Shared molecular regulation of quiescence in neural and glioma stem cells reveals therapeutic vulnerabilities.

Abstract

Quiescence, a reversible state of cell-cycle arrest, is an adaptive feature of many adult tissue stem cells, including those in the adult brain. In gliomas, brain tumour stem cells that reside in a quiescent state preferentially survive chemotherapy and radiotherapy, highlighting their critical role in therapy resistance and disease progression. To date, it remains unclear whether the molecular programs governing these states are functionally conserved between neural stem cells and brain tumour stem cells. Here, we establish novel in vitro models to study quiescence and find that glioma stem cells are markedly more resistant to entering quiescence than neural stem cells, suggesting that glioma stem cell quiescence more closely resembles a slow-cycling phenotype or shallow quiescence. Nonetheless, direct comparison of quiescent neural stem cells and quiescent/slow-cycling glioma stem cells, as they transition towards proliferation, reveals conserved gene expression trajectories, indicating shared molecular mechanisms. Furthermore, we find that pathways influencing quiescence in neural stem cells exert similar effects in glioma stem cells, underscoring the functional parallels between these populations. Finally, we identify that inhibition of TGF-{beta} signalling might provide an avenue to improve current standard-of-care treatments by targeting quiescent glioma stem cells. Note on version 2This version corrects two figure errors and expands the Methods. The conclusions of are unchanged. ImagesIn version 1, debris adjacent to the organoid had been digitally removed from the drug-treated image in Figure 6D. The unmodified original is restored here. Images within Figures 3A and 6D have each been processed identically to one another. Scattered cellular debris was present to a similar extent in control and drug-treated samples. O_FIG O_LINKSMALLFIG WIDTH=153 HEIGHT=200 SRC="FIGDIR/small/634421v2_fig6.gif" ALT="Figure 6"> View larger version (51K): org.highwire.dtl.DTLVardef@19aef83org.highwire.dtl.DTLVardef@e2c1e1org.highwire.dtl.DTLVardef@126bf22org.highwire.dtl.DTLVardef@ce7acd_HPS_FORMAT_FIGEXP M_FIG O_FLOATNOFigure 6:C_FLOATNO TGF--R1 inhibition drives GSCs out of a quiescent state. (A) QBC395 cells were cultured in BMP4 quiescence media plus inhibitors for three days. Graphs show percentage of proliferating GSCs (SOX2+Ki67+) over total GSCs (SOX2+). Data were normalised to mean of quiescence media alone. (B) QBC395 cells were cultured in palbociclib quiescence media plus inhibitors for three days. Graphs show proportion of proliferating GSCs (SOX2+Ki67+) over total GSCs (SOX2+). Data were first normalised to vehicle control for each independent experiment and then quiescence media alone. (C) Schematic of glioblastoma organoid generation. Experiments began within 1-2 weeks of derivation to maximise cellular heterogeneity. Protocol adapted from Jacob and colleagues (60). (D) Image of glioblastoma organoid treated with 30 {micro}M LY-364397 or DMSO vehicle (control) for three days, stained for SOX2 (cyan) and Ki67 (magenta). Dashed lines demarcate organoid boundary. Scale bar: 100 {micro}m. Scattered cellular debris was observed in both control and drug-treated samples to a similar extent. (E) Proportion of proliferating GSCs (SOX2+Ki67+) over total GSCs (SOX2+) after glioblastoma organoids were cultured in the presence of inhibitors. Data were normalised to mean of respective vehicle control. Graphs in (A, B, E) show mean {+/-} SEM. Experiments in (A, B) represent independent experiments on separate passages (n=3-4). Dots in (E) represent individual organoids. Statistics: one-way repeated measures ANOVA with Holm-Sidaks multiple comparisons test in (A, B). Two-way ANOVA in (E) with multiple comparisons t-test. *P < 0.05, **P < 0.01, ***P < 0.001. Panel C created with BioRender (Agreement number: FY289ELRZ5). C_FIG O_FIG O_LINKSMALLFIG WIDTH=200 HEIGHT=52 SRC="FIGDIR/small/634421v2_fig3.gif" ALT="Figure 3"> View larger version (21K): org.highwire.dtl.DTLVardef@1e04d0corg.highwire.dtl.DTLVardef@e25e26org.highwire.dtl.DTLVardef@1934ab0org.highwire.dtl.DTLVardef@10c77fb_HPS_FORMAT_FIGEXP M_FIG O_FLOATNOFigure 3:C_FLOATNO BMP4 and palbociclib treatment induces patient-derived GSCs into shallow versus deep states of quiescence, respectively. (A) Treatment with 16 ng/mL BMP4 markedly reduces the proportion of proliferating NSCs (SOX2+Ki67+) in primary adult mouse SVZ cultures. (B) Treatment with 16 ng/mL BMP4 modestly reduces the proportion of proliferating GSCs (SOX2+Ki67+) in primary patient-derived cultures (QBC395 cell line). (C) Concentration-response curve of mouse NSCs and patient-derived GSCs (QBC395 cell line) treated with BMP4, measuring proportion of proliferating stem cells (SOX2+Ki67+) relative to control media. Data were normalised to mean of control. (D) Palbociclib treatment markedly reduces the proportion of proliferating GSCs (SOX2+Ki67+) relative to control media. Data were normalised to mean of control. GSC graphs in (C, D) show mean {+/-} SEM of independent experiments conducted on separate passages (n=3-4). NSC graph in (C) shows mean {+/-} SEM of individual cell lines derived from SVZ of different mice (n=4). Scale bar: 15 {micro}m in (A, B). Cross in (D) represents concentration-response at 200 nM. Statistics in (C): two-way ANOVA reporting main effect of cell-line. C_FIG DataThe glioblastoma organoid data in Figure 6E have been updated following re-imaging and blinded re-analysis. Reported effect sizes have changed slightly; the conclusions are unchanged. Figure S2 has likewise been updated following re-imaging and re-analysis. MethodsThe Methods have been expanded to describe the analysis more fully, including data exclusion criteria, normalisation, and blinding. Minor changes to figure legends and the main text are not itemised.