OUP user menu

Behavioral and Histological Evaluation of a Focal Cerebral Infarction Rat Model Transplanted With Neurons Induced from Bone Marrow Stromal Cells

Toshiro Mimura MD, PhD, Mari Dezawa MD, PhD, Hiroshi Kanno MD, PhD, Isao Yamamoto MD, PhD
DOI: http://dx.doi.org/10.1097/01.jnen.0000190068.03009.b5 1108-1117 First published online: 1 December 2005


Neurons can be specifically induced from bone marrow stromal cells (MSCs) with extremely high efficiency using gene transfection of the Notch intracellular domain and subsequent treatment with basic-fibroblast growth factor, forskolin, and ciliary neurotrophic factor. We investigated the behavioral and histologic efficacy of such bone marrow stromal cell-derived neuronal cell (MSDNC) transplantation into a focal cerebral infarction model in rats. A left middle cerebral artery occlusion (MCAO) was performed on adult Wistar rats. MSDNC transplantation into the ipsilateral hemisphere was performed on day 7 after MCAO. The behavioral analyses were conducted on days 14, 21, 28, 35, and 36-40, and a histologic evaluation was performed on day 41. MSDNC-transplanted rats showed significant recovery compared with controls (MCAO without cell transplantation) in beam balance, limb placing, and Morris water maze tests. Histologically, transplanted cells migrated from the injection site into the ischemic boundary area, including the cortex, corpus callosum, striatum, and hippocampus. Transplanted MSDNCs were positive for MAP-2 (84% ± 8.11%), whereas only a small number of cells were positive for GFAP (1.0% ± 0.23%). The survival rates of MSDNCs and MSCs 1 month after transplantation were approximately 45% and 10%, respectively. These results suggest that use of MSDNCs may be a promising therapeutic strategy for cerebral infarction.

Key Words
  • Behavioral analysis
  • Bone marrow stromal cell
  • Immunohistochemistry
  • Middle cerebral artery occlusion (MCAO)
  • Transplantation


In the central nervous system, neural tissue that is irreversibly damaged by ischemic injury enters the process of necrosis. Although intrinsic progenitor cells proliferate and differentiation takes place, these progenitor cells are insufficient for full functional recovery (1).

Neural cell transplantation into the ischemic brain has been reported in animal models, and amelioration of somatosensory and cognitive function has often been shown in these models (2, 3). However, currently, donor cells such as neural stem cells (NSCs) can only be supplied from the fetus forebrain or the subventricular zone of the adult brain. Additional problems include an inadequate supply of NSCs on a therapeutic scale and ethical concerns regarding their use. Furthermore, there is a problem of contamination by glial cells during neural differentiation.

Bone marrow stromal cells (MSCs) can differentiate into other cell types such as osteoblasts, adipocytes, chondrocytes, and cardiomyocytes (4-8). MSCs are promising candidates for clinical application because they are easy to isolate from bone marrow aspirates and are readily expanded in vitro, and they can be used for autotransplantation without posing major ethical problems. Several reports have shown that MSCs can differentiate into neural lineages; for example, neuron-like cell induction from MSCs with some accompanying induction of glial cells has been demonstrated in vitro (9-11). However, the induction efficiency of neurons is low, and functional recovery after transplantation of induced neural cells in animal models of neuronal disorder has not been tested.

Dezawa et al recently reported that neuronal cells can be specifically induced with extremely high efficiency from MSCs by introduction of the Notch intracellular domain followed by administration of basic-fibroblast growth factor, ciliary neurotrophic facto, and forskolin without the occurrence of glial differentiation (12). The MSC-derived neuronal cells (MSDNCs) expressed neuronal markers immunohistochemically and showed neuronal properties in an electrophysiological analysis. Transplantation of these MSDNCs resulted in improvement in apomorphine-induced rotational behavior after intrastriatal implantation in a 6-hydroxydopamine rat model of Parkinson disease (12).

In this study, we induced MSDNCs from adult rat MSCs and transplanted the cells into the infarct area in a middle cerebral artery occlusion (MCAO) rat model. Behavioral tests and histologic assessment were conducted after transplantation. Our results show that MSDNCs are effective in the amelioration of damage in an ischemic rat model.

Materials and Methods

All experimental procedures were approved by the Animal Experimentation Ethics Committee of Yokohama City University School of Medicine.

Culturing of Marrow Stromal Cells and Neuronal Induction

MSCs were isolated from 8-week-old Wistar rats as described previously (13). The MSCs were labeled with green fluorescent protein (GFP) by retroviral infection using the pBabe neo-GFP vector (14).

Neuronal induction from MSCs has been described by Dezawa et al (12). Briefly, MSCs subcultured four times were transfected with a vector (pCI neo-NICD) containing the Notch1 intracellular domain (NICD) using Lipofectamine 2000 (Invitrogen Corp., Carlsbad, CA). Cells were selected by G418 treatment for 11 days. For induction of MSDNCs, NICD-transfected MSCs were subcultured once (60-70% confluence) and were incubated in alpha-MEM containing 10% FBS, 5 μM forskolin (Calbiochem, La Jolla, CA), 10 ng/mL basic-fibroblast growth factor (Peprotech, London, U.K.), and 10 ng/mL ciliary neurotrophic factor (R&D Systems, Minneapolis, MN). Five days later, cells were transplanted into the MCAO rat model. To characterize the induced MSDNCs in vitro, we performed immunocytochemistry. Anti-MAP-2ab (Sigma, St. Louis, MO), neurofilament-M (NF-M) (Chemicon, Temecula, CA) and beta-tubulin isotype-3 (β-tubulin-3) (Sigma) were used as neuronal markers.

Middle Cerebral Artery Occlusion Rat Model

Male Wistar rats weighing 200 to 250 g were kept at room temperature (24°C) with a 12-hour light-dark cycle and were given food and water freely. The MCAO procedure was a modification of the methods of Koizumi (15) and Longa (16). Briefly, under deep anesthesia induced by a mixture of 1.0% to 1.5% halothane, 10% O2, and air, a midline cervical incision was performed and the left carotid bifurcation was identified. A probe made of 4-0 nylon thread with a silicon rubber-coated head of diameter 0.3 mm was inserted into the ligated external carotid artery and advanced into the internal carotid artery to a position 16 to 18 mm from the bifurcation. During the surgery, rectal temperature was maintained between 37.5°C and 38°C using a feedback heating pad (BWT-100; Bio Research Center Co. Ltd., Tokyo, Japan). Arterial blood gas analysis was performed and p02 was maintained at 85 to 120 mm Hg through control of the anesthetic device. Reperfusion was performed 4 hours after the occlusion through a 10-mm withdrawal of the probe.


On day 7 after MCAO, rats were anesthetized with intraperitoneal injection of 50 mg/kg sodium pentobarbital and placed onto a stereotaxic frame. In a preliminary experiment, the infarct area was produced in the lateral area from approximately 3.5 mm lateral to the midline. For transplantation into the nonnecrotic brain parenchyma, the cell suspension, composed of 8,000 to 16,000 cultured cells in 3 μL of phosphate-buffered saline (PBS, pH 7.4), was stereotaxically injected into the left forebrain from the following three locations: +2 mm, 0 mm, and -2 mm anterior to the bregma, and 2 mm lateral to the midline, and at 1.2 mm depth from the cortical surface in each case. The total number of transplanted cells was 24,000 to 48,000.

Three groups of animals were prepared: 1) the control group, which received only PBS (without cell transplantation) (n = 7); 2) the MSC group, which underwent transplantation of nontreated intact MSCs (n = 10); and 3) the N-MSC group, into which MSDNCs were transplanted (n = 10).

Behavioral Testing

The severity of neurologic damage was evaluated using the following tests: a beam balance test, a limb placing test, and a Morris water maze test. The beam balance test and the limb placing test were performed on days 7 (just before transplantation), 14, 21, 28, and 35 after MCAO. The Morris water maze test was performed from days 36 to 40.

Beam Balance Test

The beam balance test is used to assess gross vestibulomotor function and was carried out as described previously (17). Scoring was based on the following criteria: balancing with a steady posture with paws on the top of the beam: a score of zero; grasping the sides of the beam and/or shaky movement: one; one or more paw(s) slipping off the beam: 2; attempting to balance on the beam, but falling off: 3; and falling off the beam with no attempt to balance or hang on: 4.

Limb Placing Test

The limb placing test examines sensorimotor integration in limb placing responses to visual, tactile, and proprioceptive stimuli and was performed as described previously (18). A proprioceptive adduction test (18) was also performed. For each test, scoring was based on the following criteria: immediate and complete placing of the limb: a score of zero; incomplete and/or delayed (>2 seconds) placing, including interspersed flailing: one; and no placing: 2. Visual, forward tactile, and lateral tactile proprioceptive stimuli were given to the right forelimb. Forward tactile and lateral tactile and proprioceptive stimuli were given to the right hindlimb. Proprioceptive adduction tests were performed for both forelimb and hindlimb. The total score ranged from zero to 18.

Morris Water Maze Test

The Morris water maze test is a useful method to assess cognitive function (19, 20). Several modifications of this test have been reported; we refer readers to the work of Fukunaga et al (20). This test was performed from day 36 to day 40 after MCAO. A pool (diameter 150 cm, depth 35 cm) was prepared with an escape platform (diameter 10 cm) located 1 cm beneath the surface of the water, which was rendered opaque and milky white. Four starting points around the edge of the pool were designated as N, E, S, and W. The platform was kept in the middle of a particular quadrant, equidistant from the center and the edge of the pool. A rat was released into the water from each starting point and allowed to swim until reaching the platform, and the time taken to reach the platform was recorded (maximum of 120 seconds). Rats were trained in the task using 2 sets of 4 trials on each of 5 consecutive days. After the first set on the fifth day, instead of the second set, a spatial probe trial was performed. This test is used to estimate short-term memory retention. The platform was removed and the rat was allowed to swim for 60 seconds. The number of times each animal crossed the area in which the platform had previously been located was measured (20). The time spent in the quadrant in which the platform had previously been located was also measured.

Histologic Analysis

The total period to finish the behavioral evaluation took one month after the cell transplantation. We intended to perform the histologic evaluation approximately one month after the transplantation. Therefore, we performed the histologic evaluation on day 41 from MCAO (34 days after the transplantation).

The rats were killed by administration of a pentobarbital overdose and then perfused transcardially with 0.9% saline followed by periodate-lysine-paraformaldehyde fixative solution (21). The brain was cut into coronal blocks of 2-mm thickness using Brain Matrix (BAS, Inc., Warwickshire, U.K.), and 10-μm-thick cryostat sections were made from each block. Sections were stained with hematoxylin and eosin to evaluate the infarct area. The images of sections were captured under a light microscope and the lesion areas were traced using Scion Image (Scion Corp., Frederick, MD). The infarct volume was calculated as described previously (22) and expressed as a percentage of the volume of the contralateral hemisphere.

For immunostaining, 14- to 16-μm cryosections were incubated with primary antibodies to MAP-2 (1:100; Boehringer Mannheim, Germany), β-tubulin-3 (1:400; Sigma), NF-M (1:200; Boehringer Mannheim), Tuj-1 (1:100; BAbCO, CA), or GFAP (1:400; Dako, Carpinteria, CA) at 4°C overnight. Alexa Fluor 546-conjugated anti-mouse IgG (Molecular Probes, Eugene, OR) (for MAP-2) or anti-rabbit IgG (Molecular Probes) (for β-tubulin-3, NF-M, Tuj-1, and GFAP) was used as the secondary antibody. TOTO-3 was used for nuclear staining. Specimens were inspected using confocal laser scanning microscopy (Radiance 2000; Bio-Rad, Hertfordshire, U.K.).

In each rat, the total number of GFP-labeled cells in the forebrain and hippocampus was determined. Ten sequential slides for each block were obtained and the mean number of GFP-labeled cells in each slide was measured. Because the approximate diameter of cell body of neuron is 20 μm (the cells less than 20 μm in diameter were excluded from the counting), we estimated the number of GFP-labeled cells in 2-mm-thick blocks could be obtained by multiplying the mean number of them by 100. The total number of GFP-labeled cells in the whole forebrain was then calculated by adding the numbers of GFP-labeled cells from all blocks. The number of GFP-labeled cells in the hippocampus was counted in the same way.

Statistical Analysis

The behavioral evaluation score and infarct volume are parametric data. For the statistical analysis among 3 groups, nonrepeated-measures analysis of variance was used. When the results were significant (p < 0.05), a Student-Newman-Keuls post hoc procedure was used at a 95% significant level using statistical software (Statistica StatSoft, Inc., Tulsa, OK). The values are presented as means ± standard deviation, unless otherwise stated.


The efficiency of retrovirus GFP introduction into MSCs was 98.0% ± 0.7%. The efficiency of NICD transfection was assessed by lipofection of pNICD-IRES2-EGFP, a GFP-containing plasmid, showing that 98.5% ± 0.8% of cells were transfected with NICD after G418 selection. MSDNCs induced from MSCs in vitro were subjected to the immunohistochemistry, showing that approximately 95% of GFP-labeled MSDNCs showed NF-M positivity and neurite extension consistent with a previous report (12) (Fig. 1). MAP-2ab and β-tubulin-3 showed similar results (data not shown).


Confocal laser scanning microscopic images in vitro. Green fluorescent protein-labeled marrow stromal cell-derived neuronal cells (B, D) showed neurofilament-M positivity (A, C). Panels (A) and (B) are in lower magnification; panels (C) and (D) in higher magnification. Scale bars = (A, B) 50 μm; (C, D) 10 μm.

One week after MCAO, and just before transplantation, severe right-sided neurologic deficits were apparent and the mean score for each behavioral test showed no statistical differences among the three groups. The results of behavioral analyses in the posttransplantation periods are summarized in Table 1 and Figures 2-5.

View this table:

Beam balance test. On day 28 after transplantation, the mean score for the neuron induced from marrow stromal cell (N-MSC) group showed a significant improvement compared with the marrow stromal cell and control groups. *, p < 0.05, **, p < 0.01.


Limb placing test. The mean scores for the neuron induced from marrow stromal cell (N-MSC) group and the marrow stromal cell (MSC) group were significantly different to that of the control group on day 28 and day 35. There was no significant difference between the N-MSC and MSC groups. *, p < 0.05, **, p < 0.01.


Morris water maze test. For the final set, the mean latency time for the neuron induced from marrow stromal cell (N-MSC) group was significantly different from those for the marrow stromal cell and control groups. *, p < 0.05.


Water maze “spatial probe trial.” Upper panel: the number of times the rat swam through the area in which the platform had previously been located. Lower panel: the time spent in the quadrant in which the platform had been located. Among the 3 groups, the best results were obtained for the neuron induced from marrow stromal cell (N-MSC) group, and statistical differences were obtained between the N-MSC group and the other groups. *, p < 0.05.

Beam Balance Test

From day 7 to day 21, the mean score was not statistically different among the 3 groups. On days 28 and 35, the mean score of the N-MSC group showed a significant improvement compared with the control (day 28: p = 0.0041, and day 35: p = 0.0001) and MSC groups (p = 0.0471 and 0.0007, respectively). Although the MSC group showed a slight improvement compared with the control group, a statistically significant difference could not be detected on days 28 and 35 (Table 1; Fig. 2).

Limb Placing Test

There were no statistical differences among the mean scores of the 3 groups from day 7 to day 21. On days 28 and 35, the N-MSC and MSC groups showed a significant improvement compared with the control group (day 28: p = 0.0011 and 0.0434, respectively, and day 35: p = 0.0002 and 0.0019, respectively). However, the mean score showed no significant difference between the N-MSC and MSC groups throughout the entire period (Table 1; Fig. 3).

Morris Water Maze Test

The mean latency time recorded in each set of 4 trials for location of the submerged escape platform is shown in Figure 4 for each of the 3 groups. The N-MSC group showed a shorter latency time than the control and MSC groups. The mean latency time for the second set on day 39 and the first set on day 40 demonstrated a significant difference between the N-MSC and control groups (p = 0.0339 and 0.0492, respectively) (Table 1; Fig. 4). Although the N-MSC group showed a tendency for shorter times to find the escape platform compared with the MSC group, there was no statistically significant difference between the groups.

In the spatial probe trial, rats in the N-MSC group showed improvement compared with the control and MSC groups (p = 0.0419 and p = 0.0453, respectively) (Table 1; Fig. 5A). The mean time spent in the quadrant in which the platform had been located was longer in the N-MSC group compared with the control and MSC groups, and a statistical difference existed between the N-MSC and control groups (p = 0.0339) (Table 1; Fig. 5B).

Histologic Study

The infarct area was located in the lateral half of the left hemisphere, including the cortex, striatum, and hippocampus, and formation of cysts and scars was observed in many of the lesioned brains (Fig. 6). The hippocampus on the lesion side was atrophic and showed a partially irregular arrangement or loss of neurons compared with the contralateral side. Infarct volumes were measured in all MCAO models. The mean infarct volumes in the N-MSC, MSC, and control groups on day 33 were 50.7% ± 10.9%, 51.0% ± 10.2%, and 50.9% ± 11.1%, respectively. There was no statistically significant difference among the 3 groups.


The drawing shows an image of the brain viewed from the upper side. The infarct area is shown in gray. Crossing lines are positioned 4, 6, and 8 mm from the pole of the frontal lobe. Color images show stained coronal sections of an infarcted brain from the neuron induced from marrow stromal cell (N-MSC) group (day 41). Arrowheads indicate the grafted sites on day 7. The characters and arrows correspond to the area indicated in Figure 7. Formation of necrotic cysts or scars was observed.


Histology of the infarcted brain in the neuron induced from marrow stromal cell (N-MSC) group on day 41. (A-C) Light microscopic images (low magnification) stained with hematoxylin and eosin. (D-F) Fluorescent microscopic images of the same areas shown in panels (A-C), respectively. Green fluorescent protein (GFP)-labeled cells were observed in panels (D-F). Scale bars = 200 μm. Squares correspond to the following LSCM images. (G-N) Tuj-1, MAP-2, β-tubulin-3, neurofilament-M (NF-M), and GFAP were labeled with Alexa546 (red color-coded), and nuclei were labeled with TOTO-3 (blue color-coded). Scale bars = 20 μm. (G) Transplanted marrow stromal cell-derived neuronal cells (MSDNCs) cell body and neurites were recognized as GFP-positive signals (arrows). (H) MAP-2-positive GFP-labeled MSDNCs were detected (arrows). (I, J) GFP-labeled cells were immunopositive for Tuj-1 (I) and β-tubulin-3 (J). Arrows indicate GFP-positive cell bodies and neurites that were simultaneously positive for Tuj-1 and β-tubulin-3. (K) GFP-labeled cells expressing MAP-2 (arrows) were detected in the hippocampus of the ipsilateral hemisphere. (L, M) In the hippocampus of the ipsilateral hemisphere, GFP-labeled cells were immunoreactive to NF-M (arrows and arrowheads), and neurite extension was observed in these cells. (N) A small number of GFP-labeled cells showed GFAP positively.

Transplanted GFP-labeled MSCs and MSDNCs were located mainly at the boundary between intact tissue and the infarct area, including the ipsilateral cortex, corpus callosum, striatum, and hippocampus (Fig. 7A-F). Infiltration of inflammatory cells into the infarct focus was observed; however, the amount of inflammatory cells in the N-MSC and MSC groups was not significantly different. In the necrotic cyst in the ischemic hemisphere, many large GFP-labeled cells, which had large or multinuclei and had no neurites were observed. Because these cells appeared to be microglia and/or macrophages but not neurons or glial cells, we excluded these cells from the cell count.

Some of confocal microscopic images showed GFP signals as vesicular appearance (Fig. 7 I, J, L, M). This is probably because GFP was introduced as retrovirus and the fluorescence was not homogenously intense as usually observed in transgenic animal-derived cells. In fact, in our previous report (14), we demonstrated that GFP introduced with the retrovirus system in MSCs showed immunoelectron microscopic localization of GFP in association with membranes and vesicles in the cytoplasm, occasionally observed as patchy.

A large number of GFP-labeled MSDNCs were immunopositive for MAP-2 and showed neurite development in the host brain (Fig. 7G, H). These cells were also immunopositive for Tuj-1 and β-tubulin-3 (Fig. 7I, J). In the ipsilateral hippocampus, many cell bodies and neurites of GFP-labeled MSDNCs were also NF-M-positive (Fig. 7K-M). A large fraction of GFP-labeled transplanted MSDNCs were positive for MAP-2 (84.0% ± 8.1%), whereas only a small number of cells were positive for GFAP (1.0% ± 0.2%) (Fig. 7N).

In contrast, the large majority of MSCs in the host brain were negative for both neuronal (MAP-2, Tuj-1, β-tubulin-3, and NF-M) and glial (GFAP) markers (data not shown). The percentages of MAP-2- and GFAP-positive cells among the GFP-labeled cells were 1.4% ± 0.2% and 4.8% ± 1.0%, respectively. However, formation of neurites in MAP-2-positive MSCs could not be found.

The number of MSDNC in the host forebrain one month after transplantation ranged from 11,920 to 14,990 (the mean was 13,250 ± 1,126). The number of MSCs in the host forebrain ranged from 4,230 to 7,560 (the mean was 5,850 ± 997). Approximately 30% to 45% of the transplanted MSDNCs were detected, whereas only 10% to 20% of transplanted MSCs were detected. Thus, the survival rate of MSDNCs in the ischemic brain was substantially higher than that for MSCs. In the hippocampus, the number of MSDNC ranged from 610 to 980 (the mean was 790 ± 160) and 89% of these cells were MAP-2-positive. The number of MSCs, in contrast, ranged from 380 to 554 (the mean was 470 ± 66) and only 0.6% was MAP-2-positive, showing that most transplanted MSCs were negative for both neuronal and glial markers. In all 3 groups, no tumor formation in the brain parenchyma was observed up to 41 days after MCAO.


The N-MSC group showed significant improvements in the behavioral assessment tests compared with the control and MSC groups. In histologic analysis, the infarct volume determined at 41 days after MCAO did not show a significant difference among the 3 groups, suggesting that both MSDNC and MSC transplantation performed 7 days after MCAO had little effect on tissue protection. However, compared with MSCs, transplanted MSDNCs demonstrated a higher survival rate in the host brain and a larger proportion of MSDNCs were positive for neuronal markers, which might have contributed to the improvements in the behavioral analysis. These results suggest that MSDNCs induced in vitro followed by transplantation is more effective than direct transplantation of nontreated MSCs.

As to the Notch protein expression in MSDNCs, NICD expression, which was clearly positive after transfection, gradually decreased according to the time course after neuronal induction, both in Western blot and reverse transcriptase-polymerase chain reaction (data not shown), is consistent with our previous report (12). This is probably the result of the degradation of introduced NICD, because NICD was introduced as plasmid DNA. These results suggest that NICD is necessary for the triggering neuronal induction in MSCs; however, once MSCs have acquired neuron-like characteristics, NICD action is no longer necessary for maintaining their characteristics.

Notch-Hes signaling is known to inhibit neuronal differentiation and instructs glial differentiation in conventional neural development (23). However, in our system, NICD introduction accelerated the induction of neuronal cells from MSCs. Although our results appear inconsistent with previous work, they do not refute the known role of Notch-Hes signals in neurogenesis. In our previous report, JAK/STAT inhibitor administration and constitutive active STAT1/3 transfection showed that downregulation of STATs was tightly associated with NICD-mediated neuronal induction, whereas Hes, downstream of Notch, was not involved in the induction event (12). Thus, our results suggest distinct cellular responses to Notch signals; for example, the repertoire of second messengers and active factors may well be different between conventional neural stem cells and/or neural progenitor cells and MSCs. It might be possible that unknown signaling pathways downstream of Notch may be involved in these events, and thus further studies are needed to identify the factor(s) involved in this phenomenon.

The MSC group demonstrated slight improvements in behavioral assessment tests compared with the control group, but these improvements were not as significant as those for the N-MSC group. According to recent reports, MSC transplantation into the ischemic brain does lead to improvement of behavioral outcome. However, the survival rate of MSCs in the host brain was below 10%, and the proportion showing neuronal differentiation was only 1% to 2% (24-26). The functional recovery in behavior after MSC transplantation is thought to be mediated by neurotrophic factors produced by MSCs or by intrinsic parenchymal cells stimulated by MSCs (24-26).

In the current study, some MSDNCs in the cortex, striatum, and hippocampus demonstrated neurite extension in the host brain, which was not observed in the MSC-transplanted rats. Hence, the significant behavioral improvements in the N-MSC group suggest the transplanted MSDNCs maintained neuronal characteristics in the host brain and contributed to functional recovery in the MCAO rat model. However, positive effects associated with trophic factors produced by transplanted MSDNCs may also be present.

The potential of other kinds of stem cells such as NSCs and umbilical cord blood cells for use in ischemic brain injury has been reported recently (27, 28). Toda et al reported that engraftment of NSCs into a hippocampus damaged by transient global ischemia results in the partial improvement of impaired spatial learning in a water maze test (27). However, only 1% to 3% of the grafted cells survived and only 3% to 9% of these expressed NeuN. Chen et al reported that intravenous administration of human umbilical cord blood cells (HUCBCs) reduces behavioral deficits after stroke in rats (28). HUCBCs contain many stem and progenitor cells, although the percentage of HUCBCs that expressed NeuN, MAP-2, and GFAP was ≤2%, ≤3%, and ≤6%, respectively. In addition, approximately only 1% to 2% of intravenously injected HUCBCs were detected in the brain. In our study, 30% to 45% of MSDNCs survived in the host brain in the N-MSC group one month after transplantation and 84% of these cells expressed neuronal markers. In our experiment, neuronal differentiation itself seemed to contribute to the cell survival in the injured brain. Although because MSDNCs had received NICD transfection followed by the G418 selection, these manipulations might select a more robustly surviving subgroup from the original culture of MSCs. In addition, by using our retrovirus introduction system, GFP expression appeared to be maintained for a longer period after transplantation. These factors might have contributed to the higher survival rate in our study. Despite these facts, both NSCs and HUCBCs are very promising sources for donor cells. However, NSCs must be harvested from adult or fetal brain, which may have associated ethical problems, and long periods, on a clinical scale, are required for proliferation of NSCs. Although HUCBCs are widely available and have already been used clinically, further investigation is necessary to assess their efficacy for ischemic stroke.

In comparison to these cell types, MSCs are good candidates in that they can easily be obtained from patients or a bone marrow bank and they can be expanded in culture with fewer ethical problems. Furthermore, autologous transplantation of MSCs or transplantation of MSCs with the same HLA subtype from a healthy donor may minimize the risk of rejection. Aging is associated with a decreased maximal lifespan and accelerates senescence of MSCs, so that the MSDNC induction ratio may be variable according to the age of source. We estimated the influence of age (6-, 8-, and 12-week-old rats) in the MSDNCs induction. The ratio was basically similar among these ages and statistical differences could not be recognized (data not shown). Nonetheless, aging should be taken into account as one of the factors affecting transplantation.

For the clinical application of the cell transplantation therapy for the injured brain, transvenous or intrathecal cell injection is preferable because it is less invasive. In the current study, we performed direct intraparenchymal injection of MSDNCs because other methods include the uncertainty of cell distribution into the ischemic brain, which influence the evaluation of donor cell's survival rate.

The results presented here suggest that our method of specific induction of neuronal cells from MSCs has great potential in MSC transplantation therapy for neurologic disorders. However, further studies are needed to ensure the long-term safety and efficacy of manipulated MSCs.


  • This study was supported by a Grant-in-Aid (no.13557120, no.15590166, and no.15016091) from the Ministry of Education, Science and Culture of Japan.


  1. 1.
  2. 2.
  3. 3.
  4. 4.
  5. 5.
  6. 6.
  7. 7.
  8. 8.
  9. 9.
  10. 10.
  11. 11.
  12. 12.
  13. 13.
  14. 14.
  15. 15.
  16. 16.
  17. 17.
  18. 18.
  19. 19.
  20. 20.
  21. 21.
  22. 22.
  23. 23.
  24. 24.
  25. 25.
  26. 26.
  27. 27.
  28. 28.
View Abstract