Author contributions: K.O., J.T., J.W., P.K., and S.W.: conception and design, collection and assembly of data, data analysis and interpretation, and manuscript writing; Y.M. and D.T.P.: collection and assembly of data and data analysis; E.M.: conception and design, collection and assembly of data, and data analysis; H.A. and J.L.: collection and assembly of data and data analysis and interpretation; O.B.: conception and design, data interpretation, financial support, and manuscript writing; O.L. and Z.K.: conception and design, data analysis and interpretation, financial support, and manuscript writing. K.O., J.T., J.W., and P.K. contributed equally to this article.
Disclosure of potential conflicts of interest is found at the end of this article.
First published online in STEM CELLSEXPRESS April 11, 2012.
Reprogramming of adult human somatic cells to induced pluripotent stem cells (iPSCs) is a novel approach to produce patient-specific cells for autologous transplantation. Whether such cells survive long-term, differentiate to functional neurons, and induce recovery in the stroke-injured brain are unclear. We have transplanted long-term self-renewing neuroepithelial-like stem cells, generated from adult human fibroblast-derived iPSCs, into the stroke-damaged mouse and rat striatum or cortex. Recovery of forepaw movements was observed already at 1 week after transplantation. Improvement was most likely not due to neuronal replacement but was associated with increased vascular endothelial growth factor levels, probably enhancing endogenous plasticity. Transplanted cells stopped proliferating, could survive without forming tumors for at least 4 months, and differentiated to morphologically mature neurons of different subtypes. Neurons in intrastriatal grafts sent axonal projections to the globus pallidus. Grafted cells exhibited electrophysiological properties of mature neurons and received synaptic input from host neurons. Our study provides the first evidence that transplantation of human iPSC-derived cells is a safe and efficient approach to promote recovery after stroke and can be used to supply the injured brain with new neurons for replacement. STEM CELLS2012;30:1120–1133
Ischemic stroke is a leading cause of mortality and disability. Ongoing studies in animal models of stroke suggest that stem cell-based approaches could provide therapeutic benefit in a clinical setting. The observed improvements may, at least partly, be due to neuronal replacement . After transplantation into the stroke-damaged rodent brain, embryonic stem cell (ESC)-derived neural stem cells (NSCs) differentiated into neurons and improved some impaired functions [2–4]. Electrophysiological recordings showed mature neuronal properties of the grafted cells and signs of synaptic integration into host brain [5, 6]. Human fetal NSCs also gave rise to mature neurons after transplantation into the stroke-damaged rat brain [7–9]. Transplanted NSCs may promote recovery without differentiating to neurons through several other mechanisms, for example, modulation of inflammation [10–12], neuroprotection , stimulation of angiogenesis [11, 13, 14], and enhancing brain plasticity [15–17].
Somatic cells can be reprogrammed to pluripotent stem cells by introduction of transcription factors . These induced pluripotent stem cells (iPSCs) can be differentiated to specific neuronal subtypes, for example, dopaminergic neurons [19, 20] and motor neurons [21, 22]. Autologous transplantation of neurons generated from iPSCs seems to be more attractive in stroke therapy than in the treatment of chronic neurodegenerative disorders, in which patient-specific cells may exhibit increased susceptibility to the pathological processes. Recently, mouse iPSCs, implanted into the stroke-damaged brain of mice, were reported to generate large numbers of neuroblasts and a few mature neurons but to form tumors . When these cells were mixed with fibrin glue and were delivered subdurally into the injured brain tissue in rats after stroke, there was a reduction of the infarct volume and some behavioral improvement . Finally, human fibroblast-derived iPSCs implanted into the striatum of stroke-damaged rats were found to improve short-term sensorimotor recovery, but whether any neurons were formed is unclear . Another recent study did not detect any effect of iPSCs on stroke-induced behavioral impairments at 4 weeks after transplantation . The long-term consequences after transplantation of human iPSC-derived cells in the stroke-damaged brain, if they can differentiate in vivo to morphologically and functionally mature neurons, establish connections to target areas, and improve behavioral deficits resembling those in stroke patients, are currently unknown.
Here, we have transplanted long-term expandable neuroepithelial-like stem (lt-NES) cells, generated from adult human fibroblast-derived iPSCs, into the stroke-damaged mouse and rat brain. We show for the first time that grafted human iPSC-derived lt-NES cells generate neurons with mature morphological and electrophysiological properties in vivo, send axonal projections throughout the host brain, receive synaptic input from surrounding host neurons, and improve motor recovery in behavioral tests relevant for human stroke.
MATERIALS AND METHODS
Generation of Human-iPSCs
For production of retroviral particles, PhoenixGP cells (ATTC #3514) were transfected with plasmids encoding for the viral glycoprotein VSV-G (Addgene 12259, Cambridge, Massachusetts, http://www.addgene.org) and the reprogramming factors (Oct4, Sox2, KLF4, and c-MYC (Addgene 17217-17220). Dermal fibroblasts were infected twice with ultracentrifuge-concentrated virus in Dulbecco's modified Eagle's medium (DMEM) high glucose media containing 10% fetal calf serum, 1% sodium pyruvate, 1% nonessential amino acids, and 1% L-glutamine (all life technologies) in the presence of 10 ng/ml FGF2. Cells were split into plates preseeded with mouse embryonic fibroblasts (MEFs). Medium was switched to human iPSC culture medium containing DMEM/F12 supplemented with 20% knockout serum replacement, 0.1 mM nonessential amino acids (all from life technologies), 1 mM L-glutamine, 0.1 mM β-mercaptoethanol, and 50 ng/ml zebrafish basic fibroblast growth factor (zbFGF) 6 days post-transduction and changed every other day. Four weeks after transduction, colonies with human pluripotent stem cell morphology were manually picked, carefully triturated, and expanded to establish human iPSC lines.
Neural Differentiation of Human iPSCs
Human iPSCs were induced to differentiate as described . Briefly, iPSC colonies were detached from the MEF layer using collagenase (life technologies), and embryonic body (EB) formation was induced by plating the colonies on nonadhesive plastic culture dishes in human iPSC media without zbFGF. Fifty percent of the media was changed every second day, and 5-day-old EBs were plated on tissue culture plates coated with 0.1 mg/ml poly(L-ornithine) (Sigma, St. Louis, Missouri, http:www.sigma-aldrich.com). Neural rosette structures started to emerge about 1 week after plating. Rosettes were carefully picked with a needle, and isolated clusters were transferred to a nonadhesive culture plate in DMEM/F12, 2 mM L-glutamine, 1.6 g/l glucose, and N2 supplement (1:100) (all life technologies). To establish lt-NES cell lines, floating rosettes were dissociated 2 days later using trypsin followed by inhibition of trypsin by trypsin inhibitor (life technologies). Cells were centrifuged for 5 minutes at 300g and plated onto poly(L-ornithine)-and 10 μg/ml laminin (Sigma)-coated plates into the same media supplemented with 10 ng/ml FGF2, 10 ng/ml epidermal growth factor (EGF) (R&D systems, Minneapolis, Minnesota, http://www.rndsystems.com), and B27 (1 μl/ml, life technologies). These iPSC-derived lt-NES cells were passaged at a ratio of 1:2-1:3 every second to third day using trypsin. All transplantation studies were performed with lt-NES cells passages 20-30.
Messenger RNA samples were isolated using mRNA extraction kit (Qiagen, Hilden, Germany, http://www.qiagen.com) according to supplier's instructions. Adult brain RNA served as positive control (Stratagen, http://www.stratagen.com). One microgram of total RNA was used for reverse transcription (RT) with the iScript cDNA synthesis kit (BioRad, Hercules, California, http://www3. bio-rad.com) following the manufacturer's protocol. Quantitative RT polymerase chain reaction (PCR) analyses were performed in triplicates on a Biorad iCycler using SYBRI-green detection method. PCR products were assessed by dissociation curve and gel electrophoresis. Data were normalized to 18S rRNA levels. Primers for vascular endothelial growth factor (VEGF): forward: 5′-cagctactgccgtccaatcg-3′, reverse: 5′-tcaccgcctcggttgt-3′.
Middle Cerebral Artery Occlusion in Mice and Rats
All experimental procedures were approved by the Malmö-Lund Ethical Committee. Stroke was induced using the intraluminal filament model of middle cerebral artery occlusion (MCAO) in Nude rats [28, 29] and C57bl6 mice [30, 31]. Briefly, the right common carotid artery (CCA) and its proximal branches were isolated. The CCA and external carotid artery (ECA) were ligated, and internal carotid artery (ICA) was temporarily clipped using a metal microvessel clip. A nylon monofilament was advanced through the ICA until resistance was felt. Thirty minutes after the start of occlusion, the nylon filament was carefully removed, and ECA was ligated permanently.
Distal MCAO in Rats
Focal ischemic injury in cerebral cortex was induced by permanent occlusion of the distal cortical branch of MCA and transient (30 minutes) occlusion of both CCAs in Nude rats, as described previously . Briefly, the temporal bone was exposed. A craniotomy of approximately 3 mm was made, the dura was carefully opened, and the cortical branch of MCA was ligated by suture. Both CCAs were isolated and temporarily ligated. At 30 minutes after the ligation, CCAs were released and surgical wounds were closed.
Intracerebral transplantation of human iPSC-derived lt-NES cells was performed stereotaxically at 1 week and 48 hours after MCAO in mice and rats, respectively. At the day of transplantation, iPSC-derived lt-NES cells transduced with lentivirus carrying green fluorescent protein (GFP) and naïve cells for rats and mice, respectively, were resuspended to a final concentration of 100,000 cells per microliter. Coordinates for transplantation sites and volume are presented in Supporting Information Table S1. Tooth bar was set at −3.3 mm. Mice were given subcutaneous injections of 10 mg/kg Cyclosporine A every other day after transplantation.
Retrograde Tracer Injection
At 9 weeks after transplantation or vehicle injection, all mice with MCAO lesions were injected iontophoretically with 2% Fluoro-Gold (FG; Biotium, Hayward, California, http://www.biotium.com), dissolved in distilled water, into globus pallidus (GP) ipsilateral to the transplant. A separate group of mice (n = 5) was transplanted with GFP-labeled lt-NES cells into the intact striatum, injected with FG into GP 2 weeks later, and sacrificed 3 weeks after transplantation, that is, 1 week after FG injection. Iontophoretic injections were made using microcapillaries with tip diameters of 20–30 μm and 5 μA positive current pulsed for 7 out of every 14 seconds during 7 minutes using a constant current generator (Midgard; Stoelting, Wood Dale, Illinois, https://www.stoeltingco.com).
Cells were fixed in 4% paraformaldehyde for 10–20 minutes at room temperature, washed in phosphate buffered saline (PBS) and incubated in blocking solution containing 10% FCS and 0.25% Triton X-100 for 1 hour at room temperature. For detection of GABA, 0,1% glutaraldehyde was added to the fixative. Mice were sacrificed at 10 weeks and rats were sacrificed at 2 weeks, 2 or 4 months after transplantation, respectively. Animals were perfused transcardially with 4% paraformaldehyde. Coronal sections (30 μm) and cells were preincubated in blocking solution (5% normal serum and 0.25% Triton X-100 in 0.1 M potassium-buffered PBS). Primary antibodies (Supporting Information Table S2) were diluted in the blocking solution and applied overnight at 4°C. Fluorophore-conjugated secondary antibodies (Molecular Probes, Eugene, Oregon, http://www.invitrogen.com or Jackson Laboratories, Bar Harbor, Maine, http:swww.jax.org) were diluted in blocking solution and applied for 2 hours followed by three rinses in potassium PBS.
The staircase test was used to assess “side-specific” skilled forelimb reaching and grasping abilities [33, 34] in a staircase apparatus (Campden Instruments, Loughborough, U.K., http://www.campden-inst.com). The corridor test,  adapted to mice , was used to assess sensorimotor impairment caused by striatal damage.
Cell Counting and VEGF Immunoreactivity
For quantification of colocalization of HuNu+ cells with different markers, approximately 1,000 HuNu+ cells in each animal were analyzed. Double-labeled cells identified in an epifluorescence/light microscope were validated with a confocal microscope (Leica, Solms, Germany, http://www.leica-camera.com). For quantification of surviving grafted cells and FG+/HuNu+ cells in mice, all HuNu+ cells in all sections were counted. Total number of HuNu+ and GFP+ cells in rats were counted stereologically by using C.A.S. T-Grid software (Olympus, Tokyo, Japan, http://www.olympus.com) and cell numbers were calculated according to the optical fractionator formula .
VEGF immunoreactivity was assessed in the astrocytes and blood vessels outside the graft in animals transplanted with lt-NES cells and in the corresponding areas of vehicle-injected animals. Similarly, VEGF immunoreactivity was measured in the graft and in the corresponding area in the vehicle-injected animals. To quantify VEGF immunoreactivity, images of different regions of interest (ROIs) in two representative sections from each brain double-stained for VEGF and glial fibrillary acidic protein (GFAP) were acquired. The area of VEGF-immunoreactivity within GFAP+ and GFAP− cells and in grafted and corresponding vehicle-injected regions was determined by image analysis using Cellsens Dimension 2010 software (Olympus). In each section, areas of immunoreactivity were identified using defined representative ranges of threshold for specific signal. Using these defined parameters, the images of each ROI were analyzed by software, which calculated the total area covered by pixels/specific immunopositive signal. The values corresponding to total fluorescence areas were averaged and expressed as the mean VEGF+ area per animal.
Rats were anesthetized with isofluorane and decapitated. Brains were rapidly removed and placed in ice-cold, gassed (95% O2, 5% CO2) solution (pH 7.2–7.4, 295–300 mosm), containing (in mM): 75 sucrose, 67 NaCl, 26 NaHCO3, 1.25 NaH2PO4, 0.5 CaCl2, 7 MgCl2, and 25 glucose. Coronal slices (300 μm), cut on a vibratome (Leica VT1200S), were placed in an incubation chamber with gassed (95% O2, 5% CO2) artificial cerebral spinal fluid (aCSF, pH 7.2–7.4, 295–300 mosm) containing (in mM): 119 NaCl, 2.5 KCl, 1.3 MgSO4, 2.5 CaCl2, 26 NaHCO3, 1.25 NaH2PO4, and 25 glucose. Slices were constantly perfused with aCSF (32°C–34°C). Recording pipettes were filled with solution (pH 7.2–7.4, 295–300 mosm) containing (in mM): 122.5 potassium gluconate, 12.5 KCl, 10.0 KOH-HEPES, 0.2 KOH-EGTA, 2.0 Mg-ATP, 0.3 Na3-GTP, and 8.0 NaCl. For measurements of spontaneous inhibitory postsynaptic currents recording pipettes contained (in mM): 135.0 CsCl, 10.0 CsOH, 0.2 CsOH-EGTA, 2.0 Mg-ATP, 0.3 Na3-GTP, 8.0 NaCl, and 5.0 lidocaine N-ethyl bromide (QX-314). Biocytin (0.5%; Sigma-Aldrich) was freshly dissolved in the pipette solution. Evoked currents were measured after electrical stimulation (two 100 μs square-wave pulses of 150–300 μA timed 50 ms apart) delivered by a single stainless steel wire enclosed in a glass pipette filled with aCSF (∼2 MΩ resistance). Voltage-gated sodium channels were blocked with 1 μM tetrodotoxin (TTX) (Tocris, Bristol, UK, http://www.tocris.com). Voltage-activated potassium channels were blocked using tetraethylammonium (TEA) (Sigma). N-methyl-D-aspartic acid (NMDA) and α-amino-3-hydroxy-5-methyl-4-isoxazolepropionic (AMPA) receptors were blocked using 50 μM D-AP5 and 5 μM 2,3-dihydroxy-6-nitro-7-sulfamoyl-benzo[f]quinoxaline-2,3-dione (NBQX), respectively. GABAA receptors were blocked using 100 μM picrotoxin (PTX) (Tocris).
Generation and In Vitro Properties of lt-NES Cells from Human iPSCs
Human iPSC lines were produced by standard retroviral transduction . Selected clones were further propagated to establish iPSC lines. The iPSC line used in this study showed morphologies indistinguishable from those of human ESCs (Fig. 1A) and expressed the pluripotency-associated markers (Fig. 1B–1D). Following transplantation into immunodeficient mice, the cells formed teratomas in vivo consisting of all three germ layers (Fig. 1E–1G) and had a normal karyotype (Fig. 1H).
A homogeneous population of iPSC-derived lt-NES cells was established using our previously established protocol . In the presence of the growth factors FGF2 and EGF, these cells can be propagated along multiple passages without losing their stem cell characteristics or their differentiation potential. Human iPSC-derived lt-NES cells (Fig. 1I) uniformly expressed the NSC-associated markers Sox2 and nestin (Fig. 1J, 1K). They assembled into rosette-like patterns with expression of the rosette-associated transcription factors PLZF and Dach1 and tight junction maker ZO-1 (Fig. 1K, 1L). Upon growth factor withdrawal and independent of passage number, human iPSC-derived lt-NES cells differentiated into neurons (>70%) and astrocytes (Fig. 1M–1P).
Transplantation of Human iPSC-Derived lt-NES Cells into the Stroke-Damaged Mouse Striatum Improves Recovery of Fine Forelimb Movements Independent of Long-Term Graft Survival
We first explored whether implantation of human iPSC-derived lt-NES cells could influence the behavioral recovery after stroke. Twenty-three mice subjected to MCAO were divided into two groups and transplanted into the striatum with human iPSC-derived lt-NES cells (n = 12) or vehicle (n = 11) at 1 week after the ischemic insult. A third group of sham-operated animals (n = 9) was given an intrastriatal vehicle injection. We evaluated fine forelimb movement using the staircase test at 2, 5, and 9 weeks after MCAO or sham surgery, that is, 1, 4, and 8 weeks after transplantation. In this test, the number of retrieved pellets reflects the ability to move the forelimb, whereas the number of eaten pellets depends on the ability of both grasping and eating pellets, requiring fine and well-coordinated forelimb movements. Vehicle-injected, stroke-affected mice showed impairment in both parameters (Fig. 2A, 2B), whereas animals with human iPSC grafts did not exhibit any deficits at any time point and performed significantly better than vehicle-treated animals. Taken together, our data indicate that the human iPSC-derived lt-NES cells gave rise to sustained improvement of fine forepaw movement, probably starting already early after transplantation. In contrast, human iPSC-derived lt-NES cell transplantation did not reverse the deficit detected in the corridor test (Fig. 2C).
We next determined the location, survival, and proliferative activity of the grafted cells. The MCAO caused a reproducible lesion of the dorsolateral part of the striatum in all mice (Fig. 3A) . Surviving grafts were found in seven out of 12 animals (Fig. 3B). The vast majority of the human cells were located in the striatum. A few HuNu+ cells were also detected in cerebral cortex and corpus callosum. The number of HuNu+ cells in each graft was 9,378 ± 2,517, that is, approximately 10% of the initial number of transplanted cells. At 10 weeks after transplantation, only few HuNu+ cells (1.6% ± 0.9%) kept proliferating and were positive for Ki67 (Fig. 3C–3E).
We observed no differences in the pattern or location of the stroke-induced damage between the human iPSC-derived lt-NES cell- and vehicle-implanted groups. At 11 weeks after the insult, the volume of the injured striatum did not differ between cell-transplanted and vehicle-injected groups (4.19 ± 0.6 mm3 and 4.69 ± 0.5 mm3, respectively). In five out of 12 animals implanted with human iPSC-derived lt-NES cells, no surviving grafts were detected 10 weeks after transplantation. We reanalyzed our behavioral data and compared the impairment and magnitude of recovery between the groups with and without surviving grafts. The two groups showed no significant difference in number of retrieved or eaten pellets or in the corridor test (data not shown). These findings suggest that long-term survival of grafted cells is not necessary for the functional recovery induced by iPSC-derived lt-NES cells during the first 2 months after transplantation.
Human iPSC-Derived lt-NES Cells Differentiate to Neurons and Establish Axonal Connections with the GP After Transplantation in Stroke-Damaged Mouse Brain
We assessed the proportion of grafted cells expressing neuronal markers such as doublecortin (DCX, a marker for neuroblasts) and HuD (a marker for young and mature neurons) at 10 weeks after transplantation (Fig. 3C–3E). The majority of grafted cells were HuNu+/HuD+ (78.5% ± 11.2%) and mainly located in the center of the graft. A smaller portion of the cells was HuNu+/DCX+ (13.4% ± 4.7%), with preferential location at the periphery of the graft. Thus, most of the grafted human iPSC-derived lt-NES cells had differentiated to neurons.
We then examined whether neurons generated from human iPSC-derived lt-NES cells send axonal projections to the appropriate location in the host brain. We injected the retrograde tracer FG into the GP (Fig. 3F) ipsilateral to the graft (Fig. 4H) at 9 weeks after transplantation (1 week prior to sacrifice). In five animals, FG staining was confined to GP (Fig. 4G). In these animals, a small fraction of HuNu+ cells in the striatal grafts was stained with FG (1.16% ± 0.16%; Fig. 3I–3M). Axons were occasionally observed to extend from HuNu+/FG+ cells (Fig. 3I). In addition, many FG+ cells were found in the remaining intact striatum ipsilateral to the lesion, demonstrating the projection of axons from host striatal neurons to GP (Fig. 3I, 3J). In the control group in which the grafted lt-NES cells were allowed to differentiate only for 3 weeks, we observed many FG+ cells throughout the whole striatum (Fig. 3N). However, we detected no FG-labeled GFP+ lt-NES cells in the transplant providing evidence that 3 weeks are not sufficient for the grafted neurons to extend their axons to GP. Our findings also provide further support for the idea that the FG labeling observed in the main experiment is due to retrograde axonal transport of FG from the GP to the grafted lt-NES cells in striatum and not due to passive diffusion. Our findings provide evidence that the neurons generated from the human iPSC-derived lt-NES cells, implanted into the stroke-damaged mouse striatum, can extend their axons to the GP, which is the main projection area for striatal medium-sized spiny neurons.
Human iPSC-Derived lt-NES Cells Induce Increased Vascular Endothelial Growth Factor Immunoreactivity in Blood Vessels Early After Transplantation in Stroke-Damaged Mouse Brain
In order to explore possible mechanisms underlying the observed functional improvements, mice were transplanted into striatum with either human iPSC-derived lt-NES cells (n = 5) or vehicle (n = 6) at 1 week after stroke and sacrificed 1 week later, that is, at the earliest time point when the enhanced recovery was observed. Stereological analysis revealed no differences in the area of neurogenic subventricular zone cells immunopositive to DCX and BrdU (data not shown). We also obtained no evidence that the grafted cells had influenced the number of activated microglia (Iba1+/ED1+) or astrocytes (GFAP+) in the striatum.
It has been reported that human fetal NSCs through secretion of VEGF can improve behavioral performance already at 1 week after implantation and 2 weeks after stroke . We found VEGF expression in the stroke-affected hemisphere of both lt-NES cell-transplanted and vehicle-injected animals, but the VEGF-immunoreactive areas were larger in animals implanted with lt-NES cells. Microscopic examination revealed that activated microglia did not express VEGF, which was confined to astrocytes and blood vessels (Fig. 4A–4F). VEGF was expressed both in the presumed astrocytic endfeet and the vessel wall (Fig. 4A–4F). The vast majority of GFAP+ astrocytes were positive for VEGF (Fig. 4A–4C) with no difference between the groups (90.9% ± 4.3% and 96.6% ± 2.8% in cell-grafted and vehicle-injected mice, respectively), and approximately 85% of VEGF immunoreactivity outside the graft was found in GFAP+ cells. Computer-based, unbiased analysis of immunostained sections revealed that in animals transplanted with lt-NES cells, the area of VEGF immunoreactivity in astrocytes was more than twofold higher as compared to vehicle-injected animals (Fig. 4G). The area of VEGF immunoreactivity in CD31+ blood vessels was only approximately 10% of that in astrocytes, and lt-NES-transplanted animals did not show any significant difference in VEGF immunoreactivity compared to vehicle-injected animals (Fig. 4G). Also many grafted lt-NES cells exhibited VEGF immunoreactivity (Fig. 4K–4O). This is in agreement with our in vitro data on VEGF gene expression in human iPSC-derived lt-NES cells (Fig. 4H). VEGF immunoreactivity in the graft was much higher than in the corresponding area in vehicle-injected animals (Fig. 4G, 4I, 4J).
We then explored whether the increased VEGF immunoreactivity observed at 2 weeks after transplantation of lt-NES cells had influenced vasculogenesis. Stereological estimation of vessel length density and measurement of CD31 immunoreactivity were performed in CD31-stained sections of lt-NES cell-grafted and vehicle-injected mice at 9 weeks after stroke. At this time point, when the lt-NES grafted animals exhibited significant behavioral improvement, we did not detect any differences between lt-NES cell-grafted and vehicle-injected groups either in vessel length density or in CD31 immunoreactivity (data not shown).
Human iPSC-Derived lt-NES Cells Survive for 4 Months and Differentiate into Neurons After Transplantation into Striatum of T-Cell-Deficient Rats
We wanted to determine whether the human iPSC-derived lt-NES cells and their progeny could survive longer than 10 weeks, differentiate to specific neuronal subtypes, and exhibit the electrophysiological properties of functional neurons. We therefore implanted the cells in Nude, T-cell-deficient adult rats in which xenotransplanted human cells can survive long-term without immunosuppression.
Rats were subjected to intrastriatal transplantation of GFP-labeled human iPSC-derived lt-NES cells at 48 hours after MCAO and were perfused 2 weeks (n = 3), and 2 (n = 8) and 4 (n = 5) months thereafter. We could not detect human iPSC-derived lt-NES cell grafts in two rats. The total number of grafted cells was estimated at 64.8% ± 12.4% and 51.2% ± 13.3% at 2 and 4 months after transplantation, respectively, as compared to the number of implanted cells. The proliferative activity in the grafts (Fig. 5A–5D), quantified as the percentage of HuNu+ cells coexpressing Ki67, was 40.2% ± 4.9% at 2 weeks after transplantation but dropped to 8.2 ± 0.5 and 1.0% ± 0.2% at 2 and 4 months, respectively. In line with this observation, we did not detect any overgrowth of grafted cells up to 4 months after transplantation.
Microscopic examination of the sections revealed that at 2 months after transplantation, the entire core of the graft and most likely the majority of grafted cells were DCX+ (Fig. 5E–5H). In contrast, at 4 months, only a few grafted cells exhibited DCX immunoreactivity (Fig. 5I–5L). Because of cross-reactivity between the HuNu and the mature neuronal marker NeuN antibodies, the neuronal phenotype of grafted cells was determined using GFP and NeuN double-immunostaining (Fig. 5M–5O). At 2 and 4 months after implantation, 65.7% ± 3.9% and 72.9% ± 4.7% of GFP+ cells in striatum, respectively, coexpressed the mature neuronal marker NeuN. Only a fraction of grafted human iPSC-derived lt-NES cells was GFAP+ astrocytes (3.8% ± 0.5% and 6.1% ± 1.5% at 2 and 4 months after transplantation, respectively; Fig. 5P–5R). These cells were distributed mainly at the periphery of the graft core.
Human iPSC-Derived Lt-NES Cells Grafted into Striatum Differentiate to Neurons of Different Phenotype
To characterize the morphological properties of the grafted cells and their progeny, we used immunocytochemical staining for GFP and for the human cytoplasmic marker SC121. The SC121 antibody labels >90% of human neural cells and does not cross-react with rat cells [7, 9]. Grafted GFP+ cells that exhibited clear morphological characteristics of migrating neuroblasts (Fig. 5T, 5X), mature neurons (Fig. 5S, 5V, 5W), or astrocytes (Fig. 5U) were more easily detectable in the areas outside the graft core. However, GFP+ cells with typical neuronal morphology were often not immunopositive for NeuN.
Many of the grafted cells showed GABA immunoreactivity (Fig. 6A–6D; 21.9% ± 1.8% and 10.9% ± 0.6% at 2 and 4 months, respectively). At 4 months after transplantation, grafted cells also stained for markers of GABAergic interneurons, parvalbumin, and calretinin (Fig. 6E–6K). Few (3.6% ± 0.7%) of the grafted cells were calretinin+ (Fig. 6I–6K), whereas a substantial proportion was HuNu+/parvalbumin+ (15.4% ± 2.6%) (Fig. 6E–6H). Interestingly, at 4 months, a small population (4.3% ± 1.3%) of the HuNu+ cells in the intrastriatal grafts expressed DARPP-32 (Fig. 6L–6O), a specific marker of medium-sized spiny striatal projection neurons.
Human iPSC-Derived lt-NES Cells Survive for 4 Months and Differentiate to Neurons After Transplantation into Cerebral Cortex of T-Cell-Deficient Rats
We analyzed the survival and neuronal differentiation of human iPSC-derived lt-NES cells implanted into the cortex of animals subjected to the occlusion of the distal branch of MCA (dMCAO), which causes cortical damage. Seven out of 10 rats that survived these procedures were killed after 2 (n = 2) or 4 (n = 5) months. The number of cells in the grafts at 2 and 4 months after transplantation was estimated at 80.8% ± 6.2% and 60.1% ± 10.6%, respectively, compared to the number of implanted cells. The percentage of cells with proliferative activity resembled that of the intrastriatal grafts (7.2% ± 1.5% and 0.9% ± 0.2%, in 2- and 4-month groups, respectively). NeuN was expressed by 71.8% ± 8.6% of the GFP+ cells and 15.5% ± 2.7% of HuNu+ cells was also GABA+ at 2 months. Many of the grafted cells displayed neuronal morphology as shown by GFP (Fig. 5V) and SC-121 (Fig. 5W) immunostainings. Although the percentage of GFP+/NeuN+ cells remained high at 4 months after transplantation (77.2% ± 3.2%), the GABA+ cells in the grafts decreased to 8.8% ± 0.8%. At this time point, we found populations of parvalbumin+ (11.9% ± 1%), calretinin+ (2.0% ± 0.5%), and calbindin+ cells (8.1% ± 0.8%; Fig. 6P–6S) in the intracortical grafts. Also when transplanted into cerebral cortex, the iPSC-derived lt-NES cells differentiated to cells expressing DARPP-32 (Fig. 6T–6W). The percentage of grafted HuNu+/DARPP-32+ cells (5.2% ± 1.5%) resembled that observed in the intrastriatal grafts.
Neurons Generated from Grafted Human iPSC-Derived lt-NES Are Functional and Receive Synaptic Input from Host Neurons
We finally studied the electrophysiological properties of cells generated from iPSC-derived lt-NES cells at 5 months after transplantation into striatum or cerebral cortex of Nude rats by performing whole-cell patch-clamp recordings in acute brain slice preparations (Fig. 7A–7G). Grafted cells were classified as neuronal or glial based on resting membrane potential and input resistance (Supporting Information Table S3). A proportion of the cells was functionally mature neurons with a resting membrane potential of approximately −50 mV and −58 mV in the striatum and cortex, respectively, the majority generated action potentials in response to depolarizing current injection (Fig. 7H, 7M). In voltage-clamp configuration, depolarizing voltage steps induced characteristic Na+ and K+ whole-cell currents, which were sensitive to the voltage-gated Na+ channel blocker, TTX, and the voltage-gated K+ channel blocker, TEA, respectively (Fig. 7I, 7N).
Spontaneous postsynaptic currents were frequently observed in grafted cells in the striatum (Fig. 7J, 7O). Spontaneous excitatory postsynaptic currents (sEPSCs) were recorded in the presence of the GABAA receptor antagonist, PTX (Fig. 7K, 7Q), and addition of glutamate receptor antagonists NBQX and D-AP5 blocked all postsynaptic currents (Fig. 7L), indicating that the cells had functional excitatory synapses. Interestingly, spontaneous inhibitory postsynaptic currents were not observed (n = 5 cells).
In addition, we detected miniature postsynaptic currents that were action potential independent (not blocked by TTX) (Fig. 7P). In two of 10 grafted cells, excitatory, AMPA receptor-mediated currents could be evoked by stimulating a cortical region remote from the transplant (Fig. 7S). This finding suggests that transplanted neurons generated from iPSC-derived lt-NES cells receive synaptic input from host neurons and functionally integrate into host brain neural circuitries.
We show here that intrastriatal transplantation of lt-NES cells, derived from human iPSCs, gives rise to improved recovery of a clinically relevant motor deficit. Although we found that at 10 weeks the majority of cells in the surviving grafts expressed neuronal markers, recovery was probably promoted earlier, before any functioning neurons could have been developed from the grafted human iPSC-derived lt-NES cells. Most likely, this early improvement is due to mechanisms other than neuronal replacement. In support of this interpretation, the enhancement of behavioral recovery in our experiment was observed irrespective of whether the grafts were surviving at 10 weeks or not. Similarly, human iPSCs improved performance in the cylinder test from 4 days after intrastriatal implantation into stroke-damaged rats .
Our hypothesis that the increased levels of VEGF in the lt-NES cell-transplanted animals could have an important role in the improved functional recovery was triggered by the study of Horie and coworkers . They transplanted a different type of cell (NSCs derived from human fetus) in Nude rats subjected to distal MCAO. They observed early functional recovery and a transient increase of blood vessel density in the peri-infarct region of animals subjected to transplantation. Secretion of VEGF by the NSCs was found to be important for the suppression of inflammation and neovascularization in the peri-infarct region probably underlying the improved recovery . Our data provide no evidence that the elevated VEGF immunoreactivity observed at 2 weeks after stroke in lt-NES cell-grafted animals gives rise to a long-lasting increase of blood vessel length density. Therefore, it is conceivable that mechanisms other than VEGF-induced vascularization and angiogenesis are responsible for the long-term behavioral improvement observed in our lt-NES cell-grafted animals. We hypothesize that VEGF leads to early and long-lasting behavioral improvement by influencing brain plasticity in the postischemic phase . In line with such a role of VEGF, it was recently reported that conditional transient increase of VEGF expression for 5–15 days leads to significant increase of long-term potentiation in the hippocampus through modulation of synaptic plasticity in mature neurons . In addition, VEGF has been implicated as a mediator of human NSC-induced effects on dendritic sprouting, axonal plasticity, and axonal transport . It should be emphasized, although, that increased VEGF signaling is only one possible explanation for the beneficial effects and other mechanisms, not explored in this study, could be responsible for the observed improved behavioral performance.
One can speculate that the VEGF immunoreactivity observed in the host tissue is due to VEGF released from the grafted lt-NES cells. However, the widespread distribution of VEGF immunoreactivity far away from the transplant without any clear gradient, with very little if any extracellular appearance and with close association to astrocytes and blood vessels, makes this suggestion unlikely.
The studies with human iPSCs in stroke published so far provided no evidence for the formation of neuroblasts or morphologically mature neurons [25, 26]. We found that at 4 months after human iPSC-derived lt-NES cells had been implanted into the striatum or cerebral cortex of T-cell-deficient rats in two different models of stroke, 73% and 77%, respectively, of grafted cells expressed the mature neuronal marker NeuN. These data indicate that grafted iPSC-derived lt-NES cells can efficiently differentiate to mature neurons in vivo and survive long-term in the stroke-damaged brain.
The grafted cells exhibited characteristic neuronal morphologies and expressed markers providing evidence that they had differentiated to various subtypes of neurons expressing GABA, parvalbumin, calretinin, or calbindin. At 4 months, the grafts also contained a population of cells expressing the specific striatal projection neuron marker DARPP-32 not only in the striatum but also in the cortex, where only weakly stained DARPP-32+ neurons are normally seen . Two important conclusions of relevance for cell replacement strategies can be drawn from our data. First, that the spontaneous neuronal differentiation of the grafted human iPSC-derived lt-NES cells in vivo is determined by intrinsic mechanisms and not significantly affected by external cues. Second, that effective neuronal replacement in different stroke-damaged brain regions will require directed in vitro differentiation of the human iPSC-derived lt-NES cells to precursors, which give rise to the lost types of neurons after transplantation.
Our electrophysiological data show, for the first time, that transplanted NSCs generated from reprogrammed adult human fibroblast-derived iPSCs can differentiate into functional neurons in vivo and receive proper connections from the host brain. Similarly, grafted human ESC-derived NSCs have previously been shown to differentiate to cells with electrophysiological properties of neurons and synaptic input from host neurons in a stroke-injured environment . We found here that at about 5 months after transplantation into striatum or cortex, grafted cells exhibited the electrophysiological properties of mature neurons. The findings of sEPSCs, which were blocked by glutamate receptor antagonists, and of AMPA-receptor-mediated currents evoked by stimulating a cortical region remote from the intracortical graft, provide evidence that the grafted neurons had developed excitatory synaptic input and were, at least partly, functionally integrated into host neural circuitries.
Our findings using the retrograde tracer FG indicated that grafted cells have capacity to establish axonal projection to the GP. This brain region is a major projection area of striatal neurons, and following a stroke, there is extensive injury to the striatopallidal system. Therefore, the ability of human-derived iPSCs to reestablish this system could potentially be important for restoration of function. Our data further illustrate the remarkable capacity of neurons derived from xenografted human iPSCs , ESCs , and fetal brain tissue  to extend specific, long axonal projections to their appropriate target areas in the rodent brain.
Tumor formation and graft overgrowth [45–47] from pluripotent stem cells and their derivatives are major hurdles for the application of stem cell therapy in stroke and other brain diseases. In previous studies, human iPSCs showed high tumorigenicity after intracerebral implantation in stroke-damaged brain . Here, we used transplantation of iPSCs-derived lt-NES cells into immunocompromised, T-cell-deficient rats subjected to stroke and a 4 months observation period to ensure lack of rejection of the xenotransplanted cells and to create optimum environment for assessing their tumorigenicity. No tumor formation or transplant overgrowth was detected in any animal grafted with iPSC-derived lt-NES cells. Our data suggest that predifferentiation of iPSCs and generation of long-term self-renewing neural cell lines is an effective strategy for minimizing the risk for tumor formation.
Our experimental study provides evidence that transplantation of human fibroblast-derived iPSCs is a new, promising approach for promoting functional recovery after ischemic stroke. Most importantly, our findings demonstrate for the first time the potential of iPSCs for neuronal replacement in the stroke-damaged brain. The iPSC-derived grafts survived long-term and contained a high proportion of cells with morphological and electrophysiological properties of neurons. These neurons received afferent inputs from the host brain and extended their axons to an appropriate target area. Before any clinical application can be considered, the mechanisms regulating the differentiation of the iPSC-derived lt-NES cells into specific neuron types in the stroke-damaged brain have to be much better understood and effectively controlled, and the behavioral recovery induced by neuronal replacement has to be optimized.
We thank Dr. Vladimer Darsalia and Vivek Verma for their contribution in initial stages of the study, Dr. Andreas Toft Sørensen for advice regarding electrophysiological recordings, Camilla Ekenstierna for technical assistance, and the Institute of Human Genetics (LIFE & BRAIN Center) for single nucleotide polymorphism analysis. This work was supported by the Swedish Research Council, the Swedish Government funding for the Strategic Research Areas of Stem Cells and Regenerative Medicine (StemTherapy), AFA insurance, EU 7th Framework Programs “European Stroke Network” (Grant number 201024), “NeuroStemCell” (Grant number 22943), TargetBraIn (Grant number 279017), BMBF Grants 01GN0813 and 01GN1009B, DFG Grant SFB-TR3 D2, and the Hertie Foundation. D.T.P. was supported by grant from Instituto de Salud Carlos III (Spain).
DISCLOSURE OF POTENTIAL CONFLICTS OF INTEREST
The authors indicate no potential conflicts of interest.