Differentiation of radial glia-like cells from embryonic stem cells

Radial glial cells play important roles in neural development. They provide support and guidance for neuronal migration and give rise to neurons and glia. In vitro, neurons, astrocytes, and oligodendrocytes can be generated from neural and embryonic stem cells, but the generation of radial glial cells from these stem cells has not yet been reported. Since the differentiation of radial glial cells is indispensable during brain development, we hypothesize that stem cells also generate radial glial cells during in vitro neural differentiation. To test this hypothesis, we utilized five different clones of mouse embryonic (ES) and embryonal carcinoma (EC) stem cell lines to investigate the differentiation of radial glial cells during in vitro neural differentiation. Here, we demonstrate that radial glia-like cells can be generated from ES/EC cell lines. These ES/EC cell-derived radial glia-like cells are similar in morphology to radial glial cells in vivo, i.e., they are bipolar with an unbranched long process and a short process. They also express several cytoskeletal markers, such as nestin, RC2, and/or GFAP, that are characteristics of radial glial cells in vivo. The processes of these in vitro generated radial glia-like cells are organized into parallel arrays that resemble the radial glial scaffolds in neocortical development. Since radial glia-like cells were observed in all five clones of ES/EC cells tested, we suggest that the differentiation of radial glial cells may be a common pathway during in vitro neural differentiation of ES cells. This novel in vitro model system should facilitate the investigation of regulation of radial glial cell differentiation and its biological function.     

Further characterization of embryonic stem cell-derived radial glial cells

Previously, we showed that radial glia-like (RG) cells differentiated from embryonic stem (ES) cells after retinoic acid induction (Liour and Yu, 2003: Glia 42:109-117). In the present study, we demonstrate that the production of RG cells from ES cells is independent of the neural differentiation protocol used. These ES cell-derived RG (ES-RG) cells are similar in morphology to RG cells in vivo and express several characteristic RG cell markers. The processes of these ES-RG cells are organized into radial arrays similar to the RG scaffold in developing CNS. Expression of Pax6, along with other circumstantial data, suggests that at least some of these ES-RG cells are neural progenitors. The progression of neurogenesis into gliogenesis during the in vitro neural differentiation of ES cells recapitulates the in vivo developmental process. The identification of two cell surface markers, SSEA-1 and GM1, on both the native embryonic RG cells and ES-RG cells, may facilitate purification of radial glial cells for future studies and cell therapy. Overall, our study suggests that differentiation of radial glial cells is a common pathway during the neural differentiation of ES cells.     

Double lox targeting for neural cell transgenesis

ES cells differentiated along the neural lineage in vitro are an attractive model system. Here we have developed ES cell lines that are suitable for inserting transgenes at a single chromosomal site. ES cell line CE1 (for Cassette Exchange) contains one "acceptor" module (CE1) that allows for efficient double lox targeting. The site is also permissive for gene expression in neural progenitor cells, as well as differentiated neurons and glia. Line CE2 was derived by swapping a puromycin resistance cassette into CE1. Neural progenitors derived from this line are puromycin-resistant. A beta-actin/GFP expression cassette was inserted into the CE1 site to create CE3. The CE3 cell line was differentiated into neural cells and displayed strong EGFP expression in neural progenitors, differentiated neurons and glia. Differentiated CE3 ES cells (4-/4+ RA) were transplanted into the injured rat somatosensory cortex where many of the transplanted cells survived and differentiated into neurons expressing GFP. This strategy for creating sets of transgenic lines with multiple cassettes inserted into a single chromosomal site provides a powerful tool for studying development and function of ES cell-derived neural cells. Many of these will be useful in transplantation research.     

Tau EGFP embryonic stem cells: an efficient tool for neuronal lineage selection and transplantation

Pluripotency and the capacity for continuous self-renewal make embryonic stem (ES) cells an attractive donor source for cell-replacement strategies. A key prerequisite for a therapeutic application of ES cells is the generation of defined somatic cell populations. Here we demonstrate that a targeted insertion of the EGFP gene into the tau locus permits efficient fluorescence-activated cell sorting (FACS)-based lineage selection of ES cell-derived neurons. After in vitro differentiation of heterozygous tau EGFP ES cells into multipotent neural precursors, EGFP is selectively induced in postmitotic neurons of various neurotransmitter phenotypes. By using FACS, ES cell-derived neurons can be enriched to purities of more than 90%. Because neuron-specific EGFP fluorescence is also observed upon transplantation of ES cell-derived neural precursors, the tau EGFP mutant represents a useful tool for the in vivo analysis of grafted ES cell-derived neurons.     

GFAP-GFP neural progenitors are antigenically homogeneous and anchored in their enclosed mosaic niche

Study of the different stages of postnatal neurogenesis relies on using antigenic markers and transgenic mice. In particular, neural stem cells that express GFAP are studied using mice expressing GFP under the human GFAP promoter (GFAP-GFP). However, it remains unclear whether GFP and the commonly used progenitor markers label different cell populations in the neurogenic subventricular zone (SVZ) and its rostral extension into the olfactory bulb (i.e. rostral migratory stream, RMS). Here, we found that all GFP-fluorescent cells express GFAP, the radial glia marker brain lipid-binding protein (BLBP), Lewis X (LeX), and the astrocytic marker GLAST. Faint GFP fluorescence could be detected in a few cells expressing EGF receptors (EGFRs), Olig2, or S100, suggesting that GFAP-GFP cells generate these diverse cell types. GFP-fluorescent cells were slowly cycling, as shown by their long-term retention of BrdU, and less than 10% expressed the proliferative markers Ki67 and Mcm2. The majority of EGFR-expressing cells and Olig2-expressing cells were cycling. NG2 and EGFR identified distinct progenitor populations while Olig2 labeled a subset of EGFR-expressing cells. The entire neurogenic zone contained a mosaic of different cell types and was ensheathed by processes of GFAP-expressing cells and NG2 cells. Finally, using time-lapse imaging in acute slices, we show that GFP-fluorescent cells are stationary within the SVZ. Our findings collectively highlight the cellular mosaic of the neurogenic niche, show that the slowly-cycling GFAP-expressing cells are stationary and generate distinct intermediate progenitors.     

Cyclin D1 is required for proliferation of olig2‐expressing progenitor cells in the injured cerebral cortex

Little is known about the molecular mechanisms driving proliferation of glial cells after an insult to the central nervous system (CNS). To test the hypothesis that the G1 regulator cyclin D1 is critical for injury‐induced cell division of glial cells, we applied an injury model that causes brain damage within a well‐defined region. For this, we injected the neurotoxin ibotenic acid into the prefrontal cortex of adult mice, which leads to a local nerve cell loss but does not affect the survival of glial cells. Here, we show that cyclin D1 immunoreativity increases drastically after neurotoxin injection. We find that the cyclin D1‐immunopositive (cyclin D1+) cell population within the lesioned area consists to a large extent of Olig2+ oligodendrocyte progenitor cells. Analysis of cyclin D1‐deficient mice demonstrates that the proliferation rate of Olig2+ cells diminishes upon loss of cyclin D1. Further, we show that cyclin‐dependent kinase (cdk) 4, but not cdk6 or cdk2, is essential for driving cell division of Olig2‐expressing cells in our injury model. These data suggest that distinct cell cycle proteins regulate proliferation of Olig2+ progenitor cells following a CNS insult.     

Neurotransmitter phenotype differentiation and synapse formation of neural precursors engrafting in amyloid-β(1-40) injured rat hippocampus

Alzheimer's disease (AD) is characterized by the dysfunction or loss of a vulnerable group of neurons. At present, only a few options exist for treating neurodegenerative diseases effectively. Advances in stem cell research have raised the hope and possibility for therapy in neurodegenerative diseases. In AD transgenic animal models, stem cell transplantation has been demonstrated to reverse behavioral deficits. Our recent study demonstrates that neural precursor cells, derived from embryonic stem (ES) cells, improve memory dysfunction in rats caused by injections of amyloid-β peptide (1-40) (Aβ????) in the dorsal hippocampus. However, the underlying mechanisms remain unknown. The present study tests a murine ES cell-based transplantation approach in rats subjected to Aβ???? injection into the hippocampus dentate gyrus. Efficacy of cell therapy with regard to graft survival, neuronal yield and diversity, synapse formation of the grafted cells, and the behavioral improvements was determined after transplanting ES cell-derived neural precursors into the hippocampus of adult rats. Here, we show that grafted cells can survive, and differentiate with high yield into immunohistochemically mature glial cells and neurons of diverse neurotransmitter-subtypes. More importantly, transplanted cells demonstrate characteristics of proper synapse formation between host and grafted neural cells. Thus, our observations show that an ES cell-based transplantation approach may be promising in the treatment of AD.     

Peering into early neurogenesis with embryonic stem cells

Embryonic stem (ES) cells transplanted into the early mouse embryo have the capacity to differentiate into all cell types of the nervous system. A simplified culture system has been developed in which single ES cells transform into neural progenitor cells that go on to form neurospheres. This system is ideally suited for mechanistic studies of the earliest stages of neurogenesis. In this model, signaling via fibroblast growth factor and bone morphogenetic protein family members is important for the first steps of neural progenitor differentiation.     

Olig genes are expressed in a heterogeneous population of precursor cells in the developing spinal cord

Recent results from multiple laboratories have identified Olig genes as important in regulating glial differentiation. Here we show that Olig2 expression at early stages of development (prior to E16.5) identifies a domain in the developing spinal cord, which contains a heterogeneous population of progenitors that includes stem cells and glial progenitors. We show that Nkx2.2 and Olig2, which are present initially in nonoverlapping domains, are coexpressed at later stages, likely due to a second wave of Olig expression. We find that Olig1, like Olig2, is present in cells that coexpress astrocytic and radial glial markers and that Olig1/2 double knockouts lead to a loss of oligodendrocytes with preservation of NG2 expression. These results coupled with previously published data indicate that Olig1/2 and Nkx2.2, while clearly important in regulating early progenitor cell differentiation, do not unambiguously demonstrate the existence of an oligodendrocyte-neuron precursor or negate the existing retroviral lineage and clonal analysis data that suggest the existence of other types of precursors such as oligodendrocyte-astrocyte precursors or neuronal precursors.     

Transplantation of Human-Induced Pluripotent Stem Cell-Derived Neural Precursors into Early-Stage Zebrafish Embryos

Induced pluripotent stem cells (iPS cells) generated from somatic cells through reprogramming hold great promises for regenerative medicine. However, how reprogrammed cells survive, behave in vivo, and interact with host cells after transplantation still remains to be addressed. There is a significant need for animal models that allow in vivo tracking of transplanted cells in real time. In this regard, the zebrafish, a tropical freshwater fish, provides significant advantage as it is optically transparent and can be imaged in high resolution using confocal microscopy. The principal goal of this study was to optimize the protocol for successful short-term and immunosuppression-free transplantation of human iPS cell-derived neural progenitor cells into zebrafish and to test their ability to differentiate in this animal model. To address this aim, we isolated human iPS cell-derived neural progenitor cells from human fibroblasts and grafted them into (a) early (blastocyst)-stage wild-type AB zebrafish embryos or (b) 3-day-old Tg(gfap:GFP) zebrafish embryos (intracranial injection). We found that transplanted human neuronal progenitor cells can be effectively grafted and that they differentiate and survive in zebrafish for more than 2 weeks, validating the model as an ideal platform for in vivo screening experiments. We conclude that zebrafish provides an excellent model for studying iPS cell-derived cells in vivo.     

Loss of gene expression in lentivirus- and retrovirus-transduced neural progenitor cells is correlated to migration and differentiation in the adult spinal cord

Gene transfer into multipotent neural progenitor cells (NPC) and stem cells may provide for a cell replacement therapy and allow the delivery of therapeutic proteins into the degenerating or injured nervous system. Previously, murine leukemia virus-based retroviral vectors expressing GFP from an internal EF-1alpha promoter and lentiviral vectors expressing GFP from a hybrid CMV/beta-actin promoter have been described to be resistant to stem cell specific gene silencing. Therefore, we investigated whether these viral vectors allow stable in vivo gene expression in genetically modified NPC isolated from the adult rat spinal cord. In vitro, NPC genetically modified to express GFP using the described retroviral vector showed strong GFP expression in undifferentiated NPC. However, in vitro differentiation resulted in the loss of GFP expression in 50% of cells. Grafting of BrdU-prelabeled NPC to the spinal cord resulted in a loss of GFP expression in 70% and 95% of surviving NPC at 7 and 28 days post-grafting, respectively. The loss in gene expression was paralleled by the differentiation of NPC into a glial phenotype. Transgene downregulation although less profound was also observed in cells modified with lentiviral vectors, whereas in vivo lentiviral gene transfer resulted in stable transgene expression for up to 16 months. Thus, in vivo gene expression in genetically engineered neural progenitor cells is temporally limited and mostly restricted to undifferentiated NPC using the viral vectors tested.     

Adult neural progenitor cell grafts survive after acute spinal cord injury and integrate along axonal pathways

The main rationale for cell-based therapies following spinal cord injury are: (i) replacement of degenerated spinal cord parenchyma by an axon growth supporting scaffold; (ii) remyelination of regenerating axons; and (iii), local delivery of growth promoting molecules. A potential source to meet these requirements is adult neural progenitor cells, which were examined in the present study. Fibroblast growth factor 2-responsive adult spinal cord-derived syngenic neural progenitor cells were either genetically modified in vitro to express green fluorescent protein (GFP) using retroviral vectors or prelabelled with bromodeoxyuridine (BrdU). Neural progenitor cells revealed antigenic properties of neurons and glial cells in vitro confirming their multipotency. This differentiation pattern was unaffected by retroviral transduction. GFP-expressing or BrdU-prelabelled neural progenitor cells were grafted as neurospheres directly into the acutely injured rat cervical spinal cord. Animals with lesions only served as controls. Three weeks postoperatively, grafted neural progenitor cells integrated along axonal profiles surrounding the lesion site. In contrast to observations in culture, grafted neural progenitor cells differentiated only into astro- and oligodendroglial lineages, supporting the notion that the adult spinal cord provides molecular cues for glial, but not for neuronal, differentiation. This study demonstrates that adult neural progenitor cells will survive after transplantation into the acutely injured spinal cord. The observed oligodendroglial and astroglial differentiation and integration along axonal pathways represent important prerequisites for potential remyelination and support of axonal regrowth.     

Fluoxetine promotes gliogenesis during neural differentiation in mouse embryonic stem cells

Selective serotonin reuptake inhibitors (SSRIs) are commonly prescribed for treatment of mood disorders and depression, even during pregnancy and lactation. SSRIs are thought to be much safer than tricyclic antidepressants, with a low risk of embryonic toxicity. Several recent studies, however, have reported that fetal exposure to SSRIs increases the risk of adverse effects during fetal and neonatal development. This is consistent with our previous finding that fluoxetine, a prototypical SSRI, profoundly affected the viability of cultured embryonic stem (ES) cells as well as their ability to differentiate into cardiomyocytes. Furthermore, we found that fluoxetine induced fluctuations in ectodermal marker gene expression during ES cell differentiation, which suggests that fluoxetine may affect neural development. In the present study, we investigated the effects of fluoxetine on the process of differentiation from ES cells into neural cells using the stromal cell‐derived inducing activity (SDIA) method. Fluoxetine treatment was found to enhance the expression of glial marker genes following neural differentiation, as observed by immunocytochemical analysis or quantitative RT‐PCR. The promoter activity of glial marker genes was also significantly enhanced when cells were treated with fluoxetine, as observed by luciferase reporter assay. The expression of neuronal markers during ES cell differentiation into neural cells, on the other hand, was inhibited by fluoxetine treatment. In addition, FACS analysis revealed an increased population of glial cells in the differentiating ES cells treated with fluoxetine. These results suggest that fluoxetine could facilitate the differentiation of mouse ES cells into glial cell lineage, which may affect fetal neural development. © 2010 Wiley‐Liss, Inc.