TP53 mutated glioblastoma stem-like cell cultures are sensitive to dual mTORC1/2 inhibition while resistance in TP53 wild type cultures can be overcome by combined inhibition of mTORC1/2 and Bcl-2

Abstract

Background: Glioblastoma is the most malignant tumor of the central nervous system and still lacks effective treatment. This study explores mutational biomarkers of 11 drugs targeting either the RTK/Ras/PI3K, the p53 or the Rb pathway using 25 patient-derived glioblastoma stem-like cell cultures (GSCs).

Results: We found that TP53 mutated GSCs were approximately 3.5 fold more sensitive to dual inhibition of mammalian target of rapamycin complex 1 and 2 (mTORC1/2) compared to wild type GSCs. We identified that Bcl-2(Thr56/Ser70) phosphorylation contributed to the resistance of TP53 wild type GSCs against dual mTORC1/2 inhibition. The Bcl-2 inhibitor ABT-263 (navitoclax) increased sensitivity to the mTORC1/2 inhibitor AZD8055 in TP53 wild type GSCs, while sensitivity to AZD8055 in TP53 mutated GSCs remained unchanged.

Conclusion: Our data suggest that Bcl-2 confers resistance to mTORC1/2 inhibitors in TP53 wild type GSCs and that combined inhibition of both mTORC1/2 and Bcl-2 is worthwhile to explore further in TP53 wild type glioblastomas, whereas in TP53 mutated glioblastomas dual mTORC1/2 inhibitors should be explored.

Keywords: brain tumor; genetic biomarkers; personalized medicine; resistance; small molecule kinase inhibitors.

Figure 1. GI50 values of 25 GSCs for a panel of small molecule compounds

A. Boxplot and dotplot in which each dot represents the GI50 value (μM) of a GSC to a specific compound. B. Supervised clustering of Z-transformed GI50 values (μM) was performed across the pathway-classified compounds. Unsupervised clustering was performed across the GSCs by complete linkage using euclidean distance. White, missing value; black rectangle, cluster of GSCs resistant to several compounds targeting the RTK/Ras/PI3K or Rb pathway.

Figure 2. TP53^mut GSCs are uniformly sensitive to dual mTORC1/2 inhibition and not to mTORC1 inhibition

A, B. Boxplot and dotplot in which each dot, stratified by their TP53 mutation status, represents the GI50 values (μM) of AZD2014 or AZD8055 (dual mTORC1/2 inhibitors) for GSCs. C, D. Live-image monitoring of proliferation in response to increasing concentrations of AZD8055. E. Spearman correlation of the GI50 values (μM) of different mTORC1 and dual mTORC1/2 inhibitors for 10 GSCs. F. Dose-response curves of the same 10 GSCs. The colors indicate the TP53 mutation status. Pink, TP53^mut; blue, TP53^wt; MUT, mutated; WT, wild type.

Figure 3. Kinome profiling identifies potential phosphosites implicated in resistance to AZD8055

A. Schematic depiction of the approach for dynamic kinome profiling. B. Waterfall plot of the 180 phosphosites. The values above and below 0 indicate respectively hyper- and hypophosphorylated phosphosites in the resistant TP53^wt (red) GSC relative to the sensitive TP53^mut (green) GSCs after exposure to AZD8055. C. Increasing doses of ABT-263 and AZD8055 were tested in the AZD8055-resistant TP53^wt GSC (GS281) and two AZD8055-sensitive TP53^mut GSCs (GS149 and GS186c).