Local Calcium Elevation and Cell Elongation Initiate Guided Motility in Electrically Stimulated Osteoblast-Like Cells

Background Investigation of the mechanisms of guided cell migration can contribute to our understanding of many crucial biological processes, such as development and regeneration. Endogenous and exogenous direct current electric fields (dcEF) are known to induce directional cell migration, however the initial cellular responses to electrical stimulation are poorly understood. Ion fluxes, besides regulating intracellular homeostasis, have been implicated in many biological events, including regeneration. Therefore understanding intracellular ion kinetics during EF-directed cell migration can provide useful information for development and regeneration. Methodology/Principal Findings We analyzed the initial events during migration of two osteogenic cell types, rat calvarial and human SaOS-2 cells, exposed to strong (10–15 V/cm) and weak (≤5 V/cm) dcEFs. Cell elongation and perpendicular orientation to the EF vector occurred in a time- and voltage-dependent manner. Calvarial osteoblasts migrated to the cathode as they formed new filopodia or lamellipodia and reorganized their cytoskeleton on the cathodal side. SaOS-2 cells showed similar responses except towards the anode. Strong dcEFs triggered a rapid increase in intracellular calcium levels, whereas a steady state level of intracellular calcium was observed in weaker fields. Interestingly, we found that dcEF-induced intracellular calcium elevation was initiated with a local rise on opposite sides in calvarial and SaOS-2 cells, which may explain their preferred directionality. In calcium-free conditions, dcEFs induced neither intracellular calcium elevation nor directed migration, indicating an important role for calcium ions. Blocking studies using cadmium chloride revealed that voltage-gated calcium channels (VGCCs) are involved in dcEF-induced intracellular calcium elevation. Conclusion/Significance Taken together, these data form a time scale of the morphological and physiological rearrangements underlying EF-guided migration of osteoblast-like cell types and reveal a requirement for calcium in these reactions. We show for the first time here that dcEFs trigger different patterns of intracellular calcium elevation and positional shifting in osteogenic cell types that migrate in opposite directions.


Introduction
Guided cell migration is essential for embryonic development, tissue formation, inflammation and wound-healing [1][2][3]. In addition to chemical (chemotaxis) or mechanical (contact) mechanisms, guided cell migration is influenced by electric potential (electrotaxis) [4][5][6][7]. Exogenous dcEFs of physiological strength induce a variety of cellular responses [8][9][10]. Although the general effects of electrical stimulation on various cell types are well known, the exact cascades translating exogenous and endogenous electrical signals into a variety of intracellular responses are still poorly understood. One important candidate for this translation is calcium. Calcium is involved in maintaining cell physiology and is a pluripotent signaling molecule that holds a crucial place in many cell biological pathways [11][12][13]. In particular, elevation of cytoplasmic calcium serves as a rapid response to various factors including electrical stimulation. Thus, regulation of intracellular calcium ([Ca 2+ ] i ) levels via non-invasive electrical stimulation may be important for controlling cellular responses during migration [14]. To address this issue here, we observe cell morphological rearrangements and [Ca 2+ ] i dynamics in response to dcEFs.
In this study we use two osteogenic cell types, rat calvarial and human SaOS-2 cells [15,16]. These are good models for directional migration studies since they prefer opposite movement directions in response to electrical stimulation. Immunofluorescence, vital staining, differential interference contrast (DIC) microscopy and time-lapse video microscopy techniques were used to analyze the short-and long-term cellular responses to electrostimulation. We offer novel observations of the initial cellular response to electrical stimulation, in terms of cytoskeletal reorganization and ion fluxes, and suggest an early role for intracellular calcium in directed cell migration.

DcEF-induced elongation and reorientation are time-and voltage-dependent
We first observed the cell morphological changes induced by exposure to low and high voltage dcEFs for different durations. The field strengths used are consistent with those applied in similar studies on osteoblast cells [17][18][19]. After 5 h of exposure to a 5 V/ cm dcEF, most of the calvarial and SaOS-2 cells elongated and orientated perpendicular to the field vector ( Fig. 1E and F), in contrast to the control cells (Fig. 1A, B, C and D). Overall, the percentage of perpendicular-orientated calvarial osteoblasts significantly increased with the duration of dcEF stimulation, from 560.8% (Mean6standard deviation, start) to 5662.2% (Mean6 standard deviation, 5 h, Fig. 1G). The lower percentage of orientation to start was due to some cells remained round, did not have a significantly elongated axis and hence, were difficult to score. Increasing numbers of perpendicular-orientated SaOS-2 cells were also seen with longer durations of dcEF stimulation (Fig. 1H), but this effect became significant only after 3 h of exposure. In both cell types, the change in cell orientation became significant in a dcEF of 2 V/cm (2162%, Mean6standard deviation) and highly increased at 5 V/cm (5261.6%, Mean6 standard deviation) ( Fig. 1I and J). Overall, reorientation of the cells increased proportionally with both duration ( Fig. 1G and H, n cell = 1200-1500.) and strength ( Fig. 1I and J, n cell = 1200-1500) of the dcEF. This experiment was done in the presence of Ca 2+ .

DcEF-induced directional migration is Ca 2+ -dependent
To determine if Ca 2+ is important for dcEF-induced directed migration, we compared cell behavior in the presence and absence of Ca 2+ . In the presence of Ca 2+ , the two osteogenic cell types migrated in opposite directions when exposed to a dcEF, calvarial cells to the cathode (Fig. 2C, Movie S1, Fig. S1) and SaOS-2 cells to the anode (Fig. 2D, Movie S2, Fig. S2). Unexposed cells moved randomly ( Fig. 2A and B). Without Ca 2+ , both SaOS-2 ( Fig. 2D') and calvarial (Fig. 2C') osteoblasts did not show a preferred direction, similar to the control cells ( Fig. 2A' and B'). The data showing the values of directedness in degree has given in the supporting information (Dataset S1). The presence of Ca 2+ increased the net linear displacement from 6.08 mm to 12.42 mm for SaOS-2 cells and from 10.86 mm to 14.72 mm for calvarial cells exposed to a 5 V/cm dcEF for 5 h (Fig. 2E). Net linear displacement was defined by the starting and end points connected by a straight line. Migration speeds calculated from the original trajectories were 20 mm/hr (calvarial) and 7.3 mm/hr (SaOS-2) for unexposed cells, and 32.1 mm/hr (calvarial) and 15.5 mm/hr (SaOS-2) for dcEF-exposed cells (Fig. 2F). DcEFs triggered an increase in distance and speed of calvarial osteoblast movement by 3663% and 6165.2%, while SaOS-2 cells increased their distance and speed by 10468% and 111610.5% respectively. Without Ca 2+ , dcEFs did not cause distinguishable differences in migration speed and distance compared to the control cells ( Fig. 2E' and F'). Mean values were evaluated from at least 3 individual experiments for each case (n cells = 17-22). In total, this data suggests that dcEFs guide osteoblast migration via a Ca 2+dependent mechanism.
DcEF-induced redistribution of adhesion structures, elongation and de novo formation of membrane protrusions is Ca 2+ -dependent To further understand the Ca 2+ -dependent mechanism by which dcEFs direct migration, we looked at indicators of morphological rearrangement. Control cells ( Fig. 3A and B) displayed a relatively homogenous distribution of focal adhesions, as marked by vinculin, and randomly arranged actin filaments. In the presence of Ca 2+ , dcEF-exposed SaOS-2 and calvarial cells had concentrations of vinculin spots near the cell edges or in the direction of movement ( Fig. 3C and D). Parallel organized filamentous actin and new membrane protrusions (lamellipodia or filopodia) were also more noticeable at the leading edge of both cell types after dcEF exposure ( Fig. 3C-F). When Ca 2+ ions were depleted from the media, none of these changes were observed ( Fig. 3A'-F'). This indicates that Ca 2+ is required for the

Strong dcEFs elevate [Ca 2+ ] i by involving VGCCs
To address whether dcEF-induced cell migration involves regulating intracellular calcium levels, we monitored [Ca 2+ ] i dynamics in response to different strength dcEFs. After the onset of exposure to a 14 V/cm dcEF in the presence of Ca 2+ , calvarial osteoblasts (

DcEFs induce cells to initially contract and relocate opposite to migration direction
We noticed that Calvarial (n = 17) and SaOS-2 (n = 21) cells ( Fig. 6A and D) underwent a significant contraction within 1 minute of exposure to dcEF (Figs. 6B and E). Here, dcEF strength of 10 V/cm was chosen as representative as this effect was also observed with other strengths of dcEF (1-14 V/cm). Calvarial osteoblasts contracted towards the anode (Fig. 6C, Movie S5) and SaOS-2 cells towards the cathode (Fig. 6F, Movie S6). Interestingly, in both cell types the relocation (positional shift) caused by the contraction was always opposite to the preferred migratory direction. This suggests that there is a refractory period during which the cells mechanically resist dcEF forces.

Discussion
Here we describe the initial events during dcEF-induced directional cell migration in osteoblasts and show that Ca 2+ is essential (Fig. 7). This is in contrast to other cell types reported to be regulated independent of Ca 2+ , such as NIH 3T3 and SV101 murine fibroblastic cell lines [20]. In particular, we find that dcEFs require Ca 2+ to induce morphological changes that enable migration. A Ca 2+ -dependent mechanism has previously been suggested to explain dcEF-induced cell shape changes, orientation and displacement of mouse embryo fibroblasts [21]. Inhibition of lamellipod formation and cell locomotion has also been reported in dcEF-guided fish keratinocytes exposed to Ca 2+ channel antagonists [22].
Our data suggests that regulation of intracellular Ca 2+ levels is a key for dcEF induction of cell migration responses. Notably, directed motility and intracellular Ca 2+ kinetics in osteoblast-like cells can be altered by different strength dcEFs. This effect might be related to VGCC activation occurring only at strong dcEFs in osteoblasts, as a similar situation has been described for hFOB 1.19 osteoblasts [23]. Inhibiting VGCCs using CdCl 2 significantly reduced [Ca 2+ ] i peak magnitude in calvarial and SaOS-2 cells. Thus, VGCCs may contribute to Ca 2+ entry across the plasma membrane at strong dcEFs.
We show for the first time here that dcEFs trigger different patterns of intracellular Ca 2+ elevation in osteogenic cell types that migrate in opposite directions. In combination with our other data, this leads us to propose a model of how dcEFs affect local calcium kinetics to guide migration of a SaOS-2 cell. The resting potential of osteoblast-like cells is around 220 to 230 mV [24,25]. A dcEF of 10 V/cm will affect a 50 mm diameter (in average) SaOS-2 cell by hyperpolarizing the membrane facing the anode by approximately 50 mV and depolarizing the cathodal side by the same amount [26]. Increasing [Ca 2+ ] i can activate myosin light chain kinase, which in turn triggers actin-activated myosin ATPase, a major regulator of cell contraction [27]. As a consequence, the cathodal side will contract and propel the SaOS-2 cell towards the anode. A cathode-facing membrane depolarization may open VGCCs and the following Ca 2+ influx will enhance the local [Ca 2+ ] i , causing cell migration towards the anode. Stretch-activated cation channels, including calciumrelated ones involved in mechanotransduction [28], are present in osteoblasts and are crucial for maintaining bone density [29]. Thus, stretch-activated calcium channels may especially contribute to the local [Ca 2+ ] i increase due to the mechanical strain induced by the EF, and cause a depolarization-activated calcium response [30] at the rear-end (cathode) of the SaOS-2 cell. Similar mechanism, but in an opposite direction, will translate the local increase in [Ca 2+ ] i at the anode facing side of a calvarial cell into movement towards the cathode. Nevertheless, how and why [Ca 2+ ] i rises on the anodal side for calvarial and cathodal side in SaOS-2 cells is still an issue to be discussed. One possible basis for this opposite effect could be the status of VGCCs in the two cell lines used. Because of CdCl 2 being a non-specific VGCC inhibitor, its use can only address the VGCCs in general, but can not reveal VGCC subtypes involve in [Ca 2+ ] i rise in two cell lines exposed to strong dcEF. Additionally, Ca 2+ -permeant mechanosensitive channels might be secondary modulators to VGCCs contributing to the differential directionality of response. However, it is likely that another mechanism keeps [Ca 2+ ] i in a steady state during weak (5 V/cm) dcEF-directed motility. Hereby, Ca 2+ ions, but not necessarily local Ca 2+ elevations, also seem to be involved in directed motility since the morphology and preferred migration direction of the cells was disrupted when Ca 2+ was depleted from the media during the exposure to weak dcEF. A mechanism involving the local activation of some other channels/ proteins might modulate directed cell migration in weak dcEFs.  Interestingly, we observed a rapid contraction and a slight relocation with a preferred directionality within the first minutes post-application of dcEFs (1-14 V/cm) in both calvarial and SaOS-2 cells. Each cell type showed a positional shift that was opposite to its preferred migrational direction, a behavior that we term the refractory period. During this time it is possible that the cells mechanically resist the dcEF forces until they reach the appropriate orientation. Considering that both the morphological reorganization and the positional shift noted here have preferred directions, they may be predetermining the migration direction and hence, required for electrically-guided cell migration. We suggest that the contraction and the shifting may also be occurring in response to the local calcium elevations at rear-end of the cells since both events refer to the same cell side with regard to the migration direction for the given cell type.
Cell migration with a preferred directionality may involve the local activation of various charged receptors or ion channels that subsequently contribute to migratory processes at the leading edge of the cell [7]. Moreover, as many molecules have charges on them, it is likely that electrical stimulation directly induces the activity of e.g. membrane proteins participating in ion transport or other molecules acting as a voltage sensor such as ciVSP, a recently identified voltage sensing phosphatase [31]. Therefore, investigating the roles of such proteins and associated mechanisms for translating electrical signals into cell biological events producing guided cell migration is our future aim.
In conclusion, the findings here provide insight into the involvement of cell-and dcEF-specific cues in guided migration. Our overview of the time scale of electrically-directed migration in osteoblast-like cells, presented for the first time here, helps to understand how physiological electrical signals can initiate directional sensing in the cells. This will contribute to our understanding of important biological events in which guided migration is crucial, such as bone remodeling.

Osteoblast cell culture
Primary osteoblastic cells were isolated from fetal rat calvaria as previously described [32]. Cultures were maintained in DMEM/ Ham's F12 medium (Gibco BRL, Karlsruhe, Germany) containing 12% fetal calf serum, 2.3 mM Mg 2+ , 100 mg/mL penicillin, 100 mg/ml streptomycin sulfate and 1.25% L-glutamine at 37uC and humidified 5% CO 2 . Passages between 5 and 7 were taken for experiments. Human osteosarcoma cells (SaOS-2), non-transformed cells with osteoblastic properties, were obtained from the American Type Culture Collection (ATCC HTB 85) and cultured in McCoy's 5A medium (Gibco BRL, Karlsruhe, Germany) containing 15% fetal calf serum and 1.25% L-glutamine. SaOS-2 cells were used up to passage number 50 since they are suspected to loose their osteoblast-phenotypic features in later passages [33]. 100 mL of 20000 cells/mL was seeded into the channel of an ibiTreat -m Slide I (Ibidi, Munich, Germany) for EF applications.

Application of dcEFs
Direct current provided by an electrophoresis power supply (Amersham Pharmacia Biotech, Freiburg, Germany) was applied to the cells through two platinum electrodes (0.2 mm diameter, Agar scientific, Essex, UK) immersed in 0.9% NaCl-filled beakers that were connected to the media-filled reservoirs of the m Slide by two approximately 20 cm long agar bridges (2% agar in PBS). HEPES buffered salt solution (HBSS; 10 mM HEPES, 135 mM NaCl, 1.2 mM CaCl 2 , 1.2 mM MgCl 2 , 5 mM KCl and 0.8% D-Glucose, pH 7.4) or cell culture media was used for ion imaging or migration assays, respectively. Agar bridges were used to avoid contact between electrode products and the cells [2]. Field strengths were measured during the experiment by a voltmeter (VoltcraftH Mebtechnik, Hirschau, Germany). Cells were exposed to the different dcEF strengths ranging from 1 to 14 V/cm for 5 min up to 5 h at 37uC in a chamber (Solent Scientific, Segensworth UK) covering the whole inverted microscope or in the incubator. Same conditions were used for controls but without application of a dcEF.

Quantification of cell orientation and migration
Cells were monitored with an Olympus IX81 inverted microscope equipped with DIC components and an integrated vital microscopy chamber (Olympus, Hamburg, Germany). Cells with an angle between 70u and 110u to the EF vector were considered as perpendicular. A total of 1000-1500 cells were scored for each condition. To quantify cell migration, time-lapse DIC videos were processed in a detecting program, TrackIT. Diagrams indicating the direction of cell migration were evaluated in Microsoft Power Point from original cell trajectories detected by TrackIT. Graphics of speed and distance were prepared in Excel from the original data generated in TrackIT.

Monitoring intracellular calcium
Fura-2 AM (Invitrogen, Karlsruhe, Germany), a Ca 2+ -specific vital dye, was used to record intracellular calcium levels. Cells were washed once with HBSS and then loaded with 3 mM Fura-2 AM, 0.05% Pluronic F-127 (Sigma-Aldrich, Munich, Germany) and 3% FCS in HBSS at 37uC for 15 min. After washing, cells were incubated further with HBSS containing 3% FCS for 10 min, then rinsed with HBSS and examined using an inverted microscope. Ca 2+ free conditions were ensured by using 1.2 mM EGTA (Applichem, Darmstadt, Germany) and 10 mM Thapsigargin (Calbiochem, Darmstadt, Germany) to define the requirement of extracellular calcium ions and calcium ions released from the intracellular stores. For VGCC blocking studies, intracellular calcium stores were first depleted using Thapsigargin to ensure that an eventual elevation in [Ca 2+ ] i induced by the dcEFs was only due to the influx through the cell membrane. Afterwards the cells were further incubated with 50 mM CdCl 2 (Sigma-Aldrich, Munich, Germany). A DcEF was applied approximately 3061 sec after the start of the fluorescence recording during the experiments. The graphic legends with the number of series indicate the region of interests (ROIs) inserted in the cells in the given microscopic field. Data was recorded and analyzed using the imaging software Olympus CellˆR.

Statistical analysis
Statistical calculations were performed either in Excel using Student's t-test or in SPSS using one-way analysis of variance (ANOVA) with Bonferroni test. P-values,0.05 were considered as significant (*). n cell , n ROI and n exp , represent the total number of scored cells, total number of detected ROIs and the total number of independent experiments, respectively. Data were collected from at least three individual experiments.

Supporting Information
Movie S1 DcEF-directed motility in calvarial osteoblasts. DIC time-lapse video showing dcEF-directed motility of calvarial osteoblasts exposed to 5 V/cm for 5 h. Calvarial osteoblasts elongate perpendicular to EF vector and form membrane protrusions towards cathode (2)  Movie S2 DcEF-directed motility in SaOS-2 osteoblast-like cells. DIC time-lapse video showing dcEF-directed motility of SaOS-2 cells exposed to 5 V/cm for 5 h. SaOS-2 cells elongate perpendicular to EF vector and form membrane protrusions towards anode (+) during EF-induced directional migration.  (2). EF was switched on and off for 5 and 1 minutes, respectively during the experiment. The relocation (positional shift) caused by the contraction is opposite to the preferred migratory direction, which is towards anode (+) for these cells. Found at: doi:10.1371/journal.pone.0006131.s006 (9.14 MB MOV) Figure S1 Migration direction of calvarial osteoblasts exposed to dcEF. Original images showing directions, pathways and linear displacements of calvarial osteoblasts towards cathode (at 180u). Cells were exposed to 5 V/cm for 5 h. Data was generated using cell tracking program Olympus cellˆR-TrackIT. Found at: doi:10.1371/journal.pone.0006131.s007 (1.69 MB TIF) Figure S2 Migration direction of SaOS-2 osteoblast-like cells exposed to dcEF. Original images showing directions, pathways and linear displacements of SaOS-2 cells towards anode (at 0u). Cells were exposed to 5 V/cm for 5 h. Data was generated using cell tracking program Olympus cellˆR-TrackIT. Dataset S1 Directedness of EF-guided osteoblast-like cells in the presence or absence of Ca2+ ions. Values between 90u-270u corresponds to the directions towards cathodal-and those between 0u-90u or 270u-360u towards anodal side. Ncell = 15-22 for each condition. Found at: doi:10.1371/journal.pone.0006131.s010 (0.02 MB XLS)