Generation of Human iPSC-Derived Astrocytes with a mature star-shaped phenotype for CNS modeling

Abstract

The generation of astrocytes from human induced pluripotent stem cells has been hampered by either prolonged differentiation-spanning over two months-or by shorter protocols that generate immature astrocytes, devoid of salient mature astrocytic traits pivotal for central nervous system (CNS) modeling. We directed stable hiPSC-derived neuroepithelial stem cells to human iPSC-derived Astrocytes (hiAstrocytes) with a high percentage of star-shaped cells by orchestrating an astrocytic-tuned culturing environment in 28 days. We employed RT-qPCR and ICC to validate the astrocytic commitment of the neuroepithelial stem cells. To evaluate the inflammatory phenotype, we challenged the hiAstrocytes with the pro-inflammatory cytokine IL-1β (interleukin 1 beta) and quantitatively assessed the secretion profile of astrocyte-associated cytokines and the expression of intercellular adhesion molecule 1 (ICAM-1). Finally, we quantitatively assessed the capacity of hiAstrocytes to synthesize and export the antioxidant glutathione. In under 28 days, the generated cells express canonical and mature astrocytic markers, denoted by the expression of GFAP, AQP4 and ALDH1L1. In addition, the notion of a mature phenotype is reinforced by the expression of both astrocytic glutamate transporters EAAT1 and EAAT2. Thus, hiAstrocytes have a mature phenotype that encompasses traits critical in CNS modeling, including glutathione synthesis and secretion, upregulation of ICAM-1 and a cytokine secretion profile on a par with human fetal astrocytes. This protocol generates a multifaceted astrocytic model suitable for in vitro CNS disease modeling and personalized medicine.

Keywords: EAAT1; Glutathione; ICAM-1; NES; Neuroepithelial Stem Cells; hiAstrocytes; hiPSCs.

Fig. 1

Neuroepithelial stem cells generate star-shaped astrocytes in 28 days under astrocytogenic conditions. a) Schematic presentation of the duration of the protocol used to derive neuroepithelial stem cells and hiAstrocytes. b) Brightfield images of NES C9 for differentiation days 0,7,14, 21, 26 and 27. Scale bar 100 μm. c) Brightfield images of hiAstrocytes C9 (i) hiAstrocytes C7 (ii) and hiAstrocytes AF22 (iii) on day 26 of differentiation. Scale bar 50 μm

Fig. 2

HiAstrocytes have a distinct astrocytic mRNA expression profile that differs from spontaneously differentiated cells. mRNA analysis (ΔΔCt values, vs. NES C9, normalized to GAPDH) of a) SOX1, NES, NFIA and DCX and b) astrocyte-specific markers CD44, S100B, ALDH1L1, GFAP and AQP4 for hiAstrocytes C9 (purple) and HFA (orange). Data shown from n = 3 independent experiments. Error bars represent ± SD. Statistical analysis was done on the ΔCt values by using multiple unpaired student’s test (Holm-Šídák method) with Welch correction. *p < 0.05. c) Heatmap mRNA levels (ΔΔCt, vs. each respective NES line) of hiAstrocytes and SDCs from NES C9, C7 and AF22 d) PCA plot and e) Hierarchical clustering of NES C9, C7 and AF22, their corresponding differentiated astrocytes (hiAstro), spontaneously differentiated cells (SDCs) and HFA, color scale in ΔCt, GOI-house-keeping gene, highest expression, ΔCt = -1,17 for VIMENTIN in hiAstro C9, and no expression, ΔCt = 15.8 for RELN in NES C9). HFA from one isolation were used. Normalized ΔCt values were used for the PCA plot and hierarchical clustering, all Ct values above 35 were assigned to 35 to avoid overestimation of the results

Fig. 3

Quantification of astrocytic markers shows a superior protein profile compared to HFA. a) ICC images of astrocytic markers S100B, CD44, GFAP, AQP4 and ALDH1L1 for hiAstrocytes C9, C7 and AF22 and HFA. Scale bar 50 μm. b) Quantification of astrocytic markers S100B, CD44, GFAP, AQP4 and ALDH1L1; each dot represents one field of view. c) Violin plot depicting the average intensity (per cell) of S100B for hiAstrocytes. [C9 (1161 cells analyzed), hiAstrocytes C7 (1725 cells analyzed), hiAstrocytes AF22 (1295 cells analyzed) and HFA (2614 cells analyzed)]

Fig. 4

Quantification of astrocytic processes shows a more elaborate, star-shaped phenotype. ICC images of cytoskeletal marker VIMENTIN for a) hiAstrocytes C9, C7 and AF22 and b) HFA. Scale bar 50 μm. c) Quantification of astrocytic processes for hiAstrocytes C9, C7 and AF22, and HFA. Each dot represents the average of four fields of view (10x)

Fig. 5

HiAstrocytes exhibit functional EAAT1- and EAAT2-mediated glutamate uptake. a) Glutamate uptake assay for HFA (black), hiAstrocytes C9 (pink) and NES C9 (green), n = 3 technical replicates. Error bars represent ± SD. Statistical analysis was done using two-way ANOVA followed by Tukey’s post-hoc test. *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001. b) mRNA analysis of glutamate transporters SLC1A2 and SLC1A3 for hiAstrocytes C9 and HFA, n = 3 independent experiments. Error bars represent ± SD. Statistical analysis was done on the ΔCt values by using multiple unpaired student’s test (Holm-Šídák method) with Welch correction. *p < 0.05. c) ICC images of glutamate transporters EAAT1 and EAAT2 for hiAstrocytes C9 and HFA

Fig. 6

HiAstrocytes harbor unique inflammatory potency and antioxidant properties. Secretion levels of a) IL-8 and b) IL-6 in NES C9, C7 and AF22, their corresponding astrocyte differentiation (hiAstrocytes C9, C7 and AF22) and HFA under basal (grey) and inflammatory conditions (IL-1β 50 ng/ml for 24 h, blue), n = 3 technical replicates. c) ICC images of ICAM-1 for hiAstrocytes C9, C7 and AF22, and HFA under basal and inflammatory conditions (IL-1β 50 ng/ml for 24 h). Scale bar 50 μm. d) Quantification of ICAM-1⁺ cells under basal (grey) condition or under inflammatory stimuli (IL-1β 50 ng/ml for 24 h, blue). Glutathione levels e) intracellularly and f) extracellularly in NES C9, C7 and AF22, their corresponding astrocyte differentiation (hiAstrocytes C9, C7 and AF22), and HFA, n = 3 technical replicates. Error bars represent ± SD. Statistical analysis was done using two-way ANOVA followed by Tukey’s post-hoc test. *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001