Electrical stimulation drives chondrogenesis of mesenchymal stem cells in the absence of exogenous growth factors

Electrical stimulation (ES) is known to guide the development and regeneration of many tissues. However, although preclinical and clinical studies have demonstrated superior effects of ES on cartilage repair, the effects of ES on chondrogenesis remain elusive. Since mesenchyme stem cells (MSCs) have high therapeutic potential for cartilage regeneration, we investigated the actions of ES during chondrogenesis of MSCs. Herein, we demonstrate for the first time that ES enhances expression levels of chondrogenic markers, such as type II collagen, aggrecan, and Sox9, and decreases type I collagen levels, thereby inducing differentiation of MSCs into hyaline chondrogenic cells without the addition of exogenous growth factors. ES also induced MSC condensation and subsequent chondrogenesis by driving Ca2+/ATP oscillations, which are known to be essential for prechondrogenic condensation. In subsequent experiments, the effects of ES on ATP oscillations and chondrogenesis were dependent on extracellular ATP signaling via P2X4 receptors, and ES induced significant increases in TGF-β1 and BMP2 expression. However, the inhibition of TGF-β signaling blocked ES-driven condensation, whereas the inhibition of BMP signaling did not, indicating that TGF-β signaling but not BMP signaling mediates ES-driven condensation. These findings may contribute to the development of electrotherapeutic strategies for cartilage repair using MSCs.


ES induces calcium/ATP oscillations and MSC condensation.
Flow cytometry analysis showed that the expanded MSCs were positive for typical MSC markers (Sca-1, CD44, CD73) but showed low expression of markers of hematopoietic stem cells (CD34), macrophages (CD11b), and granulocytes (CD45), which confirmed that the expanded MSCs exibited the characteristics of MSCs ( Supplementary Fig. S1). To determine whether ES induces ATP oscillations in MSCs, we monitored temporal changes in intracellular ATP levels using a bioluminescent ATP-dependent luciferase (Luc) reporter gene fused to a constitutive ACTIN promoter (P ACTIN -Luc). Following transfection of MSCs with P ACTIN -Luc, bioluminescence intensity was measured in real-time during ES of 0, 1, 5 or 25 V/cm at 5 Hz (Fig. 1a). In these experiments, ES of 5 V/cm induced ATP oscillations of ~ 5 min periods, whereas ES of 0, 1, or 25 V/cm did not (Fig. 1b). Because ATP oscillations were driven by changes in Ca 2+ concentrations during chondrogenesis 38 , we examined Ca 2+ oscillations using the bioluminescent Ca 2+ reporter Aequorin (AQ) gene fused to a CMV promoter (P CMV -AQ) 44 . ES of 5 V/cm consistently induced Ca 2+ oscillations, whereas ES of 0, 1, or 25 V/cm did not (Fig. 1c), suggesting that optimized ES can drive fluctuations of both Ca 2+ and ATP. We previously showed that growth factors such as TGF-β s and insulin induce prechondrogenic condensation by generating Ca 2+ /ATP oscillations 39 . Consistently, ES of 5 V/cm which induced Ca 2+ /ATP oscillations led to compact condensation of MSCs, whereas ES of 0, 1, or 25 V/cm did not (Fig. 1d). Moreover, time-course observations showed that ES of 5 V/cm induced gradual aggregation of MSCs into compact structures within 3 days, corresponding with the effects of chondrogenic medium (CM) supplemented with growth factors such as TGF-β s and insulin ( Supplementary Fig. S2). These results indicate that ES induces prechondrogenic condensation by driving Ca 2+ /ATP oscillations, even in the absence of exogenous growth factors.

ES induces MSC chondrogenesis.
In further experiments, the effects of optimized ES on chondrogenic differentiation were examined. Since ES for 3 days had little effect on cell damage (< 5%) but induced significant cell death (almost 50%) after 7 days ( Supplementary Fig. S3), ES was performed for 3 days. Gene expression of chondrogenic markers such as type II collagen (COL2A1), aggrecan (AGC), and SRY (Sex Determining Region Y)-Box 9 (SOX9) [45][46][47] was analyzed at 1 and 3-day of ES treatment and 7-day post-ES treatment. Quantitative real-time RT-PCR analyses showed that ES significantly enhanced gene expression of chondrogenic markers within 3 days of ES treatment, revealing a 66-fold increase in COL2A1, a 43-fold increase in AGC, and a 35-fold increase in SOX9 expression at 3-day of ES (Fig. 2a). Moreover, increases of chondrogenic marker expression in MSCs treated with ES for 3 days were much higher than those in the CM exposed cells, and were greater than or equal to expression levels in MCSs that were fully differentiated into chondrocytes following treatment with CM for 14 days (Fig. 2a). These data suggest that ES induces MSC chondrogenesis more effectively than CM. Moreover, it was found that the chondrogenic markers were highly expressed for as long as 7 days after the last ES treatment (Fig. 2a), which confirmed that chondrogenesis was induced in MSCs by ES. In addition, ES led to significant decreases in the expression of type I collagen (COL1; Fig. 2b). This result indicates that ES induces differentiation of MSCs into not fibrocartilaginous tissues but hyaline cartilaginous tissues [48][49][50][51] . In contrast, ES did not significantly change the expression of the osteogenic marker alkaline phosphatase (ALP), or the adipogenic marker adipocyte protein 2 (aP2; Fig. 2b Additionally, immunostaining and alcian blue staining analyses showed significantly higher expression of type II collagen and GAGs in ES-treated MSCs compared with control cells (Fig. 3a-c). Taken together, these data suggest that ES induces MSC chondrogenesis for hyaline cartilage regeneration even in the absence of exogenous growth factors. Extracellular ATP signaling via P2X 4 receptor mediates ATP oscillations, condensation, and subsequent chondrogenesis following ES. Extracellular ATP signaling via the P2X 4 receptor reportedly plays key roles in prechondrogenic condensation by mediating ATP oscillations 36 . In the present study, ES enhanced the expression of P2X 4 receptor mRNA in MSCs (Fig. 4a), suggesting that extracellular ATP signaling via the P2X 4 receptor is involved in ATP oscillations, MSC condensation, and chondrogenesis following ES. In agreement, the P2X 4 purinergic receptor inhibitor 5-BDBD inhibited ES-driven ATP oscillations (Fig. 4b). In subsequent experiments, it was examined whether extracellular ATP signaling via the P2X 4 receptor was associated with ES-driven condensation and subsequent chondrogenesis. 5-BDBD almost completely inhibited ES-driven condensation, and apyrase significantly suppressed this process (Fig. 4c). In addition, ES did not enhance the expression of the chondrogenic markers COL2A1, AGC, and SOX9 in MSCs treated with either apyrase or 5-BDBD (Fig. 4d). Hence, extracellular ATP signaling via P2X 4 receptors mediates ES-driven MSC condensation and chondrogenesis.

Intercellular communications mediates ES-driven chondrogenesis.
It was known that ES influences intercellular communications such as paracrine signaling 52 and gap junction 53 . We found that BFA, which blocks classical secretion of paracrine factors, suppressed ES-driven condensation and ES-driven increases of COL2A1, AGC, and SOX9 expression (Fig. 5a,b). In addition, the gap-junction inhibitor carbenoxolone also suppressed ES-driven condensation and ES-driven increases of expression of the chondrogenic markers (Fig. 5a,b). This result indicates that ES induces chondrogenesis by activating the release of paracrine factors and the gap-junction activity.
TGF-β signaling mediates MSC condensation and chondrogenesis following ES. TGF-β signaling reportedly induces prechondrogenic condensation and chondrogenesis through ATP oscillations 39 . The present study showed that ES led to much higher mRNA expression of TGF-β 1 (74 fold) than expression in control cells (Fig. 6a), suggesting that TGF-β signaling is involved in ES-driven MSC condensation and chondrogenesis. In agreement, inhibition of TGF-β signaling by SB-431542 almost completely blocked ES-driven condensation and significantly suppressed ES-driven increases of COL2A1, AGC, and SOX9 expression (Fig. 6b,c). Although these data indicate that ES induces MSC condensation and chondrogenesis by activating TGF-β signaling, SB-431542 did not completely suppress ES-driven induction of chondrogenic markers (Fig. 6c), suggesting that other growth factors and cytokines also mediate the actions of ES.

BMP signaling mediates ES-induced chondrogenesis, but not ES-induced condensation. BMPs
have been shown to play important roles in cartilage development 54,55 . Moreover, the present experiments showed that in comparison with non-treated controls, ES increased BMP2 expression by 42 fold (Fig. 7a). In addition, the inhibitor of BMP signaling noggin suppressed ES-driven increases in COL2A1, AGC, and SOX9 mRNA expression (Fig. 7c). However, noggin did not suppress ES-driven condensation (Fig. 7b), indicating that BMP signaling mediates ES-driven chondrogenesis, but not ES-driven condensation.

Discussion
ES is a versatile treatment that remains poorly understood in the context of stem cell-based therapy. Herein, we demonstrate that ES significantly enhances the expression of chondrogenic markers (Figs 2a and 3a,b), but significantly decreases COL1 expression in MSCs (Fig. 2b). These data indicate that ES induces MSC differentiation into hyaline chondrogenic cells, and provide evidence of the potential of electrically stimulated MSCs to efficiently regenerate hyaline cartilage in the absence of additional exogenous chemical factors.
Our previous results demonstrated that ATP oscillations driven by chondrogenic growth factors such as TGF beta and insulin play essential roles for prechondrogenic condensation that is the initial step of chondrogenesis by inducing oscillatory expression of proteins involved in actin dynamics, cell migration, and adhesion which leads to collective migration and adhesion 38,56 . The present results demonstrate that ES generates Ca 2+ /ATP oscillations in MSCs even in the absence of exogenous growth factors (Fig. 1b). Since ES directly regulates voltage-gated Ca 2+ channels 57 , ES can drive Ca 2+ oscillations by modulating voltage-gated Ca 2+ channels. In addition, since extracellular ATP signaling modulates Ca 2+ flux by producing diacylglycerol and inositol 1,4,5-triphosphate, activating protein kinase C, and by mobilizing intracellular Ca 2+ in multiple cell types 58 , ES can induces Ca 2+ oscillation by extracellular ATP signaling via the P2X 4 receptor, which is supported by the present result that P2X 4 ATP signaling mediates the actions of ES (Fig. 4). Increased Ca 2+ levels activate ATP-consuming processes such as ion pumping and exocytosis 59 , decrease glucose consumption by inhibiting glycolytic enzymes 60 , and decrease mitochondrial ATP production by abolishing mitochondrial membrane potential 61 , indicating the negative effects of Ca 2+ on ATP levels. Accordingly, Ca 2+ oscillations can drive ATP oscillations. In addition, previous studies demonstrated that pulsed electrical fields or pulsed electromagnetic fields modulate cAMP levels by activating adenosine receptors such as A 2A , A 2b , and A 3 receptors, which leads to activation of anti-inflammatory pathways and cellular proliferation in cartilage [62][63][64][65] . Our previous results showed that ATP oscillations are dependent on cAMP dynamics 38,40 . These results suggest that ES drive ATP oscillations by modulating cAMP levels.
We demonstrated that pharmacological inhibition of P2X 4 -mediated ATP oscillations suppressed ES-driven condensation (Fig. 4b,c). Previous study showed that Ca 2+ /ATP oscillations induced synchronized secretion of adhesion molecules and prechondrogenic condensation 38 . In agreement, extracellular ATP signaling reportedly mediates chemotaxis and morphological changes from spread to spherical shapes, and Ca 2+ oscillations play critical roles in cell-cell communications that lead to platelet aggregation [66][67][68] . Hence, ES may induce synchronized secretion of adhesion molecules and paracrine signaling, cell migration, and spherical morphogenesis by activating extracellular ATP signaling and Ca 2+ /ATP oscillations, leading to prechondrogenic condensation.
In the present study, ES induced chondrogenesis by stimulating both TGF-β and BMP signaling (Figs 6 and 7) 69-71 . It was known that TGF-β signaling reportedly stimulated prechondrogenic condensation by inducing the production of fibronectin and N-cadherin, and subsequently enhanced the expression of chondrogenic markers in various in vitro models 69,70 , and BMPs also promote chondrogenesis and regulate formation of cartilage elements in the limb 71 . Moreover, BMP signaling was shown to enhances TGF-β -induced chondrogenesis 72 . In addition, ES activates voltage-sensitive sodium and calcium ion channels to induce Ca 2+ influx 57 . Hence, because Ca 2+ influx activates exocytotic secretion 73 , increased Ca 2+ influx following ES may enhance secretion of TGF-β s and BMPs, likely contributing significantly to the induction of MSC chondrogenesis. These facts can explain why ES led to stronger and more rapid induction of chondrogenesis than CM supplemented with TGF-β 1 (Fig. 2a).
Many studies have shown that TGF-β signaling precedes BMP signaling and effectively initiates MSC condensation, leading to increases in the size and numbers of MSC aggregates, while BMP signaling is more effective in aggregated MSCs than in low density MSCs and increases sizes but not numbers of MSC aggregates [69][70][71]74 . We also previously demonstrated that TGF-β signaling but not BMP signaling drives ATP oscillations, leading to prechondrogenic condensation 39 . These data suggest differential effects of TGF-β and BMP signaling pathways on chondrogenesis. Consistent with these results, the present result showed that pharmacological inhibition of TGF-β signaling suppressed ES-driven condensation (Fig. 6b), whereas inhibition of BMP signaling did not (Fig. 7b), indicating that ES-driven condensation is mediated by TGF-β signaling, but is not mediated by BMP signaling. TGF-β signaling has been shown to enhance extracellular ATP levels and thus activate extracellular ATP signaling 75 . Accordingly, TGF-β signaling is stimulated by ES and then activates P2X 4 signaling to consequently induce MSC condensation, which suggests that P2X 4 signaling mediates the differential effects between TGF-β and BMP signaling on chondrogenesis.
Based upon the findings from previous studies and the present study, the actions of ES for MSC chondrogenesis could be proposed: ES drives ATP/Ca 2+ oscillations, leading to MSC condensation through TGF-β signaling and P2X 4 signaling, and subsequently induces chondrogenesis through TGF-β signaling, BMP signaling and P2X 4 signaling (Fig. 8). In summary, in this paper we demonstrate for the first time that ES drives Ca 2+ /ATP oscillations, leading to MSC chondrogenesis in the absence of exogenous cytokine or growth factor supplements, and optimized ES regimes for induction of MSC chondrogenesis. Subsequently, we showed that P2X 4 signaling mediates ES-driven ATP oscillations and chondrogenesis, and TGF-β and BMP signaling both mediates ES-driven chondrogenesis but have differential effects on ES-driven condensation. These data will facilitate the development of a novel ES-based technology for cell therapy and ES-based rehabilitation for cartilage repair. However, further studies are required to establish ES-based therapeutic strategies with the potential to overcome limitations of cartilage repair.
Lactate Dehydrogenase (LDH) Release Assays. LDH release assays were performed to assess the cytotoxicity of ES using LDH-cytotoxicity assay kits (DoGen, Korea) according to the manufacturer's instructions. After ES for 3 or 7 days, supernatants from each dish were transferred to fresh, flat bottom 96-well culture plates containing 100-μ L reaction mixtures, and were incubated for 30 min at room temperature. Formazan absorbance was then measured at 480 nm using a microplate reader (TECAN, Switzerland).

Real-time PCR analysis.
Total RNA was isolated from various MSCs cultures using the Direct-zol ™ RNA MiniPrep (Zymo Research Corporation, Irvine, CA, U.S.A.) according to the manufacturer's protocol. RNA concentrations were determined using a NanoDrop spectrophotometer (NanoDrop Technologies, Wilmington, DE, USA), and reverse transcription reactions were performed using 0.2 μ g of total RNA with a TOPscriptTM cDNA synthesis kit (enzynomics, Daejeon, Korea). The real-time PCRs for beta-actin, collagen II, and aggrecan were performed using the TOPrealTM qPCR 2X Pre MIX (enzynomics). Primer sequences are listed in Table 1.
Real-time PCRs were performed using a StepOnePlus ™ instrument (Applied Biosystems, Grand Island, NY, USA) at 95 °C for 15 min followed by 40 cycles of denaturation at 95 °C for 10 s, extension at 60 °C for 15 s, and annealing at 72 °C for 15 s. Gene expression levels were normalized to that of beta-actin and relative gene expression was calculated using the ddCT method.
Immunofluorescence staining and alcian blue staining. MSCs were fixed in 4% paraformaldehyde for 20 min at room temperature and were washed three times in phosphate buffered saline (PBS). Some samples were dehydrated through a graded ethanol series, infiltrated with xylene, embedded in paraffin, and sectioned at a thickness of 7-μ m. After blocking in PBS containing 5% goat serum and 0.3% Triton X-100 for 60 min at room temperature, cells were incubated with rabbit anti-type II collagen antibody (1:500; EnoGene Biotech, New York, NY, USA) at 4 °C overnight, were washed three times in PBS containing 0.1% Triton X-100, and were then incubated with Alexa488-conjugated secondary antibody (1:200; Invitrogen) for 60 min at room temperature in the dark. Subsequently, cells were washed three times in PBS containing 0.1% Triton X-100 and nuclei were stained with Hoechst 33258 (Dojindo, Tokyo, Japan). To visualize accumulation of sulfated glycosaminoglycans (GAGs), cells were rinsed with PBS, fixed in paraformaldehyde for 20 min, stained with Alcian Blue Solution (pH 2.5; Nacalai tesque, INC, Japan) overnight at room temperature, and were then rinsed with distilled water three times. Accumulations of glycosaminoglycans were captured using a digital camera (Olympus, Tokyo, Japan). Expression levels of type II collagen and GAGs were quantified using immunofluorescence and alcian blue intensity profiles with the NIH IMAGE J program, and data were transferred into Microsoft Excel for further analyses.

Statistical analysis.
The results are presented as means ± SD for all samples. The statistical differences between groups were analyzed by Students t-test, and multiple comparisons were performed by Fisher's protected least significant difference (PLSD) or Dunnett's test. A value of p < 0.05 was considered to indicate statistical significance.