Trametinib and U0126 were purchased from MedChemExpress (Monmouth Junction, NJ, United States). Dimethyl sulfoxide was used to dissolve trametinib powder to an initial concentration of 10 mM and preserved in −80°C. The CCK-8 kit was obtained from Dojindo (Rockville, MD, United States). The Annexin V-FITC/PI apoptosis detection kit and Annexin V-PE/7-ADD apoptosis detection kit were purchased from Nanjing Vazyme Biotech (Nanjing, China). Penicillin and streptomycin were obtained from Gibco (Thermo Fisher Scientific, Waltham, MA, United States). DMEM were purchased from CORNING (Corning Life Sciences, Bedford, MA, Unied States). The fetal bovine serum (FBS) was purchased from TBD (Tianjin, China). Lipofectamine 3,000 and TRIzol™ Reagent were obtained from Life Technologies (Carlsbad, CA, United States). XF96 seahorse assay plate was purchased from Agilent (Santa Clara, CA, United States). The antibodies used were as follows: anti-PKM2 antibody (#15822-AP), anti-c-myc antibody (#67447-1-lg), anti-GLUT1 antibody (#21829-1-AP), anti-LDHA antibody (#19987-1-AP), anti-β-actin antibody (#66009-1-lg), MEK1/2 polyclonal antibody (#11049-1-AP), ERK1/2 polyclonal antibody (#16443-1-AP), and anti-PCNA antibody (10205-2-AP) from Proteintech (Wuhan, China); and anti-Ki67 antibody (ab279653) from Abcam (Cambridge, United Kingdom).

The cell lines U87 and U251 (human glioma cells) were obtained from Shanghai Gene Chemistry (Shanghai, China). The cells were cytogenetically assessed using STR technology for authentication. Dulbecco’s modified Eagle medium (DMEM) containing 10% serum and 1% penicillin-streptomycin was used to culture the cells, which were placed in an incubator containing 5% carbon dioxide at 37°C. The PKM2 overexpression and c-myc-shRNA (sh-myc) plasmid was synthesized by GenePharma (Shanghai, China). The target sequence of sh-myc was 5′-GAG​AAT​GTC​AAG​AGG​CGA​ACA-3′, whereas the target sequence of the negative control (sh-vector) was 5′-TTC​TCC​GAA​CGT​GTC​ACG​T-3’. To establish stably transfected U87 and U251, the two cell lines were planted into 24-well plates, respectively. When 80% confluence occurred, cells were transfected with the vectors: PKM2 or sh-myc plasmids. Lipofectamine 3,000 Reagent (Thermo Fisher Scientific) was used for plasmid transfection. We added 0.75 μl lipofectamine 3,000 reagent and 2 μl P3000 reagent to 25 μl opti-MEM Medium according to the instructions, then added 1 μg DNA, and the mixture was added to the 24-well plate, and then incubated at 37°C with 5% CO2 for 24 h, observed the fluorescence of the cells. Suitable cells were screened by G418 and puromycin, while transfection efficiency was measured by qPCR and western blot.

The CCK-8 kit was used in accordance with the manufacturer’s instructions to detect trametinib activity against U87 and U251 cells. The two lines were seeded into 96-well plates at a density of 3×10⁵ per well. The treatment time of trametinib on cells was 6, 12, 24, 36, 48, 60, and 72 h, respectively. The CCK8 solution was mixed with DMEM at a ratio of 1:9, and used to replace the solution in the wells. The cells were then incubated at 37°C with 5% CO₂ for 2 h. Absorbance was measured at 450 nm wavelength to quantify the cell viability.

The cells were first treated with trametinib for separate times in 12-well plates, digested with trypsin, and then resuspended in PBS. To measure apoptosis, the cells were stained by annexin V-FITC and PI, under dark conditions according to the manufacturer’s instructions, and analyzed with a flow cytometer.

The HoloMonitor M4 culture system (PHIAB, Lund, Sweden) was used to detect the migration ability of glioma cells treated by trametinib. The trametinib treated cells were put into the system for 24 h, before being analyzed using the Hstudio M4 software.

The invasion ability of the cells was assessed by the transwell method. Cells (5×10⁵) were first seeded into a transwell chamber with Matrigel at the bottom (Corning Life Sciences, Bedford, MA, Unied States). Then, 100 μl serum-free DMEM was used in the transwell chamber, and 600 μl DMEM containing 20% FBS was used in a 24-well plate, while the chamber was incubated for 48 h. The lower layer was fixed with 4% formaldehyde, and then stained with Giemsa staining solution (Leagene Biotechnology, Beijing, China). Five fields were randomly selected under the microscope to count the number of invaded cells.

An XFe96 Analyzer (Agilent) was used to evaluate the level of aerobic glycolysis of cells, according to the manufacturer’s instructions. Subsequently, 50 nM trametinib was treated for 6, 12, 48, and 72 h, respectively. U87 and U251 cells were seeded into XFe96 cell culture plates at a density of 3,000 cells per well. The cells were left to evenly distribute the cells at 25°C for 1 h, and then incubated at 37°C with 5% CO₂ overnight. The probe card device was hydrated the day before the experiment, before being placed in a 37°C CO₂-free incubator overnight. On the day of the experiment, pyruvate, glutamine, and glucose were added to the basic culture medium according to the kit instructions at final concentrations of 1 mM, 2 mM, and 10 mM, respectively. The culture solution was then replaced with the test solution, and then placed in a 37°C CO₂-free incubator for 45 min. The oligomycin, glucose, and 2-deoxyglucose were added following the manufacturer’s instructions. Briefly, the oligomycin, glucose, and 2-deoxyglucose were added to the probe board port, and then the extracellular acidification rate (ECAR) of glioma cells was measured by the XFe96 Extracellular Flux Analyzer (Seahorse Bioscience, Billerica, MA, United States) in real-time.

Cells were lysed with lysis buffer mixed (1:100) with RIPA buffer (Beyotime, Shanghai, China) and PMSF (Beyotime). The cells broken by the ultrasonic disruptor were then centrifuged at 5,000 g at 4 °C for 60 min. The concentration of the supernatant was measured with the BCA kit (Beyotime). The protein sample in the SDS-PAGE gel (Beyotime) was separated with 120 V voltage in SDS-PAGE Electrophoresis Buffer with Tris-Gly. The separated protein was transferred to the PVDF membrane in the Western Transfer Buffer. The PVDF membrane was then placed in the blocking solution for 30 min, and sequentially incubated with primary and secondary antibodies. The western blot visualization was enhanced with an ECL Kit (Beyotime). The β-actin was used to calculate the relative integral density value.

TRIzol reagent (Life Technologies Corporation, Carlsbad, CA, United States) was used to lyse the cells and extract RNA from the cells. PKM2, c-myc, GLUT1, LDHA, and β-actin mRNA was reverse transcribed into cDNA by the HiScriptⅢ RT SuperMix for qPCR kit (Vazyme, Nanjing, China), stained using the ChanQ universal SYBR qPCR Master Mix kit (Vazyme), and detected with the 7,500 Fast RT-PCR System (Applied Biosystems, Waltham, MA, USA).

Crystal structures of PKM2 (DB code:3GR4) were downloaded from the RCSB database (https://www.rcsb.org/). The protein structure was determined using AutoDock 4.2 to remove the crystal water and add hydrogens as preparation before docking. The docking site is located at the junction of ATP and FBP. Discovery studio 2.5 software was used to enhance the image after docking.

The animal study was reviewed and approved by the Animal Ethical Committee of Shengjing Hospital of China Medical University (2020PS590K). Trypsin was used to digest the adherent cells cultured in DMEM containing 10% FBS. Live cells were washed and resuspended in PBS. Each 100 ml cell suspension contained 4×10⁶ cells. Cell suspension (100 μl) was injected into the skin with a syringe under the right armpit of BALB/c female nude mice. These mice weighed 14–15 g and were obtained from Beijing Huafukang Biotechnology (China). The mice (n = 5 animals per group) were randomized to six treatment groups. They were injected with either untransfected cells, vector-transfected cells, PKM2-overexpressed cells, PKM2+sh-NC cells, or PKM2+sh-myc cells. Trametinib was administered daily at a concentration of 1 mg/kg via oral gavage. The tumor volume was measured using an electronic caliper every 3–4 days and calculated by the formula: length × (width/2) × 0.5.

Paraffin tissue slices were placed in Xylene solution for 20 min, and then hydrated with anhydrous ethanol. The hydrated slices were placed in a Citrate Antigen Retrieval Solution (Beyotime). Slices were treated by endogenous peroxidase inhibitors for 30 min. ERK1/2 polyclonal antibody (diluted 1:50), PKM2 polyclonal antibody (diluted 1:100), and anti-Ki67 antibody (diluted 1:1,000) were placed on the surface of the slices, which were then placed in a 4°C refrigerator for 12 h. Thereafter, the slices were placed at 25°C for 30 min, adding Biotin labeled lamb anti-mouse/rabbit IgG polymer (MXB Biotechnologies, Fuzhou, China) for 30 min. All slices received streptomycin antibiotic protein-peroxidase (MXB Biotechnologies) for 10 min at 25°C. The Dab HorseraDish peroxidase Color Development Kit (Beyotime) was used to treat the glass sheet, before adding Hematoxylin Staining Solution (Beyotime), and observing the slides under a microscope.

All quantitative data were expressed as mean ± standard deviation (SD). These data were the results of at least three independent experiments. The images of these data were displayed using GraphPad Prism 8 software (GraphPad Software, San Diego, CA, United States). The data of the two groups were compared by Student’s t-tests (unpaired, two-tailed). The one-way analysis of variance (ANOVA) (followed by Tukey’s post-hoc tests) was used to compare the difference between the data and the control. p < 0.05 was regarded to be statistically significant: *p < 0.05; **p < 0.01.

First, we studied the effect of trametinib on the proliferation of U87 and U251 cells. The viability rate of cells was detected by CCK-8. Cells were treated with trametinib at 2, 20, 50, 100, and 200 nM, respectively, for 6, 12, 24, 36, 48, 60, and 72 h, respectively. Trametinib significantly inhibited the growth of U87 (Figure 1A) and U251 cells (Figure 1B). Interestingly, with 24 h treatment, both U87 cells and U251 cells had unknown resistance to drugs of various concentrations. At this time point, the inhibitory effect of drugs on glioma cells did not increase with the increased treatment time as expected, but decreased slightly compared with 12 h of treatment. For treatment times of 48, 60, and 72 h, significant inhibition of the two cell lines occurred by various concentrations of trametinib with significant dose dependence. These results showed that trametinib significantly inhibited the viability of U87 and U251 cells. Considering that the blood-brain barrier will physiologically reduce drug entry from plasma into brain cells, we chose 50 nM trametinib for 48 h treatment to conduct the following experiments.

To further study the role of apoptosis in trametinib toxicity to glioma cells, flow cytometry was used. The apoptosis rate of U87 and U251 cells increased obviously after treatment with 50 nM trametinib for more than 48 h (Figures 1C,D). It can be seen from the figures that the apoptosis of U87 and U251 cells induced by trametinib is mainly caused by late apoptosis. As the time of drug action increases, the cells in the UR quadrant in Figure 1C gradually increase, while the cells in the LR quadrant are almost no change. Therefore, trametinib can induce late apoptosis of glioma cells without the occurrence of early apoptosis. Simultaneously, we used Hstudio M4 software and a transwell experiment to study the effect of trametinib on cell migration and invasion. The migration and invasion ability of glioma cells significantly decreased with increased treatment time when U87 and U251 cells were treated for 6, 12, 24, 48, and 72 h, respectively (Figures 1E,F).

Aerobic glycolysis is one of the most significant signs of tumor formation, including gliomas. Therefore, we used the seahorse XF96 to detect the effect of trametinib on the glycolysis level in glioma cells. We chose treatment times of 6, 12, 48, and 72 h in U87 and U251 cells. With the increased treatment time of trametinib on glioma cells, the glycolytic activity and glycolytic reserve of U87 cells (Figure 2A) and U251 cells (Figure 2B) significantly decreased. Accordingly, western blotting and qRT-PCR showed that the glycolytic marker proteins GLUT1 and LDHA significantly decreased in transcription (Supplementary Figure S3A) and translation (Figure 2C). These results indicated that trametinib could inhibit the aerobic glycolysis level of U87 and U251 cells with increased treatment time.

To study the mechanism of trametinib-induced glioma cell death, we predicted the possible targets of trametinib with SwissTargetPrediction software and performed a KEGG pathway enrichment analysis. We found that “pathways in cancer” was enriched in the targets of trametinib (Supplementary Figure S1). Furthermore, aerobic glycolysis of tumor cells is an essential sign of tumor formation. As the key rate-limiting enzyme in aerobic glycolysis, PKM2 plays a vital role in the metabolism of cancer cells. According to our research showing trametinib inhibiting glycolysis of glioma cells, we suspect that trametinib interacts with PKM2. We obtained the protein 3D structure of PKM2 (3GR4) from the Protein Data Bank website. Then, using AutoDock 4.2, the energy minimization structures of PKM2 and trametinib were analyzed by molecular docking. We docked trametinib with the FBP binding bag (Supplementary Figure S2A) and ADP binding bag (Figure 3A), respectively, and compared the binding positions of FBP (Supplementary Figure S2B) and ADP (Figure 3B). This revealed that trametinib may have targeted a binding effect on the ADP binding bag of PKM2, thus inhibiting the activity of PKM2. Previous studies have found that ERK1/2 can promote the translocation of PKM2 from the cytoplasm to the nucleus (Yang et al., 2012a). As an inhibitor of MEK1/2, whether trametinib can affect the MAPK pathway by inhibiting the expression of MEK, thereby affecting the function of ERK1/2, reducing the translocation of PKM2 to the nucleus, thus affecting its function remained an open question. Curiously, we found that the expression of MEK1/2 and ERK1/2 in glioma cells after the treatment of trametinib decreased (Figure 2C), and using the MEK/ERK inhibitor U0126 to block ERK1/2 showed that trametinib can further reduce the expression of PKM2 (Figure 2D). To study the effect of trametinib on PKM2 in glioma, the two cell lines were treated with trametinib for 6, 12, 24, 48, and 72 h, respectively. Glioma cells treated with trametinib were detected by either western blotting or qRT-PCR. Because PKM2 can be transferred into the nucleus, to explore the influence of PKM2 on the inhibitory of trametinib, we studied the changes in protein expression of PKM2 in the cytoplasm and nucleus as a result of various trametinib treatment times. The results showed that trametinib had obvious inhibitory effect on PKM2 at both the protein level (Figure 3C) and transcription level (Figure 3D). Trametinib also significantly increased the protein expression of PKM2 in the cytoplasm, but significantly decreased the protein expression of PKM2 in the nucleus (Figure 3E). Interestingly, we also found that trametinib has obvious inhibitory effect on c-myc at both the protein level (Figure 3C) and transcription level (Supplementary Figure S3B). These results indicate that trametinib can not only affect the mRNA expression of PKM2 at the transcription level, but also affects PKM2 at the protein level through direct binding. Trametinib can thus inhibit the transfer of PKM2 protein from the cytoplasm to the nucleus. Simultaneously, trametinib also inhibited the transcription and translation of the protooncogene c-myc.

Numerous studies have shown that trametinib can significantly suppress the expression of PKM2 in U87 and U251 cells. To research the influence of PKM2 on migration, invasion, and apoptosis in glioma, we constructed stably expressed PKM2 overexpressed transfection strains for two cell lines, and evaluated the transfection efficiency by qRT-PCR (Figure 4A). Compared with the vector group, the expression of PKM2 mRNA in PKM2 overexpression group was significantly higher. Furthermore, PKM2 expression in the vector group was no different from that in the blank group. Meanwhile, the results of the Hstudio M4 software and transwell experiment showed that the migration and invasion ability of the two cell lines significantly decreased after 50 nM trametinib treatment for 48 h. However, glioma cells overexpressing PKM2 partially resisted the inhibition of migration and invasion by trametinib (Figures 4D,E). Similarly, seahorse XF96 energy level detection and flow cytometry were used to measure the glycolysis level and apoptosis rate of glioma cells. The glycolysis level of glioma cells significantly decreased, whereas the apoptosis rate significantly increased; the PKM2 overexpressed glioma cells could also resist the glycolysis level changes and apoptosis changes of trametinib treatment (Figure 4F and Figure 4C). Western blotting (Figure 4B) and qRT-PCR (Supplementary Figure S3C) were used to show that GLUT1 and LDHA expression could be significantly restored in glioma cells with PKM2 overexpression. These discoveries showed that overexpression of PKM2 can reverse the influence of trametinib on migration, invasion, glycolysis level, and apoptosis rate in glioma cells, which also indicates that PKM2 plays an essential role in the mechanism by which trametinib kills glioma.

Numerous studies showed that phosphorylated PKM2 can be transferred into the nucleus to regulate the transcription level of c-myc (Yang et al., 2012a). C-myc has also been shown to increase glycolysis by directly binding to GLUT1 and LDHA promoter regions for transcriptional regulation (Dang et al., 2009). Therefore, we wanted to explore whether trametinib can regulate the transcription level of c-myc by regulating PKM2—thus regulating GLUT1 and LDHA expression. Therefore, we further observed the changes of c-myc expression in U87 and U251 treated with trametinib. These discoveries showed that trametinib could significantly inhibit c-myc expression in U87 and U251 cell lines (Figure 3C and Supplementary Figure S3B). Further study found that overexpression of PKM2 could significantly increase the expression of c-myc with trametinib treatment. To further explore the role of c-myc in killing glioma cells and inhibiting the glycolysis level by trametinib treatment, we transfected sh-myc into the stable strains of two cell lines with PKM2 overexpression. Next, we examined the migration (Figure 5A), invasion ability (Figure 5B), and glycolysis level (Figure 5C) of these cells. We found that the migration, invasion ability, and glycolysis level of cells were reversed by sh-myc. Expression of GLUT1 and LDHA was also reversed by sh-myc at the transcription level (Figure 5E) and protein level (Figure 5D). These results indicate that trametinib can regulate the glycolysis level, migration, and invasion ability of glioma cells through the PKM2/c-myc axis.

To research the effect of trametinib on tumor growth in vivo, we established a transplanted tumor in an athymic nude mice model. U87 and U251 cell lines were planted in nude mice, and each cell line was divided into Control, Tra, Tra + vector, Tra + PKM2+sh-vector, and Tra + PKM2+sh-myc groups. In the first 2 weeks, there was no significant difference between each group. However, the results showed that the tumor size of Tra group was significantly smaller than that of control group. Compared with the Tra group, the tumor size of Tra + vector group had no obvious difference. The Tra + PKM2 group implanted with the PKM2 overexpression stable strain could partially resist the therapeutic effect of trametinib. Compared with the Tra + PKM2+sh-vector group, the Tra + PKM2+sh-myc group could reduce the antagonistic effect of PKM2 overexpression on trametinib (Figures 6A,B). The results of immunohistochemistry showed that trametinib could significantly inhibit the levels of Ki67, PKM2, and ERK in glioma cells (Figure 6C). These results indicate that the overexpression of PKM2 can alleviate the therapeutic effect of trametinib on tumors, while the knock-down of c-myc gene can restore the therapeutic effect of trametinib. In summary, these results illustrate that trametinib can inhibit the size of the glioblastoma xenograft and expression of Ki67, ERK and PKM2 in vivo—which achieve the purpose of treating glioma through the PKM2/c-myc axis.