In the present study, we evaluated the effect of EA on the survival and differentiation of MSC transplantation, axonal regeneration, as well as functional recovery in rats with transected spinal cords. We found that the level of cAMP and NT-3, the number of vital and differentiated MSCs, the number of 5-HT-positive and CGRP-positive fibers in and near the lesion site of the injured spinal cord were all significantly increased in the MSCs+EA group as compared to the groups that underwent MSC transplantation or EA treatment alone. Furthermore, evidence from BBB scales and spinal cord evoked potentials demonstrated a significant functional recovery in the MSCs +EA group.
MSCs transplantation combined with EA treatment increased the cAMP level in the lesion site of the spinal cord, which is consistent with our previous study . The underlying mechanism of cAMP elevation is unclear. However, it is possible that the pulsed electric field of EA therapy could cause depolarization of neurons [46, 47], which would cause Ca2+ influx via L-type Ca2+ channels, followed by Ca2+-induced elevation of intracellular cAMP levels via Ca2+/calmodulin-dependent adenylyl cyclase pathway . Moreover, intracellular Ca2+ elevation caused by neuronal depolarization may stimulate an autocrine neurotrophic mechanism, leading to the synthesis and release of neurotrophic factors, NT-3 and BDNF, by the neurons themselves [49, 50]. Thus, EA may stimulate the depolarization of neurons, which causes the opening of certain voltage-gated ion channels of neuroglia cells, which subsequently stimulates a rise in the intracellular cAMP level and autocrine release of neurotrophic factors. cAMP elevation can increase recruitment of the TrkB receptor to the plasma membrane of retinal ganglion cells , suggesting that cAMP may promote neuronal survival by increasing neurotrophin receptor availability and signaling.
In previous studies, we found that EA treatment increases the amount of NT-3 in the spinal cord tissue surrounding the lesion site 2 weeks after spinal cord transection [37, 39]. In addition, when studying the effects of MSC grafts combined with EA treatment on a spinal cord injury, we obtained similar results in which EA increased NT-3 levels in the spinal cord tissue surrounding the lesion site (unpublished data). These data suggest that EA treatment may stimulate NT-3 secretion from neuroglial cells and neurons in tissue adjacent to the lesion site. Both our previous study and the present study show the NT-3 level 2 weeks after spinal cord transection is higher than the level 4 weeks after transection, but the EA and MSCs+EA groups consistently maintained higher levels of NT-3 during this time compared to the Op-control or MSCs groups. Interestingly, in the present study, we found that EA combined with MSC transplantation significantly increased the quantity of NT-3 within the lesion site as compared to EA treatment alone. Our results also showed that some grafted MSCs were NT-3 immunopositive cells in the MSCs+EA group (data not shown). We propose that increased NT-3 content in the MSCs+EA treatment group is the result of a synergistic effect of EA treatment and MSC transplantation, as some studies have reported that transplanted MSCs can produce NT-3  or stimulate neuroglial cells to produce neurotrophic factors  in the central nervous system. MSCs may also secrete cytokines and growth factors such as NGF, BDNF, and VEGF [23, 53].
In this study, we found that the survival and differentiation of MSCs into neuron-like cells and oligodendrocytes were significantly increased in the MSCs+EA group, corresponding to a similar elevation in NT-3 and cAMP levels. This data suggests that EA promotes the survival and differentiation of grafted MSCs by elevating neurotrophic factors (such as NT-3) and the level of cAMP in injured spinal cord tissue. NT-3 is a significant member of the neurotrophic factors family and plays an important role during nervous system development, neuronal survival and differentiation, and neuronal repair via a signal transduction pathway . In particular, NT-3 can induce both the survival and proliferation of oligodendrocytes by differential involvement of the transcription factor CREB [55, 56]. In vivo studies of animal models have suggested that NT-3 may play an important role in regulating the quantity of oligodendrocytes and myelin regeneration following a CNS injury and demyelination [51, 57]. In addition, some studies have demonstrated that cAMP provides a powerful survival signal for neurons [42, 58].
However, several observations have raised questions regarding neuronal transdifferentiation of MSCs and suggested that the transdifferentiation was attributed to cell fusion with host cells. Several laboratories have proposed that MSCs fuse with the host cells, including the neurons, and acquire their phenotypes, which simulate their transdifferentiation into host cells [59–62]. While cell fusion could explain a good fraction of this apparent neural transdifferentiation of transplanted MSCs, there continues to be new evidence to suggest that, at least in some particular experimental paradigms, transdifferentiation does indeed occur [63–65]. Munoz-Elias et al.  transplanted adult rat bone marrow stromal cells into embryonic brains and observed that donor cells entered ventricular germinal zones, expressed neural progenitor traits, migrated to distant brain regions, and expressed site-specific neuronal proteins without cell fusion phenomena. Recently, Hokari et al.  found that bone marrow stromal cells might have the potential not only to differentiate into neurons, but also may fuse spontaneously with host neurons within 24 hr of cell-mixing coculture commencement. However, we never observed multiple nuclei originating from MSCs in the double-labeled cells during our study. Additionally, the NF150, β-tublin III, and 5-HT positive cells differentiated from grafted MSCs had few/no processes, which is not the typical morphology of host neurons. Furthermore, previous studies have suggested that this fusion is a rare event [59, 60]. Thus, we believe MSC-derived neuron-phenotypic cells were mostly transdifferentiated from donor cells, although we could not absolutely rule out host cell fusion. In addition, we found that the MSCs+EA group had more double-labeled neuron-phenotypic cells from MSCs than the MSCs group. We are not clear whether EA promotes grafted MSCs to transdifferentiate or fuse with host cells, so we will further study the effect of EA on the fate of grafted MSCs in the future.
We also found that 5-HT and CGRP-positive immunostained axons were regenerated in or across the lesion site into the caudal spinal cord at different degrees, most prominently in the MSCs combined with EA treatment group. It is known that scar formation by glia proliferation, lack of tropic support and inhibitory molecules  are key factors that block axons from regenerating into the injured spinal cord. However, several studies indicate that EA treatment can increase the tissue cAMP level and the expression of neurotrophic factors, such as NT-3, BDNF, NGF, and GDNF [68, 69]. Similarly, several studies [70, 71] have demonstrated that MSCs also secrete a variety of growth factors and cytokines, which can promote axonal growth in vitro as well as in vivo [43, 72, 73]. In addition, Yang et al. reported that EA treatment can inhibit the reactive proliferation of astrocytes after spinal cord injury and prevent the formation of a glial scar . Therefore, combining MSC transplantation and EA treatment may synergistically modify the hostile environment in the lesion site to promote axonal regrowth by increasing the amount of neurotrophic factors and the cAMP level and inhibiting glia scar formation.
The detection and behavioral analysis of spinal cord evoked potentials suggest that MSC transplantation combined with EA treatment efficiently improves neuronal function recovery. MSCs transdifferentiate into neurons and oligodendrocytes to replace the damaged or dead neural cells or repair myelin sheaths in the injured spinal cord, and increase the number of 5-HT-fibers passing through the lesion site into the caudal spinal tissue, which may be the morphological basis of the functional outcomes. However, whether or not these "neural cells" neurons were in fact functionally replaced damaged spinal cord neurons needs to be confirmed.