Exploring Novel Therapeutic Opportunities for Glioblastoma Using Patient-Derived Cell Cultures

Abstract

Glioblastomas (GBM) are the most common, primary brain tumors in adults. Despite advances in neurosurgery and radio- and chemotherapy, the median survival of GBM patients is 15 months. Recent large-scale genomic, transcriptomic and epigenetic analyses have shown the cellular and molecular heterogeneity of GBMs, which hampers the outcomes of standard therapies. We have established 13 GBM-derived cell cultures from fresh tumor specimens and characterized them molecularly using RNA-seq, immunoblotting and immunocytochemistry. Evaluation of proneural (OLIG2, IDH1^R132H, TP53 and PDGFRα), classical (EGFR) and mesenchymal markers (CHI3L1/YKL40, CD44 and phospho-STAT3), and the expression of pluripotency (SOX2, OLIG2, NESTIN) and differentiation (GFAP, MAP2, β-Tubulin III) markers revealed the striking intertumor heterogeneity of primary GBM cell cultures. Upregulated expression of VIMENTIN, N-CADHERIN and CD44 at the mRNA/protein levels suggested increased epithelial-to-mesenchymal transition (EMT) in most studied cell cultures. The effects of temozolomide (TMZ) or doxorubicin (DOX) were tested in three GBM-derived cell cultures with different methylation status of the MGMT promoter. Amongst TMZ- or DOX-treated cultures, the strongest accumulation of the apoptotic markers caspase 7 and PARP were found in WG4 cells with methylated MGMT, suggesting that its methylation status predicts vulnerability to both drugs. As many GBM-derived cells showed high EGFR levels, we tested the effects of AG1478, an EGFR inhibitor, on downstream signaling pathways. AG1478 caused decreased levels of phospho-STAT3, and thus inhibition of active STAT3 augmented antitumor effects of DOX and TMZ in cells with methylated and intermediate status of MGMT. Altogether, our findings show that GBM-derived cell cultures mimic the considerable tumor heterogeneity, and that identifying patient-specific signaling vulnerabilities can assist in overcoming therapy resistance, by providing personalized combinatorial treatment recommendations.

Keywords: EGFR inhibitor (AG1478); EMT; MGMT; STAT3; cancer stem cells; doxorubicin; glioblastoma; temozolomide.

Figure 1

Characterization of glioma patient cohort and transcriptomic and protein-based subtyping of patient-derived primary cell cultures. (A,B) Graphical representation of sex (A) and age (B) of patients in the analyzed glioma cohort. Red line represents the mean age. (C) The proliferation capacity of glioblastoma-patient-derived primary cell cultures and normal human astrocytes (NHA) determined by MTT metabolism assay, n = 3, mean ± SD. (D) Doubling time of primary glioma cell cultures and NHA as a non-malignant control. (E) Transcriptomic data of patient-derived cell cultures represented as a heatmap with Verhaak signatures (MES, CL and PN) for glioblastoma subtypes. (F) Representative immunoblots illustrating the markers of glioblastoma subtypes: IDH1 ^R132H, CHI3L1, pSTAT3, STAT3 (MES); EGFR (CL); and PDGFRα, TP53 (PN) in cultures of human glioma cells, NHA and NTERA-2 cells (NTERA). (G) Quantification of immunoblots. The level of a protein of interest was compared to the level of NHA. NHA cells were set equal to 1 and marked by a solid black line. β-actin was used as a loading control. NHA served as a non-malignant control, whereas NTERA served as a positive control for stemness properties, n = 2, mean ± SD. The whole blots and molecular weight markers are shown in Supplementary Figure S8.

Figure 2

Stemness and differentiation markers across primary cell cultures. (A) RNA-seq of primary cell cultures represented as heatmap with stemness (indicated by blue color) and differentiation (red color) related genes. (B) Expression of selected differentiation (GFAP, TUBBIII) and stemness (SOX2, NESTIN) related genes in human patient-derived primary cultures and NHA as a control. The RT-qPCR data are shown as delta Ct values relative to the 18S expression. (C) Representative immunoblots showing GFAP, β-TUB III and SOX2 levels in glioma primary cultures, NHA and NTERA cells. (D) Quantification of immunoblots. The level of a protein of interest in control cells equals 1 and is marked by a solid black line. β-actin was used as a loading control. NHA serves as a non-malignant control, whereas NTERA serves as a positive control with stemness properties. Statistical analysis was performed using one way ANOVA with Dunnett’s post hoc test to NHA cells (* p < 0.05, *** p < 0.001), n = 2, mean ± SD. (E) Representative immunofluorescent staining of WG4, WG13, WG14 and WG17 cells displaying the stemness properties, toward the stemness (NESTIN, SOX2) and differentiation (GFAP, β-TUB III, MAP2) markers. White arrows indicate positive nuclear staining. Scale bar: 100 µm. The whole blots and molecular weight markers are shown in Supplementary Figure S9.

Figure 3

Characterization of stemness properties of GBM-derived sphere and adherent cell cultures. (A) Morphology of WG14 and L0125 cells in the presence of serum containing media (marked as FBS) or in the growth factors containing media (marked as sph). L0125 serves as a control cell line with stemness properties. (B) Representative immunoblots showing levels of GFAP, β-TUB III (differentiation markers) and OLIG2, SOX2, NANOG, OCT4 (stemness markers) in glioma primary cell cultures, NHA, NTERA, L0125 and L0627 cells. (C) Quantification of immunoblots. The level of a protein of interest in NHA equals 1 and is marked by a solid black line. β-ACTIN was used as a loading control. Statistical analysis was performed using a t-test comparing values in sph and FBS groups (* p < 0.05, *** p < 0.001), n ≥ 2, ±SD. (D) Expression of chosen differentiation (GFAP, TUBBIII) and stemness (OLIG2, SOX2, NESTIN) related genes in WG14 human patient-derived primary cells, L0125 and L0627 cells, cultured with FBS or with defined media (sph). The RT-qPCR data are shown as delta Ct values relative to the 18S expression. Statistical analysis was performed using a t-test (* p < 0.05, ** p < 0.01, *** p < 0.001), n ≥ 2, mean ± SD. The whole blots and molecular weight markers are shown in Supplementary Figure S10.

Figure 4

Epithelial mesenchymal transition (EMT) related genes and proteins in human primary glioma cell cultures. (A) Representative immunoblots showing the EMT markers levels in glioma primary cultures, NHA and NTERA cells. (B) Quantification of immunoblots. The level of a protein of interest in NHA cells equals 1 and is marked by a solid black line. β-ACTIN was used as a loading control. Statistical analysis was performed using one way ANOVA with Dunnett’s post hoc test comparing values obtained in tumor cells to NHA cells (* p < 0.05, ** p < 0.01,), n = 2, mean ± SD. (C) Quantification of glioma cell migration using an in vitro scratch assay. Results are presented as the percentage of scratch coverage after 18 h, n = 3, mean ± SD. The whole blots and molecular weight markers are shown in Supplementary Figure S11.

Figure 5

MGMT promoter methylation and the impact of TMZ on primary glioma cell cultures. (A) Evaluation of the MGMT promoter methylation in primary glioma cell cultures and (B) a summary table. (C) MGMT expression in primary glioma cell cultures based on RNA-seq analysis. (D) Microscopic images of WG4, WG14 and WG9 treated with TMZ (1 mM, 72 h). Scale bar: 100 µm. (E) Viability of WG4, WG14 and WG9 cells after TMZ treatment, determined by MTT metabolism test. The viability of untreated cells was set as 100% and marked with a black solid line. Statistical analysis was performed on raw data using a t-test in the comparison of treated to control groups (* p < 0.05,), n = 5, mean ± SD. (F) Representative immunoblots showing the levels of apoptotic proteins: cleaved caspase 3, cleaved caspase 7 and cleaved PARP (cl. CASP 3, cl. CASP 7, cl. PARP, respectively) in TMZ-treated WG4, WG14 and WG9 cells. (G) Quantification of immunoblots. The level of a protein of interest in control cells equals 1 and is marked by a solid black line. β-actin was used as a loading control. Statistical significance was determined by one-way ANOVA followed by Dunnett’s post hoc test in the comparison of treated to untreated cells (* p < 0.05, *** p < 0.001) or between the cell lines (## p < 0.01, ### p < 0.001), n = 3, mean ± SD. The whole blots and molecular weight markers are shown in Supplementary Figure S12.

Figure 6

Treatment with the EGFR inhibitor AG1478 modifies sensitivity of glioma cells to TMZ. (A) Representative immunoblots of phosphorylated proteins involved in EGFR signaling pathways in WG4, WG14 and WG9 cells in the presence (6 h) and absence of 10 µM AG1478 (AG). (B) The densitometric quantification. The level of a protein of interest in control cells equals 1 and is marked by a solid black line. β-ACTIN was used as a loading control. Statistical significance was determined by a t-test in comparison of treated to non-treated cells (* p < 0.05, ** p < 0.01, n ≥ 3, mean ± SD. (C) Viability of WG4, WG14 and WG9 cells after 10 µM AG alone or combined with 1 mM TMZ (AG + TMZ) for 72 h was determined with a PrestoBlue test. Viability of the control group was set as 100% and marked by a black solid line. Statistical significance was determined on raw data by one-way ANOVA followed by Dunnett’s post hoc test in the comparison of treated to untreated cells (* p < 0.05, ** p < 0.01), or by one-way ANOVA followed by uncorrected Fisher’s LSD test between the groups: AG or TMZ vs. AG + TMZ (# p < 0.05, ## p < 0.01), n = 3, mean ± SD. (D) Representative immunoblots detecting the apoptosis markers: cleaved caspase 7 and cleaved PARP (cl. CASP 7, cl. PARP, respectively), and phospho-STAT3 (pSTAT3) and STAT3 of WG4, WG14 and WG9 cells treated with 10 µM AG, 1 mM TMZ or with a combination of AG + TMZ for 72 h. (E) Quantification of immunoblots. The level of protein of interest in control cells equals 1 and is marked by a solid black line. β-ACTIN was as a loading control. Statistical significance was determined by one-way ANOVA followed by Dunnett’s post hoc test in comparison of treated to untreated control cells (* p < 0.05, ** p < 0.01, *** p < 0.001) or by one-way ANOVA followed by uncorrected Fisher’s LSD test between the groups: AG or TMZ vs. AG + TMZ (# p < 0.05, ## p < 0.01), n ≥ 3, mean ± SD. The whole blots and molecular weight markers are shown in Supplementary Figures S13 and S14.

Figure 7

AG1478-enhanced cytotoxicity exerted by DOX in primary glioma cell cultures. (A) Representative immunoblots showing the apoptosis markers: cleaved caspase 7 and cleaved PARP (cl. CASP 7, cl. PARP, respectively) in WG4, WG14 and WG9 cells treated with DOX for 48 h. (B) Densitometric quantification of effects of DOX at the high doses: 0.5 and 1 mM. The level of a protein of interest in control cells equals 1 and is marked by a solid black line. β-ACTIN was used as a loading control. Statistical significance was determined by one-way ANOVA followed by Dunnett’s post hoc test (* p < 0.05, ** p < 0.01, *** p < 0.001), n = 3, mean ± SD. (C) Cell viability of WG4, WG14 and WG9 cells after 10 µM AG, 0.5 mM DOX or combined AG + DOX treatment for 48 h, determined by PrestoBlue test. Viability of the control group was set as 100% and marked by a black solid line. Statistical significance was determined on raw data by one-way ANOVA followed by Dunnett’s post hoc test in comparison to untreated control cells (* p < 0.05, ** p < 0.01, *** p < 0.001) or by one-way ANOVA followed by uncorrected Fisher’s LSD test between the groups: AG or DOX vs. AG + DOX (## p < 0.01), n ≥ 3, mean ± SD. (D) Representative immunoblots detecting the apoptosis markers: cleaved caspase 7 and cleaved PARP (cl. CASP 7, cl. PARP, respectively) of WG4, WG14 and WG9 cells treated with 10 µM AG, 0.5 mM DOX or with a combination of AG + DOX for 48 h with (E) the densitometric quantification. The level of a protein of interest in control cells equals 1 and is marked by a solid black line. Statistical significance was determined by one-way ANOVA followed by Dunnett’s post hoc test in comparison to untreated control cells (* p < 0.05) or by one-way ANOVA followed by uncorrected Fisher’s LSD test between the groups: AG or DOX vs. AG + DOX (# p < 0.05), n = 3, mean ± SD. The whole blots and molecular weight markers are shown in Supplementary Figures S15 and S16.