Establishment of human iPSC-based models for the study and targeting of glioma initiating cells

Abstract

Glioma tumour-initiating cells (GTICs) can originate upon the transformation of neural progenitor cells (NPCs). Studies on GTICs have focused on primary tumours from which GTICs could be isolated and the use of human embryonic material. Recently, the somatic genomic landscape of human gliomas has been reported. RTK (receptor tyrosine kinase) and p53 signalling were found dysregulated in ∼90% and 86% of all primary tumours analysed, respectively. Here we report on the use of human-induced pluripotent stem cells (hiPSCs) for modelling gliomagenesis. Dysregulation of RTK and p53 signalling in hiPSC-derived NPCs (iNPCs) recapitulates GTIC properties in vitro. In vivo transplantation of transformed iNPCs leads to highly aggressive tumours containing undifferentiated stem cells and their differentiated derivatives. Metabolic modulation compromises GTIC viability. Last, screening of 101 anti-cancer compounds identifies three molecules specifically targeting transformed iNPCs and primary GTICs. Together, our results highlight the potential of hiPSCs for studying human tumourigenesis.

Figure 1. Transformation of human iNPCs results in the acquisition of a GTIC-like phenotype in vitro.

(a) Flow cytometry analysis for the indicated markers in different transformed and wild-type iNPC groups (n=4/group with technical duplicates). (b) Transformation of human iNPCs leads to increased self-renewal properties only in cells where PI3K and MAPK signalling is dysregulated by overexpression of Ras/EGFR/Src mutant genes as highlighted by single-cell self-renewal assays (n=3/group with 24 technical replicates). (c) iNPC transformation results in the upregulation of endogenous NANOG as highlighted by western blot analyses. (d) NANOG expression is upregulated in primary human adult glioma samples. (e) Unsupervised cluster analysis of mRNA expression data highlighting the similarities between transformed iNPCs and primary GTICs. Please note that _p53KDiNPCs are more similar to _WTiNPCs than to primary GTICs. (f,g) Transformation of human iNPCs induces glycolytic (f) and oxidative phosphorylation changes as indicated by Seahorse analysis (g), as measured by ECAR and OCR (n=4/group with four technical replicates). Data are represented as mean ±s.d. P values were calculated by Student's t-test or Mann–Whitney test when appropriate and represented as follows: *P<0.05. Scale bars, 200 μm (d).

Figure 2. Transformation of human iNPCs results in the acquisition of GTIC-like properties in vivo.

(a) Representative pictures demonstrating the formation of vascularized tumours only in animals receiving transformed, and not wild-type, human iNPCs. (b) Haematoxylin-eosin staining demonstrating the presence of highly aggressive brain tumours upon orthotopic transplantation of transformed human iNPCs into the murine brain. (c) Immunofluorescence analysis demonstrating the presence of undifferentiated SOX2+ cells (red) as well as differentiation into the three major neural lineages upon transplantation of transformed human iNPCs. Please note the higher cellularity in tumours derived upon transplantation of _{p53KD-Ras/EGFR/Src}iNPCs. HUNU (green) indicates human nuclear antigen staining; O4 (green) indicates oligodendrocyte differentiation; Tuj1 (red) indicates neuronal differentiation; and GFAP (red) indicates glial differentiation. pERK (purple) and pAKT (purple) were only detected in tumours generated upon transplantation of iNPCs-overexpressing mutant versions of Ras/EGFR/Src. Cells were counterstained with DAPI (blue) and their proliferative state monitored by Ki67 immunostaining (red). (d) Brain tumours derived from human _{p53KD-Ras/EGFR/Src}iNPCs demonstrated the presence of human vessels derived from the injected GTIC-like cells as indicated by CD31 and HUNU co-localization. n=5 animals/condition. A minimum of three brain sections/animal/condition were analysed. Scale bars, 200 μm (b); 100 μm or 25 μm as indicated (c); and 25 μm (d).

Figure 3. Human iNPCs transformation leads to differential SOX2 binding and metabolic reprogramming.

(a) Chromatin Immunoprecipitation with SOX2 antibodies followed by deep-sequencing demonstrates differential binding of SOX2 in _WTiNPCs as compared with transformed iNPCs and primary GTICs. (b,c) Modulation of PI3K and MAPK signalling pathways with LY294002 (LY) and PD098059 (PD), respectively, reverts the metabolic changes observed in transformed iNPCs as indicated by seahorse analysis. MAPK inhibition restores glycolytic activities (b), whereas inhibition of both, PI3K and MAPK, reduced oxidative phosphorylation to the levels observed in _WTiNPCs (c) (n=4 group/condition with four technical replicates). (d) Inhibition of glycolysis by 2-DG compromises self-renewal properties in transformed iNPCs (n=4 group with four technical replicates). (e) Chemical induced glycolysis inhibition (2-DG) and ROS production (Rot/AA) compromises transformed iNPC viability in MTS assays (n=4 group/condition with four technical replicates). Data are represented as mean±s.d. P values were calculated by Student's t-test. *P<0.05.

Figure 4. Anti-cancer compound screening identifies chemicals specifically targeting transformed iNPCs and primary GTICs.

(a) Annexin V staining, as measured by flow cytometry, validates the role of lead compounds at the indicated concentrations and demonstrates a specific effect against primary GTICs and transformed iNPCs for three of the compounds analysed (Nelarabine, Letrozole and Capecitabine). In contrast, Cabazitaxel demonstrated group-dependent effects and targeting of _WTiNPCs. Data are presented as a ratio to the values obtained in the respective DMSO control groups. Negative values indicate increase viability in where basal Annexin V staining was higher in DMSO samples as compared with those treated with specific compounds; (n=3 group/condition with three technical replicates). (b,c) Treatment of transformed iNPCs and primary GTICs with the identified compounds and metabolic modulators compromises glioma stem cell properties including migration (b, n=3 group/condition with three technical replicates) and self-renewal potential in single-cell assays (c; n=3 group/condition with 24 technical replicates). (d) Tumour size measured on brain organotypic cultures for each indicated condition. Data are independently plotted for each organotypic brain slice analysed (n=>3 condition with technical triplicates). (e) Representative pictures of organotypic brain slices injected with primary GTICs and treated with each of the indicated compound. Brain slices were immunostained with the indicated markers as follows: DAPI (blue), GFP (green) and Ki67 (red). Data are represented as mean±s.d. P values were calculated by Student's t-test or Mann–Whitney test when appropriate. *P<0.05. Scale bars, 100 μm (e).