Toll-like receptors and toll-like receptor-targeted immunotherapy against glioma

All TLRs, with the exception of TLR3, either partly or fully depend upon the adaptor protein MyD88 for signaling activity [104]. Ligand binding to the LRRs in the ectodomain of TLRs affects the association of the cytoplasmic TIR domains and the MyD88. Activation of MyD88 results in myddosome formation by interacting with IL-1R-associated kinase-4 (IRAK-4), which then phosphorylates IRAK1 and IRAK2. This in turn promotes the activation of tumor necrosis factor receptor-associated factor 6 (TRAF-6) which acts as a ubiquitin ligase and triggers two major pathways. First, TRAF6 combines with ubiquitin-conjugating enzymes UBC13 and UEV1A and adds polyubiquitin chains to NF-κB essential modulator (NEMO) or IkappaB kinase gamma (IKKγ). IKKγ together with IKKα and IKKβ forms the IKK complex, which mediates rapid phosphorylation of IκB proteins, leading to release nuclear translocation and activation of NF-κB [105]. Second, TRAF6 activates the transforming growth factor β-activated kinase 1 (TAK1), which associates with two adaptor proteins: TAK1-binding proteins 1 (TAB1) and TAB2. TAK1 exerts double action: It phosphorylates IKKβ to activate NF-κB signaling or induces a cascade of mitogen-activated kinases (MAPKs) to promote phosphorylation of c-Jun kinases (JNKs), p38 MAPK, extracellular signal-regulated kinases 1 (Erk1) and Erk2, leading to activation of AP-1 transcription factor. These two signaling arms collaborate in controlling genes expression for inflammation mediators and extracellular matrix remodeling [43, 106].

In parallel, TLR activation also stimulates IFN signaling. NF-κB binds and upregulates the expression of the promoters of interferon regulatory factor 1 (IRF1), IRF2, IRF5, and IRF8 genes, forming a cross talk between the TLR and IFN pathways [107]. The transcription factor IRF1 directly associates with MyD88, IRF5 associates with both MyD88 and TRAF6, and TRAF6 also facilitates IRF7 in response to TLR7 and TLR9 activation. The IKK complex also participates in IRF3 and IRF7 activation, which induces type I IFN and host defense system. Ultimately, activation of the IRFs, NF-κB, and AP-1 transcription factors induces transcription of proinflammatory cytokine encoding genes, such as IL-6, IL-8, IL-18, IL-1β, TNFα, as well as PD-1 [108, 109].

Aside from MyD-dependent signaling, ligand binding to TLR3 and TLR4 also activates IRF3 through TIR-domain-containing adaptor-inducing interferon-β (TRIF) and TRAF3. TRAF3 subsequently induces IFN expression through TBK1. A second pathway mediated by TLR3 and TLR4 involves TRAF6 recruitment via interaction with receptor-interacting serine/threonine protein (RIP), which triggers a late phase of NF-κB signaling, initiating further production of proinflammatory molecules [58] (Fig. 4).

The expression patterns and signaling mechanisms make TLRs potential targets for glioma therapy, where multiple routes may be targeted to aid the development of effective clinical strategies. TLR agonists have been reported to initiate or suppress immune responses in the glioma environment upon binding to specific TLRs (Table 3). In particular, local administration of TLR agonists is of particular interest for immunotherapy (Fig. 5) (Table 4).

Immune checkpoints are known as double-edged swords for their ability to protect against autoimmunity, while also providing a path for immune escape in tumorigenesis. PD-1 and PD- PD-L1 have been implicated in tumor invasion by impairing antitumor response through exhaustion of effect CD8⁺ T-cells and NK cells [182]. In GBM, the expression of PD-L1 on tumor tissue promotes PD-1 activation in microglia that leads to diminished T-cell response [183]. Up to date, checkpoint blockade monotherapy is effective only in a small fraction of patients with glioma due to reasons such as low levels of PD-L1 expression, especially in IDH-mutant glioma, and exhausted tumor infiltrating T-cells [184, 185], and the results were merely comparable with standard TMZ and radiotherapy (NCT02337491, NCT02617589). Therefore, combination strategies are urgently required to unlock further therapeutic responses.

TLR activation affects PD-L1 expression in two ways: The MyD88-dependent signaling pathway promotes NF-κB-mediated cytokine transcription, PD-L1 gene transcription, and secretion; and the MyD88-independent cascade leads to late-phase NF-κB activation and type I IFN secretion, which not only enhances cancer immunotherapy, but also upregulates PD-L1 expression on tumor cells through interferon-α/β receptor (IFNAR) signaling [183] (Fig. 6). This suggests that PD-L1-expressed cells can be beneficial for TLR-based therapies in glioma and other cancer types.

Poly I:C-induced TLR3 activation has been reported to prime TME by stimulating PD-L1 and PD-L2 expressions on GBM cells, along with the secretion of proinflammatory cytokines such as IFNs, CXCLs and CCLs to attract CD4⁺ and CD8⁺ T-cell infiltration. This immune response can be effectively reinforced via blockade of elevated tumoral PD-L1 [186]. TLR3 activation also effectively promoted DC maturation and T-cell proliferation to enhance the antitumor response to anti-PD-1, resulted in increased survival rate in GBM mouse models [187].

Similarly, TLR4 overexpression was shown to promote PD-L1 expression in GBM patients via the MyD88-independent pathway, leading to autocrine induction of regional immunosuppression and poor prognosis. LPS, HMGB1, and HSPs produced in the GBM microenvironment which act as TLR4 agonists and further initiate signaling to promote PD-L1 expression [188]. High expression of PD-L1 on myeloid cells resulted in shorter OS in GBM patients treated with the HSP peptide vaccine [189]. In glioma-bearing mice, a mixture of DCs, CpG-ODN, and GSC lysate as a source of GSC-associated antigens was applied as vaccine (STDENVANT) to upregulate the expressions of PD-1 and PD-L1 on effector T-cells, DCs, and glioma tissues, leading to enhanced antigen presenting. Combining STDENVANT and anti-PD-L1 antibody attenuated Treg accumulation in the brain and showed an even greater survival advantage [190]. The results suggest that combining TLR activation to boost PD-L1 expression, alongside checkpoint blockade, provides a potential therapeutic benefit in glioma treatment. However, sufficient clinical studies are required to confirm its efficiency and safety in human.

Although limited number of studies are available on TLR immunotherapy against glioma, more results have been obtained from other cancer types and may be used as references. For example, in mice bearing melanoma, activation of TLR1/TLR2 dimer on APCs by agonist diprovocim was reported to induce strong cytotoxic T lymphocytes response to co-administered immune checkpoint inhibitor. This was mediated by proinflammatory cytokine production, antigen-specific eradication of tumor cells, in addition to anti-PD-L1 which further permitted antitumor T-cell activation [191]. In vivo study showed that intratumoral treatment of Poly I:C complexed with polyethylenimine (BO-112, to induce cancer cell apoptosis)-controlled tumor progression in melanoma-bearing mice along with increased PD-1/PD-L1 and CD137 expression. Combinatorial treatment of Poly I:C, BO-112, and checkpoint inhibitors (anti-PD-L1 and anti-CD137 monoclonal antibodies) enhanced the survival benefits through elevation of CD8⁺ T lymphocyte infiltration to the TME [192]. TLR4 expression was closely associated with PD-L1 expression in various cancer types [193‐195]. TLR4-HSP86 axis was reported to induce T-cell PD-L1 expression and mediated tumor-cell-derived extracellular vesicle-generated MDSCs to suppress immune response. Disruption of the TLR4 signaling reduced MDSC in the TME and PD-L1 expression [195]. In non-small lung cancer patients, although no correlation between TLR7 and PD-L1 expression was observed, high TLR7 expression was closely associated with poor clinical response to immune checkpoint inhibitors and resulted in worse PFS and OS [196]. TLR7-dependent overexpression of PD-L1 was reported in monocytes treated with chronic lymphocytic leukemia-derived exosomes followed by production of various proinflammatory cytokines and chemokines [197]. Although TLR7 blockade with immunotherapy shows a therapeutic potential, it is important to be aware that TLR7/8 activation can also induce immunosuppressive MDSC via NF-κB signaling; therefore, inhibition of MDSC is critical for TLR7/8 agonist-based immunotherapy. Both in vitro and in vivo studies in urothelial carcinoma and melanoma showed that triple combination of TLR7/8-based nanovaccine, anti-PD-L1 antibody, and sunitinib (to reduce MDSCs/Tregs) significantly enhanced T-cell response and improved therapeutic efficacy [198].

As TLRs elicit both pro- and antitumor effects, understanding its functional regulations in tumor cells is crucial for the success of TLR-based immunotherapy. More clinical information of TLR-based immunotherapy in multiple cancer types was recently reviewed [199, 200].

Clinical trials revealed that the TLR agonists were generally well tolerated. However, side effects raised from treatments still affect the neurological function and life quality of patients. The most common adverse effects of Poly I:C administration include temporary injection site discomfort, grade 1/2 flu-like symptoms (fever, myalgia, nausea, vomiting, fatigue, and headache) [140, 143], which usually disappear within 1–2 week. Occasional Grade 3 fever was observed (NCT03392545) and was limited to 48 h after treatment. Grade 1 leukopenia was reported in some patients, and no instances of autoimmunity were encountered [140]. Poly ICLC combined with radiation and TMZ in adults with GBM (NCT00262730) showed that grade 3/4 toxicities include neutropenia, leukopenia, thrombocytopenia, and rash, which were considered common treatment-related toxicity. In children, the grade 1/2 adverse events caused by Poly ICLC and GAA vaccination include gastrointestinal toxicity, anemia, lymphopenia, and allergic reaction, while grade 3/4 toxicity was also reported such as urticaria, non-hematological toxicity, and hematologic toxicity [144, 145]. Mild elevation of liver transaminase after Poly ICLC administration was detected both in children and in adults [143]. No severe toxicity or agonist-related death was reported.

The toxicity associated with HSSP-96 administration was also shown to be minimal. Grade 3 fatigue, low fever, and cutaneous pruritus were the most common adverse events, grade 3 hemiplegia happened occasionally, while no grade 3/4 vaccine-related adverse events were noted [157, 158]. In recurrent GBM, HSSP-96 gave rise to grade 3–4 adverse event and subdural hematoma associated with surgical resection [158]. No death event attributed to the vaccine was reported.

Unlike other agonists which were usually applied through skin absorption or muscular injection, CpG-ODN can be administrated subcutaneously, intrathecally, or intracranially [173]. Application of CpG-ODN alone revealed grade 2 common adverse effects such as lymphopenia, anemia, and neutropenia, local erythema at injection sites, fever, and seizure [173]. CpG-28 was well tolerated, transient neurological worsening or fatigue was the most significant toxicity observed, while increased steroid levels and relatively severe arachnoiditis happened occasionally but was well tolerated [172]. Although most clinical trials reported reasonably tolerated toxicity, one needs to keep in mind that most studies were limited by the number of enrolled patients (mostly under 100 or much less), and larger randomized clinical studies need to be performed to overcome the limitation of small sample size and to fully evaluate the efficacy of therapy before one could confidently declare.

Alongside TLR-targeting agonists and immune checkpoints, the roles of other molecules have been investigated in TLR-mediated tumorigenesis, providing a wide spectrum of possible therapeutic options (Table 5). TLR-targeting factors effectively suppress glioma growth in vitro and/or in vivo studies include zoledronic acid [201], annexin A2 [202], ortho-vanillin [75], cannabinoids [114], transcription factor interferon regulatory factor 3 [203], paeoniflorin [204], prosaposin suppression [205], and chloroquine [126]. However, further preclinical studies are required to clarify the therapeutic effects of these agents.

Lastly, it is noteworthy that some microRNAs (miRNA) have significantly elevated expression in glioma tissues and cell lines [206] and are closely associated with TLR-regulated immune responses, suggesting that these miRNAs could be potential targets for modern immunotherapy. For example, long noncoding RNA UBE2R2-AS1 targeted TLR4 mRNA by binding to miR-877-3p, resulting in suppression of glioma cell growth, migration, and invasiveness [207]. A subset of let-7 miRNA that functions as TLR7 ligands was shown to induce microglial release of inflammatory cytokines and antigen presentation to suppress tumor cell migration and glioma growth [208]. Micro-RNA miR-21 limited PDCD4 expression via TLR4-signaling in response to LPS by inhibition of NF-κB and elevated IL-10 production [209]. LPS and miR-34a have been shown to induce the expression of miR-132, miR-155, and miR146a that downregulated TLR4 signaling and suppressed tumor invasion and migration [210, 211]. However, the suppressive roles of these miRNAs in glioma require further clarification. Recent review on TLR-targeted miRNA development is available [212], and a manually curated database NoncoRNA (http://​www.​ncdtcdb.​cn:​8080/​NoncoRNA/​) is accessible to search for potential therapeutic RNA targets in the treatment of human cancers including glioma [213].

Glioma, especially HGG, is presented as a challenging tumor type due to its high invasiveness, short OS, recurrence, and resistance to standard therapies. TLR signaling and their broad spectrum of interactions provide attractive approaches for glioma immunotherapy. Applications of TLR agonists to suppress pro-tumorigenic signaling and undesired GAM accumulation, to modulate TLR-expressing GSCs and immune cells for the creation of antitumor microenvironment, to induce PD-1/PD-L1 expression, all showed how many processes could be directly or indirectly modulated by targeting TLR signaling. Recent clinical studies on TLR-based immunotherapy have presented effective responses in some glioma patients, either used as stand-alone treatment, or combined with radiation therapy, chemotherapy, immune vaccination, or checkpoint inhibitor against the highly immunosuppressive glioma microenvironment. Although positive results have been observed, further clinical observations are required to overcome current deficiencies and create the best therapeutic strategies that can bring new hope for glioma patients. We anticipate that TLR-based treatments in the future will provide promising results for successful treatment of glioma.