Direct Conversion of Human Fibroblasts into Neuronal Restricted Progenitors*

Background: Neuronal restricted progenitors have not been generated from fibroblasts by transdifferentiation. Results: Human induced neuronal restricted progenitors (hiNRPs) were efficiently generated from fibroblasts by transfection of three defined factors: Sox2, c-Myc, and either Brn2 or Brn4. Conclusion: Unipotent neuronal restricted progenitors can be rapidly and efficiently produced from fibroblasts. Significance: This novel method will provide a new source of neurons for cellular replacement therapy of human neurodegenerative diseases. Neuronal restricted progenitors (NRPs) represent a type of transitional intermediate cells that lie between multipotent neural progenitors and terminal differentiated neurons during neurogenesis. These NRPs have the ability to self-renew and differentiate into neurons, but not into glial cells, which is considered an advantage for cellular therapy of human neurodegenerative diseases. However, difficulty in the extraction of highly purified NRPs from normal nervous tissue prevents further studies and applications. In this study, we report the conversion of human fetal fibroblasts into human induced NRPs (hiNRPs) in 11 days by using just three defined factors: Sox2, c-Myc, and either Brn2 or Brn4. The hiNRPs exhibited distinct neuronal characteristics, including cell morphology, multiple neuronal marker expression, self-renewal capacity, and a genome-wide transcriptional profile. Moreover, hiNRPs were able to differentiate into various terminal neurons with functional membrane properties but not glial cells. Direct generation of hiNRPs from somatic cells will provide a new source of cells for cellular replacement therapy of human neurodegenerative diseases.


Neuronal restricted progenitors (NRPs) represent a type of transitional intermediate cells that lie between multipotent neural progenitors and terminal differentiated neurons during
neurogenesis. These NRPs have the ability to self-renew and differentiate into neurons, but not into glial cells, which is considered an advantage for cellular therapy of human neurodegenerative diseases. However, difficulty in the extraction of highly purified NRPs from normal nervous tissue prevents further studies and applications. In this study, we report the conversion of human fetal fibroblasts into human induced NRPs (hiNRPs) in 11 days by using just three defined factors: Sox2, c-Myc, and either Brn2 or Brn4. The hiNRPs exhibited distinct neuronal characteristics, including cell morphology, multiple neuronal marker expression, self-renewal capacity, and a genome-wide transcriptional profile. Moreover, hiNRPs were able to differentiate into various terminal neurons with functional membrane properties but not glial cells. Direct generation of hiNRPs from somatic cells will provide a new source of cells for cellular replacement therapy of human neurodegenerative diseases.
ES cells and induced pluripotent stem cells are expected to be promising sources of cells for cell therapy of human diseases because of their ability to differentiate into various cell types. However, the tumorigenic potential and impurity of the differentiated cell types increase the risk for clinical application (1).
Mouse and human fibroblasts have been directly induced into terminal-differentiated neurons with different combina-tions of transcription factors (2)(3)(4)(5)(6)(7)(8). When induced neurons (iNs) 2 were transplanted into a host, only a few cells could survive and function because of their limited ability to proliferate. As a result, the treatment effectiveness of iN transplantation is not ideal. Many studies have focused on the generation of multilineage neural stem cells (NSCs) or neural progenitors (NPs) from fibroblasts (9 -14). NSCs and NPs can differentiate into neurons and glial cells, the two major types of cells in the nervous system (15). However, studies show that NSCs are more likely to differentiate into glial cells rather than functional neurons after transplantation (16,17), which is a disadvantage for neuron replacement therapy of neurodegenerative diseases.
In neurogenesis, another major type of cell, called neuronal restricted progenitor (NRP, also known as neuroblast), which has the abilities of proliferation and migration through the rostral migratory stream in the nervous system (18,19), can develop into neurons, rather than glial cells or other cell types, in vivo and in vitro (20,21). When injected into the subventricular zone, NRPs can migrate extensively and incorporate with the different regions of the brain to differentiate into various subtypes of neurons, contributing to brain plasticity and repair (19). However, the traditional acquisition of well purified NRPs through isolation from normal nervous tissue is difficult and cumbersome (18,22), which makes it impossible to acquire sufficient cells for clinical and commercial application.
In this study, we attempt to establish an approach to directly convert human fetal fibroblasts (HFFs) into human induced neuronal restricted progenitors (hiNRPs). To change fibroblasts into hiNRPs, three processes must be considered. The first one is to use factors to convert the fibroblasts into stem cells with proliferative features. Previous reports showed that Sox2, Klf4, and c-Myc were critical for proliferation and NSC induction (10,12,13). The second one is to choose the factors to promote fibroblasts to acquire the characters of NPs. Bmi1, TLX, and FoxG1 have been proven to be key factors in NP cell induction (11,23,24). The third one is to make the induced cells achieve the capacity to become neurons. The POU III family Brn2 and Brn4 conferred to the cells the tendency to become neurons (5,25). Therefore, we chose these eight factors for initial transdifferentiation trials and successfully produced hiNRPs. After a series of further experiments, we found that, by using just three defined factors (Sox2, c-Myc, and either Brn2 or Brn4), HFFs were able to be converted into hiNRPs. The successful generation of hiNRPs from somatic cells may provide a new source of neurons for replacement therapy of human neurodegenerative diseases such as Parkinson disease, Alzheimer disease, and Huntington chorea.
Generation of hiNRPs-H1 ES cell-derived human neural progenitor (hNP) cell lines (a gift from Dr. Guangjin Pan, GIBH) were used as control for characterization of hiNRPs. Human fetal fibroblasts were derived from an 8-week-old fetus retrieved from elective termination of pregnancy following local ethical approval. Fibroblast culture procedures were done as described previously (27). The time line of hiNRP induction is shown in Fig. 1A. HFFs were seeded at 1 ϫ 10 4 cells/well in 12-well plates 1 day before infection. Viruses were added to each well. The multiplicity of infection (viral titer/cell number) was about 20ϳ30 for each lentivirus. The next day, fresh fibroblast medium was added to replace the previous medium. Two days later, the induction medium was changed to complete hiNRP medium containing KnockOut TM DMEM/F-12 supplemented with StemPro NSC SFM supplement, Glutamax (2 mM), human basic FGF (bFGF, 10 ng/ml) and EGF (10 ng/ml).
Reverse Transcriptional PCR and Quantitative PCR (qPCR)-Total RNA was extracted using a Total RNA Kit II (Omega). Then, first-strand cDNA was synthesized using a PrimeScript RT reagent kit (Takara). Synthesized cDNA was subjected to RT-PCR using Premix Ex Taq TM version 2.0 (Takara) with specific primers. qPCR was performed with SYBR Premix Ex Taq TM (Takara) on CFX96 (Bio-Rad) according to the protocol of the manufacturer.
Methylation Analysis-To determine the methylation level of the nestin enhancer in different cell types, about 500 ng of genome DNA was treated with an EZ DNA Methylation TM kit (Zymo Research) according to the protocol of the manufacturer, followed by PCR. Amplified products were purified and cloned into pMD-18T vectors (Takara). Vectors were transformed into Escherichia coli. Nine to 13 clones for each sample were picked randomly and sequenced.
Karyotype Analysis-The hiNRPs were cultured in 100-mm dishes for 2 days and incubated with 50 g/ml demecolcine (Dahui Biotech) for 1 h. Cells were treated with Accutase (Sigma) and collected by centrifugation at 200 ϫ g for 5 min. Cell precipitation was resuspended in 8 ml of 0.075 M KCl and incubated for 20 min at 37°C, followed by lysis with a hypotonic buffer. Cells were then fixed in acetic acid/methanol (v/v ϭ 1:3). Metaphase chromosomes were stained with 5% Giemsa (Invitrogen) for 15 min. Cells were dropped on a cold slide and incubated at 75°C for 3 h. Finally, metaphase state chromosomes were photographed under an Olympus BX51 microscope and analyzed using Karyo 3.0 software.
Microarray Assay-Fibroblasts, hNPs, and hiNRPs were collected in TRIzol (Invitrogen), followed by total RNA extraction. RNA integrity was assessed by standard denaturing agarose gel electrophoresis. Double-stranded cDNA was synthesized from total RNA and labeled with Cy3 using a NimbleGen One-Color DNA labeling kit according to the protocol of the manufacturer. Array hybridization was performed with the NimbleGen hybridization system. The arrays were scanned with an Agilent Scanner G2505C microarray scanner. Data were further analyzed using Agilent GeneSpring GX software (version 12.1).
Differentiation of hiNRPs in Vitro-The differentiation of hiNRPs was performed as reported previously (12), with slight modifications. For generation of astrocytes, cells were cultured on Matrigel-coated culture dishes in complete hiNRP medium with the growth factors bFGF and EGF. The next day, the medium was replaced by astrocyte differentiation medium (DMEM supplemented with 1% N2, 2 mM Glutamax-I TM , and 1% FBS) and cultured for 9 days. Medium shifting was conducted every other day. For differentiation of oligodendrocytes, hiNRPs were cultured on poly-L-ornithine/laminin-coated (Sigma) dishes for 1 day. Then, the medium was replaced by an oligodendrocyte medium containing Neurobasal medium supplemented with 2% B27, 2 mM Glutamax-I TM , and 30 ng/ml T3 (Sigma) and cultured for 9 -12 days. For the production of terminal neurons from hiNRPs, the cells were plated on poly-L-ornithine/laminin-coated plates in complete hiNRP medium for 1 day, followed by a medium change to Neurobasal medium supplemented with 2% B27 and 2 mM Glutamax-I TM . To improve neuron formation, 100 M dbcAMP (Sigma), 1 M retinoic acid (Sigma), 10 ng/ml NT-3 (PeproTech), or glial cellderived neurotrophic factor (GDNF; R&D Systems) was added into the differentiation medium. The medium was changed every other day for at least 2 weeks.
Electrophysiology-Whole cell patch clamp was performed on hiNRP-derived neurons using an Axopatch 200B amplifier and Digidata 1440A interface. Data were analyzed by pClamp 10.2 software. Neurons derived from hiNRPs were attached to coverslips that were submerged in a recording chamber continually perfused with artificial cerebrospinal fluid equilibrated with 95% O 2 and 5% CO 2 . Neurons with big soma and smooth surfaces were recorded. Patch electrodes exhibited resistances of 4 M⍀ to 9 M⍀ when filled with a solution containing 140.0 mM potassium methanesulfonate, 10.0 mM HEPES, 5.0 mM NaCl, 1.0 mM CaCl 2 , 0.2 mM EGTA, 3.0 mM ATP-Na 2 , and 0.4 mM GTP-Na 2 (pH 7.3, adjusted with KOH). The voltage was clamped at Ϫ80 mV.
For the voltage-gated current (Na ϩ and K ϩ currents) recording, the cells were delivered, in 500-ms steps, from Ϫ80 to ϩ80 mV. For the action potential recording, a series of 300-ms hyperpolarizing and depolarizing step currents from Ϫ20 to ϩ50 picoamperes (pA) was injected to elicit action potentials (APs). Spontaneous postsynaptic currents were recorded without current injection.
Transplantation and Immunohistochemistry-The hiNRPs were labeled with green fluorescence by transfection with the lentiviral vector FUGW. GFP-hiNRPs were collected at a density of 1 ϫ 10 5 cells/l for transplantation. Wild-type C57BL/6 mice (8 weeks, 25-30 g) were anesthetized and fixed on a Kopf stereotaxic frame. Up to 2.5 l of cell suspension was microinjected into the lateral ventricle (anteroposterior, ϩ1.5 mm; mediolateral, Ϯ0.9 mm; dorsoventral, 3 mm below skull) for ϳ5 min using a Hamilton 7005KH 10-l syringe. After microinjection, the mice were placed under a lamp and warmed for 4 h. Then they were immunosuppressed with cyclosporine A (10 mg/kg) each day.
Three to four weeks after transplantation, the mice were anesthetized and perfused with PBS (pH 7.4) followed by 4% paraformaldehyde. The brains were collected and post-fixed for 24 h at 4°C. After being embedded in low melting point agarose, the brains were cut into 40-m consecutive sagittal sections using a Leica VT 1000S vibratome. Free floating sections were immersed into 80°C antigen retrieval solution (Beyotime) for 25 min and washed three times in PBS. Then, the sections were permeabilized and blocked using TBS (0.05 M Tris, 150 mM NaCl, and 0.5% Triton X-100) supplemented with 0.1% sodium azide, 1.0% BSA, and 5.0% normal goat serum for 1 h. Sections were incubated with primary antibodies for 48 h on a shaker at 4°C, washed in TBS three times, and incubated with fluorescence-labeled secondary antibody at room temperature for 2 h. Finally, the sections were washed with PBS and stained with DAPI solution (2 g/ml). At last, sections were observed under Zeiss 710 NLO confocal microscope and Olympus IX 51 fluorescence microscope.

Conversion of Human Fibroblasts into hiNRPs with Eight
Transcription Factors-The protocol for generation of hiNPRs from human fibroblasts is shown in Fig. 1A. We chose eight key transcription factors (Sox2, c-Myc, Klf4, TLX, Bmi1, Brn2, Brn4, and FoxG1) as the transdifferentiation inducers for pilot studies. After 5-7 days of the transfection of these factors into HFFs, colonies with small and round cells formed. The cells in the colonies proliferated more rapidly than the original fibroblasts. After culture for 5 days more, the large colonies with compacted cells were picked and seeded to a new plate. The cells were able to maintain a strong self-renewal ability after being passaged more than 15 times and being cultured for more than 2 months on Matrigel-coated plates in the presence of bFGF and EGF. On passage 15, nestin antibody was used to conduct immunostaining. We found that nestin was homogeneously expressed in the induced cells but not in fibroblasts (Fig. 1B). We chose three colonies of the nestin-positive cells to test their differentiation capacities. After cells were allowed to differentiate in neuron-generating medium for 5-6 days, cells with neuronal morphology formed. After being cultured for another 8 -10 days, the resulting cells were Tuj1-positive, displaying their neuronal identity. However, when the nestin-positive cells were cultured in astrocyte-differentiating conditions for 2 weeks, they were not able to become GFAP-positive glial cells, whereas, as a control test, ES cell-derived NPs could differentiate into both neurons in neuron-differentiating medium and glial cells in glial cell-differentiating medium (Fig. 1C). These results indicate that the nestin-positive cells are not multipotent neural progenitors but neuronal restricted progenitors.
Narrowing Down the Candidate Factors-To determine whether all eight factors are necessary for inducing fibroblasts into NRPs, we limited the number of factors by omitting the factors one by one in the next series of induction trials. After 11 days of induction, formation of the NRP-like colonies was not affected when TLX, FoxG1, Bmi1, or Klf4 was absent. NRP-like colonies were not able to form in the absence of either Sox2 or c-Myc. Therefore, both Sox2 and c-Myc are necessary factors for the trans-differentiation. We found that, when only Sox2 and c-Myc were used to conduct the induction, no NRP-like colonies were observed. However, when either Brn2 or Brn4 was transferred into fibroblasts together with Sox2 and c-Myc, NRP colonies formed, meaning that Brn2 and Brn4 can replace each other and are indispensable for the trans-differentiation from fibroblasts to hiNRPs (Fig. 1D). These results confirm that the direct conversion of HFFs into hiNRPs can be achieved by transfection with only three factors: c-Myc, Sox2, and Brn2 or Brn4 (MSB2 or MSB4 for short).
Characterization of 3F-hiNRPs-After trans-differentiation with the three transcription factors, many NRP-like colonies emerged. The cells in 3 factors induced hiNRPs (3F-hiNRPs) colonies were round and small and could be easily distinguished from the original fibroblasts (Fig. 2, A and B). The colonies were sensitive to the surrounding environment at the first five passages. The hiNRPs were able to grow and proliferate when cultured as monolayers on Matrigel-coated plates (Fig. 2C). But when grown in suspension, they were more likely to die and could hardly be expanded for further analysis (Fig. 2D). Most initial NRP-like colonies were difficult to maintain over five passages. Some of them turned back to a fibroblast-like morphology, and some of them differentiated into typical neuronal morphologies (Fig. 2, E and F). We named these cells intermediate differentiated hiNRPs (Diff-hiNRPs). However, about 18% (11 of 59) of colonies with a homogeneous population could maintain their NRP characteristics for more than 20 passages (Fig. 2, G and H).
We then tested the expression status of introduced foreign genes and found that the exogenous Sox2, c-Myc, Brn2, and Brn4 were expressed in most hiNRP cell lines. RT-PCR and qPCR analyses showed that hiNRPs expressed many neural and neuronal markers, such as Sox2, nestin, Msi1, CD133, N-CAM, DCX, Tuj1, and MAP2, but were negative for the neural stem cell-specific marker Pax6 (Fig. 3, A and B). These results were also confirmed by immunofluorescence analysis, which showed that the hiNRPs could express neuronal-specific markers such as nestin, Sox2, DCX, Tuj1, MAP2, and Msi1 ( Fig.  3C) while being silenced for Pax6 (Fig. 3D), which was expressed in NPs (data not shown). In addition, hiNRPs did not express Oct4, NeuN, and GFAP either, the markers for ES cells, mature neurons, and glia, respectively (Fig. 3D). Chemokine receptor CXCR4, a key regulator of NPC migration, was also expressed in hiNRPs (Fig. 3C), suggesting that hiNRPs have a migratory capacity. Most of the cells expressed the proliferation marker Ki-67, even after 25 passages (Fig. 3C), indicating that hiNRPs could maintain self-renewal ability for a long time. All hiNRPs showed a normal diploid and male chromosomal karyotype at the 25th passage (Fig. 3E). These hiNRPs were hypodermically injected into four nude mice, and no teratoma was formed in any of them, even after 10 months (data not shown).
The conserved region of the second nestin intron (located at ϩ3314 to ϩ3918), which is part of the nestin enhancer, was largely methylated in hiNRPs, consistent with hNPs, but in a highly methylated status in human fibroblasts (Fig. 3F). The demethylation of the nestin enhancer reactivated the nestin promoter, resulting in nestin expression in hiNRPs. Genome-wide Transcriptional Profiling of 3F-hiNRPs-We compared the global gene expression pattern of 3F-hiNRPs with the original fibroblasts and human ES-derived hNPs to characterize the entire transcriptome. Both hierarchical clustering and pairwise scatter plots revealed that the global gene expression profiles of MSB2-hiNRPs and MSB4-hiNRPs were almost the same but different from those of parental fibroblasts, Diff-hiNRPs, and hNPs. The global gene expression profiles from both types of hiNRPs were closer to hNPs than to the parental fibroblasts. The profiles of Diff-hiNRPs were in the intermediate state between hiNRPs and the parental fibroblasts (Fig. 4A). The hiNRPs and control NPs showed a high expression of neuronal-specific genes such as nestin, DCX, and Sox11, which were less expressed in fibroblasts. Fibroblast-specific genes such as Col1a1, Dkk3, and Sphk1 were significantly down-regulated in hiNRPs.
Differentiation Potential of 3F-hiNRPs in Vitro-To test the differentiating potential of 3F-hiNRPs, we seeded hiNRPs at passage 15 onto polyornithine/laminin-or Matrigel-coated plates in neuronal medium without growth factors and in the presence of cAMP, NT-3, GDNF, or retinoic acid (RA). Cells with the morphology of typical mature neurons formed after 2-4 weeks of culture. Neuron-like cells expressed the neuronal markers Tuj1, MAP2, neuron-specific enolase, and neurofilament light subunit. Some of them expressed the synaptic protein synapsin 1 (Fig. 5A). Furthermore, neuron-like-cells positive for glutamine, GABA, tyrosine hydroxylase, 5-hydroxytryptamine, or choline acetyltransferase could be found among the hiNRP-derived cells (Fig. 5B), suggesting that hiNRPs could differentiate into neurons with various neurotransmitter phenotypes in vitro. Cells positive for the astrocyte markers S100 and GFAP were not found, even after being cultured for 2-4 weeks within the astrocyte differentiation medium. Similarly, cells expressing the oligodendrocyte marker O4 were not found after being cultured for the same time in oligodendrocyte differentiation medium. Nevertheless, many differentiated cells expressed the neuronal marker Tuj1 under either serum or T3 medium conditions (Fig. 5C). This indicates that hiNRPs are confined to the neuronal lineage no matter which induction medium was used.
To test whether hiNRP-derived neurons exhibit functional membrane properties, whole cell patch clamp recordings were performed after 4 -6 weeks of hiNRP differentiation (Fig. 5D). The voltage clamp mode records showed rapidly inactivating inward currents (Na ϩ ) and persistent outward currents (K ϩ ) in response to depolarizing voltage steps (Fig. 5E). The neurons also exhibited strong spontaneous postsynaptic currents (Fig.  5F). The neurons generated repetitive traces of action potentials through current clamp records (Fig. 5G). Therefore, the hiNRP-derived neurons exhibited functional membrane properties of mature neurons in vitro. Differentiation Potential after Transplantation of 3F-hiNRPs into Mouse Brain-HiNRP cell lines expressing GFP were established by transfection with the EGFP gene. These GFPexpressing cells were then transplanted into the left lateral ventricle of 2-month-old mice to test their survival ability and differentiation potential in vivo. After 4 weeks of transplantation, the grafts survived and integrated in the host brain around the lateral ventricle (Fig. 6A). Furthermore, the integrated grafts migrated inside into the brain (Fig. 6B, top panel) and formed many long neurites (Fig. 6B, bottom panel). The progenitor marker nestin was expressed in a few grafts, and the neuronal markers MAP2, neurofilament light subunit, and NeuN were observed in most grafts (Fig. 6C), indicating that hiNRPs could sustain self-renewal and differentiate into terminal neu-rons in the brain. Nevertheless, GFP-positive cells showed negative results for GFAP expression (Fig. 6C), indicating that the grafts had been induced to the neuronal lineage but could not form glial cells in vivo.

DISCUSSION
Our study shows that hiNRPs can be generated directly and efficiently from fibroblasts. In previous efforts, NRP-like cells have been induced from human ES cells (28). The small amount of residual pluripotent stem cells among the induced cells can be a safety concern because of the risk of their tumorigenic tendency (29,30). NRPs could be sorted from NSC progeny as well (31), but the resource is limited and the procedures are cumbersome. We provide a simple method to obtain high-pu- rity NRPs that can specifically differentiate into neurons rather than glial cells in vitro and in vivo.
We initially induced hiNRPs using eight transcription factors (Bmi1, TLX, FoxG1, Klf4, Sox2, c-Myc, Brn2, and Brn4), all of which were proven to be key factors for NSC/NP or neuron induction. We successfully reached the goal of acquiring NRPs by using these eight transcription factors in the initial trials. In the next step, we narrowed down the candidate factors to figure out which factors were essential in the generation of hiNRPs. We found that three factors (Sox2, c-Myc, and either Brn2 or Brn4) were sufficient to induce the NRPs from fibroblasts.
The factors we used are similar to a previous report (24) in which five factors (Sox2, c-Myc, Bmi1, TLX, and Brn2) were used to induce NPCs. The different fate of cells might be because we used human fibroblasts instead of mouse fibroblasts. A report from Lujan et al. (11) also uses mouse fibroblasts as starting cells. They found that of nine Foxg1/Sox2 colonies, eight could differentiate into Tuj1 ϩ and MAP2 ϩ neurons but not GFAP ϩ astrocytes, whereas only one could differ-entiate into both neurons and GFAP ϩ astrocytes. They also found that only one of four Foxg1/Brn2 NPC colonies showed a tripotent character. It indicated that their induced neural progenitors mixed with large number of unipotent NRPs. As for human cells, Giorgetti et al. (32) induced blood cord cells into neuronal cells, not NPCs, by using only Sox2 and c-Myc. The cells exhibited a neuron-restricted phenotype, which indirectly supports our findings that human neuronal restricted progenitors can also be generated by the related factors we used.
We first showed that NRPs could be generated from human fibroblasts by Sox2, c-Myc, and either Brn2 or Brn4. On the basis of their functions, reported previously, we speculated that Sox2 and c-Myc induce fibroblasts into progenitors with the capacity of proliferation. Then, Brn2 or Brn4 drives progenitors toward neuronal lineage cells. Sox2 plays an essential role in the maintenance of both ES cells and NSCs and prevents cell differentiation. As one member of Myc family, c-Myc can significantly improve proliferation and promote reprogramming. The Myc family is also a pivotal target of Wnt-␤-catenin signaling, which regulates the neuronal differentiation of NPCs and the expansion of basal progenitors (33). Notably, Brn2 and Brn4 are both expressed in early born neurons and play a key role in the initiation of neuronal differentiation (34,35). They share a high homology in their primary structure with each other and have complementary functions (36,37). Brn-2 expression is restricted to late neural precursor cells and a wide range of postmitotic neurons (38). Brn2 promotes neurogenesis by upregulating other proneural genes (e.g. Tbr2) as well as by diminishing the Notch-directed transcription of Hes5 (39). The level of Brn4 is significantly up-regulated in newborn neurons originating from embryonic striatal NSCs stimulated by IGF-1 and BDNF (35). Overexpression of Brn4 elevates neuronal differentiation and maturation from NSCs, whereas suppression of expression of Brn4 using RNAi markedly decreases the number of newborn neurons (25,40,41). Brn4 coexpresses with MAP2 or ␤-tubulin-III in neurons, whereas almost all GFAP-positive astrocytes have a weak immunoreactive intensity of Brn4 (25). Taken together, Brn2 and Brn4 are required for neuronal differentiation and development.
The specific gene expression pattern (Nestin ϩ /PAX6 Ϫ / DCX ϩ ) in our hiNRPs is similar to that in NRPs. Nestin is a broad-spectrum marker for NSCs/NPs and NRPs, whereas Pax6 is specifically expressed in NSCs/NPs and substantially down-regulated along with the initiation of differentiation into intermediate progenitor cells (42,43). In our nestin-positive hiNRPs, absence of Pax6 indicated that multipotent NSCs/NPs did not exist in the hiNRP colonies. DCX has a binding site of transcription factors Brn2 and Brn3 in its promoter region (44) and is transiently and specifically expressed in proliferating FIGURE 5. In vitro differentiation of 3F-hiNRPs. A, after differentiation of hiNRPs for 2-4 weeks, neuronal-like cells with synapses formed. Immunostaining showed that differentiated hiNRPs were positive for pan-neuronal markers (Tuj1 and MAP2) and mature neuronal markers (neuron-specific enolase, Syn1, and neurofilament light subunit). Scale bars ϭ 50 m. B, differentiated hiNRPs were also positive for subtype-specific neuronal markers (tyrosine hydroxylase (TH), 5-hydroxytryptamine (5-HT), choline acetyltransferase (ChAT), glutamate (Glu), and GABA). Scale bars ϭ 50 m. C, most of the differentiated hiNRPs formed neuronal cells rather than glial cells even under a standard glial differentiation procedure (positive for Tuj1 and negative for S100, GFAP, and O4). Scale bars ϭ 50 m. D, a patched neuron differentiated from hiNRPs in vitro. E, representative recordings of voltage-gated ion channels (K ϩ , left and Na ϩ , right) were recorded. F, representative traces of spontaneous postsynaptic currents (PSCs) were recorded. G, representative traces of action potentials of an hiNRP-derived neuron in response to step current injections (Ϫ20 to 50 pA). The membrane potential was maintained at approximately 80 mV. The bottom panel shows a single current trace at 40 pA of injected current. neuronal precursors during neurogenesis (45). The expression of DCX in our hiNRPs further confirmed their neuronal precursor identity.
The functions of hiNRPs were further validated by a series of in vitro and in vivo experiments. When cultured in vitro in a neuron-differentiation condition, hiNRPs were able to differentiate into a variety of subtypes of mature neurons, including cholinergic, serotonergic, dopaminergic, and GABAergic neurons with functional membrane properties. Similarly, when transplanted to the lateral ventricle, hiNRPs were able to differentiate into NeuN-positive terminal neurons by responding to the induction of the surrounding environment. These differentiated terminal neurons were found in a variety of locations, displaying their migration ability in the brain.
Transdifferentiation of hiNRPs from human fibroblasts with a high efficiency provides a new approach for creation of a large number of neurons, which can be used for neuronal developmental studies as well as therapeutic strategies addressing neurodegenerative diseases.