Electrical Stimulation Enhances Migratory Ability of Transplanted Bone Marrow Stromal Cells in a Rodent Ischemic Stroke Model

Ischemic stroke is a major cause of long term disability and mortality in the world because of its high incidence and narrow therapeutic window. Recently, cell transplantation therapy has emerged as a new therapeutic option for ischemic stroke. In particular, bone marrow stromal cells (BMSCs) transplantation is promising therapy [1, 2]. Various studies have been completed concerning BMSCs transplantation route and timing in both basic and clinical studies, yet none of these investigations have explored cell migratory potentials [3-6]. There are currently no effective methods of guiding transplanted cells toward the targeted site. Ischemic stroke itself has the effect of guiding transplanted cells [7-9], however, only a small number of cells reach the ischemic penumbra by trans-arterial transplantation, which is the less invasive method of cell transplantation. Moreover, intravenous transplantation, the least invasive method, delivers even fewer cells due to the majority of transplanted cells being trapped in peripheral organs [10-12]. Although intracerebral transplantation can deliver the entire cell transplant into the targeted site, this method creates a new injury to the brain tissue and cannot distribute the cells widely. The ability to effectively direct a greater number of cells to the ischemic penumbra may allow cell transplantation to become a more potent therapy option.

Independent of transplantation therapies, electrical stimulation has become an important therapeutic strategy for central nervous system disorders such as Parkinson’s disease and stroke. There are some reports that electric fields may direct various types of cells in vitro [13-15]. Furthermore, electrical stimulation enhances migration of endogenous neural stem cells and cell proliferation in vivo [16, 17]. However, to date no in vivo studies exist investigating the ability of electric current to guide the transplanted cells within the ischemic brain.

In this study, we applied electrical simulation to the infarcted hemisphere and examined the migratory ability of transplanted allogenic BMSCs in a rat model of ischemic stroke.

This study was conducted in accordance with the guidelines of the Institutional Animal Care and Use Committee of Okayama University Graduate School of Medicine. The protocol was specifically approved by the Institutional Animal Care and Use Committee of Okayama University Graduate School of Medicine (protocol #OKU- 2015390). Adult male Wistar rats were used in this study. For euthanasia, pentobarbital was used. MCAO was carried out under general anesthesia (1 % halothane in 69 % N₂O and 30 % O₂). All efforts were made to minimize suffering animals.

Adult male Wistar rats (CLEA JAPAN, Inc., Tokyo, Japan) weighing 200 to 250 g at the beginning of the experiment were used in this study. They were housed singularly per cage in a temperature and humidity-controlled room, maintained on a 12-hour light/dark cycle, with free access to food and water. Body weight (BW) was recorded at day 0, 1, 4, 8 and 15.

Rats were euthanized by sodium pentobarbital (100mg/kg) with subsequent removal of the femoral bones (n=6). After the femoral bone marrow was flushed, cells were cultured at 2×10⁴ cells/cm² in DMEM (Gibco, Cergy Pontoise, France) supplemented with 10 % fetal calf serum (Gibco) and 1 % (v/v) penicillin/ streptomycin (Gibco). Cells were cultured at 37 °C in a fully humidified atmosphere with 10 % CO₂. The medium was changed twice a week. In the present study, we used BMSCs after 3 passages.

Transient Middle Cerebral Artery Occlusion. MCAO was carried out according to the intraluminal suture method used in our previous study [18-20]. Under general anesthesia (1% halothane in 69 % N₂O and 30 % O₂) and a 4–0 monofilament nylon suture with silicone-coated tip (Xantopren L blue & ACTIVATOR Universal Liquid, Heraeus Kulzer GmbH & Co. KG, Hanau, Germany) was inserted through an arteriotomy of the right common carotid artery into the right MCA. After MCAO for 90 minutes, the filament was withdrawn and common carotid artery was ligated with 3–0 silk suture.

Cell labeling and intra-cerebral BMSC transplantation. Before cell transplantation, BMSCs were labeled with Q-Tracker 625 Cell Labeling kit (Q-dot nanocrystal, a red fluorescent marker; Invitrogen, Carlsbad, CA) for characterization of BMSCs according to the manufacturer’s protocol. Once the BMSCs were cultured using Q-Tracker 625 Cell Labeling kit, Q-dot nanocrystals were delivered into the cytoplasm of BMSCs. Thereafter, the cells were detached with 0.05 % trypsin-EDTA, and 2.5×10⁵ cells were diluted with 4µL of phosphate-buffered saline (PBS). The procedure of intracerebral transplantation in rats was as follows: At 1 day after MCAO, the rats were deeply anesthetized with pentobarbital (35 mg/kg, intraperitoneally) and placed in a stereotaxic instrument (Narishige, Tokyo, Japan). Approximately 2.5×10⁵ BMSCs in 4µL PBS was injected into the left corpus callosum with a 24 G Hamilton syringe. The injection coordinates were anterior-posterior (AP) 1mm from the bregma, mediolateral (ML) 2.5 mm to the sagittal suture, and dorsoventral (DV) 2.5 mm from the dura.

Electrode implantation and electrical stimulation. Following the cell transplantation, craniectomy (3.5 mm×3.5 mm) was performed at the right frontal bone which centered on anterior 1.00 mm and lateral 2.5 mm from the bregma. Subsequently, a disc electrode with a diameter of 3 mm was implanted on the right frontal epidural space and which was embedded in the skull with dental cement. A counter electrode was placed in the extra-cranial space. The electrode was connected to an implantable pulse generator unit (ISE1000SA; UNIMEC, Tokyo, Japan). All the systems were embedded under the skin. Rats were randomly divided into the stimulation group (n=10), and control group (n=8). The home cage of the rats was placed on a stimulation commander (ISE1010C; UNIMEC) which was connected to an electronic stimulator and isolator (SEN-7203 and SS-202J; NIHON KOHDEN, Tokyo, Japan). The stimulation system allows the rats to move freely in their home cage within the electrical stimulation. Rats of the stimulation group were continuously stimulated for 2 weeks after electrode implantation. Stimulating pulses were square-wave pulses and preset at the duration of 100µs constant current. The parameter of stimulating pulses was adjusted to the preset current (100 µA) and frequency (100 Hz). The stimulation pulses were checked by analog-to-digital converter and software (PowerLab 8/30 and Chart5; AD Instruments, Dunedin, New Zealand). We used a monopolar setting to reduce the risk of tissue damage and to stimulate more diffusely [16, 21]. After the 2 weeks of stimulation, the electric stimulation was discontinued (Fig. 1).

mNSS. Behavioral testing was performed at day 0, 1, 4, 8 and 15 using the modified Neurological Severity Score (mNSS) and cylinder test (n=18). The mNSS was used to assess motor function, sensory disturbance, reflex, and balance. In this score, one score point is awarded for the inability to perform the test or for the lack of reflex. Therefore, the higher score indicates more severe injury. Neurological function was graded on a scale of 0 to 18 (normal score: 0; maximal deficit score: 18) [22-24]. Only rats showing 7–12 points in the mNSS 1 day post-reperfusion were used in this study.

Cylinder test. The cylinder test was performed the same timing as mNSS to assess the degree of forepaw asymmetry. Rats were placed in a transparent cylinder (diameter: 20 cm, height: 30 cm) for 3 minutes and the number of forepaw contacts to the cylinder wall was counted. The score of cylinder test in this study was calculated as a contralateral bias, that is, [(the number of contacts with the non-paretic limb)-(the number of contacts with the paretic limb)/ (the number of total contacts) ×100] [25-28]. The behavioral tests were performed by the two co-authors blind to the group of rats.

Measurement of the infarction area. All rats were euthanized with overdosed pentobarbital (100 mg/kg) at 2 weeks after MCAO, and perfused transcardially with 200 ml of cold PBS and 200 ml of 4 % paraformaldehyde (PFA) in PBS. Brains were removed and post-fixed in the same fixative overnight at 4 °C, and subsequently stored in 30 % sucrose in PBS until completely submerged. Coronal sections were cut at 30-µm thickness with a freezing microtome (-20 °C). These sections were mounted onto slides and Nissl stain was performed to detect infarction area [29]. Infarction area was measured at the site of cell transplantation using a computerized image analysis (Image J; National Institutes of Health, Bethesda, USA). We evaluated infarction area ratio by the following method: infarction area ratio = LT-(RT-RI) / LT%, where LT is the area of the left hemisphere in mm², RT is the area of the right hemisphere in mm2 and RI is the infarction area in mm2 [19, 23, 30].

Measurement of cell migration. Cell migration was assessed at each 30-µm coronal sections between 360 µm anterior and posterior to the cell transplantation site. These sections were counterstained with 4, 6-diamidino-2-phenylindole (DAPI) and visualized with fluorescence microscopy (Keyence, Osaka, Japan) to detect MSCs labeled by Q-dot (n=12). At the point of histological analysis, 4 rats of stimulation group and 2 rats of control group were excluded. Because there were spilled BMSCs at the brain surface. Therefore, twelve rats were enrolled in migration assessment. Maximum migration distance was defined as the distance between cells transplanted site (2500 µm to the left from the midline) of the section and the most migrated Q-dot and DAPI double positive cell. Maximum migration area was defined as the area of polygon which formed by connecting the distributed Q-dot and DAPI double positive cells with lines in a section. This area was measured at two slices – 360 µm anterior and posterior to the cell transplantation site. The average area of anterior and posterior slices was calculated and evaluated. Maximum migration distance and area were also measured using Image J.

Immunohistochemistry for C-X-C chemokine receptor type 4 (CXCR4). In order to explore the expression of CXCR4 in transplanted BMSCs, immunohistochemistry was performed. Coronal section of cell transplantation was used. Sections were washed five times for 3 min in PBS. For CXCR4 staining, sections were incubated overnight at 4 °C with rabbit anti CXCR4 antibody (1: 500; Abcam, Cambridge, USA) with 10 % normal horse serum and washed five times in PBS. Sections were then incubated with corresponding secondary antibodies; Alexa fluor 488 fragment of goat anti-rabbit IgG (H + L) (1: 500; Thermo Fisher Scientific) for 1 hour at 4 degrees C in a dark chamber after several rinses with PBS, followed by DAPI (0.5 mg/mL) for 5 min. The sections were extensively washed with PBS, cover-slipped, and then visualized with fluorescence microscopy. Next, we evaluated the CXCR4 expression of transplanted BMSCs in each group (n=4).

To assess the efficacy of electrical stimulation on the rat’s brain, we performed ELISA analysis. Rats were performed right MCAO at day 0 (n=28). And the next day, they were randomly divided into 1 week group (each n=7, both stimulation and control) and 2 weeks group (each n=7, both stimulation and control) after mNSS assessment. Then, rats in stimulation group were placed electrode on the right epidural space and continuous electrical stimulation was performed until day 8 (1week group) or day15 (2 weeks group). Rats in control group were also placed the same stimulation system, but were not stimulated. At day 8 or day 15, rats were euthanized and performed ELISA analysis for stromal cell-derived factor 1 alpha (SDF-1α) (Fig. 2). In this experiment of ELISA analysis, rats were not received BMSCs transplantation to exclude the trophic effect of BMSCs. Brains from rats were quickly removed after decapitation using anesthesia with overdosed pentobarbital (100 mg/kg, i.p.). Brains were sliced at the thickness of 3 mm. The brain tissues of the bilateral cortex and striatum were punched out using a biopsy punch tool (3 mm hole, Kai Corporation and Kai industries co., ltd, Tokyo, Japan). Brain tissues were then homogenized in N-PER (Thermo Fisher Scientific, Waltham, USA) with Protease Inhibitor Cocktail (Sigma-Aldrich, St. Louis, USA), centrifuged at 10, 000 rpm for 10 minutes at 4 °C and the supernatant was obtained. The SDF-1α levels of tissues were measured with a SDF-1α ELISA assay kit (R&D Systems, Minneapolis, USA)

Data were analyzed using IBM SPSS Statistics version 20.0 (IBM, Armonk, USA) and presented as the means ± standard deviations. The data of BW, mNSS and cylinder test were statistically evaluated using repeated measures ANOVA. The data of histology, infarction area and ELISA were evaluated using Student’s t-test. Statistical significance was preset at a p value<0.05.

Body weight was reduced at day 1, and displayed a tendency to increase up to 15 days in both control and stimulation group (Fig. 3A). There was no significant difference of body weight between rats in stimulation and control groups (repeated measure ANOVA; F _{(1, 16)} =0.650, p=0.432).

mNSS. ANOVA revealed significant treatment effects in mNSS (repeated measure ANOVA; F _(1,16) =27.853, p<0.001; Fig. 3B). There was no significant difference in the mNSS evaluation between stimulation group and control group at day 0 and 1. Functional improvement was observed in the stimulation group compared to the control group at day 4 (control group: 9.6 ± 1.2; stimulation group: 8.1 ± 1.3, p=0.020, respectively), day 7 (control group: 8.8 ± 1.0; stimulation group: 6.3 ± 1.1, p<0.001, respectively) and day 15 (control group: 8.5 ± 1.2; stimulation group: 4.8 ± 1.5, p<0.001, respectively).

Cylinder test. Rats in the stimulation group appeared to perform better in the cylinder test than those in control groups. ANOVA revealed significant treatment effects in cylinder test (repeated measure ANOVA; F _(1,16) =9.701, p=0.007; Fig. 3C). At day 15, the stimulation group significantly improved in the cylinder test compared to the control group (control group 17.24 ± 9.27 %; stimulation group -1.43 ± 11.76 %, p=0.002, respectively).

Figure 4A shows Nissl stain of infarcted brain tissue and infarction area ratio. Infarction area of stroke rats in stimulation group were significantly decreased compared to control group (control group 64.5 ± 12.0 %; stimulation group 36.2 ± 9.9 %, p<0.001, respectively; Fig. 4B).

Migration distance of transplanted BMSCs was significantly extended in stimulation group compared to control group (control group 856.5 ± 355.4 µm; stimulation group 1961.2 ± 894.1 µm, p=0.016, respectively; Fig. 5A, C). Similarly, migration area of transplanted BMSCs was significantly extended in stimulation group (control group 0.49 ± 026 mm²; stimulation group 1.14 ± 0.43 mm², p=0.025, respectively; Fig. 5B, D).

Transplanted BMSCs were expressed CXCR4 in both control and stimulation group (Fig. 6A, B).

The SDF-1α levels were increased in the right (infarcted) cortex and striatum of rats in both stimulation and control groups, but there were no significant differences between the two groups. However only in the 1 week stimulation group, there was significant difference in SDF-1α levels between the right (infarcted) and left (intact) hemisphere. That is, SDF-1 concentration gradient of stimulation group (stimulation side > non-stimulation side) was significantly larger than that of control group. Both the cortex and the striatum of 1 week stimulation group showed significant concentration gradients of the SDF-1α (right cortex 1.19 ± 0.30 ng/ml; left cortex 0.81 ± 0.20 ng/ml, p=0.015, right striatum 1.44 ± 0.41 ng/ml; left striatum 0.92 ± 0.20 ng/ml, p=0.011, respectively; Fig. 7A). In the 2 weeks stimulation group, there were no statistical differences in left versus right or stimulation versus control groups (Fig. 7B). In comparing SDF-1α data within the stimulated groups between the two different time points, the SDF-1α level tended to be higher in the infarcted hemisphere (Fig. 8A, B), especially in the right striatum, which displayed a significant elevation of the SDF-1α level at 1 week compared to 2 weeks (1 week 1.44 ±0.41 ng/ml; 2 weeks 0.91 ±0.31 ng/ml, p=0.019, respectively; Fig. 8B).

The present study demonstrated that epidural cathodal electrical stimulation on the rat model of ischemic stroke treated with BMSCs transplantation enhanced migratory ability of transplanted cells. The rats in the stimulation group showed a decrease of brain infarction area and functional amelioration. In stimulation group, the SDF-1α concentration gradient was significantly larger than that of non-stimulation group.

Some studies have indicated that direct current electrical fields or electrical stimulation enhance migratory ability of various type of cells like NSCs and MSCs in vitro [13-15]. In the present study, BMSCs implanted in the contralateral corpus callosum migrated towards the epidural cathode electrode. Reports exist suggesting that human bone marrow-derived MSCs showed strong anodal electrotaxis, whereas rat bone marrow derived MSCs migrated towards the cathode side [15]. As mentioned previously, direction of cell migration is cell-type specific. The mechanism of electrotaxis has been reported to be the result of an influx of Ca2+ on the anodal side and a decrease of Ca2+ concentration on the cathodal side, making cell movement toward the cathodal side. However, the existence of voltage-gated Ca2+ channel on the cell will change the direction of cell movement [31]. This may explain the difference of migration direction and/or speed by various cell types.

Electrotaxis was also demonstrated by in vivo studies. Epidural stimulation of the motor cortex enhanced neurogenesis in the subventricular zone (SVZ) and migration of endogenous cell to the stimulated cortex. Some of migrated cells have expressed the neural phenotype [16, 17]. Similarly, in the transcranial direct current stimulation (tDCS), cathodal stimulation induces the recruitment of proliferating neural stem cells [32]. For migration of transplanted cells, Keuters reported that tDCS promotes the mobility of cells [33]. In our study, migration area of the transplanted cells was also wider in the stimulation group. This suggests that electrical stimulation enhances not only directional migration, but also unidirectional migration of transplanted cells. On the other hand, because of circumjacent tissue construction in vivo and possibility of less migratory ability of transplanted cells, directed migratory activity of transplanted cells was not strongly affected by tDCS [33]. As suggested above, the mechanism of the enhanced migratory ability in this study is likely not only electrotaxis; there may exist other mechanisms that enhance migratory ability of transplanted cells.

In the rat model of ischemic stroke, locally produced SDF-1α and CXCR4 expressed on the BMSCs surface plays a critical role in the migration of transplanted BMSCs [34-36]. In this study, SDF-1α concentration gradient of stimulation group was significantly larger than that of non-stimulated group. BMSCs are known to express CXCR4 [37], and transplanted BMSCs of both control and stimulation group expressed CXCR4 in this study. CXCR4 positive cells migrate towards the gradient of SDF-1α in vitro [38]. In vivo, pre-differentiated cerebellar neurons, which express CXCR4, migrate along the SDF-1α gradient [39]. This study demonstrated that electrical stimulation increases SDF-1α concentration gradient (stimulation side > non-stimulation side). It suggests that increased SDF-1α concentration may play a vital role in migration of transplanted BMSCs.

In the ischemic brain, the expression of SDF-1α peaks within a few days, and is upregulated up to 14 days [36, 40]. Here, the concentration of SDF-1α was significantly higher at 1 week compared to 2 weeks in the right striatum of stimulation group. This suggests that electrical stimulation increased the expression of SDF-1α a few days after ischemic stroke. While electrical simulation enhances expression of various cytokines like a vascular endothelial growth factor (VEGF), Glial cell line-derived neurotrophic factor (GDNF) and brain-derived neurotrophic factor (BDNF) [18, 41], there are no reports about the effect of electrostimulation on expression of SDF-1α in the brain. Interestingly, the expression of SDF-1α was significantly increased in the anal sphincter of rats immediately after 1 hour of electrical simulation [42]. This suggests that expression of SDF-1α may be enhanced in brain tissues at an earlier stage. Regarding expression of SDF-1α, it may be sufficient to stimulate for a short period in the acute phase.

In the stimulation group, functional recovery was significantly enhanced and infarction area significantly decreased. In the very-acute phase of ischemic stroke, it has been reported that cathodal electrical stimulation has neuroprotective effect and decreases the infarction area by reducing cellular depolarization [43]. Cathodal electrical stimulation also preserves cortical neurons and promotes functional recovery [44]. In addition, we used implantable electrical stimulator which can stimulate rats continuously and allow rats to move freely in their home cages. A combination of electrical stimulation and rehabilitation is thought to augment these endogenous plasticity processes [45, 46]. Therefore, rats may have showed more amelioration of motor function due to the merit of stimulation system used in this study.

This study demonstrates that electrical stimulation enhances the migratory ability of transplanted bone marrow stromal cells in a rat model of ischemic stroke. Our results suggest that electrical stimulation promotes not only electrotaxis of transplanted cells, but also chemotaxis in the ischemic stroke brain. The increased SDF-1α concentration gradient is at least one reason for enhanced migration of BMSCs by electrical stimulation.

We thank Masako Arao and Yoshie Ukai for technical assistance in the conduct of the experiments, and Marci Crowley and Michael Liska for help in manuscript preparation. Author contributions are as follows: conceived and designed the experiments: JM, TY, TA and ID; performed the experiments: JM, MU, IK, KK, KK, MO, HT, TS and AT; analyzed the data: JM, TY and CVB; contributed reagents/materials/analysis tools: JM, TY, and ID, wrote the manuscript: JM, TY, NT, ID and CVB.

No conflict of interests exists.