Efficient derivation of functional astrocytes from human induced pluripotent stem cells (hiPSCs)

Abstract

Astrocytes are specialized glial cell types of the central nervous system (CNS) with remarkably high abundance, morphological and functional diversity. Astrocytes maintain neural metabolic support, synapse regulation, blood-brain barrier integrity and immunological homeostasis through intricate interactions with other cells, including neurons, microglia, pericytes and lymphocytes. Due to their extensive intercellular crosstalks, astrocytes are also implicated in the pathogenesis of CNS disorders, such as ALS (amyotrophic lateral sclerosis), Parkinson's disease and Alzheimer's disease. Despite the critical importance of astrocytes in neurodegeneration and neuroinflammation are recognized, the lack of suitable in vitro systems limits their availability for modeling human brain pathologies. Here, we report the time-efficient, reproducible generation of astrocytes from human induced pluripotent stem cells (hiPSCs). Our hiPSC-derived astrocytes expressed characteristic astrocyte markers, such as GFAP, S100b, ALDH1L1 and AQP4. Furthermore, hiPSC-derived astrocytes displayed spontaneous calcium transients and responded to inflammatory stimuli by the secretion of type A1 and type A2 astrocyte-related cytokines.

Fig 1. Representative cell morphology changes during astrocytes differentiation.

(A) Overview of the in vitro method used for astrocytes generation from hiPSC-derived NPCs. (B) Light microscopic images were taken before initiating astrocyte induction (Day 0), Day 7 of astrocyte induction, Day 21 of astrocyte induction and Day 42 of astrocyte maturation. (Scalebar: 50 μm).

Fig 2. The expression of astrocytic markers.

(A) Western blotting (WB) was used to analyze the typical astrocytic marker protein expression (GFAP, VIM, and ALDHL1) during the differentiation of astrocytes from iPSCs. After protein separation by SDS-PAGE, the proteins were blotted onto a PVDF membrane for antibody detection. Chemiluminescence detection was used to visualize the labeled bands using β-tubulin as a loading control. The average intensities of the signals ± SD were calculated from WB images of two independent experiments with two individual cell lines (BIOT.009 and BIOT.021) and compared to iPSC (Student’s t-test, *p < 0.05, ** p < 0.01, n = 4). The original uncropped blots are presented in S1 Fig. (B) Representative images show the expression of AP markers CD44 (green) and NFIA (red), detected by immunocytochemistry. (Scalebar: 100 μm). (C) Representative immunocytochemical staining of GFAP (green), NFIA (red) and AQP4 (yellow) in Day 42 astrocytes. On Day 42, the expression of astrocyte markers ALDH1L1 (D) and GS (E) were also detected (yellow, scalebar: 100 μm). DAPI was used as a nuclear counterstain.

Fig 3. The degree of differentiation of the astrocytes derived from hiPSCs.

(A) Heatmap showing basemean expression profile of the 52 marker genes from hiPSCs-derived astrocytes from the study of Perriot et al. 2018 [23] and TCW et al. 2017 [20] and fetal astrocytes from the study of Zhang et al. 2016 [34]. The 52 markers are listed on the x-axis, and hiPSCs-derived astrocytes on the y-axis. Results are expressed as the Z-score of the log2 basemean values. (B) Heatmap showing the percentage of generated astrocytic cultures sharing the expression signatures with the human adult brain tissue extracted during surgery [34]. Cellular deconvolution was carried out using BrainDeconvShiny [32] with the assistance of dtangle [33].

Fig 4. Cytokine/chemokine profiling.

(A) The applied array is suitable for detecting 105 human cytokines, chemokines, growth factors and other soluble proteins (S3 Table). Left panels show their secretion from control, non-stimulated astrocyte cultures, right panels show the secretion from astrocyte cultures stimulated with IL-1β and TNF-α. Framed spots in red indicate the stimulator molecules (IL-1β and TNF-α) and the spots with black frame indicate the cytokines that were secreted in higher amounts upon stimulation. (B) List of the top-24 cytokines/chemokines whose secretion levels were altered following IL-1β+TNFα treatment. Spot signal intensity was quantified using GelQuant.NET software provided by biochemlabsolutions.com. Newly expressed molecules that were not detectable in the cell culture supernatant samples without treatment are indicated with > 100-fold change. IL-1β and TNF-α used for pro-inflammatory stimulation were not listed.

Fig 5. Day 21 astroglial progenitor cells and Day 42 astrocytes conduct calcium waves.

Astroglial progenitor cells (AP) at Day 21 and astrocytes (ACs) at Day 42 were loaded with the calcium-sensitive dye Fluo-4. Fluorescence intensity changes were recorded by time-lapse imaging (1 image/second for 5 minutes). Curves representing fluorescence change (DF/F0) over time—with signal width (SW) and amplitude (AMP) values—were obtained in CaSian. APs from the BIOT.009 (A) and BIOT.021 (B) cell lines displayed slow spikes with high SWs, whereas the maturated astrocytes from BIOT.009 (C) and BIOT.021 (D) lines produced rhythmic calcium waves with higher frequencies.