Introduction

Glioblastoma (GBM) is highly malignant and grows rapidly, and existing treatments are ineffective1. Despite the continuous improvement of comprehensive treatments such as surgery, radiotherapy and chemotherapy, the survival of GBM has only increased by a few months2,3. Therefore, it is crucial to understand the mechanism of GBM malignant development, which will help understand the disease and develop treatments.

Mesenchymal stem cells are a type of cell with the ability to repair and differentiate, and cancer-associated mesenchymal stem cells have similar functions to mesenchymal stem cells in normal tissues and have the ability to differentiate into corresponding tumor cells4,5. When MSCs are recruited into the tumor microenvironment of GBM, they are transformed into glioma associated mesenchymal stem cells (GAMSC) by cellular crosstalk6. GAMSC are a unique subpopulation of tumor cells with stem-like characteristics that can initiate and maintain tumor growth, and most of them are resistant to anticancer treatment, which is considered to be one of the important reasons for the malignancy and development of GBM7,8,9,10. Although the effect of GAMSC on the malignancy of GBM has been reported, its mode of action is still unclear and needs further exploration.

Intercellular communication is a crucial mechanism of cell-to-cell interaction and a vital connection in the communication between cancer and its surrounding microenvironment11. There are many mechanisms for intercellular communication, including neurotransmitters, cytokines/chemokines, gap junctions, tunneling nanotubes, and exosomes12. Exosomes are one of the most important subtypes of extracellular vesicles involved in intercellular communication. They can not only carry their own unique biomacromolecules, but also carry bioactive molecules, including cytokines, RNA, etc.12,13. Therefore, we consider that GAMSCs may promote the malignant progression of GBM through some means of intercellular communication.

Abnormal cellular metabolism is a prominent feature of tumors and is considered a hallmark of tumorigenesis14. Warburg effect is the main way tumor cells produce energy, and its essence is the process of converting mitochondrial oxidative phosphorylation of glucose into aerobic glycolysis15. It is worth noting that abnormal energy metabolism has been found in GBM16. Pyruvate dehydrogenase A1 (PDHA1), a hub for glycolysis, is involved in cancer metabolism. Previous reports have shown that its inhibition can increase the malignancy of cancer cells17,18,19. But its role and mechanism in GBM remain unclear. Therefore, we decided to explore whether GAMSCs promote glycolysis in GBM through some means of intercellular communication.

In this study, we revealed that GAMSC-derived exosomes can promote the proliferation, migration, and invasion of GBM cells. The miR-21-5p molecules carried in GAMSC-EXO can directly act on PDHA1 to promote glycolysis of GBM.

Methods and materials

Cell culture

U87 and U251 cells that had been tested for contamination were obtained from the Pricella Co, Ltd (Wuhan, China). GAMSC were obtained and established according to our previous protocols5,20 and cultivated by mesenchymal Stem Cell Culture Medium (Cyagen, CA, USA). GAMSC was transfected using Lipofectamine2000 (Invitrogen, MA, USA) with miR-21-5p mimics, inhibitor, or negative controls (NC) (RiboBio, Guangzhou, China). GBM cells were transfected using Lipofectamine 2000 with siPDHA1 or control (RiboBio, Guangzhou, China). At 37 °C with 5% CO2, cells were cultured in DMEM medium (HyClone, Logan, Utah, USA). Penicillin–streptomycin (1%) and fetal bovine serum (10%) (Gibco, Grand Island, USA) were added to the medium.

Identification of GAMSC

GAMSC suspensions were incubated with antibodies (CD44, CD73, CD90, and negative control cocktail) of Human MSC Analysis Kit (BD, NJ, USA) for cell surface marker characterization by flow cytometry (BD LSRFortessa™ X-20, NJ, USA). The multilineage differentiative capacity of GAMSC was assessed with Alizarin Red S, Oil Red O, or Alcian Blue staining after culture using Osteogenic differentiation complete medium, lipogenic differentiation induction medium, or chondrogenic differentiation medium (Cyagen, CA, USA), separately.

Luciferase reporter gene assay

PDHA1 sequences carrying the miR-21-5p binding 3′UTR site and its mutant counterpart were constructed by PCR and subsequently cloned into pmirGLO Dual-Luciferase Vector. PDHA1 constructs or oligonucleotides of miR-21-5p mimics or inhibitor were cotransfected into GBM cells with 70% by Lipo-3000 (Thermo, MA, USA) and cultured for 24 h. Then, the dual luciferase assay (Promega, WI, USA) was used to measure luciferase activity in 96-well plates.

Isolation and characteristic assays of exosomes

Before conditioned medium collection, the basal medium was treated with GAMSC that was 80% confluent for 24 h. Exosomes were extracted from the conditioned media by different centrifugation procedures. The supernatant was collected after centrifugation at 2,000 g for 35 min and then 12,000 g for 50 min, filtrated with a 0.45-μm filtering membrane, and ultracentrifuged twice at 110,000 g for 80 min. Finally, a PBS resuspension of exosome sediment was used for further assays.

The concentrations of exosomes were determined with a protein assay kit based on the BCA method. The levels of the exosome surface markers TSG101; CD81 and calnexin were measured using western blot as described below. For the transmission electron microscopy (TEM) assays, 5–6 μl exosomes were dropped onto a copper grid, stained with 2% phosphotungstic acid for 1 min, dried naturally, and then observed using an 80 kV transmission electron microscope (FEI, MA, USA). For the cell uptake assay, purified exosomes were labelled using PKH26 (Sigma-Aldrich, MA, USA) according to the manufacturer’s protocol. Then, GBM cells were incubated with labelled exosomes for 8 h at 37 °C protected from light. After fixation and staining with Hoechst 33,342 (Beyotime, Jiangsu, China) for 15 min, the uptake level of labelled exosomes by GBM cells was observed using a confocal microscope (Nikon, NY, USA).

Establishment of mice models

All animal procedures were performed under the guide for the care and use of laboratory animals and the approval (IACUC: S2448) of the Centre of Experimental Animal Tongji Medical College of Huazhong University of Science and Technology (Wuhan, China), and we confirmed that all methods are reported in accordance with ARRIVE guidelines. Specific-pathogen-free male nude mice aged about 6wk and weighed around 20 g were purchased from Vitalriver (China). Using a stereotaxic instrument, 2 μl of GBM cell suspension (about 1 × 105 cells pretreated with or without GAMSC-exo (around 1–2 × 109 exosomes/mL)) was injected into the nude mice at a depth of about 2 mm, about 2.5 mm in front of the anterior bregma. The neurological symptoms, tumor growth, and survival time of the mice were recorded regularly. After surgery, all animals were supplied with antibiotics and allowed unrestricted food, water, and activity. The injections were conducted on a week after surgery. The mice were euthanized by cervical dislocation following inhaled isoflurane, and the GBM samples were harvested. Immunohistochemistry and immunofluorescence analysis was performed using PKM2 antibody (Proteintech, USA, 15822-1-AP) to determine the expression level.

Statistical analyses

All statistical analyses were conducted utilizing GraphPad Prism version 9.0 (La Jolla, CA, USA) and SPSS version 22.0 statistical softwares (Armonk, NY, USA). The experimental data are expressed as the mean values ± standard deviations (SDs). Normality of the data was assessed through the use of histograms, the Shapiro–Wilk test, and the Kolmogorov–Smirnov test. For comparisons involving two or more groups of normally distributed variables, one-way ANOVA-Turkey test and unpaired Student’s t test were utilized respectively. Kruskal–Wallis test was applied for the analysis of three groups with non-normal distribution parameters. Survival analysis was performed using Kaplan‒Meier methods with the log-rank test. All in vitro experiments and assays were repeated at least three times. P value < 0.05 was considered statistically significant.

Results

Identification of Glioma associated mesenchymal stem cells (GAMSC)

GAMSCs were extracted and cultured using our previous method5,20 and identified when the culture distribution was approximately 80–90% (Fig. 1A). Flow cytometry results showed that these cells were positive for the MSC markers CD73(99.3%), CD90(93.9%), and CD44 (91.6%), and were negative for the hematopoietic stem cell markers CD45, CD34, and HLA-DR (5.2%, mixed as a negative cocktail) (Fig. 1B). Alizarin Red S, Oil Red O, and Alcian Blue staining showed that the isolated cells were induced to undergo osteogenic, adipogenic, and chondrogenic differentiation (Fig. 1C). Together, the above results show that we have successfully extracted GAMSC.

Fig. 1
figure 1

Identification of Glioma associated mesenchymal stem cells (GAMSC). (A) Micrographs of human GAMSC. Upper bar = 20 μm, Lower bar = 100 μm. (B) Flow cytometry graphs of the surface markers CD73, CD90, and CD44 and the negative cocktail of GAMSC. (C) Representative graph of Alizarin Red S, Oil Red O, or Alcian Blue staining assays of GAMSC lipogenic, osteogenic, and chondrogenic differentiation ability, (Bar = 50 μm).

Full size image

GAMSC-derived conditioned medium promotes GBM cell proliferation, migration, and invasion

It has been reported that stem cells regulate recipient cells’ biological behavior through cell interactions21,22. To further explore the influence of cell interactions on the tumor-related phenotypes, we harvested the conditioned medium (CM) of GAMSCs and treated GBM cells with CM in clone formation assays, MTT assays, Wound healing assay and Transwell assays. The clone formation assay findings revealed that the rate of cell proliferation in the CM group was markedly increased in comparison to the basal medium (BM) group (Fig. 2A). The MTT results showed that from 0 to 72 h, the value of CM group at 490 nm is significantly higher than that of CM group (Fig. 2B). The results of Transwell assay indicated that CM could promote GBM cell migration and invasion (Fig. 2C and D). In addition, the wound healing assay also verified that CM from GAMSC could promote GBM cell migration (Fig. 2E). In conclusion, the above results suggested that there is a substance in CM that promotes the migration, invasion and proliferation of GBM cells.

Fig. 2
figure 2

GAMSC-derived conditioned medium promotes GBM cell proliferation, migration, and invasion. (A)The colony formation images and quantification of GBM cells with BM or CM. (B) MTT assay results of GBM cells with BM or CM. *P < 0.05 vs. BM. Transwell results of migration (C) and invasion (D) ability of GBM cells. *P < 0.05 vs. BM. (bar = 200 μm). (E) Wound healing of GBM cells at 0 h, 24 h (with BM or CM) (bar = 200 μm).

Full size image

GAMSC-EXO promotes GBM cell proliferation, migration and invasion

We considered that exosomes might be the substance in CM from GAMSC that could promote the tumor phenotype of GBM cells. Therefore, we extracted and characterized exosomes in GAMSC-derived CM. The nanoparticle size profile and the transmission electron microscopy showed that the isolated exosomes were round or elliptical membranous vesicles with clear margins (Fig. 3A). The western blot results showed that the exosomes expressed the marker proteins TSG101 and CD81 and lacked the endoplasmic reticulum protein calnexin (Fig. 3B). Then, we labelled GAMSC-EXO with PKH26, a red fluorescent lipidic marker. The confocal laser microscopy results showed that after 8 h of incubation, PKH26-labeled GAMSC-EXO were localized in the cytoplasm of GBM cells (Fig. 3C). The clone formation assay results indicated that the quantity of cell proliferation in the BM + EXO group was markedly increased in comparison to the BM group (Fig. 3D). The MTT results showed that from 0 to 72 h, the value of BM + EXO group at 490 nm is significantly higher than that of CM group (Fig. 3E). The results of Transwell assay indicated that EXO could promote GBM cell migration and invasion (Fig. 3F and G). The results revealed that GAMSC-EXO could promotes GBM cell migration, invasion and proliferation.

Fig. 3
figure 3

GAMSC-EXO promotes GBM cell proliferation, migration and invasion. (A) Electron microscopy micrographs of GAMSC-EXO (Bar = 100 nm). (B) Western blot results for the exosomal surface markers TSG101, CD81, and calnexin. (original blots data are presented in supplementary file) (C) Confocal microscopy images of GAMSC-EXO uptake by GBM cells. GAMSC-EXO were labelled with PKH26 (red) and the nucleus of GBM cells were labelled with Hoechst 33,342 (blue). (Bar = 10 μm). (D) The colony formation images and quantification of GBM cells with BM or BM + GAMSC-EXO. (E) MTT assay results of GBM cells with BM or BM + GAMSC-EXO. *P < 0.05 vs. BM. Transwell results of migration (F) and invasion (G) ability of GBM cells. *P < 0.05 vs. BM (bar = 200 μm).

Full size image

GAMSC-EXO promotes glycolysis in GBM cells

Glycolysis was the main way to produce ATP in tumor cells, also known as the Warburg effect15. In order to explore the effect of GAMSC-EXO on glycolysis, the glycolysis-specific indicator kinase isozyme type M2 (PKM2) was detected by western blot and qPCR. The results showed that GAMSC-EXO could promote the expression of PKM2 at both protein and gene levels (Fig. 4A and B). In addition, we examined the effects of GAMSC-EXO on glucose utilization, lactate production, and intracellular ATP levels in GBM cells. As the results shown in Fig. 4C–E, GAMSC-EXO could promote the glucose utilization, lactate production and intracellular ATP levels in GBM cells. We evaluated the effect of GAMSC-EXO on mitochondrial oxidative respiration via measuring oxygen consumption rate (OCR) in GBM cells. As shown in Fig. 4F, the OCR of the mitochondrial oxidative respiration phase in the GAMSC-EXO group was increased in comparison to the NC group. The basal respiration and maximum respiration of the GAMSC-EXO group were markedly enhanced compared to the NC group (Fig. 4G). Besides, extracellular acidification rate assay (ECAR) was used to further exam the status of overall glycolytic flux in GBM cells. As showed in Fig. 4H and I, the results were similar with the OCR results. Moreover, we introduced the glycolysis inhibitor 2-DG into the rescue experiments. The results of colony formation assay in Figure s1A showed that GAMSC-EXO can promote GBM cell proliferation, but this effect can be partially inhibited by 2-DG. The MTT assay results indicated that 2-DG can inhibit the proliferative effect of GAMSC-EXO on GBM cells (Figure s1B). Also, the Transwell results revealed that the influence of GAMSC-EXO on enhancing the invasion of GBM cells could be inhibited by 2-DG (Figure s1C). The above results demonstrated that GAMSC-EXO was able to promote the invasion and proliferation of GBM cells by enhancing glycolysis. In order to analyze the regulatory effect of GAMSC-EXO on the glycolytic capacity of GBM cells, we cultured GBM cells with D-glucose-13C and detected the role of 13C in the glycolysis process by LC/MS. In Figure s3 B and C, the glucose-6-phosphate (m + 0 and m + 6) and acetyl-CoA (m + 0 and m + 3) measurements showed that cells in the GAMSC-EXO group exhibited enhanced glycolytic flux. In addition, the detected intermediates of glycolytic flux were also increased in the GAMSC-EXO group in comparison to the NC group (Figure s3 D-G). Together, these results suggested that GAMSC-EXO enhanced glycolysis in GBM cells.

Fig. 4
figure 4

GAMSC-EXO promotes glycolysis in GBM cells. The PKM2 expression in GBM cells (with GAMSC-EXO or not) was evaluated by western blot (original blots data are presented in supplementary file) (A) and RT-PCR (B) *P < 0.05 vs. NC. Statistical analysis of glucose utilization (C), lactate production rate (D), and intracellular ATP level (E) in the NC and GAMSC-EXO group, *P < 0.05 vs. NC. Oxygen consumption rate (F), basal respiration and maximal respiration (G) of GBM cells in the NC and GAMSC-EXO group, *P < 0.05 vs. NC. Extracellular acidification rate (H), glycolysis and glycolytic capacity (I) in the NC and GAMSC-EXO group, *P < 0.05 vs. NC.

Full size image

GAMSC-EXO can promote GBM growth and glycolysis in vivo

We further observed the effects of GAMSC-EXO on GBM in in vivo experiments (Fig. 5A). After transplantation GBM cells transfected with fluorescence, we detected the fluorescence intensity of tumors through in vivo fluorescence imaging of mice. As displayed in Fig. 5B and C, the fluorescence intensity of the GAMSC-EXO group was markedly higher than that of the NC (three replicate samples in each group). Besides, survival analysis (Kaplan–Meier) in Fig. 5D showed that the survival time of the GAMSC-EXO group was significantly shorter in comparison to NC group. The above results demonstrated that GAMSC-EXO could promote the growth of GBM and accelerated host death. In addition, IHC and IF analysis were performed on the obtained tissue sections, and the results displayed that the expression of PKM2 in the GAMSC-EXO group was increased in comparison to the NC (Fig. 5E and F), indicating that GAMSC-EXO could enhance the expression of PKM2 in vivo.

Fig. 5
figure 5

GAMSC-EXO can promote GBM growth and glycolysis in vivo. (A) Photo of animal model. (B) In vivo fluorescence imaging of mice (Blank: injected saline, NC: luciferin GBM cells, GAMSC-EXO group: luciferin GBM cells pretreated with GAMSC-EXO). (C) Biofluorescence quantitative analysis of GBM, *P < 0.05 vs. NC. (D)Survival analysis (Kaplan–Meier) curve graph (P < 0.05). IHC (E) and IF (F) images of PKM2-stained tissue microarray (TMA) in the NC and GAMSC-EXO group, scale bars are displayed in the figure.

Full size image

Knockdown of PDHA1 can promote glycolysis in GBM cells

In order to further clarify whether PDHA1 was involved in the regulation of glycolysis in GBM cells by GAMSC-EXO. we evaluated the PDHA1 expression in GBM cells by western blot and qPCR. The results showed that GAMSC-EXO treatment significantly decreased the PDHA1 expression level in GBM cells (Fig. 6A and B). Then, we knockdown the expression of PDHA1 in GBM cells by small interfering RNA (siRNA) and performed further experiments (Fig. 6C and D). First, we examined the effects of PDHA1 on glucose utilization, lactate production and intracellular ATP levels in GBM cells. As shown in Fig. 6E–G, compared with the NC group, the glucose utilization rate, lactate production rate and ATP level in cells were significantly increased in the PDHA1 knockdown group (siPDHA1). Then, the results of OCR suggested that the OCR of the mitochondrial oxidative respiration phase in siPDHA1 group was significantly increased, and basal respiration and maximum respiration were also enhanced in comparison to the NC (Fig. 6H and I). In addition, ECAR was used to further test the status of overall glycolytic flux in GBM cells. As displayed in Fig. 6J and K, the results of ECAR were similar with the OCR results.

Fig. 6
figure 6

Knockdown of PDHA1 can promote glycolysis in GBM cells. The PDHA1 expression in GBM cells (with GAMSC-EXO or not) was assessed by western blot (original blots data are presented in supplementary file) (A) and RT-PCR (B) *P < 0.05 vs. NC. The PDHA1expression in GBM cells (with siNC or siPDHA1 or not) was evaluated by western blot (original blots data are presented in supplementary file) (C) and RT-PCR (D) *P < 0.05 vs. siNC. ns: no sense. Statistical analysis of glucose utilization (E), lactate production rate (F), and intracellular ATP level (G) in the NC and siPDHA1 group, *P < 0.05 vs. NC. Oxygen consumption rate (H), basal respiration and maximal respiration analysis (I) of GBM cells in the NC and siPDHA1 group, *P < 0.05 vs. NC. Extracellular acidification rate (J), glycolysis and glycolytic capacity analysis (K) in the NC and siPDHA1 group, *P < 0.05 vs. NC.

Full size image

Identifying miR-21-5p from GAMSC-EXO as a direct repressor of PDHA1

To evaluate potential GAMSC-EXO miRNAs that modulate PDHA1 expression, we first screened the top 20 miRNAs with the highest abundance in GAMSC-EXO from GSE206675 (Fig. 7A). Then we conducted a Venn diagram of potential upstream miRNAs that regulate PDHA1 expression based on three miRNA databases, including mirWalk, miRDB, and Tarbase. As shown in Fig. 7B, miR-21-5p was identified as a direct repressor of PDHA1 in GAMSC-EXO. To further demonstrate the direct regulation of miR-21-5p on PDHA1, we constructed two recombinant plasmids of PDHA1 3’UTR sequences: the wild type (WT) containing binding sites of miR-21-5p, the mutant forms with mutation of the binding sites (MUT) (Fig. 7C). Then, dual-luciferase reporter constructs carrying WT or MUT PDHA1 3’UTR sequences were separately co-transfected into GBM cells with miR-21-5p mimics, inhibitor, or the respective controls. The results showed that compared with the respective control groups, the reporter luciferase activities in the miR-21-5p mimics or inhibitor groups were significantly reduced or increased in GBM cells with WT PDHA1 reporters, respectively, while the reporter luciferase activities were not different among groups in GBM cells with MUT PDHA1 reporters (Fig. 7D). This result indicated that miR-21-5p could directly bind the 3’UTR sequence of PDHA1 and negatively regulate its mRNA levels in GBM cells. In addition, we transfected GAMSC with miR-21-5p mimics/inhibitors and its negative control miR-21-5p mimics/inhibitors-NC, extracted exosomes from the conditioned medium of GAMSC in each group, and detected the transfection efficiency by RT-PCR. The results showed in Fig. 7E that compared with the control group, miR-21-5p mimics could upregulate the expression level of miR-21-5p in GAMSC-EXO, and miR-221 inhibitors could inhibit the expression level of miR-21-5p in GAMSC-EXO. Moreover, we treated GBM cells with GAMSC-EXO mimics/inhibitor and detected the expression level of PDHA1 in GBM cells by western blot. The western blot results suggested that GAMSC-EXO miR-21-5p could inhibit the PDHA1 expression in GBM cells (Fig. 7F).

Fig. 7
figure 7

Identifying miR-21-5p from GAMSC-EXO as a direct repressor of PDHA1. (A) Top 20 abundant miRNAs in GAMSC-EXO from GSE206675. (B) Venn diagram analysis of the upstream miRNAs that regulate PDHA1 expression, indicating the overlap according to the top 20 abundant miRNAs in GAMSC-EXO and three miRNA (C) The potential binding sites of miR-21-5p to the 3′-UTR of PDHA1 and the mutation sites of MUT. (D) The effect of miR-21-5P on PDHA1 expression in GBM cells. (E) RT-PCR results: Transfection efficiency of miR-221 mimics and inhibitors. (F) The expression of PDHA1 in GBM cells was evaluated by western blot (original blots data are presented in supplementary file). *P < 0.05 vs. NC.

Full size image

GAMSC-EXO miR-21-5p promotes glycolysis in GBM cells

In order to further verify whether GAMSC-EXO miR-21-5p is associated with the glycolysis process in GBM cells. we examined the effects of GAMSC-EXO miR-21-5p on glucose utilization, lactate production and intracellular ATP levels in U87 cells. In Fig. 8A–C, compared with the NC group, GAMSC-EXO miR-21-5p mimics group (mimics) had a higher glucose utilization, lactate production and intracellular ATP levels, whereas those of the GAMSC-EXO miR-21-5p inhibitor group (inhibitor) were lower. Furthermore, the OCR of the mitochondrial oxidative respiration phase in mimics group was significantly increased and basal respiration and maximum respiration were also enhanced, whereas those of the inhibitor group were lower than the NC group (Fig. 8D and E). Moreover, the ECAR results displayed that compared with NC group, mimics group had a higher glycolytic flux, but those of the inhibitor group were lower (Fig. 8F and G). Therefore, the above results indicated that GAMSC-EXO miR-21-5p could enhance glycolysis in U87 cells. In order to explore whether the same effect exists in other GBM cells, we performed the same experiment in U251 cells. The results showed that GAMSC-EXO miR-21-5p could also enhance glycolysis in U251 cells (Figure S2).

Fig. 8
figure 8

GAMSC-EXO miR-21-5p promotes glycolysis in U87 cells. Statistical analysis of glucose utilization (A), lactate production rate (B), and intracellular ATP level (C) in the NC, mimics and inhibitor group, *P < 0.05 vs. NC. Oxygen consumption rate (D), basal respiration and maximal respiration analysis (E) of U87 cells in the NC, mimics and inhibitor group, *P < 0.05 vs. NC. Extracellular acidification rate (F), glycolysis and glycolytic capacity analysis (G) in the NC, mimics and inhibitor group, *P < 0.05 vs. NC.

Full size image

Discussion

GAMSCs are a type of mesenchymal stem cells responsible for the growth, chemoresistance, and recurrence of gliomas5,7,23. In this study, we revealed the role of GAMSC-EXOs in regulating glycolysis in GBM cells. Our discoveries offer profound insights into the metabolic reprogramming of GBM, facilitated by GAMSC-EXO, which contribute to the aggressive characteristics of this malignancy.

Intercellular communication serves as a crucial mechanism for cellular interaction. Exosomes, as a common intercellular communication medium, are widely involved in the regulation of biological activities between cells24,25,26, which open up a wide range of possibilities for intercellular communication27. Although this study explored the regulation of GBM cells by exosomes derived from GAMSCs, it is also a potential possibility that the biological process of exosome release by GAMSCs is regulated. We speculate that GBM may also act as a regulatory cell to release certain substances that interact with GAMSCs, thereby regulating the process of GAMSCs releasing exosomes, thus forming a closed loop with our current research. This study showed that GAMSC-EXO significantly enhanced glycolytic activity in GBM cells, which is consistent with the concept of the Warburg effect28, that is, cancer cells heavily rely on glycolysis to generate energy regardless of oxygen conditions29. This shows that the glycolysis pathway plays a crucial role in the survival of tumor cells, and it provides most of the energy source for tumor cells.

PDHA1 is one of the main components of the pyruvate dehydrogenase complex and can inhibit the Warburg effect. It has been indicated that diminished expression of PDHA1 is one of the factors contributing to the increased malignancy of cancer cells.17. MicroRNAs, a type of small non-coding RNAs, can suppress protein expression levels. Many studies have shown that the dysregulation of MicroRNAs can lead to imbalances in many biological processes, thus causing the occurrence of many diseases30,31,32. In this study, we found that the downregulation of PDHA1 in GBM is directly regulated by miRNA-21-5p. GAMSC-EXO directly provides miR-21-5p to GBM, which explains why GAMSC promotes GBM malignancy from the perspective of cell–cell interaction.

This study only focused on the effect of GAMSC-EXO on GBM cell glycolysis, which is also a limitation of this study. In fact, previous studies by our team found that GAMSCs were involved in changes in the GBM microenvironment (including immunity and angiogenesis)20,33,34, but did not explain whether these changes in the tumor microenvironment were caused by extracellular vesicles produced by GAMSCs. In addition, in addition to GAMSCs, whether extracellular vesicles from other cell sources have similar or different effects as GAMSCs is also worthy of further exploration. From the perspective of cell-to-cell interactions, what drives these cells to affect the progression of GBM is also worth thinking about, which requires further improvement and exploration.

Together, this study demonstrated that GAMSCs provide miR-21-5p to GBM by secreting exosomes, thereby reducing the expression level of PDHA1 in GBM, promoting GBM glycolysis and cancer cell phenotypes such as proliferation and invasion.