Gene targeting of the transcription factor Mohawk in rats causes heterotopic ossification of Achilles tendon via failed tenogenesis

Significance

Molecular mechanisms of tendon development and homeostasis are not well understood. Generation and analysis of Mkx^−/− rats revealed new functions of Mohawk (Mkx) in mediating cellular responses to mechanical stress. An Mkx-ChIP assay in rat tendon-derived cells with Mkx expression suggested that this factor may associate with both tendon- and cartilage-related genes to orchestrate tendon cell differentiation and maintenance. These findings advance our understanding of tendon physiology and pathology.

Abstract

Cell-based or pharmacological approaches for promoting tendon repair are currently not available because the molecular mechanisms of tendon development and healing are not well understood. Although analysis of knockout mice provides many critical insights, small animals such as mice have some limitations. In particular, precise physiological examination for mechanical load and the ability to obtain a sufficient number of primary tendon cells for molecular biology studies are challenging using mice. Here, we generated Mohawk (Mkx)^−/− rats by using CRISPR/Cas9, which showed not only systemic hypoplasia of tendons similar to Mkx^−/− mice, but also earlier heterotopic ossification of the Achilles tendon compared with Mkx^−/− mice. Analysis of tendon-derived cells (TDCs) revealed that Mkx deficiency accelerated chondrogenic and osteogenic differentiation, whereas Mkx overexpression suppressed chondrogenic, osteogenic, and adipogenic differentiation. Furthermore, mechanical stretch stimulation of Mkx^−/− TDCs led to chondrogenic differentiation, whereas the same stimulation in Mkx^+/+ TDCs led to formation of tenocytes. ChIP-seq of Mkx overexpressing TDCs revealed significant peaks in tenogenic-related genes, such as collagen type (Col)1a1 and Col3a1, and chondrogenic differentiation-related genes, such as SRY-box (Sox)5, Sox6, and Sox9. Our results demonstrate that Mkx has a dual role, including accelerating tendon differentiation and preventing chondrogenic/osteogenic differentiation. This molecular network of Mkx provides a basis for tendon physiology and tissue engineering.

Tendons play a critical role in the musculoskeletal system by connecting muscle to bone to transmit mechanical loads and enable movement. Tendon injuries and damage are repaired slowly and incompletely because of poor intrinsic healing capacity, which in part results from tissue hypocellularity and hypovascularity (1). Even after surgical tendon repair, a standard treatment for tendon rupture, clinical outcomes are not satisfactory because of recurrent rupture or adhesions (2). To develop cell-based or pharmacological approaches for promoting tendon repair, the molecular mechanism of tendon development and regeneration must be determined; however, the key genome network for tendon differentiation and homeostasis has not been well characterized.

We, along with other researchers, recently reported the tendon-specific expression and functions of the transcription factor Mohawk (Mkx), which regulates tendon-related gene expression (3, 4). Mkx knockout mice showed general tendon hypoplasia (5, 6), suggesting that Mkx plays an important role during tendon development. Moreover, overexpression of Mkx in mesenchymal stem cells (MSC) elevates tendon-related markers, and transplantation of these cells increases the diameter of collagen fibers in tendons (7, 8), suggesting the potential application of Mkx in cell therapy for tendon injury.

Although the results from analysis of Mkx knockout mice have provided critical information about tendon development, the utility of mice as an animal model has some limitations. In particular, regenerative experiments for tendon repair with precise surgical interventions are challenging, and most reports of cell therapy for tendon repair used animals that were larger than mice (7, 9). For physiological experiments, such as treadmill exercise to test the effect of mechanical load on tendons/ligaments, rats are preferable for analyzing the exact responses to the stress because they are physiologically more similar to humans than mice (10). It is also difficult to obtain a sufficient number of primary tenocytes from mice for tendon/ligament research.

To overcome these limitations, rats are frequently used in musculoskeletal research. However, technical challenges related to the isolation and culture of ES cells posed difficulties in generating genetically modified rats (10, 11). Recent developments in gene-editing technologies, such as zinc-finger nuclease (ZFN) (12), transcription activator-like effector nuclease (TALEN) (13), and clustered regularly interspaced short palindromic repeats/CRISPR associated proteins (CRISPR/Cas9) facilitate the generation of genetically modified rats in a one-step injection (14). The objective of this study was to investigate the function of Mkx in vitro and in vivo, including developmental phenotype and molecular targets, through the generation of Mkx knockout rats.

Results

Embryonic and Adult Rat Tendons Express Mkx.

Mouse tendons express Mkx in the embryonic and adult stages (3, 5, 15). To confirm that the same is true in rats, the expression of Mkx was analyzed at several developmental stages using whole-mount in situ hybridization. Embryonic day (E) 11.5 embryos showed expression of Mkx mRNA in the dermomyotome dorsomedial lip, forelimb, and hind limb (Fig. S1A). In E15.5 embryos, the expression of Mkx was identical to the tendon tract (15). In the postnatal stage, quantitative RT-PCR (RT-qPCR) revealed that tendon showed significantly higher expression of Mkx than other tissues (Fig. S1B).

Fig. S1.

Expression of Mkx in rats and Primers of Mkx for RT-qPCR. (A) Whole-mount in situ hybridization of Mkx at E11.5 embryo (Left) and E15.5 embryo (Right). (B) RT-qPCR analysis for Mkx in several organs of 3-wk-old Wistar rats. GAPDH was used as an internal control. Error bars, SEM (n = 3). (C) Scheme and sequence lists of Mkx primers for RT-qPCR. (D) RT-qPCR analysis in the patellar tendon of 3-wk-old Mkx^+/+ or Mkx^−/− rats with above primers. GAPDH was used as an internal control. Error bars, SEM (n = 3). ***P < 0.005.

Mkx Knockout Rats Show “Wavy Tails” and Systemic Hypoplasia of Tendons.

Expression of Mkx correlating with tendon development suggested that Mkx is a tenogenesis-related transcription factor not only in mice but also in rats. To evaluate the function of Mkx, we generated Mkx knockout rats using a CRISPR/Cas9 system. The mRNA of hCas9 and guide RNA (gRNA), which targets the second exon of the Mkx gene, were injected into 96 rat zygotes (Fig. 1A) (14). Twenty-six rats were obtained; direct sequencing of genomic DNA revealed that nine rats (34%) contained mutations at the target site. Of these nine chimeric rats (F0), three were crossed with Wistar rats and germ-line transmission was confirmed. Three lines of F1 rats were obtained. Putative off-target sites (16) of CRISPR/Cas9 were evaluated by direct sequencing, which showed no mutations (Table S1). Two primer pairs were designed to analyze the mRNA expression of Mkx in the patellar tendon. RT-qPCR with primers that flanked the target site showed significantly decreased Mkx expression. When one primer was designed precisely to be on the target site, the Mkx expression was reduced in Mkx^−/− rats, and these primer pairs were used for subsequent experiments (Fig. S1 C and D).

Fig. 1.

Generation of Mkx^−/− rats. (A) Target site of gRNA and result of direct sequencing. (B) Tendons of Mkx^+/+ or Mkx^−/− rats (8-wk-old). (C) H&E staining (Upper) and Picrosirius red staining (Lower) of the patellar tendons (black arrowhead) in 2-wk-old Mkx^+/+ or Mkx^−/− rats. (D) Tensile strength of the patellar tendon in Mkx^+/+ or Mkx^−/− rats. Absolute value of tensile strength (Left) and tensile strength per unit area (Right). Error bars, SEM (n = 3). (E) RT-qPCR analysis in the patellar tendon of 3-wk-old Mkx^+/+ or Mkx^−/− rats. GAPDH was used as an internal control. Error bars, SEM (n = 3). **P < 0.01; ***P < 0.005.

Table S1.

Off target analysis of Mkx^−/− rats

Site name	Sequence	Chromosome	Primers for sequence
Target	ACTCTTGGCTCTAGG	Ch17	Forward 5′–3′	Reverse 5′–3′
Off 1	ACTCTTGGCTCTAGG	Ch 6	GTAGCATTCAGTAGCTTCCCAGTTCAGAC	CTGCCCTTCGACTTGACTTTACTC
Off 2	GCTCTTGGCTCTAGG	Ch 5	CTCACTTTGGAGATCTGGAAGAAATGTTTGC	CTTTCTTGACATTCTTTCTATTACCAGGCAC
Off 3	CCTCTTGGCTCTAGG	Ch 6	GAAGTCATTCCCACATGTCCACATTATCATAG	CAGTAACAGACACAGTGGAAGAAAATCAGTAC
Off 4	GCTCTTGGCTCTAGG	Ch 6	GAGAGATCTTTGGACTGTCAAAACTACTAC	CTGAAATCCTGCCAGGTAAATGGTG
Off 5	TCTCTTGGCTCTAGG	Ch 7	GGACAAGCCTAAGCAGCTTACTAGAAG	GCAGAATTGGATTTGGCTTCTGGG
Off 6	CCTCTTGGCTCTAGG	Ch 8	GGTACCACATGGTCCAGAGGCAG	CTCCAGTGGCTCTTTGGAGCC
Off 7	CCTCTTGGCTCTAGG	Ch 8	GCTCTTCATTCCTGCTTTGGTATCTG	GCCACCTTATCATGTCTTCCTAAAACAGA
Off 8	ACTCTTGGCTCTAGG	Ch 9	CGTGCTTCTAGTCTGTGAATTATTCGC	GTCAATGAAGGGGCTTCTGTATTTGATG
Off 9	TCTCTTGGCTCTAGG	Ch 9	GACCTTCATCCCCAATGTAACAGG	GCATTGGCTCATGGTTCTGGAG
Off 10	ACTCTTGGCTCTAGG	Ch 9	CTGAGAAAAGGCTTTAACGAGTCCATG	GAGGGTCCTCCTAAAATGCCATTATAGT
Off 11	ACCCTTGGCTCTAGG	Ch14	CCCTATCATCATTCTCCTCGAGGTCTAG	GGAGACACATGGGGCATTTTCCTC

Off indicates off-target candidate sites. Mismatches from the target are in bold and underlined.

Mkx^−/− mice are known to have a “wavy tail” phenotype, which becomes more apparent when the mice are running (6), but Mkx^−/− rats have a more profound wavy tail phenotype without running (Fig. S2). Mkx^−/− rats manifest general hypoplasia of tendons, such as the flexor and extensor tendons of limbs, tail tendons, and patellar tendons while maintaining collagen orientation (Fig. 1 B and C). The tensile strength of the patellar tendons was decreased in the Mkx^−/− rats even after normalization with their cross-sectional area (Fig. 1D). These wavy tail phenotypes and general hypoplasia of the tendon were confirmed in all three knockout lines (Fig. S2). One of these lines (14-deletion) was used for subsequent experiments. mRNA levels of tendon-related markers, such as tenomodulin (Tnmd), collagen type 1 α 1 (Col1a1), fibromodulin (Fmod), decorin (Dcn), and tenascin XB (Tnxb), were decreased in the patellar tendons of Mkx^−/− rats (Fig. 1E). Scleraxis, known as a tenogenesis-related transcription factor (17), was also decreased in Mkx^−/− rats but the difference was smaller compared with the other tendon markers. Osteogenesis- and chondrogenesis-related genes, such as Col2a1, Acan, Runx2, Alpl, and IBSP, were elevated in Mkx^−/− rats. Transmission electron microscopy (TEM) showed that the collagen fibril diameter in the tail tendon of Mkx^−/− rats was uniformly smaller than that in Mkx^+/+ rats (Fig. S3). These phenotypes of Mkx^−/− rats were similar to those of Mkx^−/− mice (5, 6), but were severe and more readily observed because of their larger sizes.

Fig. S2.

Sequences and phenotypes of three lines of Mkx^−/− rats.

Fig. S3.

Diameter of collagen fibrils in Mkx^+/+ or Mkx^−/− rats. (A) TEM view of collagen fibrils in the tail tendons of Mkx^+/+ or Mkx^−/− rats. (B) Diameter of collagen fibrils. Error bars, SEM (n = 100). ***P < 0.005.

Heterotopic Ossification in Achilles Tendon of Mkx^−/− Rats.

Mkx^−/− rats showed elevated osteogenic and chondrogenic markers in the patellar tendon, suggesting heterotopic ossification in the tendon. To check general ossification of Mkx^−/− rats, microcomputed tomography (microCT) was performed, which revealed early heterotopic ossification in the Achilles tendon of Mkx^−/− rats (Fig. S4A). Heterotopic ossification occurred in 20% of 3-wk-old Mkx^−/− rats and in all 5-wk-old Mkx^−/− rats (Fig. S4B). The cross-sectional area of the ossifications gradually increased until the age of 15 weeks in Mkx^−/− rats (Fig. S4C). No heterotopic ossification in other tendons or ligaments was observed in 15-wk-old Mkx^−/− rats (Fig. S5). Histochemistry of the Achilles tendon revealed that postnatal day 9 (P0) Mkx^−/− rats already had chondral lesions in their Achilles tendon (Fig. 2A). Endochondral ossification occurred from the middle of the chondral lesion in 3- to 4-wk-old animals. To compare rats and mice, Venus knockin homozygous mutant mice (5) were also analyzed. P0 and 4-wk-old Mkx^−/− mice had no chondral lesion or ossification in the Achilles tendon (Fig. S4D). RT-qPCR on RNA from the Achilles tendon revealed that expression of chondrogenic genes increased in the 2-wk-old Mkx^−/− rat Achilles tendon (Fig. 2B). The mRNA levels of osteogenic-related genes increased in Mkx^−/− rats from 2 to 4 wk of age. The expression of bone morphogenetic protein (BMP) pathway-related genes, such as BMPr1a, BMPr2, Smad1, and Smad5 was also elevated in Mkx^−/− rats. Immunohistochemistry also showed expression of chondrogenic/osteogenic genes at the site of heterotopic ossification in Mkx^−/− rats (Fig. S4E). To evaluate the influence of heterotopic ossification on the physiological function, gait analysis was performed, which revealed a significant decrease in the maximum ankle plantar flexion of Mkx^−/− rats (Fig. S6 and Movies S1 and S2).

Fig. S4.

Achilles tendon of Mkx^−/− mice and rats. (A) Micro-CT of the Achilles tendons of Mkx^+/+ or Mkx^−/− rats. (B) Proportion of ossification in the Achilles tendon of Mkx^−/− rats (n = 10). (C) Cross-sectional area of ossification in the Achilles tendon of Mkx^−/− rats. Error bars, SEM (n = 7). (D) Achilles tendon of P0 and 4-wk-old Mkx^−/− mice. (E) Immunohistochemistry of the Achilles tendon in 3-wk-old Mkx^+/+ or Mkx^−/− rats. Dashed line indicates the border between normal tendon and ossified region.

Fig. S5.

Micro-CT of 15-wk-old Mkx^−/− rats.

Fig. 2.

Heterotopic ossification in the Achilles tendon of Mkx^−/− rats. (A) H&E staining, Safranin O-Fast green staining, and Alizarin red staining of the Achilles tendons in P0, 3-wk-old, and 4-wk-old Mkx^+/+ or Mkx^−/− rats. (B) RT-qPCR analysis in the Achilles tendon of 2-wk-old and 4-wk-old Mkx^+/+ or Mkx^−/− rats. Gapdh was used as an internal control. Error bars, SEM (n = 3); ns, not significant. *P < 0.05; **P < 0.01; ***P < 0.005.

Fig. S6.

Gait analysis of Mkx^+/+ or Mkx^−/− rat. Gait analysis of the ankle joint angle in Mkx^+/+ or Mkx^−/− rats. MT, metatarsal head. Error bars, SEM (n = 3). ***P < 0.005.

Mkx Deficiency Accelerates Chondrogenic and Osteogenic Differentiation of Tendon-Derived Cells.

Endochondral ossification involves mesenchymal stem cell condensation and chondrocytic differentiation (18). Endochondral ossification in the Achilles tendon of Mkx^−/− rats prompted us to test whether Mkx is critical for restricting mesenchymal cell differentiation into tenocytes and for preventing chondrocytic differentiation. It was previously reported that a stem cell population, known as tendon stem/progenitor cells (TSPCs), were enriched in tendon tissues (19).

We isolated and cultured tendon-derived cells (TDCs) from the patellar tendons of 3-wk-old Mkx^+/+ and Mkx^−/− rats (Fig. 3A). There were no differences in the morphology of Mkx^+/+ and Mkx^−/− TDCs (Fig. S7A) and FACS analysis revealed that these cell preparations did not contain hematopoietic stem cells (Fig. S7B). These TDCs were capable of osteogenic, chondrogenic, and adipogenic differentiation, supporting the idea that the TDCs contained stem/progenitor populations.

Fig. 3.

Mkx regulates chondrogenic, osteogenic, and adipogenic differentiation in TDCs. (A) Protocol for isolation of TDCs. (B) Appearance of pellets of Mkx^+/+ or Mkx^−/− TDCs after chondrogenic differentiation (Alcian blue staining). (C) Diameter of the pellets of Mkx^+/+ or Mkx^−/− TDCs. Error bars, SEM (n = 3). (D) Appearance of the wells of Mkx^+/+ or Mkx^−/− TDCs after osteogenic differentiation (Alizarin red staining). (E) Relative absorbance at 450 nm of Alizarin red dye elution. Error bars, SEM (n = 3). *P < 0.05; **P < 0.01; ***P < 0.005.

Fig. S7.

Analysis of TDCs. (A) Morphology of Mkx^+/+ or Mkx^−/− TDCs. (B) Flow cytometry analysis of the expression of CD90 and CD45 in Mkx^+/+ or Mkx^−/− TDCs. (C) RT-qPCR analysis in Mkx^+/+ or Mkx^−/− TDCs after chondrogenic differentiation. GAPDH was used as an internal control. Error bars, SEM (n = 3). (D) RT-qPCR analysis in Mkx^+/+ or Mkx^−/− TDCs after osteogenic differentiation. GAPDH was used as an internal control. Error bars, SEM (n = 3). (E) Mkx^+/+ or Mkx^−/− TDCs after adipogenic differentiation and AdipoRed staining. (F) Ratio of the AdipoRed staining-positive cells. Error bars, SEM (n = 3). (G) Immunocytochemistry of Mkx^−/− TDCs after retrovirus infection. (H) Real-time PCR analysis in Mkx^−/− TDCs after retrovirus infection. GAPDH was used as an internal control. Error bars, SEM (n = 3). (I) Western blot of FLAG and β-actin in Mkx-overexpressing TDCs. (J) RT-qPCR analysis in Mkx-overexpressing TDCs after chondrogenic differentiation. GAPDH was used as an internal control. Error bars, SEM (n = 3). (K) Appearance of the wells of Mkx-overexpressing TDCs after osteogenic differentiation and Alizarin red staining. (L) Relative absorbance at 450 nm with Alizarin red dye elution. Error bars, SEM (n = 3). (M) RT-qPCR analysis in Mkx-overexpressing TDCs after osteogenic differentiation. GAPDH was used as an internal control. Error bars, SEM (n = 3). (N) Adipogenic differentiation and AdipoRed staining of Mkx-overexpressing TDCs. (O) Ratio of cells staining positive for AdipoRed. Error bars, SEM (n = 3). (P) Luciferase analysis of ChIP-seq peak sites. Error bars, SEM (n = 3). *P < 0.05; **P < 0.01; ***P < 0.005.

Under chondrogenic differentiation conditions, the pellets of Mkx^−/− TDCs were larger than those of Mkx^+/+ TDCs (Fig. 3 B and C). RT-qPCR revealed that the expression of chondrogenic markers was higher in Mkx^−/− TDCs after 1 wk of chondrogenic differentiation (Fig. S7C). Under osteogenic differentiation, Alizarin red staining revealed greater ossification in Mkx^−/− TDCs than in Mkx^+/+ TDCs at 14 d (Fig. 3 D and E). The difference in ossification between Mkx^+/+ and Mkx^−/− decreased at 21 d, but remained significant. RT-qPCR showed higher expression of osteogenic genes in Mkx^−/− TDCs (Fig. S7D) and there were no differences in adipogenic differentiation between Mkx^+/+ and Mkx^−/− TDCs (Fig. S7 E and F). These data suggest that Mkx deficiency leads to enhanced osteogenic and chondrogenic differentiation of TSPCs.

Mkx Overexpression Suppressed Chondrogenic, Osteogenic, and Adipogenic Differentiation of Mkx^−/− TDCs.

To rescue the loss of Mkx, Mkx^−/− TDCs were retrovirally transduced with the Mkx coding sequence containing a FLAG tag. Retrovirus-encoding Venus protein was also used as a control. Induction of FLAG-tagged Mkx was confirmed by RT-qPCR, Western blotting, and immunocytochemistry (Fig. S7 G–I).

Mkx-transduced Mkx^−/− TDCs showed lower expression of chondrogenic markers after chondrogenic differentiation (Fig. S7J). In the osteogenic differentiation condition, Mkx-transduced TDCs had fewer calcium deposits (Fig. S7 K and L) and reduced osteogenic markers (Fig. S7M). Mkx-transduced TDCs lost the ability to differentiate into adipocytes (Fig. S7 N and O).

Mechanical Stretch Stimulation of Mkx^−/− TDCs Leads to Chondrogenic Differentiation.

Tendons respond to appropriate mechanical strains by increasing collagen production in tenocytes (20) and mechanical strains promote MSC differentiation into tenocytes (21). As shown above, Mkx^−/− TDCs showed a strong ability to differentiate into osteocytes and chondrocytes, suggesting that the loss of Mkx affects the response to mechanical stress of TDCs. To investigate this theory, TDCs were subjected to mechanical stretch stimulation (Fig. 4). After 4% monoaxial cyclic elongation for 6 h, Mkx^+/+ TDCs showed elevated levels of tendon-related genes, such as Mkx, Col1a1, and Col3a1, indicating the tenogenic differentiation. However, the same mechanical stimulation of Mkx^−/− tendon-derived cells increased chondrogenic markers, such as SRY-box (Sox)6, Sox9, and Acan, rather than tendon-related genes.

Fig. 4.

Mechanical stretch stimulation of Mkx^−/− TDCs leads to chondrogenic differentiation. (A) Protocol of mechanical stretch stimulation. (B) Real-time PCR analysis in Mkx^+/+ or Mkx^−/− TDCs after mechanical stretch stimulation. GAPDH was used as an internal control. Error bars, SEM (n = 3). *P < 0.05; **P < 0.01; ***P < 0.005.

Both Tendon-Related and Chondrogenic Differentiation-Related Genes Are Putative Targets of Mkx.

Although Mkx appears to be a critical transcription factor for tendon development and homeostasis, a genome-wide approach to identify the direct targets of Mkx in tenocytes has not yet been reported, partly because of the difficulty of assembling a sufficient number of samples in mice. In the experiments described above (Fig. 3 and Fig. S7 G–O), we successfully rescued the Mkx^−/− TDC phenotype by overexpression of Mkx. Because a ChIP-grade antibody for mice and rats is not yet available, we attempted to perform ChIP-sequencing using the Mkx^−/− TDCs overexpressing tagged Mkx. The hemagglutinin (HA) tag-fused Mkx binding region in Mkx^−/− TDCs was analyzed by ChIP-seq in the next-generation sequencer MiSeq (Fig. 5A). We obtained 6,356,463 sequence reads with the anti-HA antibody ChIP sample and 7,541,415 sequence reads from input DNA. The sequencing data were aligned with the rat genome (rn6) using bowtie software (22), resulting in 4,177,212 read maps of ChIP and 4,961,690 read maps of input samples. By using the mapped sequence reads, Mkx binding regions were detected by model-based analysis of ChIP-seq (MACS) (23) using the default parameters. The following analysis revealed 6,000 peaks of putative Mkx binding sites. Within these peaks, protein-coding genes were selected and functional annotation was performed using DAVID v6.7 (https://david.ncifcrf.gov/home.jsp). Among the functional annotation chart, “skeletal system development (GO: 0001501)” was listed with a P value 7.6E-11, which included tendon-related genes, such as Col1a1, Col3a1, and Mkx. Interestingly, chondrogenic differentiation-related genes, such as Sox5, Sox6, and Sox9 were also included in the ontology (Fig. 5B). Moreover, other tendon-related genes, such as tenascin C (Tnc) and Fmod, were also included within these peaks. De novo motif discovery of these peaks using MEME-ChIP software (24) revealed three binding motifs, and two of them contained A-C-A, which is the putative binding site of mouse Mkx (25) (Fig. 5C). To support the physical interaction between Mkx and promoters, peaks in the promoter regions of Mkx, Col1a1, and Col3a1 were inserted into a thymidine kinase (TK) minimal promoter luciferase vector, and luciferase assays were performed after coexpression of Mkx tagged with the VP16 effector. As a result, higher luciferase activity was observed with VP16-Mkx expression compared with the control (Fig. S7P). This result indicates that these promoter regions may interact with Mkx either directly or indirectly.

Fig. 5.

ChIP-seq of HA-Mkx–overexpressing TDCs revealed that Sox5, Sox6, and Sox9 are putative targets of Mkx. (A) Protocols of ChIP-seq in HA-Mkx–overexpressing TDCs. (B) ChIP-seq peaks of Mkx. Red bars indicate the peak. The distance from the transcription start site is shown above each peak. (C) De novo motif analysis of ChIP-seq.

Discussion

Although the rat is a preferable experimental animal compared with the mouse in several medical and biological research fields, including that of the musculoskeletal system, limited information is available regarding the use of knockout rats. This limitation is because of the technical difficulty in manipulating rat ES cells (26) and germ-line stem cells (27) for targeted genome deletion compared with that in the widely used mouse ES cells (11). In this regard, recent genome-editing technologies, such as ZFN, TALEN, and CRISPR/Cas9, are powerful strategies by which gene knockouts can be generated in various species of animals without using ES cells and homologous recombination (10). Here, we successfully generated genetically modified rats with deletion of Mkx with CRISPR/Cas9. The analysis of Mkx^−/− rats not only confirms but also extends our knowledge on Mkx-dependent tendon differentiation and regulation, by allowing us to observe a more severe phenotype of heterotopic ossification in the Achilles tendon, to perform the physiological assessment of the ankle joint angle during ambulation, and to collect sufficient amounts of primary tenocytes from knockout rats for mechanistic analyses and ChIP-seq.

In human, heterotopic ossification is a substantial medical problem because it is associated with pain and dysfunction (28). In systemic heterotopic ossification, a fibrodysplasia ossificans progressiva-like phenotype (29) and ossification of the posterior longitudinal ligament of the spine (30) are important hereditary diseases, although these ossifications occur only several years after birth. BMP4 transgenic mice (31) and Npps^−/− mice, referred to as tip-toe-walking mice (32), have been considered as animal models of these human diseases, with a fibrodysplasia ossificans progressiva-like phenotype and posterior longitudinal ligament of the spine, respectively. Biglycan (Bgn)- and Fmod-deficient mice showed heterotopic ossification in not only the Achilles tendon, but also around the knee joint (33). These Bgn/Fmod-deficient mice also showed decreased diameter of collagen fibrils, similar to that observed in Mkx knockout mice and rats. Our data also indicate that Fmod expression in the patellar tendons of Mkx^−/− rats was decreased. Furthermore, ChIP-seq of Mkx showed significant peaks around Fmod, which suggest an interaction between Fmod and Mkx.

Although many patients with Achilles tendon heterotopic ossification have a history of trauma or surgery, hereditary factors are thought to be involved in the etiology of this disorder (34). Excessive stress on the Achilles tendon, as well as surgical intervention or injection of growth factors (35–37) have been shown to cause heterotopic ossification of the tendon.

Here, we observed Achilles tendon ectopic ossification in neonatal Mkx^−/− rats. This ectopic ossification was related to an endochondral ossification program; at birth, chondrogenesis occurs in the center of the Achilles tendon and then cartilage tissues are replaced by bone tissues. The precise molecular mechanisms of this heterotopic ossification of the tendon are not fully understood; however, our data support that idea that mesenchymal cells (i.e., TSPCs), which should differentiate into tenocytes, may lose their fate without Mkx and can differentiate into chondrocytes during embryogenesis. Several factors, such as Indian hedgehog, insulin-like growth factors, and BMPs, regulate the behavior of chondrocytes during endochondral ossification (38, 39). The up-regulation of BMP pathway-related genes in the Achilles tendon of Mkx^−/− rats suggests that Mkx regulates the BMP pathway.

Mechanical stress affects tendon development before and after birth (40), and excessive mechanical stress can lead to ossification of the tendon (28, 41). It is also reported that mechanical strain promotes the differentiation of MSCs into tenocytes in vitro (21, 42). In this regard, we recently found that Gtf2ird1 translocates into the nucleus in response to mechanical strain and activates the Mkx promoter through chromatin regulation (43). Here, our mechanical stretch experiments with TDCs support the idea that mechanical stimulation causes Mkx^−/− TDCs but not wild-type cells to undergo chondrogenic differentiation. The reason why rats showed a more severe phenotype than mice with Achilles tendon ossification may be explained as follows. Rats are larger than mice, which may increase mechanical stimulation to the Achilles tendon during embryogenesis and more readily stimulate Mkx^−/− TSPCs to undergo chondrogenic differentiation. In Mkx^−/− rats, ectopic chondrogenesis in the Achilles tendon was observed in the embryonic and neonatal stages; however, increased chondrogenic marker gene expression was terminated by 4 wk, suggesting that additional chondrogenic changes may not occur in the Achilles tendon after birth. The discrepancy of cartilaginous change phenotypes between the embryonic stage and after birth may reflect the pluripotency of TSPCs in embryos and adults. In this regard, it is of great interest to examine whether there would be a difference in TSPCs from Mkx^−/− rat embryos and adults with mechanical loading. In addition, whether excessive exercise promotes ectopic ossification in other tendons in mature Mkx^−/− rats should be examined in the future.

Overexpression of Mkx has been shown to promote tendon-related gene expression and to repress gene expression characteristic of other cell lineages (7, 8, 25, 44). Consistent with these previous reports, our study also showed that osteogenic and chondrogenic differentiation occurs more readily in TDCs from an Mkx^−/− background than in those from an Mkx^+/+ background (Fig. 3 B–E). Regulation of Mkx expression was shown to affect expression of essential extracellular matrix genes of tendon tissues, such as Col1a1, decorin, and Tnc, and chondrogenesis master genes. These in vitro observations, as well as the in vivo ectopic ossification phenotype, indicate that the potential function of Mkx is to regulate cell fate of TSPCs via repressing chondrogenenic factors. Our ChIP-seq data revealed that Mkx interacts with both extracellular matrix genes and chondrogenic genes. Further detailed experiments are needed to clarify the precise molecular mechanisms of how Mkx coordinates expression of this diverse set of genes.

Therefore, we show in Mkx knockout rats that Mkx plays a critical role in TSPC differentiation to tenocytes and development of tendon tissues. These findings indicate that Mkx can be applied as a therapeutic target for tendon repair or tissue engineering. In addition, the Mkx knockout rats represent a powerful animal model for further research on musculoskeletal tissues and diseases.

Materials and Methods

Detailed materials and methods were described in SI Materials and Methods. The list of primer sequences for RT-PCR is shown in Table S2.

Table S2.

Primers for RT-qPCR and Plasmid

Primer	Forward 5′–3′	Reverse 5′–3′
Primers for RT-qPCR
Acan	CTGTCTATCTGCACGCCAACC	CCTCTTCACCACCCACTCCGA
Alpl	GCACAACATCAAGGACATCG	TCAGTTCTGTTCTTGGGGTACAT
Bmpr1a	GCGTTTCACACACACGACCTC	GTACATGGCTCCAGAAGTGCTG
Bmpr2	CTTTGCCCTCCTGCTTCTTGG	CCAAGGTCTTGTTGATACGGGTC
Col10a1	CCAGGACACAATACTTCATCCCATACC	CCAGGAATGCCTTGTTCTCCTCTTAC
Col1a1	GTCCGAGGTCCTAATGGAGATGC	GGTCCAGGGAATCCGATGT
Col2a1	CCAGGTCCTGCTGGAAAA	CCTCTTTCTCCGGCCTTT
Col3a1	GAGGAATGGGTGGCTATCCT	GGTATCCAGGAGAACCAGGAG
Col4a1	CCATGGTCAGGACTTGGGTA	AAGGGCATGGTGCTGAACT
Comp	GATGCCTGCGACGACGACATAGATG	CCATCACCATCGCTGTCACTCTG
Dcn	GACTCCACGACAATGAGATCACC	GTTGCCATCCAGATGCAGTTC
Fmod	CAAGGCAACAGGATCAATGAG	CTGCAGCTTGGAGAAGTTCA
Gapdh	GGCAAGTTCAATGGCACAGT	TGGTGAAGACGCCAGTAGACTC
Ibsp	GAGATGGCGATAGTTCGGAGGAG	CTCCGCCTCCTGGTCTTCATTC
Opn	GCAGCTCAGAGGAGAAGGCG	CCTGACCCATCTCAGAAGCAGA
Runx2	CCACAGAGCTATTAAAGTGACAGTG	AACAAACTAGGTTTAGAGTCATCAAGC
Scx	GCACCTTCTGCCTCAGCAAC	TTCTGTCACGGTCTTTGCTCA
Smad1	GGTTCCAAGCAGAAGGAAGTCTG	GTTGTACTCGCTGTGCCTCG
Smad5	CAACTCAACCATCGAGAACACCAG	GCTGTCACTAAGACACTCGGC
Sox6	TTCTTCACTGTGGGGCAAC	GCACCAGGATACACAACACCT
Sox9	ATCTTCAAGGCGCTGCAA	CGGTGGACCCTGAGATTG
Tnmd	CTACAGCAATGGCGAGAAGAAGAAG	GACCTACAAAGTAGATGCCAGTGTATC
Tnxb	GACCACAAGTACAAGATGAACC	GGTCCAGGAGAGGCTCA
Primers for cloning to pGL4.12 TK-luciferase vector
Col1a1	aaaaaaggtaccCTTTGTTTTTGAGGGCCTGTGCTG	ttttttctcgagAGCACCGTCAGGAAAAGAAAATTCAC
Col3a1	aaaaaaggtaccCTGTCTAGCCTCTGATGGTGGAATG	ttttttctcgagCCAGCCAAACAGAACATGTAGGAC
Mkx	aaaaaaggtaccGCTCCGGACAGCTTCTCCTATTG	ttttttctcgagCTGGGTGAGCCTAGGGTTCAG

Preparation of hCas9 and gRNA.

The preparation of hCas9 and gRNA has been described previously (45). Briefly, the gRNA, the target of which was designed to the second exon of the Mkx gene, was constructed by inverse PCR. In vitro RNA synthesis and purification were performed.

Generation of Mkx Knockout Rats and Genotyping.

All animal experiments were approved by the Institutional Animal Care and Use Committee at the Tokyo Medical and Dental University. The gRNA and hCas9 mixture RNA was microinjected into the cytoplasm of Wistar rat zygotes by the UNITECH Corporation. The resulting chimeric offspring were crossed with Wistar rats and germ-line transmission was confirmed by sequencing.

Tensile Testing.

Patellar tendons from 6-mo-old wild-type or Mkx^−/− rats (n = 3) were pulled at a constant strain rate of 0.05 mm/s by a uniaxial materials testing system (5).

Isolation of Rat TDCs.

Patellar tendons of 3-wk-old wild-type or Mkx^−/− rats were dissected. The samples were cut into small pieces and digested with collagenase (Sigma). After filtration with a nylon filter, digested cells were cultured. All experiments were performed until passage 5.

Adipogenic/Osteogenic Differentiation.

TDCs were plated into 24-well plates at 37 °C. After 24 h, the medium was changed to Adipogenesis Induction Medium (Lonza) and incubated for 7 d for adipogenic differentiation. The medium was changed to Osteogenesis Induction Medium (Lonza) and incubated for 14 d, and Alizarin red staining was performed for osteogenic differentiation.

Chondrogenic Differentiation and Alcian Blue Staining.

TDCs were suspended in Chondrogenic Incomplete Medium (Lonza). After centrifuging and changing the medium to Chondrogenic Complete Medium (Lonza), the pellets were incubated for 21 d and Alcian blue staining was performed.

Retrovirus Infection.

Venus, FLAG-Mkx, or HA-Mkx was inserted to the MIGR vector (Addgene). The plasmids were transfected into PLAT-E cells. After 24 h, the filtered supernatant was used to infect TDCs (P1) derived from an Mkx^−/− rat.

Mechanical Stretch Stimulation.

Cells were seeded into the elastic silicon rubber chambers 12 h before stretching. The chambers were set on a monoaxial stretching device (STB-140, Strex) and monoaxial cyclic strain was applied for 6 h.

ChIP.

After fixation, the cells were washed with cell lysis buffer containing protease inhibitors, and resuspended in nuclear lysis buffer containing protease inhibitors. Chromatin was fragmented to 100–400 base pairs using sonication. The solution was then incubated with HA or normal rabbit IgG antibodies bound to beads. The immunoprecipitates were eluted from the beads, incubated to reverse the cross-linking, and purified for DNA analysis.

ChIP-seq Library Preparation and Data Analysis.

DNA libraries for next-generation sequencing were prepared using the TruSeq ChIP Sample Preparation kit (Illumina) from 3-ng ChIP DNA or input DNA and sequenced on a MiSeq (Illumina). ChIP DNA-enriched regions were detected by MACS v1.4.2 with default parameters. De novo motif discovery was performed with MEME-ChIP (24) using the default parameter.

Luciferase Reporter Assays.

HEK293FT cells were seeded in 96-well plates at 30% confluence and were transfected with pcDNA-VP16-Mkx or control pcDNA-Venus along with firefly and Renilla luciferase reporters. Thirty-six hours after transfection, luciferase activity was measured. The results were normalized to Renilla luciferase activity.

Statistical Analysis.

The two-tailed independent Student’s t test was used to calculate the P values.

SI Materials and Methods

Preparation of non–RI-RNA Probe and Whole-Mount in Situ Hybridization.

To prepare the non–RI-RNA probe, RT-PCR was performed using RNA from the Achilles tendon of a Wistar rat (Sankyo Labo Service Corporation) using the following primers: F-GAAGGTGAGGCACAAGCGAC and R-GATGATGGAGACACCAGTTCTTCTTCA. After RT-PCR, the PCR product was cloned into a cloning vector. In vitro transcription was performed using the cloned vector as a template with the following primers: F-AATACGACTCACTATAGGG (T7-promoter) and R-GAAGGTGAGGCACAAGCGAC. A digoxigenin (DIG)-labeled antisense riboprobe was transcribed using the DIG-RNA labeling mix (Roche) following the manufacturer’s instructions. The whole-mount in situ hybridization procedure has been described previously (5).

Quantitative Real-Time PCR.

Ten-week-old male Wistar rats were purchased from Sankyo Labo Service Corporation. RNA was extracted using Isogen (Nippongene) from the tendon, muscle, bone, cartilage, brain, eye, thyroid, thymus, lung, heart, liver, spleen, kidney, and prostate. After reverse transcription using ReverTra Ace (Toyobo), qPCR was performed using THUNDERBIRD SYBR qPCR Mix (Toyobo). Gene-expression changes were quantified using the ΔΔC_T method. This experiment was performed with three independent samples and confirmed based on reproducibility. The list of primer sequences for RT-PCR is shown in Table S2. GAPDH was used as an internal control.

Preparation of hCas9 and gRNA.

The preparation of hCas9 and gRNA has been described previously (45). Briefly, the hCas9 and gRNA cloning vectors (14) were purchased from Addgene. The gRNA target was designed to the second exon of the Mkx gene (5′-AGAAGATACTCTTGGCTCTAGG-3′). The gRNA expression vector was constructed by inverse PCR using the following primers (F: 5′-AGAAGATACTCTTGGCTCTGTTTTAGAGCTAGAAATAGCAAG-3′; and R: 5′-AACAGAGCCAAGAGTATCTTCTCGGTGTTTCGTCCTTTCCAC-3′). After digestion with DpnI, the PCR product was transformed into Escherichia coli (DH5a) cells. Insertion of the Mkx target was confirmed by sequencing. The T7 promoter sequence was added to the 5′ end of the forward primers to prepare the template using the following primers (hCas9-F: 5′-TAATACGACTCACTATAGGGAGAATGGACAAGAAGTACTCCATTGG-3′; hCas9-R: 5′-TCACACCTTCCTCTTCTTC-3′; gRNA-F: 5′-TTAATACGACTCACTATAGGAGAAGATACTCTTGGCTCT-3′; and gRNA-R: 5′-AAAAGCACCGACTCGGTGCC-3′). In vitro RNA synthesis and purification were performed using the mMessage mMachine T7 kit and MEGAclear kit (Life Technologies) according to the manufacturer’s instructions. The RNA of gRNA and hCas9 were mixed together at 250 ng/μL each (total 500 ng/μL).

Generation of Mkx Knockout Rats and Genotyping.

All animal experiments were approved by the Institutional Animal Care and Use Committee at the Tokyo Medical and Dental University. The gRNA and hCas9 mixture RNA was microinjected into the cytoplasm of Wistar rat zygotes by the UNITECH Corporation. Genomic DNA was extracted from tail tips of rat pups and PCR was performed using the following primers (F: 5′-GTGACAACCCGTACCCTACGAAGACTG-3′; and R: 5′-CCACGTTAGAGGAGTCTGGATTACCTGC-3′). PCR products were purified using ExoSAP-IT (Affymetrix) and direct sequencing was performed. The resulting chimeric offspring were crossed with Wistar rats and germ-line transmission was confirmed by sequencing.

Off-Target Analysis.

According to previous reports (16), putative off-target sites were defined as having 1 or 0 mismatch in the seed region (the 12-base region directly upstream of the PAM sequence) and were identified using BLAT (https://genome.ucsc.edu/cgi-bin/hgBlat?command=start). After PCR with the primers of putative off-target sites (Fig. S2), direct sequencing was performed.

Histochemistry and Immunohistochemistry.

After killing of the rats, rat tissues were fixed with 4% (wt/vol) PFA/PBS for 2.5 h, embedded in SCEM (Section-Lab), and frozen. Undecalcified frozen sections that were 3-μm thick were obtained with adhesive films (Section-Lab) (46). After washing with ethanol and PBS, sections were stained with H&E (Wako), Safranin-O-Fast green stain (Wako), Alizarin red S (Sigma), or Picrosirius red stain kit (Polysciences).

For immunohistochemistry, frozen sections with adhesive films were washed with ethanol and PBS. After incubating in 10% (wt/vol) EDTA/PBS (pH 7.1–7.2) for 1 h, the sections were blocked in Blocking One solution (Nacalai Tesque) for 1 h. After blocking, the samples were incubated with anti-Runx2 antibody (1:500; ab23981, Abcam), anti-Sox6 antibody (1:500; ab30455, Abcam), anti-Sox9 antibody (1:500; ab5535, Abcam), and antiosteopontin antibody (1:500; ab8448, Abcam) overnight at 4 °C. The sections were then incubated with Alexa Fluor 488 donkey anti-rabbit antibody (1:500; Life Technologies) and Hoechst 33342 (1:2,000; Life Technologies) for 1 h.

Tensile Testing.

Hindlimbs of 6-mo-old wild-type or Mkx^−/− rats (n = 3) were obtained. The femurs and quadriceps muscles were removed and the quadriceps tendons, patella, patellar tendons, and the tibia complexes were obtained. The width and thickness of the central one-third of the patellar tendons were measured using a precision caliper to obtain the cross-sectional areas. A uniaxial materials testing system (Autograph AGS-X, Shimadzu) was used. To facilitate gripping during testing, the proximal end of the patella and quadriceps tendons were sandwiched with sandpaper using glue and clamped. The specimens were pulled at a constant strain rate of 0.05 mm/s. All samples broke within the gauge length. Force data were collected using Trapezium (Shimadzu) software every 0.01 s (5).

Gait Analysis.

The right hindlimbs of 6-mo-old wild-type or Mkx^−/− rats (n = 3) were shaved, and three anatomical landmarks (the lower third of the tibia, lateral malleolus, and fifth metatarsal head) were marked with a black ink marker (47). Ankle joint angle was defined as the angle between these three markers. Movements of the right hindlimb markers were recorded during treadmill walking at a speed of 15 m/min using a high-speed digital image camera (Sports CoachingCAM, Logical Product) at 240 images per second for 20 s. Dartfish software (Dartfish) was used to recognize the markers and to calculate the ankle joint angles (48). Maximum ankle plantar flexion per gait cycle was subsequently calculated.

Transmission Electron Microscopy.

The tail tendons of wild-type or Mkx^−/− rats were fixed with 2.5% (vol/vol) glutaraldehyde solution for 2 h. After washing with 0.1 M PBS for 1 h, postfixation was performed with 1% OsO₄ solution for 2 h. After dehydration through an ethanol series, the samples were embedded in Epon812 and cut into ultrathin (90 nm) sections. Sections were double-stained with uranyl acetate and lead citrate, and then examined by TEM (H-7100, Hitachi) (49). Collagen fibril (n = 100) diameters were measured three times in different views for fibril diameter analysis.

MicroCT.

After a general anesthesia with isoflurane, wild-type or Mkx^−/− rats were analyzed by microCT (R-mCT2, Rigaku). To calculate the amount of ossification of the Achilles tendons, the maximum cross-sectional area of sagittal plane was obtained and measured using ImageJ.

Isolation of Rat TDCs.

Patellar tendons of 3-wk-old wild-type or Mkx^−/− rats were dissected and stored in PBS. The samples were cut into small pieces and digested with 3 mg/mL collagenase (Sigma) for 1 h at 37 °C. After filtration with a 70-μm nylon filter for single-cell suspension, digested cells were cultured in α-MEM+10% (vol/vol) FBS, 100 U/mL penicillin, and 100 μg/mL streptomycin at 37 °C with 5% CO₂ incubation. After 48 h from the initial plating, the dishes were washed twice with PBS. The adherent cells (P0) were cultured for ∼10 d until 80% confluency and trypsinized. All experiments were performed until passage 5.

Flow Cytometry.

FITC-conjugated anti-rat CD90 and PE-cy7-conjugated anti-rat CD45 (BD Biosciences) antibodies were used. Flow cytometry was performed on a FACSCalibur HG flow cytometer with CELL Quest Pro software (BD Biosciences). Dead cells were excluded by gating of forward and side scatters.

Adipogenic Differentiation.

TDCs (2 × 10⁴ cells) were plated into 24-well plates at 37 °C. After 24 h, the medium was changed to Adipogenesis Induction Medium (Lonza) and the cells were incubated for 7 d, changing the medium every 3–4 d. The cells were stained with AdipoRed (Lonza) according to the manufacturer’s instructions. The ratio of adipocytes to all cells was counted in three different views.

Osteogenic Differentiation and Alizarin Red Staining.

Tendon-derived cells (2 × 10⁴ cells) were plated into 24-well plates at 37 °C. After 24 h, the medium was changed to Osteogenesis Induction Medium (Lonza) and incubated at 37 °C, changing the medium every 3–4 d. Cells were retrieved with ISOGEN (Nippongene) for RT-PCR on day 14. Other cells were fixed with 10% neutral-buffered formalin for 30 min and stained with Alizarin Red (PG Research) for 30 min on days 14 and 21. The absorbance at 450 nm of Alizarin red dye elution was determined in a microplate reader (ARVOx3, PerkinElmer) according to the manufacturer’s instructions.

Chondrogenic Differentiation and Alcian Blue Staining.

Tendon-derived cells (2 × 10⁵ cells) were suspended in Chondrogenic Incomplete Medium (Lonza) in a 15-mL conical tube. The tube was centrifuged at 18.0 × g for 3 min. After changing the medium to Chondrogenic Complete Medium (Lonza), the pellets were incubated at 37 °C. Cells were retrieved using ISOGEN (Nippongene) for RT-PCR on day 7. Other pellets were fixed on day 21 with 10% (vol/vol) neutral-buffered formalin for 60 min and stained with 100 μg/mL Alcian Blue 8 GX (Sigma) in 40% (vol/vol) acetic acid and 60% (vol/vol) ethanol overnight at 37 °C in the dark. After washing with PBS, pellets were destained with 40% (vol/vol) acetic acid and 60% (vol/vol) ethanol for 20 min.

Retrovirus Infection.

Venus, FLAG-Mkx, or HA-Mkx was inserted to the MIGR vector (Addgene). The plasmids were transfected into PLAT-E cells (50) using FugeneHD (Promega). After 24 h, the filtered supernatant was allowed to infect TDCs (P1) derived from an Mkx^−/− rat. This procedure was repeated three times every 12 h. After retrovirus infection, selection with 5 μg/mL of puromycin was performed for 5 d.

Immunocytochemistry.

For immunocytochemistry, cells were fixed with 4% (wt/vol) PFA/PBS for 15 min at room temperature. After immersion in Blocking One solution (Nacalai Tesque), they were incubated with anti-FLAG antibody (F1804, Sigma) overnight at 4 °C. The cells were then incubated with rabbit Alexa Fluor 488 donkey anti-rabbit antibody (1:500; Life Technologies) and Hoechst 33342 (1:2,000; Life Technologies) for 1 h.

Western Blot Analysis.

Cells were trypsinized, suspended in RIPA buffer (50 mM Tris⋅HCl, 150 mM NaCl, 0.5% DOC, 0.1% SDS, and 1% Nonidet P-40) and incubated at 4 °C for 30 min. After centrifugation, supernatant were analyzed. DC Protein Assay (Bio-Rad) was used for normalization of total soluble protein. After separation by SDS/PAGE, total soluble protein was transferred to a PVDF membrane. After blocking for 30 min with Blocking One solution (Nacalai Tesque), the membranes were incubated with anti-FLAG antibody (F1804, Sigma) or anti–β-actin antibody (A5316, Sigma) at 4 °C overnight and then incubated for 1 h with HRP-conjugated anti-rabbit IgG antibody (A6154, Sigma) or HRP-conjugated anti-mouse IgG antibody (A2304, Sigma). The blots were developed using Chemi-Lumi One (Nacalai Tesque).

Mechanical Stretch Stimulation.

First, 10-cm² elastic silicon rubber chambers (STB-CH-10, Strex) were coated with 0.1 mg/mL collagen type 1 (Cellmatrix type1-c, Nitta Gelatin) in 1 mM HCl overnight. After washing with PBS, cells (2 × 10⁵ in 2 mL) were seeded into the chambers 12 h before stretching. The chambers were set on a monoaxial stretching device (STB-140, Strex) and monoaxial cyclic strain (4%, 0.5 Hz) was applied for 6 h. After stretching, the mRNA of the TDCs was isolated immediately using ISOGEN (Nippongene). TDCs cultured on the chambers without stretching were used as controls.

ChIP.

After fixation with 1% formaldehyde for 10 min at room temperature, the cells were washed with cell lysis buffer [5 mM Pipes (pH 8.0), 85 mM KCl, and 0.5% Nonidet P-40] containing protease inhibitors, and resuspended in nuclear lysis buffer [1% SDS, 50 mM Tris⋅HCl (pH 8.0), and 10 mM EDTA] containing protease inhibitors. Chromatin was fragmented to 100–400 base pairs using the Covaris Model S220 (Covaris). The solution was then incubated with HA (ab9110, Abcam) or normal rabbit IgG (Santa Cruz Biotechnology) antibodies bound to Dynabeads Protein A (Dynal Biotech, Thermo Scientific) overnight at 4 °C. The beads were washed twice and the immunoprecipitates were eluted from the beads at 65 °C in the lysis buffer. Eluates were then incubated for 8 h at 65 °C to reverse the cross-linking before the addition of 0.5 mg of protease K per milliliter for 2 h at 55 °C. DNA was purified using the MinElute PCR purification kit (Qiagen). The quantities of ChIP DNA were measured using a Qubit 2.0 Fluorometer with the Qubit dsDNA HS Assay kit (Life Technologies).

ChIP-seq Library Preparation and Data Analysis.

DNA libraries for next-generation sequencing were prepared using the TruSeq ChIP Sample Preparation kit (Illumina) from 3 ng ChIP DNA or input DNA according to the manufacturer’s instructions and sequenced on an MiSeq (Illumina).

We acquired 95 bases for each paired-end sequence data, and only unique sequences on the genome were mapped to the rat (rn6) using bowtie software (bowtie-bio.sourceforge.net/index.shtml) v1.0.1. The mapped sam format files were converted to bam files using samtools (samtools.sourceforge.net/) v0.1.19 and ChIP DNA-enriched regions were detected by MACS (liulab.dfci.harvard.edu/MACS/) v1.4.2 with default parameters. The sequence data were visualized using IGV (https://www.broadinstitute.org/igv/). De novo motif discovery was performed with MEME-ChIP (24) using the default parameter.

Plasmid.

A MultiSite Gateway Three- Fragment Vector Construction Kit (Life Technologies) was used to obtain the VP16–Mkx fusion protein. To generate entry clones, 3xVP16, rat Mkx, and 3xFLAG were cloned to pDONR P4-P1r, pDONR 221, and pDONR p2R-P3 by BP clonase, respectively. Entry vectors were then cloned into the pcDNA3.1 R4R3 destination vector using LR clonase to generate pcDNA-VP16-Mkx. To generate the pGL4.12 TK-luciferase vector, the TK promoter was obtained from the phRL-TK vector (Promega) using BglII and HindIII digestion, and inserted into the pGL4.12 luciferase vector. Genomic regions of Col1a1, Col3a1, and Mkx were amplified using PCR from the genome of Wistar rats and cloned into the pGL4.12 TK-luciferase vector in the KpnI and XhoI sites. CMV promoter-driven Renilla luciferase was constructed. The primers used are listed in the lower half of Table S2. The plasmids were then sequenced to check for mutations.

Luciferase Reporter Assays.

HEK293FT cells were seeded in 96-well plates at 30% confluence and were transfected with pcDNA-VP16-Mkx or control pcDNA-Venus along with firefly and Renilla luciferase reporters using Polyethylenimine “Max” (Polysciences). Thirty-six hours after transfection, luciferase activity was measured using the Dual-Luciferase Reporter Assay System (Promega) according to the manufacturer’s instructions, and the results were normalized to Renilla luciferase activity.

Statistical Analysis.

The two-tailed independent Student’s t test was used to calculate the P values.