Electroacupuncture facilitates the integration of a grafted TrkC‐modified mesenchymal stem cell‐derived neural network into transected spinal cord in rats via increasing neurotrophin‐3

Abstract Aims This study was aimed to investigate whether electroacupuncture (EA) would increase the secretion of neurotrophin‐3 (NT‐3) from injured spinal cord tissue, and, if so, whether the increased NT‐3 would promote the survival, differentiation, and migration of grafted tyrosine kinase C (TrkC)‐modified mesenchymal stem cell (MSC)‐derived neural network cells. We next sought to determine if the latter would integrate with the host spinal cord neural circuit to improve the neurological function of injured spinal cord. Methods After NT‐3‐modified Schwann cells (SCs) and TrkC‐modified MSCs were co‐cultured in a gelatin sponge scaffold for 14 days, the MSCs differentiated into neuron‐like cells that formed a MSC‐derived neural network (MN) implant. On this basis, we combined the MN implantation with EA in a rat model of spinal cord injury (SCI) and performed immunohistochemical staining, neural tracing, electrophysiology, and behavioral testing after 8 weeks. Results Electroacupuncture application enhanced the production of endogenous NT‐3 in damaged spinal cord tissues. The increase in local NT‐3 production promoted the survival, migration, and maintenance of the grafted MN, which expressed NT‐3 high‐affinity TrkC. The combination of MN implantation and EA application improved cortical motor‐evoked potential relay and facilitated the locomotor performance of the paralyzed hindlimb compared with those of controls. These results suggest that the MN was better integrated into the host spinal cord neural network after EA treatment compared with control treatment. Conclusions Electroacupuncture as an adjuvant therapy for TrkC‐modified MSC‐derived MN, acted by increasing the local production of NT‐3, which accelerated neural network reconstruction and restoration of spinal cord function following SCI.


| INTRODUC TI ON
Spinal cord injury (SCI) results in damage to spinal cord neurons, disrupting axonal tracts, and causing the dysfunction or loss of locomotor function, sensation, and autonomic functions below the injured spinal cord segments. To date, no effective therapy has been established for SCI. This is likely due to the multidimensional pathophysiological changes and the obstructive microenvironment that develop in the injury region. 1 A combinatory treatment may result in a better therapeutic response relative to monotherapy by addressing multiple pathological aspects or mechanisms associated with SCI, such as inflammatory reactions, glial scar formation, and the insufficient supply of neural growth factors. 2,3 For functional recovery after SCI, a potential therapeutic strategy that has been proposed is the fabrication of a neuronal relay between the injured axonal tract projection fibers and the denervated neurons, which represents a necessary primary step for repairing damaged neural connectivity.
To achieve this goal, implantation of exogenous stem cell-derived neurons into the injured spinal cord tissue has been the focus of many studies. This is because it is highly unlikely that the lost neurons can be replaced by the host neural stem cells, which tend to differentiate into neuroglial cells at the injury sites. 4,5 The direct transplantation of exogenous stem-cell-derived neurons or neuronal progenitor cells into the injured spinal cord may help compensate for and replace lost host neurons. Mesenchymal stem cells (MSCs), which are easy to obtain, ethically uncomplicated, and relatively safe, have emerged as a promising candidate for SCI treatment, with the potential to differentiate into neuron-like cells. Although this approach remains controversial, experimental evidence has suggested that MSCs, after suitable modifications, can display phenotypic and functional neuronal properties. 6,7 We have previously reported that an MSC-derived neural network (MN) implantation strategy was able to facilitate the recovery of limb motor function in a rat model following thorough spinal cord transection. 8  Electroacupuncture (EA), which is a traditional Chinese medicinal therapy, has been widely used and applied during clinical practice and in animal models for the treatment of SCI. [9][10][11][12][13][14] The application of EA has been demonstrated to have neural protective and antiinflammatory effects, which would improve the suitability of the damaged areas for axonal regeneration after SCI. In addition, because the NT-3 levels in the injured spinal cord may represent a supportive factor during SCI repair, the potential that EA might enhance NT-3 levels may support EA as a potential therapeutic strategy for SCI. The differentiation rate of donor NSCs and the induction of MSCs into neuron-like cells at the site of SCI were reported to improve significantly following the application of EA. 11,15,16 However, implanted prototype MSCs exhibited a low rate of development into neuron-like cells.
These findings suggested that the combination of tissue engineering, such as neural network transplantation, together with EA therapy may represent an optimal approach for achieving increased survival and neuron-like differentiation among implanted MSCs, which we explored in this study. We report here that the application of EA can increase the NT-3 contents in transected spinal cord tissue, especially in the host tissue adjacent to the injury/graft site.
More importantly, EA promoted the survival, differentiation, and migration of donor MSC-derived neuron-like cells, which is essential for the improved formation of potential neuronal relays at the injury/ graft site of the spinal cord.

| Animal care
Eighty-four Sprague-Dawley (SD) rats (Adult, female, 220-250 g) were used for all animal experiments in this study. All experimental protocols and animal handling procedures were approved by the MN was better integrated into the host spinal cord neural network after EA treatment compared with control treatment.
Conclusions: Electroacupuncture as an adjuvant therapy for TrkC-modified MSCderived MN, acted by increasing the local production of NT-3, which accelerated neural network reconstruction and restoration of spinal cord function following SCI.

| Isolation and culture of MSCs and SCs
Mesenchymal stem cells were isolated, as previously described, from green fluorescent protein (GFP) transgenic SD rats (Osaka University, Osaka, Japan), which ubiquitously express GFP in all tissues. 18 Briefly, one-week-old rats (n = 10) were sacrificed, their femurs were removed, and flushed of bone marrow, which was cultured in low-glucose Dulbecco's modified Eagle medium (L-DMEM, Gibco), supplemented with 10% fetal bovine serum (FBS, TBD Co, Tianjin, China) and 4 mM L-glutamine (Invitrogen, USA), in a 5% CO 2 incubator at 37°C. When adherent cells reached 80% confluence, they were passaged (1:3) into different culture flasks.
MSCs from passages 3 to 5 were used for all experiments in this study.
To isolate Schwann cells (SCs), five-day-old neonatal SD rats (not GFP transgenic rats, n = 30) were decapitated and sterilized. The sciatic nerves and brachial plexus were dissected, and the adherent connective tissue and epineurium were removed under a dissecting microscope. The nerves were cut into small pieces (< 2 mm) and digested with 0.16% collagenase (Sigma-Aldrich, St. Louis, MO) at 37°C for 15 min. The dissociated tissue was placed on culture dishes coated with poly-L-lysine and containing DMEM/F12 medium containing 10% FBS for 30 min at 37°C in a 5% CO 2 humidified atmosphere. After 30 min, an additional 2 ml of medium was added to each culture dish. The medium was changed every 2 days. After 5-7 days, when the cells reached 80% confluency, they were subcultured and purified using differential velocity adherent methods.
Based on our recent study, 95-96% of the cells were estimated to be SCs. 19

| MSC and SC transfection and seeding on a 3D gelatin sponge scaffold
Briefly, in vitro, the MSCs and SCs were transduced with recombinant adenoviruses (Advs) containing the TrkC gene (Adv-TrkC) and the NT-3 gene (Adv-NT-3), respectively, to induce the overexpression of TrkC and the oversecretion of NT-3. After 3 h, Adv-TrkC, administered at a multiplicity of infection (MOI) =300, yielded 81.32% transduced MSCs, whereas Adv-NT-3 administered at an MOI =100 yielded 78.19% transduced SCs, with excellent viability. When higher viral titers were used, the cells tended to show pathological changes. The medium was replaced with DMEM/F12 supplemented with 10% FBS, and the cells were incubated for 24 h at 37°C.
A three-dimensional (3D) gelatin sponge scaffold with a 3-mm diameter and a 2-mm length was prepared, as previously described. 8 Scaffolds were seeded with 1 × 10 5 total cells (equal numbers of MSCs and SCs) to generate MNs and were cultured in 10 µl culture medium and incubated at 37°C for 14 days; the culture medium was changed every 2 days.

| Spinal cord injury and transplantation
Three days before surgery, the rats received a subcutaneous cyclosporin A injection in the belly (1 mg/100 g per rat). They were anesthetized with 1% pentobarbital sodium (40 mg/kg, i.p.). A laminectomy was performed to expose the T9 and T10 spinal cord segments, and the dura was slit vertically with a pair of micro-forceps and micro-scissors. A pair of angled micro-scissors was used to fully transect the spinal cord, and a 2-mm segment of the spinal cord was removed at the T10 spinal cord level. Then, either the generated MNs (the MN group) or gelatin sponge scaffolds containing no cells (the GS group) were used to fill the spinal cord gap. 8 After the surgical incisions were sutured, the rats received extensive postoperative care, including the intramuscular injection of penicillin (50,000 U/kg/day) for 3 days. The manual emiction was performed on the experimental rats twice daily until automatic micturition was re-established. Cyclosporin A was administered once daily for 8 weeks. China) were inserted at a depth of 5 mm into the GV acupoints. 16 The two pairs of needles were connected to the output terminals of an EA apparatus (model number G6805-2A, Shanghai Medical Electronic Apparatus Company, China). EA was applied using alternating strings of dense sparse waves at alternating frequencies (60 Hz for 1.05 s and 2 Hz for 2.85 s, pulse width of 0.5 ms). The positive and negative electrodes used to connect the two pairs of needles were alternated with each treatment. During the EA process, the current intensity was tested in the animal's body between the acupoint pair across the graft location, which indicated a current intensity of approximately 5 µA.

| Assessment of locomotor function
After surgery, the rat hindlimb function was assessed weekly, using the Basso, Beattie, and Bresnahan (BBB) open-field locomotor test 21 to evaluate voluntary movement and body weight support and a modified inclined-grid climbing test to qualitatively assess the accuracy of foot placement and coordination, which differentiates local reflex activity from voluntary movement. Two independent investigators blinded to the experimental treatments determined the BBB scores.

| Electrophysiology
Evoked potentials (EPs) were measured and recorded (n = 5/group), as previously described, to evaluate motor axonal conduction. 22 Under general anesthesia, the sensorimotor cortex (SMC) and sciatic nerve of rats were exposed. Stimulating electrodes (NeuroExam M-800 Data Acquisition Analysis System, MEDCOM, Zhuhai, China) were placed in the SMC (located 2 mm lateral to the midline and 2 mm caudal to Bregma), and the recording electrodes were connected to the sciatic nerve. The amplitude of the cortical motor-EP (CMEP) was calibrated and then recorded.

| CTB retrograde tracing
Two months after surgery, the rats were anesthetized with 1% pentobarbital sodium (40 mg/kg, i.p.), and their sciatic nerves were exposed under sterile conditions. Using a dissecting stereomicroscope (Leica Microsystems, Inc., Wetzlar, Germany), the needle tip of a 30 g needle Hamilton syringe (Hamilton Co., Reno, USA) was inserted 10 mm into the sciatic nerve along its longitudinal axis and then withdrawn 2 mm to create a potential pool for injection. Then, 2 µl of 2% cholera toxin B (CTB) conjugated to Alexa Fluor 555 (CTB-555, Life, USA) was slowly injected, as previously described. 23 The sciatic nerve proximal to the injection site was lightly crushed with a pair of blunt forceps to facilitate CTB-555 uptake by the nerve fibers. The injection site was thoroughly rinsed using a sterile saline-soaked cotton-tipped stick, and the wound was sutured. The rats were sacrificed 1 week after tracer injection.

| Pseudorabies virus (PRV-CMV-mRFP) retrograde tracing
The same injection procedure described for CTB retrograde tracing was used to slowly inject 1 µl of pseudorabies virus (PRV, 2.5 × 10 9 PFU/ml; Brainvta, China), after which the needle was maintained in place for 5 min. The sciatic nerve was lightly crushed with a pair of blunt forceps proximal to the injection site to maximize PRV contacts with the nerve fibers. The injection site was thoroughly rinsed with a sterile saline-soaked cotton-tipped stick, and the wound was sutured.
Twenty rats (n = 5/group) were sacrificed 6 days after injection. were centrifuged for 10 min at 18800 g at 4°C and used for Western blot analysis. The relative expression of target protein was indicated by the target protein/GAPDH (loading control) intensity ratio as reported in a previous study. 24 The rats were sacrificed at the end of 2 or 8 weeks following scaffold transplantation. All rats were deeply anesthetized with 1% pentobarbital sodium (50 mg/kg, i.p.) and transcardially perfused with normal saline containing 0.002% NaNO 2 and 0.002% heparin, followed by a fixative containing 4% paraformaldehyde (PFA) in 0.1 M PBS (pH 7.4). The spinal cord was dissected and post-fixed for 24 h in the same fixative, followed by 30% phosphate-buffered sucrose at 4°C for 48 h. Samples were frozen and embedded in OCT compound. The T8-T12 successive segments of the spinal cord were cut into longitudinal 25-µm-thick sections.

| Immunofluorescence staining
Specific proteins in the obtained spinal cord tissue sections were detected by immunofluorescence staining (IFS). Sections were immersed in 0.01 M phosphate-buffered saline (PBS) three times for 5 min and then blocked with 10% goat serum for 30 min at 37°C. The sections were incubated with primary antibodies in 0.01 M PBS containing 0.3% Triton X-100 at 4°C for 12 h. The tissue sections were washed with PBS three times for 5 min and incubated with secondary antibodies for 1 h at 37°C. The slides were observed, and images were captured using a fluorescence microscope (Leica) or confocal microscope (Carl Zeiss). A summary of the antibodies used can be found in Table S1.

| In situ hybridization
Twenty-five rats (n = 5/group) were transcardially perfused 2 weeks after SCI. The spinal cord was dissected from the T8 to T12 successive segments, post-fixed in 4% paraformaldehyde for 2-4 h and saturated in 30% sucrose overnight at 4°C; 1% diethylpyrocarbonate was added to the above solution to prevent mRNA degradation. The After brief staining with Hoechst33342, the sections were covered with coverslips and examined under a fluorescence microscope.  For the quantification of NT-3 mRNA or NT-3 protein in vivo, positive cells were quantified using ImageJ, as described previously. 27 Setting the epicenter of the SCI site as the origin, the injury site, and adjacent areas were separated into eight regions. NT-3 mRNA positive cells were converted into an area of interest (AOI), and the pixel area of each AOI was automatically calculated by ImageJ. The staining density for each region, representing the total pixels in each AOI, was then calculated.

| Statistical analysis
All statistical analyses were performed using the statistical software SPSS version 20.0. Data are reported as the mean ±standard deviations (SD). The data were analyzed using one-way analysis of variance (ANOVA). The Shapiro-Wilk's test and Levene's test were conducted to confirm normality of the data. If equal variances were found, the least significant difference test was applied; otherwise, the data with heterogeneity of variance were analyzed using nonparametric test (Kruskal-Wallis test). When two groups were compared, the unpaired Student's t-test was used. p < 0.05 was considered significant, and p < 0.01 was considered highly significant.

MSC-derived neuron-like cells
To assess whether EA could promote the de novo formation of synaptic contacts between the transplanted MSC-derived neuron-like cells, western blot, immunofluorescence staining, and immunoelectron microscopy assays were performed. The results showed that donor Map2-positive neuron-like cells displayed only moderate expression levels of synaptophysin (SYN) in the injury/graft site of the spinal cord in the MN group (Figures 2A, a1-3); however, in the MN+EA group, many donor Map2-positive neuron-like cells emitted intense SYN-positive immunofluorescence ( Figures 2B, b1-3).

| Pseudorabies virus tracing
To explore whether EA promoted the integration of transplanted

| Improvements in paralyzed hindlimb motor function
Behavioral observations, electrophysiological detection, the BBB open-field motor assessment, and modified grid climbing tests were then performed to assess the functional recovery of rats subjected to the various treatments. In normal animals, stimulation at certain points of the sensorimotor cortex could evoke CMEPs with short latencies and large response amplitudes ( Figure 5A-C). Small CMEPs could be evoked in the GS group ( Figure 5A). In the GS+EA, MN, and MN+EA groups, EA treatment or MN transplantation markedly improved CMEP performance compared with that in the GS group ( Figure 5B,C, p < 0.05), and MN+EA treatment further shortened the latency and increased the amplitudes of CMEPs ( Figure 5B,C, p < 0.05). BBB open-field motor assessments showed that the hindlimbs of all rats were completely paralyzed following T10 spinal cord transection. Over 8 weeks after operation and transplantation, hindlimb locomotor performance showed progressive improvements in all groups. At 8 wpi, the mean BBB score in the MN+EA group was 8.92 ± 0.81. Some animals in the MN+EA group showed the extensive movement of all three hindlimb joints and interval plantar support of the paw, with BBB scores as high as 10 ( Figure 5D). At 8 wpi, the 45° sloping grid climbing method was used to evaluate the spontaneous placement reflex triggered by direct touch (Figure 5E; Video S1 shows the grid climb test in the GS, GS+EA, MN, and MN+EA groups). The hindlimbs of rats in the GS group were dragged behind when rats climbed the sloping grid with their forelimbs. In the GS+EA and MN groups, the rats climbed up the grid using the backs of their paws. The rats in the MN+EA group, however, stepped on the grid with the plantar surfaces of their hind feet and displayed the coordinated movement of their forelegs and hindlegs.

| EA increases NT-3 levels in the injured spinal cord
To evaluate NT-3 expression in the injured spinal cord following various treatments, the injury/graft site and adjacent segment tissues were processed for immunofluorescence and in situ hybridization stainings. NT-3 levels were calculated by measuring the standard-  (Figures 6B and S8).

| Transwell migration of MSCs
We explored the effects of NT-3 on the migration capability of TrkC-MSCs in vitro. As shown in Figure S9 when grafted into SCI sites in immunodeficient rats, first expressed markers of neuronal maturity and continued maturation over a period of 1.5 years. 31 We believe that a cell therapy regimen that combines a bioscaffold-based cell delivery method with EA treatment, as was adopted in this study, can greatly improve donor cell survival in the hostile SCI microenvironment. 9 Using this strategy, the donor cell population showed a better survival rate and maintained a neuronlike phenotype for at least eight weeks after EA. Another finding from this study was the importance of the NT-3 level in the injured spinal cord, which was markedly augmented by EA treatment. 11,25 Our recent study has also verified that EA could stimulate the spinal Studies have demonstrated that NT-3 can attenuate neuroinflammation. 34,35 It is well documented that vigorous inflammatory reaction occurs readily following the SCI. Appropriate inflammatory response has beneficial effects notably in the early stage of SCI, such as the removal of cellular debris through phagocytosis.
However, excessive and prolonged inflammatory response can cause damage to the surrounding tissues and is unfavorable for the spontaneous regeneration and functional recovery. 36 Several studies have reported that EA or MSC transplantation can effectively regulate the inflammatory responses; more importantly, it protects the injured spinal cord from the secondary pathological reaction. The therapeutic approaches include the EA, cell therapies, monoclonal antibodies among others. [37][38][39] It has been reported that EA can reduce inflammation via exciting specific neural pathways. This would further activate particular receptors in the spleen that suppressed pro-inflammatory molecules. 37 In addition, MSCs can reverse the phenotype of M1 macrophage/microglia towards anti-inflammatory M2 subpopulation in vitro and in vivo. 39 In this study, we found that the combination of EA and MN could reduce the expression of CD68 protein in the injured spinal cord at 8 weeks SCI. The above when taken together with the present results suggest that EA in combination with MN transplantation can exert a synergistic effect by inhibiting the inflammatory response in SCI.
Very interestingly, in a separate recent study, it has been reported that MSC transplantion downregulates the NT-3 expression level in ischemia-injured brain. 40 The discrepancy in results may be attributed to differences in phenotype of transplanted MSCs used and also the use of different animal models. Further studies are clearly desirable to decipher the underlying mechanisms. Besides the neurotrophin and immune regulation, 41 emerging evidence has shown that MSC transplantion can protect and improve the neuron function after injury via exosomes 42 and mitochondrial transfer. 43 Notwithstanding of the above, it is suggested that MN+EA treatment as demonstrated in the present study can repair the SCI effectively which is likely through multiple effects. All in all, it is suggested that EA can facilitate the cell survival, migration, differentiation, and

CO N FLI C T O F I NTE R E S T
The authors have no financial conflicts of interest.

AUTH O R S' CO NTR I B UTI O N S
YSZ and YD designed and supervised the study. YY, HYX, DQW, GHW, LJW, HJ, JWR, BJ, and YQW performed the experiments and collected the data. YY, HYX, YD, and YSZ summarized, analyzed, and plotted the data, and drafted the manuscript. YD, YSZ, and XZ wrote and finalized the manuscript. All authors have given approval to the final version of the manuscript.

DATA AVA I L A B I L I T Y S TAT E M E N T
The data that support the findings of this study are available from the corresponding author upon reasonable request.