TIGIT Expression on Activated NK Cells Correlates with Greater Anti-Tumor Activity but Promotes Functional Decline upon Lung Cancer Exposure: Implications for Adoptive Cell Therapy and TIGIT-Targeted Therapies

Abstract

Treatments targeting TIGIT have gained a lot of attention due to strong preclinical and early clinical results, particularly with anti-PD-(L)1 therapeutics. However, this combination has failed to meet progression-free survival endpoints in phase III trials. Most of our understanding of TIGIT comes from studies of T cell function. Yet, this inhibitory receptor is often upregulated to the same, or higher, extent on NK cells in cancers. Studies in murine models have demonstrated that TIGIT inhibits NK cells and promotes exhaustion, with its effects on tumor control also being dependent on NK cells. However, there are limited studies assessing the role of TIGIT on the function of human NK cells (hNK), particularly in lung cancer. Most studies used NK cell lines or tested TIGIT blockade to reactivate exhausted cells obtained from cancer patients. For therapeutic advancement, a better understanding of TIGIT in the context of activated hNK cells is crucial, which is different than exhausted NK cells, and critical in the context of adoptive NK cell therapeutics that may be combined with TIGIT blockade. In this study, the effect of TIGIT blockade on the anti-tumor activities of human ex vivo-expanded NK cells was evaluated in vitro in the context of lung cancer. TIGIT expression was higher on activated and/or expanded NK cells compared to resting NK cells. More TIGIT⁺ NK cells expressed major activating receptors and exerted anti-tumor response as compared to TIGIT^- cells, indicating that NK cells with greater anti-tumor function express more TIGIT. However, long-term TIGIT engagement upon exposure to PVR⁺ tumors downregulated the cytotoxic function of expanded NK cells while the inclusion of TIGIT blockade increased cytotoxicity, restored the effector functions against PVR-positive targets, and upregulated immune inflammation-related gene sets. These combined results indicate that TIGIT blockade can preserve the activation state of NK cells during exposure to PVR⁺ tumors. These results support the notion that a functional NK cell compartment is critical for anti-tumor response and anti-TIGIT/adoptive NK cell combinations have the potential to improve outcomes.

Keywords: NK cell; TIGIT blockade; cancer immune cell therapy; checkpoint blockade immunotherapy; lung cancer.

Figure 1

Cytokine-activated or PM21-particle-expanded NK cells highly expressed TIGIT. NK cells were either expanded with PM21 particles from T-cell-depleted PBMCs (PM21-NK cells) or were isolated from PBMCs by negative selection and analyzed for TIGIT expression directly (resting NK cells) or after overnight stimulation with cytokines. TIGIT expression was analyzed on the mRNA level by qPCR and on the protein level by flow cytometry. (A) TIGIT RNA level was on average 5.6 ± 2.6-fold greater in PM21-particle-expanded NK cells compared to resting NK cells (N = 8 donors). TIGIT expression was measured on the protein level by flow cytometry. (B) A representative histogram overlay is shown comparing PM21-NK cells (red) and resting NK cells (black) to isotype control (gray). (C) Summary data from 10 donors showed TIGIT protein expression significantly increased in PM21-NK cells (red triangles) compared to donor-matched resting NK cells (black circles). (D) Flow cytometry was also used to analyze TIGIT expression on NK cells activated by other methods, with dot plots showing examples of TIGIT expression on resting and activated cells as indicated above the graphs (red—anti-TIGIT vs. black—isotype ctrl.). (E) Consistent with PM21-NK cells (triangles), TIGIT expression is also increased in NK cells activated overnight with 1000 U/mL of IL-2 (squares) or 10 µg/mL IL-12, 100 µg/mL IL-15, and 50 µg/mL IL-18 (diamonds) compared to resting NK cells (circles) (N = 5–10 donors). Data are presented as a mean with error bars representing standard deviation (SD), scatter plots with donor-pair lines, or scatter plots with mean and SD. Paired two-tailed Student’s t-tests were used to compare TIGIT expression in PM21-NK cells to that in resting cells and one-way ANOVA corrected for multiple comparisons using a Turkey post hoc test was used to compare TIGIT expression across different NK cell activation methods using GraphPad Prism software v. 9.3.1. p-values are shown as ** if p < 0.01, and **** if p < 0.0001.

Figure 2

TIGIT⁺ NK cells have increased expression of NK cell receptors compared to TIGIT⁻ NK cells and have enhanced anti-tumor function. NK cells from four donors were expanded with PM21 particles from T-cell-depleted PBMCs for 12 days. Expression of NK cell activating and inhibitory receptors was determined by flow cytometry and gated on TIGIT⁻ or TIGIT⁺ NK cells. (A) Representative flow cytometry dot plots with gating are shown comparing NK cells with isotype control (gray) or the indicated receptor-specific antibody (red). The percentages of NK cells expressing (B) activating receptors CD16, NKp30, NKp46, DNAM1, and NKG2D and (C) inhibitory receptors CD96, TIM3, NKG2A, LAG3, and PD1 were determined for TIGIT⁺ NK cells (red triangles) and TIGIT⁻ NK cells (black circles) (N = 4 donors, avg. of duplicates). PM21-NK cells were stimulated with either PVR⁻ or PVR⁺-K562 cells and the percentage of TIGIT⁻ or TIGIT⁺ NK cells expressing IFNγ, TNFα, or CD107a was determined by flow cytometry with (D) representative dot plots shown for stimulation with PVR⁻ for each effector function measured. Significantly more TIGIT⁺-PM21-NK cells (red triangles) produced IFNγ and TNFα, and expressed CD107a after stimulation with either PVR⁻ (E) or PVR⁺ (F) K562 cells compared to TIGIT⁻-PM21-NK cells (black circles) (N = 4 donors, avg. of 2–3 replicates). Data are presented as scatter plots with donor-pair lines. Statistical significance was determined by multiple paired t-tests. p-values are shown as * if p < 0.05, ** if p < 0.01 and *** if p < 0.001.

Figure 3

TIGIT blockade did not enhance PM21-NK cell cytotoxicity against A549 lung tumor monolayers. NK cells were expanded with PM21 particles from T-cell-depleted PBMCs obtained from multiple donors for 14–16 days. Cytotoxicity against A549-NLR lung cancer cells in a monolayer, measured by kinetic live-cell imaging, was not significantly improved in the presence of anti-TIGIT antibodies compared to isotype control. (A) Representative cytotoxicity time courses from one donor are shown in the presence of 0.3:1 NK cells:A549 cells either with isotype control (black circles) or in the presence of anti-TIGIT antibodies (red triangles). (B) Summary plot comparing cytotoxicity at 24, 48, and 72 h for NK cells from multiple donors at 0.3:1 of NK:A549 cells is shown (N = 3). Concentration-dependent cytotoxicity curves are shown for one donor at multiple NK cells:A549 cell ratios at 24 h (C), 48 h (D), and 72 h (E). Statistical significance was determined by multiple paired t-tests and represented as ns if p > 0.05.

Figure 4

TIGIT blockade enhanced PM21-NK cell cytotoxicity against A549 lung tumor spheroids. Expanded NK cells were co-cultured with A549 tumor spheroids for 7 days. NK cell cytotoxicity was determined by kinetic live-cell imaging. Representative images at 10× magnification from the live-cell imaging cytotoxicity assay of NLR-expressing A549 cancer cell spheroids incubated with 10,000 NK cells in the presence of anti-TIGIT or isotype control after 0, 48, 96, and 144 h showed increased NK cell cytotoxicity in the presence of anti-TIGIT antibodies against A549 spheroids (A). Representative cytotoxicity time courses from one donor are shown with isotype control (black circles) or anti-TIGIT (red triangles) antibodies present (B). Summary plot comparing NK cell cytotoxicity from multiple donors at 72 h at 1:1 NK:A549 ratio shows that TIGIT blockade significantly increased cytotoxicity (N = 6 donors, avg. of 2–3 replicates) (C), and relative cytotoxicity increased for each donor when normalized to isotype control (D). Data are presented as scatter plots with donor-pair lines or as mean with error bars representing standard deviation. Statistical significance was determined by multiple paired t-tests. For concentration-dependent cytotoxicity curves, the area under the curve (AUC) was determined and compared by unpaired t-tests. p-values are shown as **** if p < 0.0001.

Figure 5

TIGIT blockade enhanced NK-cell-mediated killing in multiple 3D lung tumor spheroid models. NK cells were expanded with PM21 particles from T-cell-depleted PBMCs obtained from multiple donors for 14–16 days. Expanded NK cells were co-cultured with NCI-H1299-NLR, NCI-H358-NLR, or NCI-1975-NLR lung tumor spheroids for 7 days. NK cell cytotoxicity time courses were determined by kinetic live-cell imaging. Representative cytotoxicity curves from one donor are shown against H1299 (A), H358 (B), and H1975 (C) spheroids either with isotype control (black circles) or TIGIT (red triangles) antibodies present. Summary plot comparing NK cell cytotoxicity from multiple donors at 72 h shows that TIGIT blockade significantly increased cytotoxicity (D) and enhanced relative cytotoxicity when cytotoxicity with anti-TIGIT for each donor was normalized to isotype control (E) (E:T ratios used were 3:1 for NCI-H1299, 1:1 for NCI-H358, and 1:3 for NCI-H1975; N = 6 donors, avg. of 3 replicates). Data are presented as scatter plots with donor-pair lines or as mean with error bars representing standard deviation. Statistical significance was determined by multiple paired t-tests. p-values are shown as * if p < 0.05 and ** if p < 0.01.

Figure 6

TIGIT blockade prevents PVR-mediated NK cell exhaustion during spheroid exposure. NK cells were expanded from T-cell-depleted PBMCs for 14–16 days. Expanded NK cells were co-cultured with A549 spheroids for 7 days in the presence of anti-TIGIT or isotype control. After 7 days of co-culture, NK cells were stimulated with K562 cancer cells with or without PVR expression for 4–6 h in the presence of Brefeldin A and Golgi Stop, and NK cell expression of surface CD107a, IFNγ, and TNFα was analyzed with flow cytometry. Unexposed PM21-NK cells, either unstimulated or stimulated, were used as controls. Additionally, NK cells were selected after co-culture with A549 spheroids and used for RNA extraction and transcriptomic analysis. A schematic of the experiment is shown in (A). IFNγ, TNFα, and CD107a expression on NK cells was determined by flow cytometry, and representative histograms are shown overlaying unexposed stimulated PM21-NK cells (light gray fill with black outline) with PM21-NK cells that were tumor exposed in the presence of a blocking antibody isotype control (dark gray fill with black outline) or anti-TIGIT (red fill and outline) and stimulated with PVR⁺ K652 cells for all shown conditions (B). Unstimulated, unexposed NK cells (open squares), unexposed stimulated NK cells (gray filled diamonds), NK cells that were tumor-exposed in the presence of isotype (dark gray filled circles) or anti-TIGIT antibodies (red filled triangles) are shown for either re-stimulation with PVR⁻-K562 cells or PVR⁺-K562 cells. TIGIT blockade preserved IFNγ (C), TNFα (D), and CD107a (E) expression to levels of unexposed NK cells, when re-challenged with PVR⁺-K562 cells (N = 4–6 donors, avg. of 2 replicates). GSEA analysis of RNA-seq data from 3 donors shows that TIGIT blockade upregulated IFNγ, TNFα, and other related inflammation response gene sets. Summary graphs show upregulated gene sets upon TIGIT blockade based on their −log₁₀(FDR) (F). Data are presented as scatter plots or bar graphs with error bars representing standard deviation. Statistical significance was determined by 2-way ANOVA with p-values shown as * if p < 0.05, ** if p < 0.01, *** if p < 0.001, and **** if p < 0.0001 for comparing each exposure and stimulation condition. Statistical significance was determined by unpaired t-tests with p-values shown as ns p > 0.05, # if p < 0.05, or ## p < 0.01.

Figure 7

Activated NK cells infiltrated into LUAD tumors upregulate TIGIT and correlate with better survival. The TIMER2.0 webserver was used to analyze the lung adenocarcinoma (LUAD) cohort of The Cancer Genome Atlas (TCGA) (n = 515) to determine if there are correlations between NK cell levels, TIGIT/PVR expression, and outcomes. The analysis revealed that TIGIT and its ligands PVRL2 and PVRL4 are significantly upregulated in LUAD tumors compared to normal tissue (A). TIGIT expression positively correlated with activated NK cells, but not resting NK cells, based on absolute CIBSORT deconvolution of immune infiltrates and gene expression in the cohort (B). Additionally, higher levels of activated NK cell infiltration correlated with better survival (C) while higher levels of PVR on tumors correlated with poorer survival (D). Statistical significance in (A) was determined by the Wilcoxon test; p-values are shown as * if p < 0.05 or *** if p <0.001). Scatter plots with purity-adjusted spearman’s rho and p-value are shown in (B). The hazard ratio (HR) and p-value for Kaplan-Meier curves are shown in (C,D).

Figure 8

TIGIT expression in the context of NK cell activation state and TIGIT blockade. TIGIT expression is increased on activated NK cells compared to resting. Activated TIGIT⁺ NK cells have a better anti-tumor response as compared to TIGIT⁻ NK cells. However, chronic TIGIT engagement with its ligands in the tumor microenvironment leads to the functional decline of NK cells, which can be prevented with anti-TIGIT.