Identification of the gliogenic state of human neural stem cells to optimize in vitro astrocyte differentiation

Highlights

• The potential of glia differentiation increases with successive stem cell divisions.

• NSC, which underwent high cell divisions adopt a pre-glial cell phenotype.

• SUZ12, SMAD4 and STAT3 are potential upstream regulators of the gliogenic program.

• Targeted differentiation leads to pure mature and functional astrocyte cultures.

Abstract

Background

Human preclinical models are crucial for advancing biomedical research. In particular consistent and robust protocols for astrocyte differentiation in the human system are rare.

New method

We performed a transcriptional characterization of human gliogenesis using embryonic H9- derived hNSCs. Based on these findings we established a fast and highly efficient protocol for the differentiation of mature human astrocytes. We could reproduce these results in induced pluripotent stem cell (iPSC)-derived NSCs.

Results

We identified an increasing propensity of NSCs to give rise to astrocytes with repeated cell passaging. The gliogenic phenotype of NSCs was marked by a down-regulation of stem cell factors (e.g. SOX1, SOX2, EGFR) and an increase of glia-associated factors (e.g. NFIX, SOX9, PDGFRa). Using late passage NSCs, rapid and robust astrocyte differentiation can be achieved within 28 days.

Comparison with existing method(s)

In published protocols it usually takes around three months to yield in mature astrocytes. The difficulty, expense and time associated with generating astrocytes in vitro represents a major roadblock for glial cell research. We show that rapid and robust astrocyte differentiation can be achieved within 28 days. We describe here by an extensive sequential transcriptome analysis of hNSCs the characterization of the signature of a novel gliogenic stem cell population. The transcriptomic signature might serve to identify the proper divisional maturity.

Conclusions

This work sheds light on the factors associated with rapid NSC differentiation into glial cells. These findings contribute to understand human gliogenesis and to develop novel preclinical models that will help to study CNS disease such as Multiple Sclerosis.

Introduction

Studying human neural cells in health and disease necessitates the development of diverse, flexible and reproducible methods. The potential to differentiate stem cells, i.e. induced pluripotent stem cells (iPSCs) of individual humans, into diverse cell types has allowed the development of diverse human in vitro models for a whole host of organ systems and diseases (Takahashi et al., 2007). However, in vitro models which focus on human neuroglial cells are uncommon. According to current published standards, the generation of mature astrocytes requires an in vitro differentiation time of over three months (Krencik et al., 2011, Roybon et al., 2013). The difficulty, expense and time associated with generating astrocytes in vitro represent a major roadblock for glial cell research.

NSCs—often referred to in vivo as neural progenitor cells (NPCs)—are multipotent cells with the capacity to self-renew. In vertebrates these cells have been proposed as emerging in the early neural plate (Temple, 2001) and mark the origin of neurogenesis and most gliogenesis (apart from microglia). Mechanisms controlling neuronal and glial differentiation—aside from being clearly tightly temporally regulated—are also sensitive to the extracellular environment and paracrine factors as well as intrinsic regulatory mechanisms (Kessaris et al., 2001, Temple, 2001). In vivo, it has been shown that the differentiation of glial cells occurs later than the development of neurons (Bayer and Altman, 1991). This prompted researchers to model these processes using systems biology and bioinformatic approaches (Qian et al., 2000), wherein models explaining the timed generation of different cell subsets can be generalized as either relying on extrinsic or intrinsic mechanisms. In extrinsic models, the potential of a stem cell is similar in early and late divisions and the development of one population before the other is driven by external cues that drive certain processes, e.g. signaling that promotes neuronal differentiation over glial differentiation in the early phase. In intrinsic models the potential of a stem cell changes over time, drawing support from the observation that most cells from early divisions will develop into neurons, and cells from later phases are more likely to give rise to glial cells.

In vivo, astrogenesis is regulated by both cell intrinsic mechanism such as epigenetic chromatin modification and by extrinsic signals including growth factors and cytokines (Hirabayashi and Gotoh, 2010, Namihira and Nakashima, 2013, Rowitch and Kriegstein, 2010). During astrocyte development in mice, late NPCs as well as early neurons release cytokines such as CNTF (Bonni et al., 1997) and cardiotrophin 1 (CT1) (Barnabé-Heider et al., 2005)—which activate the JAK/STAT signaling cascade—and LIF—which induces the activation of BMP-SMAD pathways (Nakashima et al., 1999a). This leads to the expression of STAT3 and SMAD (Nakashima et al., 1999b) inducing the expression of GFAP, S100B (He et al., 2005) and NFIA, which in turn drives EAAT1 expression (Deneen et al., 2006).

We illustrate here that repeated cell passaging drives antagonistic regulatory programmes for gliogenic vs. neurogenic fate decision in hNSCs. Transcription factors for stemness are downregulated along with those leading into the neuronal lineage, others involved in glial cell differentiation are upregulated in a timed process, dependent on cell division, and in the absence of external stimuli. Thus, we demonstrate that hNSC cultures in vitro recapitulate the “neuron first, glia second” principle of differentiation by sequential activation of pro-neuronal, anti-stemness and pro-glial programmes.