatment modifies bone growth. Although the role of fibroblast growth factor (FGF) in skeletal development is quite well understood, little is known about the intracellular signaling that mediates the overactivation of FGFR3. Our data suggest that phosphorylated ERK1/2 and p38 and total Sox9 proteins are regulated in response to FGF signaling. Both ERK1/2 and p38 phosphorylation levels and the Sox9 expression level are elevated in the growth plates of Fgfr3 Y367C/+ mice, emphasizing the dysregulation of these proteins in Fgfr3-related skeletal dwarfism. Literature data show that the expression of a constitutively active mutant of Mek1 rescues the skeletal overgrowth phenotype in Fgfr3-deficient mice and that p38, Sox9 and ERK1/2 upregulate hypertrophic differentiation. 31,40 Here, we visualized in silico that in the FGFR3 ATP-binding pocket, (-)-epicatechin clearly remained bound to ERK1, ERK2 and p38. Interestingly, in vivo, (-)-epicatechin treatment decreased the activation of p38, ERK1/2, and Sox9 and was associated with a greater hypertrophic chondrocyte volume, which is known to be a major determinant of the longitudinal bone growth rate. We further showed that (-)-epicatechin interacts with the tyrosine kinase pocket of ERK1/2 and p38, which suggests the presence of a direct link to the compound's effects on the MAPK signaling pathway.
Our results also showed how (-)-epicatechin rescues proliferating columnar chondrocytes with a flattened, stacked appearance. The clonal expansion of chondrocytes resulted in bone growth. At the cellular level, (-)-epicatechin rescued primary cilium elongation defects and thus modified cell division and columnar zone elongation. We therefore examined (-)-epicatechin as a treatment to directly affect growth plate organization and proliferation by modifying the length and the ϕ angle of the primary cilia in the cartilage tissue, thus improving bone growth. We previously demonstrated that sustained FGFR3 activity was associated with shorter cilia, 7,9 abnormal chondrocyte homeostasis, and disturbance of the growth plate's columnar organization via chondrocyte rotation. 9,10,13 Our in vitro and in vivo results show that (-)-epicatechin was associated with longer chondrocyte primary cilia. Our morphometric studies of the growth plate showed that (-)-epicatechin treatment modified chondrocyte alignment and the length and position of the primary cilia in the growth plate. These findings were supported by our transcriptomic analysis, highlighting (i) the abnormal expression of several genes involved in ciliogenesis and the Ihh and PKA signaling pathways and (ii) their normalization after (-)-epicatechin treatment. We were therefore not surprised to see that the expression levels of genes linked to skeletal ciliopathies were abnormal in models of ACH compared to controls; these genes included Dync2li1 (the pathogenic gene in short rib polydactyly syndrome) and Kif22 (the pathogenic gene in spondyloepimetaphyseal dysplasia with joint laxity). 41,42 Given the essential role of primary cilia in Hh signaling and the known interactions between members of the FGF and Hh pathways, 3 it is likely that inhibition of chondrocyte proliferation by upregulated FGF signaling is caused (at least in part) by inactivation of Ihh signaling. 43 Our transcriptomic data confirmed that the Hh signaling pathway is dysregulated in Fgfr3 Y367C/+ mice. Ptch1 is essential for limb development 44,45 and interaction with Smoothened and Hh signaling through the Gli1 regulator. [46][47][48][49] This upregulation of Ptch1 is in agreement with data from another mouse model of an Fgfr3 gain-of-function mutation. 50 The present study demonstrated that (-)-epicatechin treatment controls the Ihh-related primary cilia signaling pathway by acting on the Ihh/ Ptch1/Gli1/Smo cascade in cartilage. We also highlight a putative role of the PKA pathway and Adcy7 in Fgfr3 Y367C/+ growth plate Significance * **** **** **** **** **** The percent growth gains of the naso-anal, femur, tibia, humerus, ulna, and radius lengths with detailed values for (-)-epicatechin-injected Fgfr3 Y367C/+ animals. Data are presented as the mean ns Not significant *P < 0.05; **P < 0.01; ***P < 0.001; ****P < 0.000 1 by two-tailed, unpaired t test Epicatechin restores cilium defects and promotes bone growth L Martin et al.
cartilage. Given that PKA was found to regulate Hh signaling in primary cilia and adenylate cyclase was clearly identified in skeletal primary cilium, 51 we hypothesize that PKA interferes with Ihh signaling in FGFR3-related disorders. Taken together, our results demonstrate that (-)-epicatechin restores the impairments in the various genes controlling cilium elongation by acting on both the Ihh and PKA signaling pathways involved in bone growth regulation.
A better mechanistic understanding of the regulation of primary cilia formation in vivo provides foundations for therapeutic interventions in FGFR3-related disorders. Although the current treatment for ACH restores chondrocyte differentiation via the MAPK pathway, it fails to restore the columnar arrangement of chondrocytes. 15,16 This is in contrast with (-)-epicatechin, which (in addition to chondrocyte differentiation) restores the columnar arrangement by rescuing defects in the primary cilium through regulation of primary cilia-related genes and thus enabling growth plate elongation. These data emphasize the relevance of (-)-epicatechin's action to control both proliferation and differentiation. We consider that modulation of the FGFR3-activated signaling pathways by (-)-epicatechin provides a rationale for developing treatments for ACH. Moreover, strict control of FGFR3 signaling by (-)-epicatechin provides a rationale for developing alternative therapeutic approaches with (-)-epicatechin alone or in combination with other relevant ACH treatments.

Ethics statement
All animal procedures and protocols were approved by the local animal care and use committee (approval number APAFIS#24826-2018080216094268 v5) and were carried out in compliance with EU directive 2010/63/EU for animals.

Transcriptomics
Details of the Fgfr3 Y367C/+ mouse models with a C57BL/6 background were previously described. 23 Three pairs of Fgfr3 Y367C/+ and Fgfr3 +/+ control littermates were produced at four time points (d7, d14, d21, and d28). Mouse genotypes were ascertained by PCR as described previously. 23 Mouse femoral heads were collected at slaughter and frozen in liquid nitrogen. The samples were crushed, and RNA was extracted using a Qiagen RNeasy Mini kit and Qiagen's animal tissue protocol. All RNA samples had an RNA integrity number greater than 8 and were analyzed on Affymetrix Mouse Genome 430 2.0 arrays by our service provider (PartnerChip, Evry, France). Processed and raw data were submitted to the Gene Expression Omnibus database (ID: GSE145821). Statistical analyses were performed using BioConductor 2.5, normalized using the gcrma package and were log2-transformed. Probesets differentially expressed by Fgfr3 Y367C/+ vs. Fgfr3 +/+ animals during their development were identified by computing contrasts with a genotype*timepoint design. To define the gene expression profiles, we also computed the interactions and linear, quadratic and cubic polynomial contrasts. P values were adjusted for multiple testing using the eBayes function, and the false discovery rate was computed using the decideTests function. Annotations were retrieved from the MGI database. Gene Ontology annotations were used to identify the genes involved in cilium organization and function. Genes with at least 3 types of proof were also retrieved from Cildb. 24 Heatmaps were generated online using Morpheus software.
Chromatography and ESI-TOF mass spectrometry detection The compounds in the Theobroma cacao extract fractions were separated as described previously, 18 with one modification: the HPLC system was coupled to a TOF mass spectrometer equipped with an ESI interface, operating in negative ion mode with a capillary voltage of +3.5 kV. The optimum values of the other parameters were as follows: drying gas temperature, 200°C; drying gas flow, 10 L·min −1 and nebulizing gas pressure, 2.3 bar. The mass range for detection was 50-1 200 m/z. To ensure repeatability, samples were injected in triplicate.

NMR analyses
Each sample was dissolved in DMSO-d6 and transferred into an oven-dried 5 mm NMR tube. The NMR spectra were recorded at 293 ± 0.1 K on a Bruker Avance III 600 spectrometer operating at a proton frequency of 600.13 MHz and fitted with a 5 mm QCI quadruple resonance pulse field gradient cryoprobe. The multiplicities observed were labeled a s = singlet, d = doublet, dd = doublet of doublets, t = triplet, m = multiplet, and bs = broad singlet. Each sample was measured with 8 dummy scans prior to 128 scans. The acquisition parameters were as follows: size of free induction decay (FID) = 64 K, spectral width = 20.5 ppm, acquisition time = 2.73 s, relaxation delay = 10 s, receiver gain = 20.2, and FID resolution = 0.25 Hz. A presaturation pulse sequence (Bruker 1D noesygppr1d) was used to suppress the residual H 2 O signal via irradiation at the H 2 O frequency during the recycling and mixing times. The resulting spectra were automatically phased, baseline-corrected, and calibrated against the trimethylsilyl-2,2,3,3-tetradeuteropropionate signal at 0.0 ppm. The t1 time was set to 4 µs, and the mixing time (d8) was set to 10 ms. The spectrometer transmitter was locked onto the DMSO-d6 frequency. Spectra were processed using TopSpin TM software (version 3.1, Bruker, Germany). 1H-1H total correlation spectroscopy (TOCSY) spectra, 1H-13C heteronuclear single quantum coherence (HSQC) spectra, and 1H-13C heteronuclear multiple bond coherence (HMBC) spectra were recorded using standard Bruker sequences. The TOCSY spectra were obtained by applying a relaxation delay of 2.0 s, a spectral width in both dimensions of 7 194.25 Hz, and a receiver gain of 64.0. These spectra were then processed using a sine-bell window function (shifted sine bell = 2.0). The HSQC spectra were acquired using a relaxation delay of 1.0 s and a spectral width of 7 211.54 Hz in F2 and 24 900.71 Hz in F1. A quadratic sine window function (shifted sine bell = 2.0) was used for HSQC spectra. The HMBC spectra were recorded with the same parameters as those for the HSQC spectra, except that the spectral width in F1 was 37 729.71 Hz. The coupling constants was set to 145 Hz for the HSQC experiments and 145 and 8 Hz (long range) for the HMBC experiments.
Primary and immortalized human chondrocyte and mouse chondrocyte cultures Primary and immortalized mutant cells were obtained and cultured as described previously. 9,22 Human and mouse cells were incubated with FGF2 (100 ng·mL −1 ) for 5 min and then with Theobroma cacao extract fraction 5 (100 µg·mL −1 ) or (-)-epicatechin (80 µg·mL −1 ) (Sigma-Aldrich, HW101708-1) for 30 min. Cultured chondrocytes were immunolabeled as previously described. 9 Total RNA extraction and RT-qPCR Reverse transcriptase quantitative polymerase chain reaction (RT-qPCR) was used to evaluate RNA expression after (-)-epicatechin treatment. RNA was extracted using an RNeasy Mini Kit (Qiagen) according to the manufacturer's instructions. The measured quantity of RNA (500 ng) was reverse-transcribed using Super-Script™ II Reverse Transcriptase (Invitrogen). RT-qPCR was performed with a ViiA-7 system (Applied Biosystems) using SYBR Green (Life Technologies) for fluorescence detection. qPCR optimization was performed to avoid off-target products and artifacts. This optimization was carried out with negative and positive control samples. We also examined the specificity and identity of the amplified product by melting curve and the amplicon size. Primers were designed using the Primer3Plus website. RT-qPCR data were analyzed using the 2 −ΔΔCT method, Epicatechin restores cilium defects and promotes bone growth L Martin et al.
and β-actin was used as the housekeeping control. The primer sequences used are listed in Table 2.

Computational details
The starting coordinates of ERK1, ERK2, and p38 were downloaded from the Protein Data Bank with the following respective accession codes: 4QTB, 5NGU, and 4L8M. The ligand (-)-epicatechin was built using Maestro software and optimized with Amber software. Molecular docking calculations were performed with AutoDock 4.2 software. We selected the docking pose that presented the lowest free energy of binding in the most populated cluster. The stability of the complexes formed by (-)-epicatechin and the kinases ERK1, ERK2, p38, and FGFR3 were further explored through molecular dynamics simulations using Amber software and the FF14SB/gaff2 force field. For each system, the simulation started with energy optimization of the water solvent followed by an overall minimization and a heating phase (from 0 to 300 K). Next, molecular dynamics was performed with the NTP ensemble, leading to a 50 ns trajectory. After the conventional 50 ns period of molecular dynamics, all systems underwent an additional 50 ns of Gaussian accelerated molecular dynamics.
Histological and immunohistochemical analyses of frozen sections Femur explants or femurs isolated from the mice at P16 were fixed in methanol chilled at −20°C for 5 h or 24 h, respectively. After incubation in 0.5 mol·L −1 EDTA, pH 8 for 72 h or 2 weeks, femurs were placed in 30% sucrose for 24 h, transferred to OCT compound at room temperature, and frozen in isopentane at −45°C. The 50 μm tissue sections were permeabilized with 0.3% Triton X-100 for 30 min and immunolabeled with rabbit IgG anti-Arl13b (Proteintech #17711-1-AP, IF 1:100) or mouse IgG 1 anti-γtubulin (Sigma-Aldrich #T6557, 1:100) primary antibodies. The primary antibodies were detected with goat anti-mouse IgG 1 coupled to Alexa Fluor 488 (Life Technologies, 1:400) and antirabbit IgG coupled to Alexa Fluor 647 (Life Technologies, 1:400). Tissue sections were mounted with DAPI-Fluoromount G ® (CliniSciences). Three-dimensional images of the growth plate were obtained using a spinning disc confocal microscope. Images were displayed using FIJI and the FigureJ plugin.

Statistical analysis
Differences between experimental groups were assessed by analysis of variance with Tukey's post hoc test or a Mann-Whitney U test. The threshold for statistical significance was set to P ≤ 0.05. Statistical analyses were performed using GraphPad Prism software. A paired Student's t test was used to compare two treatments in the same cell population. An unpaired Student's t test was used to compare groups of mice or different primary chondrocyte preparations.