MicroRNA-34a Modulates Cytoskeletal Dynamics through Regulating RhoA/Rac1 Cross-talk in Chondroblasts*

Background: JNK signaling involved in regulation of chondrogenic differentiation contributes modulation of miR-34a. Results: JNK signaling modulates miR-34a level and regulates stress fiber formation in chondroblasts. Conclusion: miR-34a regulates RhoA/Rac1 cross-talk and negatively modulates the actin cytoskeleton reorganization during chondrogenesis. Significance: This study provides new insights into understanding the regulatory role of miR-34a in the process of chondrogenic differentiation. MicroRNAs (miRNAs) have been implicated in various cellular processes, such as cell fate determination, cell death, and tumorigenesis. In the present study, we investigated the role of miRNA-34a (miR-34a) in the reorganization of the actin cytoskeleton, which is essential for chondrocyte differentiation. miRNA arrays to identify genes that appeared to be up-regulated or down-regulated during chondrogenesis were applied with chondrogenic progenitors treated with JNK inhibitor. PNA-based antisense oligonucleotides and miRNA precursor were used for investigation of the functional roles of miR-34a. We found that, in chick chondroprogenitors treated with JNK inhibitor, which suppresses chondrogenic differentiation, the expression levels of miR-34a and RhoA1 are up-regulated through modulation of Rac1 expression. Blockade of miR-34a via the use of PNA-based antisense oligonucleotides was associated with decreased protein expression of RhoA (a known modulator of stress fiber expression), down-regulation of stress fibers, up-regulation of Rac1, and recovery of protein level of type II collagen. miR-34a regulates RhoA/Rac1 cross-talk and negatively modulates reorganization of the actin cytoskeleton, which is one of the essential processes for establishing chondrocyte-specific morphology.

Chondrogenesis, which is a prerequisite for cartilage formation in the developing limb, involves mesenchymal cell recruitment/migration, condensation of progenitors, and chondrocytic differentiation and maturation (1,2). During development, most of our bones form through endochondral ossification, in which bones are first laid down as cartilage precursor (3,4).
One of the common features of all chondrocytes is the interdependence of cell shape and differentiation status (5)(6)(7). Chondrogenesis is characterized by drastic changes in the shape of the cell, which transitions from a fibroblastoid shape to a round or polygonal morphology (5). The molecular mechanisms responsible for this transition are largely unknown, but the reorganization of actin filaments is known to be a critical regulatory factor for chondrogenesis (6,7). Precursor cells and dedifferentiated chondrocytes are characterized by a relatively fibrillar organization among their actin filaments, whereas chondrocytes display a mostly cortical organization of actin (8,9). Several culture methods have been devised to solve the problem of phenotypic instability. Chitosan is a non-toxic, biocompatible, and biodegradable compound, has film-forming properties, and mimics the natural environment in living articular cartilage matrix. The general effect of chitosan is to maintain the round cell shape typical of chondrocytes (56). Our laboratory showed that chondrocytes exhibit rounded morphology when attached to the chitosan surface and this controlling in cell shape induces chondrogenic differentiation.
Mitogen-activated protein kinase (MAPK) cascades play essential roles in transducing extracellular signals to cytoplasmic and nuclear effectors and regulate a wide variety of cellular functions, including cell proliferation, differentiation, and stress responses. Our previous reports and others have shown that various MAPKs, including ERK and p38MAPK, regulate mesenchymal cell chondrogenesis (10 -12). During chondrogenesis of chick limb bud mesenchymal cells, p38MAPK promotes chondrogenesis by modulating the expression of cell adhesion molecules (e.g. N-cadherin, fibronectin, and integrin ␣5/␤1) at the post-precartilage condensation stages (10,12). Extensive research has defined the roles of two major MAPK signaling pathways, those mediated by extracellular signal-regulated kinase 1/2 (ERK1/2) and p38 MAPK, in the successive stages of chondrogenic differentiation. However, relatively little is known about the involvement of another MAPK signaling pathway, that mediated by c-Jun N-terminal kinase (JNK), in the regulation of cartilage formation. Recently, a few studies have suggested that the JNK signaling pathway (also known as the stress-activated pathway) is involved in the differentiation of articular chondrocytes (13,14). However, the results from these studies are contradictory. In articular chondrocytes, Wnt-3a caused dedifferentiation of chondrocytes by up-regulating c-Jun expression and the JNK-mediated phosphorylation of c-Jun, resulting in activation of the c-Jun/activator protein (14). In contrast, treatment with transforming growth factor-␤ superfamily members promoted cartilage-specific gene expression during in vitro chondrogenic differentiation of mesenchymal progenitor cells from bone marrow and trabecular bone through activation of p38, ERK1, and JNK (15). In addition, although JNK signaling appears to be involved in chondrogenic differentiation, the precise pathways and their effects have not yet been fully elucidated.
MicroRNAs (miRNAs) 3 are evolutionarily conserved small non-coding RNAs that regulate gene expression and play important roles in diverse biological functions, including cell differentiation, tumorigenesis, apoptosis, and metabolism (16 -20). For miRNA biogenesis, the miRNA-encoding genes are transcribed mainly by RNA polymerase II as long primary transcripts, which are then processed by the nuclear RNase, Drosha, to produce precursor miRNAs that are subsequently exported to the cytoplasm. The precursor miRNAs are further processed into mature miRNAs by the cytoplasmic RNase, Dicer (21). Functionally, miRNAs recognize and bind to partially complementary sites in the 3Ј-UTRs of their target mRNAs, resulting in either translational repression or degradation of the target mRNAs (22). However, although miRNAs play important roles in a wide variety of biological functions, relatively little is known regarding the regulation of miRNA expression.
Dicer, an essential component for miRNA biogenesis, is known to be involved in the regulation of chondrocyte proliferation and differentiation during skeletal development (23), suggesting a possible important role of miRNA in limb development. Recent studies have indicated that miRNAs are also important for tissue morphogenesis, and several miRNAs, including lin-4, lin-7, and miR-196, have been shown to play roles in limb development. In particular, miR-196 is thought to be involved in specifying hind limb development (24). However, it is not yet known which miRNA(s) could be the key player(s) in limb development. In this study, we show that miR-34a is a key modulator of cytoskeletal dynamics through interaction between RhoA and Rac1 during chondrogenesis.

EXPERIMENTAL PROCEDURES
Cell Culture and Treatments-Mesenchymal cells derived from the distal tips of Hamburger-Hamilton (HH) stage 22/23 embryo leg buds of fertilized White Leghorn chicken eggs were micromass cultured as described previously (10). Briefly, the cells were suspended at a density of 2 ϫ 10 7 cells/ml in Ham's F-12 medium containing 10% fetal bovine serum (FBS), 100 IU/ml penicillin and 100 g/ml streptomycin (Invitrogen). The cells were plated in three drops (15 l each) dispensed into 35-mm culture dishes or 19 drops (15 l each) into 60-mm Corning culture dishes and incubated for 1 h at 37°C under 5% CO 2 to allow attachment. The cells were maintained in 1 ml of culture medium in the absence or presence of 3 g/ml JNK inhibitor II (Calbiochem) or 10 M staurosporine (STSN; Sigma).
For dedifferentiated chondrocyte, after 7 days of culture, the cells were dissociated with 0.25% trypsin, 1.0 mM EDTA, and collagenase and replated at a density of 1 ϫ 10 5 cells/cm 2 . These cells were subcultured three more times at 2-4-day intervals. Then cells were seeded at 5000 cells/cm 2 onto uncoated or chitosan-coated 35-mm culture dishes. The culture medium was changed 24 h after cell seeding and every 48 h thereafter.
Analysis of Cell Condensation and Differentiation-Chondrogenic differentiation was measured by Alcian blue staining of sulfated cartilage glycosaminoglycans. To demonstrate the deposition of cartilage matrix proteoglycans, representative cultures were collected at day 5 of incubation and stained with 0.5% Alcian blue 8GX, pH 1.0. Alcian blue bound to sulfated glycosaminoglycans was extracted with 6 M guanidine HCl and quantified by measuring the absorbance of the extracts at 600 nm. Binding of peanut agglutinin (PA) was used as a specific marker for precartilage condensation. Briefly, cultures were rinsed twice with 0.02 M PBS, pH 7.2, fixed in methanol/acetone (1:1) for 1 min, air-dried, and then incubated with 100 g/ml biotinylated PA (Sigma) for 1 h. Bound PA was visualized using the VECTASTAIN ABC and DAB substrate solution kit (Vector Laboratories Inc., Burlingame, CA).
Cell Viability Assay-Cell viability was assayed using CellTiter-Glo luminescent cell viability assay kit (Promega) based on the quantitation of ATP present in metabolically active cells or viable cells (15). An equal volume of reconstituted CellTiter-Glo reagent was added to 100 ml of cell suspension. The contents were mixed on an orbital shaker for 2 min to induce cell lysis and incubated at room temperature for 10 min to stabilize the luminescent signal, and the luminescence was recorded (15).
Western Blot Analysis-Cells were lysed in buffer (50 mM Tris-HCl, pH 7.4, containing 150 mM NaCl, 1% Nonidet P-40, 0.1% SDS, 0.1% deoxycholic acid, 10 mM NaF, 10 mM Na 4 P 2 O 7 , 0.4 mM Na 3 VO 4 , and protease inhibitors) for 30 min on ice. For Western blot analysis using conditioned medium, chick leg bud mesenchymal cells were incubated in serum-free medium for 24 h after drug treatment. Total protein content of the cells was determined by the bicinchoninic acid method (Pierce). Loading amounts were standardized to the middle protein concentration for all samples. Then the conditioned medium was concentrated with 10% trichloroacetic acid (TCA) precipitation. The precipitate was dissolved in 1ϫ sample buffer (62.5 mM Tris-HCl, pH 6.8, containing 10% glycerol, 2% SDS, and 0.005% bromphenol blue). Proteins (30 g) were separated by 10% polyacrylamide gel electrophoresis containing 0.1% SDS and transferred to nitrocellulose membrane (Schleicher and Schuell). The membranes were incubated for 1 h at room temperature in blocking buffer (20 mM Tris-HCl, 137 mM NaCl, pH 8.0, containing 0.1% Tween and 3% nonfat dry milk) and probed with antibodies against JNK and phospho-JNK (Cell Signaling Technology, Beverly, MA); type II collagen (Chemicon, Temecula, CA); RhoA, Rac1, and GAPDH (Santa Cruz Biotechnology, Inc., Santa Cruz, CA); type I collagen, N-cadherin, and Sox-9 (Calbiochem); and HSP70 (Stressgen, San Diego, CA). The blots were developed with a peroxidase-conjugated secondary antibody, and reacted proteins were visualized using the electrochemiluminescence (ECL) system (Pierce).
Immunocytochemistry-Cells grown on coverslips were washed three times with phosphate-buffered saline (PBS) and then fixed and permeabilized. For actin staining, each culture was stained with Alexa488-phalloidin (Molecular Probes, Inc., Eugene, OR) prepared in PBS containing 1% (v/v) bovine serum albumin for 1 h at room temperature in a light-proof box. Cultures were then washed three times with water and mounted with Gel/Mount (Biomedia, Foster City, CA).
Chitosan Coating-One gram of chitosan powder (Sigma) was dissolved in 2% (v/v) acetic acid solution to give a final concentration of 0.1% (w/v). The solution was stirred overnight and autoclaved for 10 min until the material was completely dissolved. Two milliliters of sterile chitosan solution was used to evenly coat the surface of each 35-mm culture dish and then evaporated to dryness for at least 24 h at room temperature in a laminar clean bench. The coated dishes were neutralized by the addition of 1 ml of 0.5 M NaOH followed by several washings with Hanks' balanced salt solution.
RNA Preparation, miRNA Microarray Hybridization-Total RNA was isolated from micromass-cultured limb mesenchymal cells at 24 h in the absence or presence of JNK inhibitor II using the mirVana miRNA isolation kit (Ambion). For the miRNA microarray, RNA samples were processed and hybridized to PANArra miRNA (Panagene), containing 200 miRNA probes. Briefly, the total RNA (400 ng) denaturation mixture was incubated at 95°C for 5 min and hybridized at 55°C for 4 h in hybridization mixture. After washing the microarray slides, ligation mixture was added and incubated at 37°C for 2 h, and the microarray slide was scanned with a GenePix4000B array scanner (Axon Instruments). RNU6B was used as an endogenous control (internal control), and data were normalized with the expression level of RNU6B. Data were analyzed using analysis of variance with a multiple comparison corrected p value less than 0.005 and a false discovery rate set to 0.05. miRNA and mRNA Real-time Quantitative RT-PCR-miRNA and mRNA expression was independently quantified using the TaqMan MicroRNA and TaqMan gene expression assays, respectively (Applied Biosystems) according to the manufacturer's protocols. miRNA expression was normalized to RNU43 small nuclear RNA endogenous controls. For assessment of mRNA, transcripts were quantified by real-time quantitative polymerase chain reaction (RT-PCR) and normalized with respect to gapdh expression. The oligonucleotides used as primers were as follows: RhoA, 5Ј-TGGGATACAGCAGGACAGGA-3Ј (antisense) and 5Ј-AAAACCTCCCTCACACCGTC-3Ј (sense); Rac1, 5Ј-ATGACAGACTACGCCCACTC-3Ј (antisense) and 5Ј-CCA-CTTAGCACGGACATTTT-3Ј (sense); gapdh, 5Ј-GATGGG TGTCAACCATGAGAAA-3 (antisense) and 5Ј-ATCAAAGGT GGAAGAATG GCTG-3Ј (sense).
Induction of PNA-based miRNA Inhibitor or miR-34a Precursor Oligonucleotides-The PNA-based antisense oligonucleotides (ASOs) contained an O-linker at the N terminus of the PNA to improve solubility were purchased from Panagene. 200 nM PNA-based ASO (PNA34a, UGGCAGUGUCUUAGCUG-GUUGU) or 50 nM miR-34a precursor oligonucleotides (UGGCAGUGUCUUAGCUGGUUGU; Ambion) were electroporated into isolated mesenchymal cells using a square wave generator (BTX-830, Gentronics (San Diego, CA)) with 20-ms, 200 square pulses. A scrambled PNA-based ASO or scrambled miRNA was used as a negative control.
Reporter Vectors and DNA Constructs-The 3Ј-UTR of Rac1 (2341 bp) was PCR-amplified using primers 5Ј-GCACGT-GTTCCCGACATAAC3Ј and 5Ј-GGCGAGGTAAAGAGGGT-CAG-3Ј (nucleotides 514 -1259) and cloned downstream of the CMV-driven firefly luciferase cassette in the pMIR-Report vector (Ambion). A reporter vector containing a directly matched miRNA-binding site oligonucleotide (ϳ51 bp) for miR-34a was used as the positive control. For miRNA target validation, limb mesenchymal cells were electroporated using a square wave generator (20-ms, 200 square pulses; BTX-830, Gentronics (San Diego, CA)) with 25-50 ng of each firefly luciferase reporter construct, 150 -175 ng of empty pcDNA3 vector, 200 ng of pcDNA3 harboring the Renilla luciferase gene (transfection control), and 30 pmol of miR-34a precursor or negative control for miR-34a (Ambion). At 24 h post-transfection, the activities of firefly and Renilla luciferase were assayed (Promega, San Luis Obispo, CA). Normalized relative light units were used to represent the ratios of firefly luciferase activity versus Renilla luciferase activity.
Statistical Analysis-Statistical analysis was performed using the SPSS program for Windows, Standard Version (version 18.0, SPSS Inc., Chicago, IL). The significance level was determined by a nonparametric Mann-Whitney U test, a Kruskal-Wallis Test, or a t test. A value of p Ͻ 0.05 was regarded as statistically significant.

A JNK Inhibitor Suppresses Chondrogenesis of Chick Limb Mesenchymal Cells and Stimulates miR-34a Expression-Pre-
viously, our laboratory demonstrated the involvement of JNK signaling in chondrogenesis (57). That is, the phosphorylation of JNK was increased during the early stages of normal chondrogenic differentiation. Precartilage condensation and chondrogenic differentiation were analyzed by PA and Alcian blue staining, respectively. The PA and Alcian blue staining intensities (Fig. 1A); transcription of type II collagen protein, a marker for chondrogenic differentiation; transcription of the two major proteoglycans, aggrecan and versican; and the translational level of type II collagen and condensation-related mole-cules, such as Sox-9 and N-cadherin (Fig. 1B), were decreased in JNK inhibitor-treated cells versus untreated controls and were suppressed by blockage of JNK signaling.
miRNAs are known to modulate a variety of cellular pathways related to cell development and differentiation (16 -18, 25). To identify miRNAs that could be differentially expressed in JNK inhibitor-treated cells versus untreated controls, we performed a miRNA array screening with 200 unique miRNAs in their mature forms. From this array, we identified 14 up-regulated miRNAs and 12 down-regulated miRNAs following treatment with a JNK inhibitor, and we choose highly conserved miRNAs between human and chicken by comparing their mature sequence (supplemental Table 1). Among them, miR-34a was substantially induced by JNK signaling (determined using a p value of 0.01 as a cut-off for significance). During in vitro culture, the induction of miR-34a was decreased with the progression of chondrogenic differentiation (Fig. 1C), suggesting that miR-34a plays a role in a later stage of chondrogenic differentiation, such as precartilage condensation and chondrocyte commitment, possibly not in proliferation of chondrogenesis in this study. Treatment with the JNK inhibitor significantly increased the expression level of miR-34a as confirmed by RT-PCR (Fig. 1D). In order to examine the possible involvement of miR-34a in chondrogenic differentiation of chick limb mesenchymal cells, we exposed mesenchymal cells to 200 nM PNA-based ASOs against miR-34a, which confirmed a signifi-cant decrease in the expression level of miR-34a by RT-PCR (Fig. 1D); treated the cells with or without the JNK inhibitor; and assessed chondrogenesis by Alcian blue staining on day 5 post-treatment and by expression level of type II collagen (Fig.  1E). The decreased intensities and expression of PA and Alcian blue staining and type II collagen by JNK inhibitor were recovered by co-treatment with anti-miR-34a oligonucleotides (Fig.  1E). Conversely, overexpression of the miR-34a precursor oligonucleotides leads to a suppression of precartilage condensation ( Fig. 2A, top) and chondrogenic differentiation ( Fig. 2A, bottom) without affecting cell death (Fig. 2B).
The cytoskeleton plays fundamental roles in cell survival, migration, division, and chromosome and organelle movement. Cytochalasin, a cytoskeleton-interrupting reagent, has varying effects on different types of cells, suggesting that the cytoskeleton can have critical yet cell type-dependent functions (26,27). To examine the possible interaction between miR-34a and cytoskeletal reorganization, we used phalloidin staining to observe actin distribution in chick limb mesenchymal cells (Fig.  2C). Actin stress fibers were intensified in JNK inhibitor-treated cells versus untreated controls, but this effect was at least partially rescued by co-treatment with the anti-miR-34a oligonucleotides.

miR-34a Modulates Rearrangement of Actin Cytoskeleton by Regulating RhoA during Chondrogenic Differentiation of Chick
Limb Mesenchymal Cells-Rho GTPases, a family of small GTP-binding proteins, is known to be involved in cell cytoskel-

FIGURE 1. miR-34 is involved in chondrogenic differentiation of limb bud mesenchymal cells. A, chondroprogenitor cells were treated with or without 5
M JNK inhibitor and stained with PA at day 3 of culture or with Alcian blue at day 5 of culture. B, change in RNA levels of aggrecan, versican, type I collagen, and type II collagen (left) and in protein level of type II collagen, Sox-9, and N-cadherin was determined by Western blotting or PCR analysis. C, the induction of miR-34a was measured during in vitro culture periods. D, chondroprogenitor cells were treated with 5 M JNK inhibitor or 100 nM anti-miR-34a oligonucleotides (anti-miR-34a). Cells without treatment of anti-miR-34a (Con, Ϫ, JNKinhibitor) were treated with scrambled PNA-based ASOs. The expression of miR-34a was measured with real-time PCR. E, cells were stained with Alcian blue at day 5 of culture, and chondrogenesis was quantified by measuring the absorbance of bound Alcian blue at 600 nm (top). Change in protein level of type II collagen was determined by Western blotting at the indicated culture days (bottom). The data shown are representative of at least four independent experiments. The mean is plotted, and the error bars represent 95% confidence interval (lower/ upper limit). *, statistically different from control cells (p Ͻ 0.005). The diameter of typical standard culture is 5 mm. eton organization. Animals have three Rho isoforms, RhoA, RhoB, and RhoC. RhoA and RhoC play key roles in the regulation of actomyosin contractility and in cell locomotion (28). Regulation of RhoA activity has been shown to influence chondrogenesis. Overexpression of RhoA in ATDC5 cells inhibits glycosaminoglycan synthesis (29), and repressing RhoA signaling enhances chondrogenesis in micromass cultures of limb bud mesenchymal cells (29). Therefore, we tested that the modulation of a stress fiber formation by JNK inhibition or miR-34a inhibition is due to alteration of RhoA level. JNK inhibition was up-regulated at the RhoA protein level (Fig. 3A). Interestingly, RhoA was down-regulated in miR-34a-knockdown cells and up-regulated in miR-34a-overexpressed cells compared with controls (Fig. 3B). Overexpression of wild-type RhoA in chick limb mesenchymal cell induced stress fiber formation. The level of RhoA after electroporation was confirmed by Western blotting with anti-HA antibody (Fig. 3C, left). Furthermore, cotreatment of anti-miR-34a oligonucleotides reduced stress fiber formation induced by RhoA, whereas co-treatment of miR-34a precursor oligonucleotides resulted in the most significant induction of stress fibers (Fig. 3C, right). Finally, chondrogenesis in limb bud micromass cultures similarly entailed a loss of RhoA protein and stress fibers (Fig. 3D).
A recent study showed that dedifferentiated chondrocytes cultured in alginate gel experience losses of RhoA protein and stress fibers concomitant with chondrocyte differentiation (30). STSN is known to stimulate chondrogenic differentiation and to inhibit stress fiber formation (31) and cartilage differentiation (32). Treatment with STSN reduced the level of miR-34a and RhoA in chick limb mesenchymal cell culture at day 4 (Fig. 3E). Co-treatment of anti-miR-34a oli-gonucleotides reduced stress fiber formation induced by RhoA or miR-34a precursor oligonucleotide (Fig. 3F). Consistent with this observation, chondrogenesis in limb bud micromass cultures similarly entailed a loss of RhoA protein and stress fibers (Fig. 3G).
Chondrocyte differentiation, which is characterized in part by rounded cell morphology, is closely associated with chondrogenesis. To see whether miRNA is involved in a redifferentiation program because miR-34a seems to be involved in the modulation of actin cytoskeletal reorganization by targeting RhoA, differentiated chondrocytes obtained from day 7 micromass cultures of chick limb mesenchymal cells were subcultured to monolayers and passaged every 2-3 days until passage 4. During this time, we monitored the progressive loss of differentiated phenotypes. The passage 4 cells were then transferred to chitosan membrane, a partially deacetylated product of chitin that has film-forming properties, and their redifferentiation was examined. The presence of chitosan maintained the rounded cell shape typical of chondrocytes (Fig. 4A), and decreased levels of miR-34a (Fig. 4B, left) and RhoA (Fig. 4B,  right), and restored type II collagen expression (Fig. 4B, right). Interestingly, inhibition of miR-34a induced the cells to morphologically transition to cortical actin (Fig. 4A) and restored type II collagen expression (Fig. 4B), indicating that redifferentiation had occurred. Similar to cells cultured on chitosan membrane, the reduction of RhoA protein level was also observed in cells treated with anti-miR-34a oligonucleotides (Fig. 4).
miR-34a Is Involved in Rac1-RhoA Cross-talk in Chondrogenic Progenitors-Previously, Rac-Rho cross-talk has been suggested in several studies (33,34). This coordinated and FIGURE 2. miR-34a alters stress fiber formation in limb mesenchymal cells. A, cells were cultured at a density of 2 ϫ 10 7 cells/ml with or without 50 nM miR-34a precursors (mir-34a). Scrambled PNA-based ASOs were used as control (Con). Cells were stained with PA at day 3 of culture (top, a-d; c and d are high magnification images of a and b, respectively) and Alcian blue at day 5 of culture (bottom, a-d; c and d are high magnification images of a and b, respectively), and chondrogenesis was quantified by measuring the absorbance of bound Alcian blue at 600 nm (right). The mean is plotted, and the error bars represent 95% confidence interval (lower/upper limit). *, statistically different from control cells (p Ͻ 0.005). B, total cell numbers were counted, and cell viability was analyzed using the CellTiter-Gloா luminescent cell viability assay at 2 days of culture. C, cells were treated with 50 nM miR-34a precursors or 5 M JNK inhibitor and stained with Alexa488-conjugated phalloidin. Cells without treatment of anti-miR-34a (Con, JNKinhibitor) were treated with scrambled PNA-based ASOs. The data shown are representative of at least four independent experiments. The diameter of typical standard culture is 5 mm. Dedifferentiated chondrocytes were cultured on chitosan membrane and treated with or without anti-miR-34a oligonucleotides. Dedifferentiated chondrocytes were cultured on chitosan membrane or were treated with anti-miR-34a. Scrambled PNA-based ASOs were used as a control (Con). A, cells were stained with Alexa488-conjugated phalloidin. B, the expression of miR-34a was measured with real-time PCR (left), and change in protein levels of type II collagen and RhoA was examined by Western blotting (right). The data shown are representative of at least four independent experiments. The mean is plotted, and the error bars represent 95% confidence interval (lower/upper limit). *, statistically different from control cells (p Ͻ 0.005). Change in protein level of RhoA was determined by Western blotting at 2 days of culture. B, cells were treated with 100 nM anti-miR-34a oligonucleotides (anti-mir-34a) or 50 nM miR-34a precursors (mir-34a). Scrambled PNA-based ASOs were used as control (RRRQRRKKR-OO-ATTAATGTCGGACAA). Change in protein level of RhoA was determined by Western blotting at 2 days of culture. Cells were stained with Alexa488-conjugated phalloidin. Cells were electroporated with wild type RhoA expression vector (RhoA) and cultured with 100 nM anti-miR-34a oligonucleotides (anti-mir-34a) or 50 nM miR-34a precursors (mir-34a). C, wild-type RhoA was introduced into cells by electroporation, and the level of RhoA after electroporation was confirmed by Western blotting with anti-HA antibody (top). Cells without treatment of anti-miR-34a or miR-34a precursor (Con, Ϫ, RhoA) were treated with scrambled PNA-based ASOs (bottom). D, cells were stained with Alcian blue at day 5 of culture, and chondrogenesis was quantified by measuring the absorbance of bound Alcian blue at 600 nm. E, cells were cultured with or without 10 M STSN. The expression of miR-34a was measured with real-time PCR (left), and RNA level of RhoA was measured by reverse transcription PCR (right). GAPDH was used as a loading control. Cells were electroporated with wild type RhoA expression vector (RhoA) and cultured with or without 10 M STSN. Cells without treatment of anti-miR-34a or miR-34a (Con, STSN, STSN/RhoA) were treated with scrambled PNA-based ASOs. F, cells were stained with Alexa488-conjugated phalloidin. G, cells were stained with Alcian blue at day 5 of culture, and chondrogenesis was quantified by measuring the absorbance of bound Alcian blue at 600 nm. The data shown are representative of at least four independent experiments. The mean is plotted, and the error bars represent 95% confidence interval (lower/upper limit). *, statistically different from control cells (p Ͻ 0.005). The diameter of standard culture is 5 mm.
opposed activity of Rac1 and RhoA is known to be crucial to cellular dynamics, the former promoting membrane protrusion and cell polarity and spreading and the second promoting cytoskeleton contractility and tail retraction. Therefore, we examined whether JNK signaling is involved in regulation of Rac1. Treatment of JNK inhibitor blocked the translational level of Rac1 (Fig. 5A). To verify whether Rac1 is one of targets for miR-34a, we checked the expression level of Rac1 by anti-miR-34a or miR-34a precursor oligonucleotides (Fig. 5B). We observed a significant reduction in Rac1 protein expression among miR-34a precursor-treated chondroprogenitor cells, whereas these levels were recovered in cells that had been treated with anti-miR-34a oligonucleotides. To further confirm that Rac1 is a target of miR-34a, we cloned the entire 3Ј-UTR of Rac1 into a luciferase reporter vector, electroporated the vector into chondrogenic progenitor cells along with miR-34a precursor or a cognate non-targeting negative control, and assayed cell lysates for luciferase expression. We found that cells transfected with the Rac1 3Ј-UTR-driven vector plus miR-34a precursor exhibited significantly less luciferase activity compared with cells that received the reporter plus the non-targeting negative control (Fig. 5C and supplemental Figs. 1 and 2), suggesting that miR-34a targets Rac1 and represses its expression during JNK-mediated signaling.
To confirm the correlation among Rac1, RhoA, and miR-34a during chondrogenic differentiation, the expression level of Rac1 and RhoA was examined. The expression of Rac1 was increased with the progression of chondrogenic differentiation (Fig. 5D) along with a decrease in the induction of miR-34a ( Fig.   1D and supplemental Fig. 3). However, the induction of RhoA was decreased with the progression of chondrogenic differentiation.

DISCUSSION
miRNAs play key roles in diverse regulatory pathways, including development (25)(26)(27)(28)(29)(30)35), cell proliferation, differentiation (16 -18), apoptosis (19,20), tumorigenesis (36,37), and many other physiological or pathological processes (38,39). Recent studies have indicated that microRNAs are also important for tissue morphogenesis (40,41). However, the precise roles of miRNAs in cartilage biology are largely unknown, and many of the individual miRNA targets that are likely to be important within chondroblasts remain yet unidentified. Here, we characterized one specific miRNA involved in the morphological transition of chondroprogenitors from fibrillar organization to cortical organization, which is a necessary step for chondrogenic differentiation (42,43). More specifically, we showed that miR-34a was up-regulated by chondro-inhibition, whereas miR-34a knockdown stimulated the transition of cells from a fibroblastoid morphology to a rounded or polygonal morphology. In addition, we found that this occurred at least in part through regulating RhoA, which is known to modulate the organization of stress fibers and focal adhesions (30,44). Previously, miR-143 and miR-145 were shown to act as integral components of the network that regulates cytoskeletal remodeling and phenotypic switching of smooth muscle cells during vascular disease (45). The miR-200 family of miRNAs has been shown to directly target WAVE3 to regulate remodeling of the FIGURE 5. miR-34a regulates RhoA/Rac1 cross-talk. A, chondroprogenitor cells were treated with 5 M JNK inhibitor, and change in protein level of Rac1 was determined by Western blotting at day 3 of culture. B, cells were treated with 100 nM anti-miR-34a oligonucleotides (anti-mir-34a) or 50 nM miR-34a precursors (mir-34a). Scrambled PNA-based ASOs were used as a control (Con). Change in protein level of Rac1 was determined by Western blotting at day 3 of culture. C, luciferase reporter assays of cells harboring a vector construct driven by the human Rac1 3Ј-UTR, with or without forced expression of miR-34a. D, the induction of Rac1 and RhoA was measured during in vitro culture periods. The data shown are representative of at least four independent experiments. The mean is plotted, and the error bars represent 95% confidence interval (lower/upper limit). *, statistically different from control cells (p Ͻ 0.005).
actin cytoskeleton and the epithelial-mesenchymal transition in cancer cells (46). Previously, miR-34 had been shown to help regulate the cell cycle and cell growth by targeting proteins, such as Cdk, cyclin E2, c-Met, and Bcl-2 (47,48). miR-34a inhibits the invasiveness of cervical carcinoma and choriocarcinoma cells through targeting Notch1 and Jagged1 (49) and migration of hepatocellular carcinoma cells by down-regulation of c-Met (50). During human monocyte-derived dendritic cell differentiation, miR-34a targets WNT1 and JAG1, factors involved in stalling monocyte-derived dendritic cell differentiation (51).
Adding to these previous findings, we herein show for the first time that miR-34a regulates rearrangement of the actin cytoskeleton by regulating RhoA/Rac1 cross-talk during the chondrogenic differentiation of chick limb mesenchymal cells. Actin filaments are associated with the plasma membrane at sites where the cells form connections with the extracellular matrix or other cells. Well spread cells exert tension on the adhering extracellular matrix via fully extended stress fibers (52,53). Stress fibers directly terminate at focal adhesions, amid accumulations of numerous structural proteins that are responsible for connecting the cell membrane to the underlying substrates. Chondrogenesis is characterized by a drastic change in cell shape, which transitions from a fibroblastoid shape to a round or polygonal morphology. We previously showed that activation of matrix metalloprotease-2 (MMP-2) is responsible for reorganization of the actin cytoskeleton to a cortical pattern, with cell rounding occurring in parallel (10). We also previously showed that dedifferentiated chondrocytes are capable of regaining their differentiated chondrocytic phenotypes on a chitosan membrane and that this change is associated with increased expression of type II collagen protein, which is a typical differentiation marker for chondrogenesis (54). However, the regulatory molecules involved in this sort of morphological transition are largely unknown. Here, we report that actin filaments in undifferentiated mesenchymal cells or cells treated with a JNK inhibitor may be seen bundled as stress fibers. However, when these cells were treated with anti-miR-34a oligonucleotides, they regained a rounded shape and re-expressed sulfated proteoglycans, suggesting that miR-34 may trigger the disassembly of the actin stress fibers, followed by their replacement with cortically distributed microfilaments. Rho proteins are small GTPases that are involved in modulating the organization of stress fibers and focal adhesions (30,34,43,44). Blocking the Rho signaling pathway leads to enhanced chondrogenesis among mesenchymal cells, suggesting that Rho plays a negative role in chondrocytic differentiation (29). Rac1 signaling is a required signaling pathway to generate normal N-cadherin-dependent cellular junctions that are known to be essential for chondrogenesis (55). Culture of dedifferentiated chondrocytes in alginate gel has been shown to be associated with decreased RhoA protein levels and a loss of stress fibers, concomitant with the re-expression of the chondrocytic differentiation program. A mutual antagonism between the Rac and Rho GTPases has been observed in several cellular settings, suggesting reciprocal control of Rac and Rho small GTPases. Coordinated and opposed activity of Rac1 and RhoA is crucial to cellular dynamics, the former promoting membrane protru-sion, cell polarity, and spreading (33,34). Here, we report for the first time that the forced overexpression of miR-34a in chick limb mesenchymal cells can induce intensified stress fibers, whereas blockade of miR-34a induced morphological transition to chondrocyte-specific morphology through modulation of RhoA/Rac1 cross-talk. Based on our findings, it seems possible that intensified stress fibers seen in JNK inhibitor-treated cells may be due to suppression of miR-34-targeted Rac1 expression. Our data suggest that miR-34a regulates RhoA/ Rac1 cross-talk and negatively modulates reorganization of the actin cytoskeleton, which is one of the essential processes for establishing chondrocyte-specific morphology.