TIGIT as an emerging immune checkpoint

TIGIT expression

In both mice and humans, TIGIT is expressed on NK cells and T cells, including CD4 ⁺ T cells, CD8 ⁺ T cells and T _regs 22 , 23 , 24 , 25 . TIGIT expression is usually low in naive cells, but both T cells and NK cells have been shown to up‐regulate TIGIT upon activation 22 . Consequently, in naive mice and healthy individuals, T _regs , memory and activated T cells and NK cells show the highest expression of TIGIT 22 , 25 .

Cell‐extrinsic mechanism

Early studies have suggested a cell‐extrinsic mechanism based on the observation that neither TIGIT‐specific small interfering RNA (siRNA) nor anti‐TIGIT mAbs affected human memory CD4 ⁺ T cell responses to anti‐CD3 stimulation 22 . However, when T cells were cultured with autologous CD11c ⁺ DCs, the addition of anti‐TIGIT mAbs increased T cell proliferation and IFN‐γ production. DCs are antigen‐presenting cells (APCs) that are crucial for the priming of T cell responses 49 . The quality of the T cell response induced depends upon the type of DCs, as well as on their maturation level. Yu et al . demonstrated that TIGIT’s interaction with CD155 modulated cytokine production by DCs 22 . Upon TIGIT ligation, CD155 signalling in human monocyte‐derived DCs led to increased secretion of IL‐10 and decreased secretion of proinflammatory cytokine IL‐12. These data suggested that TIGIT–CD155 interactions promote tolerogenic DCs that down‐regulate T cell responses. Moreover, the ability of TIGIT‐Fc to relieve T cell‐mediated delayed‐type hypersensitivity symptoms in wild‐type but not Il10 ^−/− mice suggested that this mechanism of action is conserved across species. The regulatory role of TIGIT as a ligand that promotes suppressive functions of CD155‐expressing myeloid cells was confirmed by another group 50 . This second study indicated that, in mice, TIGIT promotes the polarization of CD155‐expressing type 1 proinflammatory macrophages into IL‐10‐secreting type 2 macrophages.

Cell‐intrinsic mechanisms

In addition to its role as a regulator of DCs and macrophages, TIGIT also suppresses T cell functions in a cell‐intrinsic manner. Multiple studies have shown that agonistic anti‐TIGIT mAbs inhibit anti‐CD3/anti‐CD28 mAb‐mediated human and mouse T cell proliferation and cytokine production in the absence of APCs 25 , 46 , 51 . Further, a recent study demonstrated that melanoma cells expressing a truncated version of CD155 suppressed CD8 ⁺ T cell IFN‐γ production in a similar manner as cells expressing wild‐type CD155 52 . This indicated that TIGIT–CD155 interaction can inhibit T cell functions without downstream signalling via CD155. For cell‐intrinsic mechanism of action it was hypothesized that, given the high affinity of TIGIT for CD155, TIGIT may inhibit T cells by out‐competing DNAM‐1 for the binding of CD155. This was first suggested by the observation that TIGIT knock‐down in human CD4 ⁺ T cells increased their expression of T‐bet and IFN‐γ, and this could be overcome by DNAM‐1 blockade 46 . Similarly, in another study, TIGIT was found to suppress mouse CD8 ⁺ T cell responses in a DNAM‐1‐dependent manner 45 . In addition, time‐resolved fluorescence resonance energy transfer (TR‐FRET) analysis revealed TIGIT’s ability to interact with DNAM‐1 on the surface of human T cells and to disrupt DNAM‐1 cis‐homodimerization, suggesting another mechanism by which TIGIT interferes with DNAM‐1‐mediated co‐stimulation 45 .

Besides preventing DNAM‐1 signalling, TIGIT can also directly transmit inhibitory signals via its cytoplasmic tail. Most experiments investigating intracellular TIGIT signalling have been performed using the human NK cell line YTS, transfected with human or mouse TIGIT. Two publications from the same group established that the ITIM motif is essential for human TIGIT signalling, whereas mouse TIGIT inhibition can be mediated by either the ITIM motif or the ITT motif alone 24 , 26 . Indeed, while human TIGIT with a mutated or truncated ITIM motif failed to inhibit NK cell cytotoxic activity 24 , mouse TIGIT function was lost only when tyrosine residues in both the ITIM and ITT‐like motifs were mutated 26 . Moreover, another group suggested an important role for the intracellular ITT‐like motif in human TIGIT and highlighted two different signalling pathways interfering with NK cell cytotoxicity or IFN‐γ production 53 , 54 . Following ligand binding, ITT‐like motif becomes phosphorylated and binds cytosolic adapter growth factor receptor‐bound protein 2 (Grb2), which then recruits SH2‐containing inositol phosphatase‐1 (SHIP‐1), leading to the inhibition of phosphoinositide 3‐kinase (PI3K) and mitogen‐activated protein kinase (MAPK) signalling cascade and the down‐regulation of NK cell killing activity 53 . In addition, association of adaptor β‐arrestin 2 with phosphorylated TIGIT mediates recruitment of SHIP‐1 through the ITT‐like motif and impairs TNF factor receptor (TNFR)‐associated factor 6 (TRAF6) auto‐ubiquitination to abolish nuclear factor kappa B (NF‐κB) activation, leading to inhibition of IFN‐γ production 54 . Only one study investigated the role of direct TIGIT signalling in T cells. Whole‐genome microarray analysis on mouse T cells showed that TIGIT engagement suppressed T cell activation via down‐regulating T cell receptor (TCR) expression, together with several other molecules involved in TCR and CD28 signalling 51 .

TIGIT correlates with dismal clinical outcomes in cancer

TIGIT expression on TILs of melanoma patients or on peripheral blood CD8 ⁺ T cells of gastric cancer patients has been associated with metastases development and poor survival 59 , 61 , 68 . Moreover, a strong correlation has been observed between TIGIT expression on peripheral blood CD8 ⁺ T cells and AML relapse post‐transplantation 56 . In endometrial cancer, high levels of TIGIT on tumour‐resident NK cells have been associated with disease severity 65 . Finally, a high TIGIT/DNAM‐1 ratio on tumour‐infiltrating T _regs was shown to correlate with poor clinical outcome following ICB targeting PD‐1 and/or CTLA‐4 63 .

TIGIT blockade in mouse pre‐clinical models

In two independent studies, anti‐TIGIT mAbs used as single agents were found insufficient to suppress the growth of already established subcutaneous tumours in mice 45 , 69 . By contrast, Zhang et al . recently reported that early treatment with anti‐TIGIT mAbs (i.e. started before the tumour being established) delayed the growth of CT26 subcutaneous tumours and methylcholanthrene (MCA)‐induced fibrocarcinomas 64 . The same group showed that anti‐TIGIT mAbs protected mice against 4T1 or B16 experimental metastasis. Interestingly, Zhang et al . proposed that NK cells were involved in the protection observed following TIGIT‐blockade in all these different models. Similarly, we found that blocking TIGIT significantly decreased tumour burden in an aggressive mouse myeloma model (Vk12653) and increased survival in two additional mouse myeloma models, Vk12598 and 5TGM1 58 . However, in opposition to Zhang et al ., we observed only minor TIGIT expression on NK cells in MM‐bearing mice and we found that TIGIT blockade protected against MM in a CD8 ⁺ T cell‐dependent manner. The promising potential of targeting TIGIT in MM was confirmed by another group, who established that anti‐TIGIT mAbs protected mice against Vk12653 myeloma relapse following haematopoietic stem cell transplantation 66 .

Several groups have shown that the limited efficacy of anti‐TIGIT mAbs against established subcutaneous tumours could be overcome by combining TIGIT blockade with other therapies, notably with ICB of the PD‐1/PD‐L1 pathway 45 , 69 . In the MC38 model, co‐blockade of TIGIT and PD‐1 was associated with enhanced effector cell functions of both CD4 ⁺ and CD8 ⁺ T cells compared to either therapy alone; and TIGIT/PD‐1 co‐blockade led to a 100% cure rate 69 . In addition, combination of anti‐TIGIT mAbs with PD‐1 or PD‐L1 blockade induced the regression of CT26 and EMT6 tumours, with most mice experiencing complete response; this phenomenon was dependent upon CD8 ⁺ T cells and DNAM‐1 45 . In this last study, even though single therapies increased CD8 ⁺ T cell IFN‐γ production in the tumour‐infiltrating lymph node, only TIGIT/PD‐L1 co‐blockade significantly increased CD8 ⁺ T cell IFN‐γ production in the tumour. Further supporting the potent anti‐tumour effect of TIGIT/PD‐1 co‐blockade, two groups have shown that this combination therapy protected mice against orthotopically implanted GL261 glioblastoma 69 , 70 . Finally, co‐blockade of TIGIT and PD‐L1 in combination with radiotherapy resulted in 90% cure rates of mice bearing CT26 subcutaneous tumours 71 .