mRNA-Driven Generation of Transgene-Free Neural Stem Cells from Human Urine-Derived Cells

Abstract

Human neural stem cells (NSCs) hold enormous promise for neurological disorders, typically requiring their expandable and differentiable properties for regeneration of damaged neural tissues. Despite the therapeutic potential of induced NSCs (iNSCs), a major challenge for clinical feasibility is the presence of integrated transgenes in the host genome, contributing to the risk for undesired genotoxicity and tumorigenesis. Here, we describe the advanced transgene-free generation of iNSCs from human urine-derived cells (HUCs) by combining a cocktail of defined small molecules with self-replicable mRNA delivery. The established iNSCs were completely transgene-free in their cytosol and genome and further resembled human embryonic stem cell-derived NSCs in the morphology, biological characteristics, global gene expression, and potential to differentiate into functional neurons, astrocytes, and oligodendrocytes. Moreover, iNSC colonies were observed within eight days under optimized conditions, and no teratomas formed in vivo, implying the absence of pluripotent cells. This study proposes an approach to generate transplantable iNSCs that can be broadly applied for neurological disorders in a safe, efficient, and patient-specific manner.

Keywords: direct conversion; induced neural stem cells (iNSCs); neurological diseases; reprogramming; self-replicative mRNA; small molecules.

Figure 1

Optimized-conditioning by small molecules for generating induced neural stem cells (iNSCs) from human urine-derived cells (HUCs). (A) GFP (Green fluorescence protein) expression in HUCs at 24 and 48 h after electroporation. (B) FACS (Fluorescence-activated cell sorting) analysis showed high transfection efficiency to HUCs using synthetic mRNA encoding GFP. (C) Small molecules for generating iNSCs from HUCs. (D) Treatment with combinations of small molecules (P: purmorphamine, F: forskolin, V: ascorbic acid, N: sodium butyrate) increases efficiency for generating iNSC colonies. (E) Exclusion of a small molecule decreases the efficiency for generating iNSC colonies. (F) Efficiency for generating iNSC with or without B18R protein during induction. (G) qRT-PCR analysis of PLZF and endogenous SOX2 expression during iNSC generation. (H) qRT-PCR analysis of OCT4, REX1, and NANOG expression during iNSC generation. (I) Hypoxic condition increases the efficiency for generating iNSC colonies. Error bars represent standard deviation. * p ≤ 0.05; ** p ≤ 0.01.

Figure 2

Generation of iNSCs from HUCs by self-replicating mRNAs with small molecules. (A) Schematic of the time course of the process used to directly convert HUCs into iNSCs. (B) Morphology of HUCs after transfection of self-replicating mRNAs. (C,D) Morphology of iNSC colonies exhibit neuroepithelial colony formation on day eight and 12. (E) Morphology of iNSCs after passaging using Accutase solution. (F–I) NSC markers (SOX1, NESTIN, PLZF, PAX6, SSEA1, and SOX2) expression in iNSC colonies, as determined by immunofluorescence analysis. (J) NSC markers are expressed in iNSCs similar with H9-NSCs, as demonstrated by RT-PCR. (K) Ki67 staining in iNSCs, as determined by immunofluorescence analysis. (L,M) OCT4 and NANOG expression in iNSCs and H9-ESCs, as determined by immunofluorescence analysis. Scale bars, 200 μm.

Figure 3

Developmental region-specific patterning of iNSCs. (A) RT-PCR analysis of forebrain, midbrain, hindbrain, and spinal cord markers in iNSCs and H9-NSCs. (B) Analysis of central nervous system (CNS) region-specific maker expression in H9-NSCs and iNSCs by RNA sequencing. FB, forebrain; MB, midbrain; HB, hindbrain; and SC, spinal cord. (C) Identification of dorsal and ventral-specific marker expression in H9-NSCs and iNSCs by RNA sequencing. (D–I) Purmorphamine treatment induces ventral markers (NKX2.2, OLIG2, and NKX6.1) and decreases dorsal markers (PAX3 and PAX6) of iNSCs, as determined by immunofluorescence analysis. (J–M) Forebrain marker (OTX2) and midbrain marker (EN1) expression in iNSCs under LSC culture medium or FGF8b treatment, as determined by immunofluorescence analysis. (O) Schematic diagram of region-specific patterning of iNSCs. Scale bars, 200 μm.

Figure 4

RNA sequencing analysis of the global gene expression profile of iNSCs. (A) Hierarchical clustering analysis of global gene expression in H9-NSCs, iNSCs, and HUCs. Two thousand two hundred and thirty-seven genes are selected with significant p-value (p < 0.05) and fold change (FC > 2). (B,C) Scatter plots of gene expression in iNSC vs. H9-NSCs or HUCs. (D) Hierarchical clustering analysis of gene expression related to neurogenesis in H9-NSCs, iNSCs, and HUCs. Two hundred ninety-nine genes are selected with significant p-value (p < 0.05) and fold change (FC > 2). (E) Venn diagrams show up-regulated genes H9-NSCs and iNSCs vs. HUCs. (F) Venn diagrams show down-regulated genes H9-NSCs and iNSCs vs. HUCs. (G) Gene ontology analysis of top 15 up-regulated genes in iNSCs vs. HUCs. (H) Gene ontology analysis of top 15 down-regulated genes in iNSCs vs. HUCs.

Figure 5

Differentiation potential of iNSCs in vitro. (A–E) Neuronal markers (TUJ1, MAP2, GABA, TH, HB9, and SYN) expression in iNSC-derived neurons, as determined by immunofluorescence analysis. (F) Image of patch clamp experiments. (G) Capacitance and resting membrane potential (RMP) were 30.74 pF ± 1.77 and −58.10 mV ± 7.62 in iNSC-derived neurons, respectively. (H) Representative trace of current clamp recordings of iNSC-derived neurons. Voltage responses were recorded from a holding potential of −80 mV using 500 ms steps with 30 pA of current injection. (I) MEA system showed spontaneous electronic spikes induced by iNSC-derived neurons. (J,K) The peak inward-currents due to Na+ channels were inhibited by addition of 5 μM Tetrodotoxin (TTX) (red, n = 3). (L,M) The outward-currents due to K⁺ channels were inhibited by addition of 10 mM TEA (blue, n = 4). For voltage-clamp recordings, sodium and potassium currents were recorded from a holding potential of −80 mV using 500 ms pulse voltage steps from −100 to + 80 in 10 mV increments. Values are mean ± SEM. * p < 0.05 vs. absence of TTX or TEA. (O) Astroglial markers (GFAP and S100β) expression in iNSC-derived astroglia, as determined by immunofluorescence analysis. (P–R) Oligodendroglial markers (PDGFRα, O4, and MBP) expression in iNSC-derived oligodendroglia, as determined by immunofluorescence analysis. Scale bars, 200 μm.

Figure 6

Injection of GFP-tagged iNSC into immunodeficient mouse brains. (A,B) Microscopy analysis of GFP-tagged iNSCs produced via lentiviral infection. (C) Schematic representation of iNSC transplantation. (D–F) At two months after transplantation, injected GFP+ iNSCs were positive for MAP2, GFAP, and MBP, as determined by immunofluorescence analysis. Scale bars, 200 μm.

Figure 7

Biosafety evaluations of iNSCs. (A,B) Karyotyping of iNSCs at early (P5) and late (P20) passages. (C) Doubling time of iNSCs at early (P8) and late (P20) passages. (D) Exogenous Venezuelan equine encephalitis (VEEs) are not detected in total RNA and genomic DNA of iNSCs, as demonstrated by RT-PCR. (E) Teratoma formation efficiency of BG01-ESCs and iNSCs. (F,G) Teratoma formations in nude mice upon injecting 1 × 10⁶ BG01-ESCs in left dorsal flank and 1 × 10⁶ iNSCs in right dorsal flank. Black arrows indicate teratoma formations two months and four months after injection. (H) Morphological analysis of injected iNSCs into brain of nude mice. All mice did not die before sacrifice and no malignant appearances are showed in their brains at one month, four months, and six months after transplantation of iNSCs.

Figure 8

Schematic illustration of generating iNSCs from HUCs.