Inhibition of epigenetic and cell cycle-related targets in glioblastoma cell lines reveals that onametostat reduces proliferation and viability in both normoxic and hypoxic conditions

Abstract

The choice of targeted therapies for treatment of glioblastoma patients is currently limited, and most glioblastoma patients die from the disease recurrence. Thus, systematic studies in simplified model systems are required to pinpoint the choice of targets for further exploration in clinical settings. Here, we report screening of 5 compounds targeting epigenetic writers or erasers and 6 compounds targeting cell cycle-regulating protein kinases against 3 glioblastoma cell lines following incubation under normoxic or hypoxic conditions. The viability/proliferation assay indicated that PRMT5 inhibitor onametostat was endowed with high potency under both normoxic and hypoxic conditions in cell lines that are strongly MGMT-positive (T98-G), weakly MGMT-positive (U-251 MG), or MGMT-negative (U-87 MG). In U-251 MG and U-87 MG cells, onametostat also affected the spheroid formation at concentrations lower than the currently used chemotherapeutic drug lomustine. In T98-G cell line, treatment with onametostat led to dramatic changes in the transcriptome profile by inducing the cell cycle arrest, suppressing RNA splicing, and down-regulating several major glioblastoma cell survival pathways. Further validation by immunostaining in three cell lines confirmed that onametostat affects cell cycle and causes reduction in nucleolar protein levels. In this way, inhibition of epigenetic targets might represent a viable strategy for glioblastoma treatment even in the case of decreased chemo- and radiation sensitivity, although further studies in clinically more relevant models are required.

Figure 1

Workflow of the current project. The figure was composed using BioRender.

Figure 2

Radar plots enabling comparison of viability pIC₅₀ values obtained for different compounds following incubation of cells under normoxic or hypoxic conditions. Cell lines: (A) T98-G, (B) U-251 MG, (C) U-87 MG; the incubation conditions are shown in the right bottom corner. The names of compounds are listed along the radar perimeter. The data was obtained by pooling all independent experiments (N ≥ 3). For clarity, no error bars are depicted and only one pIC₅₀ value is shown per compound (in case of compounds featuring the biphasic dose–response fit, only the largest pIC₅₀ value was chosen). The numbering of y-axis is shown in light grey. Please note that the y-scale ranges from 0 to 10 in (A,B), and from 0 to 12 in (C).

Figure 3

Effect of onametostat (ON) or lomustine (LO) on the spheroid formation in glioblastoma cell lines. (A) and (B) Show examples of spheroid morphology in U-251 MG cells and U-87 MG cells, respectively, following the 96-h treatment with indicated compounds in a single representative experiment; scale bar: 200 µm. (C) Summarizes normalized spheroid area and (D) summarizes normalized average intensity of propidium iodide staining per spheroid area using data pooled from 2 independent experiments (a total of 11–12 spheroids per condition); error bars indicate standard deviation. The grouped comparisons show statistical significance of differences for the parameters measured following incubation of cells with onametostat or lomustine versus 0.1% DMSO (one-way ANOVA): *** indicates P ≤ 0.001, ** indicates P ≤ 0.01, * indicates P ≤ 0.05, ns indicates not significant.

Figure 4

DEG counts in treatment comparisons and number of common DEGs in various treatments. (A) Number of DEGs identified as significantly enriched in the pairwise comparisons of differently treated cells (FDR < 0.05); ↑ indicates higher abundance in hypoxia- or onametostat-treated cells and ↓ indicates higher abundance in normoxia- or DMSO-treated cells. (B) Venn diagram of the DEGs in various treatment comparisons (FDR < 0.05 in any compared treatments). H, hypoxia; N, normoxia; ONAM, onametostat.

Figure 5

Volcano plots showing DEGs in treatment comparisons. (A) Hypoxic vs normoxic conditions in the T98-G cells treated with 0.1% DMSO; (B) Hypoxic vs normoxic conditions in the T98-G cells treated with 1 µM onametostat; (C) Onametostat- vs DMSO-treated T98-G cells following incubation in normoxia; (D) Onametostat- vs DMSO-treated T98-G cells following incubation in hypoxia. Top hits are marked with the name labels. In (A) and (B), DEGs coloured in orange feature binary logarithm of fold change values of below -1 or over 1, and DEGs coloured in blue feature the P-value cut-off of 10^–3. In (C) and (D), DEGs coloured in orange feature binary logarithm of fold change values of below -2 or over 2, and DEGs coloured in blue feature the P-value cut-off of 10^–32. H, hypoxia; N, normoxia; ONAM, onametostat.

Figure 6

Top cellular pathways identified by the Metascape platform based on the DEGs corresponding to the comparisons of the differently treated samples (FDR < 0.05). The columns indicate decimal logarithms of P-value. (A) Pathways downregulated in hypoxic vs normoxic conditions; no major overlapping pathways were found between the cells treated with 0.1% DMSO (white columns) or 1 µM onametostat (purple columns). (B) Pathways upregulated in hypoxic vs normoxic conditions; no major overlapping pathways were found between the cells treated with 0.1% DMSO (white columns) or 1 µM onametostat (purple columns). (C) Pathways downregulated in onametostat- vs DMSO-treated cells; one major overlapping pathway was found between the cells treated in normoxia (black columns) or hypoxia (light green columns). (D) Pathways upregulated in onametostat- vs DMSO-treated cells; five major overlapping pathways were found between the cells treated in normoxia (black columns) or hypoxia (light green columns). H, hypoxia; N, normoxia; ONAM, onametostat.

Figure 7

Effect of onametostat (ONAM) or lomustine (LOMU) on the abundance of cell cycle-related markers in nuclei of glioblastoma cell lines. Each box shows the range between the 2nd and 3rd quartile, while the whiskers shown the range between 5 and 95 percentile; thick solid line indicates median of each treatment and thin dotted line in the graph background indicates median of the non-treated cells. Cell lines: data for U-251 MG shown in (A,D,G,J); data for T-98G shown in (B,E,H,K); data for U-87 MG shown in (C,F,I,L). Quantified parameters: total intensity of pS10H3 staining in nucleus shown in (A–C) (N = 6 for each cell line); total intensity of AurA staining in nucleus shown in (D–F) (N = 3 for each cell line); total intensity of CCNA2 staining in nucleus shown in (G–I) (N = 3 for each cell line); total intensity of CENPF staining in nucleus shown in (J–L) (N = 3 for each cell line). The 48-h treatment conditions are listed at the bottom of the image and the total number of nuclei quantified in case of each condition is shown at the bottom of each panel. The paired comparisons show statistical significance of differences for the parameters measured following incubation of cells with onametostat or lomustine versus non-treated cells (Mann–Whitney U-test): *** indicates P ≤ 0.001, ** indicates P ≤ 0.01, * indicates P ≤ 0.05, ns indicates not significant.