Beta-Diketone–Cobalt Complexes Sensitize Glioma Stem Cells to Temozolomide Partly Through the ERK-MGMT Pathway

Glioblastoma multiforme is characterized by high invasiveness, multidrug resistance, and inevitable recurrence, and current standard treatment regimens are not curative. Even if most glioma cells are eliminated by chemotherapy and radiotherapy, glioma stem cells can survive and differentiate into new tumor cells, thereby triggering tumor recurrence and drug resistance. Therefore, inhibiting tumor invasiveness, reversing drug resistance, and effectively ablating glioma stem cells are critical for improving the prognosis of glioblastoma multiforme. Previous studies reported that the combination of β-diketone–cobalt complexes (CoAc2) and temozolomide (TMZ) has synergistic inhibitory effects on glioma cells. Therefore, we compared cell proliferation, colony-forming capacity, cell migration, and invasion of TMZ-resistant glioma cells and corresponding glioma stem cells after treatment with CoAc2 and/or TMZ. We also explored the underlying mechanism by which CoAc2 sensitizes cells to TMZ through transcriptome sequencing and related signal pathway blockade. We found that CoAc2 signicantly increased the inhibitory effect of TMZ on the proliferation, colony formation, migration, invasion, and survival of drug-resistant stem cells. By downregulating ERK pathway activity, CoAc2 inhibited the expression of O6-methylguanine-DNA methyltransferase and eventually sensitized drug-resistant glioma cells to TMZ. In conclusion, the combined use of CoAc2 and TMZ can reverse TMZ resistance and signicantly enhance its inhibitory effect on the malignant phenotype of glioma cells and glioma stem cells.


Introduction
Glioma cells were seeded into a six-well plate at a density of 1 × 10 4 cells per well, and the corresponding concentration of drugs was added after 24 h. The medium was changed every 2 days. After 10 days of culture, the cells were xed with 75% alcohol, stained with crystal violet, dried, and counted.

Protein extraction and Western blot assays
Cells were seeded into a six-well plate and treated with the drugs for 24 h, followed by rinsing with icecold PBS solution. The cells were lysed with RIPA lysis buffer (Beyotime, China). The cell supernatant was collected by centrifugation at 12,000 × g for 5 min. SDS-PAGE sample loading buffer was added, followed by boiling at 100°C for 5 min. After SDS-PAGE electrophoresis, the total protein was transferred to a PVDF membrane (Millipore, USA), blocked with 5% skim milk for 2 h, and incubated using the primary antibodies overnight at 4°C. After three rinses in TBST, the membrane was incubated with secondary antibodies for 2 h and nally developed using ECL. ImageJ software was used to analyze and quantify protein expression. The primary antibodies included phospho-p44/42 MAPK (Erk1/2) (Thr202/Tyr204) antibody, p44/42 MAPK (Erk1/2) (L34F12) mouse monoclonal antibody (mAb), MGMT (E6M7V) rabbit mAb (CST, USA), and GAPDH (0411) mouse mAb (Santa Cruz, USA). The secondary antibodies were antirabbit IgG (HRP-linked) antibody and anti-mouse IgG (HRP-linked) antibody (CST).
Transwell migration and Matrigel® invasion assays U251 TMZ cells were seeded in the upper compartment of 24-well Transwell plates at a density of 3 × 10 3 cells/well without FBS. For the invasion assay, 100 µL Matrigel® diluted at 1:3 in DMEM was placed in each well of the Transwell and incubated at 37°C for approximately 30 min for solidi cation before plating transfected cells. The lower compartment was lled with 600 µL DMEM supplemented with 10% FBS. After 48 h of incubation at 37°C in the presence of 5% CO 2 , the cells that did not migrate or invade and stayed on the upper surface of the lter were obliterated using a sterile cotton swab. The cells that migrated and invaded through the membrane into the bottom chamber were xed with 75% ethanol for 30 min and stained with crystal violet. Images captured were used to count cells in randomly selected six elds per well using a microscope (DMI3000B, LEICA). The experiment was repeated three times.
RNA library construction and sequencing U251 TMZ cells were treated with 20 µg/mL CoAc2 or the same volume of DMSO as a control. The total RNA of the cells was extracted using TRIzol after 24 h. The samples were entrusted to GENEWIZ China & Suzhou Lab for subsequent RNA sample quality testing, ribosome removal, library construction, library puri cation, library detection, library quanti cation, and sequencing cluster generation. The samples were nally sequenced and analyzed on the Illumina HiSeq X Ten platform.

Statistical analysis
All experimental data were obtained through three independent experiments and presented as the mean ± SD. Student's t-test was used to calculate the statistical signi cance of the experimental results. p < 0.05 denoted statistical signi cance.

Effects of TMZ combined with CoAc2 on the migration of U251 TMZ cells
We further determine the effect of the combined use of TMZ and CoAc2 on the aggressive behaviors of U251 TMZ cells by evaluating migration and invasion. Transwell tumor cell migration and invasion assays illustrated that TMZ had weak inhibitory effects on U251 TMZ cell migration and invasion (Fig. 3A, B).
Compared with the control group ndings, the migration and invasion rates of U251 TMZ cells after treatment with TMZ were 90.65% ± 8.92% and 94.47% ± 6.22%, respectively (Fig. 3C, D). Contrarily, CoAc2 strongly inhibited the migration and invasion of U251 TMZ cells (Fig. 3A, B). Compared with the control group ndings, the migration and invasion rates of U251 TMZ cells after treatment with CoAc2 were 65.14% ± 7.19% and 40.46% ± 5.05% (Fig. 3C, D). Compared with the control group data, the migration and invasion rates of U251 TMZ cells following treatment with both drugs were 33.54% ± 5.21% and 7.87%

Reduction of U251 TMZ GSCs induced by TMZ and CoAc2
Aggressiveness and high malignancy are highly correlated with GSCs. Therefore, we further tested whether the combination of TMZ and CoAc2 has an inhibitory effect on GSCs. First, U251 TMZ cells were induced via sphere-formation culture to form spheroid bodies, and immuno uorescence detection revealed that spheroid bodies expressed Nestin, a marker of neural stem cells, indicating that U251 TMZ cells were induced to form neurospheres (Fig. 4A). Considering the self-renewal and differentiation capabilities of tumor stem cells, we further tested the passage and differentiation capabilities of the neurospheres. U251 TMZ neurospheres cultured in an ultralow attachment plate could be passaged, and the newly formed neurospheres had no obvious morphological changes (Fig. 4B). Subsequently, U251 TMZ neurospheres were digested and placed in a serum-containing medium for adherent culture to induce cellular differentiation. These neurospheres could differentiate into mature glioma cells (Fig. 4B). These results indicate that the neurospheres that form after induction of U251 TMZ cells have stemness-related properties.
Subsequently, the neurospheres that formed after the 7-day induction of U251 TMZ cells were treated with TMZ (50 µg/mL), CoAc2 (20 µg/mL), or both for 2 weeks. The size and number of neurospheres in each group were measured and counted. We then found that TMZ or CoAc2 alone had certain inhibitory effects on the number of neurospheres, and the combination had signi cantly greater ability to inhibit the number and size of neurospheres than each drug used alone (Fig. 4C, D). Further, the cell counting kit-8 assay revealed that the U251 TMZ GSC survival rate after combined treatment with TMZ and CoAc2 was 46.22% ± 3.27%, which was signi cantly lower than that after TMZ or CoAc2treatment alone (98.35% ± 3.52% and 78.06% ± 5.71%, respectively).

Effect of CoAc2 on gene expression in U251 TMZ cells according to RNA-seq
To explore the molecular mechanism by which CoAc2 can reverse TMZ resistance, we performed RNA sequencing in U251TMZ cells treated with CoAc2. First, Gene Ontology (GO) was used to analyze the main biological functions of differential genes of U251 TMZ cells after treatment with CoAc2. The enriched GO terms included binding, cell part, and cellular process (Fig. 5A). In addition, Kyoto Encyclopedia of Genes and Genomes functional enrichment analysis was performed on differentially expressed genes.
The results identi ed MAPK signaling pathway genes as the most signi cant differentially expressed genes, and the enrichment degree was relatively large (Fig. 5B). Furthermore, differentially expressed proteins in U251 TMZ cellswere detected after treatment with TMZ (50 µg/mL), CoAc2 (20 µg/mL), or both. We found that TMZ alone had no effect on MGMT expression, whereas CoAc2 alone signi cantly downregulated phospho-Erk1/2 and MGMT expression. Both phospho-Erk1/2 and MGMT were signi cantly downregulated after combined treatment with TMZ and CoAc2.

Inhibitory effects of TMZ combined with mitogen-activated protein kinase (MEK) inhibitors on TMZ-resistant glioma cells
To verify the role of the ERK pathway in TMZ resistance, we tested the effects of the MEK inhibitorPD98059 (25 µM) and TMZ (50 µg/mL) on drug-resistant cells. The addition of CoAc2 or PD98059 to TMZ resulted in signi cant downregulation of phospho-Erk1/2 and MGMT (Fig. 6A). The migration and invasion rates of U251 TMZ cells after combined treatment with PD98059 and TMZ were 48.14% ± 2.90% and 18.38% ± 2.24%, respectively, compared with 36.69% ± 3.99% and 10.98% ± 2.45%, respectively, after treatment with CoAc2 and TMZ ( Fig. 6B-E).
Subsequently, the neurospheres that formed after 7-day induction of U251 TMZ cells were treated with TMZ (50 µg/mL) combined with PD98059 (25 µM) or CoAc2 (20 µg/mL) for 2 weeks. The size and number of neurospheres in each group were measured and counted. We found that both combinations had signi cant inhibitory effects on the number and size of neurospheres (Fig. 7A, B). CCK-8 analysis of the cells revealed that the survival rates of U251 TMZ GSCs after PD98059/TMZ and CoAc2/TMZ treatment were 56.72% ± 4.35% and 46.55% ± 2.96%, respectively (Fig. 7C).

Discussion
GBM is the most common and aggressive tumor of the central nervous system,and it carries the worst prognosis [25]. TMZ is a rst-line chemotherapy for glioma, but most patients with glioma will develop resistance [26]. Studies have revealed that CDC2 expression is negatively correlated with the TMZ sensitivity of gliomas. High CDC2 expression in patients with glioma portends a poor prognosis and short overall survival [27]. In addition, LncRNA-XIST reduces the sensitivity of gliomas to TMZ by inhibiting miR-29c and upregulating SP1 and MGMT. Patients with high LncRNA-XIST expression have short overall survival [28]. Therefore, increasing the sensitivity of glioma to TMZ may be an effective strategy to improve the overall survival of patients.
Moreover, the mechanism by which the sensitivity of glioma to TMZ is enhanced can also be used to reverse TMZ resistance in this malignancy. Our previous study found that CoAc2 can increase the sensitivity of human glioma cells to TMZ, and moreover, there is a synergistic effect between CoAc2 and TMZ. Therefore, we attempted to verify whether CoAc2 can reverse TMZ resistance in glioma. In this study, we used the method of concentration doubling to generate the TMZ-resistant cell line U251 TMZ and con rmed that these cells were resistant to TMZ but not CoAc2. Further examination demonstrated that CoAc2 both reversed resistance to TMZ and synergized its cytotoxic effects in both TMZ-sensitive and TMZ-resistant glioma cells.
GBM cannot be completely removed via surgical resection because of its high invasiveness; it is also prone to develop resistance to chemotherapy because of tumor heterogeneity and tumor stem cells [29]. Extensive evidence indicates that tumor stem cells are related to chemotherapy resistance and tumor maintenance and recurrence; therefore, we further demonstrated that the combined use of TMZ and CoAc2 signi cantly inhibited the proliferation of GSCs. In addition, the combination regimen signi cantly inhibited colony formation, migration, and invasion by U251 TMZ cells.
In this study, RNA-seq was performed on U251 TMZ cells treated with CoAc2. Based on previous ndings, we focused on the ERK pathway and found that CoAc2 decreased Erk1/2 phosphorylation and inhibited the expression of the TMZ resistance gene MGMT. Some studies have reported that MEK inhibition activates p53 that downregulates MGMT expression [30][31][32]. The methyl group of the O6 position of guanine added by TMZ is removed by MGMT, which prevents cell death by mismatch repair (MMR) [9,33,34]. Elevating ERK activation increases protein levels of stemness markers in GSC, and inhibiting ERK activation impairs sphere formation and attenuates the expression of genes associated with neural cancer stemness [35,36]. Our results suggested that CoAc2 might reverse TMZ resistance by inhibiting the activation of the ERK pathway. Thus, we used the MEK inhibitor PD98059 [37] in combination with TMZ and found that inhibiting ERK pathway activation could downregulate the expression of MGMT.
This regimen also inhibited the migration and invasion of U251 TMZ cells similarly as TMZ plus CoAc2.
However, the nding that the combination of CoAc2 and TMZ more strongly inhibited the proliferation of GSCs than TMZ plus PD98059 indicated that CoAc2 reverses TMZ resistance through multiple effects including inhibition of the ERK pathway.

Figure 7
PD98059 combined with TMZ signi cantly inhibits the proliferation of U251TMZ GSCs. (C) U251TMZ cells (5000 cells/mL) were cultured in CSC medium for 7 days to permit neurosphere formation.