Modulation of DNA methylation influences cartilage formation in murine chondrogenic models

The aim of this study was to investigate the role of DNA methylation in the regulation of in vitro and in vivo cartilage formation. Based on the data of an RNA chip-assay performed on chondrifying BMP2-overexpressing C3H10T1/2 cells, the relative expression of Tet1 (tet methylcytosine dioxygenase 1), Dnmt3a (DNA methyltransferase 3) and Ogt (O-linked N-acetylglucosamine transferase) genes was examined with RT-qPCR in mouse cell-line based and primary micromass cultures. RNA probes for in situ hybridization were used on frozen sections of 15-day-old mouse embryos. DNA methylation was inhibited with 5-azacytidine during culturing. We found very strong but gradually decreasing expression of Tet1 throughout the entire course of in vitro cartilage differentiation along with strong signals in the cartilaginous embryonic skeleton. Dnmt3a and Ogt expressions did not show significant changes with RT-qPCR and gave weak in situ hybridization signals. Inhibition of DNA methylation applied during early stages of differentiation reduced cartilage-specific gene expression and cartilage formation. In contrast, it had stimulatory effect when added to differentiated chondrocytes. Our results indicate that the DNA demethylation-inducing Tet1 is a significant epigenetic factor of chondrogenesis, and inhibition of DNA methylation exerts distinct effects in different phases of in vitro cartilage formation.


Introduction
Epigenetics refers to those reversible and heritable biological processes that regulate gene expression without the alteration of the primary DNA sequence [1]. Beside genetic factors, environmental determinants also influence the development of an organism and the differentiation of cells. Epigenetic regulation is responsible for cell-specific gene expression and for the inheritance of these unique expression patterns to daughter cells [2]. The primary target of epigenetic processes is the DNA-histone complex termed chromatin. The accessibility of a specific DNA segment for transcription factors is influenced by different epigenetic marks such as DNA methylation or histone modifications. miRNA interactions are recently discovered epigenetic factors that may alter the transcriptional regulation of mRNA [3]. DNA methylation, the most widely studied epigenetic modulation, causes gene repression or silencing by adding a methyl group from the ubiquitous methyl donor S-adenosyl methionine to the carbon 5 position of cytosine rings of the DNA. The newly formed unit is called 5-methylcytosine (5-mC). This process is catalysed by DNA methyltransferases (DNMTs) which can be classified into two groups according to their enzymatic activity. Dnmt3a and Dnmt3b are de novo methyltransferases that have a role in generating new methylation patterns during ontogenesis. Dnmt1, however, has Despite the accumulating wealth of data regarding the epigenetic regulation of gene activity in immature and mature cartilage, there are still many unanswered questions. The impact of epigenetic mechanisms on early stages of chondrogenesis and chondrocyte differentiation has not been described thoroughly, despite their high therapeutic relevance [32] [33] [34] [35]. In this study, we investigated the possible involvement of several epigenetic factors influencing DNA methylation in the regulation of chondrogenesis. We compared data obtained from chondrifying cultures of the murine embryonic mesenchymal cell line C3H10T1/2, murine primary chondrogenic cell cultures, and sections of developing whole mouse embryos. We performed a detailed expression analysis of Dnmt3a, Tet1 and Ogt genes and investigated the impact of the targeted inhibition of DNA methylation on chondrogenesis by using 5-azaC. Our results indicated Tet1 as a prominently expressed gene during both in vitro and in vivo chondrogenesis, and a developmental stage-dependent effect of 5-azaC, reflecting on the importance of DNA methylation and demethylation processes during cartilage formation.

EXPERIMENTAL MODELS 1. Primary chondrifying micromass cultures
High density cultures were established from mouse limb bud-derived mesenchymal cells following a protocol used on chicken micromass cultures with some modifications [36] [37]. First, NMRI laboratory mice were mated overnight. On the following day, successful mating was detected by confirming the presence of the vaginal plug -this day was considered as day 0 of gestation. Embryos on gestational day 11.5 (E11.5) were retrieved from the uterus. NMRI mice were sacrificed according to the ethical standards defined by the University of Debrecen Committee of Animal Research (Permission No. 2/2018/DE MÁB). After some brief washes with sterile calcium and magnesium-free phosphate buffered saline (CMF-PBS), distal parts of fore and hind limb buds were removed and pooled in sterile CMF-PBS. Limb buds were then dissociated in 0.25% trypsin-EDTA (Merck) incubated at 37 °C in a CO2 incubator (5% CO2, 80% humidity) for 20-30 min. After the addition of equal volume of foetal bovine serum (FBS; Gibco, Gaithersburg, MD, USA), cells were centrifuged for 10 min at 800 × g. The digested cells were filtered through a 40-μm pore size plastic filter unit (Corning, Tewksbury, MA, USA) in order to gain a single cell suspension of mesenchymal cells. Cells were centrifuged again for 10 min at 800 × g. The cell pellet was resuspended in high glucose (4.5 g/L) Dulbecco's modified Eagle's medium (DMEM; Sigma-Aldrich, St. Louis, MO, USA) supplemented with 10% (v/v) FBS, 0.5 mM stabile L-glutamine (Sigma-Aldrich), and antibiotics/antimicotics (penicillin, 50 U/mL; streptomycin, 50 μg/mL; fungizone, 1.25 μg/mL; TEVA, Debrecen, Hungary) at a concentration of 1.0 × 10 7 cells/mL and 100-μL droplets were inoculated into 35 mm plastic tissue culture dishes (Eppendorf, Hamburg, Germany). Cells were allowed to attach to the surface for 2 h at 37 °C in a CO2 incubator. Finally, 2 mL of culture medium was added into the culture dishes. Cultures were incubated in a conventional cell culture incubator. Day of inoculation was considered as day 0 of culturing. Cell cultures were maintained for 0, 1, 3, 4, 6, 10 or 15 days. The medium was changed on every second day.

Micromass cultures established from C3H10T1/2 BMP-2 cells
The BMP-2 overexpressing C3H10T1/2 cells were maintained as monolayer cultures in high glucose (4.5 g/L) Dulbecco's modified Eagle's medium (DMEM; Sigma-Aldrich) supplemented with 10% (v/v) FBS, 0.5 mM stabile L-glutamine, 6.6 μg/mL ampicillin, 100 μg/mL streptomycin (TEVA), 5 μg/mL puromycin (Sigma-Aldrich), and kept in a humidified CO2 incubator at 37 °C. Once monolayer cultures reached an approx. 80% confluence, cells were harvested in order to establish high density cultures, as described previously [38]. After trypsinisation, cells were centrifuged at 800 × g for 10 min. The number of cells was determined with the help of an automated cell counter (LUNA TM , Logos Biosystems, Annandale, VA, USA). For micromass cultures, cell density was 1.0 × 10 7 cells/ml. 100-μL droplets of the cell suspension were inoculated into 35 mm plastic Petri dishes (Eppendorf). This day was considered as day 0. After leaving the cells to attach to the surface (2 h, 37 °C), high glucose DMEM was added to the dishes, which was changed on every second day. The micromass cultures were maintained for up to 0, 5, 10, 15 days.

PCR ASSAY
Quantitative PCR arrays were used to monitor differential expression of 80 genes encoding epigenetic factors and chondrogenesis markers in C3H10T1/2 BMP-2 high density cell cultures. Total RNA was isolated using Direct-Zol® RNA Miniprep kit (Zymo Research, Irvine, CA, USA) from cultures harvested on designated days of culturing, and cDNA synthesis was conducted by iScript RT Supermix kit (Bio-Rad, Hercules, CA, USA) on 1 µg total RNA. PrimerQuest software was used to design primers and were manufactured by Integrated DNA Technologies (Coralville, IA, USA). RT-qPCR was performed using a CFX RT-PCR machine (Bio-Rad) and the SsoFastEvaGreen™ Supermix (Bio-Rad). Parameters of PCR were set according to the manufacturer's protocol. The RT-qPCR primer sequences are in Table S1 in the Supporting information. The specificity of the PCR primers was monitored by a post-PCR melting curve analysis, and heatmap analysis was conducted by CFX Maestro software (Bio-Rad).

RNA ISOLATION AND REVERSE TRANSCRIPTION
On the designated days of culturing, micromass cultures were washed with physiological NaCl two times and stored at -80 °C. Total RNA from C3H10T1/2 BMP-2 high density cell cultures and primary micromass cultures were isolated as previously described [39]. Briefly, cultures were mixed with TRI Reagent (Applied Biosystems, Foster City, CA, USA). 20% RNAse-free chloroform was added and samples were centrifuged at 10,000 × g at 4 °C for 15 min. After the incubation in 500 µL RNAse-free isopropanol for 1 h at -20 °C, total RNA was dissolved in RNAse-free water and stored at -80 °C. RNA concentration and purity were determined using a NanoDrop 1000 spectrophotometer (Thermo Fisher Scientific). Reverse transcription reactions were performed on 1000 ng of total RNA using the High Capacity cDNA Reverse Transcription Kit (Thermo Fisher Scientific) according to the manufacturer's protocol. cDNA was stored at -20 °C.

QUANTITATIVE REAL-TIME PCR ANALYSES
RT-qPCR reactions were performed as described previously [40]. The SYBR Greenbased system (Promega, Madison, WI, USA) was used by absolute quantification using the standard curve method. Specific primer pairs for the chondrogenic and DNA methylation-specific genes were designed by the Primer-BLAST service of NCBI and ordered from Integrated DNA Technologies (IDT, Coralville, IA, USA). Nucleotide sequences of the primer pairs are shown in Table S2 in the Supporting information. Conventional PCR was used to create standard curves for the absolute quantification. The Promega GoTaq Flexi DNA Polymerase kit (Promega, Madison, WI, USA) was applied to set up the following mixture (50 μL per each reaction): 1.25 U GoTaq polymerase; 3 mM MgCl2; 0.2 mM dNTP; 200 nM primers; and 50 ng cDNA (pooled from 6-day-old untreated samples). Amplification was performed in a programmable thermal cycler (MultiGene 96-well Gradient Thermal Cycler; Labnet International, Edison, NJ, USA) using the following thermal profile: initial denaturation (95 °C, 5 min), followed by 40 cycles of denaturation (95 °C, 15 s), annealing (58 °C, 20 s), extension (74 °C, 20 s); and final extension (74 °C, 5 min). Roche High Pure PCR Product Purification Kit (Roche, Basel, Switzerland) was used to purify the PCR products according to the instructions of the manufacturer. DNA concentration of the isolated PCR products was detected using a Nanodrop 1000 UV-Vis spectrophotometer (Thermo Fisher Scientific). Purified samples were diluted in a serial manner (10fold, starting with 1 ng/μL) to establish the standard curves. The QuantStudio 3 Real-Time PCR System (Thermo Fisher Scientific) was applied for the reactions using the GoTaq qPCR Master Mix (Promega) and 10 ng cDNA per each 10-μL reaction. The settings for the process were the following: initial denaturation (95 °C, 2 min), followed by 40 cycles of denaturation (95 °C, 5 s), annealing and extension (60 °C, 30 s), and final extension (72 °C, 20 s). Amplification was followed by a melt curve stage consisting of 3 steps: denaturation at 95 °C for 15 sec, annealing at 55 °C for 15 sec, and a dissociation step at 0.15 °C/sec increments between 55 °C and 95 °C. Amplification data were analysed using the QuantStudio Design and Analysis Software (version 1.5.1) and exported data were processed using Microsoft Excel.

DIGOXIGENIN-LABELLED RNA PROBE PREPARATION
PCR primers were designed to amplify a ~1000 bps long region from the 3'UTR of Dnmt3a, Ogt and Tet1 genes. PCR-amplified 3'UTR regions were cloned into pDrive vector (Qiagen, Germantown, MD, USA) and sequenced. Insert flanking T7 promoters were used for generating antisense probes. Sequence data of the cloned regions are in Table S3 in the Supporting information. The specific gene products of Dnmt3a, Ogt and Tet1 probes were amplified with the help of PCR from the plasmids. Amplifications were performed in a thermal cycler (Labnet MultiGene™ 96-well Gradient Thermal Cycler; Labnet International, Edison, NJ, USA) using the following settings: 95 °C, 2 min, followed by 33 cycles (denaturation, 95 °C, 15 s; annealing for 20 s at 57 °C; extension, 72 °C, 75 s) and then 72 °C, 2 min. Digoxigenin-labelled RNA probe preparation was performed as recommended by Roche, with some modifications. The amplified PCR products were isolated using a Roche High Pure PCR Product Purification Kit (Roche, Basel, Switzerland) according to the instructions of the manufacturer. DNA concentration of purified PCR products were detected with the help of a Nanodrop 1000 UV-Vis spectrophotometer (Thermo Fisher Scientific). The specific RNA labelling was developed with a DIG RNA labelling mix by in vitro transcription of DNA. First, the following components were mixed together to create the DIG RNA labelling mix: 1 µL of purified PCR product (concentration between 100-200 ng/µL); 2 µL of 10× concentrated DIG RNA Labelling Mix (Promega); 4 µL 5× Transcription Buffer (Promega); 2 µL 100 mM Dithiothreitol (DTT) (Promega); 2 µL T7 RNA Polymerase (Promega) and 9 µL nuclease-free water (NFW) (Promega) to make a total reaction volume of 20 µL. After the components were mixed together, the mixture was incubated for 2 h at 37 °C. Polymerase reaction was terminated by 2 µL 0.2 M EDTA (pH 8.0). The labelled RNA was precipitated after the addition of 2.5 µL 4 M LiCl and 75 µL pre-chilled 100% ethanol. After a brief mix with vortex, the precipitate was incubated at -80 °C overnight. On the next day, the sample was centrifuged at 13,000 × g for 15 min at 4 °C. The supernatant was discarded, and the pellet was washed with 100 µL ice-cold 70% (v/v) ethanol. The precipitate was centrifuged again at 13,000 × g for 15 min at 4 °C, and after discarding the supernatant, the sample was left to dry on room temperature for some minutes. Finally, the RNA pellet was dissolved in 75 µl hybridization buffer (containing 20× saline sodium citrate (SSC), dextran sulfate, 50× Denhardt's solution, sodium dodecyl sulfate (SDS), tRNA, and 50% (v/v) formamide; Sigma-Aldrich) and stored at -20 °C.

IN SITU HYBRIDIZATION
Whole murine embryos were collected as previously described. Briefly, NMRI mice were mated overnight, and detectable vaginal plug confirmed on the following morning, which was regarded as day 0. On gestational day 15, whole mouse embryos were retrieved from the uterus, washed in DEPC-PBS (PBS with 0.1% dietyhl-pyrocarbonate) and were fixed in 4% paraformaldehyde (PFA, dissolved in DEPC-PBS) overnight. On the following day, embryos were washed in DEPC-PBS two times for 10 min each, then immersed into 15% and 30% RNAse-free sucrose solution until they sank. After embedding the embryos into Cryomount medium (Bio-Optica, Milan, Italy), 20 µm thick frozen sections were cut in a sagittal plane using a cryostat (CM3050 S, Leica Biosystems, Buffalo Grove, IL, USA) and mounted onto Superfrost glass slides (Thermo Fisher Scientific). Sections were stored at -20 °C. We applied a nonradioactive in situ hybridization protocol described earlier, with some modifications [41]. Briefly, sections were removed from -20 °C and left at room temperature for 20 min. The glass slides were placed into a 58 °C incubator overnight for drying. On the following day, slides were removed from the incubator and left at room temperature for 20 min. Samples were fixed in 4% PFA (dissolved in DEPC-PBS) for 20 min. After washing with DEPC-PBS for 2×10 min, the remaining liquid was blotted and samples were treated with 100 µL Proteinase K solution (20 µg/mL; Promega) at 37 °C for 20 min. The slides were washed with DEPC-PBS for 2×5 min. Samples were prehybridised for 4 h at 58 °C, then the solution was changed to the hybridisation solution that contained the RNA probe (1-2 µg/mL) and the slides were incubated at 58 °C for 16 h. All components were RNAse free until this step. On the third day, slides were washed in 1× SSC at 58°C for 15 min, then in 1.5× SSC for another 15 min at 58 °C, and finally twice in 2× SSC for 2×20 min at 37 °C. Samples were treated with 0.5 µg/mL RNAse A dissolved in 2× SSC at 37 °C for 20 min. After washing in 2×SSC at room temperature for 10 min, slides were washed twice in 0.2× SSC at 58 °C for 2×30 min. Then, sections were washed twice at 58 °C for 2×15 min, then at room temperature for 10 min with PBST. Finally, samples were incubated in 10% Blocking buffer solution (Blocking buffer powder dissolved in maleic acid buffer with Tween (MABT); Roche) with α-DIG antibody (anti-digoxigenin, 1:1000; Abcam, Cambridge, UK; Cat. No.: ab420) at 4 °C overnight. Sections were then washed three times in PBT (PBS with 0.1% Triton X-100 and 2 mg/mL BSA) for 3×20 min, then twice in 1 M TRIS solution (pH 9.0) for 2×5 min. Digoxigenin antibody was visualised by incubation with TRIS-NBT/BCIP solution (20 mg/mL stock solution of nitro blue tetrazolium and 5-bromo-4-chloro-3-indoyl phosphate, dissolved in 1 M TRIS; Sigma-Aldrich) at room temperature in the dark for 2~20 hours (depending on the amount of the RNA). After the incubation time, samples were washed in PBST for 2×10 min. Finally, slides were mounted with DPX medium (Sigma-Aldrich). Photomicrographs of the sections were taken using an Olympus BX53 camera on a Nikon Eclipse E800 microscope (Nikon Corporation, Tokyo, Japan).

DIMETHYL-METHYLENE BLUE STAINING METHOD
The dimethyl-methylene blue (DMMB) staining method was used to demonstrate the amount of metachromatic cartilage ECM in mouse whole embryos and also in primary murine high density cultures. Sections of whole embryos stained with DMMB served as a control for in situ hybridization. Frozen sections were prepared as described above. After the glass slides were removed from -20 °C they were dried at room temperature for 10 min, then at 58 °C for 1 h. After washing in distilled water for 2×10 min, samples were stained with 0.1% (w/v) DMMB (Sigma-Aldrich) dissolved in distilled water for 5 min. Surplus dye was removed by washing the sections with distilled water for 3×10 min. Slides were mounted with DPX. Photomicrographs of the stained samples were taken as described above. As for HD cultures, 30-μL droplets of the cell suspensions were inoculated on the surface of 10-mm round coverglasses (Menzel-Gläser, Menzel GmbH, Braunschweig, Germany) into 24-well culture plates. On day 4 or 6 of culturing, colonies were rinsed with PBS and fixed in a 4:1 mixture of absolute ethanol and 40% formaldehyde. After rehydration in a descending series of ethanol, cultures were stained with 0.1% (w/v) DMMB dissolved in 3% (v/v) acetic acid (pH 1.8). Surplus dye was washed in acetic acid, then with distilled water. Finally, cultures were mounted with Aquatex (Sigma-Aldrich). Photomicrographs of the stained cultures were taken as described above.

MANIPULATION OF EPIGENETIC REGULATION WITH 5-AZACYTIDINE
The epigenetic modifier 5-azacytidine (5-azaC; Cat. No.: A2385; Sigma-Aldrich) was used to inhibit DNA methyltransferases and to consequently activate specific gene regions by triggering DNA demethylation [30] [42]. 5-azaC was dissolved in dimethyl sulfoxide (DMSO) at 10 mM and then applied at a final concentration of 10 µM for 72 h on culturing day 1 or 3. Primary micromass cultures were harvested on the 4 th or 6 th day of culturing, according to the treatment protocol. Control colonies were treated with equal amount of the vehicle (DMSO). Cell viability was monitored as previously described [36]. 24-well plates were used for cell culturing. 25 µL of MTT reagent (3-[4,5-dimethylthiazolyl-2]-2,5diphenyltetrazolium bromide; 5 mg/mL in PBS) were pipetted into each well on culturing day 4 or 6. Cells were incubated for 2 hours at 37 °C. Following the addition of 500 μL MTT solubilizing solution (10% Triton X-100 in 2-propanol), optical density was measured at 570 nm (Chameleon, Hidex Ltd., Turku, Finland). Measurements were carried out in 3 samples of each experimental group in 3 independent experiments. Optical density readings of the experimental groups have been normalized to those of the vehicle controls and shown as percentage changes.

CELL PROLIFERATION ASSAY WITH 3 H-THYMIDINE LABELLING
Rate of cell proliferation was examined as previously described [36]. Briefly, 1 μCi/mL 3 H-thymidine (diluted from methyl-3 H-thymidine; 185 GBq/mM, Amersham Biosciences, Budapest, Hungary) was added to cell culture medium on day 3 or 5, 16 h before the end of treatments. After washing with PBS, proteins were precipitated with icecold 5% trichloroacetic acid, and washed with PBS again. Colonies were then air-dried for one week and radioactivity was counted by a liquid scintillation counter (Chameleon, Hidex). Measurements were carried out in 9 samples of each experimental group in 3 independent experiments. Scintillation counting data of the experimental groups were normalised to those of the respective controls and presented as percentage changes.

STATISTICAL ANALYSIS
All data are representative of at least three independent experiments. Data in figures is representative of the mean ± SEM (standard error of the mean) of a single experiment. With regard to RT-qPCR reactions, one representative data set is shown out of 3 parallel experiments showing similar trends, and data were normalised to beta actin (Actb) or Succinate Dehydrogenase Complex Flavoprotein Subunit A (Sdha), as calculated by NormFinder. Statistical differences were determined using paired Student's t-test or One-Way ANOVA with Tukey HSD. The specific differences were considered statistically significant if P < 0.05. Statistical significance is indicated by asterisks as follows: P < 0.05 = *; P < 0.01 = **; P < 0.001 = ***.

Dnmt3a, Tet1 and Ogt epigenetic marker genes display a specific expression pattern in murine chondrogenic models
We first studied the expression pattern of epigenetic marker genes in different in vitro murine chondrogenic model systems. Samples for RNA assays were obtained from high density cultures established from C3H10T1/2 BMP-2 cells collected on culturing days 0, 5, 10 and 15 (corresponding to the main stages of chondrogenesis in vitro), in order to examine the expressional peaks of epigenetic markers at the mRNA level. The results of the RNA assay clearly showed the expression of every gene of interest (Figure 1). Interestingly, many of the epigenetic marker genes in connection with DNA methylation were upregulated at later stages of chondrogenic differentiation (culturing days 10 and 15). Three epigenetic modifiers were chosen for subsequent analysis: DNA methyltransferase 3 alpha (Dnmt3a), Tet methylcytosine dioxygenase 1 (Tet1), and O-Linked N-Acetylglucosamine (GlcNAc) Transferase (Ogt). Dnmt3a was strongly expressed on culturing day 15, while Tet1 expression peaked around day 10. It is worth noting that Ogt, which interacts with Tet1, peaked on culturing days 10 and 15. On the other hand, the expression profile of the chondrogenic markers collagen type II alpha 1 chain (Col2a1) and aggrecan (Acan) showed an earlier activation and increase in transcript levels between days 5 and 10 of culturing. Col10a1, a marker for matrix mineralization and chondrocyte hypertrophy, showed an expressional peak in the samples on days 10 and 15. The expression pattern of Col1a1 gene followed that of Col10a1, reflecting on the initiation of osteogenic differentiation in the presence of hypertrophic chondrocytes. After RNA isolation from the cell-line based micromass cultures, quantitative realtime PCR analysis was carried out to study the relative expression of the three DNA methylation-specific marker genes during chondrogenesis. The mean quantity values for Dnmt3a, Tet1 and Ogt markers were normalized to Actb. The largest increase of gene expression was on culturing days 10 and 15 for all three genes (Figure 2). The relative gene expression of Tet1 displayed the most prominent changes, with the greatest degree of upregulation on day 10 of culturing. The expression profiles showed high similarity to those detected with the RNA assay. Primary high density cultures were established from mouse embryonic limb buds and collected on designated culture days. Transcripts for the DNA methylation genes were also identified in this type of in vitro model by RT-qPCR; however, their expression profile was more varied compared to the cell line-based model. After choosing the most stably expressed normalizing gene, the mean quantity values for the three examined DNA methylation markers were normalized to the reference gene Sdha, and the normalized mean quantity was set to 1.00 on culturing day 0 for each of the genes. Tet1 showed the highest expressional fold change among the three examined genes, with peaks on day 1 and 3 of culturing. Dnmt3a transcript level was the highest on day 3 and displayed a significant downregulation by day 15. Ogt, on the contrary, was constantly expressed by the differentiating chondrocytes, except on day 15, when it was significantly downregulated (Figure 3). Data are expressed as the mean ± SD relative to culturing day 0 and normalized against the reference gene Sdha. Statistically significant differences of gene expression levels are indicated by asterisks as follows: *P < 0.05; **P < 0.01; ***P < 0.001. Representative results out of 3 independent experiments (biological replicates) showing similar trends of changes.
Next, we aimed to prove the in vivo relevance of the three examined genes in chondrogenesis in vivo. Therefore, we performed in situ hybridization analyses of Dnmt3a, Tet1 and Ogt in 15-day-old whole mouse embryo samples. This experiment revealed that all three genes were expressed in the territories of hyaline cartilage islands of the developing forelimb, vertebrae and sternum of the E15 mouse embryos. Among the three examined genes, Tet1 signals were the strongest (Figure 4. a, b, c), Dnmt3a showed a moderate expression pattern (Figure 4. d, e, f), while Ogt appeared to have a weak expression in the primitive limb and vertebrae compared to the other two epigenetic markers (Figure 4. g,  h, i). At this developmental stage, cartilage had not started to differentiate into bone, the hypertrophic cartilaginous area was minimal, and ossification has not commenced yet but the cartilaginous primordia of the future bones were well visible. The purple metachromatic colour was obtained with the application of DMMB, in order to demonstrate the cartilage elements of the developing limbs, vertebrae, skull and ribs (Figure 4. j, k, l).  (a, b, c), Tet1 (d, e, f) and Ogt (g, h, i). Sections were also stained with DMMB for cartilage-specific proteoglycans (j, k, l). Metachromatic (purple) areas in photomicrographs show polyanionic glycosaminoglycan-rich cartilage ECM. Photomicrographs of whole embryos were taken with a 4× objective (a, d, g, j). Inserts were taken with a 10× objective, which correspond to areas indicated with boxes (b, c, e, f, h, i, k, j). Note the strong expression of Dnmt3a and Tet1 in maturing chondrocytes of the developing vertebrae, limb buds and skull in the mouse embryo. Scale bar for a-d: 1 mm, for the rest: 200 µm.

ECM morphology, cell proliferation and cell viability of early and late chondrogenic stages are different after 5-azaC treatment
The DNA methylation inhibitor 5-azaC was applied on primary murine micromass cultures at 10 μM. Some high density cultures were treated during the beginning of chondrogenesis (i.e. from day 1 for 72 hours), while other cultures were treated from day 3 for 72 hours in order to investigate its effects on later stages of chondrogenesis. To visualise cartilage-specific ECM accumulation in the primary murine micromass cultures, the qualitative DMMB staining method was used on culturing days 4 and 6 at the end of the treatment protocols. The DNA methylation inhibitor attenuated the amount of metachromatic ECM produced if applied at the early stage of chondrogenesis. Interestingly, when 5-azaC was administered from culturing day 3 for 72 hours, the morphology of metachromatic cartilage nodules was similar to that of the untreated HD cultures. It is of note that the characteristic metachromatic (purple) colour was weaker in the treated colonies at day 6, indicating that the chondrocytes of these cultures probably produced less ECM compared to the controls (Figure 5. a).
We hypothesized that one of the reasons behind the attenuated ECM production could be the altered proliferative and mitochondrial activity of the chondroprogenitor cells and chondrocytes. Thus, we examined the effects of 5-azaC on cell viability and cell proliferation during chondrogenic differentiation. The assays were carried out on culturing days 4 or 6, depending on the starting day of treatment. Both treatment methods inhibited the proliferation of chondrifying cells, especially during the early stages of chondrogenesis, when this parameter was lowered by 55% (±5%), as opposed to later stages, when the rate of cell division was reduced by 37% (±7%) (Figure 5. b). We also studied the potential cytotoxic effect of 5-azaC during in vitro cartilage formation. The percentage of viable cells in the 4-day-old colonies after treatment was 90% (±2%), compared to the control group, and this was a significant decrease. In contrast, cells in 6-day-old micromass cultures showed a massive reduction in their mitochondrial activity (24±3%) (Figure 5. c). Figure 5. Effect of the DNA methylation inhibitor 5-azaC on cartilage ECM production, cell proliferation and cell viability. (a) Metachromatic staining of 4-and 6-day-old primary murine high density cultures. 5-azaC was applied from the first or the third day of culturing, respectively, for 72 hours at a final concentration of 10 µM. Metachromatic ECM accumulation was visualized by dimethyl methylene blue (DMMB) qualitative staining assay. Original magnification was 4×. Scale bar: 1000 or 500 μm. Effects of 5-azaC on cell proliferation (b) and cell viability (mitochondrial activity) (c) in primary murine high density cultures. Cell viability was determined by using MTT assay and cell proliferation was examined by 3 H-thymidine incorporation assay on day 4 or day 6, following treatment with 5-azaC. Statistically significant differences between proliferation rate and mitochondrial activity of cells in cultures that received the inhibitor versus vehicle control cultures are marked by asterisks (*P < 0.05, **P < 0.01, ***P < 0.001). Representative data out of 3 independent experiments.

Inhibition of DNA methylation influences chondrogenic marker gene expression depending on the developmental stage of chondrogenesis
In order to detect the effects of 5-azaC treatment on gene expression profiles in primary micromass cultures, RT-qPCR reactions were performed. We collected samples for total RNA isolation on culturing days 4 or 6. 5-azaC was applied for 72 hours prior to the sample collection. First, we wanted to check whether the expression of the investigated genes mediating DNA methylation was altered after the application of the inhibitor. To this end, we assessed the quantitative expression profile of Dnmt3a, Tet1 and Ogt. Our results confirmed that 5-azaC treatment significantly downregulated the expression of Dnmt3a and Ogt, while Tet1 expression was not influenced. This pattern was similar in the two different experimental groups and reflected on a transcriptional influence of 5-azaC on Dnmt3a and Ogt genes (Figure 6. a).
Next, we studied the mRNA levels of key chondrogenic marker genes with RT-qPCR. The key chondrogenic transcription factor Sox9, as well as the two major cartilage matrixspecific genes (Col2a1 and Acan) were selected. We found that the expression profiles of these genes were significantly altered after the inhibition of DNA methylation at both the early and the late stages of chondrogenesis (Figure 6. b). During the early stage of in vitro cartilage formation, all three marker genes were significantly downregulated. The largest decrease was detected for Col2a1 and Acan. On the contrary, during the later stage of chondrogenesis, Sox9 and Acan were upregulated, while Col2a1 expression remained unchanged. Figure 6. DNA methylation-specific (a) and cartilage-specific (b) marker gene expression in 4and 6-day-old primary murine chondrifying cell cultures after 5-azaC treatment. The DNA methylation inhibitor was added to the culture medium from the first or the third day of culturing, respectively, for 72 hours at a final concentration of 10 µM. Data are expressed as the mean ± SD relative to the vehicle control and normalized against the reference gene Sdha. Statistically significant differences of gene expression levels are indicated by asterisks as follows: *P < 0.05; **P < 0.01; ***P < 0.001. Representative data out of 3 independent experiments.

Discussion
DNA methylation plays an important role in joint diseases [43]. Recent studies indicate that DNA methylation may serve as a promising therapeutic target for several human joint disorders, including osteoarthritis [44]. Since the stem cell therapy-based approach represents a very attractive component in the toolkit of regenerative medicine, a better understanding of DNA methylation during early chondrogenesis is essential. To this end, we investigated the role of DNA methylation during in vitro chondrogenesis in different experimental models in the present study. We focused on the temporal expression pattern of specific regulators of DNA methylation at the mRNA level in different murine chondrogenic models, and we have also investigated the effects of the DNA methylation inhibitor 5-azaC on chondrocyte differentiation. First, we studied the osteo-chondrogenic differentiation in high density cultures of the mesenchymal cell line C3H10T1/2 in which cells constitutively express BMP-2 [45]. The cell line-based high density cultures were collected for RNA isolation on designated days of culturing, based on the specific differentiation stage of chondrocytes. In terms of chondrogenic marker expression patterns, the results of the RNA assay are in a good correlation with our earlier study which analysed the transcript levels of the same markers by conventional RT-PCR [38]. The proteins coded by the Col2a1 and Acan genes are characteristic components of the cartilage-specific ECM [46]. According to the RNA assay, these genes were upregulated around the 5 th day of culturing, corroborating our earlier results suggesting that metachromatic areas corresponding to glycosaminoglycan-rich cartilage ECM started to appear from the 3 rd day of culturing [38].
After having verified the expression of chondrogenic marker genes by the RNA assay in murine cell-line-based micromass cultures, we undertook to study the gene expression profiles of epigenetic markers. The importance of Dnmt3b enzyme in normal limb development and hypertrophic chondrocyte maturation has already been proven [14]. This was well visible with the RNA assay, where the expression of Dnmt3a and Dnmt3b genes showed strong elevation as chondrogenesis proceeded into later stages. TET enzymes contributing to the reversible nature of DNA methylation were also investigated, as recent studies pointed out that Tet1 might be a key epigenetic regulator of chondrogenesis. Although lineage-specific knockdown of Tet1 caused only minor skeletal abnormalities in the transgenic animals, significant downregulation of the cartilage matrix-specific gene expression was observed by in vitro experiments [19] [27] [47]. In line with these observations, Tet1, 2, and 3 showed intense expression in the heatmap of the RNA assay during the second half of in vitro chondrogenesis.
In addition to the cell-based model, we also employed a primary high density cell culture system established from murine limb bud-derived chondroprogenitor mesenchymal cells to study the formation of hyaline cartilage in vitro [48]. In primary micromass cultures, a moderately high Dnmt3a expression was detected at the time of the commitment of chondrogenic cells (i.e. day 3 of culturing) and a gradual decrease along with the progress of chondrogenesis was seen when RT-qPCR results were analysed. The expression of Tet1 showed significantly elevated levels compared to the other two epigenetic genes of interest. Ogt, encoding a molecular partner of TET enzymes, showed low and constant level of expression as revealed with RT-qPCR. Ogt clearly has a specific association with the TET enzymes, as it has the same expression pattern as the three TET genes according to the RNA assay. The reason behind the different quantitative gene expression profiles between the cell line-based and primary micromass cultures could be the differentiational rate and state of the cells in the high density cultures. The C3H10T1/2 BMP2 cell-based micromass cultures demonstrated a distinct macroscopic morphology compared to the primary high density cultures on culturing day 6 according to our earlier results [38], which means that the chondrogenic differentation of cell-line based cultures might be slightly slower. Additionally, primary micromass cultures represent a more heterogeneous cell population, in which low number of fibroblasts can also be present.
Next, we aimed to detect the expression of the three examined epigenetic markers at the mRNA level in cryosections of 15-day-old whole mouse embryos by in situ hybridization. We have chosen the age of the embryo by considering the stages of cartilage development as described by Rafipay et al [49]. They established that cartilage and bone formation in the developing limbs of a mouse embryo took place between E12.5 and E16.5. The first cartilaginous areas were identified in the forelimb at E12.5, while the mineralization-specific Alizarin red staining was recognizable from E15.5. Thus, we decided that E15 was the most suitable stage to examine the DNA methylation-related gene expression. At that stage, chondrocytes in mature hyaline cartilage could be easily identified, and the developmental stage of chondrocytes in E15 mouse embryos is almost the same as in the 6-day-old primary murine micromass cultures. The in situ hybridization results were in a good agreement with the findings of the quantitative real-time PCR studies. The most important similarity between the results of in situ hybridization and RT-qPCR was that in both cases, Tet1 had the strongest expression pattern among the three examined epigenetic regulators. The significant role of this epigenetic factor in osteochondrogenic differentiation was proved by the experiments in which the specific knockout of Tet1 impaired the skeletal development of the mutant mice: the examined animals showed growth irregularities or even embryonic lethality [50]. Specific knockdown of Tet1 in the ATDC5 chondroprogenitor cell line also altered the chondrogenic differentiation of the cells [13]. Based on our results and the experimental data mentioned earlier, it is reasonable to suggest that Tet1 has a major role in cartilage formation.
We have also investigated the effects of the DNA methylation inhibitor 5-azaC on chondrocyte differentiation. Based on data found in earlier studies [33] [51] [52], we applied the compound at a final concentration of 10 µM. This concentration was five-fold higher than that applied to mature chondrocytes by Duan et al [53], but it was well tolerable by the chondrogenic cells as shown by the viability assay. The cell proliferation assay indicated that 5-azaC affected two important phases of in vitro chondrogenesis: the proliferating phase between days 0 and 3 (with mostly chondroprogenitor cells and early chondroblasts present in the micromass culture), and also the differentiating phase between days 3 and 6 (with chondroblasts and mature chondrocytes that produce high amount of cartilage-specific ECM). We found that the early phase of chondrogenesis was more strongly affected by the treatment, causing extreme loss in the proliferative ability of the cells. The RT-qPCR results confirmed that 5-azaC was indeed modulating DNA methylation during in vitro chondrogenesis: the expression of Dnmt3a gene was significantly inhibited, while the expression of Tet1 gene, which has a role in DNA demethylation, was not altered. 5-azaC treatment has been documented to enhance the chondrogenic differentiation of human MSCs in vitro [54]. We proved that 5-azaC had a differentiation state-dependent inhibitory mechanism during in vitro hyaline cartilage formation. The mRNA expression levels of Sox9, Col2a1 and Acan significantly decreased when 5-azaC was applied during the early stages of chondrogenesis. However, when DNA methylation was inhibited during later stages of chondrogenesis, Sox9 and Acan became significantly upregulated. These findings are in good correlation with the results obtained with DMMB staining, where the primary murine high density cultures treated during the early phase of chondrogenesis displayed a weaker metachromasia than those which received 5-azaC during the second phase of the culturing.

Conclusions
This is the first study to report the differentiation stage-dependent transcript expression levels of key enzymes mediating DNA methylation and demethylation during in vitro chondrogenesis of primary chondrifying micromass cultures. Both enzyme systems showed increased expression during the early stages of chondrogenic differentiation, which was followed by a gradual decrease at the later phase. The two epigenetic mechanisms (i.e., DNA methylation and demethylation) are mutually exclusive; however, considering different regulatory regions (i.e., promoters and enhancers), the two processes can also take place simultaneously. The differentially methylated regions must be identified in order to provide a better insight into chondrogenesis and discover genes that may serve as drug targets. The differentiational stage-dependent effects of 5-azaC on chondrogenic cells suggest the need of careful design for therapeutic application of this compound during the intended treatment of joint diseases. Moreover, 5-azaC can also inhibit RNA methylation, which might provide another regulatory layer for chondrogenic differentiation [55]. Therefore, it is reasonable to consider that option when the effect of 5-azaC treatment is evaluated.

Supplementary Materials:
The following are available online at www.mdpi.com/xxx/s1, Table S1: Sequences of primer pairs used for the PCR assay analyses, Table S2: Sequences of primer pairs used for the RT-qPCR reactions, Table S3: Sequence data of the 3'UTR regions of Dnmt3a, Ogt and Tet1 genes with insert flanking T7 promoters for antisense probe preparation.

Data Availability Statement:
The data presented in this study are available on request from the corresponding author.