Isolation and Characterization of Cerebellum-Derived Stem Cells in Poststroke Human Brain

Patients

All procedures performed in studies involving human participants were in accordance with the ethical standards of the institutional and/or national research committee and with the 1964 Helsinki declaration and its later amendments or comparable ethical standards. Informed consent was obtained from all individual participants included in the study. The Ethics Committee of Hyogo College of Medicine reviewed and approved the study protocol (approval no.: 1776, 0385). Brain samples were obtained from poststroke areas of two patients who satisfied the following major criteria: [1] requiring both decompressive craniectomy and partial lobectomy for diffuse cerebellar infarction, [2] >20 years of age, and [3] willing to provide written informed consent. Detailed methods and criteria were described previously [14].

Histological examination

Histological analyses were performed as described [14]. In brief, tissue samples obtained from the cerebellar cortex of patients with cerebellar infarction were fixed with periodate-lysine-paraformaldehyde, cryoprotected in 30% sucrose, frozen at −80°C, and cut into 20-μm sections using a cryostat. Sections were then stained with hematoxylin and eosin (H&E) and examined by immunohistochemistry as described below.

Immunohistochemistry

Immunohistochemistry was performed as described [14]. In brief, sections were labeled with antibodies against Tuj1 (1:1,000; STEMCELL Technologies, Vancouver, BC, Canada), microtubule-associated protein 2 (MAP2, 1: 500; Sigma, St. Louis, MO), glial fibrillary acidic protein (GFAP, 1:200; Millipore, Temecula, CA), S-100β (1:100, Abcam; Cambridge, United Kingdom), nestin (1:100, Santa Cruz Biotechnology; Dallas, TX), von Willebrand factor (1:100, vWF; Santa Cruz Biotechnology), neuron-glial antigen 2 (NG2, 1:200; Millipore), and alpha smooth muscle actin (αSMA, 1:100; LifeSpan Biosciences, Seattle, WA). Primary antibodies were visualized using Alexa Fluor 488- or 555-conjugated secondary antibodies (1:500; Molecular Probes, Eugene, OR). Nuclei were counterstained with 4′,6-diamino-2-phenylindole (DAPI, 1:500; Kirkegaard & Perry Laboratories, Gaithersburg, MD). Images of brain sections were captured using a laser scanning microscope (LSM780, Carl Zeiss AG; Oberkochen, Germany) or fluorescence microscopy (Keyence BZ-X700; Keyence, Osaka, Japan). For negative controls in the immunohistochemistry experiments, we omitted the primary antibodies. No staining was observed in the control sections.

Cell culture and immunocytochemistry

The traits of cl-iSCs were investigated as described [14]. In brief, human ischemic tissue samples obtained from the cerebellar cortex were mechanically dissociated by passage through 23- and 27-G needles to yield single-cell suspensions. Suspensions were seeded on 10-cm dishes (Corning, NY) in Dulbecco's modified Eagle's medium/F12 (DMEM/F12; Invitrogen, Carlsbad, CA) containing basic fibroblast growth factor (bFGF, 20 ng/mL; Peprotech, Rocky Hill, NJ), epidermal growth factor (EGF, 20 ng/mL; Peprotech), 10% fetal bovine serum (FBS), and N2 (1%; Invitrogen). When the cells reached confluence, they were treated with trypsin and reseeded under the same conditions.

Passage 5–10 cells (cl-iSCs) were examined by immunocytochemistry, reverse transcriptase-polymerase chain reaction (RT-PCR), and FACS. Immunocytochemistry was performed using antibodies against Ki67 (1:50; BD Pharmingen, San Jose, CA), nestin (1: 100; Santa Cruz Biotechnology), αSMA (1: 200; LifeSpan Biosciences), NG2 (1: 200; Millipore), platelet-derived growth factor receptor beta (PDGFRβ, 1: 200; Santa Cruz Biotechnology), GFAP (1: 1000; DAKO, Glostrup, Denmark), CD31(1: 50; Santa Cruz Biotechnology), and CDH5 (1: 50; Santa Cruz Biotechnology), followed by Alexa Fluor 488- or 555-conjugated secondary antibodies (1: 500; Molecular Probes).

For mesoderm differentiation, cl-iSCs were incubated in medium to promote osteogenic, adipogenic, or chondrocytic differentiation according to the manufacturer's protocol (SC 006; R&D systems, Minneapolis, MN). The medium was changed every three days. Two weeks after incubation, each specimen was subjected to immunocytochemistry using antibodies (SC 006; R&D systems) against osteocalcin, fatty acid binding protein 4 (FABP4), or aggrecan. In addition, iSCs incubated in adipogenic differentiation medium were stained with the lipid-specific dye Oil Red O.

For neuronal differentiation, cl-iSCs were incubated for 1 week in DMEM/F12 containing bFGF, leukemia inhibitory factor (LIF; Millipore), 10% FBS, and N2 to promote the formation of neurosphere-like cell clusters, as previously described [14]. The medium was changed every three days. Then, cell clusters were picked up and subjected to RT-PCR and immunocytochemistry. The differentiated cells were immunostained with antibodies against Tuj1 (1:500; STEMCELL Technologies), neurofilament (NF, 1:200; Millipore), and MAP2 (1:500; Cell Signaling Technology, Danvers, MA) followed by Alexa Fluor 488- or 555-conjugated secondary antibodies (1: 500; Molecular Probes).

Human pediatric brain-derived endothelial cells (ECs, ACBRI-376; Cell Systems, Kirkland, WA), human adult bone marrow (BM)-derived MSCs (PT-2501; Lonza, Basel, Switzerland), human NSPCs derived from H9 embryonic stem cells (NSPCs, Part. No. SCC003; Millipore), human fetal brain-derived astrocytes (ACs, #1800; ScienCell Research Laboratories, Carlsbad, CA), and human fetal brain-derived pericytes (PCs, #1200; ScienCell Research Laboratories) were maintained in media specified by the supplier.

Briefly, ECs were planted in 10-cm dishes (Corning, NY) and maintained in EGM-2 medium containing vascular endothelial growth factor, bFGF, EGF, insulin-like growth factor, hydrocortisone (CC3156; Lonza), and 10% FBS. MSCs were planted in T25 flasks (Corning, NY) and maintained in mesenchymal stem basal medium (Lonza) containing MSC growth supplement (Lonza), l-glutamine, and 10% FBS. NSPCs were plated in poly-L-ornithine (20 μg/mL) and laminin (5 μg/mL) double coated dishes and maintained in ENStem-A Expansion Medium (SCM 004; Millipore) containing bFGF and l-glutamine. ACs were planted in 10-cm dishes (Corning, NY) and maintained in Astrocyte Medium (#1801; ScienCell).

The ECs were immunostained with antibodies against CD31 and CDH5 as a positive control. Extracted RNA samples from ECs, MSCs, NSPCs, ACs, and PCs were used as controls for microarray analyses of cl-iSC gene expression patterns.

Cell proliferation

To evaluate the proliferative capacity of cl-iSCs, 1 × 10⁴ cells were plated onto 12-well plates in DMEM/F12 containing bFGF, EGF, 10% FBS, and N2. Then, the numbers of cl-iSCs were counted manually over 7 days (N = 4). Doubling time was calculated as described [14,18].

Reverse transcriptase-polymerase chain reaction

Total RNA was isolated from cl-iSCs and ACs using an RNeasy Micro Kit (Qiagen, Hilden, Germany). cDNA was amplified according to the manufacturer's protocol as described [9 –11,14,19]. The primer sequences used in this study are listed in Supplementary Table S1.

Microarray analysis

Total RNA was isolated from cl-iSCs from the first patient (cl-iSCs-1) and second patient (cl-iSCs-2), as well as from ECs, MSCs, NSPCs, ACs, and PCs using an RNeasy Micro Kit (Qiagen, Hilden, Germany). In addition, total RNA was obtained from two different cerebral stroke patients (cr-iSCs-1, cr-iSCs-2) [14]. For microarray analysis, three samples were used for each cell type. Microarray analysis, including RNA quality assessment, labeling, hybridization, scanning, and data analysis, was performed by a contract service (Takara Bio, Inc., Shiga, Japan). In brief, biotinylated cRNA was synthesized using the GeneChip 3′IVT PLUS Reagent Kit (Affymetrix; Thermo Fisher Scientific, Inc., Waltham, MA) from 250 ng total RNA. Biotinylated cRNA production was checked using a NanoDrop ND-2000 Spectrophotometer. Following fragmentation, 15 μg of cRNA was hybridized for 16 h at 45°C on the GeneChip Human Genome U133 Plus 2.0 Array (Thermo Fisher Scientific, Inc.). Arrays were washed and stained in the GeneChip Fluidics Station 450 (Affymetrix) and then scanned using GeneChip Scanner 3000 7G. Single Array Analysis was calculated by Microarray Suite version 5.0 (MAS5.0) with Affymetrix default settings and global scaling as the normalization method. The trimmed mean target intensity of each array was arbitrarily set to 500. Principal component analysis (PCA) and unsupervised hierarchical clustering were performed on samples based on normalized expression of all genes (total of 54,613 ID genes).

Flow cytometric analysis

Suspended single cl-iSCs, hMSCs, and hNSPCs were labeled with antibodies against CD19, CD44, CD45, CD90, CD105, CD166 (SC017; R&D Systems, Minneapolis, MN), and CD133 (Proteintech, Rosemont, IL) according to the manufacturer's instructions. Cell subpopulations were quantified using a fluorescence-activated cell sorter (FACS, BD LSRFortessaÔX-20; BD Pharmingen) as described [14].

Electrophysiology by multielectrode arrays

Electrophysiological properties were analyzed using multielectrode arrays (MEA2100 Lite, Multi Channel Systems MCS GmbH, Reutlingen, Germany) as described [13,20,21]. In brief, cl-iSCs (80 μL of a 1 × 10⁵ cells/mL suspension) were plated onto 60 electrode dishes (60MEA200/30) precoated with Matrigel™ (Corning). During experimental recordings, the medium, including growth factors, was changed every 3 days. The spontaneous activity was recorded by the MEA2100-Lite system (Multi Channel Systems). The electrophysiological signals were detected using MC_Rack software (Multi Channel Systems), and the data were digitized with a 60-channel A/D converter at rate of 20 kHz with 16-bit resolution. Data obtained from multielectrode were converted to Axon Binary File format and analyzed using LabChart 8.0 Japanese software (AD Instruments Japan, Inc., Nagoya, Japan).

Analysis of neurogenic potential by single cell from cl-iSCs

Individual clone cells were isolated by the limited dilution method as described [22]. In brief, single-cell suspensions of cl-iSCs (passage 3) were plated on 96-well plates (1 cell per well) in DMEM/F12 containing bFGF, EGF, 10% FBS, and N2. The medium was changed every 2 days. On day 26 after incubation, expanded cl-iSCs were treated with trypsin and reseeded under the same conditions. For neurogenic differentiation, iSCs expanded from single cells were further incubated in a new medium containing DMEM/F12 plus bFGF, LIF, 10% FBS, and N2 and then examined by immunohistochemistry or multielectrode arrays as described above.

Statistical analysis

Data are presented as mean ± standard deviation. Differences were analyzed using Student's t-test. A gene was considered differentially expressed when its adjusted P value was lower than 0.05 (P < 0.05).

Clinical history of two patients with cerebellar infarction undergoing decompressive lobectomy

The first case (Case 1, Supplementary Table S2) was an 84-year-old female hospitalized at the cardiovascular division of our hospital for postoperative management of percutaneous coronary intervention (PCI). On day 6 after PCI, she suddenly complained of nausea and consciousness disturbance. She had paraparesis, and her NIH stroke scale score was 39 (severe). Diffusion-weighted magnetic resonance imaging (MRI) showed a large early hyperintensity in the basilar artery territory (Fig. 1A), and magnetic resonance angiography (MRA) showed basilar artery occlusion (Fig. 1B). Given the risk of mortality, she received mechanical thrombectomy. Although reperfusion was achieved, there was no improvement of consciousness disturbance. On the day following thrombectomy, computed tomography (CT) showed upward tentorial herniation (Fig. 1C). She received internal and external decompression 36 h after ischemic onset (day 6). Postoperative CT showed successful evacuation of the infarct areas (Fig. 1D).

OPEN IN VIEWER

FIG. 1.

Neuroimaging of two patients with cerebellar infarction. Diffusion-weighted MRI (A, E) and MRA (B) on admission. Head CT just before (C, F, G) and after surgery (D, H). (Case 1: A–D, Case 2: E–H). CT, computed tomography; MRI, magnetic resonance imaging; MRA, magnetic resonance angiography.

The second case (Case 2, Supplementary Table S2) was an 80-year-old male who visited his doctor due to gait ataxia. Diffusion-weighted MRI showed right cerebellar infarction (Fig. 1E). Thus, he was immediately transferred to our hospital. We started medical treatment, but on day 5 after admission, his consciousness got worse. Head CT revealed a large low-density area in the right cerebellar hemisphere with brainstem compression and hydrocephalus (Fig. 1F, G). Emergent decompressive craniectomy and partial lobectomy were performed 5 days after stroke onset. Postoperative CT confirmed successful decompressive lobectomy (Fig. 1H).

Histological examination of poststroke cerebellar tissue sections

We performed histological examination of poststroke brain tissue sections obtained from the first patient (Fig. 2A). H&E staining revealed numerous pyknotic cell nuclei indicative of apoptosis/necrosis in the cerebellar cortex, including the molecular, Purkinje cell, and granular layers (Fig. 2B). No mature neural cells, including neurons [Tuj1⁺ cells (Fig. 2C) and MAP2⁺ cells (Fig. 2D)] and ACs [GFAP⁺ cells (Fig. 2E) and S100β⁺ cells (Fig. 2F)], were found in the molecular layer of the cerebellar cortex. These findings suggest that most of neurons and ACs had already disappeared due to apoptosis/necrosis. This is also consistent with previous findings showing that under pathological conditions, such as after ischemic stroke, mature neural cells within ischemic areas undergo cell death within a couple of days [14,23].

OPEN IN VIEWER

FIG. 2.

Histological findings of poststroke human cerebellum obtained from the first patient. Hematoxylin and eosin staining of surgically excised tissue (A) revealed nuclear pyknotic changes within ischemic areas (B), indicating apoptotic/necrotic changes. Immunohistochemistry revealed very few Tuj1⁺ or MAP2⁺ neurons [Tuj1 (C: green ), MAP2 (D: green )] and GFAP⁺ or S100β⁺ ACs [GFAP⁺ (E: green ), S100β (F: green )]. All cells were counterstained with DAPI (C–F: blue ). In contrast, immunohistochemical analysis confirmed the presence of nestin⁺ cells [nestin (G: green ), DAPI (G: blue )] localized near vWF⁺ ECs within these regions [nestin (H: green ), vWF (H: red ), and DAPI (H: blue )]. Nestin⁺ cells expressed NG2 [nestin (I, J: green ), NG2 (I, K: red ), and DAPI (I–K: blue )] (arrows) and αSMA [nestin (L, M: green ), αSMA (L, N: red ), and DAPI (L–N: blue )] (arrows). Scale bars = 2 mm (A), 25 μm (L), and 50 μm (B–I). αSMA, alpha smooth muscle actin; MAP2, microtubule-associated protein 2; ECs, endothelial cells; ACs, astrocytes; αSMA, alpha smooth muscle actin; GFAP, glial fibrillary acidic protein; NG2, neuron-glial antigen 2.

Using a mouse model of cerebral infarction, we and others previously identified putative iSCs within ischemic areas [9 –13]. Very recently, we also demonstrated putative iSCs in the human ischemic cerebrum [14]. These findings led us to hypothesize that putative iSCs might also be present in the human ischemic cerebellum. Cerebral iSCs expressed the NSPC marker nestin [9 –13], so we investigated the presence of nestin⁺ cells in the ischemic cerebellum of the first patient. Immunohistochemistry revealed nestin⁺ cells in the molecular layer of the cerebellar cortex (Fig. 2G).

Our previous studies showed that nestin⁺ iSCs are likely derived from brain PCs following ischemia (ischemic pericytes; iPCs) , so we further investigated the localization of nestin⁺ cells. Immunohistochemistry revealed the presence of nestin⁺ cells near vWF⁺ ECs (Fig. 2H) in the molecular layer of the cerebellar cortex. In addition, nestin⁺ cells in the molecular and granular layers of the cerebellar cortex coexpressed the pericytic markers NG2 (Fig. 2I–K) and αSMA (Fig. 2L–N), respectively. Similar histological characteristics were observed in the ischemic cerebellum tissue from the second patient (Supplementary Fig. S1A–D).

Characterization of cerebellar stem cell origin and marker phenotype

We further investigated the phenotypic characteristics of these nestin⁺ cells by polymerase chain reaction (PCR) and immunocytochemistry. Ischemic brain tissue obtained from the first patient was mechanically dissociated into single cells by passage through needles, and these dissociated cells were incubated as adherent monolayer cultures under conditions promoting iSC proliferation. Adherent cells (Fig. 3A) were highly proliferative, and immunohistochemistry revealed expression of the dividing cell marker Ki67 (Fig. 3B) and nestin (Fig. 3C). These cells also expressed pericytic markers, including αSMA (Fig. 3D–F), NG2 (Fig. 3G–I), and PDGFRβ (Fig. 3J–L), results confirmed by PCR (Fig. 3P).

OPEN IN VIEWER

FIG. 3.

Isolation and characterization of human cl-iSCs from the first patient. Immunohistochemistry showed that putative cl-iSCs (A) widely expressed the proliferative marker Ki67 (B: green ) and nestin (C: green ). Cells were counterstained with DAPI (B, C: blue ). In addition to nestin (D, E, G, H, J, K, M, N: green), these cells expressed αSMA (D, F: red ), NG2 (G, I: red ), and PDGFRβ (J, L: red ), but not GFAP (M, O: red ). Cells were counterstained with DAPI (D–O: blue ). Putative cl-iSCs expressed pericytic (P) but not astrocytic markers (Q) as determined by PCR. Putative cl-iSCs did not express CD31 and CDH5 according to histochemical staining. In contrast, control human ECs expressed these markers [CD31 (R: green ), CDH5 (R: green )]. Cells were counterstained with DAPI (R: blue ). Expression patterns of EC-related genes (eg, KDR, PECAM1, CDH5, NOS3, TIE1) and PC-related genes (ACTA2, PDGFRβ, and CSPG4) were also compared between ECs and cl-iSCs obtained from the first patient by DNA microarray analysis (S). Scale bars = 50 μm (A, R), 100 μm (B, C), and 20 μm (D, G, J, M). PCs, pericytes; cl-iSCs, cerebellar-induced multipotent stem cells; PCR, polymerase chain reaction; PDGFRβ, platelet-derived growth factor receptor beta.

To exclude the possibility that these nestin⁺ cells are reactive ACs [24], we further investigated whether the nestin⁺ cells expressed the astrocytic marker GFAP but found no such expression (Fig. 3M–O). Furthermore, no GFAP mRNA expression was detected by PCR using two different GFAP primers, although GFAP mRNA expression was detectable in commercially obtained ACs using the same PCR methods (Fig. 3Q). In addition, heat mapping and scatter plot analysis showed strong expression of astrocytic markers GFAP, S100β, aquaporin4 (AQP4), and FABP7 in ACs compared to cl-iSCs (Supplementary Fig. S2A–D). These results strongly indicate that these nestin⁺ cells are not of the astrocytic lineage.

We then examined whether putative cl-iSCs contain EC populations, because it has been reported that nestin⁺ cells can also originate from vascular lineage cells, including ECs [25,26]. Immunohistochemistry revealed no detectable expression of EC markers, such as CD31 and CDH5, by cl-iSCs, although expression of these markers was observed in human ECs (ECs) used as a positive control (Fig. 3R). Scatter plot analysis comparing global gene expression profiles between cl-iSCs-1 and ECs revealed stronger expression of pericytic-related genes [eg, ACTA2 (αSMA), CSPG4 (NG2), PDGFRβ] in cl-iSCs (>2-fold higher than in ECs, red) and stronger expression of EC-related genes [eg, PECAM1 (CD31), CDH5, KDR (Flk1), TIE1, and NOS3 (eNOS)] in ECs (>2-fold higher than in cl-iSCs, green) (Fig. 3S). Similar lineage traits were confirmed in putative cl-iSCs obtained from the second patient (cl-iSCs-2) (Supplementary Fig. S3A–T).

Multipotency of cerebellar iSCs and differentiation potential to electrophysiologically active neurons

We next investigated whether cl-iSCs have multipotency. They displayed a high proliferation rate as demonstrated by a doubling time of 65.8 ± 6.3 h (Fig. 4A). In addition, cl-iSCs maintained expression of various stem cell and undifferentiated cell markers even after several passages, including nestin, c-myc, Klf4, and Sox2 (Fig. 4B). This finding suggests that cellular phenotypes regarding stemness are not variable during in vitro treatment (eg, repeated passages, chemical stimulation, and so on), at least under the tested conditions. To confirm this multipotency, cl-iSCs were incubated separately under osteoblastic, adipogenic, and chondrocytic differentiation conditions. Under these respective conditions, cl-iSCs differentiated into mesodermal lineages, including osteocalcin⁺ osteoblasts (Fig. 4C), FABP4⁺ and Oil red O⁺ adipocytes (Figs. 4D, E), and aggrecan⁺ chondrocytes (Fig. 4F).

OPEN IN VIEWER

FIG. 4.

Multipotent stem cell activity of human cl-iSCs obtained from the first patient. Putative cl-iSCs demonstrated high proliferative potential (A). PCR analysis showed that cl-iSCs expressed various genes for stem and undifferentiated cells even after several passages (P5, passage 5; P8, passage 8; P10, passage 10) (B). cl-iSCs differentiated into multilineage cells, including osteocalcin⁺ osteoblasts (C: green ), FABP4⁺ adipocytes (D: red ), Oil red O⁺ adipocytes (E: red ), and aggrecan⁺ chondrocytes (F: green ). Cells were counterstained with DAPI (C, D, F: blue ). cl-iSCs formed neurosphere-like cell clusters in floating culture (G, H). PCR analysis confirmed that they expressed neuronal genes (I). Immunohistochemistry showed that the cell clusters differentiated into Tuj1⁺ and NF⁺ neuronal cells [Tuj1 (J, K: red ), NF (J, L: green ), and DAPI (J–L: blue )]. Multielectrode array (M) recordings confirmed that cl-iSCs differentiated into electrophysiologically active neurons (N) expressing Tuj1 [Tuj1 (O: green ) and DAPI (O: blue )]. Arrows (M, O) indicate the position where the waveform (N) was observed. Scale bars = 50 μm (C–F, J), 100 μm (H, M, O). NF, neurofilament; FABP, fatty acid binding protein.

To investigate whether cl-iSCs can also differentiate into neuronal lineages, adherent cells were incubated in floating cultures under conditions that promote the formation of neurosphere-like cell clusters (Fig. 4G) [9,10,27,28]. On day 4 of incubation, cell clusters were observed (Fig. 4H), and PCR analysis revealed expression of the neuronal markers Tuj1, NF, and MAP2 (Fig. 4I). Immunohistochemistry also showed that these clusters produced Tuj1⁺ and NF⁺ neuronal cells (Fig. 4J–L). Furthermore, electrophysiological recordings using multielectrode arrays (Fig. 4M) showed that cl-iSCs exhibited the potential to differentiate into electrophysiologically active neurons (Fig. 4N). In addition, the presence of Tuj1⁺ neuronal cells on microelectrode dishes was confirmed by immunohistochemistry (Fig. 4O).

To exclude possible contamination by mature neurons in cl-iSC populations, single-cell suspensions made from cl-iSCs were diluted in medium and plated in 96-well plates as described [22]. One day after plating, wells successfully plated (1 cell per well) were marked and incubated for further analysis (Fig. 5A). Single cells gradually expanded under conditions that promote the proliferation of cl-iSCs (Fig. 5B). Then, the capacity of expanded cl-iSCs to produce functional neurons was investigated by further incubation under conditions that promote differentiation to neuronal lineages. Indeed, these cells expressed neuronal markers, including Tuj1, NF, and MAP2 (Fig. 5C–H). In addition, these cells were electrically functional (Fig. 5I–K). Collectively, these results confirm that cl-iSCs can give rise to functional neurons. Similar multipotency was confirmed in cl-iSCs obtained from the second patient (cl-iSCs-2) (Supplementary Fig. S4A–O), and multielectrode arrays revealed stimulus-evoked responses (Supplementary Fig. S5).

OPEN IN VIEWER

FIG. 5.

Neurogenic potential of single-cell suspensions from cl-iSCs. Using the limited dilution assay, single-cell suspensions obtained from cl-iSCs were plated in 96-well plates (A). Under conditions promoting proliferation, the numbers of cl-iSCs (arrows) gradually increased (B). Following incubation under conditions promoting neuronal differentiation, cl-iSCs were subjected to immunohistochemistry and multielectrode array analysis. Immunohistochemistry showed that cl-iSCs differentiated into Tuj1⁺ and NF⁺ neurons [Tuj1 (C, D: red), NF (C, E: green), and DAPI (C–E: blue)] or Tuj1⁺ and MAP⁺ neurons [Tuj1 (F, G: red), MAP2 (F, H: green), and DAPI (F–H: blue)]. Multielectrode recordings demonstrated that cl-iSCs (I) exhibited the features of electrophysiologically functional neurons (J). Immunohistochemistry confirmed that these cells expressed Tuj1 [Tuj1 (K: green), DAPI (K: blue)]. Arrows (I, K) indicate the position where the waveform (J) was observed. Scale bars = 50 μm (B), 100 μm (C, F, I, K).

Comparison of cerebellar iSCs to other types of cells, including stem cells

Our data thus far demonstrate that cl-iSCs can differentiate into both mesodermal and neural lineage cells. Thus, we next compared the traits of cl-iSCs to MSCs and NSPCs that mainly commit to the mesoderm and neural lineage, respectively. FACS analysis revealed that cl-iSCs-1 (Fig. 6A) and cl-iSCs-2 (Fig. 6B) rarely expressed hematopoietic lineage markers (CD19 and CD45) but did express various MSC markers (CD44, CD90, CD105, and CD166), similar to established MSCs (Fig. 6C). These findings are consistent with our previous results showing that cr-iSCs share several markers with MSCs [14]. NSPCs also express MSC markers [29]. However, the NSPC marker CD133 [30] was specifically expressed in established NSPCs but not cl-iSCs and MSCs (Fig. 6D).

OPEN IN VIEWER

FIG. 6.

Flow cytometric analysis of human cl-iSCs (A, B), MSCs (C), and NSPCs (D). Flow cytometric analysis demonstrated that putative human cl-iSCs from the first patient (cl-iSCs-1) and the second patient (cl-iSCs-2) expressed multiple MSC markers (CD44, CD90, CD105, and CD166), while they rarely expressed hematopoietic markers (CD19 and CD45). CD133 was exclusively expressed in NSPCs. MSCs, mesenchymal stem cells; NSPCs, neural stem/progenitor cells.

To further investigate the traits of cl-iSCs, global gene expression profiles were analyzed in cl-iSCs obtained from both patients. In scatter plot analysis comparing global gene expression profiles between cl-iSCs-1 and MSCs [log₂ (cl-iSCs-1/MSCs ratio)], ∼74% of all gene sequences were with the difference range ≥−2-fold to ≤2-fold change) [gray areas in Fig. 7A (left panel) and Supplementary Fig. S6A]. In the scatter plot analysis comparing global gene expression profiles between cl-iSCs-1 and NSPCs [log₂ (cl-iSCs-1/MSCs ratio)], ∼57% of total gene sequences were present in this range [gray areas in Fig. 7B (left panel) and Supplementary Fig. S7A]. Similarly, in global gene expression profile comparison between cl-iSCs-2 and MSCs [log₂ (cl-iSCs-2/MSCs ratio)], ∼74% of sequences were within this range [gray areas in Fig. 7A (middle panel) and Supplementary Fig. S6B]. Moreover, in the comparison between combined cl-iSCs from both patients (cl-iSCs-1,2) and MSCs [log₂ (cl-iSCs-1,2/MSCs ratio)], 77% of total gene sequences were present within this range [gray areas in Fig. 7A (right panel) and Supplementary Fig. S6C]. In the comparison of global gene expression profiles between cl-iSCs-2 and NSPCs [log₂ (cl-iSCs-2/NSPCs ratio)] and between cl-iSCs-1,2 and NSPCs [log₂ (cl-iSCs-1,2/NSPCs ratio)], ∼57% [gray areas in Fig. 7B (middle panel) and Supplementary Fig. S7B] and 59% of total gene sequences, respectively, were within the ≥−2-fold and ≤2-fold change range [gray areas in Fig. 7B (right panel) and Supplementary Fig. S7C]. These results indicate that the global gene expression pattern of cl-iSCs is closer to that of MSCs than NSPCs.

OPEN IN VIEWER

FIG. 7.

Global gene expression profiles of human cl-iSCs. Scatter plot analysis comparing global gene expression profiles between cl-iSCs and MSCs (A). Scatter plot analysis comparing global gene expression profiles between cl-iSCs and NSPCs (B). PCA mapping (C) and heat mapping analysis (D) among cl-iSCs-1, cl-iSCs-2, cr-iSCs-1, cr-iSCs-2, ECs, MSCs, NSPCs, ACs, and PCs. Scatter plot analysis comparing global gene expression profiles between cl-iSCs and cr-iSCs (E). cr-iSCs, cerebral-induced multipotent stem cells; PCA, principal component analysis.

While NSPCs are neural plate derivatives, certain brain PCs are neural crest derivatives [31]. Thus, we further compared the expression levels of genes related to neural plate or neural crest origin between NSPCs and cl-iSCs-1,2 [log₂ (cl-iSCs-1,2/NSPCs ratio)]. Expression levels of numerous neural plate-related genes (early neural genes) were lower in cl-iSCs-1,2 than NSPCs (<−2-fold change) (Supplementary Table S3), whereas many neural crest-related genes were higher in cl-iSCs-1,2 than NSPCs (>2-fold change) (Supplementary Table S4).

We further compared global gene expression patterns between cl-iSCs and cr-iSCs obtained from two different cerebral stroke patients (cr-iSCs-1, cr-iSCs-2) [14]. PCA mapping (Fig. 7C) and heat mapping (Fig. 7D) showed that the gene profiles of cl-iSCs and cr-iSCs were closer to PCs than to ECs, MSCs, NSPCs, and ACs. These data indicate that iSCs, which are likely brain PC derivatives, are present in poststroke cerebellum, as well as cerebrum.

In addition, scatter plot analysis of each pairwise comparison (cl-iSCs-1 vs. cr-iSCs-1; cl-iSCs-1 vs. cr-iSCs-2; cl-iSCs-1,2 vs. cr-iSCs-1; cl-iSCs-2 vs. cr-iSCs-1; cl-iSCs-2 vs. cr-iSCs-2; cl-iSCs-2 vs. cr-iSCs-1,2; cl-iSCs-1,2 vs. cr-iSCs-1; cl-iSCs-1,2 vs. cr-iSCs-2, cl-iSCs-1,2 vs. cr-iSCs-1,2) showed that most genes (∼74%–86%) were restricted within the −2-fold ≥ and ≤2-fold change range (Fig. 7E, Supplementary Fig. S8A–I), indicating a strong lineage resemblance between cl-iSCs and cr-iSCs.

However, PCs in forebrain regions are reported to be neural crest derivatives, while PCs in caudal regions are considered to be mesoderm derivatives [32,33]. Thus, we compared the neural crest-related gene expression levels between cl-iSCs-1,2 and cr-iSCs-1,2 [log₂ (cl-iSCs-1,2/cr-iSCs-1,2 ratio)] (Supplementary Table S5). Expression levels of ERBB3, TFAP2A, ETS1, and BGN were significantly lower in cl-iSCs than cr-iSCs, while Sox9, ID1, ID2, MGP, and FOXD3 expression levels were significantly higher in cl-iSCs than cr-iSCs. However, expression levels of most neural crest-related genes were within the −2-fold ≥ and ≤2-fold change range. Thus, we further examined the expression levels of cerebellar specific genes (Supplementary Table S6) and found that of the genes exhibiting a >1.5-fold change [log₂ (cl-iSCs-1,2/cr-iSCs-1,2 ratio)], the cerebellar granule cell marker PDEIC was significantly higher in cl-iSCs-1,2 than cr-iSCs-1,2 (P < 0.05). These results indicate that cl-iSCs resemble cr-iSCs but possess cerebellar specific traits.