TLRs are membrane-spanning proteins (3) with an extracellular N-terminal domain consisting of 19-25 tandem leucine-rich repeat motifs (2), displaying a horseshoe tertiary structure (30, 31), which is responsible for ligand recognition (2). Next to the ectodomain is a transmembrane region (3), connecting the intracellular portion: the Toll/IL-1 receptor (TIR) domain in the C-terminal tail, which is homologous to IL-1 receptor family’s cytoplasmic region (32) and required for downstream signaling (10).

All the TLRs are synthesized in the endoplasmic reticulum (ER), then transferred to Golgi apparatus, and finally migrate to the plasma membrane or intracellular endosomes (33).

Human TLRs can be classified into two groups by sub-cellular localization: 1) the cell surface TLRs, TLR1, 2, 4, 5, 6, and 10, which reside on the plasma membrane (4), traveling to phagosomes when activated (34), 2) the intracellular TLRs, TLR7, 8, 9, which are expressed on endosomes inside the cells, with the LRR domain facing the inner cavity of the compartment, sampling their ligands (31). TLR3 are also primarily expressed on endosomes, yet it has been reported to express on the cell surface of fibroblasts, epithelial cells, and macrophages as well (35, 36).

In the ligand-binding, namely, activated status, TLRs tend to be dimerized (31). TLR2 alternatively forms heterodimers with TLR1, 6, or 10 (37, 38); each combination is associated with the specificity for recognizing a particular group of ligands, making TLR2 the most competent TLR in sensing a diverse range of PAMPs. Other TLRs tend to form homodimers ( Figure 1 ) (39).

TLRs do not necessarily distinguish between self and non-self, but they respond to “danger signals” in general; as long as the “signals” are encompassed within certain molecular patterns, both no matter whether exogenous or endogenous ligands (40).

TLR ligands (TLRLs) primarily fall into three categories (1): natural exogenous ligands (PAMPs) (2); natural endogenous ligands (DAMPs and secreted ligands) (3); synthesized agents.

Exogenous ligands, namely, PAMPs originate from broadly expressed components in bacteria, fungi, protozoa, parasites, and viruses (10, 41, 42); they are common and genetically conserved because of their significance to microbial survival and infectivity (1, 43). As sentinels of the host’s defense system, TLRs sense PAMPs as invading dangers and directly activate the innate immune system (44), meanwhile indirectly activating the adaptive immune system (43), thereby providing indispensable protection against pathogens. Studies on mice lacking each TLR have shown that each TLR has a different function in PAMP recognition and immune response (45). Cell surface TLRs mainly recognize microbial membrane components such as lipids, lipoproteins, and proteins. TLRs expressed in intracellular components, such as endosomes, recognize viral or bacterial nucleic acids (46, 47). Specifically, TLR1/2 recognizes triacylated lipopeptides from Gram negative bacteria (38, 39), TLR2/6 recognizes diacylated lipopeptide from mycoplasma (38, 48), and TLR2/10 recognizes microbial components shared by TLR1 (37). TLR3 ligands are double-stranded RNAs (dsRNAs) emerging from cytosol virus replication (13, 39). TLR4 binds lipopolysaccharide (LPS) from Gram negative bacteria (39, 49) and fungal mannanes (48).TLR5 ligates with bacterial flagellin (13).TLR7 and 8 ligands are both single-stranded RNAs (ssRNAs) from viruses (13, 39).TLR9 recognizes nonmethylated CpG motifs normally found in bacterial and viral DNA (13, 39).

On the other hand, endogenously derived DAMPs from cells and ECM also activate TLRs. DAMPs are generated from cellular components or fragments of extracellular macromolecules, which are typicallyinaccessible to TLR ligand-binding regions in homeostasis, yet exposed to TLRs in the case of cell necrosis (passive release), apoptosis (pulsative release) or extracellular matrix degradation (5, 16, 43, 50) following trauma, inflammation and tissue remodeling during development (43). Specifically, TLR2 recognizes HMGB1, β-defensin-3, serum amyloid A, and biglycan (51);TLR4 recognizes HMGB1, fibrinogen, saturated fatty acids, and hyaluronic acid fragments (51);TLR3 recognizes mRNA (52);TLR7 and TLR8 recognize ssRNA (53) and antiphospholipid antibody (54);TLR9 recognizes IgG-chromatin complexes (55).

In the tumor microenvironment (TME), DAMPs can be abundantly released by solid tumors (5, 56) and have a significant impact in the TME by adopting mechanisms of tissue repair [such as cell proliferation (57, 58), angiogenesis (56), and tissue remodeling (59)], contributing to tumor progression. Alternatively, just as PAMPs, DAMPs can also act as stimulators of immune responses, showing anticancer potential in vitro as well (56).

Synthetic TLRLs on the other hand have varied levels of similarity to natural ligands (41), designed for the purpose of disease treatment by manipulating different aspects of TLR functions. The synthetic TLRLs are in active pre-clinical and clinical investigation. The advances of TLRLs, including synthetic TLRLs, in therapeutic research are reviewed in the following section 5 and Tables 1 – 3 .

In tumors, TLR2 has been found to enhance T cell immunity (253, 254) and suppress the accumulation of immunosuppressive T regulatory cells (Tregs) (62) and MDSCs (255). TLR2 also induces the anti-tumor M1-like macrophage polarization (66). The immune stimulating potential of TLR2 has been utilized in both basic and clinical therapeutic research.

Recently, there has been pre-clinical data ( Table 1 ) demonstrating tumor inhibition effects by different TLR2L, including triacylated lipopeptides, synthetic chemical compounds, glucomannan polysaccharide, naturally extracted compound, and inactivated virus. In those studies, TLR2Ls are either investigated alone or in combination with tumor vaccines, immune checkpoint inhibition (ICB), chemotherapy, photodynamic therapy(PDT), or adoptive cell transfusion (ACT). In general, enhanced immune activation and suppressed tumor cell viability were observed in these studies ( Table 1 ).

The number of clinical trials with sole TLR2 agonism in tumor immunotherapy is limited. It is noteworthy that Amplivant, the modified Pam3CSK4, can conjugate synthetic long peptides (SLPs) and elicit stronger DC and T cell stimulation in preclinical models compared to Pam3CSK4 (63). In 2021 the first-in-human phase I clinical trial NCT02821494 showed HPV16 E6 SLP-conjugated Amplivant induced robust HPV16-specific T cell immunity with good safety profile (187).XS15 is a TLR1/2 agonist, applied as an adjuvant to a multi-peptide vaccine scheme is in currently active phase I trial NCT04688385 (256).

TLR2 agonism has been applied in chimeric antigen receptor T cell (CAR-T) technology. In trial NCT02822326, a TLR2 Toll/IL-1 domain is incorporated in CAR-T, and the results showed that the tumor-targeting CD19-CAR-T2 cells augmented anti-leukemia responses in relapsed or refractory B cell acute lymphoblastic leukemia (B-ALL) patients with extramedullary involvement, competent in the eradication of extramedullary leukemia cells (257). Another similar phase I trial NCT04049513 is currently investigating the efficacy of 1928T2z CAR-T cells in treating B cell lymphoma.

In general, the Amplivant-SLP compounds are promising for further research for their early clinical data in terms of safety and efficacy of immune stimulation. Also, they have a great potential in treating different tumors as SLPs can be highly variable.

Clinical trials with multi-TLR activation, including TLR2, are discussed in the below sections.

TLR3 actively regulates multiple aspects of tumor development. Either directly or indirectly, TLR3 triggers both innate and adaptive immune responses (258, 259);, it suppresses tumor cell proliferation and promotes tumor cell apoptosis (260–262); it regulates tumor angiogenesis (263); it has also been found to enhance chemosensitivity (70) and radiosensitivity (264). By initiating these mechanisms, TLR3 agonists have shown promising therapeutic value.

In recent years, pre-clinical investigations of TLR3-targeted therapy ( Table 1 ) have focused on combination therapy of TLR3 agonism with other treatment modalities, including chemotherapy, tumor vaccines, and ICB. There have also been studies examining nanoparticle incorporated TLR3 targeted therapies. Poly (I:C) has been used in most studies as the TLR3-targeting agent, while ARNAX, a new DNA-capped dsRNA, has also been applied. These treatment strategies aim to improve immune stimulation and reduce treatment side effects, thus achieving better therapeutic results.

Three TLR3 agonists, Poly ICLC (Hiltonol), Rintatolimod (Ampligen), and BO-112 have progressed for assessment in clinical trials ( Tables 2 and 3 ) as cancer treatments.

Poly ICLC is carboxymethylcellulose, polyinosinic-polycytidylic acid, and poly-L-lysine double-stranded RNA, administratedintratumorally (i.t.), intramuscularly (i.m.), or subcutaneously (s.c.) in cancer treatments of different malignant stages, including primary, recurrent, and metastatic. The various types of cancer involved range from hematopoietic malignancies, skin cancers, sarcoma, squamous cell carcinomas, breast cancer, neurological malignancies, gastroenterological tumors, female reproductive tract malignancies, and urinary tumors ( Tables 2 and 3 ). In these cancer treatments, Poly ICLC is mostly used in combination with traditional therapies (surgery, chemotherapy, or radiotherapy), immunotherapies (tumor vaccine, immune co-stimulation, or ICB), or other synthesized anti-tumor agents. Most of the trials are active or undergoing recruitment. One completed phase II trial (NCT02423863) aiming at multiple solid tumors applied Poly ICLC intratumorally (i.t.) or intramuscularly (i.m.), showing good tolerability with local and systemic immune responses, achieving clinical benefit (188). Two trials investigating melanoma have also been completed with available results: in phase I/II trial NCT01079741, poly ICLC was combined with NY-ESO-1 protein vaccine and Montanide to treat melanoma, effectively activating humoral and T cell immunity in patients. In phase I trial NCT01585350, poly ICLC together with peptide vaccine and incomplete Freund’s adjuvant (IFA) markedly enhanced T cell responses (190). Phase I trial NCT01834248 also combined poly ICLC with NY-ESO-1 vaccine and Decitabine in treating myelodysplastic syndrome, which showed success in the induction of NY-ESO-1-specific cytotoxic CD4⁺ and CD8⁺ T cells as well as CD141^Hi cDCs (265). The completed phase I trial NCT03380871 combined Hiltonol with a tumor vaccine, pembrolizumab, and chemotherapy to treat NSCLC, which showed a good safety profile and immune stimulation. Two completed phase I trials investigating either Hiltonol combined with tumor vaccines (NCT02721043) or Hiltonol with a tumor vaccine and Nivolumab (NCT02897765), all showing good treatment safety profile and immunogenicity. In particular, the latter treatment strategy induced intra-tumoral chemotaxis of cytotoxic T cells (191). The phase I trial NCT02544880 recently completed, where Hiltonol and Mucin1 were combined to make the MUC1 vaccine; together with influenza vaccine and PDE5 inhibitor Tadalafil, MUC1 vaccine has been used to treat solid tumors. Published interim results showed that the combination of MUC1 vaccine and Tadalafil was well tolerated and induced immune activation in HNSCC patients, despite the implication of immune evasion after the treatment was also observed (192). However, trial NCT01532960, where Hiltonol was combined with a tumor vaccine in treating breast cancer, was terminated due to the lack of immune stimulation by of Hiltonol for the vaccine (266). Phase II trial NCT02873819, phase I/II trial NCT02643303, and phase I trial NCT03358719 usingHiltonol in combinatory cancer therapies have recently closed, and are awaiting results. In general, Hitonol is well tolerated and can induce immune responses; however, there is a lack of robustdata presenting its competence in bringing clinical benefit.

Rintatolimod is a modified poly (I:C), PolyI: PolyC12U, administratedintravenously (i.v.) or intraperitoneally (i.p.) in gynecological cancers, breast cancer, and colorectal carcinoma. In clinical trials, Rintatolimod is applied used in combination with chemotherapy, a tumor vaccine, ICB, or other immune-modulatory therapies. Phase I/II trial NCT01312389 and phase II trial NCT03403634 using Rintatolimod have been completed, however the results are not yet publicly available.

BO-112 is a nanoplexed form of poly (I:C), which is currently under investigation in melanoma with pembrolizumab (NCT04570332).

In conclusion, pre-clinical data of TLR3 agonists provide promising perspectives for them to move from bench to bedside, yet more research and trials are needed in optimizing TLR3Ls’ clinical effectiveness.

In tumor studies, TLR4 has been reported to enhance the function of antigen-presenting cells (APCs) (267), increase the production of pro-inflammatory cytokines as well as IFNs (268–270), and boost cytotoxic responses of CTLs and NK cells (267, 271). These immune stimulating mechanisms can be initiated by TLR4 agonists in cancer treatment.

Recently, lipopolysaccharide, lipid A derivatives, polysaccharides, and protein TLR4 agonists were investigated in therapeutic studies ( Table 1 ). These TLR4 agonists have been mostly used in combination with tumor vaccines, including DNA-, peptide- and DC-based vaccines; they have also been used in combinationwith radiotherapy and different kinds of immune-stimulating agents. Notably, many TLR4 agonists are engineered into nanoparticles, which have improved drug delivery efficiency and reduced off-target effects (88–91).

Several TLR4 agonists have completed clinical studies( Table 2 ), including LPS, GLA-SE(G100), IDC-G305, GSK1795091 (CRX-601), and the TriMix DC vaccine.

The completed phase I trial NCT01585350, investigated LPS in combination with a peptide vaccine and incomplete Freund’s adjuvant (IFA) to treat melanoma, which showed that this treatment was well tolerated and stimulated marked T cell responses (190).

GLA-SE is a glucopyranosyl lipid A-stable oil-in-water emulsion, tested in lymphoma, skin cancers, sarcoma, lung cancer, and colorectal cancer, either applied alone or in combination with surgery, chemotherapy, radiotherapy, ICB, or a tumor vaccine. In these trials, GLA-SE is administered i.t. or i.m. A phase I trial NCT02035657 applying GLA-SE in treatment against Merkel cell carcinoma has been completed. Treatment withintratumoral GLA-SE as an adjuvant to surgery and radiotherapy demonstrated safety and feasibility in Merkle cell carcinoma with increased intratumoral infiltration of CD8⁺ and CD4⁺ T cells, activation of immune-related genes, and local tumor regression (193).

IDC-G305 is a combination of NY-ESO-1 recombinant protein and GLA-SE, that has been tested in melanoma, ovarian cancer, renal cell cancer, and non-small cell lung cancer(NSCLC) (NCT02015416). This phase I trial showed that intramuscular administration of IDC-G305 was well-tolerated and can generate antigen-specific immunity (195). On the contrary, another two phase I trials NCT02387125 and NCT02609984, where NY-ESO-1 recombinant protein-GLA-SE was used with a DC-targeting tumor vaccine in treating solid tumors including sarcoma, melanoma, NSCLC, and ovarian cancer, were terminated because of lack of efficacy.

GSK1795091, a synthetic aminoalkyl glucosaminide 4-phosphate, in conjunction with anti-PD-1 pembrolizumab was investigated in advanced solid tumor treatment in trial NCT03447314; the results of this trial are yet to come.

TriMix DCs are autologous DCs incorporating mRNA encoding, CD40 ligand, CD70, and a constitutively active TLR4. The phase I/II trial NCT01530698 showed TriMix DCs, further electroporated with a tumor antigent encoding mRNA, administered intranodally, was safe but with only narrow immunological and clinical response.

In general, recent clinical trials usingTLR4 agonists did show clinical or immune stimulating efficacy in solid tumor treatment. However, there are contradictory results about the same agent in different trials. For those agents that showed promising results in early phase trials, future studies should pay attention to the optimal treatment conditions, including the therapy combination strategy and the route of administration.

TLR5 can be activated by bacterial flagellin and plays a key role in the immune homeostasis of pathogen-host interactions (272, 273). In addition, TLR5 is associated with decreased tumor cell proliferation, invasion, and metastasis (274, 275). Several drugs targeting TLR5 are under investigation.

Recently, protein-based TLR5 agonists have been investigated pre-clinically: in combination with a tumor vaccine (96), incorporated into the adoptive cell therapy (97), or used alone (98). All yielded enhanced therapeutic effects ( Table 1 ).

Entolimod and Mobilan are administrated in two phase I trials ( Table 2 ). Entolimod, namely, CBLB502 is a derivative of Salmonella flagellin. In trial NCT01527136, i.m. or s.c. administration of Entolimod was usedto treat locally advanced or metastatic solid tumors that cannot be removed by surgery. Results of this trial are yet to be available. While having the advantage of avoiding the dangerous inflammatory “cytokine storm” (276), Entolimod has a drawback that it can induce a rapid antibody neutralization (277),leading to its efficacy limitation. Pharmacological modification could be one solution to this problem. Mett et al. (277) identified and eliminated the epitopes leading to Entolimod’s neutralizing immunogenicity and refined its structure to get GP532 as a new TLR5 agonist. GP532 reduced neutralizing antibody response while preserving pro-inflammatory capacity. However, the anti-tumor potential of GP532 has not been verified in any therapeutic studies yet.

To extend the application of TLR5 agonism in TLR5 negative tumors, an adenovirus carrying TLR5 receptor and its agonist’s gene called Mobilan (M-VM3) was designed (276). In preclinical prostate cancer models, Mobilan displayed promising therapeutic efficacy (276).In phase I clinical trial NCT02654938, Mobilan was used intratumorally to treat prostate cancer; no results are available yet. A similar study showed Mobilan was well-tolerated in prostate cancer patients (278); however, more data is needed to verify its clinical benefit.

TLR 7 and 8 serve as important sentinels in response to infection, which makes them irreplaceable in the activation of mammalian innate immune cells (279). These two receptors have also been found to participate in the regulation of the host adaptive immunity (280–283). The emergence of novel small molecular agents targeting these two receptors in the last decade has greatly facilitated cancer immunotherapy studies by adopting their immune activation potential.

Since many agonists exert dual activation to both TLR7 and TLR8, we will summarize TLR7 and TLR8 agonists together in one section.

Following pre-clinical investigations TLR7 or TLR7/8 agonists can be classified by molecular nature into imidazoquinoline and its derivatives, novel synthetic small molecules, and RNA-based agents ( Table 1 ). These agonists are studied either with traditional therapies, such as chemotherapy, radiotherapy, surgery, or novel therapies, such as tumor vaccines, immune checkpoint blockade, adoptive cell therapy, and photothermal therapy ( Table 1 ). It is noteworthy that nanotechnology has become a hot research topic in the drug delivery of therapies applying TLR7/8 agonists. TLR7/8 agonists are incorporated into multi-functional nanoparticles that overcome drug solubility limitations (120, 122) or improve tumor-specific environment-targeting ability (115, 124, 125), thus improving drug delivery efficiency.

A number of TLR7 agonists that have undergone clinical investigation include imiquimod, TMX-101, RNA-LPX, BNT411, DSP-0509, LHC165, SHR2150, TQ-A3334, and RO7119929 ( Tables 2 and 3 ).

Imiquimod has already been clinically applied in the treatment for genital warts (284), low-risk superficial basal cell carcinoma (285), and Bowen’s disease (cutaneous squamous cell carcinoma in situ) (286).In clinical trials initiated in the decade, Imiquimod has been utilized in a wide range of clinical trials targeting different tumor types. There are studies attempting to widen the utilization of Imiquimod based on the current standard of care.

Many clinical studies of Imiquimod focus on Skin cancer. In basal cell carcinoma, a phase III trial (NCT05212246) is going to use Imiquimod as a single treatment in high risk populations to prevent tumor occurrence; imiquimod is also applied in two active trials in combination with curettage (NCT02242929, phase III), and sonidegib together with surgery (NCT03534947, phase II) to treat basal cell carcinoma. In an active phase I trial (NCT03370406), Imiquimod together with 5-fluorouracil is applied to treat squamous cell carcinoma. Two active phase I trial are investigating melanoma, where Imiquimod is combined with either Pembrolizumab(NCT03276832) or tumor vaccine(NCT04072900). A completed phase III trial NCT02385188 found imiquimod together with paracetamol and lidocaine induced 82.6% response rate in vulvar Paget disease but 80% of patients needed painkiller to mitigate side effects (197).

Imiquimod is also studied in intraepithelial neoplasia. In grade 2/3 cervical intraepithelial neoplasia (CIN), a phase II/III trial (NCT02130323) showed that Imiquimod induced lower HPV clearance rate compared to large loop excision (43% vs 64%), which made conization remain the standard care when intervention is needed (199).However, trial NCT03233412 showed that when adjuvant to the loop excision procedure, weekly application of imiquimod can promote regression of high-grade squamous intraepithelial lesions (200). Two active trials are investigating Imiquimod combining HPV vaccine (phase II, NCT02864147) or topical Fluorouracil (phase I, NCT03196180) in treating CIN2/3. In vaginal intraepithelial neoplasia (VIN), the results of the phase III trial NCT01861535demonstrated that: Imiquimod is safe and effective in treating VIN and should be considered as a first-line treatment alternative to surgery (198).In the active phase III trial NCT02059499, single use of Imiquimod is applied in HIV patients who have anal intraepithelial neoplasia.

In glioma, results of phase I trial NCT02454634 recently published in Nature (202) showed that topical Imiquimod combined with IDH1 peptide vaccine induced 93.3% immune responses in IDH1R132H-mutated patients while adverse effects are within grade I. The immune responses induced by the treatment were represented by enhanced IL-17 production and increased IDH1-specific T cells (202). Trial NCT00899574 focused on breast carcinoma; the results showed that single topical use of imiquimod proved beneficial for breast cancer metastasis to skin/chest wall and has been well-tolerated; additionally, this trial proved that imiquimod could induced a pro-immunogenic TME (201). There are also completed phase I/II trials using Imiquimod with radiotherapy, chemotherapy or tumor vaccine to treat breast cancer (NCT01421017, NCT02276300).

In prostate cancer, Imiquimod is applied topically with different tumor vaccines (NCT02293707, NCT02234921). The phase II trial NCT02293707 showed the combination of Imiquimod, telomerase peptide-based vaccine, and Montanide ISA-51 VG was safe in prostate cancer patients. The number of treatments was positively associated with immunological response (204). Bladder cancer is another type of malignancy where Imiquimod displayed its therapeutic potential. TMX-101 is the intravesical formulation for Imiquimod; single use of TMX-101 induced significant increase in urinary cytokines, including IL-6, IL-18, IL-1β, IL-1 receptor antagonist (IL-1ra), and vascular endothelial growth factor (VEGF), in non-muscle invasive bladder cancer, despite accompanied with mild adverse effects (phase II, NCT01731652) (205).

In general, Imiquimod showedits potential in cancer treatment, and also an increasing number of active trials using it has been initiated.

A trial with RNA-LPX (NCT02410733) showed early results with effective quick induction of IFNα and IP-10, and the induction of de novo antigen-specific T-cell responses. In this study, i.v. administration of RNA-LPX proved to be well-tolerated (116). SHR2150 and RO7119929 are oral TLR7 agonists, currently in trials in combination with other therapies in metastatic solid tumors (SHR2150, NCT04588324), and liver/biliary tract cancer (RO7119929, NCT04338685). BNT411 and DSP-0509 are synthetic small molecular TLR7 agonists administered i.v. that are being investigated in multiple types of tumors. LHC165 is a benzonapthyridine TLR7 agonist, which has been used in combination with anti-PD-1 (PDR001) to treat solid tumors.

TLR8 agonist VTX-2337, also known as Motolimod, has been tested in ovarian cancer and squamous cell cancer patients combined with chemotherapy, ICB, or small molecular inhibitors; in most studies, Motolimod was applied subcutaneously. Four studies have since completed with published results. In the first trial, a phase I trial (NCT01294293) with different histological subtypes of ovarian cancer, Motolimod was combined with doxorubicin, showing 15% patients with a complete response and 53% with stable disease (210). A similar phase II trial (NCT01666444), however, showed that the supplement of Motolimod to doxorubicin did not improve clinical outcomes in patients with recurrent or persistent epithelial ovarian, fallopian tube or primary peritoneal cancer, compared with placebo (207). Motolimod was demonstrated to augment clinical responses in patients with HNSCC of advanced stages, who have also been treated with the epidermal growth factor receptor inhibitor cetuximab (phase I trial NCT01334177) (209). However, in a trial of recurrent or metastatic SCCHN patients, the addition of Motolimod to platinum-based chemotherapy and cetuximab did not improve progression free survival (PFS) or OS; nevertheless, remarkable benefit was presented in HPV-positive patients and those with injection site reaction, suggesting the application potential of Motolimod in subset- and biomarker-selected patients (phase II trial NCT01836029) (287). Recently, a strong TLR8 agonist, combined with a HER2 monoclonal antibody was designed (sbt-6050), aiming to activate myeloid cells with the presence of HER2 positive tumor cells, and it is under investigation now in HER2 positive solid tumors with pembrolizumab (NCT04460456).

In terms of dual TLR7/8 agonists, Resiquimod, TransCon, 852A, BDB001, BDB018, NKTR-262, and BDC-1001 have undergone clinical investigation. Resiquimod has been examined by topical use in melanoma and brain tumors in combination treatment with a tumor vaccine and/or other adjuvants, however the addition of Resiquimod to NY-ESO-1 protein vaccine and IFA failed to elicit steady antigen-specific CD8+ T cell response (NCT00821652) (206). TransCon is a type of Resiquimod prodrug, now under investigation in combination with pembrolizumab in solid tumors (NCT04799054). 852A is also an imidazoquinoline TLR7/8 agonist, which has previously been tested in multiple tumors types (including renal cell carcinoma, melanoma, lung cancer, hematologic malignancies, breast cancer, ovarian cancer, and cervical cancer), showing acceptable tolerability, evident immune stimulation, and clinical benefit in some patients (288–291). The i.v. administered TLR7/8 agonist BDB001 and its refined analog BDB018 (designed to enhance immune stimulation while preserving the safety profile of BDB001) are both investigated in solid tumor treatment (NCT03486301, NCT04196530, NCT04840394). BDC-1001 is an immune stimulating antibody conjugate of an anti-HER2 monoclonal antibody with a TLF7/8 dual agonist, currently recruiting a trial enrolling patients with HER2 positive solid tumors (NCT04278144).

In summary, TLR7/8 agonists have drawn most research attention among other TLR ligands; despite that preclinical data indicate promising prospects, none of the TLR7/8 agonists has gained regulatory approval to treat cancer in human except Imiquimod. Future study could focus on optimizing delivery routes, administration schedule, and the vaccine itself (where vaccines are applied) (292).

TLR9 is preferentially expressed on the endosomal membrane of B cells and plasmacytoid dendritic cells (pDCs), and its primary ligand is unmethylated cytidine phosphoguanosine (CpG) oligonucleotides (ODNs) (293). TLR9 activation can lead to potent activation of innate (133, 294, 295) and adaptive immunity (both cell and humoral immunity) (133).

TLR9 agonists previously evaluated in pre-clinical studies are all DNA-based agents containing non-methylated CpG motifs ( Table 1 ). Most of these agonists are CpG ODNs, while some have more complex DNA structures designed for better protection from drug degradation and enhanced TLR9 stimulation capacity (164–166). The therapeutic enhancing effect of TLR9 agonists has been studied with chemotherapy, photodynamic therapy, radiofrequency ablation, vaccines, ICB, and other immune stimulatory agents. ( Table 1 ) Interestingly, in two studies (139, 140) the CpG motif has been used as myeloid cell targeting sequence, incorporated into decoy ODNs that contain STAT3-specific sequence, which can inhibit STAT3 transcription. Both studies showed the decoy ODNs effectively suppressed tumors in animal models of hematological malignancies. Similar to TLR7/8 agonists, many studies have also looked into the design of nanoparticles for TLR9 agonist drug delivery. These versatile nanoparticles have improved the ability for TME-targeting (159), specific cell-targeting (148, 150, 163), and even sub-cellular compartment-targeting (147); they have also protected against drug degradation (152, 156); plus, some conjugated to drug tracers to monitor drug tracking after administration in vivo (152, 155).

Currently the most clinically investigated TLR9 agonists ( Tables 2 , 3 ) are SD-101, CMP-001, IMO-2125. SD-101 is a class C CpG oligonucleotide, applied intratumorally in trials for advanced (metastatic, refractory, or recurrent) malignancies such as pancreatic adenocarcinoma, prostate cancer, liver metastatic uveal melanoma, and multiple lymphoma subtypes. In these early phase trials, SD-101 will be administered in combination with ICB, radiotherapy, small molecular inhibitors, or an anti-OX40 antibody.

Two phase I/II trials applying SD-101 in combination with radiotherapy and immunotherapy in treating lymphoma and other solid tumors just completed, and the results are not yet available. CMP-001, also named as Vidutolimod, is a short DNA piece packaged in protein, mimicking the virus structure. It has been administratedsubcutaneously or intratumorally with ICB, an OX40 agonist, an IgG2 agonist, surgery, and radiotherapy in multiple types of primary and metastatic solid tumors. The completed phase I trial NCT03084640 is completed, which combinedPembrolizumab with CMP-001, and compared the safety together with response rates of CMP-001 administered in different routes in intreating melanoma. Another completed phase I trial NCT03438318 combined CMP-001 with Atezolizumab and radiotherapy to treat NSCLC. In both of the above trials, CMP-001 elicited antitumor-related transcriptional signatures (2022 AACR Abstract #: LB107); more results from trial on CMP-001 are yet to be published. IMO-2125, also known as Tilsotolimod, a synthetic phosphorothioate oligodeoxyribonucleotides, has shown a good safety profile with effective activation of dendritic cells, induction of type I IFN signaling, and up-regulation of multiple immune checkpoint pathways in a phase I trial (NCT03052205) (213), where IMO-2125 was usedas a monotherapy administrated intratumorally in refractory solid tumors. However, a phase III trial (NCT03445533), where IMO-2125 and Ipilimumab were combined for treating melanoma, was terminated for lack of efficacy. Other phase I to II trials with intratumoral administration of IMO-2125 are actively recruiting, either alone or with the combination of ICB in solid tumors ( Table 3 ).

MGN1703 (Lefitolimod) is a DNA-based molecular alternative to a CpG-ODN, administratedsubcutaneously in trials and currently under investigation in combination with anti-CTLA4 in advanced cancers (NCT02668770). In trial NCT02200081, MGN1703 was used with platinum-based chemotherapy in advanced stage small-cell lung cancer where it was well tolerated but showed no impact on the main efficacy of treatment and overall survival of patients (214). DUK-CPG-001 a CpG-based TLR9 agonist is being evaluated in the treatment of myeloid and lymphoid malignancies combined with donor lymphocyte infusion enriched with NK cells in a phase II trial (NCT02452697).

Other TLR9 agonists, all CpG-based, have been assessed in clinical trials, with results available. Inhaled DV281 was tested with anti-PD-1 in advanced NSCLC, with no results publicly available yet (NCT03326752). CpG-MCL vaccine, combining CpG7909 and autologous mantle cell lymphoma-derived tumor cells, was studied to treat mantle cell lymphoma in combination with immunochemotherapy and an autologous hematopoietic stem cell transplant (NCT00490529). The trial showed moderate response to the vaccine among patients (40% developed anti-tumor CD8 T cell response, associated with favorable clinical outcomes), good treatment efficacy (89% were minimal residual disease negative), and a good safety profile (211). CpG7910 was applied in combinationwith local radiotherapy to treat recurrent low-grade lymphoma in a phase I/II trial (NCT00185965), where local administration of this agent systemically induced tumor-reactive memory CD8⁺ T cells. This regimen showed clinical feasibility (212). EMD 1201081 combined with EGFR inhibitor, however, showed no improvement to the clinical efficacy of HNSCC treatment (215).

In summary, despite that there are trials showing the immunogenicity and good tolerability of CpG-based agents, low translational rate remains an issue. This may due to the difference in TLR9 expression in animal models and humans (296). Pre-clinical models of human origins are necessary, such as primary tumor organoids. Also, one safety concern is that CpG motifs are potential to induce autoreactivity (297). There need to be trials with larger sample sizes, longer follow-up durations, and stratified patient subgroups to verify the safety and the suitable patient groups that can benefit from CpG ODN adjuvanted therapies.

In addition to single TLR activation, multi-TLR agonism is also under both pre-clinical and clinical investigation for many cancers. Multi-TLR agonism can be realized by a single agent or by a combination of several TLR agonists.

Preclinically, the most studied treatment modality with multi-TLR activation is with tumor vaccines ( Table 1 ). There is also a trend of applying nanotechnology in drug delivery in multi-TLR stimulating treatments for improved APC-targeting (173, 179) and drug tracing (170).

In clinical trials, Bacillus Calmette-Guérin (BCG), an attenuated live Mycobacterium bovis, is the most studied multi-TLR agonist. It can activate TLR2, TLR4, and TLR9 (298); it was originally applied as a vaccine against tuberculosis and later broadened its application as an adjuvant in the standard care of a part of bladder cancer patients (298). Currently, BCG is still investigated in bladder carcinoma patients in combination with other drug treatments, including an anti-PD-1/PD-L1 antibody, small molecule inhibitors, and a neoantigen encoding gene vaccine. A completed trial NCT02753309 combined rapamycin with BCG in bladder cancer treatment, which showed good tolerability in high-grade non-muscle invasive bladder cancer patients and induction of antigen-specific γδ T cell response as well as urinary cytokine production (219). BCG has also been assessed for the treatment of upper or lower urinary tract carcinoma. In treating urinary tract urothelial carcinoma, the trial NCT00794950 investigated combination therapy of BCG and sunitinib, a novel anti-angiogenesis drug. This trial has since been completed, and results showed the combination treatment was associated with a lower rate of progression and recurrence with an acceptable safety profile (218). Besides malignancies in the urinary system, BCG combinedwith chemotherapy, radiofrequency ablation, and GM-CSF have been evaluated in the treatment of liver metastatic colorectal cancer (NCT04062721).

OM-174 is a lipid A analog, which activates both TLR2 and TLR4. In a phase I trial (NCT01800812) with solid tumor patients, OM-174 as a monotherapy was well tolerated and proved to be effective for the induction of IL-8, IL-10, TNF-α, and IL-6 (despite that TNF-αand IL-6 decreased progressively in repeated treatment). Progressively increased NK cells and NK cell activity were observed in patients who received the highest dose, 1000 μg/m² (217). AS-15 is an adjuvant system, comprised of a TLR4 agonism system (MPL and QS-21) and a TLR9 agonist (CpG 7909) in a liposomal formulation. In the completed phase II trial NCT01266603, AS-15 was combined with the MAGE-A3 tumor antigen and high dose IL-2 to treat metastatic melanoma. This treatment achieved 25% response rate, which was associated with increased infiltrating T cells in the primary tumor tissue (222).

TLR3 agonist poly-ICIC and TLR7/8 agonist Resiquimod have been combined with tumor vaccines in the treatment of melanoma (NCT02126579) and advanced tumors refractory to conventional treatment (NCT00948961). In the completed trial NCT00948961, poly-ICIC and/or Resiquimod combined with NY-ESO-1-targeting DC cell vaccine. The combination therapy induced NY-ESO-1–specific IgG titers in 79% patients and NY-ESO-1–specific T cell responses in 56% patients, leading to disease stabilization and tumor regression in 15/70 patients (220).

A phase I trial NCT00068510, where poly-ICLC or imiquimod used in conjunction with an autologous tumor-pulsed DC vaccine, has reported that this regimen was safe and effective in prolonging survival for glioblastoma patients (221).

Considering the versatility of TLRs in cancer, some of which can promote tumor growth, there is also rationale for exploring TLR antagonists as a cancer therapy.

In several types of cancers arising from the digestive system, pre-clinical studies have demonstrated the therapeutic potential of TLR antagonists. OPN-301, an inhibitory anti-TLR2 antibody, suppressed tumor initiation and growth in a gastric animal model, which was associated with gene suppression of CXCL2 and TNF-α (183). Another two studies reported that IRS-954, an inhibitory DNA sequence against TLR7 and 9, and chloroquine, which inhibits TLR7 and 9 activation, significantly impeded tumor development and growth in vivo and in vitro in liver cancer (184) and cholangiocarcinoma models (185).

A subset of diffuse large B cell lymphoma patients carry the MyD88 L265P mutation, associated with overactivation of the TLR7/9/MyD88 pathway (186). A TLR7/8/9 antagonist IMO-8400 was previously usedin this subset of lymphoma patients in a completed clinical trial NCT02252146, which, showed limited therapeutic efficacy. Similarly, trial NCT02092909, studying IMO-8400 in treating Waldenstrom’s macroglobulinemia, also showed insufficient efficacy of IMO-8400, and hence it was terminated.

Recently, another synthesized ODN-based TLR7/9 antagonist HJ901 was shown to significantly inhibit tumor cell proliferation and tumor growth in cell lines or animal models carrying the MyD88 L265P mutation (186). However, these data warrant further investigation before it can be translated to the establishment of an early phase trial.

There are obviously fewer published studies on anti-cancer TLR antagonists, which generally showed less efficacy than TLR agonists.

As TLRs regulate a wide range of functions, initiating signaling pathways that may lead to both anti-tumor and pro-tumor effects, there is a concern regarding whether the application of TLR agonists will promote tumor growth or whether the anti-tumor effect will be counteracted by the pro-tumor effect they simultaneously initiated. For example, Theodoraki and colleagues have found that TLR3 agonists stimulate the TLR3-TRAF3/IRF3 pathway, which leads to the production of CTL attractants and activate MAVS/helicase pathway, which elicits Treg chemotaxis (299).

Two directions may be feasible for further investigation into this challenge. On one hand, choosing the agonist that specifically activates the anti-tumor pathways but leaves no effect on the pro-tumor pathway could help ( Figure 3 ). Theodoraki et al. (299) further discovered the difference of pathway stimulation potential in different TLR3 agonists: dsRNA Sendai Virus, Poly (I:C), and Rintatolimod. Rintatolimod can activate the TLR3 pathway only, resulting in the induction of CTL attracting cytokines IFNa, ISG-60, and CXCL10, and avoid the initiation of MAVS/helicase pathway, which then prevents Treg accumulation. In the future, more focus to elucidate how drug conformation of TLR-targeting agents influences downstream signaling is necessary, which can then guide future drug synthesis or discovery.

On the other hand, combination therapy is of utmost significance in TLR-targeted therapy. Combined therapies can enhance therapeutic effects in many cases and potentially diminish the pro-tumor impact elicited by TLR agonists. Feng et al. (300) discovered that poly (I:C) could inhibit HCC cell proliferation and lead to cell apoptosis while promoting cell migration and invasion at the same time. When combined with the toad venom constituent bufalin, cell migration and invasion were inhibited, whereas cell apoptosis and suppressed proliferation remained unaffected (300), indicating the potential of combination therapy of the two in eliminating poly (I:C)-initiated pro-tumor effect.

Despite numerous promising results emerging from pre-clinical research with TLR-targeted therapies, many failed to advance to clinical implementation. Lack of efficacy and toxicity has been an issue for TLR-targeting therapy (28, 301).

To address this, first, precise identification and investigation of patient sub-populations that respond to TLR-targeting strategies is important ( Figure 3 ). Some treatments may not have shown significant efficacy in the overall study population, but do in a subset of patients. Monk et al. (207) and Ferris et al. (208) both found in their clinical trial studies that adding Motolimod to chemotherapy did not improve survival in the overall patient population. In contrast, the sub-population with injection site reactions did significantly benefit from Motolimod. The evidence above suggests 1) lack of significance in the overall analysis does not reject the potential of a certain TLR-targeting drug; 2) there are certain groups of patients that may be more responsive to TLR-targeting agents than others. Future work calls for more translational studies incorporated into the trial designs to examine any associations between treatment responsiveness and patient characteristics, including gene expression patterns, biomarker levels, local treatment reaction, etc., thus providing clues for the screening of the target patient population.

Secondly, both lab and clinical research data support that the optimal treatment potency of immunotherapy requires combinatory approaches (27, 28), which continues to be the trend for TLR-targeted therapy. As discussed in section 5 ( Tables 1 – 3 ), many studies showed that the combination of TLR agonists with other therapies enhanced tumor retardation compared to a single treatment alone. However, there is still a gap in our knowledge regarding the timing and dosing to be applied to each treatment to achieve the optimal therapeutic result. For example, in the combinatory treatment of poly (I:C) and radiation in Lewis lung carcinoma models, Yoshida et al. (264) found using poly (I:C) one day ahead of radiotherapy yielded better tumor suppression than to use after radiation.

Pharmaceutical modification of targeting agents has great potential to improve efficacy and lower toxicity as well. Furthermore, polymer and nanoparticle formulations incorporating TLR agonists are promising avenues to explore as they may potentially increase the specificity of targeting and prevent the drug from early degradation, thereby reducing systemic toxicity. Those formulations can now be designed to load multiple agents and be multi-functional ( Figure 3 ), including but not limited to pH-specific drug-releasing (115, 124), in vivo drug tracing (152, 155), and tumor hypoxia relief (106). The advancement of nanotechnology utilizedin TLR-targeted therapy has been reviewed comprehensively elsewhere (302). However, careful attention should be paid to these new biomaterials as the bioactivities they may incur in vivo remain to be fully elucidated. Meanwhile, reducing the financial and timing costs is also a formidable challenge to face.

Besides refinement of the drug delivery system, choosing a localized delivery route may also have an influence on efficacy. This can avoid systemic inflammation to a great extent and improve intratumoral efficacy (300), especially for solid tumors.

TLR tolerance arises due to the unresponsiveness or hyporesponsiveness of TLRs upon TLR agonist stimulation after repeated, prolonged, or chronic activation, which also includes TLR cross-tolerance, where a pre-used TLR agonist incurs tolerance of another TLR (303, 304). There have been studies showing TLR tolerance in TLR2, 3, 4, 5, 7/8, and 9 (304–309). This mechanism physiologically prevents uncontrolled inflammation and autoimmunity, thus protecting healthy tissues from inflammatory damage (303, 304). In cancer, however, such tolerance leads to impaired tumoricidal effect.

To overcome this phenomenon, Bourquin et al. (310) found cycles of repeated low dose TLR7/8 agonist injection within 24h with intervals every five days, compared to one high dose injection every three days, significantly suppressed tumor growth in a mouse model (310), indicating the potential of circumventing TLR tolerance by adjusting treatment dose and time. Similarly, Tsitoura et al. (311) found the minimum and optimal dosing interval to maintain a TLR agonists’ pharmacological responsiveness. Future research needs to further explore the cyclical changes in specific TLR-TLRL induced responsiveness and consider cross-tolerance as new cancer therapies incorporating multi-TLR agonism continue to emerge ( Figure 3 ).