Identification of Time-dependent Phenotypic Shifts in Patient-derived Glioma Stem-like Cells After Chemoradiotherapy

BACKGROUND
Glioblastoma remains among the deadliest tumors without a suitable cure. The failure of virtually all clinical trials from the past 20 years can be attributed to the extensive (epi)genetic and phenotypic heterogeneity within and across GBM patients but also to the extensive plasticity of cancer stem-like cells that can easily adapt and drive resistance to treatment. A fundamental understanding of this plasticity and its underlying mechanisms is still largely missing, but it remains pivotal to identify suitable targets that can prevent cells from surviving treatment.
METHODS
In this project, we start by defining the landscape of standard-of-care-induced plasticity using multi-omic profiling of patient-derived glioma stem-like cells (PDGSCLs) after radiation therapy (RT, 2 Gy) and temozolomide (TMZ, 150 μM). In a pilot experiment, bulk RNA sequencing was performed on six PDGSCLs at different time points after RT+TMZ (16 hours, 3 days, 10 days). A single-cell RNA sequencing (scRNAseq) experiment has been set up in a similar fashion, including the 16-h and 3-d time points.
RESULTS
Analysis of the bulk transcriptomic data, including module scoring based on transcriptional cell states defined by Neftel et al. (2019), revealed a time-dependent shift in phenotype after treatment. Two of the six PDGSCLs did not show clear indications of plasticity, while two others showed minor shifts in phenotype and two PDGSCLs displayed clear shifts towards the mesenchymal (MES) phenotype, with one of them reverting back to its original state after 10 days. Pathway analysis additionally indicated an upregulation of MES-associated processes including ameboidal-type cell migration, cell-substrate adhesion and wound healing, with differentially upregulated pathways between sensitive and resistant PDGSCLs that both displayed MES-shifts.
CONCLUSIONS
Our results confirm that treatment can induce phenotypic changes in PDGSCLs in a cell line-specific manner, suggesting different mechanisms can lead to the enrichment of the same phenotype. Moreover, our time course data highlights that the timing of read-out is important as well, and comparing tumors at diagnosis vs. at recurrence might not suffice to fully understand this plastic behavior, for which these intermediate time points are more relevant. The scRNAseq experiment is currently ongoing and will inform on the gene regulatory networks responsible for the observed shifts while also highlighting the heterogeneity within these PDGSCLs with greater resolution, as seen in the patient population. This project is funded by FWO (11B2323N) and the Foundation against Cancer (F/2022/2026).