TIGIT Enhances Antigen-Specific Th2 Recall Responses and Allergic Disease

T cell Ig and ITIM domain receptor (TIGIT), also known as WUCAM (1) or Vstm3 (2, 3), has been described as an inhibitory molecule expressed on memory T, activated T, NK and regulatory T (Treg) cells (4–6). TIGIT has been found to limit autoimmunity (2, 7) and to restrain T cell responses acting as an intrinsic inhibitory molecule by decreasing TCR activation, proliferation, and cytokine secretion (5, 7). TIGIT expressed on T or NK cells can bind to its high-affinity ligand CD155 (1), also known as poliovirus receptor (8) or Necl-5 (9), expressed on the surface of APCs (4, 10, 11). Engagement of TIGIT to CD155 promotes IL-10 while it restrains IL-12 production by dendritic cells (DCs), leading to reduced T cell activation in vitro (4). TIGIT-expressing Foxp3⁺ T cells have been recently characterized as a highly functional Treg cell subset that selectively suppresses Th1 and Th17 responses (6). In addition, upregulated TIGIT expression by tumor-infiltrating CD8⁺ T cells promotes T cell exhaustion and tumor expansion (12, 13). Likewise, high TIGIT expression by CD8⁺ T cells inhibits antiviral immunity (12). Although TIGIT expression by T cells has suppressive effects on autoimmunity, antiviral, and antitumor immunity, its effects on allergic Th2 type responses remain unknown. Therefore, we addressed the role of TIGIT in Th2 immunity and specifically in Th2 recall responses that lead to allergic asthma.

A recent study demonstrated that the TIGIT locus is hypomethylated, and its gene product is overexpressed in CD4⁺ T cells of pediatric patients with allergic asthma (14). In another study, splenocytes of immunized Cd155^−/− mice secreted lower levels of IL-4 and contained fewer IL-4– and GATA-3–expressing CD4⁺ T cells (15). Of note, a recent report showed that TIGIT-expressing Foxp3⁺ T cells could not suppress allergic asthma upon transfer (6). The above findings suggest that TIGIT may play a role in Th2 immunity, possibly distinct from its role in other types of immunity. In this report, we provide evidence that TIGIT expression by Th cells and its interaction with CD155 enhances Th2 responses, and blockade of TIGIT is therapeutic for experimental allergic airway inflammation.

BALB/c and OVA-specific TCR-transgenic DO11.10 (Tcr-TG-DO11.10) mice were purchased from The Jackson Laboratory (Bar Harbor, ME). Mice were housed at the Animal Facility of the Biomedical Research Foundation of the Academy of Athens. Protocols were approved by the Bioethics Committee of Biomedical Research Foundation of the Academy of Athens and the Greek Government. All procedures were in accordance with the National Institutes of Health Statement of Compliance (Assurance) with Standards for Humane Care and Use of Laboratory Animals (A5736-01) and with the European Union Directive 86/609/EEC for the protection of animals used for experimental purposes.

Mice were immunized with 0.01 mg chicken OVA (Sigma-Aldrich) in 0.2 ml aluminum hydroxide (alum; Serva) i.p. on days 0 and 12. After the two immunizations (on day 18), draining lymph nodes (LNs) of DO11.10 mice were isolated, and CD4⁺CD44^hiCD62L⁺CD127⁺ T memory cells were sorted for cocultures with naive DCs. For allergic airway disease induction after OVA/alum immunizations, BALB/c mice were administered aerosolized OVA (5%, for 20 min) three times on days 18–20. Each mouse received 20 μg affinity-purified blocking polyclonal Ab against TIGIT (AF-7267, R&D Systems) or corresponding Ig isotype control Ab (R&D Systems) three times i.p., 2 h before challenges. Similarly, we used 20 μg/mouse purified monoclonal blocking Abs against TIGIT (1G9), CD155 (4.24.1-LEAF), or isotype control Ig Abs (MOPC-21 and RTK2758-LEAF; BioLegend). For certain experiments, we administered 250 μg/mouse monoclonal IgG2a blocking Ab against TIGIT (clone 10A7) (4, 12), kindly provided by Genentech, or the corresponding Ig isotype control Ab (R&D Systems). Mice were euthanized on day 21. We stained paraffin-embedded lung sections (4 μm) with H&E and periodic acid–Schiff (PAS) to evaluate lung infiltration and goblet cell hyperplasia, as previously described (16). We used a semiquantitative scoring system to grade the size of lung infiltrates as previously described (17). Goblet cells were counted on PAS-stained lung sections using an arbitrary scoring system as described previously (17). Images from H&E and PAS stained sections were obtained by a Leica DM LS2 optical microscope and analyzed with Leica Application Suite V3.6.0 (Leica Microsystems). We harvested, measured, and analyzed bronchoalveolar lavage (BAL) inflammatory cells, as previously described (16, 17). Serum Ab concentration of OVA-specific IgE was measured with an ELISA kit (BioLegend). We used a previously described method to isolate cells from draining LNs (18). Mediastinal LNs (MLNs) from immunized or asthmatic mice were harvested, pooled from each mice group, and stimulated ex vivo with 125 μg/ml OVA (Sigma-Aldrich). Measurements of cytokine levels were performed in culture supernatants, as previously described for LN cells (18). MLN CD4⁺ T cells were isolated with the CD4⁺ T Cell Isolation Kit (Miltenyi Biotec).

Draining LNs from immunized or naive mice were harvested. Naive T cells were isolated from BALB/c mice (for Th polarization) and naive or memory T cells from DO11.10 mice (for cocultures) with the CD4⁺ CD62L⁺ T Cell Isolation Kit (Miltenyi Biotec). FACS sorting (FACSAria III; BD Biosciences) followed for further purification with anti-mouse mAbs: CD3 (145-2C11), CD4 (GK1.5), CD62L (MEL-14), CD44 (IM7), CD25 (PC61), and CD127 (A7R34; BioLegend). DCs were enriched with the CD11c microbead kit (Miltenyi Biotec) and FACS sorting followed with anti-mouse mAbs for: CD19 (6D5) and CD11c (N418). For dead cell exclusion, 7-aminoactinomycin D was used (BD Biosciences). The two-step coculture was previously described by Alpan et al. (19). In our setting, isolated DCs from LNs and spleens of naive mice were cocultured either with T memory cells (Th2) from two times DO11.10-sensitized mice (OVA/alum) or with T cells from naive DO11.10 mice, as control, for 12 h in the presence of 100 μg/ml OVA in a cell ratio of 1:35 DC/T. After 12 h, Th2- or T naive–activated DCs were sorted and cocultured again for 72 h with naive DO11.10 T cells in the presence of 100 μg/ml OVA in a ratio 1:350 DC/T cells. Blocking Abs against TIGIT (1G9), CD155, or isotype control Ig Abs (BioLegend) were used in a concentration of 10 μg/ml in cocultures. For Th polarization cultures, anti-CD3/28 beads (Life Technologies) were added for 96 h together with 20 ng/ml rIL-2 (PeproTech) for Th0, 20 ng/ml rIL-2, 20 ng/ml rIL-12 (PeproTech), and 10 μg/ml purified anti–IL-4 (BioLegend) for Th1, 20 ng/ml rIL-2, 20 ng/ml rIL-4 (PeproTech), 10 μg/ml anti–IFN-γ (BioLegend) for Th2, and 10 ng/ml rTGF-β, 100 ng/ml rIL-6 (PeproTech), 10 μg/ml anti–IFN-γ, and anti–IL-4 for Th17. To restimulate polarized Th cells, new anti-CD3/28 beads were added for 72 h. Commercially available ELISA kits were used for detection of mouse IL-4, IL-13, IL-5, IL-10, IFN-γ, and IL-17 (R&D Systems) in all culture supernatants and BAL fluid.

Freshly isolated MLN cells were stained with combinations of fluorochrome-conjugated Abs to CD3, CD4, CD44, CD62L, CD11c, TIGIT (1G9), CD155 (TX56), IL-13Rα1 (SS12B), T1ST2 (DIH9), ICOS (C398.4A), programmed cell death-1 (PD-1; 29F.1A12) (BioLegend), and CXCR5 (L138D7) (BD Pharmingen). Before intracellular staining, CD16/32 mAb (BioLegend) was used to block FcγRII/III receptors. Intracellular cytokine expression was assessed by 25 ng/ml PMA (Sigma-Aldrich) and 1 μg/ml ionomycin calcium salt (Sigma-Aldrich) for a 5-h incubation and a Cytofix/Cytoperm Kit Plus (Golgiplug; BD Biosciences). For intracellular cytokine staining, cells were stained with surface markers and then with Abs against IL-4 (11-B11), IL-5 (TRFK5), IL-13 (eBio13A), IL-17A (TC11-18H10.1), IFN-γ (XMG1.2), and IL-10 (JES5-16E3) (eBioscience and BioLegend). For proliferation assays, naive CD4⁺ T cells from BALB/c spleens were stained with CellTrace Violet (Life Technologies) and activated with anti-CD3/28 beads for 96 h. All stained samples were analyzed in Attune Acoustic Focusing Cytometer (Applied Biosystems), and the raw data were calculated and visualized with FlowJo Software (Tree Star).

RNA extraction, quantification, and cDNA synthesis from isolated DCs or CD4⁺ T cells were performed as previously described (20). Primers were designed using the Primer 3 program and are as follows: Tigit sense primer, 5′-GTGGGATTTACAAGGGGAGA-3′ and antisense, 5′-CTTCCAGGGGATGAGAGAC-3′; Gata3 sense primer, 5′-TTACCACCTATCCGCCCTAT-3′ and antisense, 5′-ACACACTCCCTGCCTTCTG-3′; Cd226 sense primer, 5′-CAAGAACACTGGCACAAAGA-3′ and antisense, 5′-GAAACAAGCAGGAGTAGATGC-3′; c-maf sense primer, 5′-CCCTTGACAGTTTGCTTCTA-3′ and antisense, 5′-CCCATTCTGCTATCTTTGAC-3′; Rorc sense primer, 5′-TGTTTTATGGGGTTTGGGTA-3′ and antisense, 5′-AAGAGATTGTGTGCCAGAG-3′; and Tbx21 sense primer, 5′-TAGTGATTGGTTGGAGAGGA-3′ and antisense, 5′-GTAGTTCGGGCAGAGAAAG-3′ (Eurofins MWG). Primers for Cd155 were previously described (15). Real-time PCR was performed as previously described (20).

Data were analyzed using Prism Software (GraphPad) and the unpaired Student t test. The p values ≤0.05 were considered significant.

Initially, we polarized isolated naive Th cells toward Th1, Th2, or Th17 and analyzed their TIGIT expression. We observed that Th2-polarized cells harbored the most significant increase in Tigit expression (Fig. 1A). As expected, newly differentiated Th1 and Th17 cells had also upregulated expression of Tigit (Fig. 1A). However, after reactivation, only cells of the Th2 subset upregulated Tigit expression (Fig. 1B). Moreover, TIGIT⁺ Th2 cells were higher in percentages when compared with Th1 and Th17 cells (Fig. 1C). Effective polarization was verified by the relevant expression of signature transcription factors Gata3, Tbx21, and Rorc (Supplemental Fig. 1A) and IL-4, IL-13, IL-5, IFN-γ, and ΙL-17 cytokine secretion for each subset (Supplemental Fig. 1B). Highly purified naive CD4⁺ T cells used in these experiments were CD25⁻, and no Foxp3 expression was detected before and after Th2 differentiation/reactivation (data not shown).

We then investigated whether TIGIT expression in the Th2 cultures could influence their ability to produce lineage-specific cytokines. We found that TIGIT⁺ outnumbered the TIGIT⁻ IL-4– and IL-5–expressing cells in the Th2 cultures (4- and ∼2-fold, respectively) (Fig. 1D). In contrast, in the rest of the Th cell cultures, we noticed significantly less TIGIT⁺ compared with TIGIT⁻ IFN-γ–expressing cell numbers in the Th1 (Fig. 1E), and accordingly less TIGIT⁺ IL-17–expressing cells in the Th17 culture, as measured to the TIGIT⁻ fraction (Fig. 1F). Thus, the levels of TIGIT expression were elevated among Th2-differentiated cells that produced high amounts of IL-4 and IL-5, but no difference was observed in Th1 or Th17 cell cultures. Moreover, sorted TIGIT⁺ Th2 was the cell subpopulation that expressed significantly higher Gata3 and cmaf levels at 72 h of differentiation (Fig. 1G). Although significantly higher percentages of TIGIT⁺ Th2 cells were IL-4⁺ compared with the TIGIT⁻ (Fig. 1H), in IL-5⁺ cells, percentages were similar among TIGIT⁺ and TIGIT⁻ Th2 cells (Fig. 1H). Accordingly, we measured high levels of IL-4 and IL-5 secreted by TIGIT⁺ Th2 cells in the polarization culture (data not shown). Nevertheless, TIGIT⁻ Th2 were activated cells (CD44^hi), similarly to TIGIT⁺ Th2 (data not shown), and capable of producing lineage-specific cytokines. (Fig. 1D, 1H).

Based on these results, we then asked whether allergic mice upregulated TIGIT. We found significantly increased numbers of TIGIT⁺ CD4⁺ T cells in draining mediastinal LNs (MLNs) from mice with Th2-mediated allergic airway inflammation compared with sensitized control mice (Fig. 2A). Allergic mice also exhibited higher percentages of TIGIT⁺ CD4⁺ T cells in MLNs compared with sensitized mice, presenting both effector CD44^hiCD62L⁻CD127⁻ CD4⁺ T and CD44^hiCD62L⁺CD127⁺ central memory CD4⁺ T cells (Fig. 2A). In support, MLN CD4⁺ T cells from allergic mice expressed higher levels of Tigit mRNA (Fig. 2B), and a substantial percentage of TIGIT⁺CD4⁺CD44^hi T cells were expressing IL-4 (Fig. 2C). Of note, the TIGIT⁺ IL-4–expressing T cells of allergic mice were also T1ST2⁺ (data not shown).

Certain reports show that CD155, the high-affinity ligand of TIGIT, is expressed on T cells upon their activation (5) and can mediate a costimulatory signal on CD4⁺ T cells (21). We did not observe any difference in CD155⁺ CD4⁺ T cell numbers and percentages (Fig. 2D) derived from allergic MLNs compared with CD4⁺ T cells from sensitized mice. In contrast, we found that DCs derived from MLNs of allergic mice were CD155⁺, and their numbers were significantly elevated compared with DCs from sensitized mice that were mostly CD155⁻ (Fig. 2E). Percentages of CD155⁺ DCs from MLNs of allergic mice were also elevated compared with sensitized mice (Fig. 2E). These findings are in agreement with studies that demonstrate elevated TIGIT expression by activated T cells and CD155 upregulation by activated DCs (4, 5, 15, 22, 23).

We then activated naive DCs with Th2 memory cells in the presence of Ag in vitro. Th2 memory cells were isolated from LNs of OVA-sensitized mice. Upon activation, Th2 memory cells expressed high levels of Tigit (Fig. 3A). We also observed significantly elevated Cd155 mRNA expression in DCs cocultured with Th2 memory cells (Fig. 3A). In contrast, DCs cocultured with naive CD4⁺CD44⁻CD62L⁺ T cells had almost undetectable levels of Cd155 expression (Fig. 3A), indicating that the significant elevation of Cd155 expression in DCs is correlated with the higher Tigit expression in Th2 memory cells compared with Th naive cells. Subsequent coculture of these CD155-expressing DCs with naive T cells in the presence of Ag led to IL-4, IL-5, and IL-13 cytokine secretion (Fig. 3B). Th cultures induced with CD155-expressing DCs expressed significantly higher levels of Gata3 when compared with the cultures in which DCs had minimal CD155 expression (Fig. 3C). However, the expression levels of both Tbx21 and Rorc remained the same (Fig. 3C). Following their coculture with Th2-activated DCs, the newly differentiated Th2 cells exhibited also significantly increased Tigit expression (Fig. 3C). In contrast, DCs that expressed low Cd155 did not induce high Tigit expression in new Th2 cells (Fig. 3C). Expression of Cd226, the alternative CD155 ligand in Th2 cells, had no difference between the two DC groups (Fig. 3C). Thus, levels of Tigit expression on newly differentiated Th2 cells (Fig. 3C) are correlated with levels of Cd155 expression on DCs (Fig. 3A). Conclusively, the above results suggest that TIGIT on Th cells and its ligand CD155 on DCs are highly abundant under Th2 conditions.

TIGIT has been generally described as inhibitory molecule for T cell responses (4–6). We thus questioned whether elevated TIGIT expression in Th2 cells could possibly act as an inhibitor or an enhancer for Ag-specific Th2 cells in recall responses. We measured Ag-specific Th2 cytokine secretion in the presence or absence of a polyclonal TIGIT-blocking Ab (Fig. 4A) in Th2 cell differentiation cultures, in which activated DCs (by Th2 memory cells) were added. Interestingly, TIGIT blockade restrained IL-4, IL-13, IL-5, IL-10, and IFN-γ secretion by newly differentiated Th2 cells compared with Ig Ab (Fig. 4A). Inhibition of TIGIT binding also restrained Gata3 expression in newly differentiated Th2 cells, whereas Tbx21 and Rorc expression remained in the same levels (Fig. 4B).

Similarly to TIGIT blockade, use of a CD155-blocking Ab had inhibitory effects on Ag-specific Th2 cytokine secretion (Fig. 4C), as well as on Gata3 expression (Fig. 4D). Furthermore, inhibition of TIGIT or CD155 binding also restrained Tigit expression in newly differentiated Th2 cells (Fig. 4B, 4D), indicating a possible feedback loop regulation. In the same cultures, blockade of either TIGIT or CD155 caused upregulation of Cd226 expression, the alternative ligand of CD155, compared with Ig control Abs (Fig.4B, 4D). Moreover, by blocking concomitantly both TIGIT and CD155, we found even lower secretion of IL-4 and IL-13 when compared with cultures in which either TIGIT or CD155 was blocked (Supplemental Fig. 2A). This dramatic attenuation of Th2 activity upon concomitant inhibition of TIGIT and CD155 is possibly explained by the fact that there is no other interaction favored.

Therefore, TIGIT blockade mediates significant downregulation of Gata3 expression in the Th2 setting, implying that TIGIT signaling maintains Th2 responses (Figs. 4B, 4D, 5B). Importantly, neither TIGIT Ab nor CD155 Ab, when compared with Ig control Abs, elevated IFN-γ secretion that could lead to hampered Th2 responses through a switch to Th1 (Fig. 4A, 4C). In different settings, TIGIT ligation enhanced IL-10 production by T cells (5, 7, 24), which was consistent with our observation of reduced IL-10 secretion upon either TIGIT or CD155 blockade (Fig. 4A, 4C). Overall, TIGIT–CD155 interaction is crucial for Th2 induction by DCs upon Ag-specific recall responses in vitro.

We also blocked TIGIT in cultures of naive Th cells that were differentiated in Th2 cells through TCR stimulation and in the absence of DCs. Upon blockade, numbers of both TIGIT⁺/TIGIT⁻ IL-4– and IL-5–expressing cells were significantly reduced (Fig. 5A). As expected, TIGIT Ab caused reduction in Gata3 expression of Th2-differentiated cells (Fig. 5B). TIGIT blockade in the polarization cultures also affected the cell numbers of the TIGIT⁻ IL-4– and IL-5–secreting populations, indicating that possibly a soluble factor produced by TIGIT⁺ cells important for Th2 cell differentiation is inhibited. In contrast, TIGIT blockade in Th1 and Th17 cell polarization cultures had no effect on TIGIT⁺/TIGIT⁻ IFN-γ– (Fig. 5C) or TIGIT⁺/TIGIT⁻ IL-17–expressing cell numbers, respectively (Fig. 5D). These results suggest that TIGIT blockade could effectively attenuate Th2 cell polarization also in the absence of DCs, but has no apparent enhancing effect on Th1 and Th17 differentiation.

As TIGIT is highly expressed by activated Th2 cells (Figs. 1B, 2B, 2C), and CD155 is upregulated in both Th2-activated DCs (Fig. 3A) and in DCs from allergic mice (Fig. 2E), we sought to investigate the potential role of TIGIT in the in vivo Th2-driven responses and allergic airway disease. We thus administered a polyclonal purified blocking Ab against TIGIT in mice developing allergic asthma during allergen (OVA) challenges. Blockade of TIGIT conferred significant suppression of disease, with decreased histological score (Fig. 6A) and lower numbers of BAL eosinophils and macrophages, compared with mice treated with the control Ig Ab (Fig. 6B). Importantly, serum OVA-specific IgE responses were significantly decreased in TIGIT Ab-treated mice (Fig. 6C).

During the effector phase of an allergic immune response, Ig production depends on the help provided by T follicular helper cells (Tfh) to B cells, and Tfh cells are important for IgE-mediated responses in allergic disease (25). Also, Tfh cells can be derived from Th2-committed cells after antigenic challenge in vivo (26) and can be IL-4 producers (27, 28). TIGIT was firstly identified in Tfh cells that express elevated TIGIT compared with naive CD4⁺ T cells (1, 29). TIGIT blockade resulted in significantly decreased numbers of CXCR5⁺ICOS⁺PD-1⁺ Tfh MLN cells (Fig. 6D), consistent with the observed inhibition of Ag-specific IgE production (Fig. 6C).

Moreover, TIGIT blockade restrained the numbers of total CD4⁺CD44^hiCD62L⁻CD127⁻ effector T cells with inflammatory Τh2 phenotype (T1ST2⁺, IL-4⁺, IL-13Rα1⁺) in MLNs (Fig. 6D). More specifically, TIGIT blockade limited the numbers of MLN effector T1ST2⁺ and/or IL-13Rα1⁺CD4⁺ effector T cells, indicating a reduction in IL-33– and IL-13–induced signaling, crucial for allergic asthma (30, 31). IL-10⁺CD4⁺ effector T cells were also decreased (Fig. 6D).

TIGIT blockade also led to significantly lower levels of Gata3 expression in MLN CD4⁺ T cells (Fig. 6E). Accordingly, in vivo TIGIT blockade caused significantly dampened IL-4, IL-13, IL-5, and IL-10 cytokine production in culture supernatants of Ag-specific responses of MLNs (Fig. 6F), whereas no effect on IFN-γ and IL-17 secretion was observed (Fig. 6F). Similarly, BAL fluid from lungs of mice treated with TIGIT Ab had lower levels of IL-4, IL-13, and IL-5. In contrast, IL-10 and IFN-γ levels had no significant differences compared with BALs of Ig Ab-treated mice (Fig. 6G). BAL fluid had undetectable levels of IL-17 (not shown).

We also observed disease suppression when we used a monoclonal blocking Ab against TIGIT (1G9) in vivo. Treatment with 1G9 Ab resulted in decreased histological score (Fig. 7A), lower serum Ag-specific IgE production (Fig. 7B), decreased numbers of proinflammatory CD4⁺ effector T cells (T1ST2⁺, IL-4⁺, IL-13R⁺, and Tfh) in MLNs (Fig. 7C), and reduced CD4⁺ T cell Gata3 expression (Fig. 7D). 1G9 TIGIT Ab also dampened IL-4, IL-13, IL-5, and IL-10 cytokine production in culture supernatants of Ag-specific responses of MLNs (data not shown).

In order to exclude the possibility that the specific TIGIT Abs we use for blockade exert any agonistic activity on T cells, we performed proliferation assays, as previously described (2, 5, 7). Neither of the Abs was able to inhibit proliferation of anti-CD3/28–stimulated CD4⁺ T cells (Supplemental Fig. 2B, 2C). In fact, upon addition of 1G9, proliferation remained the same (Supplemental Fig. 2B), whereas the use of the polyclonal TIGIT Ab enhanced the proliferative capacity of CD4⁺ T cells (Supplemental Fig. 2C). Thus, neither of these two Abs presented any agonistic activity.

To further support our findings, we administered a widely tested mAb against TIGIT (10A7) with established blocking capacity (4, 12). The results obtained were in agreement with previous findings when the polyclonal and 1G9 TIGIT Abs were used (Fig. 6, 7). In fact, upon Ag challenges, mice administered with blocking Ab 10A7 exhibited significantly reduced histological scores, mucus secretion (Supplemental Fig. 3A), and decreased numbers of BAL cells compared with mice treated with the control Ig Ab (Supplemental Fig. 3B). Importantly, serum Ag-specific IgE responses were significantly decreased in TIGIT Ab-treated mice (Supplemental Fig. 3C). Moreover, numbers of MLN CD44^hiCD4⁺ T cells (T1ST2⁺, IL-13⁺, IL-4⁺, IL-5⁺, IL-10⁺, and Tfh) were also lower compared with controls, whereas the population of IFN-γ⁺CD44^hiCD4⁺ T cells had no difference (Supplemental Fig. 3D). 10A7 dampened IL-4 and IL-13 production in culture supernatants of Ag-specific responses of MLNs (Supplemental Fig. 3E), as well as IL-4, IL-13, and IL-5 production in BAL fluid (Supplemental Fig. 3F). In contrast, 10A7 administration had no effect on Ag-specific IFN-γ and IL-17 secretion in MLNs and BAL (Supplemental Fig. 3E, 3F).

Conclusively, TIGIT blockade upon antigenic challenge significantly limited Th2 responses and, consequently, protected from allergic asthma. In the context of Th2 recall responses, in vivo TIGIT blockade did not enhance Th1 and Th17 Ag-specific responses (Fig. 6F, Supplemental Fig. 3D).

Consistent with the in vitro blockade of CD155 (Fig. 4C, 4D), when we used the same blocking Ab against CD155 in vivo, allergic mice showed decreased Th2 responses. Specifically, serum Ag-specific IgE-production (Fig. 7F), numbers of proinflammatory CD4⁺ effector T cells (T1ST2⁺, IL-4⁺, IL-13Rα1⁺, and Tfh) in MLNs (Fig. 7G), and CD4⁺ T cell Gata3 expression were significantly reduced (Fig. 7H) in allergic mice administered with the CD155 Ab compared with those that received the Ig control Ab. Nevertheless, CD155 blockade could not confer significant protection from lung infiltration, as the unaltered histological score between CD155 Ab- and Ig Ab-treated mice revealed (Fig. 7E). Therefore, CD155 Ab administration upon antigenic challenge could limit Th2 responses in a fashion similar to TIGIT Abs.

In agreement with previous reports showing TIGIT expression in activated CD4⁺ T cells (4, 5), we initially observed that upon stimulation, newly differentiated Th2 cells, as well as memory Th2 cells, expressed high levels of TIGIT. Moreover, elevations in TIGIT expression were specifically confined to the Th2 subset upon reactivation. Secondary stimulation of Th2 cells resulted in dramatically higher TIGIT expression, whereas this was not apparent for Th1 and Th17 cells. In vivo, mice with Th2-driven allergic asthma presented high levels of TIGIT expression in a high proportion of effector T and central memory CD4⁺ T cells derived from MLNs. As TIGIT on Th cells interacts with ligands expressed on APCs (4), we asked whether blockade of these interactions could have an effect on Th2 immunity. TIGIT blockade restrained DC and Ag-driven Th2 responses in vitro. Treatment using two different types of blocking TIGIT Abs suppressed Ag-specific Th2 responses and IgE production, as well as eosinophilia and the development of allergic airway disease. Use of the monoclonal blocking TIGIT Ab 10A7 also reduced goblet cell hyperplasia and, consequently, mucus production. Thus, although TIGIT signaling appears to be inhibitory for Th cells and for autoimmunity (2, 7, 32), our findings pointed to a proallergic, Th2-enhancing role for this molecule.

TIGIT ligand CD155 expression in follicular DCs is important for effective Ab secretion (1, 29), and Cd155^−/− mice have reduced Tfh cell numbers as well as defective development of secondary humoral immune responses (29, 33). We observed that, similar to anti-TIGIT, blocking of the TIGIT ligand CD155 during antigenic challenge resulted in significant suppression of Th2 responses. A previous report showed that Cd155^−/− mice exhibited decreased percentages of splenic IL-4⁺GATA3⁺CD4⁺ T cells upon OVA/CpG immunization (15). We also observed that blocking of either CD155 or TIGIT resulted in dampened Ag-specific IgE production, decreased proinflammatory T1ST2⁺ or IL-13Rα1⁺ Th2 and Tfh cell numbers in MLNs, as well as reduced Gata3 expression in MLN CD4⁺ T cells. These results indicated the significance of the TIGIT–CD155 interaction in the Th2 context. In contrast, a recent study showed that primary OVA immunization of Cd155^−/− mice resulted in increased percentages of splenic IL-4– and IL-13–secreting CD4⁺ T cells and higher total serum Ig production compared with wild-type (21). However, immune parameters were only measured after primary sensitization and not following secondary Ag responses. In addition, although the authors mentioned that Cd155^−/− mice had exacerbated airway inflammation, the changes in eosinophilia were modest, and Ag-specific responses were not measured (21). Our findings indicated that CD155 has an enhancing role for peripheral Th2 responses during antigenic challenge. The use of a CD155-blocking Ab (34, 35) in our study provided a clearer view of its significance as a clinical target for airway inflammation. However, although anti-TIGIT inhibited eosinophilia, lung infiltration, and goblet cell hyperplasia, anti-CD155 did not have a significant effect on these features of allergy. It is possible that interactions of TIGIT with other molecules could mediate an important role for lung infiltration.

CD155 blockade inhibits only TIGIT–CD155 interaction, whereas TIGIT Ab blocks the interaction of TIGIT with other ligands, such as CD112 (Nectin-2) (2, 4). CD112 is highly expressed on eosinophils mediating mast cell degranulation, and its blockade resulted in reduced IgE-dependent activation of mast cells (36). Thus, interruption of TIGIT–CD112 interaction could be inducing reduced lung eosinophilia, observed in TIGIT blockade, but not in CD155 blockade experiments. Clarification of the possible role of CD112 in this context could provide new insights in allergy. In addition to TIGIT expressed on T cells, TIGIT Ab blocks TIGIT on other cell types that mediate allergic responses (37). For example, TIGIT is highly expressed on NK cells and is an inhibitor of NK cell cytotoxicity (3, 38, 39). However, the role of TIGIT in NK2 cells that mediate asthma is unknown and merits investigation (40). Overall, TIGIT–CD155 interaction appears to be very important for Th2 cell responses in the allergic model. The possibility of TIGIT interactions with other ligands cannot be excluded that would be important for eosinophilia and lung cell infiltration.

TIGIT blockade suppressed all known hallmarks of allergic airway inflammation. For these experiments, we used a TIGIT-specific polyclonal Ab previously shown to effectively block TIGIT in vitro (38). We alternatively tested a mAb against TIGIT (1G9) that had no agonistic activity (7). We confirmed that both Abs do not exert agonistic activity. In our experiments, 1G9 Ab almost completely inhibited all aspects of allergic response; however, the reduction in histological score and the Ag-specific IgE production were not as dramatic as the ones induced by the polyclonal TIGIT Ab. The polyclonal Ab targets several epitopes of TIGIT molecule, and this could be the reason of the described differences in efficacy. The blocking ability of both Abs (polyclonal and 1G9) were in agreement with the findings obtained by administration of the monoclonal blocking TIGIT Ab 10A7, demonstrating disease attenuation. In fact, administration of 10A7 showed more effective dampening of allergic inflammation that led to significant reduction of goblet cell hyperplasia and mucus secretion.

Moreover, TIGIT activation in Th cells using an agonistic TIGIT Ab has been found to restrain IFN-γ secretion (5) and enhance Il10 gene expression (6). Our experiments using blocking nonagonistic Abs for TIGIT in vivo led to unaltered levels of IFN-γ, as well as downregulated IL-10 secretion in MLNs.

A previous report described a TIGIT-expressing Foxp3⁺ Treg cell subset (6). TIGIT expression by these cells renders them suppressive in autoimmunity but not in allergic disease upon transfer (6). The explanation provided for the lack of efficacy of TIGIT⁺ Treg cells in the Th2 context was that these cells produce Fgl2 that is contributing to allergic responses (41). In addition, TIGIT⁻Foxp3⁺ Treg cells were suppressive for Th2-mediated allergic disease, and this was explained by reduced Fgl2 expression by these cells (6). It is very likely that in vivo blockade of TIGIT has an effect on both Treg and Th2 cells. Our preliminary experiments showed that suppression of allergic disease by either TIGIT or CD155 Ab was correlated with reduction in levels of Fgl2 expression in MLN cells (data not shown). Whether downregulation of Fgl2 expression is confined to Treg and/or Th2 cells remains to be elucidated.

Previously, T memory–DC–Th cell crosstalk was described to mediate oral tolerance (19). Our in vitro experiments demonstrated this axis to be of high importance in the context of Th2 secondary responses. Actually, we showed that DCs, after conditioning by Th2 memory cells, can promote differentiation of new Th cells toward Th2, and this important DC–Th crosstalk was regulated by the CD155–TIGIT interaction. CD155-expressing DCs were highly efficient for induction of TIGIT-expressing Th2 cells. In agreement with Yu et al. (4) showing that TIGIT expressed on human memory T cells interacts with CD155 expressed on DCs, our data demonstrated for the first time, to our knowledge, that CD155 expressed on activated DCs induces TIGIT expression on newly differentiated Th2 cells. As expected, TIGIT operated as a Th2 cell enhancer in vitro, similarly to the in vivo Th2 response.

Also, TIGIT blockade in Th2 differentiation cultures without DC addition inhibited Th2 cell polarization. In contrast, a previous study demonstrated that the use of a TIGIT agonistic Ab in human memory T cell cultures exhibited an inhibitory role of TIGIT signaling for T-bet, Rorc, and Gata3 expression (5). In that study, the authors used human memory CD4⁺ T cells and performed Th2 cultures without IFN-γ blockade, whereas Gata3 expression was measured at an early time point (5). Moreover, blockade of a receptor, compared with its agonistic triggering, does not always lead to opposite effects. For example, deficiency of TIGIT expression by T cells was reported to significantly inhibit IL-10 (5, 7), whereas agonistic activation of TIGIT failed to increase their IL-10 production (5). These results are consistent with our findings in which TIGIT blockade had no apparent enhancing effect on Th1 and Th17 differentiation, whereas agonistic TIGIT triggering has inhibitory effects upon certain conditions (5).

Furthermore, we found another CD155 ligand, Cd226 (or DNAM-1) receptor to be expressed in newly differentiated Th2 cells. Cd226 expression was elevated in Th2 cells in vitro upon disruption of CD155–TIGIT interaction. Blocking of the CD155–TIGIT interaction led, apart from dampening Tigit expression, to enhanced CD226 expression. A recent study described CD226/CD155 interaction on the T cell surface to be Th2 suppressing, as it downregulates IL-4, IL-13 levels, and GATA3 expression in human nonpolarized T cells via enhancement of Th1 and Th17 responses (42). Thus, in our setting, if activation of CD226 by CD155 was inhibitory for Th2 responses, then blockade of either of these molecules would lead to Th2 enhancement. However, we observed exactly the opposite, as blockade of CD155 limited Th2 responses and IFN-γ production. In addition, simultaneous blockade of both TIGIT and CD155 had similar effects to either TIGIT or CD155 blockade and dampened Th2 responses. Thus, it is unlikely that interaction of CD226 with CD155 could restrain Th2 responses in our setting. Supportively, a previous study reported that Th2 differentiation was not regulated through CD226 stimulation (43), and importantly, CD226 blockade had no effect on Th2 polarization (42).

The ITIM domain of TIGIT expressed on NK cells was found to be responsible for inhibition of their cytotoxicity (3). This observation led to the current belief that TIGIT could also act as an inhibitory molecule for T cells, through its ITIM domain, although there is no proof for its functionality in T cells. A study, which sought to investigate whether TIGIT acts as an inhibitory molecule for T cells through its ITIM domain, demonstrated that ITIM domain was not responsible for TIGIT-mediated inhibition for T cell responses and suggested that it was not a functional domain of TIGIT expressed on T cells (4). This conclusion was further supported by a study where gene expression in T cells upon their incubation with agonistic TIGIT Ab was not consistent with ITIM domain-induced inhibition (7). Also, TIGIT engagement did not interfere with the TCR-induced signaling cascade, but downregulated TCR activation pathways by reducing TCR expression (7). Subsequently, TCR downregulation could possibly alter Ag density thresholds required for productive T cell activation, as suggested (32). Overall, it appears that ITIM domain of TIGIT in T cells does not function as expected. Previous studies measured inhibition of T cell activation by TIGIT in other than Th2 contexts, such as delayed-type hypersensitivity, experimental autoimmune encephalomyelitis, and collagen-induced arthritis (2, 4, 7), characterized by Th1 cell induction in which, in contrast to Th2, stronger TCR stimulation operates for Th cell activation (44, 45). In contrast, dampened TCR signaling does not restrain Th2 differentiation, and conditions of reduced TCR activation (such as low antigenic dose) favor Th2 cell responses (44, 45). Thus, it is probable that TIGIT-mediated dampening of TCR activation in the presence of Ag potentiates stronger Th2 responses, and this could be an explanation for the Th2-enhancing activity of TIGIT we found upon recall allergic responses. Moreover, microarray analysis revealed that although TIGIT engagement downregulated TCR expression, this was not inhibiting for several other important cellular processes (7). In fact, TIGIT promoted T cell maintenance and survival by driving expression of several cytokine receptors (IL-2R, IL-7R, and IL-15R) and antiapoptotic molecules (7). These findings explain the capacity of TIGIT to promote expression of activating T cell molecules and are in no contrast to our results that indicate enhancement of Th2 cell activation upon recall responses.

Overall, in this study, we demonstrate that TIGIT has a stimulatory role in Th2 responses and allergic disease, which is distinct from its inhibitory role in autoimmunity and antitumor immunity (2, 7, 13). Our findings suggest a direct effect of TIGIT on Th2 responses, pointing to TIGIT as a potential therapeutic target for asthma. These results may also enhance our understanding of antitumor immunity as TIGIT blockade is used to boost effective responses in cancer (12, 13). Importantly, in the Th2 context, we observed decreased levels of IFN-γ secretion upon TIGIT/CD155 blockade in vitro and stable IFN-γ and IL-17 secretion upon TIGIT blockade in vivo. Therefore, we demonstrated that blockade of TIGIT to lower Th2 responses is a selective strategy that does not enhance concomitant Th1 and Th17 responses, which are very important for maintenance of antiviral, antibacterial, and antifungal immunity.

We thank Genentech for kindly providing the blocking 10A7 Ab. We also thank M. Bessa, F. Gargoulas, I. Skordos, E. Chala, M. Tzioras, E. Christakou, and E. Papaioannou for technical assistance.

J.L.G. is employed by Genentech, a corporation that develops and markets drugs for profit. The other authors have no financial conflicts of interest.