The U87 MG (ATCC® HTB-14™) human glioblastoma cell line was obtained from the American Type Culture Collection (ATCC) and was grown in Eagle's supplemented minimum essential medium (EMEM; Lot No. 2062257), further supplemented with heat-inactivated 10 % fetal bovine serum, 1 mM L-glutamine (Gibco Life Technologies), 100 U/mL penicillin, and 100 μg/mL streptomycin (Invitrogen). Cells were cultured in a humidified incubator at 37 °C and 5 % CO₂.

A multi-drug mixture was prepared with 5-fluorouracil (5-Fu) (CAS 51-21-8, F6627, Sigma Aldrich, USA), cisplatin (CP) (CAS 15663-27-1; sc-200896, Santa Cruz, USA), and paclitaxel (PX) (CAS 33069-62-4, T7191, Sigma Aldrich, USA), which have different modes of action and are extensively used for tumor sensitization. In the first step, we determined sub-lethal doses of each chemotherapeutic by testing a range of concentrations: 1.0-1000 µM for 5-Fu, 0.78-200 µM for CP, and 0.5-500 nM for PX. We calculated the IC₅₀ and IC₂₀ doses of these drugs by probit analysis using data obtained from MTT [3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide] assays (Table 1(Tab. 1)). To create the multi-drug therapy, we used the IC₂₀ values obtained over 48 hours for 5-Fu and CP and 72 hours for PX; this combination was applied to U87 cells and yielded cell viability of 64.67±14.04 % after 72 hours of incubation. We concluded this combination dosage appropriate for establishing an in vitro toxicity model.

The chemotherapy agents were dissolved in DMSO (dimethyl sulfoxide), prepared fresh for each treatment, then diluted with molecular grade water to a final DMSO concentration of 0.5 % and used in combined application. The multi-drug mixture was initially applied to 75 cm² cell culture flasks. After every 72 hours, the cells were passaged to ensure cell proliferation and then given 24 hours to reach confluence and adhere, after which drug administration was repeated. In total, ten passages and combined treatment cycles were carried out, paralleled by control cells that received an equal concentration of vehicle. After completing the ten passages (which took approximately 40 days), the cells were then maintained for five more passages in complete EMEM medium without any drug application so as to eliminate any temporary response. Cells were harvested at the 15^th passage and the MTT test was performed with etoposide (EP) and temozolomide (TMZ); these cells were used in all other experimental assays (Figure 2(Fig. 2)).

U87 MG cells were seeded into 96-well plates (3000 cells/per well) and incubated overnight under standard cell culture conditions (see above). The cells were then treated with selected concentrations of drugs at a final vehicle concentration of 0.5 % DMSO (Merck, USA). After incubation for 24, 48, or 72 h, the cells were treated with (3-4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide (MTT, 5 mg/mL; Thermo Fisher Scientific, USA) for 4 h. Next, the cell medium was replaced with DMSO (102950 Supelco, USA) and the absorbance was read at both 492 and 570 nm using a reference wavelength of 650 nm (Thermo Multiskan GO, USA). IC₅₀ and IC₂₀ values were calculated by probit analysis using SPSS 20 software.

Multidrug-treated U87-R or untreated U87 cells (approx. 15000 per well) were seeded onto 24-well plates. After 18 h, the respective IC₅₀ doses of EP and TMZ were each administered for 6 and 24 hours. Percentages of apoptotic, necrotic, and live cells were determined by a Tali® image-based cytometer using the Tali® Apoptosis Kit (Life Technologies, USA); additionally, live-cell imaging was performed and the results analyzed according to a protocol described previously (Doğanlar et al., 2020[15], 2021[16]) using the NucBlue® Live Ready Probes® Reagent, CM-H₂DCFDA dye (Life Technologies, USA), and acridine orange-ethidium bromide (AO/EB) (Sigma Aldrich, USA).

Multidrug-treated U87-R or untreated U87 cells were seeded in 25-ml tissue culture flasks and allowed to grow until reaching 85-90 % confluence. The adhered cells were then scratched with a sterile 200-μL pipette tip and the debris washed away with EMEM. Next, we added 1 mL of serum-free EMEM and exposed the cells to IC₅₀ concentrations of EP and TMZ (or vehicle). Wound closure was determined according to the protocol established by Doğanlar and colleagues (2020[15]).

Total cellular RNA was extracted using a PureLink™ RNA Mini Kit (Thermo Scientific, USA) according to the manufacturer's protocol. The cDNA libraries were prepared using the High-Capacity cDNA Reverse Transcription Kit (Thermo Scientific, USA), and qRT-PCR was performed using Power SYBR™ Green PCR Master Mix (Life Technologies, USA) on a Quant Studio 5 real-time PCR machine (Life Technologies, USA) to determine relative fold changes in target gene expression. Primers and PCR conditions are given in Table 2(Tab. 2). Fold change was calculated using the 2^-ΔΔCt method, with β-actin and GAPDH used as multiple internal controls.

Total protein was extracted using the RIPA Lysis Buffer System (sc-24948, Santa Cruz, USA) according to the kit protocol. Protein levels were determined by Western blot assay using the NuPAGE™ 4 to 12 %, Bis-Tris, 1.0 mm Mini Protein Gel (NP0321) with the WesternBreeze™ Chemiluminescent Kit (ThermoFisher Scientific, USA) according to the method of Doganlar and colleagues (2021[16]). Seven primary antibodies were utilized: GST tag antibody (GST 3-4C), ABCC1 (NB400-156), ABCC2 antibody (NBP1-69023), ABCG2 antibody (BXP-21), LC3B antibody (NB100-2220) (Novus Biological, USA), and actin antibody (ACTN05 (C4)) (ThermoFisher Scientific). Protein expression levels were normalized to β-actin, and band intensity was measured and calculated as per our previous study (Doğanlar et al., 2021[14]).

Differences between multidrug-treated U87-R and untreated U87 groups in percentages of live, dead, and apoptotic cells, cell viability, and relative fold change in gene expression were assessed by one-way analysis of variance (ANOVA) with the Tukey HSD test or Tamhane's T2 as appropriate for the normalized test results. Additionally, the Mann-Whitney U test or independent samples t-test was used in two-group comparisons. Results having p-values of less than 0.05 were considered significant. We analyzed and plotted our results using SPSS 20 and GraphPad Prism 7.

To determine whether glioblastoma cells given prolonged exposure to sub-lethal doses of 5-Fu, cisplatin, and paclitaxel develop multi-drug resistance, we first tested the cells with drugs having structures and modes of action distinctly different from those agents. Specifically, after five additional passages without drugs to eliminate spontaneous resistance, we treated the U87 cells with increasing concentrations of EP or TMZ for 72 h and determined cell proliferation by MTT test (Figure 3(Fig. 3)). Cells subjected to the sub-lethal chemotherapeutic mix showed strong resistance to EP and TMZ; at the 72-hour mark, U87-R cells exhibited respective IC₅₀ values for EP and TMP that were approximately 4.36- and 1.98-fold greater than those of control cells. Our results thus demonstrate that repeated application of combined chemotherapy to U87 cells induces development of resistance to both of the unrelated drugs EP and TMZ.

In the second step, we focused on determining the mechanism of this resistance in the U87-R population. The main detoxification mechanisms and efflux pumps involved in MDR are ATP-binding cassette membrane transporters (ABC transporters) and glutathione S-transferase. To assess whether the observed EP and TMZ resistance originates from an MDR mechanism, we treated U87-R cells with the IC₅₀ value of EP or TMZ, and investigated at 6 and 24 h the molecular activation of MRP1/ABCC1, ABCC2, BRCP/ ABCG2, and GST (Figure 4(Fig. 4)).

Our results indicate that both EP and TMZ significantly activate MDR markers in a time-dependent manner. In resistant U87-R cells, ABC transporters and GST alike had significantly increased gene expression compared to normal U87 cells after six hours of drug administration (except ABCC2 for TMZ treatment). After 24-hour treatment, however, MDR markers were reduced in TMZ-treated U87-R cells compared to controls, while gene expression levels increased significantly in all groups treated with EP. Western blot analysis supported our gene expression results; the highest ABC transporter protein levels after TMZ treatment were observed in the U87-R group at six hours, while EP treatment resulted in especially evident increases after 24 hours.

Oxidative stress, cell cycle arrest, and deterioration in nuclear morphology are major cellular events that initiate apoptosis following treatment with TMZ or EP. To qualitatively assess the efficiency of TMZ and EP treatment on U87 and U87-R populations, we stained live cells with CM-H₂DCFDA (for oxidative stress), Nucblue, and the dual stain AO/EB (for apoptosis). In the U87 population, both TMZ and EP treatment induced strong oxidative stress, caused apoptotic body formation, and triggered the apoptosis response. In U87-R cells, however, oxidative stress, apoptotic body formation, and especially early and late apoptotic cells were all significantly reduced (Figure 5a(Fig. 5)). Next, we quantified these findings via determining the percentages of live, necrotic, and apoptotic cells using Tali image-based cytometry with Annexin V:propidium iodide (PI) staining. Following TMZ and EP treatment, we observed a 13.4-15.7 % increase in non-stained live cells among U87-R cells compared to U87 cells, along with similar reductions of 15.05-17.91 % in Annexin V-positive (apoptotic) and Annexin V:PI-positive (late apoptotic) cells, confirming reduction of early and late apoptosis in the resistant U87-R population. In the same experimental conditions, the two populations did not differ in rate of necrotic death (characterized by PI-positive cells), but necrotic responses to EP and TMZ differed, with 1.4-1.6 times higher necrotic death being observed with EP treatment compared to TMZ treatment (Figure 5b(Fig. 5)).

To further demonstrate inhibition of apoptosis in the resistant population, U87 and U87-R cells were treated with IC₅₀ doses of TMZ and EP for 6 and 24 h. As shown in Figure 5c(Fig. 5), treatment of U87-R cells with either drug resulted in induction of BCL2 and significant decrease of the BAX/BCL2 ratio. In agreement with this initial anti-apoptotic signal, the U87-R cell population exhibited significantly decreased caspase-3 gene expression, indicating a weak mitochondrial apoptosis signal. To investigate the effects of TMZ and EP treatment on the extrinsic apoptosis pathway in U87 and U87-R populations, we also measured relative gene expression of caspase-8 and death receptor 4 (DR4) and 5 (DR5), which are the main players in that pathway. While DR4 expression was not significantly different between U87 and U87-R cells, DR5 increased significantly at six hours after TMZ treatment and 24 hours after EP treatment in U87-R cells relative to U87 cells. Meanwhile, strong caspase-8 expression was observed in both groups, especially after EP treatment, with increases of 13.1-fold in U87 cells and 21.4-fold in U87-R cells. However, Tali image-based cytometer results indicated a significantly reduced proportion of apoptotic U87-R cells, while the necrotic cell population was increased (Figure 5a(Fig. 5)). These data suggest that in the context of EP treatment, caspase-8 may promote necrosis induced by the TNF-α/caspase-8 axis in both U87 and U87-R cells. Taken together, these findings support that the mitochondrial (intrinsic) apoptosis pathway is significantly less active in the resistant U87-R cell population compared to U87 cells given the same IC₅₀ doses of TMZ and EP.

To assess the autophagy signal induced by IC₅₀ doses of TMZ and EP in U87 or U87-R brain tumor cell lines, we investigated expression of autophagy pathway genes in the ULK/BECLIN1, ATG12/ATG7, and ATG3/ LC-II/autophagasome axes. We also determined regulation of the autophagy biomarker LC3-I/LC3-II using Premo Autophagy Sensor molecular probe staining and Western blot assay. According to our results, TMZ and EP treatments in U87 cells significantly decreased cell population and induced strong autophagy signals, the latter also supported by increased LC3-II protein level. However, the resistant U87-R cells clearly exhibited dense cell population and less autophagosome formation, with Western blot analysis revealing a lower level of LC3-II protein (Figure 6a(Fig. 6)). Compared to the U87 group, the resistant U87-R population also demonstrated significant increase in ULK gene expression after six hours of TMZ treatment, while BECLIN1 decreased significantly after 24 hours of EP application. Meanwhile, while resistant cells exhibited increased ATG7 and ATG12 gene expression upon short-term administration of chemotherapy agents, levels were lower than in U87 cells, especially after 24 hours. A similar situation was seen for ATG3 and also for LC3-II, which encodes the key molecule in the ATG3/LC3-II axis (Figure 6b(Fig. 6)).

To investigate the MDR-associated invasiveness of GBM using our drug-treated chemo-resistant cell model, we first measured relative expression of genes in the HIF1-α/VEGF/MMPs axis, which are linked to hypoxia and angiogenesis (Figure 7a(Fig. 7)). We found expression of this axis to be significantly higher in the resistant U87-R population compared to normal U87 cells. Overall, the highest expression values were found after 24-hour treatment with EP, with resistant cells exhibiting respective increases of 12.8-fold, 32.4-fold, 9.3-fold, and 19.6-fold for VEGF, HIF1-α, MMP3, and MMP9. Furthermore, wound healing assays revealed the U87-R population to have considerably higher rates of both cell migration and wound closure than U87 cells under the same treatment conditions (Figure 7b(Fig. 7)).

See also the Supplementary data.

Oğuzhan Doğanlar: Conceptualization, funding acquisition, project administration, data curation, formal analysis, investigation, writing - review & editing. Zeynep Banu Doğanlar: Project administration, data curation, formal analysis, investigation. Suat Erdoğan: Project administration. Emre Delen: Investigation

This study was funded by the Trakya University Scientific Research Fund (TUBAP 2017/51).

The authors declare that there is no conflict of interest. This article does not contain any studies with the use of humans and animals as study objects.