TIGIT: a novel immunotherapy target moving from bench to bedside

Abstract

Treatment strategies for patients with advanced solid tumors have traditionally been based on three different paradigms: surgery, cytotoxics (chemotherapy or radiation therapy) and targeted therapies. Immunotherapy has emerged as a novel treatment paradigm in our armamentarium. Unfortunately, most patients still do not benefit from immunotherapy. These patients often have “cold tumors” characterized by a paucity of effector T cells in the tumor microenvironment, low mutational load, low neoantigen burden and often an immunosuppressive tumor microenvironment. TIGIT is an immunoreceptor inhibitory checkpoint that has been implicated in tumor immunosurveillance. Expression of TIGIT has been demonstrated in both NK cells and T cells and plays a role in their activation and maturation. TIGIT competes with immunoactivator receptor CD226 (DNAM-1) for the same set of ligands: CD155 (PVR or poliovirus receptor) and CD112 (Nectin-2 or PVRL2). TIGIT’s role in tumor immunosurveillance is analogous to the PD-1/PD-L1 axis in tumor immunosuppression. Both TIGIT and PD-1 are upregulated in a variety of different cancers. Anti-TIGIT antibodies have demonstrated synergy with anti-PD-1/PD-L1 antibodies in pre-clinical models. Currently, there are multiple first-in-man phase I trials hoping to exploit this new pathway and improve response rates with existing immunotherapies.

Introduction

Treatment strategies for patients with advanced solid tumors have traditionally been based on three different paradigms: surgery, cytotoxics (chemotherapy or radiation therapy) and targeted therapies. Immunotherapy has emerged as a novel treatment paradigm in our armamentarium. At the Huntsman Cancer Institute, 38% of phase 1 trials are now immunotherapy studies. Following the initial success of checkpoint inhibitors in melanoma, indications for this class of drugs have expanded to different tumor types. At least six checkpoint inhibitors have been approved by the FDA for different indications including: melanoma, non-small cell lung cancer, squamous cell carcinoma of the head and neck, Hodgkin lymphoma, bladder cancer and Merkel cell carcinoma [1–6]. More recently, pembrolizumab became the first tissue-agnostic regulatory approval for patients with mismatch repair deficient advanced cancers that have exhausted standard of care [7]. Unfortunately, a vast majority of patients still do not benefit from immunotherapy. These patients often have “cold tumors” (MSI-low colorectal cancer, pancreatic cancer, sarcoma) characterized by a paucity of effector T cells in the tumor microenvironment, low mutational load, low neoantigen burden and often an immunosuppressive tumor microenvironment. Different strategies are undergoing clinical testing to expand indications of immunotherapy to patients with “cold” tumors. This includes a combination of checkpoint inhibitors with targeted therapies, vaccines, radiation or chemotherapy as well as T cell therapies. T cell immunoreceptor with immunoglobulin and immunoreceptor tyrosine-based inhibitory motif domain (TIGIT) is an immunoreceptor inhibitory checkpoint that has been implicated in tumor immunosurveillance [8–11]. Anti-TIGIT antibodies have demonstrated synergy with anti-PD-1/PD-L1 antibodies in pre-clinical models [12–14]. Currently, there are multiple first-in-man phase I trials hoping to exploit this new pathway and improve response rates with existing immunotherapies. Here, we will review TIGIT pathway inhibitors undergoing clinical development as they represent an additional opportunity for combination therapy.

TIGIT pathway

TIGIT was first identified in 2009 as an immune checkpoint that inhibits activation of T cells and NK cells [11, 15, 16]. The TIGIT receptor has an IgV domain, a transmembrane domain and an immunoreceptor tyrosine-based domain [11]. TIGIT has been found to be upregulated in tumor antigen-specific CD8+ T-cells and CD8+ TILs involved in adaptive immune tumor surveillance [17]. TIGIT is also expressed on NK cells, binding of which inhibits NK cytotoxicity directly [16]. This downregulation of NK cell activity occurs through the phosphorylation of immunoreceptor tail tyrosine (ITT)-like domain of the cytoplasmic tail of TIGIT. The ITT-like domain is phosphorylated at Tyr225 and binds Grb2, recruiting SHIP1 to terminate PI3K and MAPK signaling in the NK cell [18]. NK cell maturation is dependent on the presence of the TIGIT receptor and essential for development of self-tolerance [19, 20].

TIGIT competes with immunoactivator receptor CD226 or DNAX accessory molecule-1 (DNAM-1) for the same set of ligands: CD155 (PVR or poliovirus receptor) and CD112 [Nectin-2 or poliovirus receptor-related 2 (PVRL2)] [8]. DNAM-1 is expressed on monocytes, T cells and NK cells [21]. In opposition to TIGIT’s immunosuppressive effects, DNAM-1 enhances cytotoxicity of T lymphocytes and NK cells [22]. DNAM-1’s interactions with PVR and Nectin-2 were found to enhance NK-mediated lysis of tumor cells. This was demonstrated by antibody mediated masking of receptors and its ligands [23]. Without the presence of PVR ligand, NK cells become hypo-responsive [19]. DNAM-1 binds PVR and likely provides positive costimulation to induce interferon-γ (IFN-γ) production [24]. Conversely, TIGIT binding of PVR suppresses IFN-γ production leading to downregulation of NK cells [25].

Multiple mechanisms have been demonstrated to explain TIGIT’s ability to suppress immune function. First described was TIGIT’s role in production of pro- and anti-inflammatory cytokines. Following TIGIT ligation, PVR is phosphorylated which modulates Erk activity and cytokine production via monocyte dendritic cells. This enhances production of IL-10 and diminishes release of IL12p40 [11, 24]. IL10 is an anti-inflammatory cytokine that contributes to T cell exhaustion and immunosuppression [11]. Knock down of TIGIT using RNA interference induced T cell proliferation and increased IFN-γ production [11, 24]. Blocking of TIGIT on cytokine-induced killer cells using an anti-TIGIT functional antibody increased proliferation and cytotoxic targeting of tumor cells expressing PVR and also increased IFN-γ, IL-6 and TNF-α production [8]. This was also demonstrated in siRNA knockdown of TIGIT on CD8+ T cells derived from AML patients which lead to a significant increase in TNF-α and IFN-γ production and decreased susceptibility to apoptosis of CD8+ T cells [26].

A second mechanism in which TIGIT inhibits immunosurveillance is through both competition and direct inhibition of DNAM-1 [13]. Before the discovery of TIGIT, it was known that DNAM-1 binds ligands PVR and Nectin [27, 28]. TIGIT was found to bind PVR with a higher affinity compared to DNAM-1; therefore, outcompeting with DNAM-1 for binding and inducing T-cell suppression [11]. Additionally, it was demonstrated that TIGIT directly interacts on the cell surface with DNAM-1, impairing its ability to homodimerize and thus activate immunosurveillance. TIGIT intrinsically may affect T-cell proliferation even without antigen and antigen presenting cells present. In mouse models, agonistic anti-TIGIT antibodies inhibited T cell proliferation in the absence of any other cell type [15].

TIGIT is expressed on a subset of Foxp3+ Treg cells in isolated human lymphocyte populations [11]. Foxp3 is essential for maturation of naïve T-cell development toward regulatory T cell phenotype [29]. It was demonstrated that TIGIT-deficient T-cells had lower Foxp3+ expression [30]. Additionally, TIGIT+ Tregs co-express an immunosuppressive gene signature including CTLA-4 and PD-1 [30]. Activation of TIGIT+ Treg cells through agonistic anti-TIGIT antibodies induces expression of fibrinogen-like protein 2 (FGL2). Neutralization of FGL2 lead to TIGIT+ Treg-cell suppressive abilities to similar levels as TIGIT− Treg cells. TIGIT- Treg cells, unlike TIGIT+ Treg cells, are able to suppress differentiation of Th2 cells; however, when FGL2 expression is lost; TIGIT+ Treg cells suppress differentiation of Th2 cells similar to TIGIT− Treg cells [30]. Therefore, TIGIT+ Treg cells shift the balance of activity away from proinflammatory Th1 and Th17 cells to Th2 cells via FGL2 allowing for specialized suppressive effects in inflammatory tissue. TIGIT activation in Tregs release a gene program to inactivate CD8+ T cells. In a murine model, mice with TIGIT deficient Tregs resulted in significantly delayed tumor growth compared to wild type Tregs, whereas mice with TIGIT deficient CD8+ T cells had similar tumor growth compared to wild type CD8+ T cells [9]. It was proposed that upregulation of IL-10 production by TIGIT+ Tregs and CD8+ T cells leads to the dysfunctional phenotype of CD8+ cytotoxic TILs [9].

CD96 is a member of the immunoglobulin gene superfamily and is an important receptor that allows adhesive interactions of NK and T cells in immune response [31, 32]. CD96 has been found to have similar immunosuppressant effects as TIGIT, binding to its ligand PVR with an affinity higher than DNAM-1 but weaker than TIGIT [11]. CD96’s effect on tumor surveillance was initially demonstrated in CD96−/− mice, which were found to have fewer lung metastases than DNAM-1 −/− mice injected with melanoma cell lines [33]. Furthermore, they demonstrated that TIGIT−/− knockout mice treated with anti-CD96 mAb also had reduced lung metastases compared to TIGIT−/− knockout alone. Like TIGIT, CD96 is found to be expressed on NK cells and T-cells in humans [32]. It was proposed that CD96/TIGIT/DNAM-1/PVR form a dynamic axis of inhibitory signals from TIGIT and CD96 opposing stimulatory signals from DNAM-1 [34].

TIGIT’s role in tumor biology

TIGIT and PD-1 were found to be upregulated in a 15-gene signature of multiple tumor-associated T cells [13]. Specifically, this was seen in colon, endometroid, breast and renal clear cell carcinoma. The correlation between TIGIT and PD-1 expression in tumor samples suggests both receptors are partners in inducing T cell exhaustion. Furthermore, upregulation of TIGIT and downregulation of DNAM-1 in CD8+ TILs was seen in advanced melanoma patients, and most of these cells co-expressed PD-1 [17]. PVR, a TIGIT ligand, was found to be strongly expressed on tumor cells in patients with cutaneous T cell lymphoma in parallel with TIGIT expression [35]. Additionally, it was found that DNAM-1 expression was decreased on NK cells and CD8+ cells in the peripheral blood of these patients suggesting the antagonistic nature of DNAM-1 and TIGIT in tumor suppression [35].

DNAM-1 appears to serve a role in immunosurveillance for cancer detection. Mice deficient with DNAM-1 had enhanced development of fibrosarcoma and papilloma when exposed to carcinogens [36]. PVR and Nectin-2 appear to serve as backup redundancy with each other as PVR-deficient mice demonstrated compensatory tumor surveillance using Nectin-2 binding [22]. Increased levels of soluble PVR were seen in serum samples from patients with lung, esophageal, gastric and pancreaticobiliary malignancies compared to healthy controls [37]. Counterintuitively, TIGIT was found to be the immunologic-related gene associated with the best relapse-free survival and overall survival in a retrospective study of triple-negative breast cancer [38]. Of note, they were unable to confirm this finding with other datasets.

TIGIT’s role in suppression of circulating T lymphocytes (CTL) was also demonstrated in melanoma. PVR was found to be constitutively expressed in melanoma cell lines. IFN-γ producing cells were found to have higher PD-1 and TIGIT expression and decreased DNAM-1 expression. Overexpression of PVR was found to significantly suppress CTL activation [10]. Furthermore, patients with metastatic melanoma expressed a higher TIGIT/CD226 ratio on Tregs in the tumor microenvironment compared to that seen in the peripheral blood [39]. Subgroup analysis of patients with the highest TIGIT/CD226 ratio that were treated with anti-PD1 and/or anti-CTLA-4 agents had a significantly worse PFS of 2 months versus 12 months (p = 0.039) [39]. Eleven AML patients who received allogenic stem cell transplants were found to have increased TIGIT expressed on CD8+ T cells compared to controls. A higher rate of primary refractory disease was seen in the high-TIGIT expressers [26].

Interestingly, TIGIT’s role in the tumor microenvironment may also be intertwined with the microbiome. Fusobacterium nucleatum, a bacterium indigenous to the oral cavity, was found to be inversely associated with CD3+ T-cell density in colorectal carcinoma tissue, suggesting a role in tumorigenesis [40]. F. nucleatum was found in humans to produce a protein, Fap2, that directly interacts with TIGIT, causing inhibition of NK cell cytotoxicity [41, 42]. Further establishing TIGIT’s role in colorectal cancer, PVR specifically has been found to be overexpressed in colorectal cancers [43].

Synergy of anti-TIGIT therapy with other immunotherapies

Despite the expanding indications for immunotherapy antibodies in the treatment of cancers, most patients still do not respond to these agents. Combination of immunotherapies using different mechanisms of actions are ongoing [44]. Most notably, in CheckMate-067, a randomized phase 3 study in patients with metastatic melanoma comparing nivolumab versus nivolumab plus ipilimumab versus ipilimumab, improved overall survival (OS) was seen in the combination arm compared to ipilimumab monotherapy [45]. However, improved progression-free or overall survival was not demonstrated between monotherapy nivolumab and combination nivolumab and ipilimumab arms (the study was not powered for statistical comparisons between these two arms). Early data from a variety of phase 1 studies testing IDO inhibitors in combination with checkpoint inhibitors suggest that the combination may increase response rates when compared to historical controls [46].

In preclinical trials, anti-TIGIT candidate drug OMP-313M32 demonstrated a statistically significant reduction of tumor volume in human melanoma PDX in humanized NOD scid gamma (NSG) mice [47]. The poliovirus receptor-related immunoglobulin (PVRIG) is a recently discovered inhibitory immune checkpoint expressed on the surface of T and NK cells [48]. In syngeneic tumor models, COM701, a PVRIG inhibitor, increased effector CD8 T cell activation and synergized with PD-1 and TIGIT inhibitor in an in vivo model leading to tumor growth inhibition and increased survival [12].

Inhibitors used in combination, as discussed above, may increase response rates with anti-TIGIT therapies (Fig. 1). Increased expression of TIGIT and PD-1 has been demonstrated in human non-small cell lung cancer and advanced melanoma-associated CD8+ T cells suggesting a role for synergy [13, 17]. In harvested human metastatic melanoma tumors, addition of anti-TIGIT and anti-PD-1 blocking antibodies resulted in increased proliferation of CD8+ TILs with increased degranulation of these cells as compared to those treated with anti-TIGIT or anti-PD1 antibodies alone [17]. Synergy of anti-TIGIT agents has also been demonstrated with anti-PD-L1 antibodies. Mice treated with CT26 cell lines, analogous to human aggressive, undifferentiated, refractory human colorectal carcinoma, showed response to combination anti-TIGIT and PD-L1 inhibitors. This response was not demonstrated with monotherapy of either agent [13]. In melanoma cell lines, TILs had slightly enhanced activation by addition of anti-TIGIT and anti-PD-L1 antibodies, also suggesting synergy [10]. OMP-313R12, a novel TIGIT antibody, induced tumor growth inhibition in a murine colon carcinoma model (CT26 WT) [47]. Combination OMP-313R12 and anti-PD-L1 had statistically significant improved overall survival in mice models as compared to controls [14]. First-in-man clinical trials with anti-TIGIT either alone or in combination with other immunotherapies are ongoing (Table 1).

Fig. 1.

Mechanism of action of ongoing pre-clinical and phase I trial drugs

Table 1.

Ongoing clinical trials in August 2018

Trial sponsor	Agent	Clinicaltrials.gov identifier	Estimated enrollment	Type of trial	Combination immunotherapy
Genentech	Anti-TIGIT monoclonal antibody (MTIG7192A)	NCT02794571	300	Phase I	PD-L1 (atezolizumab)
Bristol-Myers Squibb	Anti-TIGIT monoclonal antibody (BMS-986207)	NCT02913313	170	Phase I/II	PD-1 (nivolumab)
OncoMed pharmaceuticals	Anti-TIGIT monoclonal antibody (OMP-313M32)	NCT03119428	30	Phase I	N/A
Merck	Anti-TIGIT monoclonal antibody (MK-7684)	NCT02964013	240	Phase 1	N/A
Arcus biosciences	Anti-TIGIT monoclonal antibody (AB154)	NCT03628677	42	Phase 1	PD-1 (AB122)
Compugen [49]	Anti-TIGIT antibody (CGEN-15137)	N/A	N/A	Pre-clinical	Anti-PVRIG antibody (CGEN-15029)
Cascadian therapeutics [50]	Anti-TIGIT antibody (CASC-TIGIT)	N/A	N/A	Pre-clinical	N/A

Anticipated adverse events

In pre-clinical models of TIGIT knock out 2e2 TCR transgenic mice, which are known to develop spontaneous autoimmune optic neuritis, all developed experimental autoimmune encephalomyelitis [15]. However, this contrasts with CTLA-4-deficient mice who universally develop fatal lymphoproliferative disease [51]. Preclinical toxicology studies across different TIGIT molecules have shown excellent safety profiles with no toxicities that can be directly attributed to the TIGIT antibody (unpublished data).

Translation of TIGIT to the clinic

Current clinical trials using TIGIT inhibitors alone or in combination with anti-PD-1/PD-L1 therapy are all phase I designed to determine dosing and safety (Table 1). Once we have phase I data of these agents, planning for phase II trials will need to target specific cancers that have the highest likelihood of response. Tumors with higher mutational burden and have subsequently demonstrated response to PD-1/PD-L1 inhibitors (i.e., melanoma, NSCLC, bladder) would be natural targets [52]. Increased expression of TIGIT and synergy between anti-TIGIT and anti-PD-1/PD-L1 antibodies has already been shown in preclinical models of melanoma, NSCLC and colon cancer [10, 13, 17]. As discussed above, colon cancer’s association with the bacterium F. nucleatum and its interaction with TIGIT make it a potential target for anti-TIGIT therapies as well [42, 43]. However, one study looking at human cancers found that some of the highest TIGIT expression was in tumors not associated with response to anti-PD-1/PD-L1 therapy including triple-negative breast cancer, T-cell lymphoma and cervical cancer suggesting we may need to consider a broader range of tumor targets [49].

Potential target populations for TIGIT therapy would be those who failed PD-1/PD-L1 monotherapy in which these agents are already approved. One may hypothesize that anti-TIGIT antibodies would be dependent on presence of significant levels of TILs; therefore, targeting MSI/MMR “hot” tumors that do not respond or lose response to PD-1/PD-L1 may be most efficacious. Approximately, one-third of patients with MMR-high mutations do not respond to anti-PD-1 therapy [53]. The hope would be for a synergistic effect like that seen in the CheckMate-067 trial combining anti-PD-1 and anti-CTLA-4 antibodies where improved response rates in advanced melanoma was demonstrated [45]. However, recent negative data from the ECHO-301/KEYNOTE-252 trial, a randomized phase 3 trial with nivolumab and epacadostat (IDO inhibitor) in metastatic melanoma suggests that perhaps a better understanding of immunotherapy combinations is needed before launching large randomized trials [54]. The role of TIGIT inhibition in “cold” tumors will depend on its ability to potentially overcome T cell exclusion.

Useful biomarkers for predicting response to therapy will be of paramount importance especially as we conduct trials with multiple immunotherapies where we anticipate increased immune-mediated adverse events. Current clinically available biomarkers in the case of checkpoint inhibitors have been unsatisfactory. The most commonly used biomarker, PD-L1 by IHC, has been inconsistent with tumor response to these therapies [55]. TIGIT expression or a potential gene signature predictive of TIGIT driven immunosupppresion could be useful marker of disease that is refractory to standard cytotoxic agents for targeting patients for anti-TIGIT therapies. In AML patients, TIGIT expression in circulating CD8+ T cells correlated with a higher rate of primary refractory disease to cytotoxic therapy [26].

Optimal biologic dosing of targeted therapies such as checkpoint inhibitors does not necessarily correlate with maximum-tolerated dosing (MTD) as it is often the case with cytotoxic agents [56]. Therefore, correlative biomarkers of biologic activity may be more useful than MTD for determining dosing of checkpoint inhibitors. A proposed mechanism of PD-1 inhibitor resistance is through mutations in interferon pathway genes such as JAK1 and JAK2 [57]. Ongoing clinical trials are evaluating an 18-gene signature of IFN-γ pathway activation as a predictor of response to PD-1 inhibition [58]. Given IFN-γ is an indirect marker of TIGIT and PVR binding, IFN-γ concentration could be a good marker of biologic activity of anti-TIGIT agents [25]. Additionally, murine models have shown that qRT-PCR analysis of known markers of CD4+/CD8+ T cell and NK cell activity (CD3e, CD8a, NCR1, IFN-γ, GZMA, CD226) in both the tumor and whole blood showed dose-dependent correlation with anti-TIGIT antibody and would be informative exploratory correlatives in human trials [49].

Whether anti-TIGIT monotherapy results in anti-tumor activity in humans is currently unknown. Pending results of ongoing phase I trials may help elucidate if monotherapy should be attempted in larger phase II trials. In pre-clinical murine models, anti-TIGIT monotherapy has shown independent activity [14, 47]. Furthermore, anti-TIGIT monotherapy has a dose-dependent association with decrease in tumor size and infiltration and activation of CD8+ and CD4+ T cells in the tumor microenvironment [49]. The safety profile of monotherapy compared to dual therapy with anti-PD-1/PD-L1 inhibitors from ongoing phase I trials will also determine if dual therapy should be evaluated in larger phase II trials.

Conclusion

Immunotherapy has rapidly revolutionized the cancer treatment landscape across different tumor types. Although some immune-related toxicities could be potentially life-threatening if not identified early, the toxicity profile of immunotherapy is generally better than the mainstay of cytotoxic agents. In addition, immunotherapy has offered unprecedented long-term responses in advanced and metastatic disease. However, even among sensitive tumor types, a large group of patients will not respond to these therapies. TIGIT has demonstrated great promise in preclinical models as a new target for cancer immunotherapy and potentially may work synergistically to expand the activity of approved checkpoint inhibitors. Synergy of anti-TIGIT and anti-PD1/PD-L1 antibodies has been demonstrated in in vitro and mice models. Ongoing first-in-man phase I trials will help answer if TIGIT has a place in the next generation of immunotherapy.

Abbreviations

FGL2

Fibrinogen-like protein 2

ITT

Immunoreceptor tail tyrosine

MTD

Maximum-tolerated dosing

PVR

Poliovirus receptor

PVRL2

Poliovirus receptor-related 2

TIGIT

T cell immunoreceptor with immunoglobulin and immunoreceptor tyrosine-based inhibitory motif domain

Author contributions

BLS was primary author of the manuscript. IG-L aided with additional content and editing.

Funding

No relevant funding.

Compliance with ethical standards

Conflict of interest

The authors declare they have no conflicts of interest.