BMP, MEK, and WNT inhibition with NGN2 expression for rapid generation of hiPSC-derived neurons amenable to regional patterning

Abstract

Human induced pluripotent stem cells (hiPSCs) are a promising tool for studying neurological diseases and developing therapies for neurodegenerative diseases. Differentiation of hiPSCs into neurons can be achieved by dual SMAD inhibition (dSMADi) or by induced neurogenin 2 (NGN2) overexpression ("iNGN2"). Starting directly from hiPSCs, iNGN2 shortens the time to a neuronal stage but leads to neurons partially resembling peripheral or posterior fates while dSMADi more faithfully recapitulates telencephalic development. To modify the iNGN2 approach, we applied an accelerated induction paradigm that is dependent on the inhibition of BMP, MEK, and WNT pathways ("BMWi"), to commit hiPSCs into a telencephalic fate before iNGN2. The resulting neurons showed strong expression of telencephalic markers, with decreased levels of peripheral and posterior marker genes compared to iNGN2 alone. The resulting telencephalic neurons are suitable for a tau aggregation assay. Furthermore, we could demonstrate that during BMWi treatment, the cells are amenable to additional regional patterning cues. This allowed the generation of neurons from different regions of the CNS and peripheral nervous system (PNS), which will significantly facilitate in vitro modeling of a range of neurodevelopmental and neurodegenerative disorders.

Keywords: Alzheimer's disease; Parkinson's disease; cortical neurons; disease modeling; dopaminergic neurons; human induced pluripotent stem cells; motoneurons; neural development; neurodegenerative diseases.

Graphical abstract

Figure 1

Identifying an induction paradigm for neuroectoderm (A) Schematic of the signaling pathways and their inhibitors (red) used in this study. (B) Time course of gene expression measured by real-time qPCR after treatment of hiPSCs (hiPSC_1–6) with different combinations of inhibitors from day 0 to day 6 (results are shown as means ± SEM; N = 6 different cell lines, see also Figures S1B and S2). (C) Representative IF staining (hiPSC_5) after 6 days of treatment. (D) Gene expression after 6 days of treatment with the inhibitor combination indicated. 6 different cell lines (hiPSC_1–6) were used, and a total of 12 independent differentiations were performed (N = 6 cell lines, each with n = 2 independent differentiations). (E) IF staining of hiPSC_5 after treatment with BMWi for 4 and 6 days (scale bars: 200 μm, insert 3× zoom-in, ^∗p < 0.05, ^∗∗p < 0.01, ^∗∗∗p < 0.001; ● = hiPSC_1, ▲ = hiPSC_2, ■ = hiPSC_3, ♦ = hiPSC_4, ★ = hiPSC_5, ⬣ = hiPSC_6).

Figure 2

Combined BMWi and iNGN2 accelerate differentiation of hiPSCs into telencephalic neurons (A) Differentiation protocols used. Abbreviations: RO: RO4929097 γ-secretase inhibitor; DOX, doxycycline, induces NGN2 overexpression; ROCKi, Y-27632 ROCK inhibitor; dSMADi (LDN193189 + SB431542): inhibition of ALK2/3 and ALK4/5/7, dual SMAD inhibition; CHIR, CHIR99021 GSK3 inhibitor/WNT activator. Medium compositions: see Methods. (B) Gene expression at comparable times of different differentiation protocols. Neurons were replated in NMM+ or NMM-. dSMADi neurons were used as a telencephalic standard and prepared according to the original protocol. 5 independent differentiations (different cell lines) were carried out (sBMWi, BMWi, iNGN2; hiPSC_1–5) and as reference 3 cell lines (hiPSC_7–9) with the dSMADi protocol (means ± SEM). Expression levels were normalized to the average expression of undifferentiated hiPSCs. (C) Representative IF images of mature neuron cultures (hiPSC_5) of BMWi neurons day 14 after replating (scale bars: 200 μm, insert 3× zoom-in) (^∗p < 0.05, ^∗∗p < 0.01, ^∗∗∗p < 0.001).

Figure 3

BMWi neurons show telencephalic transcriptome and electrophysiological activity profiles (A) RNA-seq results showing the comparison of iNGN2 and BMWi protocols. The neurons were either plated in NMM+ or NMM-. BMWi cells were also replated in NMM-, and the medium was changed to NMM+ after 7 days (BMWi NMM-/+ final). The sequencing was carried out with 3 different cell lines with at least 3 technical replicates (hiPSC_1–3). Heatmap showing the expression levels of selected genes, color-coded by different regions/types of interest. Hierarchical clustering shows a clear separation of the different time periods of the protocols. Genes showing differential expression between iNGN2 and BMWi protocols are indicated in green boxes (see also Figure S6A schematic protocol, S6B PCA, S6C heatmap 500 most variable genes, and S6E heatmap subtype clustering). (B) Fold changes (FCs) between NPCs at the replating step. Selected marker genes are shown. (C) Differential gene analysis of iNGN2 neurons (NMM+ and NMM-) vs. BMWi neurons (NMM-). Selected marker genes are shown. (D) RNA-seq comparing BMWi and dSMAD neurons at different time points of maturation. Samples were taken 2, 3, and 4 weeks after final replating of the respective protocol (dSMAD samples 2 and 4 weeks). Sequencing was performed on 3 different cell lines of BMWi neurons (hiPSC_1,3,5) and 4 different cell lines of dSMAD neurons (hiPSC_1,3,5,7) (all conditions with 3 technical replicates). Heatmap showing the expression levels of selected genes, color-coded by different brain regions/marker types of interest, as shown in (A). Hierarchical clustering confirms a clear separation of the different protocols. Genes not detected across samples were indicated with dark-blue color, at higher intensity than minimum expression levels. (E) Results of snRNA-seq of iNGN2 neurons, BMWi neurons, BMWi NPCs of day 6, and hiPSCs (hiPSC_3). Samples of the final neurons were taken 14 days after final replating of the respective protocol. In UMAP plots, the expression of some neural marker genes is shown. iNGN2 and BMWi neurons were also compared regarding neuronal subtype marker genes in a bubble plot. (F) Development of electrophysiological activity of the iNGN2 and BMWi neuron cultures. Medium of the neurons replated in NMM- was switched to NMM+ after 7 days. The mean firing rate of the action potentials (Hz) is plotted against the days after the final replating, across 5 different cell lines (hiPSC_1–5) each with 6 technical replicates (results are shown as means ± SEM). The activity of the neurons was recorded for 8 min. Insert shows hiPSC_5 neurons on MEA plate (day 29, NMM-). (G) Network activity pattern of the MEA measurements from (F). 2-min sections are shown (day 49, hiPSC_4).

Figure 4

Tau seeding in telencephalic neurons (A) Representative IF images of tau sPHF-seeded telencephalic neurons of the BMWi and dSMADi protocol 4 weeks post seeding stained with MC1 and pan-neuronal marker MAP2 (scale bars: 100 μm). (B) Quantification of tau aggregates (MC1-positive area [μm2]) and number of healthy nuclei in dSMADi and BMWi neurons 3 and 4 weeks post seeding with sPHFs. The number of healthy nuclei and the MC1-positive area (μm²) were determined using high-content imaging (means ± SEM from N = 4 independent experiments with each n = 3 replicate wells) (^∗p < 0.05, ^∗∗p < 0.01, ^∗∗∗p < 0.001).

Figure 5

Patterning during BMWi treatment (A) Schematic representation of the BMWi protocol with additional patterning. (B) Gene expression of different regional genes to assess patterning along the dorsoventral axes. The experiments were carried out with 4 different cell lines (hiPSC_1–4, means ± SEM) and normalized to expression of housekeeping genes. (C) Representative IF images (hiPSC_3) of NPCs expressing either the neuroectodermal precursor marker PAX6 or the floorplate marker FOXA2 with additional ventral patterning by SAG (scale bars: 200 μm, insert 3× zoom-in). sBMWi and BMWi NPCs on day 4/day 6 are shown. Quantification of PAX6 and FOXA2 of 3 cell lines is also shown (hiPSC_1,2,4, means ± SEM) (^∗p < 0.05, ^∗∗p < 0.01, ^∗∗∗p < 0.001; ● = hiPSC_1, ▲ = hiPSC_2, ■ = hiPSC_3, ♦ = hiPSC_4, ★ = hiPSC_5).

Figure 6

Patterning of neuronal subtypes (A) Schematic representation of ventral midbrain BWMi patterning. (B) Gene expression of relevant marker genes for ventral midbrain (N = 4 cell lines, hiPSC_1–4). (C) Representative IF images (day 20, hiPSC_3) of ventral midbrain-patterned neurons. (D) Schematic representation of the BMWi MN protocol. (E) Comparison of gene expression at different time points of the BMWi MN protocol (N = 3 cell lines, hiPSC_1–3) and the protocol for differentiation of MN according to Du et al (hiPSC_1,8). (F) Representative IF images of BMWi MN (day 22, hiPSC_1) (all scale bars: 200 μm, insert 3× zoom-in, all results normalized to housekeeping gene expression, means ± SEM) (^∗p < 0.05, ^∗∗p < 0.01, ^∗∗∗p < 0.001; ● = hiPSC_1, ▲ = hiPSC_2, ■ = hiPSC_3, ♦ = hiPSC_4, ◻ = hiPSC_8).

Figure 7

Patterning of DRG (A) Schematic representation of the shortBMi with sensory neuron patterning protocol (sensBMi). (B) Representative IF images of the expression of sensory neuron progenitors after 4 and 6 days of the neural crest marker SOX10 (hiPSC_5, scale bars: 200 μm, insert 3× zoom-in). (C) Gene expression of relevant marker genes for sensory neurons normalized to average expression in hiPSCs (N = 5 different cell lines, hiPSC_1–5, means ± SEM). (D) Representative IF images of the expression of sensory neuron marker genes on day 20 (hiPSC_5) (scale bars: 200 μm, insert 3× zoom-in). (E) MEA assay of sensory sBMi neurons 3 weeks after replating (N = 5 different cell lines each n = 2/3 technical replicates, hiPSC_1–5). 100 nM CAP diluted in NMM+ was added during measurement. As control, just NMM+ with solvent was added (^∗p < 0.05, ^∗∗p < 0.01, ^∗∗∗p < 0.001; ● = hiPSC_1, ▲ = hiPSC_2, ■ = hiPSC_3, ♦ = hiPSC_4, ★ = hiPSC_5).