Macrophages and microglia are important cell types in the immune system. Macrophages are primarily derived from bone marrow-resident haematopoietic stem cells (112). After entering the blood, mature mononuclear macrophages can settle in different tissues, such as the liver, lung, brain, lymph nodes and other organs and tissues, at which time they will become macrophages. Macrophages are involved in phagocytosis and clearance of pathogenic microorganisms, necrotic tissues, and secretion of a variety of inflammatory mediators involved in immune regulation and tissue repair (113–115). Microglia are induced by the colony-stimulating factor 1 receptor (CSF1R) and are generated from red myeloid progenitors of the yolk sac. They are self-renewing and reside in the CNS for a long time (116). Microglia play an important role in maintaining the homeostasis of the nervous system, including phagocytosis (117), promoting synapse formation (118), and supplying nutrients (119).

In the TME, macrophages can manifest as the M1 type (characterized by inflammatory and antitumour responses) or M2 type (involved in the repair of damaged tissues and anti-inflammation), but the TME tends to induce the differentiation of macrophages towards the M2 type (120–122). Anti-inflammatory factors released by tumours, such as interleukin (IL)-4, IL-10, and transforming growth factor-β (TGF-β), can promote the transformation of macrophages into the M2 type (123). M2 macrophages similarly release growth factors, such as vascular endothelial growth factor (VEGF), epidermal growth factor (EGF), and fibroblast growth factor (FGF), which promote tumour cell proliferation, invasion, and metastasis (124).

Glioma-infiltrated macrophages and microglia are collectively referred to as GAMs, which represent the largest population of cells infiltrating tumours, accounting for more than 1/3 of the total tumour mass (125, 126). GAMs play an important role in the glioma TME and promote tumour progression. First, anti-inflammatory factors such as IL-10 and TGF-β produced by GAMs inhibit the function of other immune cells in the TME and weaken the antitumour immune response (127, 128). Second, GAMs also secrete matrix metallopeptidase 2 (MMP2) and MMP9, which are able to breakdown matrix proteins, such as collagen and fibronectin, thereby enabling glioma cells to penetrate and invade the surrounding stromal tissues (129). Finally, GAMs also secrete proangiogenic molecules, such as VEGF and CXC motif chemokine ligand 2 (CXCL2), which have been shown to promote glioma growth and metastasis (129, 130).

In the glioma TME, ROS generally induce the generation of M2 GAMs (96, 131). ROS modulator 1 (Romo1), a membrane protein located on mitochondria, was found to regulate mitochondrial ROS (mtROS) production in GBMs (132). In GBM mouse models, overexpression of Romo1 induces ROS generation via mTORC1 signalling, which in turn promotes the polarization of bone marrow-derived macrophages (BMDMs) to the M2 type, resulting in a significant suppressive TME (133). In addition, a prognostic model of human GBMs and ROS showed that high expression of ROS-related genes such as HSPB1, LSP1 and PTX3 was closely associated with M2 macrophages and correlated with shorter survival of GBM patients. This suggests that ROS-related genes may be potential targets for GBM treatment. Therefore, inhibiting the polarization of macrophages towards M2 type and promoting the polarization towards M1 type may be beneficial for the treatment of GBMs (134). Besides, GAMs could survive in a high ROS environment mainly due to the action of antioxidant enzymes. A study of GBM tissues in humans and mice showed that the active antioxidant enzyme GPX1 was expressed at higher levels in GAMs than in GBM cells, resulting in GAMs being able to survive in a high ROS environment (135). It is known that GPX1 plays an important role in H₂O₂ detoxification (136). In summary, the antioxidant enzymes in GAMs protect them from ROS damage, which is necessary for the formation of M2 GAMs.

MDSCs, which are mainly differentiated from haematopoietic stem cells in the bone marrow, are a group of myeloid cells with heterogeneous and immature characteristics (137). In normal organisms, the levels of MDSCs in the peripheral blood tend to be very low (138). MDSCs have certain immunomodulatory effects, which can regulate the inflammatory response, inhibit overactivated immune cells, prevent excessive immune responses, and reduce tissue damage (105, 139).

Upon tumour stimulation, MDSCs are activated and released into peripheral blood and tissues. However, MDSCs often suppress the immune response and cause tumour escape (140, 141). The suppressive effect of MDSCs is mainly manifested by inhibiting the activity of other immune cells, including macrophages (142), CD4⁺ T cells (106), CD8⁺ T cells (143), NK cells (144), and DCs (145). First, MDSCs can highly express arginase-1 (ARG-1), which can convert arginine to uric acid and ornithine, thereby reducing the concentration of arginine in the internal environment (105). Arginine deficiency leads to limited activation of immune cells such as T cells (146) and NK cells (147), thereby impairing the immune response and promoting tumour development and metastasis. Second, MDSCs can secrete immunosuppressive cytokines, such as TGF-β and IL-10, which can inhibit the secretion of IL-12 by macrophages, thereby blocking the activity of cytotoxic T lymphocytes (CTLs) (148, 149). TGFβ-1 secreted by MDSCs also promotes the transformation of CD4⁺ T cells into immunosuppressive Tregs (150). Finally, MDSCs can also express immunosuppressive ligands, such as programmed death ligand-1 (PD-L1), which in turn suppresses T-cell priming and activation (151).

In gliomas, MDSCs comprise approximately 30% to 50% of the tumour entity (152). The increase in MDSCs is thought to be associated with glioma progression and immune escape (153). Generally, MDSCs can be divided into two main subsets based on their phenotype and function: monocytic (mMDSCs) and granulocyte/polymorphonuclear (gMDSCs) (154). Specifically, mMDSCs represent the major subset in the GBM TME. mMDSCs in the GBM TME of humans and mice expressed higher levels of adhesion molecules, such as integrin β1 and dipeptidyl peptidase 4 (DPP4), leading to enhanced cell adhesion and further promoting tumour migration and invasion (155). In addition, MDSCs can promote angiogenesis through the release of VEGF (156), as well as the release of cytokines such as IL-10, IL-6, and TGF-β under hypoxic conditions (157), thereby promoting glioma growth and invasion.

In the TME, MDSCs can survive in a high ROS environment because of their high expression of NRF2. On the one hand, NRF2 upregulated anti-OS genes in MDSCs and protected MDSCs from OS damage. On the other hand, NRF2 enhances the immunosuppressive activity of MDSCs by increasing their ability to produce ROS (158). In gliomas, ROS in MDSCs play an important role in maintaining the function of MDSCs (159). ROS can prevent the differentiation of MDSCs and promote the formation of an immunosuppressive TME (160). Specifically, ROS maintain the undifferentiated state of MDSCs by inhibiting the differentiation of MDSCs into mature immune cells such as macrophages and DCs (145, 161, 162). This undifferentiated state allows MDSCs to continue expressing immunosuppressive molecules such as TGF-β, IL-10 and PD-L1 (163). In addition to being able to impair the antigen presentation capacity of DCs (164–166), these immunosuppressive molecules can also inhibit the activity of T cells and induce the differentiation of T cells into Tregs (141, 167). Collectively, high levels of ROS play an important role in maintaining the undifferentiated state of MDSCs, which in turn mediates the immunosuppressive TME. Therefore, targeting MDSCs may become a promising therapeutic strategy.

T cells are members of the adaptive immune system that respond to antigens presented by antigen-presenting cells such as DCs and macrophages (168). T cells can be divided into CD4⁺ T cells and CD8⁺ T cells based on surface markers and function (169). CD4⁺ T cells have antigen receptors on their surface, which can recognize antigen fragments presented by MHC class II molecules on the surface of antigen-presenting cells and exert immune functions by activating other types of immune cells. CD8⁺ T cells generally refer to CTLs that can directly kill infected cells by MHC class I molecules (170). Regulatory T cells (Tregs) are a subset of CD4⁺ T cells that express the transcription factor forkhead box protein 3 (FOXP3) and play a role in inhibiting pathological immune responses and maintaining homeostasis in the body (171).

In tumours, CD4⁺ T cells mainly activate other immune cells, such as CD8⁺ T cells and NK cells, to enhance the immune response (172). CD4⁺ T cells can secrete cytokines, such as interferon-gamma (IFN-γ) and tumour necrosis factor-alpha (TNF-α), which directly kills tumour cells (173). CD8⁺ T cells carry specific T-cell antigen receptors (TCRs) that recognize and bind to tumour cells expressing specific antigens, thereby releasing cytotoxins, such as perforin and granzyme, to directly kill tumour cells (174). In addition, CD8⁺ T cells can also secrete cytokines, such as IFN-γ and TNF-α, which directly inhibit tumour cell growth and proliferation (175). Tregs play an indispensable role in maintaining the homeostasis of the immune system. Tregs suppress other immune cells and prevent excessive immune responses by producing inhibitory cytokines such as IL-10, IL-35, and TGF-β. However, overactive Tregs in turn limit the antitumour ability of immune cells (176).

In gliomas, T-cell dysfunction is often caused by the strong immune escape ability of glioma cells. Some studies have shown that human GBM cells are capable of producing the immunosuppressive factor TGF-β (177), which inhibits T-cell activation, thereby weakening the immune response (178). In addition, the human glioma TME has many immunosuppressive cells, such as M2 macrophages and Tregs, whose presence usually inhibits the activity and function of T cells and is associated with reduced overall survival of patients (179–181). Moreover, human glioma cells often express immune escape sites on the surface, such as PD-L1 and B7 homologue 3 (B7-H3), which can bind to immune checkpoint receptors, thereby inhibiting the activity and function of T cells (182).

ROS play an important role in regulating T-cell function and activity. Low levels of ROS can promote the activation and proliferation of T cells to enhance immune responses (183). However, higher levels of ROS can inhibit the secretion of cytokines by T cells and induce apoptosis (184). In the TME, excessive ROS may induce apoptosis of T cells, leading to decreased antitumour ability. For example, ROS produced by neutrophils or tumour cells can be transferred to T cells and cause OS, thereby causing hyporeactivity of T cells in cancer patients (185). In mouse glioma models, the administration of hyperbaric oxygen (HBO) can induce the generation of ROS in the thymus, which subsequently inhibits the generation of CD3⁺ T cells and promotes glioma growth in vivo (186). Therefore, high levels of ROS in the TME may lead to impaired T-cell function, which in turn enhances tumour escape. Targeting ROS in the TME to enhance the killing ability of T cells may be a potential therapeutic option.

Neutrophils are derived from haematopoietic stem cells in the bone marrow through differentiation and maturation. When neutrophils mature, they enter the circulation and are distributed throughout the body through the blood (187, 188). Neutrophils are important immune cells of the body that are capable of engulfing and eliminating pathogens, such as bacteria and viruses (189, 190). In addition, when tissues are injured or infected, neutrophils rapidly migrate to the damaged site and release cytokines and chemokines to trigger local inflammation (191).

The role of neutrophils in tumours is complex. On the one hand, neutrophils are capable of killing tumour cells by releasing cytotoxic ROS (192) and by direct cell contact (193). On the other hand, neutrophils also promote tumour growth and metastasis by secreting immunosuppressive factors such as TGF-β, IL-6 and IL-8 (194) and interacting with circulating tumour cells (195). Furthermore, neutrophil extracellular traps play an important role in tumour progression. In the early stages of tumour invasion and metastasis, neutrophils can release neutrophil extracellular traps (NETs), which include DNA, tissue proteins and other substances. NETs can form channels suitable for tumour cell migration and protect tumour cells from immune system attack (196, 197).

Neutrophils, as mediators of inflammation, are early markers of GBM progression (198). Overall, neutrophils promote tumour growth, invasion, and angiogenesis. Neutrophils contribute to glioma infiltration by secreting elastase (199). Furthermore, neutrophils may also become resistant to antineoplastic therapy. In patients receiving anti-VEGF therapy, neutrophils contribute to glioma resistance to anti-VEGF therapy by increasing S100A4 expression and angiogenesis in glioma tissues (200). S100A4 is known to be a biomarker expressed in glioma stem-like cells (GSCs) that induces the tumorigenic activity of neutrophils (201). In addition, some studies have shown that the expression of MDSCs is increased in the peripheral blood of GBM patients, of which the neutrophilic MDSC subset accounts for the largest proportion, accounting for approximately 60% (102). Neutrophilic MDSCs derived from the peripheral blood of GBM patients can inhibit T-cell proliferation in vitro, which is related to the high expression of PD-L1 on effector memory CD4⁺ T cells (106).

Several studies have shown that ROS are important factors in promoting the formation of NETs (202–204). In chronic granulomas, NOXs are activated by protein kinase C (PKC) and produces ROS. These ROS can act as signalling molecules, causing neutrophils to release DNA and form a mesh-like structure, which then combines with the adhered granule proteins to form NETs (202). Furthermore, in primary mouse and human neutrophils, members of the MAPK family, such as c-Jun N-terminal kinases (JNKs) (203), extracellular signal-regulated kinases (ERKs) and p38 (204), can activate NOXs to generate ROS, which in turn induces the production of NETs. Similarly, ROS are similarly closely related to NETs in GBM TME. A study in human GBMs showed that NETs promote IL-8 secretion in GBMs by stimulating the NF-κB signalling pathway, which in turn stimulates endothelial cells to generate blood vessels to deliver essential nutrients and oxygen to the tumour site (101). When IL-8 binds to C-X-C motif receptor 2 (CXCR2) on neutrophils, it mediates the formation of NETs through the CXCR2/PI3K/AKT/ROS axis. This positive feedback loop stimulates the interaction between NETs and GBM cells and leads to profound changes in the TME (101). Recent studies in murine models of GBMs have additionally demonstrated that neutrophils promote the necrosis of GBM cells by transferring particles containing myeloperoxidase into these cells. This phenomenon induces OS, which is a result of the iron-dependent accumulation of lipid peroxides in GBM cells (205).

DCs, which differentiate from bone marrow haematopoietic stem cells through common DC progenitors (CDPs), play an important immunomodulatory role by presenting antigens (206). In tissues, DCs are usually naturally present and are considered to serve as a bridge connecting innate and adaptive immunity and are able to promote the transformation of innate to adaptive immune responses (207). Innate immunity refers to the immunity possessed by individuals at birth, which has a wide range of effects and is not triggered by specific antigens (208). Adaptive immunity is mainly the ability to respond to and adapt to a specific antigen or pathogen, which is achieved through T-cell-mediated cellular immunity and antibody-mediated humoral immunity (209).

Tumour formation is often accompanied by the expression of tumour antigens. Tumour antigens can be captured and processed by DCs and subsequently presented to naive T cells to induce their proliferation and differentiation into effector cells, such as CD8⁺ T cells, which subsequently kill the tumours (210). Furthermore, DCs can produce a variety of immune stimulating factors, such as cytokines and chemokines, which induce DCs and NK cells to reach inflammatory sites (211), as well as induce the activation of tumour-specific T cells (212).

Antigens released by glioma cells can be captured and processed by DCs, presented to T cells, and activate effector T-cell function (213). However, glioma cells can often evade immune surveillance by inhibiting the maturation of DCs. Studies have shown that tumour-conditioned medium (TCM) collected from the supernatant of human primary glioma cells can upregulate the expression of suppressor of cytokine signalling 1 (SOCS1) in DCs and then inhibit the NF-κB signalling pathway, thereby limiting the maturation of DCs. The subsequent suppression of T-cell activity, as well as IFN-γ secretion, results in immune escape of glioma cells (214). A previous study revealed that mouse gliomas have the ability to secrete cell factors including TGF-β and IL-10 (215). These factors are known to impede the maturation and functionality of DCs within the TME (216).

In addition, recent studies on the role of DCs in the progression of human GBMs have focused on the maintenance of DC homeostasis. Overexpression of NRF2 in DCs leads to the inhibition of DC maturation and subsequently reduces effector T-cell activation, which may be related to the decrease in ROS levels mediated by NRF2. In contrast, inhibition of NRF2 promotes the maturation of CD80⁺ and CD86⁺ DCs (217).

NK cells belong to a type of lymphocyte that can eliminate tumour cells without specific antigens and are an important part of innate immunity (218). NK cells are derived from bone marrow haematopoietic stem cells and enter the circulation after maturation (219). Approximately 5-15% of lymphocytes in normal blood are NK cells (220).

NK cells have the ability to kill tumour cells. First, NK cells kill tumour cells by making direct contact with tumour cells that express specific ligands. There are a variety of activated receptors on the surface of NK cells, such as natural killer group 2 member D (NKG2D) and natural cytotoxicity receptors (NCRs; NKp46, NKp44 and NKp30) (221), etc. Among them, NKG2D is one of the most studied receptors and is able to recognize ligands on the surface of tumour cells, such as major histocompatibility complex class I polypeptide-related sequence A and B (MICA/B) and UL16-binding protein (ULBP). This recognition activates NK cells and prompts them to kill tumour cells (222). Second, NK cells can kill tumour cells through the antibody-dependent cellular cytotoxicity (ADCC) mechanism. When the antigens on the surface of tumour cells are labelled with specific antibodies, NK cells can bind to the specific antibodies through the CD16 (FcγRIIIa) receptor on their surface. Activated NK cells then release particles containing perforin and granzyme, which trigger apoptosis of antibody-labelled tumour cells (223, 224). Furthermore, NK cells can produce cytokines, such as IFN-γ and TNF-α, which can enter tumour cells and thus kill them (225). Moreover, IFN-γ released by NK cells can also inhibit tumour angiogenesis, thereby impeding tumour nutrient supply (226).

NK cells also have a killing effect on gliomas. First, NK cells can kill glioma cells by secreting perforin and granzyme B upon induction by IFN-β (227). Second, NK cells kill gliomas by specific activating receptors on their surface. When NKG2D and DNAX accessory molecule-1 (DNAM-1) on the surface of human NK cells bind to their ligands on the surface of GBM cells, they can trigger NK cell cytotoxicity and cause GBM cell death (228, 229). However, human GBM-derived TGF-β may lead to downregulation of NKG2D receptors on the surface of NK cells and contribute to GBM cell survival, suggesting that blocking TGF-β may be beneficial in the treatment of GBMs (230). Similarly, NK cells can also kill GBM cells through ADCC. Cetuximab is a monoclonal antibody that binds epidermal growth factor receptor (EGFR) on tumour cells (231). When administered, cetuximab binds to a human GBM surface antigen (EGFRvIII) and activates fragment crystallizable (Fc) receptors on NK cells, leading to NK cell-mediated cytotoxicity against GBM cells (232).

NK cells are often particularly sensitive to the cytotoxic effects of ROS, and their antitumour activity is often inhibited by ROS in the solid tumour TME, whereas antioxidant therapy may partially restore NK cell function (110, 233). Previous studies have shown that high levels of ROS in rats with fibrosarcoma can limit the adhesion of NK cells to similarly charged tumour cells by promoting the accumulation of anionic charges on their surface. This disadvantage can be prevented by antioxidant molecules such as CAT and SOD (234). Similarly, in vitro, the CD20 antibody rituximab triggered monocyte ROS production, which in turn inhibited the ADCC effect of NK cells on human primary leukaemia cells. However, antioxidant treatment (histamine dihydrochloride and diphenylene iodonium chloride) partially restored the ADCC effect of NK cells (235). At present, although some studies have suggested the inhibitory effect of ROS on NK cell activity in the TME, the study of ROS in NK cells in the glioma TME is still limited. Further studies will help to understand the effect of ROS on NK cell function in the glioma TME.

The threshold of ROS is very important in cancer therapy. When ROS produced by tumour cells exceed a certain threshold and cannot be detoxified by antioxidants, it results in high levels of OS, which drives cancer cell death or cause them to become more sensitive to treatment. However, a low level of ROS in tumour cells contributes to their growth, proliferation, invasion and metastasis (236). Therefore, tumour cells need to maintain their ROS levels to maintain their survival and invasive abilities (237).

A study of tumour tissues and blood samples from glioma patients found that abnormal increases in ROS caused DNA damage in glioma cells, resulting in high expression of the DNA damage marker 8-oxo-deoxyguanosine (8-oxo-dG) and low expression of the epigenetic marker 5-methylcytosine (m5C). This is associated with increased malignancy of gliomas (22). In mouse models, glioma cells can overexpress aquaporin 8 (AQP8), which increases ROS levels, resulting in decreased expression of phosphatase and tensin homolog (PTEN) and increased expression of phosphorylated (p)-AKT, thereby promoting the growth and proliferation of gliomas (23). Moreover, the production of a significant amount of ROS induced by 12-O-tetradecanoylphorbol-13-acetate (TPA) can activate the MAPK pathway and cyclooxygenase-2 (COX-2)/prostaglandin E2 (PGE2) pathways, subsequently enhancing the in vitro migration and invasion capability of glioma cells (24). In contrast, high levels of ROS activate regulated glioma cell death programs, including apoptosis, necrosis, autophagy, ferroptosis, etc. For example, in vitro, salinomycin can activate p53, trigger the opening of mitochondrial permeability transition pore (mPTP), and induce the production and accumulation of mtROS, leading to the necrosis of glioma cells (25). The activation of transient receptor potential mucolipin 1 (TRPML1) inhibits autophagy in glioma cells in vitro, leading to ROS production and subsequent induction of apoptosis (238). Similarly, the increase in ROS induced by isoaaptamine also leads to apoptosis and autophagy in GBM cells in vitro (26). Furthermore, the heat shock protein 90 (HSP90) and dynamin-related protein 1 (DRP1) increase ROS production by regulating the expression of acyl-coenzyme A synthetase long-chain family member 4 (ACSL4) and mitochondrial morphology, leading to ferroptosis of gliomas in mice in vitro and in vivo (239).

Glioma stem-like cells (GSCs) are a subpopulation of GBM cells with stem cell characteristics. They have self-renewal, tumourigenicity and multidirectional differentiation potential and are closely related to the occurrence, development, treatment resistance and recurrence of GBMs (240). Many studies have confirmed that ROS are involved in the proliferation, self-renewal and differentiation of GSCs (241, 242). In a study conducted on human-derived GSCs, it was found that TGF-β upregulated the expression of the NOX4 gene, leading to the generation of ROS. Consequently, this ROS generation promoted GSC proliferation and maintained their stem cell state (241). Other studies have shown that serum stimulation in an in vitro environment is able to cause an increase in mitochondrial ROS within GSCs and modulate differentiation signalling pathways in GSCs. Interestingly, in the in vivo environment, increased ROS could greatly enhance glioma formation, which may be related to the activation of the NF-κB pathway by ROS (243). Compared with other tumour cells, GSCs have stronger antioxidant capacity (47). In vitro, the highly expressed antioxidant protein PRDX4 was able to mitigate OS in GSCs by reducing ROS generated by the protein folding process (47, 244). Furthermore, GSCs also inhibit mitochondrial respiration by increasing the expression of mitochondrial uncoupling protein 2 (UCP2), thereby alleviating OS caused by high levels of intracellular ROS and ensuring their own survival (245).

However, the understanding of the ROS threshold in cancer cells is still unclear. Measurement of the ROS threshold requires the consideration of multiple factors, including the concentration and type of ROS, the activity of intracellular antioxidant enzymes, and the type and physiological state of tumour cells (246–249). Therefore, more studies are needed to fully assess ROS thresholds and determine their impact on tumour cells. Overall, both ROS and thresholds play a crucial role in glioma cells. This provides a new research direction for ROS-based glioma therapy.

Excessive ROS can induce tumour death, so amplifying the effect of ROS may be a good way to kill tumours. For example, the use of PDT (251), SDT (252), CDT (253), can be beneficial therapies for GBMs. This part mainly summarizes the research progress of PDT, SDT and CDT in the treatment of gliomas, and discusses the application of nanodrug delivery platforms in them.

Antioxidants are a class of compounds that inhibit oxidation by scavenging ROS and reducing OS, and they can help reduce or block oxidative reactions in cells (360). When OS occurs, antioxidants interact with ROS to capture and neutralize ROS, thereby protecting cells from oxidative damage (361). Common antioxidants include vitamin C, vitamin E, α-carotene, selenium, etc., which can be obtained through food intake or supplements (362). Furthermore, the use of antioxidants can inhibit tumorigenesis by preventing OS caused by various causes, and the mechanism is to repair damaged DNA and inhibit cancer occurrence, including gene mutations, oxidative chromosomal damage, and lipid peroxidation of cell membranes (131, 363).

There is considerable evidence that intake of antioxidants may help reduce the risk of gliomas (364). For example, CoQ10 can act as a ROS scavenger to increase the sensitivity of gliomas to TMZ, thereby inhibiting the invasion of glioma cells in vitro and in vivo. Mechanistically, CoQ10 can integrate into the mitochondrial membrane and reduce ROS production. It also reduced the expression of MMP9 and epithelial-mesenchymal transition (EMT) markers (28, 365). Naringenin is an antioxidant. Naringenin supplementation for 1 month can reduce lipid peroxidation and decrease the expression of PKC, NF-κB, cyclin D1(CCND1) and cyclin-dependent kinase 4 (CDK4), thereby inhibiting the proliferation of glioma cells in mouse models (27). Astaxanthin is a natural carotenoid, and adonixanthin is a product of its formation (366). Studies have confirmed that both have strong antioxidant capacity, which can cross the BBB and protect brain tissues from ischaemia or hemorrhage (367, 368). In mouse glioma models, astaxanthin and adonixanthin intake increased p38 phosphorylation in glioma cells, leading to cell cycle arrest. Furthermore, adonixanthin was able to reduce the expression of MMP2 and fibronectin downstream of ERK1/2 and AKT signalling pathways and inhibit invasion and metastasis in both in vitro and in vivo GBM models (369). Chrysin is a kind of flavonoid with antioxidant properties. The p38-MAPK pathway is activated in rat glioma cells treated with chrysin, resulting in the accumulation of p21 (WAF1/CIP1) protein, decreased activities of CDK2 and CDK4, and cell cycle arrest in G1 phase (32). Similarly, hypoxia-inducible factor-1alpha (HIF-1α) expression was blocked when the antioxidant melatonin was used, resulting in a significant inhibition of MMP2 and VEGF expression, thereby inhibiting GBM cell migration and invasion in vitro (370). Moreover, antioxidants quercetin (QE), baicalein (BE) and myricetin (ME) effectively downregulated ROS and MMP9 and inhibited glioma cell invasion/migration events in vitro (24).

However, some studies have shown that intake of antioxidants, such as carotenoids (371), vitamin E (372), and coffee (373), is not associated with the risk of developing gliomas. This may be related to factors such as bioavailability, dose, BBB permeability, and tumour heterogeneity (364). Therefore, the role of ROS scavenging using antioxidants in glioma therapy still needs to be confirmed by more studies.