Abstract
Cellular reprogramming using small molecules (SMs) without genetic modification provides a promising strategy for generating target cells for cell-based therapy. Human adipose-derived stem cells (hADSCs) are a desirable cell source for clinical application due to their self-renewal capacity, easy obtainability and the lack of safety concerns, such as tumor formation. However, methods to convert hADSCs into neural cells, such as neural stem cells (NSCs), are inefficient, and few if any studies have achieved efficient reprogramming of hADSCs into functional neurons. Here, we developed highly efficient induction protocols to generate NSC-like cells (iNSCs), neuron-like cells (iNs) and GABAergic neuron-like cells (iGNs) from hADSCs via SM-mediated inhibition of SMAD signaling without genetic manipulation. All induced cells adopted morphological, molecular and functional features of their bona fide counterparts. Electrophysiological data demonstrated that iNs and iGNs exhibited electrophysiological properties of neurons and formed neural networks in vitro. Microarray analysis further confirmed that iNSCs and iGNs underwent lineage switch toward a neural fate. Together, these studies provide rapid, reproducible and robust protocols for efficient generation of functional iNSCs, iNs and iGNs from hADSCs, which have utility for modeling disease pathophysiology and providing cell-therapy sources of neurological disorders.
Similar content being viewed by others
Introduction
Expandable human neural stem cells (NSCs) and functional γ-aminobutyric acid-secreting neurons (GABAergic neurons or interneurons) derived from stem cells are invaluable cell sources for treating nervous system injuries and neurodegenerative diseases. Human embryonic stem cells (ESCs) and induced pluripotent stem cells (iPSCs) can differentiate into all types of cells, making them useful in biomedical research. However, these two stem cells have limitations on clinical application. Besides the ethical concerns in ESCs derivation, iPSCs have safety issues related to exogenous reprogramming factors which may cause the activation of oncogenic and/or unexpected pathways and technical issues about low efficiency with slow reprogramming process. To overcome these limitation, alternative approaches, such as chemicals, microRNAs and direct conversion, have been employed1,2,3. Despite of considerable improvements, there are continued hurdles to get iPSCs safely to the clinic.
Compared to ESCs and/or iPSCs, human adipose-derived mesenchymal stem cells (hADSCs), which represent a type of adult mesenchymal stem cells, are a suitable source for clinical application, because they can be transplanted across allogeneic barriers, reduced ethical concern with obtaining adipose tissue and strong self-renewal capacity4. They also can be isolated repeatedly and obtained from patients with an easy and noninvasive procedure4. In addition, it has been reported that hADSCs have the neuronal differentiation potential5. Moreover, a recent study has reported that hADSCs can be transdifferentiated into NSC- and neuron-like cells, which can generate tetrodotoxin-sensitive sodium current and/or outward potassium current as well as glia-like cells in vitro 6. Although these results imply the neuronal transdifferentiation capacity of hADSCs, further functional properties, such as action potential or spontaneous postsynaptic current, should be evaluated. However, despite effort to induce hADSCs into neurons or GABAergic neurons, as far as we are aware, there is not an efficient protocol to generate functional neurons, including GABAergic neurons from hADSCs.
Synergistic inhibition of SMAD signaling enhanced neuronal induction efficiency of ESCs and iPSCs by applying small molecules (SMs), the Lefty/Actin/TGFβ pathway inhibitor SB431542 and bone morphogenic protein (BMP) inhibitor noggin together with LDN1931897. Several recent studies also demonstrated that fibroblasts could be directly converted into functional human cardiomyocytes and mouse NSCs by SMs8, 9. The SMs also promote neural conversion of hADSCs although no electrophysiological properties of induced cells are presented10. However, as far as we are aware, there is no published report of generating human NSCs, functional neurons and/or GABAergic neurons from hADSCs by SMs without genetic modification.
Here, we developed an optimal protocol for inducing hADSCs into NSC-like cells (iNSCs) with >85% efficiency using step-wise procedure by SMs without genetic manipulation, which expressed the key NSC markers, such as Pax6, Sox1, Nestin and Vimentin 11. We also confirmed that the iNSCs were expandable and could maintain their gene expression signatures during serially passaged culture. In addition, we established a highly efficient method for inducing hADSCs into functional neuron-like cells (iNs) within 6 weeks using brain-derived neurotrophic factor (BDNF) and sonic hedgehog (SHH) Smoothened (Smo) activator purmorphamine. Next, we further modified a protocol to generate functional GABAergic neuron-like cells (iGNs) within 7 weeks using BDNF and purmorphamine or cyclic AMP (cAMP) analog dibutyryl-cAMP (dbcAMP). Both the iNs and the iGNs exhibited electrophysiological properties unique to bona fide neurons and GABAergic neurons, respectively, and had gene expression profiles similar to neuronally differentiated human primary NSCs. Taken together, our rapid, reproducible and robust protocols enable the efficient generation of high-quality expandable iNSCs, functional iNs and iGNs from hADSCs, which can be greatly useful for safe and effective cell-based therapy and in vitro disease modeling of various neurological and psychiatric disorders.
Results
Induction of hADSCs into NSC-like cells using small molecules
hADSCs isolated from three male donors were characterized by flow cytometry to confirm their identity (Supplementary Table S1) and induced into iNSCs using a 3-step NSC induction protocol (Fig. 1a). We determined the lowest concentration of knock out serum replacement (KOSR) required to produce approximately 60% of cells stained positive for neural cell adhesion molecule (NCAM) using flow cytometry analysis (Supplementary Fig. S1). Thus, we used 3% of KOSR for NSC induction of hADSCs for subsequent studies. To further augment neural induction, we supplemented the cultures at STEP 1 with SM inhibitors of TFG-β and BMP signaling pathways; SB431542 (SB), noggin (N) and LDN193189 (L), which are active in early neural development. Real-time quantitative PCR analysis (qPCR) for expressed levels of NSC markers, such as Sox1, Musashi-1 and Ascl1, revealed that NSC induction was most efficient when all three compounds were present in the induction media (Supplementary Fig. S2).
Figure 1a illustrates the experimental scheme, which consists of three steps (STEP 1–3), used to confirm the effects of SMs on NSC induction of hADSCs. hADSCs cultured in STEP 1 media with (+) or without (−) SMs exhibited different morphologies (Fig. 1b). Furthermore, after STEP 3, cells in STEP 1 (−) SM media showed significantly higher cell density (Fig. 1b). Real-time qPCR further showed that expression levels of NSC markers associated with early neural development and neural differentiation, such as Nestin, Sox1, Sox2, Pax6, Musashi-1, Vimentin, Emx1, Gli3, FoxG1, Ascl1, Tuj1 and Olig2, were significantly increased in hADSCs cultured in STEP 1 (+) SM media as compared to hADSCs and those in STEP 1 (−) SM media (Fig. 1c). The analysis also revealed that markers of medial ganglionic eminence (MGE)-like ventral forebrain progenitors, such as Nkx2.1 and Dlx2, were significantly increased in hADSCs cultured in STEP 1 (+) SM media as compared to hADSCs and those in STEP 1 (−) SM media (Fig. 1c). Flow cytometry analyses showed that the proportion of NCAM-positive cells in STEP 1 (+) SM media was 85.4% compared to 58.6% in STEP 1 (−) SM media. Furthermore, the proportion of NCAM−, Nestin- and Ki67-positive cells in STEP 1 (+) SM media was 85.4%, 92.9% and 76.8%, respectively, by flow cytometry analysis (Fig. 1d). Taken together, these results suggest that our induction protocol can generate iNSCs from hADSCs with >85% efficiency without the need for genetic manipulation.
Characterization of iNSCs derived from hADSCs
As shown in Fig. 2a, iNSCs derived from hADSCs grew as neurospheres in suspension culture, which is a characteristic property of NSCs12. It is well-established that bona fide NSCs highly express the intermediate filament Nestin and the transcription factor Sox2 13. Immunocytochemistry confirmed expression of both Nestin and Sox2 proteins in hADSC-derived iNSCs, whereas expression of these proteins were barely detectable in native hADSCs (Fig. 2b). We further determined the mRNA expression of early neural markers at each phase (STEP 1–3) of the iNSC induction process by qPCR. Expression levels of Nestin, Sox1 and Olig2 gradually increased throughout the process of iNSC induction (between STEP 1 and 3), whereas Sox2, Pax6, Musashi-1, Vimentin, Emx1, Gli3, FoxG1, Nkx2.1, Dlx2, Ascl1 and Tuj1 levels were markedly upregulated at STEP 2 and then subsequently were downregulated during STEP3, such that Vimentin, Emx1, Gli3, FoxG1, Dlx2 and Ascl1 expressions at STEP 2 were statistically higher than those at STEP 3 (Fig. 2c). To determine whether iNSCs survive and maintain their marker expressions under ex vivo culture condition, iNSCs suspended in neurobasal media were transplanted onto the ventral horn of organotypic rat spinal cord slice cultures. One week after transplantation, iNSCs still maintained expressions of SOX2 and TuJ1 under these conditions (Fig. 2d). Taken together, these results suggest that iNSCs acquired morphological and molecular properties of bona fide human NSCs.
Characterization of neuron-like cells induced from hADSCs
We next sought to induce hADSCs into mature neurons by further culturing iNSCs in neuronal induction medium containing purmorphamine and BDNF (Fig. 3a). After 7 days of culture in the latter, the cells exhibited neuron-like structures characterized by distinct bipolar or multipolar neurite outgrowth at the periphery of their cell bodies (Supplementary Fig. S3). Real-time qPCR revealed that iNSCs showed increased expressions of TuJ1, MAP2, Dlx5, Lhx6, GAD67 and the sodium ion channel SCN5A after 7 day-exposure to neuronal induction media. Interestingly, FoxG1 was expressed at significantly lower levels in induced neurons (iNs) as compared to iNSCs, Pax6, Nkx2.1 and Dlx2 mRNA levels were maintained in iNs, demonstrating similar patterns with in vivo neuronal development (Fig. 3b).
Immunocytochemistry analysis showed that iNs mainly expressed immature and mature neuronal markers. For example, more than 75% of cells stained double positive for the immature neuronal markers TuJ1 and Pax6, and the mature neuronal markers MAP2 and NeuN (Fig. 3c,d). In contrast, a relatively small fraction (27–33%) of cells expressed astrocyte markers GFAP and S100, indicating the presence of astrocytes at various developmental stages in the culture. Moreover, approximately 17% of cells expressed the oligodendrocyte marker OLIG2, whereas very few cells were double positive for OLIG2 and MAP2 (Fig. 3c,d). iNSCs could be further induced into cells that expressed markers of all three lineages; neurons, astrocytes and oligodendrocytes. However, our data did not unequivocally demonstrate that these cells were iNSCs, since this was not confirmed at the single cell level. Nevertheless, the data suggest that our induction protocol is capable of producing from hADSCs all three cell lineages; mixed neuronal populations and glial cells.
To assess whether iNs exhibit functional properties of neurons, we performed whole-cell patch clamp analysis. In current clamp mode, iNs successfully elicited action potentials (APs) when depolarized by current injection (from −20 to 20 pA). Moreover, some cells showed a depolarizing “sag” in membrane voltage produced in response to injection of hyperpolarizing current. In addition, a spontaneous postsynaptic current (sPSC) was evident in recorded iNs, indicating that these iNs were connected to each other by synapses and formed a neural network (Fig. 3e).
Optimized induction of hADSCs into GABAergic neuron-like cells (iGNs) and functional characterization of iGNs
As described above (Fig. 2), iNSCs at STEP 2 expressed high levels of NSC-specific markers, such as Nestin and Sox2. In addition, the levels of several transcription factors important for GABAergic neuronal differentiation, including Dlx2 and Ascl1, were substantially decreased at STEP 3. Thus, we decided to use iNSCs at STEP 2 for GABAergic neuronal induction and added dbcAMP into the culture to prevent cell death of induced neurons (Fig. 4a). At days in vitro (DIV) 25, the majority of cells expressing neural markers TuJ1 or MAP2 were also positive for NKX2.1 (approx. 78%), DLX2 (approx. 84%) and LHX6 (approx. 69%), specific markers for MGE originating from the ventral telencephalon ventricular zone (Fig. 4b,c).
To further investigate the GABAergic neuronal properties of iGNs, real-time qPCR was performed to evaluate the mRNA levels of various GABAergic neuronal markers at DIV 25 and DIV 32. The level of MAP2, a marker of mature neurons, increased substantially at DIV 25 compared to native hADSCs (P < 0.001) and further increased at DIV 32 (Fig. 4d). Similarly, the expression of the sodium ion channel subtype markers SCN1A and SCN5A significantly increased in both hADSCs and iGNs at DIV 32 as compared to DIV 25 (Fig. 4d). In addition, while iGNs matured from DIV 25 to DIV 32, the early and late MGE transcription factor Nkx2.1, Dlx2, Dlx5 and Lhx6 respectively, displayed an expression pattern consistent with that observed during MGE progenitor differentiation into GABAergic neurons in vivo (Fig. 4d). Interestingly, in addition to vesicular GABA transporter vGat, the expression of the vesicular glutamate transporter vGlut2 and astrocyte intermediate filament GFAP significantly increased in both hADSCs and iGNs at DIV 32 compared to DIV 25, suggesting that some iNSCs were also induced into glutamatergic neuron-like cells and astrocyte-like cells in the culture (Fig. 4d).
For iGNs to exhibit functional properties of GABAergic neurons, it is essential to have GABA release from the presynaptic neurons and expression of GABA receptors on postsynaptic neurons. GABAA receptor, one of two classes of GABA receptors, is an ionotropic receptor consisting of various types of subunits, including GABRA1, GABRA2 and GABRA5. The significant increase of GABRA2 and GABRA5 mRNA levels was continuous until DIV 32 (Fig. 4d). However, the significantly increased GABRA1 mRNA level at DIV 25 decreased up to the level similar to controls. Glutamic acid decarboxylase (GAD) is an enzyme that converts glutamate to GABA, existing in two isoforms GAD65 and GAD67. GAD65 and GAD67 synthesize GABA for functionally different purpose; GAD65 for neurotransmission and GAD67 for neuronal activity. The mRNA level of GAD65 was significantly enhanced at DIV 25 and sustained at DIV 32 compared to native hADSCs (Fig. 4d). However, the mRNA level of GAD67 was not increased until DIV 32, indicating that the neuronal activity, such as synaptogenesis, exceedingly occurred during neuronal maturation (Fig. 4d). GABAergic cortical interneurons can be distinguished into several subgroups based on the molecular markers, including calbindin2 (CALB2); also known as calretinin (CR), somatostatin (SST) and parvalbumin (PV). In addition, MGE progenitors mostly differentiate into SST− and PV−expressing interneurons14. Consistent with this developmental pattern, the mRNA expression of CALB2, SST and PVALB significantly increased in iGNs at DIV 32 compared to both hADSCs and iGNs at DIV 25, indicating that our protocol can generate iGNs with these GABAergic interneuron subtypes. These results suggest that iGNs may adopt the feature of MGE-derived inhibitory GABAergic neurons.
We also observed NF-M-positive elongated neurites radiating from the neurosphere-like cell cluster and extended neurite outgrowth forming neural networks in the culture (Fig. 5a). Consistent with these data, immunostaining revealed that many MAP2-positive iGNs expressed GABAergic neuron-specific markers GABA (approx. 47%) and GAD (approx. 67%) (Fig. 5b,c). In addition, some MAP2-positive iGNs were labeled with the presynaptic inhibitory synapse marker vGAT (Fig. 5d). MAP2-positive iGNs were also immunopositive for the postsynaptic excitatory synapse marker PSD95, which is co-labeled with the synaptic protein synaptophysin (SYP). Both PSD95 and SYP exhibited punctuated staining indicative of possible synaptic connections (Fig. 5e).
In order to verify functional synaptic formation, electrophysiological properties of iGNs were examined at DIV 32. Before measuring inhibitory postsynaptic current (IPSC), intrinsic electrical properties were measured by eliciting APs. The cells exhibited a fast Na2+ inward current during depolarization of membrane potential. Next, IPSC was measured in the presence of D-APV (50 μM) and 6-cyano-7-nitroquinoxaline-2,3-dione (CNQX, 10 μM), NMDAR antagonist and AMPAR antagonist, respectively. The presence of IPSCs was frequently observed and the isolated IPSCs were further blocked by bicuculline (BIC, 10 μM), indicating that iGNs exhibited electrophysiological features of functional GABAergic neurons (Fig. 5f).
Microarray analysis of iNSCs and iGNs induced from hADSCs
To further validate the induction protocol, gene expression profiles of cells at different time points during the induction process were analyzed by microarray analysis (Fig. 6a). Furthermore, to examine the stability of iNSC phenotype, cells at different passage numbers (P0, P1 and P2) were included in microarray analysis. ReN VM and CX cells induced to mature neural phenotypes were used as positive controls of human neurons.
Gene Ontology (GO) analysis revealed that genes involved in biological processes of nervous system development, generation of neurons and neuron differentiation were highly expressed in both iGNs and differentiated ReNcells, indicating that iGNs acquired the gene expression signatures of mature neurons (Fig. 6b). During induction of hADSCs into iNSCs and iGNs, genes associated with regulation of cell differentiation/development were upregulated, whereas genes involved in hADSC functions, such as cell adhesion and extracellular matrix organization, were downregulated (Fig. 6b). Principle component analysis showed that iGNs/differentiated ReNcells, iNSCs P0/P1/P2, and hADSCs were clustered and each cluster was well-distinct with one another (Fig. 6c). In addition, the heatmap analysis of 8 neural cell- and 4 hADSC-enriched gene expressions demonstrated that neural genes were upregulated, but hADSC genes were significantly downregulated in iNSCs, iGNs and ReNcells (Fig. 6d). Furthermore, we performed REACTOME pathway-enrichment analysis and identified REACT_13685:Synaptic Transmission as a statistically significant pathway related to the differentially expressed genes in iGNs. Taken together, these results indicate the efficient cellular reprogramming of hADSCs into neural cells.
Discussion
Our study demonstrates for the first time that hADSCs without genetic manipulation can be induced into expandable iNSCs with high efficiency, 85.4% of NCAM+ cells. This occurred with SM inhibitors of Lefty/Actin/TGFβ and BMP pathways. Additionally, hADSCs can be induced into functional iNs and/or iGNs with a neuronal purity greater than 75% and 46%, respectively.
iNSCs could self-renew and form neurospheres in suspension culture as bona fide NSCs. Moreover, iNSCs exhibited features of MGE-like ventral forebrain progenitors, e.g. they exhibited significantly increased expression of Nkx2.1 and Dlx2 compared to native hADSCs. These MGE-like progenitors were further induced into mature neuron-like cells. We also observed cells expressing glial markers, such as GFAP, S100 and OLIG2, indicating that iNSCs could be further induced into neurons, astrocytes and oligodendrocytes (Fig. 7). Due to their self-renewal capacity, iNSCs can be cryopreserved and may be used as a stable intermediate for the generation of neurons and glia. In our study, iNSCs can maintain their gene expression signatures during serially passaged culture and after cryopreservation. In addition, iNSCs can be transplanted ex vivo with good viability making them suitable for clinical applications. For example, we showed that iNSCs transplanted into spinal cord slice cultures maintained expressions of SOX2 and TuJ1 up to one week post-transplantation (Fig. 2d). Therefore, iNSCs may represent an alternative cellular therapeutic for neurodegenerative diseases, although further evaluation is needed using appropriate translational models to assess their potential efficacy.
The potency of iNSCs was evaluated by differentiating into mature neurons with purmorphamine targeting Smo to activate the SHH pathway and BDNF15, 16. Although it has been reported that sole treatment with BDNF can induce hMSCs into neuron-like cells5, 17, as far as we are aware, there is no report of the use of both purmorphamine and BDNF to induce hADSCs into neurons. Here, we developed an efficient protocol for neuronal induction of hADSCs by activating both SHH and BDNF pathways. In addition to morphological and molecular features of neurons, iNs exhibit excitable properties of neurons and sPSC, indicating possible synaptic networks in the culture (Fig. 3e).
Importantly, our study demonstrates for the first time that functional iGNs can be induced from hADSCs without genetic modification, such as transfection of proneural transcription factors. Although iN is a useful source for the research of molecular and/or developmental mechanisms of neurodegenerative disorders, it is hardly applicable to neurological pathologies caused by impaired excitation and inhibition (E/I) balance in neuronal networks. Dysfunction of GABAergic neurons is involved in many neurological and psychiatric disorders due to their crucial role in E/I balance in the brain18, 19. Quite a few studies have reported that GABAergic interneuron transplantation exhibited therapeutic effects in animal models of neurological disorders, such as epilepsy and Parkinson’s disease, by modifying neural circuits, suggesting a possible application of cell-based therapy for human neurological and neurodegenerative disorders19. Also, iGNs with specific pathological traits can facilitate studies of pathogenesis and in vitro disease modeling.
A recent study demonstrates that human fibroblasts can be converted into functional GABAergic neurons by introducing five factors (FoxG1, Sox2, Ascl1, Dlx5 and Lhx6)20. In our study, the mRNA levels of FoxG1, Sox2, Dlx2 and Ascl1 at STEP 2 of iNSC induction were significantly upregulated compared to those in hADSCs and even higher than those at STEP 3. In addition, iNs expressed significantly higher mRNA levels of cortical interneuron transcription factors Dlx5 and Lhx6 and early GABA transcription factor GAD67 compared to native hADSCs and iNSCs, indicating that iNs can be possibly further differentiated into GABAergic neurons. Thus, we selected STEP 2 cells to optimize the protocol for GABAergic neuronal induction by step-wise procedures: neuronal induction medium consisting of purmorphamine and BDNF, and further differentiation with dbcAMP and BDNF. Many MAP2-positive iGNs expressed GABA (approx. 47%) and GAD (approx. 67%), and some MAP2-positive iGNs were immunopositive for vGAT, PSD95 and SYP, indicating possible inhibitory and excitatory synaptic network formation in vitro. Microarray analysis also demonstrated that iGNs acquire gene expression signatures of mature neurons and highly express genes involved in synaptic transmission. Consistently, whole-cell patch clamp recordings show that iGNs generate evoked sodium currents together with action potentials and spontaneous IPSCs, confirming the existence of functional GABAergic neurons. Our simple, rapid and robust protocol enables us to generate functional iGNs from hADSCs and requires even less time than the iGN-induction protocol using hiPSCs in a recent study3. Although our results illustrate the development of an optimal method for generating functional iGNs from hADSCs without genetic manipulation, further studies are needed to fully characterize the functional attributes of iGNs as well as their therapeutic potential, if any.
A recent study has demonstrated that a cocktail of nine SMs can convert mouse fibroblasts into mouse NSCs, suggesting the feasibility of generating human NSCs and neurons using SMs only9. Moreover, the SMs also promote differentiation of hADSCs into neuron-like cells expressing Sox1, Pax6 and NF-H 10. SMs are ideal for generating target cells similar to human bona fide in vivo counterparts because they are safe, stable and cost-effective. Also, SMs are easy to manipulate and allow precise temporal control of conversion. In addition, SMs provide insights on the underlying mechanisms of cell conversion by targeting specific pathways. However, especially for neural induction, relatively low efficiency of chemical induction has limited broad application compared to induction with genetic manipulation, such as forced expression of proneural genes. However, introducing exogenous genes may cause variability of target cells depending on transfection efficiency and safety concerns, such as activation of unexpected pathways. Thus, it is safer and more stable to generate neural cells by activating endogenous neural genes using SMs.
However, as far as we are aware, there is no published report of the generation of human NSCs and/or functional GABAergic neurons from hADSCs using SMs without genetic modification. Here, for the first time, we developed an efficient protocol for inducing hADSCs into iNSCs, iNs and iGNs through synergistic inhibition of the SMAD signaling pathway by applying three SMs, SB431542, noggin and LDN193189. The SMs promoted mRNA expression of genes related to neural development, such as Sox1, Pax6 and Musashi-1 (Fig. 1c). The neural tube, the primordium of the central nervous system, consists of neuroepithelial cells expressing Pax6 and Sox1, which are crucial factors in neural cell fate determination and/or differentiation21. In addition, the Musashi-1 is selectively expressed in mammalian NSCs, and plays an important role in maintenance of NSC state and differentiation22. Dual inhibition of SMAD signaling during iNSC induction also upregulates genes associated with NSCs and/or neural progenitors, such as Vimentin, FoxG1, Ascl1 and Olig2, which leads to enhanced Tuj1, a neuron-specific microtubule expressed in immature neurons23. Taken together, our results demonstrate efficient and reproducible generation of functional iNSCs, iNs and iGNs using only SMs without genetic manipulation of hADSCs, which are known to be refractory to neural conversion.
Methods
Isolation and culture of hADSCs
Adipose tissue was obtained through liposuction24 from the healthy volunteers (23–26 years old) provided with informed consent. All procedures were performed in accordance with the Declaration of Helsinki and were approved by the Institutional Review Board of Seoul National University Hospital. Isolated hADSCs were maintained in culture medium consisting of alpha-minimum essential medium (α-MEM; Hyclone) supplemented with 10% fetal bovine serum (Hyclone) and 100 μg/ml penicillin and streptomycin (Gibco)25. Cells were subcultured every 3–4 days at a density of 1,400 cells/cm2 and maintained at 37 °C with 5% CO2. hADSCs at passage 4 were used for all experiments.
Induction of hADSCs into iNSCs
hADSCs were seeded on gelatin-coated dishes (10,000 cells/cm2) in NSC induction medium consisting of Dulbecco’s modified Eagle’s medium F12 (DMEM/F12; Gibco) supplement with 3% KOSR (Gibco), 1% Glutamax (Gibco), 1% non-essential amino acid (Gibco) and 4 ng/mL basic fibroblast growth factor (bFGF; Peprotech) with or without three small molecules; 10 μM SB431542 (Sigma), 100 ng/ml noggin (R&D) and 0.5 μM LDN193289 (Stemgent), for 8 days (STEP1). Next, cells were cultured in neurobasal medium (Gibco):DMEM/F12(1:1) supplemented with 2% B27 (Gibco), 1% N2 (Gibco), 1% Glutamax (Gibco) and 200 μM ascorbic acid (Sigma) for 5 days (STEP2). The medium was replaced with STEP2 medium supplemented with 20 ng/mL epidermal growth factor (EGF; Peprotech) and 20 ng/mL bFGF for another 7 days (STEP3).
Induction of iNSCs into iNs and iGNs
iNSCs were dissociated using TryPLE select (Gibco) and replated (50,000 cells/cm2) on plastic cover slips (Thermanox; Thermo) coated with 5 μg/ml poly-L-ornithine (PLO; Sigma) and 50 μg/ml fibronectin (FN; Sigma). On the next day, the medium was replaced with neural induction medium; STEP2 medium containing 1 μM purmorphamine (Tocris) and 10 ng/ml brain-derived neurotrophic factor (BDNF; R&D) for 12–14 days.
For GABAergic neuronal induction, the cells at STEP2 were incubated with the neural induction medium for 12–14 days, followed by STEP2 medium consisting of 50 μM dbcAMP (Sigma) and 20 ng/ml BDNF for another 12–14 days.
Culture and neuronal differentiation of ReNcell VM and CX
Two immortalized stem cell lines, ReNcell-VM from midbrain (Millipore) and -CX from cortex (Millipore), were cultured on 1 μg/ml PLO and 10 μg/ml FN-coated dishes in growth medium of DMEM/F12 supplemented with 1% N2, 1% Glutamax, 20 ng/mL EGF and 20 ng/mL bFGF. Subculture was performed every 3–4 days using Accutase (Gibco) and reattached (50,000 cells/cm2).
For neuronal differentiation, cells (passage 20–30) were seeded in dishes (50,000 cells/cm2) for spontaneous aggregation and expanded in growth medium for 5–7 days. Aggregates were replated on PLO/FN-coated plates and then differentiated for 10 days using growth medium containing 1 mM dbcAMP and 2 ng/ml GDNF (Peprotech)26.
Quantitative real-time PCR
Total RNA was extracted using the TRIzol® reagent (Invitrogen). After DNase treatment, cDNA was synthesized by M-MLV reverse transcriptase (Promega) at 42 °C for 1 h. Quantification of genes was performed using SYBR FAST qPCR Kits (KAPA biosystems). PCR amplification was generated using gene-specific primers (Supplementary Table S2). The target gene expression was normalized to endogenous GAPDH, and the level of genes was determined by the comparative Ct method. The Ct value is the cycle number at which the fluorescence level reaches threshold. The ΔCt value is determined by subtracting the Ct value of the GAPDH control from the Ct value of the target gene [ΔCt = Ct(Target) − Ct(GAPDH)]. This relative value of target genes to endogenous reference is described as the fold-change of GAPDH = 2−ΔCt.
Immunocytochemistry
Cells were fixed with 4% paraformaldehyde (PFA) in PBS, permeabilized and blocked with 0.1% Triton-X100 plus and 5% normal goat serum in PBS. Then, they were incubated overnight at 4°C with the following primary antibodies: mouse anti-DLX2, 1:500 (Santa Cruz), rabbit anti-GABA, 1:1000 (Sigma), rabbit anti-GAD65/67, 1:250 (Millipore), mouse anti-GFAP, 1:200 (Millipore), mouse anti-MAP2, 1:200 (Chemicon), rabbit anti-MAP2, 1:1000 (Millipore), mouse anti-nestin, 1:200 (Chemicon), mouse anti-NeuN, 1:500 (Millipore), rabbit anti-neurofilament M, 1:200 (Millipore), mouse anti-NKX2.1, 1:1000 (Millipore), rabbit anti-OLIG2, 1:1000 (a gift from Dr. Charles Stiles, Harvard Medical School), mouse anti-PAX6, 1:250 (Millipore), mouse anti-PSD95, 1:500 (Millipore), rabbit anti-S100, 1:250 (Dako), mouse anti-SOX2, 1:500 (Millipore), rabbit anti-synaptophysin, 1:250 (Sigma), rabbit anti-TuJ1, 1:1000 (BioLegend) and mouse anti-vGAT, 1:200 (Synaptic Systems). After washing with PBS, cells were incubated with the secondary antibodies, Alexa Fluor® 555 anti-mouse IgG (Molecular Probes) and Alexa Fluor® 488 anti-mouse IgG (Molecular Probes). Cells were then counter-stained with 4,6-diamidino-2-phenylindole (DAPI) (Santa Cruz). The images were captured using a confocal laser-scanning microscope (LSM700; Zeiss) and digital inverted fluorescence microscope (DM5000B; Leica).
Fluorescence-activated cell sorter (FACS) analysis
hADSCs were incubated with each antibody against human CD14, CD34, CD45, HLA-DR (FITC; BD), CD44, CD73, CD90, CD105, CD166 (PE; BD) and CCR2 (Alexa Fluor; Serotec) for 1 h at 4 °C. Corresponding mouse isotype antibodies were used as controls. The single color-stained cells were washed with PBS and fixed with 1% PFA in PBS. MSC immunotypes were determined by flow cytometry on a FACS Calibur System (BD) and the percentage of expressed cell surface antigens was calculated for 10,000 gated-cell events.
To quantify the percentage of iNSCs induced from hADSCs, iNSCs were incubated with NCAM-conjugated FITC, Nestin-conjugated PE (BD) and Ki67-conjugated FITC (eBioscience) for 1 h at 4 °C. Cells without antibody binding were used as controls. FACS experiment and data analysis were carried out according to manufacturer's instructions (FACS Calibur System, BD).
Slice culture and cell transplantation analysis
All procedures were approved by the Institutional Animal Care and Use Committee of Seoul National University. The organotypic slice culture was performed as previously described27,28,29. Cells (6 × 103 cells/μl) prepared in neurobasal medium were transplanted onto the slice using aspirator tube assembly for microcapillary pipette (Sigma). In 7 days, the slices were fixed with 4% PFA overnight at 4°C, permeabilized and blocked with 0.1% Triton X-100 in 3% BSA. They were incubated overnight at 4°C with mouse anti-SOX2, 1:500, rabbit anti-TuJ1, 1:1000 and mouse anti-human nuclei, 1:500 (Millipore). Incubation with the secondary antibodies was performed as described above. The z-stack images were captured using a laser-scanning confocal microscope at 3–4 µm intervals.
Electrophysiology
AP firing and spontaneous post synaptic current were recorded from cultured iN at DIV 22–32 days. The cover slip was placed in recording solution containing (in mM) 150 NaCl, 10 HEPES, 3 KCl, 2 CaCl2, 2 MgCl2, 5.5 glucose and 20 sucrose with pH 7.3. The recorded cell was visualized either directly via the microscope’s optics, or indirectly via a high-resolution CCD camera system (Orca Flash 2.1, Hamamatsu) that received the output of a CCD camera attached to the microscope’s video port. Whole-cell patch clamp recordings were obtained using borosilicate glass pipettes (resistance 4–8 MΩ) prepared using a 2-stage vertical pipette puller (Narishige PC-10). Voltage and current clamp recordings were acquired using Multiclamp 700A (Axon Instruments). Signals were filtered at 1–2 kHz. All recordings under these protocols were digitized at 5 kHz and analyzed using pCLAMP 10 software (Axon Instruments). Experiments with a holding current of more than −100 pA or in which there was a change in input resistance >30% of the control were rejected. Whole-cell recordings were carried out with K-gluconate-based internal solution containing (in mM) 140 K-gluconate, 7 NaCl, 10 HEPES, 4 MgATP and 0.3 Na3-GTP, in which the pH was adjusted to 7.2 with KOH (OSM = 289).
Microarray data analysis
Gene expression profile was analyzed using Illumina Human HT-12 v4.0 Expression BeadChip™ containing 47,306 probes. Raw data preprocessing for the Illumina beadchip microarray, including variance stabilizing transform and quantile normalization (for all 18 samples), was performed by the lumi package (Illumina) of the R Statistical package30, 31. Filtering transcripts with detection p-values < 0.05 in at least 50% of samples returned 20,213 probes for the following analysis. Probes associated with the same gene were summarized at the gene level expression by calculating median expression levels of the probes. Finally, we obtained expression profiles of 12,269 genes for each sample.
Differential gene expression analysis
An empirical Bayes moderated t-statistics using the limma R package was used to determine differentially expressed genes in iNSC, iGN and ReNcell groups compared to control hADSC samples after correcting for multiple hypothesis testing by the Bonferroni method (p < 0.05), for which the absolute cutoff value of log fold change equaled 1.032.
Heatmap representation of the genes that was identified by Kruskal-Wallis test followed by multiple comparison tests according to Conover p-value < 0.005. Heatmap was generated using the function heatplot in the ‘made4’ package of the Bioconductor project33. The distance between genes was calculated by 1–Pearson correlation, and subsequent clustering using the average method for two-dimensional hierarchical clustering.
Functional annotation and pathway enrichment analysis
Differentially expressed genes were annotated with GO and Reactome pathways with the following enrichment analyses34, 35 using RDAVIDWebService (Benjamini-Hochberg FDR < 0.05 after one-tailed Fisher’s exact tests)36.
Statistical analysis
All quantitative data were presented as mean ± standard error of mean (SEM) and evaluated using one-way analysis of variance (ANOVA) with Newman-Keuls post hoc. Statistical analysis was performed using the language R (R Development Core Team 2010). Probability (P)-values < 0.05 were considered statistically significant.
References
Bao, X. et al. MicroRNAs in somatic cell reprogramming. Curr. Opin. Cell Biol. 25, 208–214 (2013).
Liu, Y. et al. Directed differentiation of forebrain GABA interneurons from human pluripotent stem cells. Nat. Protoc. 8, 1670–1679 (2013).
Sun, A. X. et al. Direct Induction and Functional Maturation of Forebrain GABAergic Neurons from Human Pluripotent Stem Cells. Cell Rep. 16, 1942–1953 (2016).
Kern, S., Eichler, H., Stoeve, J., Klüter, H. & Bieback, K. Comparative analysis of mesenchymal stem cells from bone marrow, umbilical cord blood, or adipose tissue. Stem Cells 24, 1294–1301 (2006).
Anghileri, E. et al. Neuronal differentiation potential of human adipose-derived mesenchymal stem cells. Stem Cells Dev. 17, 909–916 (2008).
Feng, N. et al. Generation of highly purified neural stem cells from human adipose-derived mesenchymal stem cells by Sox1 activation. Stem Cells Dev. 23, 515–529 (2014).
Chambers, S. M. et al. Highly efficient neural conversion of human ES and iPS cells by dual inhibition of SMAD signaling. Nat. Biotechnol. 27, 275–280 (2009).
Cao, N. et al. Conversion of human fibroblasts into functional cardiomyocytes by small molecules. Science 352, 1216–1220 (2016).
Zhang, M. et al. Pharmacological Reprogramming of Fibroblasts into Neural Stem Cells by Signaling-Directed Transcriptional Activation. Cell Stem Cell 18, 653–667 (2016).
Madhu, V., Dighe, A. S., Cui, Q. & Deal, D. N. Dual Inhibition of Activin/Nodal/TGF-β and BMP Signaling Pathways by SB431542 and Dorsomorphin Induces Neuronal Differentiation of Human Adipose Derived Stem Cells. Stem Cells Int. 2016 (2015).
Pankratz, M. T. et al. Directed neural differentiation of human embryonic stem cells via an obligated primitive anterior stage. Stem Cells 25, 1511–1520 (2007).
Gage, F. H., Ray, J. & Fisher, L. J. Isolation, characterization, and use of stem cells from the CNS. Annu. Rev. Neurosci. 18, 159–192 (1995).
Ahmed, S. et al. Transcription factors and neural stem cell self-renewal, growth and differentiation. Cell Adh. Migr. 3, 412–424 (2009).
Le Magueresse, C. & Monyer, H. GABAergic interneurons shape the functional maturation of the cortex. Neuron 77, 388–405 (2013).
Sinha, S. & Chen, J. K. Purmorphamine activates the Hedgehog pathway by targeting Smoothened. Nat. Chem. Biol. 2, 29–30 (2006).
Huang, E. J. & Reichardt, L. F. Neurotrophins: roles in neuronal development and function. Ann. Rev. Neurosci. 24, 677 (2001).
Yu, J. H. et al. GABAergic neuronal differentiation induced by brain-derived neurotrophic factor in human mesenchymal stem cells. Ani. Cells Syst. 18, 17–24 (2014).
Markram, H. et al. Interneurons of the neocortical inhibitory system. Nat. Rev. Neurosci. 5, 793–807 (2004).
Southwell, D. G. et al. Interneurons from embryonic development to cell-based therapy. Science 344, 1240622 (2014).
Colasante, G. et al. Rapid conversion of fibroblasts into functional forebrain GABAergic interneurons by direct genetic reprogramming. Cell Stem Cell 17, 719–734 (2015).
Zhang, S. C. Neural subtype specification from embryonic stem cells. Brain Pathol. 16, 132–142 (2006).
Okano, H., Imai, T. & Okabe, M. Musashi: a translational regulator of cell fate. J. Cell Sci. 115, (1355–1359 (2002).
Memberg, S. P. & Hall, A. K. Dividing neuron precursors express neuron‐specific tubulin. J. Neurobiol. 27, 26–43 (1995).
Shiffman, M. A. & Di Giuseppe, A. Liposuction: principles and practice. (Springer Science & Business Media, 2007).
Shim, G. et al. Enhanced survival of transplanted human adipose-derived stem cells by co-delivery with liposomal apoptosome inhibitor in fibrin gel matrix. Eur. J. Pharm. Biopharm. 85, 673–681 (2013).
Donato, R. et al. Differential development of neuronal physiological responsiveness in two human neural stem cell lines. BMC Neurosci. 8, 1 (2007).
Park, H. W. et al. Human mesenchymal stem cell‐derived Schwann cell‐like cells exhibit neurotrophic effects, via distinct growth factor production, in a model of spinal cord injury. Glia 58, 1118–1132 (2010).
Park, H. W. et al. Directed induction of functional motor neuron-like cells from genetically engineered human mesenchymal stem cells. PLoS ONE 7, e35244 (2012).
Cho, J.-S. et al. Transplantation of mesenchymal stem cells enhances axonal outgrowth and cell survival in an organotypic spinal cord slice culture. Neurosci. Lett. 454, 43–48 (2009).
Du, P., Kibbe, W. A. & Lin, S. M. lumi: a pipeline for processing Illumina microarray. Bioinformatics 24, 1547–1548 (2008).
Gentleman, R. C. et al. Bioconductor: open software development for computational biology and bioinformatics. Genome Biol. 5, 1 (2004).
Smyth, G. K. In Bioinformatics and computational biology solutions using R and Bioconductor 397–420 (Springer, 2005).
Culhane, A. C., Thioulouse, J., Perrière, G. & Higgins, D. G. MADE4: an R package for multivariate analysis of gene expression data. Bioinformatics 21, 2789–2790 (2005).
Ashburner, M. et al. Gene Ontology: tool for the unification of biology. Nat. Genet. 25, 25–29 (2000).
Croft, D. et al. The Reactome pathway knowledgebase. Nucleic Acids Res. 42, D472–D477 (2014).
Fresno, C. & Fernández, E. A. RDAVIDWebService: a versatile R interface to DAVID. Bioinformatics 29, 2810–2811 (2013).
Acknowledgements
This research was supported by grants of the Basic Science Research Program (2012R1A1A2007945, 2015R1D1A1A01056950) and the Creative Research Initiative Program (2015R1A3A2066619) from the National Research Foundation of Korea, and by a grant of the Korean Health Technology R&D Project (HI12C0417(A120476)) from the Ministry of Health & Welfare, Republic of Korea.
Author information
Authors and Affiliations
Contributions
J.S.K. and M.-S.C. designed the study; J.P., N.L., J.L., E.K.C. and M.K.K. performed the experiments; J.L. and J.H.K. performed the microarray data analysis; M.S.B., M.-W.C., S.W.K. and C.J.L. analyzed the data; J.P. and M.-S.C. wrote the manuscript.
Corresponding author
Ethics declarations
Competing Interests
The authors declare that they have no competing interests.
Additional information
Publisher's note: Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Electronic supplementary material
Rights and permissions
Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons license, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons license and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this license, visit http://creativecommons.org/licenses/by/4.0/.
About this article
Cite this article
Park, J., Lee, N., Lee, J. et al. Small molecule-based lineage switch of human adipose-derived stem cells into neural stem cells and functional GABAergic neurons. Sci Rep 7, 10166 (2017). https://doi.org/10.1038/s41598-017-10394-y
Received:
Accepted:
Published:
DOI: https://doi.org/10.1038/s41598-017-10394-y
This article is cited by
-
Mouse Neural Stem Cell Differentiation and Human Adipose Mesenchymal Stem Cell Transdifferentiation Into Neuron- and Oligodendrocyte-like Cells With Myelination Potential
Stem Cell Reviews and Reports (2022)
-
Neural stem cells derived from primitive mesenchymal stem cells reversed disease symptoms and promoted neurogenesis in an experimental autoimmune encephalomyelitis mouse model of multiple sclerosis
Stem Cell Research & Therapy (2021)
-
Mesenchymal stem cell therapy for ischemic stroke: A look into treatment mechanism and therapeutic potential
Journal of Neurology (2021)
Comments
By submitting a comment you agree to abide by our Terms and Community Guidelines. If you find something abusive or that does not comply with our terms or guidelines please flag it as inappropriate.