Analyzing the Effect of the K27M Mutation in Histone H3.3 on Gene Expression in Pluripotent Stem Cells and Neural Stem Cells.. Analyzing the Effect of the K27M Mutation in Histone H3.3 on Gene Expression in Pluripotent Stem Cells and Neural Stem Cells.

Diffuse intrinsic pontine glioma (DIPG) is an aggressive childhood tumor of the brainstem with currently no curative treatment available. The vast majority of DIPGs carry a mutation in histone H3 leading to a lysine 27-to-methionine exchange (H3K27M). The H3K27M mutation has been shown to inhibit EZH2, the catalytic subunit of the polycomb repressive complex 2, which presumably leads to global reduction of H3K27 trimethylation and upregulation of tumor-driving genes. However, the exact pathomechanism and cellular context remain elusive. Recent single cell transcriptome data suggested the DIPG cell of origin to be of oligodendroglial lineage, whereas in vivo studies demonstrated neural stem or progenitor cells (NPCs, NSCs) to be susceptible for malignant transformation upon H3K27M expression. Here, we used targeted human induced pluripotent stem cells (iPSCs) carrying an inducible H3.3-K27M allele in the endogenous H3F3A locus together with specific differentiation methods to study the effects of the mutation in disease-relevant cell identities of the neural lineage. Phenotypic, transcriptomic, and chromatin immunoprecipitation sequencing (ChIP-seq) analyses revealed a distinct role of H3.3-K27M for deregulating bivalent promoter-associated developmental genes producing diverse outcomes in different cell types. While being fatal for iPSCs, H3.3-K27M increased proliferation in NSCs and to a lesser extent in oligodendroglial progenitor cells (OPCs) but not in astroglial-restricted precursors. Importantly, we demonstrated that only NSCs but not OPCs gave rise to tumors upon induction of the H3.3-K27M and TP53 inactivation in an orthotopic xenograft model faithfully recapitulating human DIPGs. Thus, we provide evidence that indeed only NSCs can develop H3.3-K27M-driven DIPG-like tumors but that the mutation actively drives acquisition of oligodendroglial characteristics, thus explaining the different observations of cell identity. Identification of the originating cell type may help to determine specific vulnerabilities of this tumor for developing targeted therapies. Overall design: We sought to investigate the gene expression changes in induced pluripotent stem cells (iPSCs) and neural stem cells (NSCs) upon endogenous expression of the H3.3-K27M mutation. Further, we wanted to perform transcriptome profiling of iPSC-NSC-derived xenograft tumors. For this purpose, we generated human iPSCs carrying a Cre-inducible H3.3-K27M reading frame in one copy of the H3F3A locus. Lentiviral infection delivering Flp enzyme expression loops out the selection marker from the initial targeting cassette and preserves the wildtype H3.3 variant (control). Lentiviral delivery of Cre enzyme loops out the wildtype H3.3 reading frame and activates expression of the H3.3-K27M variant (mutation). Lentiviral infection of Cre and Flp was conducted at the iPSC stage for two engineered lines from two individuals as well as after NSC differentiation of one engineered iPSC line. For xenograft experiments, iPSC-derived NSCs were infected with lentivirus expressing Cre-GFP and retrovirus expressing shRNA against TP53 before orthotopic injection into immunodeficient mice. Total RNA from control and mutant iPSCs was extracted between 5-6 days after lentiviral infection encompassing two biological replicates per cell line. Total RNA from control and mutant iPSC-derived NSCs was extracted around day 10, 18, 26, and 30 after lentiviral infection encompassing one biological replicate per time point. Total RNA from iPSC-NSC-derived tumors was extracted from lesion areas after sacrificing sick animals according to defined endpoints of tumor burden. Two tumors were harvested for molecular analysis. Transcriptome analysis were performed on Affymetrix HG-U133_Plus 2 arrays.