While numerous studies have focused on IFNs produced by immune cells in the TME, glioma stem cells also retain intrinsic type I IFN signaling that enables immune evasion from host cells. The Gl261, CT2A, and SMA-560 glioma cell lines uniformly express IFNAR1 and IFNAR2. Thus, autocrine type I IFN signaling is likely intact within these tumor cells lines, and ablation of IFNAR in the Gl261 cell line affected in vitro and in vivo tumor cell growth (37). Furthermore, glioma stem cells evade host type I IFN suppression through the downregulation of STAT1 mediated by the MBD3/NuRD complex (38). Inactivation of STAT signaling provides a means of escape for glioma tumor cells from IFN suppression. Another study showed that glioma stem cells exhibit a differential cell-intrinsic type I and type II IFN signature. This cell-intrinsic IFN signaling among glioma stem cells regulates tumor cell proliferation and correlates with mesenchymal phenotypes (39). In addition to the effects of tumor IFN signaling, other cell-intrinsic characteristics of tumor cells affect tumor immunogenicity. For example, high levels of PD-L1, expressed by cancer cells, inhibit cytotoxic type I IFN responses and sustain chronic responses that enhance the IFN-related damage resistance signature (IRDS). This, in turn, prevents cancer cell death (40).

Apart from tumor intrinsic IFN signaling, exogenous type I interferons are able to inhibit glioma cells. IFNβ suppressed growth of glioma stem cells by downregulating cell proliferation and ribosome pathways (25). Also, pulse stimulation of GSCs with IFNβ reduced sphere formation capacity and migratory signature (41). In addition, treatment of IFNβ sensitized GSCs to temozolomide, which was associated with upregulation of XAF1 and TRAIL death ligands (42). Altogether, tumor-derived interferon signaling can promote immune evasion from host cells while exogenously administered interferons can inhibit GSC growth.

Astrocytes account for approximately half of the cells in the brain and contribute to proper CNS development. Astrocytes maintain the overall structure of the blood-brain barrier through interactions with pericytes and endothelial cells (43, 44). Within the brain TME, astrocytes interact with tumor cells via gap junctions, ion channels, and the release of gliotransmitters and extracellular vesicles (45). During GBM development, astrocytes are activated and the TME polarizes reactive astrocytes to tumor-supporting glial cells. Reactive astrocytes in the GBM compartment produce tumor-promoting factors such as TGFβ and STAT3 that promote tumor metastasis (45). A comparison of high- and low-grade gliomas revealed a subpopulation of astrocytes expressing high levels of osteopontin (SPP1) in high-grade gliomas, suggesting a correlation with poor survival among glioma patients (46, 47).

Type I IFN signaling in astrocytes regulate immune responses in the CNS through cell-intrinsic and extrinsic manners. For example, in viral neuroinfections, astrocytes produce type I IFNs (48), indicating their role as IFN producers in the CNS and potentially in the brain TME as well. Type I IFN signaling is critical in astrocytes for synaptic plasticity and memory formation mediated through the astrocytic glutamate-aspartate transporter (GLAST). IFNAR loss in astrocytes impairs proper hippocampus and cognitive functions through glutamate uptake modulation (49). Tumor-associated astrocytes in the GBM TME are anti-inflammatory and are enriched in gene sets associated with JAK/STAT pathway activation and IFNα and IFNγ responses. These tumor-associated astrocytes were later suggested to mediate specific re-programming of microglia cells (50).

Resident microglia are brain tissue-resident macrophages that make up the main immune cell populations in the brain. These cells participate in various physiological processes in the brain including neurogenesis and axon growth (51). Recent advances in imaging mass cytometry have provided insight into the spatial organization of the GBM TME, revealing that tissue-resident microglia and monocyte-derived macrophages comprise the dominant immune populations across human GBM samples (52). There is a heterogenous population of these glioma-associated microglia (GAMs) that are divided into antitumoral M1 and immunosuppressive M2 phenotypes (53). M1 GAMs are distinguished by their ability to secrete proinflammatory cytokines such as IL-23, IL-12, IL-6, and IL-1β along with the production of reactive oxygen species (ROS). Expression of STAT1 by M1 GAMs promotes their antitumoral function due to the tumor suppressor activities associated with STAT1 expression (54). M2 GAMs, on the other hand, are characterized by low expression levels of MHC II, IL-12, and IL-23 and secretion of anti-inflammatory cytokines such as TGFβ and IL-10, which are important for the prevention of destructive immune responses (51).

Type I IFNs allow microglia to exert multifaceted roles in the CNS, including maintaining homeostasis and recovering from injuries (55). The overall immune landscape during the early stages of GBM is predominated by M1 proinflammatory GAMs, characterized by an upregulation of genes associated with proinflammatory processes such as Tnf, Il1b, Il1a, and Cxcl10 (56). Despite the growing evidence of GAM functions in the brain TME, there are still unanswered questions regarding how type I IFNs affect microglial function in CNS diseases and, in particular, gliomas. Attempts to narrow this knowledge gap can be seen in a recent study demonstrating that peripherally derived IFNα can directly transduce signaling effects across the blood-brain barrier on resident microglia. Direct microglia IFN signaling is involved in the transcriptomic changes that are seen in the brain parenchyma as a result of peripherally administered IFNα (57). In addition, adenoviral-mediated IFNβ expression in an orthotopic Gl261 GBM model showed an activated microglia phenotype, in correlation with increased tumor cell death and improved mouse survival (58).

Growing evidence suggests that monocytes play a role in the development and progression of cancers, including GBM. The current understanding regarding the effects of type I IFNs on monocytes in cancerous settings is based on its immunostimulatory effects. The frequency of suppressive monocytes is often associated with resistance to immune checkpoint blockade (ICB) therapy. Cancer cell-derived type I IFNs polarize a subset of intratumoral monocytes to function as immunostimulatory mediators. Type I IFN signaling is necessary for tumor-associated monocytes to drive CD8⁺ T-cell proliferation as well as restrain the immunosuppressive activities of different monocyte subsets (59). In addition, a recent study conducted by Zemek et al. showed that inflammatory monocytes are the primary source of IFNβ in responding tumors and that type I IFN signaling within tumor-infiltrating monocytes contributes to T cell expansion (60). In addition, Ochoka et al. suggested that monocyte/macrophage clusters show a higher expression of type I IFN-related genes in relation to other cell clusters, indicating that monocytes are affected by type I IFN signaling within the GBM TME (61)

In malignant gliomas, circulating monocytes migrate to the tumor niche and differentiate into either TAMs or tumor-associated microglia (62). A reduction in classical and non-classical monocytes was observed in glioma patients compared to healthy volunteers. Non-classical monocytes from the glioma patients exhibited a proinflammatory cytokine profile (TNFα and IL-12), suggesting a role of monocytes in antitumor immune responses against gliomas (63). Monocytes have become potential therapeutic candidates as IFN gene-delivery vehicles into the brain TME. Tie-2 expressing monocytes, due to their homing ability to the TME, are efficient cellular vehicles for the delivery of IFNs via lentiviral transduction methods (64, 65).

TAMs make up a prevalent portion of the heterogeneous population of immune cells in the brain TME. Depending on their functions, TAMs are divided into M1 and M2 subtypes. In general, M1 TAMs carry out antitumoral functions that facilitate Th1 responses and secrete proinflammatory cytokines such as TNFα. M2 TAMs promote tumor progression through the secretion of anti-inflammatory cytokines, such as TGFβ, that promote Treg and tumor cell proliferation (66). These M1 and M2 TAMs can also be distinguished based on whether they reside in the tumor periphery or the hypoxic tumor core, respectively (67).

A recent study identified a subset of proinflammatory CD169⁺-expressing TAMs that contribute to antitumor immune responses against GBM. These cells are enriched in gene sets for responses to IFNα, suggesting that CD169⁺ TAMs are activated by type I IFNs (68). Transcriptomic profiling of myeloid cells in Gl261 glioma-bearing mice shows that glioma-associated macrophages are enriched in gene sets related to responses to IFNβ and several upregulated gene sets that overlap with activated microglia populations (61). Taken together, heterogenous populations of TAMs in high-grade glioma TMEs exhibit a type I IFN signature that plays a role in shifting TAMs toward protumoral or antitumoral myeloid cells.

Several studies have detailed the roles of DCs within the brain TME. Experimental data suggest an interplay between DCs, microglia, macrophages, and T cells (69). The main role of tumor-infiltrating DCs is to take up and present tumor-associated antigens to CTLs (70). In a microenvironment where there is a high level of immunosuppression, such as within the brain TME, DCs are usually in an inhibitory or immature state (71). Differentiation of DCs to other subsets is inhibited by immunosuppressive cytokines that are secreted by glioma cells. For example, prostaglandin E2 is produced by glioma cells that, in turn, promotes IL-10 production by DCs to inhibit effector T cell responses (71). In another study, the role of the cDC1 subset was characterized during an endogenous immune response to brain tumors. cDC1s are required for neoantigen-specific responses and responses to ICBs in syngeneic models of GBM (72).

One of the main immunostimulatory actions of type I IFNs on DCs is to enhance their cross-presentation ability to CD8⁺ T cells. Transcriptional profiling of intratumoral DCs revealed that tumor-infiltrating DCs exist in different functional states. Activation of CD11b⁺ DCs is characterized by an ISG signature, and the “ISG-DCs” can activate CD8⁺ T cells. Sustained IFN signaling derived from local tumor cells promotes this ISG-like state within CD11b⁺ intratumoral DCs. Thus, these DCs present tumor-derived peptides via MHC I dressing (73).

The antitumoral activities of plasmacytoid DCs (pDCs) have been studied in various solid tumor models. The hallmark of pDCs is their ability to produce large amounts of type I IFN. Their role in antiglioma immunosurveillance was described in a study that used transgenic mice to deplete pDCs. Using a syngeneic murine model of glioma, pDC depletion leads to increased survival of mice bearing intracranial Gl261 tumors. Additionally, there is a notable change in immune cell infiltrates. Analysis of the immune cell profile revealed a decrease in immunosuppressive regulatory T cells upon pDC depletion in Gl261 tumor-bearing mice (74), suggesting a role of pDCs in antitumor immunity against GBM.

CTLs are the main executors in the regulation of antitumor immune responses and make up the backbone of cancer immunotherapy (75). These effector T cells have been extensively studied in viral infections, and their detailed roles in antitumor responses against brain tumors have recently gained much attention. A positive GBM prognosis is associated with a high ratio of CD8⁺ T cells to CD4⁺ T cells. Among CD8⁺ tumor-infiltrating lymphocytes, a subset of CD8⁺Tbet⁺ cells indicates the presence of activated proliferating CD8⁺ T cells within the TME (76). Although cytotoxic CD8⁺ T cells are known for their effector functions, these cells may exist in a dysfunctional state when exposed to the brain TME. Chronic immune activation leads to terminal CD8⁺ T-cell exhaustion mediated by IL-10 release from the myeloid compartment of the TME (77). Other factors that render T cells dysfunctional in the brain TME include high levels of hypoxia (78), limited glucose availability (79), and production of oncometabolites (80). Thus, the role of CD8⁺ effector T cells cannot be ignored during antitumor immunity against brain tumors. Various cytokines and chemokines mediate and regulate the activity of these CTLs, including type I IFNs.

The immunomodulatory effects of type I IFNs on CD8⁺ T cells are pleiotropic (81). Type I IFNs directly enhance the cytotoxic functions of CD8⁺ T cells, stimulating the production of IFNγ and TNFα and enhancing the secretion of granzyme A, perforin, and granzyme B. IFNs with IL-2 can function as a third signal for naïve CD8⁺ T-cell differentiation (82) and protect T cells from natural killer (NK) cell-mediated attack (83). T cells that lack IFNAR are more susceptible to NK cell-mediated attack through the expression of natural cytotoxicity triggering receptor 1 (NCR1) (84). Ablation of IFNAR in mouse models further supports the role of type I IFNs in the pathogenesis of gliomas. Glioma-induced Ifnar1 ^-/- mice show decreased infiltration of CD8⁺ T cells and reduced potency in their cytotoxic functions (19).

CD4⁺ T cells are differential coordinators of adaptive antitumoral immunity and are best known for their helper functions (85). Once stimulated by antigens presented by MHC II, different helper T cell subsets orchestrate the direction of the immune response through cytokine secretion (86). Helper CD4⁺ T cells typically promote CTL function through the activation of DCs and regulate the myeloid compartment (85). The role of CD4⁺ T cells in regulating antitumor immune responses has been overlooked due to the clinical success of CD8⁺ Tcell-based immunotherapies, and only recently has the importance of CD4⁺ T cells in driving antitumor immunity been recognized. CD4⁺ T cells have been implicated in ICBs in brain tumors. CD4⁺ T-cell deficiency drives CD8⁺ Tcell exhaustion leading to unresponsiveness to PD-1 blockade, as demonstrated through CD4 depletion in a Gl261-induced glioma mouse model (87, 88).

Interestingly, there has been recent evidence that similar to cytotoxic CD8⁺ T cells, cytotoxic CD4⁺ T cells retain antitumoral activity (89). CD4⁺ CTLs, similar to their CD8⁺ CTL counterparts, express granzymes and perforins to carry out tumor-killing functions (89). The direct cytotoxic potential of CD4⁺ T cells in the TME has been shown in human cancers where tumor-specific CD4⁺ T cells portray similarities to classical CTLs. Transcripts for cytolytic-related molecules, such as IFNγ, CCL4, and CCL5, are enriched in these cytotoxic CD4⁺ T cells (90). Thus, CD4⁺ CTLs may play an important role in antitumor immunity against gliomas.

CD4⁺ CAR T cells targeting GBM-associated IL-13 receptor α2 outperform their CD8⁺ CAR T cells counterparts in antitumor activity and exhibit less activation-induced exhaustion features (91). Increased expression of granzyme A, granzyme B, and perforin by intratumoral CD4⁺ CTLs in gliomas suggests these cells to be a positive prognostic marker for glioma survival (89). Conversely, CD4⁺ T cells can also be negatively associated with ICB therapy. Transcriptomic analysis of tumor-specific CD4⁺ T cells revealed a type I IFN response signature with an upregulated exhaustion signature characterized by increased PD-1 levels (92).

Regulatory T cells (Tregs), a subset of CD4⁺ T cells, are usually characterized by FOXP3⁺ expression. CD4⁺CD25^highFOXP3⁺ Tregs function in maintaining self-tolerance, and previous studies have demonstrated their contribution to the suppression of antitumor immune responses (93). Tregs, in addition to M2-like macrophages/microglia, predominate the suppressive immune cell populations within the GBM TME, contributing to ICB resistance and dampening CTL responses (94). Evidence strongly suggests that increased Treg populations are associated with higher histological grades in gliomas, with a significant amount of regulatory T cell infiltration in GBM (95). Immunosuppressive functions of Tregs are perturbed upon αPD1 treatment as was shown using a GBM model. Tregs exhibit an anergic phenotype characterized by the expression of CD73 and FR4. These regulatory T cells express glucocorticoid-induced tumor necrosis factor (GITR) that, when blocked, enhances CD4⁺ effector functions and reduces the immunosuppressive effects of Tregs. The combination of αGITR and αPD1 essentially converts GBM Tregs to Th1-like effector cells (94). In addition, the depletion of Tregs in the TME is associated with improved survival. Administration of the IL-2Rα antibody daclizumab to GBM patients depletes Tregs leading to enhanced antitumor responses (96).

Thus, methods to suppress Treg function within the brain TME may result in a better prognosis for GBM survival, and type I IFNs may provide leverage to overcome Treg-mediated immunosuppression. Rather than a direct inhibitory effect, type I IFNs may indirectly render Tregs dysfunctional through chemokine regulation. Intratumoral IFNα gene delivery into tumors enhances antitumor immune responses by reducing infiltration of regulatory T cells and production of IL-6, a Treg-inhibiting cytokine (97). Another indirect effect of type I IFN-mediated Treg suppression occurs when intratumoral IFNα delivery downregulates tumor expression of CCL17 that trafficks Tregs to tumor centers (93).

NK cells are the innate counterpart of CTLs due to their ability to lyse malignant cells (98). NK cells can kill cancerous cells independently of T cells through secretion of granules containing membrane-disrupting proteins and granzymes. In a recent study, NK cells represented the highest percentage of total infiltrating immune cells in brain cancer, suggesting an indispensable role for these cells (99). NK cells can alter the phenotype of glioma stem-like cells and reduce chemotherapy resistance (100), which is important when considering future immunotherapy options for brain tumors. Despite the immune-activating effects of NK cells, they can also be rendered dysfunctional in the GBM TME (101). For instance, tumor growth factors (TGFβ) from the sera of glioma patients downregulates the expression of NKG2D on NK cells and CD8⁺ T cells, rendering them less efficient in killing tumor cells (102). Thus, attempts to restore NK cell functions are being explored in emerging NK cell-based therapies. Single-cell RNA sequencing analysis of infiltrating NK cells in a Gl261 syngeneic glioma model revealed the downregulation of IFN genes and reduced expression of activation markers (103), suggesting a role for type I IFNs in NK cells in brain tumors.

NK cells are directly activated by type I IFNs resulting in enhanced cytotoxic functions, production of inflammatory cytokines, or crosstalk with antigen-presenting cells (84, 100, 104). NK cell-mediated tumor surveillance is hindered when IFNAR is abolished among these cells, hindering their cytotoxic capabilities as demonstrated in lymphoma and melanoma models (105). There are still unsolved questions concerning the detailed mechanisms of how IFNs regulate NK cells within the brain TME.

Exogenous administration of recombinant IFNs is used for treating a range of cancers including hairy cell leukemia, B- and T-cell lymphomas, and solid tumors, such as melanoma and renal cell carcinoma ( Figure 3A ). Dating as far back as 1986, recombinant IFNα received United States Food and Drug Administration (US FDA) approval for treating hematological malignancies, becoming the first therapeutic drug to treat cancer patients (97, 106, 107). Since then, many clinical trials have shown the efficacy of using recombinant type I IFNs in treating patients with solid tumors (106). The efficacy of recombinant IFN therapy has been demonstrated in a randomized clinical trial involving high-grade glioma patients. The overall survival was improved upon treatment with temozolomide and IFNα, demonstrating the benefits of combination treatment over the conventional use of temozolomide alone (108). In addition, the combination of IFNα with BCNU and radiation therapy showed increased overall survival among high-grade glioma patients, thus demonstrating the safe and feasible use of IFNs with other cancer therapies (109).

Recombinant IFN therapy is not a new concept, yet this type of immunotherapy has its setbacks. Virtually all patients treated with recombinant IFNα experience adverse side effects during the course of treatment. Administration of recombinant IFNs can provoke several reactions including cardiovascular, respiratory, endocrine, and metabolic problems (110). Thus, although there has been some success in treating solid cancers including gliomas, the adverse reactions associated with this type of treatment call for a safer and more efficient method for use of IFNs in immunotherapies. Lessons from clinical trials have shown that there are still unresolved issues that need to be addressed. These include the optimal dose-scheduling of IFN treatment. It is important to understand what levels are required to reach optimal interaction with radiation therapy to optimize treatment (111). In addition, optimizing the scheduling of IFN treatment in regards to duration and timing is needed in order to avoid prolonged damage associated with radiosensitization (111).

Though not necessarily explored in gliomas, alternative strategies to overcome the limitations of recombinant interferon therapy have been explored by fusing tumor antigens with interferons in different tumor models (112, 113). Research into improving recombinant interferon therapy in gliomas has not yet been extensively studied. But given the success of combining tumor targeting antibodies with modified forms of interferons in other solid tumors, similar techniques may also be applied to treat gliomas.

Type I IFNs have broad implications in various physiological processes, especially in the context of immune cells. Thus, given the powerful functions of these cytokines, there have been many recent attempts to harness them as therapeutic drugs. Yet the pleiotropic effects and systemic toxicity that occur when using IFNs as therapeutic drugs have been major hurdles (114). Thus, novel approaches for delivering IFNs to the TME have been explored.

Recent advances in gene therapy show that targeted delivery of immunoactivating cytokines into the TME can endure and overcome TME immunosuppression, while tightly regulating the expression of the delivered cytokines ( Figure 3B ). Birocchi et al. demonstrated that the inducible release of IFNα into the GBM TME perturbed tumor growth and essentially reprogrammed the immune microenvironment toward a proinflammatory and antitumoral state. This reprogrammed immune landscape was characterized by proinflammatory TAMs that exhibited a gene signature associated with poor prognosis in human GBM. This inducible system was able to limit overall systemic toxicity by tightly regulating the release of IFNα into the TME (65). Additionally, Palma et al. showed that angiopoietin receptor Tie2-expressing monocytes have a unique capacity to home to tumors. Therefore, using a lentiviral vector system, these monocytes were engineered to express IFNα and were delivered to the GBM TME, where improved antitumor activity against brain tumors was observed (64).

Other novel approaches to deliver type I IFNs into tumors include arming PD-L1 antibodies with IFNα to overcome PD-L1-mediated immune suppression, while at the same time, increasing antigen cross-presentation to enhance T cell activation (115). Although this technique is therapeutically beneficial in advanced tumors, such as colon cancer and lymphomas, this approach has not been explored in brain tumors.

In general, recent breakthroughs in cancer therapy using ICB have attracted significant attention. Recent success in ICB therapy targeting PD-1 and CTLA-4 has provided a breakthrough in the field of immunotherapy ( Figure 3C ) (116). Despite some clinical success, however, ICB has not been fully effective in treating high-grade gliomas such as GBM. The immune checkpoint receptors PD-1 and CTLA-4 are expressed on the surfaces of T cells and inactivate T cells once in contact with tumor cells (116). Engagement of PD-1 with its ligands PD-L1 and PD-L2 maintains peripheral tolerance and compromises antitumor immunity (117). Blocking the interaction between these immune checkpoint receptors and tumor cells, thereby preventing T cell inactivation, is the essence of ICB therapy. However, three controlled trials using the monoclonal anti-PD1 antibody nivolumab have failed to improve survival among GBM patients (118). Resistance to ICB therapy, in part, can be credited to sustained intrinsic IFN signaling among tumor cells. Mechanisms of ICB resistance can also be attributed to PD-L1 independent adaptive resistance that is associated with type I IFN gene expression. As demonstrated by Benci et al., PD-L1 independent resistance is essentially orchestrated by chronic IFN signaling (119). Therefore, new potential biomarkers to evaluate the responsiveness of ICB among GBM patients are needed. The biomarkers of ICB resistance may include PD-L1 positivity on tumor and immune cells, an IFNγ gene signature, and the level of IFN signaling that exists within cancerous cells (118). The combined use of ICB and exogenous IFN administration may perhaps improve the clinical efficacy against brain tumors.

DCs have gained interest in their use as therapeutic cancer vaccines. Due to the ability of DCs to process and present antigens to CTLs and helper T cells, DC vaccines exploit this function for therapeutic use ( Figure 3D ). In general, patients are vaccinated with tumor-associated antigen (TAA)-loaded DCs to initiate an antitumoral T cell response (120). These DCs have either been generated from monocyte precursors or CD34⁺ hematopoietic precursors and expanded in vivo (121). Increased cytotoxic T cell infiltration with reduced levels of TGFβ in brain tumor regions were seen in dendritic cell-vaccinated patients (122). Combination of autologous tumor lysate-loaded dendritic cell therapy combined with standard therapy showed an improvement in survival of patients with recurrent glioblastoma (123).

In addition, several studies have demonstrated that type I IFNs markedly enhance the maturation, proliferation, and activation of DCs. These so-called ‘IFN-DCs’ have been implicated in high-grade gliomas. LPS-stimulated IFN-DCs from high-grade glioma patients exhibit high levels of the costimulatory molecule CD86 and MHC II antigens, similar to donor IFN-DCs (124). Thus, IFN-DCs can be promising therapeutic candidates to use in DC-based vaccines to treat GBM.