Therapeutic Effect of BDNF-Overexpressing Human Neural Stem Cells (HB1.F3.BDNF) in a Rodent Model of Middle Cerebral Artery Occlusion

Abstract

Ischemic stroke mainly caused by middle cerebral artery occlusion (MCAo) represents the major type of stroke; however, there are still very limited therapeutic options for the stroke-damaged patients. In this study, we evaluated the neurogenic and therapeutic potentials of human neural stem cells (NSCs) overexpressing brain-derived neurotrophic factor (HB1.F3.BDNF) following transplantation into a rodent model of MCAo. F3.BDNF human NSCs (F3.BDNF) were transplanted into the contralateral side of striatum at 7 days after MCAo, and the transplanted animals were monitored up to 8 weeks using animal MRI and various behavioral tests before they were sacrificed for immunohistochemical analysis. Interestingly, animal MRI results indicate that the majority of contralaterally transplanted neural stem cells were migrated to the peri-infarct area, showing a pathotropism. Transplanted animals exhibited significant behavioral improvements in stepping, rotarod, and modified neurological severity score (mNSS) tests. We also found that the transplanted human cells were colocalized with nestin, DCX, MAP2, DARPP-32, TH, GAD65/67-positive cells, of which results can be correlated with neural regeneration and behavioral recovery in the transplanted animals. More importantly, we were able to detect high levels of human BDNF protein expression, presumably derived from the transplanted F3.BDNF. Taken together, these results provide strong evidence that human neural stem cells (F3.BDNF) are effective in treating stroke animal models.

Introduction

Ischemic stroke is caused by blockade of the blood supply to part of the brain, which causes the death of multiple neuronal cell types, as well as astrocytes, oligodenderocytes, and endothelial cells in the brain (27,39). As a result, most patients suffer from persistent motor, sensory, or cognitive impairments after stroke, and there are still very limited therapeutic options for stroke-damaged patients. Neural stem cells (NSCs) or neural precursor cells (NPCs) are known to survive and ameliorate neurological deficits when they are engrafted in animal models of various neurological diseases (17,26,27). We have previously generated a v-myc immortalized clonal human neural stem cell line (HB1.F3), which was shown to differentiate into neurons and glial cells (7,18) and to improve neurological functions when transplanted into animal models of rat focal ischemia (8,9) and mouse intracerebral hemorrhage (ICH) (14,22). However, due to low survival rate of grafted F3 NSCs (22), functional outcome was insufficient. To overcome this problem, we have recently generated another cell line overexpressing human BDNF (brain-derived neurotrophic factor) (23).

BDNF is a member of the neurotrophin growth factor family that consists of nerve growth factor (NGF), neurotrophin 3 (NT-3), and neurotrophins 4/5 (NT-4/5) (12). It is known that ischemic insult to the brain induces a marked increase in BDNF expression, as a part of the neuroprotective response (2,24). It has also been shown that intravenously administering BDNF prior to focal ischemia injury (36), intracerebral infusion of BDNF following ischemia (38), or transplantation of mesenchymal stem cells overexpressing BDNF (21,31) all resulted in a significant reduction of the infarct volume and/or improvement of behavioral recovery. Therefore, it is highly conceivable that BDNF overexpression will greatly improve the existing limitations of F3 NSCs.

As expected, we found that F3.BDNF cells produce six times more BDNF protein, compared with F3 cells alone (23). Moreover, we also observed graft survival and behavioral recovery were significantly improved when F3.BDNF cells were transplanted into a mouse model of ICH, compared with F3 cell alone-transplanted group (23). In the present study, we extended our previous results using a mouse ICH model to a rodent model of ischemic stroke (middle cerebral artery occlusion, MCAo), with different perspectives. In particular, we employed animal magnetic resonance imaging (MRI) in order to track the fates of F3.BDNF cells following transplantation into the contralateral side of infarct region.

Materials and Methods

Culture and Differentiation of Human Neural Stem Cells Overexpressing BDNF (F3.BDNF)

Our immortalized human neural stem cells HB1.F3 (F3) and BDNF-overexpressing F3 (F3.BDNF) cells were cultured and maintained as described previously (7,18,23). Briefly, they were grown in 0.2% gelatin-coated tissue culture dishes (BD Biosciences, San Jose, CA, USA) containing Dulbecco's modified Eagle's medium (DMEM) with high glucose containing l-glutamine, supplemented with 10% FBS, 100 U/ml penicillin, and 100 mg/ml streptomycin (all from Invitrogen, Carlsbad, CA, USA). To induce neural differentiation, F3.BDNF cells were allowed to form neurospheres in suspension culture for 6 days (early differentiation period) in neural differentiation medium containing DMEM/F12 (Invitrogen), 100 U/ml penicillin, 100 mg/ml streptomycin, 1% nonessential amino acid (0.1 mM, Invitrogen), 0.01% β-mercaptoethanol (0.1 mM, Invitrogen), N2 supplements (Invitrogen), and 0.2 mM ascorbic acid (Sigma, St. Louis, MO, USA). Once neurospheres were formed, they were attached to tissue culture dishes (BD Biosciences) that were precoated with poly-l-ornithine (15 mg/ml, Sigma) and fibronectin (1 mg/ml, Sigma) and were maintained in neural differentiation medium supplemented with 20 ng/ml BDNF (R&D Systems, Minneapolis, MN, USA) for an additional 6 days (late differentiation period).

RNA Isolation and Reverse Transcription-Polymerase Chain Reaction (RT-PCR) Analysis

Total RNA was isolated from cells using Trizol RNA extraction method (Gibco, Gaithersburg, MD, USA). We synthesized cDNA using Moloney Murine Leukemia Virus (M-MLV) reverse transcriptase (Promega, Madison, WI, USA) at 42°C for 1 h. PCR amplification was carried out using Taq polymerase according to the manufacturer's instructions (Intron Biotechnology, Kyungki-Do, Korea). Primer sequences, number of cycles, and annealing temperatures are described in Table 1. Glyceraldehyde-3-phosphate dehydrogenase (GAPDH), a housekeeping gene, was used as an internal loading control.

Table 1.

RT-PCR Primers Used in This Study

Gene Name	Forward (F) and Reverse (R) Primer Sequences	Product Size (bp)	Annealing Temp. (°C)	No. Cycles	GenBank Accession No.
Human BDNF	F: 5′-TGGCTGACACTTTCGAACAC-3′	391	57	30	M61176.1
	R: 5′-ATGGGATTGCACTTGGTCTC-3′
Sox2	F: 5′-GCTGCAAAAGAGAACACCAA-3′	232	58	30	NM_003106
	R: 5′-CTTCCTGCAAAGCTCCTACC-3′
Nestin	F: 5′-TCCAGAAACTCAAGCACCA-3′	183	59	30	NM_006617
	R: 5′-AAATTCTCCAGGTTCCATGC-3′
Musashi	F: 5′-ACAGCCCAAGATGGTGACTC-3′	191	59	30	NM_002442
	R: 5′-CCACGATGTCCTCACTCTCA-3′
NCAM	F: 5′-GATGCCGGAGAGTACATCTG-3′	385	55	30	NM_000615.6
	R: 5′-GGTCGATGGATGGTGAAGAG-3′
III β-tubulin	F: 5′-ATGAGGGAGATCGTGCACAT-3′	239	59	30	BC000748.2
	R: 5′-GCCCCTGAGCGGACACTGT-3′
Map2	F: 5′-AGAGGCATCTGCACACTCAC-3′	171	59	30	U01828
	R: 5′-CCGTTGATCCCATTCTCTTT-3′
Darpp-32	F: 5′-GAGCTTCGGGAGCTGGGTTA-3′	233	55	30	AF464196.1
	R: 5′-AGTGCTGGGTCTTCCACTTG-3′
GFAP	F: 5′-GAAGCTGCTAGAGGGCGAGG-3′	207	55	30	BC041765.1
	R: 5′-GTCCTGCCTCACATCACATC A-3′
GAPDH	F: 5′-GTCATACCAGGAAATGAGCT-3′	422	60	30	BC083511.1
	R: 5′-TGACCACAGTCCATGCCATC-3′

BDNF, brain-derived neurotrophic factor; SOX2, sex-determining region Y box 2; NCAM, neural cell adhesion molecule; Map2, microtubule-associated protein 2; Darpp-32, dopamine- and cAMP-regulated phosphoprotein, Mr 32 kDa; GFAP, glial fibrillary acidic protein; GAPDH, glyceraldehyde 3-phosphate dehydrogenase.

Western Blot Analysis

Proteins were separated by 12% sodium dodecyl sulfate polyacrylamide gel electrophoresis (SDS-PAGE; Bio-Rad, Hercules, CA, USA) and transferred to a membrane (Immobilon-P; Millipore, Billerica, MA, USA). The membranes were blocked in blocking solution (Tris-buffered saline, TBS; Abelbio, Seoul, Korea) containing 0.1% (v/v) Tween-20 (Sigma) and 5% (v/v) nonfat skim milk (BD Biosciences) for 1 h and then incubated overnight with antibodies against human BDNF (1:1,000, R&D Systems) and β-actin (1:1,000, Santa Cruz Biotechnology, Dallas, TX, USA). The membranes were washed extensively in TBS containing 0.1% (v/v) Tween-20 before incubation with a secondary anti-mouse (1:1,000) or anti-rabbit (1:3,000) antibody conjugated to horseradish peroxidase (Millipore) for 1 h. After washing with PBS, protein was then visualized using ECL solution (GE Healthcare, Little Chalfont, Buckinghamshire, UK) and X-ray film.

Immunocytochemical Analysis

Immunocytochemical analyses were performed to characterize the marker expression in undifferentiated F3.BDNF cells and their neuronally differentiated cell types. The following primary antibodies were used: human-specific BDNF (1:200, R&D Systems), human-specific nestin (1:200, Chemicon, Temecula, CA, USA), sex-determining region Y box 2 (SOX2; 1:200, Chemicon), type III β-tubulin (Tuj1) (1:500, Chemicon), microtubule-associated protein 2 (MAP2; 1:200, Chemicon), and glial fibrillary acidic protein (GFAP; 1:500, BD Biosciences). Secondary antibodies used were goat anti-mouse IgG-conjugated Alexa-555 (1:200, Molecular Probes, Eugene, OR, USA), goat anti-rabbit IgG-conjugated Alexa-488 (1:200, Molecular Probes), and goat anti-mouse IgM-conjugated Alexa-555 (1:200, Molecular Probes). After washing three times in PBS, cells were counterstained with 4′,6-diamidino-2-phenylindole (DAPI; 1:1,000, Roche, Indianapolis, IN, USA). A confocal laser-scanning microscope imaging system (LSM510, Carl Zeiss, Inc., Thornwood, NY, USA) was used to study the staining patterns and take photographs.

MCAo Animal Model and Cell Transplantation

All animal experiments were carried out in accordance with the CHA University IACUC (Institutional Animal Care and Use Committee; IACUC090012). We induced MCAo experimently according to the method of Longa (28) using adult male Sprague-Dawley rats (Orient, Seoul, Korea), weighing 270-300 g. We inserted a blunt-ended monofilament (4-0, Ethicon, Livingston, Scotland, UK) into the internal carotid artery approximately 19 mm from the carotid bifurcation in order to occlude the middle cerebral artery (MCA) for 90 min, after which the monofilament was carefully removed. Acute neurological assessment during the first 24 h after 90 min of MCAo, which includes forelimb and hindlimb placement (10) and circling behavior (5), was employed to determine and select the appropriate stroke animal models

Animals which scored 3 points or lower (normal: 5 points) were selected for transplantation experiments. Seven days after MCAo, eight rats were injected with 2 μl of F3.BDNF cells (200,000 cells/μl) into the contralateral side of the infarct region using a Hamilton syringe (Reno, NV, USA), targeting the following coordinates: AP +1.0 mm, ML +3.0 mm, and DV −5.0 mm from bregma. In the control group (n = 7), 2 ml of vehicle (DMEM) was injected in parallel. In order to carry out animal MRI experiments, the same cells were labeled with ferumoxide (Feridex^®; Taejoon Pharma, Seoul, Korea)-protamine sulfate (Sigma) complexes, as described previously (16). Briefly, 10 μg/ml ferumoxide and protamine sulfate in DMEM (Dulbecco's modified Eagle's medium, Invitrogen) without serum were prepared and mixed for 30 min at room temperature. An equal volume was added to the cells, and they were incubated overnight at 37°C. To immunosuppress the transplanted animals, cyclosporine A (10 mg/kg, Sigma) was administered intraperitoneally 24 h before transplantation and then daily for up to 8 weeks.

Behavioral Tests

To determine the validity of the MCAo model, as well as the effect of stem cell transplantation, we performed rotarod test (14), stepping test (32), and modified neurological severity score (mNSS) test (34) every week and apomorphine-induced rotation test (37) every 2 weeks following transplantation. Baseline scores for each test were obtained 1-3 days before transplantation.

Statistical Analysis

Statistical analysis on the behavior data was performed using the Statistical Analysis System (Version 9.1.3; SAS, Seoul, Korea) on a CHA University mainframe computer. All values were presented as means ± SD (standard deviation). Statistical significance was analyzed using one-way ANOVA (analysis of variance), followed by Fisher's least significant difference (lSD) test as a post hoc test. A value of p < 0.05 was regarded as significant, and p < 0.001 was regarded as highly significant.

MRI Detection of Transplanted Cells

For animal MRI analysis, a 4.7T Bio Spec (Bruker, Ettlingen, Germany) was employed using T2- and T2*-weighted imaging techniques. The MRI setting and detection protocols were the same as described previously (16). To detect the Feridex^®-labeled F3.BDNF cells, we performed Prussian blue staining [cells were incubated at room temperature in 4% potassium ferrocyanide (Sigma) and 20% hydrochloric acid (Sigma) for 30 min], followed by immunohistochemical staining using antibodies against human-specific nuclei or mitochondria, in conjunction with markers representing various cell types or lineages. Prior to transplantation, we verified that the cells were completely labeled with Feridex^®, by examining the Prussian blue staining pattern.

Immunohistochemical Analysis

For immunohistochemical analysis, the transplanted animals were sacrificed at 8 weeks following transplantation and were perfused and fixed with 4% paraformaldehyde (Sigma). Coronal sections of the brain were cut at 40 mm using a cryostat (Microm, Walldorf, Germany). Primary antibodies used for immunohistochemistry were as follows: human-specific nuclei (1:200, Chemicon), human-specific mitochondria (1:200, Chemicon), human-specific nestin (Chemicon), doublecortin (DCX) (1:200, Cell Signaling Technology, Danvers, MA, USA), MAP2 (1:200, Chemicon), neuronal nuclei (NeuN; 1:500, Chemicon), tyrosine hydroxylase (TH; 1:1000, Pel-Freez, Rodgers, AR, USA), glutamic acid decarboxylase 65/67 (GAD65/67; 1:200, Chemicon), dopamine- and cAMPregulated phosphoprotein, Mr 32 kDa (DARPP-32; 1:100, Cell Signaling), chemokine (C-X-C motif) receptor 4 (CXCR4; 1:40, R&D Systems), human-specific Von Willebrand factor (vWF; 1:250, DAKO, Carpinteria, CA, USA), proliferating nuclear antigen (PCNA; 1:50, Santa Cruz Biotechnology), and human-specific BDNF (1:200, R&D Systems). Procedures for secondary antibody treatment, counterstaining, and confocal analysis were the same as described in immunocytochemical analysis (see above).

Results

Characterization of Human Neural Stem Cells Overexpressing BDNF (F3.BDNF)

We first investigated the expression levels of BDNF mRNA and protein in F3.BDNF neural stem cells using semiquantitative RT-PCR, Western blots, and immunocytochemical methods. For a control, F3 neural stem cell line without overexpression of BDNF was used. As expected, we found a significant increase of BDNF mRNA (Fig. 1A) and protein (Fig. 1B) in F3.BDNF cells, compared with F3 cells. We also detected high levels of BDNF protein in F3.BDNF cells by immunocytochemical method (Fig. 1C). These results strongly indicate that BDNF mRNA and protein are overexpressed in F3.BDNF cells.

Figure 1.

Characterization of human neural stem cells overexpressing BDNF (F3.BDNF). (A) Semiquantitative RT-PCR analysis showing increased expression of brain-derived neurotrophic factor (BDNF) mRNA. (B) Western blot analysis showing the mature form of BDNF protein (~15 kDa). (C) Immunocytochemical analysis showing abundant expression of BDNF protein in F3.BDNF cells. GAPDH, glyceraldehyde 3-pjhosphate dehydrogenase; DAPI, 4′,6-diamidino-2-phenylindole.

Neural Properties of F3.BDNF Cells

Since neural stem cells have already committed to neural lineages, they have been conventionally used for neural transplantation studies without further differentiation into more specific type neurons (Fig. 2). However, in order to examine the neuronal phenotypes of F3.BDNF cells, we induced undifferentiated cells (Fig. 2A) to form neurospheres in suspension culture (Fig. 2B), after which they were attached to tissue culture dishes precoated with poly-l-ornithine and fibronectin for further neuronal differentiation (Fig. 2C). At the undifferentiated stage, semiquantitative RT-PCR analysis showed that expression of Sox2 was prominent, compared with other neural stem cell markers including nestin and Musashi, whereas the expression of general neural marker, NCAM, and mature neuron marker, Map2, was relatively low. BDNF mRNA was detected at relatively high levels (Fig. 2D). Immunocytochemical staining further showed that SOX2 expression (Fig. 2F) was slightly higher than nestin expression (Fig. 2E). By contrast, MAP2 expression (Fig. 2G) was relatively low. We also examined the marker expression at the differentiated stage. Semiquantitative RT-PCR analysis showed the expression of more mature neuronal markers, such as Map2 and Darpp-32, was detected at high levels. Importantly, BDNF mRNA was continuously expressed after neuronal differentiation. By contrast, expression of early neurons (type III β-tubulin; Tuj1 antibody) or astrocytes (GFAP) was relatively low. Immunocytochemical staining showed the expression level of MAP2 (Fig. 2J) was higher than those of type III β-tubulin (Fig. 2I) and GFAP (Fig. 2K). Taken together, these results indicate that undifferentiated F3.BDNF cells mainly consist of neural precursor cells or early-stage neurons, whereas differentiated cells contain more mature neurons.

Figure 2.

Neural properties of F3.BDNF cells. Experimental scheme showing a step-wise neuronal differentiation of F3.BDNF cells from (A) undifferentiated stage to (B) neurosphere stage and (C) mature neuron stage. Semiquantitative RT-PCR (D) and immunocytochemical analyses (E–G) showing marker expression in undifferentiated F3.BDNF cells. Semiquantitative RT-PCR (H) and immunocytochemical analyses (I–K) showing marker expression in neurally differentiated F3.BDNF cells. Scale bar: 50 mm. SOX2, sex-determining region Y box 2; NCAM, neural cell adhesion molecule; Map2, microtubule associated preotein 2; Darpp-32, dopamine- and cAMP-regulated phosphoprotein, Mr 32 kDa; GFAP, glial fibrillary acidic protein; Tuj1, antibody for III β-tubulin.

Intracerebral Transplantation of F3.BDNF Cells Into a Rodent Model of MCAo

Figure 3 outlines the experimental scheme for making stroke animal models, cell transplantation, behavioral tests, and immunohistochemical analysis. MCAo was induced by 90-min reperfusion, and F3.BDNF cells were transplanted into the contralateral side of ischemic lesion at one week following MCAo induction. Among various routes for stem cell transplantation, contralateral injection, as opposed to ipsilateral injection, was chosen as this will minimize the secondary damage to the injured brain, which may be caused by direct injection into the lesion side. Furthermore, since we routinely use animal MRI to track the fates of transplanted stem cells, contralateral injection will provide much cleaner results than ipsilateral injection, in which the needle track near the lesion area can often gives rise to false positive or ambiguous MRI signals. After transplantation, we performed behavioral tests, including stepping, rotarod, and mNSS tests weekly up to 8 weeks until the rats were sacrificed for immunohistochemical analysis.

Figure 3.

Experimental scheme showing the timetable for MCAo animal model production, cell transplantation, behavioral tests, MRI examination, and immunohistochemical analysis. MCAo, middle cerebral artery occlusion; MRI, magnetic rosnance imaging; Tx, transplantation; IHC, immunohistochemistry; mNSS, modified neurological severity score.

Behavioral Improvements in F3.BDNF-Transplanted MCAo Rats

To assess whether F3.BDNF cells could exert functional effects after intracerebral transplantation, we grafted 400,000 cells in a volume of 2 ml into the contralateral side of striatum (n = 8) at 1 week after MCAo surgery was made. In sham control rats (n = 7), we only injected vehicle (DMEM). To evaluate functional effects of the grafts, we employed three behavioral tests (i.e., stepping, rotarod and mNSS tests) up to 8 weeks following transplantation. We observed significant behavioral recovery (p < 0.001) at as early as 3 weeks (in the stepping and rotarod tests) (Fig. 4A, B) and 4 weeks (in the mNSS test) (Fig. 4C) after transplantation surgery. Improvements of behavioral scores in each test indicate amelioration of forelimb akinesia (stepping test), unbalanced motor performance (rotarod test), and defective sensorimotor functions (mNSS test) in stroke-damaged rats, following transplantation of F3.BDNF cells.

Figure 4.

Behavioral improvements following transplantation of F3.BDNF cells into MCAo models. Stepping test (A), rotarod test (B), and mNSS test (C) were carried out every week up to 8 weeks. *p < 0.05; **p < 0.001.

Formation of Neural Tissues From Transplanted F3.BDNF Cells

After completion of the behavioral tests at 8 weeks, we killed the rats and analyzed their brains histologically. To identify the transplanted human cells, we used antibodies against human-specific nuclear antigen (hNu) and human mitochondria (hMito). Using confocal microscopy and double immunostaining, we observed high proportions of contralaterally transplanted F3.BDNF cells migrated to the stroke-damaged area in the ipsilateral side. We found that some of the transplanted cells still remained as nestin-positive neural stem cells (Fig. 5A), whereas the others formed DCX-positive migrating neuroblasts (Fig. 5B) or MAP2-positive mature neurons (Fig. 5C). Importantly, they also formed TH-positive neurons (Fig. 5D) and GAD65/67-positive neurons (Fig. 5E), indicating the formation of dopamine and GABA, respectively, which are two important neurotransmitters regulating striatal function. In conjunction with these histological findings, we observed that the transplanted cells expressed DAPRPP-32, a marker for striatal projection neurons (Fig. 5F). As expected, ipsilaterally migrated F3.BDNF cells expressed CXCR4 (Fig. 5G), a marker for chemokine receptor 4, which is to known to respond to inflammatory signals, such as stromal-derived factor-1 (SDF-1), following ischemic brain injury. Interestingly, we observed small proportions of transplanted F3.BDNF cells formed vascular structures, exemplified by the induction of vWF expression (Fig. 5H). Finally, we detected small proportions of F3.BDNF cells were still proliferating, judged by expression of PCNA-positive cells (Fig. 5I). However, we never found any signs of graft overgrowth or tumor formation in all 15 transplanted rats 8 weeks after transplantation.

Figure 5.

Immunohistochemical analysis at 8 weeks following transplantation, showing formation of various neuronal cell types from transplanted F3.BDNF cells. Colocalization of human-specific markers: human nuclei (hNu) (A–F, H, I) or human mitochondria (hMito) (G) with markers for neural precursor cells (nestin, A), migrating neuroblasts (DCX, B), mature neurons (MAP2, C), dopaminergic neurons (TH, D), GABAergic neurons (GAD65/67, E), medium spiny projection neurons (DARPP-32, F), a chemokine receptor (CXCR4, G), vascular endothelial cells (vWF, H), and proliferating cells (PCNA, I). DAPI was used to counterstain the cells. Arrows, colocalized marker expression. Scale bar: 20 mm. DCX, doublecortin; TH, tyrosine hydroxylase; GAD65/67, glutamate decarboxylase 65/67; CXCR4, chemokine (C-X-C motif) receptor 4; vWF, Von Willebrand factor; PCNA, proliferating cell nuclear antigen.

Migration and Distribution of Transplanted F3.BDNF Cells

To monitor the fates of transplanted cells following transplantation into contralateral side of stroke-damaged brains, we labeled the cells with Feridex^® and carried out MRI analysis using a 4.7T animal MRI. Figure 6A shows coronal views of some representative T2- and T2*-weighted MR images taken near cell injection site (AP: +0.3 mm from bregma), as well as in the posterior region (AP: −4.44 mm from bregma). Interestingly, migration of contralaterally transplanted F3.BDNF cells to the infarct area was already detectable at as early as 5 days after transplantation. We also observed that the MR signal intensity increased gradually when examined at 18 and 32 days after transplantation. Interestingly, the MR signal was also detected in the posterior region, demonstrating the extensive migration of transplanted cells. To visualize the MRI signals histologically, we performed Prussian blue staining on brain section and found intensive staining in the infarct area (Fig. 6B, C). To confirm whether Prussian blue-stained cells correspond to the transplanted cells, we carried out immunohistochemical staining using an antibody against human-specific nuclear antigen (hNu) and found that they are indeed derived from transplanted human cells (Fig. 6D). Figure 6E shows a diagram illustrating the location and distribution of hNu-positive cells. Interestingly, they were mainly detected on the border of infarct area, as well as in the corpus callosum, which may serve as the major route for cell migration towards different hemisphere of the brain. Small proportions of hNu-positive cells were also detected in the cerebral cortex and the ventral side of putamen, but it is not clear whether they also participate in the migration process. Interestingly, immunohistochemical staining indicates that human BDNF proteins, which were likely to be produced from the transplanted F3.BDNF cells, were distributed consistently with hNu-positive staining patterns (Fig. 6E).

Figure 6.

In vivo tracking of F3.BDNF cells using 4.7T animal MRI following transplantation into the contralateral side of infarct in a rodent model of MCAo. (A) Representative T2- and T2*-weighted images showing coronal views of transplanted and migrated cells in antero-posterior sequences (from 0.3 to −4.44 mm, measured from bregma) taken at 5 days (5D), 18 days (18D), and 32 days (32D) following transplantation. Note the migration of Feridex^®-labeled transplanted cells to the infarct area (arrowheads at AP: +0.3 mm) and the increase of signal intensity at later stages. Accumulation of signals is also detectable in the posterior region (arrowheads at AP: −4.44 mm). (B) Prussian blue staining showing the detection of Feridex^®-labeled transplanted cells in the infarct area. (C) Higher magnification of inset (C) in (B), showing Prussian blue-stained cells. (D) Immunohistochemical staining showing the accumulation of hNu-positive cells along the infarct border area in inset (D) of (B). (E) Schematic diagram showing the distribution of contralaterally transplanted F3.BDNF cells, based on hNu staining patterns. Cells were detected in various regions including corpus callosum (CC), but they were strongly localized in the infarct and peri-infarct areas, exhibiting their extensive tropism toward the ischemic injury. (F) Immunohistochemical staining showing the distribution of human BDNF proteins that are derived from transplanted F3.BDNF cells in various regions in the brain. Central figure shows neuronal nuclei (NeuN) staining visualized by DAB staining, which also depicts infarct area and various locations where BDNF staining patterns were analyzed. Insets show high magnifications of NeuN staining patterns intact and infarcted striatum, respectively. Note the abundant BDNF expression in the (C) region (infarct area), in consistent with hNu staining pattern, as shown in (D). Scale bar: 100 mm.

Discussion

In this study, we showed that neural stem cells overexpressing BDNF (F3.BDNF) are highly effective in treating a rodent model of ischemic stroke (MCAo). This is a continuation of our recent study using a mouse model of hemorrhagic stroke (ICH) (23). While our earlier study focused on the relative advantages of using F3.BDNF cells as opposed to F3 cells alone, in terms of BDNF production, graft survival, behavioral recovery, and apoptosis, the present study directly addressed the therapeutic effects of F3.BDNF cells in a rodent model of MCAo, the most common type of stroke.

Firstly, we showed that BDNF mRNA and protein are abundantly produced in F3.BDNF cells using semiquantitative RT-PCR, Western, and immunocytochemical analyses (Fig. 1), further confirming the previous observations (23). In particular, we showed that mature form of BDNF protein is processed using Western blot analysis (Fig. 1B).

Secondly, we addressed the neural properties of F3.BDNF cells in vitro by neurosphere formation in suspension culture and the subsequent attachment to tissue culture dishes precoated with poly-l-ornithine and fibronectin (Fig. 2).

Previous studies using neural stem cells mostly analyzed the neural or glial marker expression at undifferentiated stage, in which relatively early stage neurons or neural precursors are detectable. By contrast, we were able to detect more maturetype neurons, such as MAP2, DARPP-32-expressing neurons in the differentiated cell populations (Fig. 2H), which represents the neurogenic potential of F3.BDNF cells at more differentiated stages. Interestingly, we observed that BDNF is still expressed at this stage.

Thirdly, we carried out detailed analysis on the in vivo functions of F3.BDNF cells following transplantation into the contralateral side to the infarct region streotoxically. Our previous MRI study using human bone marrow-derived mesenchymal stem cells (16) as well as F3 cells (6) demonstrated that most stem cells have the capacity to migrate to the injury site, exhibiting a pathotropism (15). Based on these observations, we avoided ipsilateral injection that will inevitably introduce a secondary injury to the infarct area by causing tissue damage and inflammatory response, while contralateral injection into healthy striatum will introduce less transplantation-associated tissue damage. Moreover, contralateral injection can minimize the misinterpretation of MRI results, since the injection site itself in the ipsilateral region can be often interpreted as false positive.

Transplanted animals were monitored up to 8 weeks using various behavioral tests and animal MRI before they were sacrificed for immunohistochemical analysis (Fig. 3). They exhibited significant behavioral improvements in stepping, rotarod, and modified neurological severity score (mNSS) tests, indicating significant amelioration of forelimb akinesia, motor performance, and sensorimotor functions can be achieved by transplanting F3.BDNF cells into stroke-damaged rat brains (Fig. 4). Immunohistological analysis confirmed that the transplanted human cells were colocalized with nestin, DCX, MAP2, DARPP-32, TH, GAD65/67-positive cells (Fig. 5). In particular, formation of DARPP-32-positive striatal projection neurons (Fig. 5F) is closely related to the improvement of striatal functions in stroke-damaged brains following transplantation (3).

We found that the great majority of transplanted cells express CXCR4, a marker for chemokine receptor, which responds to stroke-induced inflammatory signals, such as SDF-1 (13,29), and the strong expression of CXCR4 in F3.BDNF cells (Fig. 5G) represents their migratory capacity towards the infarct area (Fig. 6). Importantly, we found that small portions of transplanted F3.BDNF cells formed vWF-positive vascular structures (Fig. 5H). We have previously reported that F3.BDNF cells can stimulate proliferation of host microvessels in mouse ICH models (23), but the present study demonstrates the graft-induced microvessel formation. Cross-talk between neural cells and endothelial cells has been suggested in many studies, and this result indicate neurotrophic factor BDNF, which is released from transplanted F3.BDNF cells, may switch their cell fate from neural stem cells into endothelial or vascular lineages to some extent in the stroke-damaged brain (25). We also detected small populations of PCNA-positive cells (Fig. 5I), without any signs of graft overgrowth or tumor formation. For this, we speculate that the transplanted neural stem cells may undergo proliferation until they obtain sufficient numbers of cells in order to migrate to and engraft to appropriate environments for neural tissue formation and tissue repair. More importantly, we were able to detect high levels of human BDNF protein expression, presumably derived from the transplanted F3.BDNF cells. The secreted BDNF may provide milieu to protect the stroke-damaged ischemic area, to promote the proliferation of exogenous and/or endogenous stem cells, or to facilitate neuronal differentiation and restoration. It is also possible that the secreted BDNF plays an important role in modulating inflammatory response, which may be closely involved in the disease process (25,39).

In our animal MRI study, we found that migration of contralaterally transplanted F3.BDNF cells to the infarct area was already detectable at as early as 5 days, and the signal intensity increased further at 18 and 32 days after transplantation. In addition to tangential migration across brain hemispheres, we also observed the accumulation of cells in the posterior region of infarct area, indicating an antero-posterior migration of cells (Fig. 6A). Using both Prussian staining and immunohistochemical staining using an antibody against hNu, we were able to track the migration and distribution patterns (Fig. 6B–E). Feridex^®-labeled cells were detected in various regions including corpus callosum that may serve as a major route for tangential migration, but they were strongly localized in the infarct and peri-infarct areas, exhibiting their extensive tropism toward the ischemic injury.

Figure 7 summarizes the possible action mode of F3.BDNF cells following transplantation into contralateral side of infarct region, in which the transplanted cells migrate to the injury site extensively and differentiate into neuronal cell types and produce BDNF proteins, which contribute to the functional recovery (Fig. 7).

Figure 7.

Schematic diagram showing the action mode of F3.BDNF cells following transplantation into contralateral side of infarct region.

Clinical trials with stroke patients have been performed so far using various cell sources, from immortalized human teratocarcinoma cell line to autologous bone marrow-derived mesenchymal stem cells (BM-MSCs) (1,4,11,19,20,30). Outcomes have been variable with some reports describing no benefit whilst others have indicated transient clinical improvement (1,35). More recently, another clinical trial in stroke patients is ongoing by ReNeuron using clonal, conditionally immortalized NSCs (ReN001) from isolated human fetal cortex, which were previously shown to ameliorate motor impairments in the rat stroke model (33). In this case, the cells were administered through the intracerebroventricular route, and the preliminary data suggest that it may be safe (35). It is noteworthy that one of the F3 derivatives that overexpress cytosine deaminase, called HB1.F3.CD cells, is currently used in the phase I clinical in brain tumor patients (1). Therefore, it will be theoretically possible to use F3.BDNF cells for clinical purposes if they were prepared under GMP-grade facility. Apart from neural or nonneural cell sources discussed so far, patient-derived iPSCs (induced pluripotent stem cells) may become an important autologous cell source for treating stroke patients in the future.

In summary, we have carried out detailed analysis on the therapeutic potential of human neural stem cells overexpressing BDNF (HB1.F3.BDNF), and our results show they are very effective in treating stroke animal models and may serve as an experimental platform to develop clinical applications in the future.