Differential Expression of Immune Checkpoints TIM-3, LAG-3, TIGIT, and Siglec-7 on Circulating Natural Killer Cells – Insights from Healthy Donors Compared to Gastric Cancer Patients

Abstract

Introduction

The complex, multifactorial nature of gastric cancer presents significant challenges in the development of effective immunotherapies. Targeting immune checkpoints has emerged as a promising strategy, with blockade therapies demonstrating clinical success. However, resistance in a subset of patients emphasizes the need for alternative approaches. Exploration of novel immune checkpoints, particularly on natural killer (NK) cells, could enhance the efficacy and potency of immunotherapy, offering new avenues for overcoming resistance and improving patient outcomes. NK cells are crucial in the primary defense against viral infections, tumor development, and metastasis. The cytotoxic function of NK cells is finely regulated by a complex array of activating and inhibitory receptors, including checkpoint receptors. Malignantly transformed cells can impair NK-cell activity by expressing soluble or membrane-bound checkpoint ligands, thereby modulating immune responses to support tumor progression.

Methods

To investigate this dilemma, we simulated in vitro activation by NK-cell co-incubation with K562 cells and analyzed expression of TIM-3, LAG-3, TIGIT, and Siglec-7. After that, we analyzed the checkpoint expression of circulating NK cells from 35 healthy donors and compared it to their expression in patients with gastric cancer (n = 21) using flow cytometry.

Results

In healthy donors, we observed that 25–97% of all circulating NK cells expressed TIM-3, TIGIT and Siglec-7, while only a small fraction of 0.6% expressed LAG-3. Co-incubation of peripheral blood mononuclear cells from healthy donors with K562 cells resulted in heightened expression levels of TIM-3 and TIGIT on NK cells. Conversely, NK cells in patients with gastric cancer showed an increased LAG-3 and reduced Siglec-7 expression.

Conclusion

Our findings suggest the potential of LAG-3 as a next-generation checkpoint molecule, alongside Siglec-7. Especially targeting the sialic acid-Siglec-7 axis may offer promising therapeutic strategies for various cancer types in the future.

Introduction

Gastric cancer represents the fifth most prevalent form of cancer and the third most significant cause of cancer-related mortality [1]. The advent of checkpoint immunotherapy has significantly advanced the field of cancer treatment by augmenting the immune system’s capacity to combat tumors. However, the issue of resistance to immune checkpoint inhibitors (ICIs) remains unresolved, impeding the utilization of these medications for a considerable proportion of cancer patients. The identification of effective strategies to address this resistance could facilitate a more profound comprehension of the underlying mechanisms and, consequently, enhance the efficacy of therapeutic interventions [2].

In 2018, the Nobel Prize in Physiology and Medicine was awarded to James P. Allison and Tasuku Honjo for their seminal research contributing to the development of ICI immunotherapies (ICTs) [3–5]. Indeed, the treatment of malignant neoplastic diseases such as gastric cancer, Hodgkin lymphoma, metastatic melanoma, and many others with ICIs (also called immune checkpoint [IC] blockers) has been proven to be clinically effective (recently reviewed in [6, 7]), leading to the approval of IC for over 43 different indications by the US Food and Drug Administration (FDA) [8, 9]. To date, all FDA-approved ICIs comprise monoclonal antibodies [9], among which ipilimumab was the first to be approved by the FDA in 2011, followed by nivolumab and pembrolizumab in 2014 [10].

Throughout tumor progression, the immune system permanently selects more and more resistant tumor subclones based on a process called cancer immunoediting [11]. Consequently, such aberrant cells acquire, due to genetic instability, immune tolerance and suppressive abilities to evade their demise. The expression of IC ligands that attenuate the activation and function of immune cells is such a mechanism. Normally, under physiological conditions, upregulation of ICs prevents autoimmunity by keeping immune responses in check [10]. Accordingly, the mode of action of checkpoint inhibitor-based immunotherapies is to unchain the patient’s own immune system to attack aberrant tissues by blocking the immunosuppressive signals provided by cancer cells via the administration of ICIs [6]. For instance, ipilimumab binds to the IC molecule “cytotoxic T lymphocyte antigen 4” (CTLA-4), thereby preventing a competitive inhibition of the CD28 costimulatory receptor for the ligands B7-1 (CD80) or B7-2 (CD86) expressed on tumor cells [12]. Another relevant IC molecule, which can be blocked by pembrolizumab or nivolumab, is the PD (programmed death)-1 receptor expressed on immunocompetent cells, which binds the programmed death-ligand (PD-L)-1 on cancer cells [12].

While ICIs targeting the PD-L1/PD-1 and CTLA-4 pathways have demonstrated significant clinical success, a considerable subset of patients remains resistant to these therapies [13]. This underscores the need to explore alternative ICs, particularly those expressed on natural killer (NK) cells, to enhance the effectiveness of immunotherapy. Furthermore, combining NK cell-targeting ICTs with established checkpoint inhibitors that modulate T-cell responses, such as those targeting the CTLA-4 or PD-1 axis, may enhance anti-tumor effects. This dual-checkpoint inhibition approach, including T and NK cells, has the potential to generate a more robust and synergistic immune response against malignancies [14]. So far, only ICIs blocking these two pathways have been FDA approved. However, there are many more receptors expressed on immune cells also functioning as ICs. In this study, we have focused on the IC receptors called lymphocyte activation gene 3 protein (LAG-3), T-cell immunoglobulin and mucin domain-containing protein 3 (TIM-3), T-cell immunoreceptor with Ig and ITIM domains (TIGIT), and sialic acid-binding immunoglobulin-like lectin 7 (Siglec-7), which are present on the surface of NK cells, and analyzed the expression of these ICs via flow cytometry. To our knowledge, it has not yet been investigated whether the expression of the named ICs on circulating NK cells differs between patients with gastric cancer and healthy donors. Moreover, these four ICs have diagnostic potential regarding cancer patients [15]. Since monoclonal antibodies against LAG-3 [16], Siglec-7 [17], TIM-3 [18], and TIGIT [19] are under development and/or the subject of various clinical trials, these ICs are also considered potential targets for ICT.

As cytotoxic immune cells against infected cells and cancer cells, NK cells play a vital role in the body’s immune defense. They are considered to be part of the innate immune system and function as a first line of defense against viral infections or cancer [20]. Human NK cells are recognized as cells that are CD56-positive while, at the same time, CD3-negative. Typically, they can be further distinguished into two subgroups depending on the expression level of CD56 (CD56^dim or CD56^bright) on the cell surface [21]. Moreover, they make up to 13 percent of the peripheral blood lymphocytes in humans [22]. As cytotoxic cells, NK cells have two major modes of action to kill infected or malignantly transformed cells: (1) targeted secretion of lytic granules, which contain granzymes and perforin, and (2) inducing apoptosis through death receptors [23]. The first mode of action can also be triggered via antibodies bound to target cells (antibody-dependent cellular cytotoxicity) [24]. The activation and effector function of NK cells rely on signals from activating and inhibitory receptors. In brief, healthy cells express MHC class I molecules on their surface, which interact with inhibitory receptors on NK cells, promoting self-tolerance. However, tumor or virus-infected cells often stop expressing MHC class I, reducing the inhibitory signal for NK cells. At the same time, cellular stress caused by viral infection or associated with tumor development triggers the upregulation of ligands for activating receptors on these cells. Following NK-cell activation, inhibitory receptors, such as checkpoint receptors, are expressed. Consequently, the signal from activating receptors on NK cells tips the equilibrium toward activation, leading to the elimination of target cells [25].

In the present study, we conducted an in vitro simulation of NK-cell activation and performed an analysis to determine the expression of TIGIT, TIM-3, LAG-3, and Siglec-7 on mature and CD56 ^bright NK cells, both before and after co-incubation with K562 cells. Subsequently, we compared the expression of the aforementioned ICs on circulating human NK cells obtained from healthy donors to the expression on NK cells from patients suffering from gastric cancer and observed a divergent expression of ICs attributable to the mature CD56^dim subpopulation of NK cells. Finally, based on our findings, we discuss whether the four ICs studied here can be utilized as diagnostic biomarkers and targets to be addressed via immunotherapies in cancer.

Methods

Study Subjects

At the Institute of Clinical Immunology at the University of Leipzig and at the Institute of Transfusion Medicine of the University of Leipzig Medical Center, blood samples were obtained from healthy adult donors (n = 35) (shown in online suppl. Table S1; for all online suppl. material, see https://doi.org/10.1159/000545429). Additionally, blood samples collected from patients suffering from gastric adenocarcinoma (n = 9) or adenocarcinoma of the gastroesophageal junction (n = 12) were analyzed (shown in online suppl. Table S2). Those samples were taken at the Leipzig University Cancer Clinic. Ethics vote, approval of sample collection, and processing are detailed in the Statement of Ethics.

K562 Cell Culture and Vitality Test with Violet Proliferation Dye 450 Staining

K562 cells, provided by the Institute of Clinical Immunology at the University of Leipzig, were cultured in RPMI Medium 1640 (Gibco, Paisley, UK) supplemented with 10% FBS (Gibco, Paisley, UK) and 1% Gentamycin (Sigma-Aldrich, USA). They were kept at a concentration of 0.5–1 × 10⁶ cells/mL. Before using the cells in experiments, 5 × 10⁶ cells were washed and diluted in 1 mL of sterile phosphate-buffered saline (PBS) (Gibco, Paisley, UK) to be stained with 1 µm of Violet Proliferation Dye 450 (VPD 450) (BD Biosciences, USA) for 15 min at 37°C. After that, cells were washed in PBS and re-suspended in complete medium consisting of Iscove’s Modified Dulbecco’s Medium [IMDM] (Gibco, Paisley, UK) supplemented with 10% FBS, 1% penicillin-streptomycin (Gibco, Paisley, UK) and 50 µm 2-Mercaptoethanol (Gibco, Paisley, UK).

Stimulation of Peripheral Blood Mononuclear Cells with K562

Peripheral blood mononuclear cells (PBMCs) were isolated from fresh heparinized peripheral blood by density gradient centrifugation over Pancoll (Pan-Biotech, Germany). Isolated PBMCs were resuspended in complete medium at a concentration of 2 × 10⁶ cells/mL. 2 × 10⁵ PBMCs/well were seeded into a 96 well round bottom plate and stimulated with 2 × 10⁵ K562 cells/well for 24 h (effector to target cell ratio [E:T] = 10:1) in a humidified atmosphere of 5% CO₂ at 37°C. PBMCs co-incubated with K562 cells for 24 h were washed with sterile PBS before they were stained.

Flow Cytometry Analysis of TIM-3, LAG-3, TIGIT, and Siglec-7 Expression on Human NK Cells

Measurement of Unstimulated NK Cells from Healthy Donors

Two separate antibody cocktails, “Mix #1” and “Mix #2,” were used for measuring the expression of TIM-3, LAG-3, TIGIT and Siglec-7 on circulating NK cells from healthy donors. Both antibody cocktails shared a common backbone that allowed the identification of NK cells (shown in online suppl. Fig. S1A–D). This backbone consisted of CD3 V500 (2.5 µL, clone UCHT1, BD Biosciences), CD16 FITC (2.5 µL, clone 3G8, BioLegend), CD45 APC-H7 (2.5 µL, clone 2D1, BD Biosciences) and CD56 PE-Cy^TM7 (2.5 µL, clone NCAM16.2, BD Biosciences). For “Mix #1” we added CD223 (LAG-3) PE (5 µL, clone 11C3C65, BioLegend) and CD366 (TIM-3) APC (5 µL, clone F38-2E2, BioLegend) and for “Mix #2” we added CD328 (Siglec-7) PE (5 µL, clone 6–434, BioLegend) and TIGIT Alexa Fluor^® 647 (5 µL, clone A15153G, BioLegend). Those drop-ins allowed the detection of either TIM-3 and LAG-3 (“Mix #1”) or Siglec-7 and TIGIT (“Mix #2”) (shown in online suppl. Fig. 1E–H). Each antibody cocktail was used to stain 100 µL of EDTA whole blood. After 15 min of cell surface staining at room temperature in the dark, erythrocytes were lysed by incubation with BD FACS^™ Lysing Solution (BD Biosciences, USA) for 10 min. Afterward, the remaining leukocytes were washed (500 × g, 5 min) in PBS supplemented with 0.5% FBS and 0.05% NaN₃ (Merck, USA) (“washing buffer”) and then fixated with 300 µL PBS containing 1% formaldehyde (Merck, USA). FACSCanto II-based eight-color flow cytometry was conducted to measure the samples. The system was set up with three lasers: a violet laser with a wavelength of λ = 405 nm, a blue laser with λ = 488 nm, and a red laser λ = 647 nm (shown in online suppl. Fig. S3). BD FACSDiva software (Version 8.0.1) was used for acquisition of events.

Measurement of K562-Stimulated NK Cells

PBMCs co-incubated with K562 cells were stained with the same two antibody cocktails as described above. After 15 min of staining, cells were washed with PBS supplemented with 3% FBS and then fixated.

Determination of K562 Viability

After co-incubating PBMCs and K562 cells for 24 h, cells were harvested and washed with sterile PBS. They were then stained with 1 µL of Fixable Viability Dye eFluor™ 780 (eBioscience, USA). After 30 min of staining protected from light at 4°C, they were washed with PBS supplemented with 3% FBS and then again stained with CD3 V500 (2.5 µL, clone UCHT1, BD Biosciences), CD56 PE-Cy^TM7 (2.5 µL, clone NCAM16.2, BD Biosciences), and CD69 APC (20 µL, clone FN50, BD Biosciences) (shown in online suppl. Fig. S2). After washing them again with PBS + 3% FBS, they were fixated. Mouse IgG1 isotype control APC (20 µL, clone MOPC-21, BD Biosciences) was used for the assessment of background staining.

Measurement of NK Cells from Patients with Adenocarcinoma

NK cells from whole blood from patients with adenocarcinoma (see Study Subjects) were analyzed following the procedure outlined in section Measurement of Unstimulated NK Cells from Healthy Donors.

Isotype Control

PE Mouse IgG1 isotype control (5 µL, clone 679.1Mc7, Beckmann Coulter), Alexa Fluor^® 647 Mouse IgG2a isotype control (5 µL, clone MOPC-173, BioLegend), APC Mouse IgG1 isotype control (5 µL, clone 679.1Mc7, Beckmann Coulter), and FITC Mouse IgG1 isotype control (20 µL, clone MOPC-21, BD Biosciences) were used for the assessment of background staining.

Statistical Analysis

The statistical analysis was performed using SigmaPlot 14.0 (Systat Software Inc, USA). Measured results of each set of experiments are given as median and interquartile range (IQR). The Mann-Whitney U test was performed to compare differences between healthy donors and patients. Otherwise, Wilcoxon’s test was used for comparison of paired samples. p values ≤0.050 were deemed statistically significant (n.s. – not significant, *p ≤ 0.050, **p ≤ 0.010, ***p ≤ 0.001). Correlation analysis was used to identify Spearman’s correlation coefficient (r_s).

Results

Differential Expression of TIM-3, LAG-3, TIGT, and Siglec-7 on Circulating Human NK Cells

We used flow cytometry to measure the expression of TIM-3, LAG-3, TIGIT, and Siglec-7 on circulating NK cells from 35 healthy donors. To achieve a more detailed analysis, we compared the expression on mature CD56^dim NK cells to the expression on immature CD56^bright NK cells (shown in Table 1).

Table 1.

Comparison of mature CD56^dim NK cells to immature CD56^bright NK cells from a healthy control group (n = 35) in terms of the median fluorescence intensity (MFI) and percentage of TIM-3, LAG-3, TIGIT, or Siglec-7 positive NK cells

	All NK cells	CD56dim	CD56bright
Healthy donors, median (IQR)
TIM-3
MFI	709 (472–927)	716 (473–939)	573 (431–718)
%	24.9 (11.8–38)	24.5 (12.7–38.4)	19.3 (12.3–29.4)
LAG-3
MFI	46 (41–57)	46 (41–57)	57 (32–69)
%	0.6 (0.3–0.9)	0.6 (0.3–1)	0.4 (0–1)
TIGIT
MFI	878 (788–1,073)	895 (804–1,136)	556 (462–740)
%	29.2 (23–41.3)	28.9 (23.7–43.5)	25.5 (15.3–32.7)
Siglec-7
MFI	10,694 (7,660–12,478)	10,907 (7,420–12,618)	5,042 (4,298–6,372)
%	97 (93.2–98.7)	97 (93–98.7)	96.5 (95–97.8)

TIM-3 was detectable on nearly one quarter of all circulating NK cells (median, 24.9%; IQR, 11.8%, 38%) (shown in Fig. 1a). However, we observed high interindividual differences (shown in Fig. 1b). Notably, the surface detection of TIM-3 on CD56^dim NK cells showed a 25% greater MFI than on CD56^bright NK cells (p = 0.022) (shown in Fig. 1c). This could indicate a higher surface density on CD56^dim NK cells, as CD56^dim and CD56^bright NK cells do not show a significant difference with regard to the portion of TIM-3⁺ cells (p = 0.076) (shown in Fig. 1d).

Fig. 1.

Determination of checkpoint expression on circulating NK cells from healthy donors (n = 35). The portion of TIM-3⁺ (a), LAG-3⁺ (e), TIGIT⁺ (i), or Siglec-7⁺ (m). NK cells was measured via flow cytometry. Representative histograms portray interindividual differences in the expression of those ICs (b, f, j, n). The MFI of all four checkpoints (c, g, k, o) and the portion of checkpoint positive cells (d, h, l, p) were compared between the CD56^dim and CD56^bright subpopulation. Connected data points represent the same donor. *p ≤ 0.050; ***p ≤ 0.001. n.s., not significant (Wilcoxon’s test).

We only detected a very small fraction of LAG-3⁺ NK cells (median, 0.6%; IQR, 0.3%, 0.9%) (shown in Fig. 1e, f). Neither CD56^dim nor CD56^bright NK cells appear to express significant amounts of LAG-3 on their surface when in a resting state (shown in Fig. 1g–h).

TIGIT expression also varied strongly among different individuals but was clearly detectable on some circulating NK cells (median, 29.2%; IQR, 23%, 41.3%) (shown in Fig. 1i, j). Mature CD56^dim NK cells showed a 60.97% higher MFI than CD56^bright NK cells (p ≤ 0.001) (shown in Fig. 1k) and expressed TIGIT more frequently (p = 0.040). Nevertheless, it is worth to note that TIGIT⁺ CD56^dim NK cells showed a lower MFI of TIGIT than TIGIT⁺ CD56^bright NK cells (p ≤ 0.001) (shown in online suppl. Fig. S4).

Siglec-7 was expressed on the vast majority of NK cells (median, 97%; IQR, 93.2%, 98.7%) (shown in Fig. 1m, n). Its MFI on CD56^dim NK cells is 2.16 times greater than that on CD56^bright NK cells (p ≤ 0.001) (shown in Fig. 1o). Given the fact that both subsets show an equal portion of Siglec-7⁺ cells (p = 0.310) (shown in Fig. 1p), this could point to a higher Siglec-7 expression density on mature CD56^dim NK cells.

Killing of Target Cells Is Associated with Increased Expression of TIM-3

To investigate how the expression of TIM-3, LAG-3, TIGIT, and Siglec-7 on NK cells is impacted by the execution of cytotoxic effector functions, we co-incubated PBMC with HLA class I – deficient K562 cells and used CD69 as activation marker. After 24 h of co-incubation, the percentage of dead K562 cells increased by 72% (p = 0.002) (shown in Fig. 2a). NK cells co-incubated with K562 also showed a 3.5 times higher MFI of CD69 than NK cells co-incubated with medium alone (p ≤ 0.001) (shown in Fig. 2b; Table 2). The MFI of CD69 on NK cells correlated strongly with the percentage of dead K562 cells (r_s = 0.673, p = 0.0209) (shown in Fig. 2d). The MFI of CD69 on CD56^dim NK cells was more than twice as high as the MFI on CD56^bright NK cells (p ≤ 0.001) (shown in Fig. 2c; Table 2).

Fig. 2.

PBMCs and K562 cells were co-incubated for 24 h at an E:T-ratio of 10:1 (n = 11). The portion of dead K562 cells was measured and compared to 24 h of K562 incubation with medium alone (a). The MFI of CD69 was measured on all NK cells (b), as well as on the CD56^dim and CD56^bright subsets (c), and compared to 24 h of PBMC incubation with medium alone. The MFI of CD69 on NK cells correlated with the percentage of dead K562 cells (d). **p ≤ 0.010, ***p ≤ 0.001. n.s., not significant (Wilcoxon’s test); r_s, Spearman’s correlation coefficient.

Table 2.

Comparison between the MFI of CD69, TIM-3, LAG-3, TIGIT, or Siglec-7 on NK cells co-incubated with medium or HLA class I – deficient K562 cells (n = 11)

	Medium, median (IQR)	K562, median (IQR)	p value
All NK cells
CD69	88 (82–97)	308 (200–516)	<0.001
TIM-3	669 (606–729)	886 (817–1,017)	<0.001
LAG-3	75 (71–97)	77 (71–86)	0.831
TIGIT	802 (629–875)	1,255 (1,093–1,736)	<0.001
Siglec-7	9,289 (7,325–10,827)	9,203 (7,551–11,802)	0.365
CD56dim
CD69	90 (82–97)	330 (242–572)	<0.001
TIM-3	723 (641–754)	897 (819–1,039)	<0.001
LAG-3	80 (74–97)	82 (73–86)	0.846
TIGIT	891 (840–1,039)	1,353 (1,233–1,782)	<0.001
Siglec-7	10,514 (7,940–12,394)	10,419 (8,193–12,170)	0.175
CD56bright
CD69	88 (82–111)	150 (138–190)	<0.001
TIM-3	544 (474–577)	801 (726–1,076)	<0.001
LAG-3	65 (50–98)	66 (55–85)	0.966
TIGIT	309 (278–363)	345 (310–439)	<0.001
Siglec-7	4,079 (2,833–4,923)	4,456 (3,100–5,215)	0.206

Confrontation with HLA class I – deficient target cells resulted in a 32.44% increased MFI of TIM-3 on NK cells (p ≤ 0.001) (shown in Fig. 3a). The MFI of TIM-3 correlated strongly with both the MFI of CD69 (r_s = 0.682; p = 0.0186) and the percentage of dead K562 cells (r_s = 0.682; p = 0.0186) (shown in Fig. 3e, f). Stimulated as well as unstimulated CD56^dim NK cells possessed a higher MFI of TIM-3 than their CD56^bright counterparts (p ≤ 0.001 [PBMCs + Medium]; p = 0.010 [PBMCs + K562]) (shown in Fig. 3g; Table 4). Both subpopulations showed comparable stimulation indices (p = 0.147), which indicates that CD56^dim and CD56^bright NK cells increased the MFI of TIM-3 to an equal extent (shown in online suppl. Table S3).

Fig. 3.

PBMCs and K562 cells were co-incubated for 24 h at an E:T-Ratio of 10:1 (n = 11). The MFI of TIM-3 (a), TIGIT (b), Siglec-7 (c), or LAG-3 (d) on NK cells was measured via flow cytometry. The MFI of TIM-3 correlated strongly with both the MFI of CD69 (e) and the percentage of dead K562 cells (f). The MFIs of TIM-3 (g) and TIGIT (h) were also compared between the CD56^dim (left) and the CD56^bright (right) subpopulation (**p ≤ 0.010, ***p ≤ 0.001. n.s., not significant (Wilcoxon’s test); r_s, Spearman’s correlation coefficient).

Table 4.

Impact of K562 and adenocarcinoma on the MFI of TIM-3, LAG-3, TIGIT, or Siglec-7 on NK cells

The MFI of TIGIT increased by 56.48% when NK cells were incubated with K562 cells (p ≤ 0.001) (shown in Fig. 3b; Table 4). Unlike TIM-3, TIGITs MFI correlated neither with the MFI of CD69 nor with the percentage of dead K562 cells. Even though unstimulated CD56^dim NK cells already showed a higher MFI of TIGIT (p ≤ 0.001), they still increased the expression at a greater rate than CD56^bright NK cells, as was indicated by a higher stimulation index (=MFI on stimulated cells/MFI on unstimulated cells) (p ≤ 0.001) (shown in Fig. 3h, online suppl. Table S3).

The MFIs of both LAG-3 (p = 0.831) (shown in Fig. 3d) and Siglec-7 (p = 0.365) (shown in Fig. 3c) were not affected by the presence of K562 cells. This remained to be true when CD56^dim (p = 0.846 [LAG-3]; p = 0.175 [Siglec-7]) and CD56^bright NK cells (p = 0.966 [LAG-3]; p = 0.206 [Siglec-7]) were looked at separately (shown in Table 4).

NK Cells from Gastric Cancer Patients Express More LAG-3 and Less Siglec-7

Some forms of cancer are known to be accompanied by altered expression patterns of ICs on NK cells. We therefore investigated whether NK cells from patients with adenocarcinoma of the stomach or gastroesophageal junction show different TIM-3, LAG-3, TIGIT, or Siglec-7 expression levels when compared to NK cells from healthy donors (shown in Tables 3, 4).

Table 3.

Comparison of a healthy control group (n = 35) to adenocarcinoma patients (n = 21) regarding the MFI and percentage of TIM-3, LAG-3, TIGIT or Siglec-7 positive NK cells

	Healthy donors, median (IQR)	Adenocarcinoma, median (IQR)	p value
All NK cells
TIM-3
MFI	709 (472–927)	672 (503–893)	0.933
%	24.9 (11.8–38)	27.5 (16.15–38.45)	0.407
LAG-3
MFI	46 (41–57)	55 (45–72.5)	0.036
%	0.6 (0.3–0.9)	1.5 (0.55–3.1)	0.011
TIGIT
MFI	878 (788–1,073)	782 (681.5–1,038.5)	0.198
%	29.2 (23–41.3)	25.9 (15.05–38.9)	0.243
Siglec-7
MFI	10,694 (7,660–12,478)	8,068 (5,637.5–9,754.5)	0.022
%	97 (93.2–98.7)	93.3 (84.9–96.35)	0.022
CD56dim
TIM-3
MFI	716 (473–939)	674 (502.5–899)	0.800
%	24.5 (12.7–38.4)	27.8 (16.15–38.6)	0.441
LAG-3
MFI	46 (41–57)	57 (42.5–69.5)	0.030
%	0.6 (0.3–1)	1.2 (0.55–2)	0.029
TIGIT
MFI	895 (804–1,136)	783 (689–1,046)	0.220
%	28.9 (23.7–43.5)	25.6 (15.05–39.3)	0.250
Siglec-7
MFI	10,907 (7,420–12,618)	8,309 (5,653–10,023)	0.029
%	97 (93–98.7)	93.2 (84.55–96.45)	0.022
CD56bright
TIM-3
MFI	573 (431–718)	618 (497–790)	0.374
%	19.3 (12.3–29.4)	25.7 (13.55–34.2)	0.176
LAG-3
MFI	57 (32–69)	42 (24–91)	0.919
%	0.4 (0–1)	0.8 (0–9)	0.149
TIGIT
MFI	556 (462–740)	546 (462–614.5)	0.493
%	25.5 (15.3–32.7)	20.5 (15.5–31.85)	0.565
Siglec-7
MFI	5,042 (4,298–6,372)	5,927 (4,676.5–6,387)	0.318
%	96.5 (95–97.8)	95.6 (92.65–97.05)	0.058

The MFI of TIM-3 (p = 0.933) and TIGIT (p = 0.198) did not differ between the two groups (shown in Fig. 4a, b). A separate comparison of the CD56^dim or CD56^bright subpopulation did also not reveal a significantly altered expression, neither for TIM-3 (p = 0.800 [CD56^dim]; p = 0.374 [CD56^bright]) nor for TIGIT (p = 0.220 [CD56^dim]; p = 0.493 [CD56^bright]). NK cells of patients did show a 19.57% higher MFI of LAG-3 (p = 0.036) (shown in Fig. 4c) and a 24.56% lower MFI of Siglec-7 (p = 0.022) (shown in Fig. 4d). Both of those differences were driven by the CD56^dim subset, while expression on the CD56^bright subset remained unchanged (shown in Fig. 4e, f).

Fig. 4.

The MFI of TIM-3 (a), TIGIT (b), LAG-3 (c), or Siglec-7 (d) on NK cells from patients with adenocarcinoma (AC, grey; n = 21) was compared to a group of health donors (HD, white; n = 35). The MFIs of LAG-3 (e) and Siglec-7 (f) were also compared between the CD56^dim (left) or the CD56^bright (right) subpopulation. *p ≤ 0.050. n.s., not significant (Mann-Whitney U test).

Discussion

In general, activated NK cells appear to display certain differences in their phenotype compared to resting NK cells, characterized by the upregulation of inhibitory receptors and/or the downregulation of activating receptors. These phenotypic alterations in NK cells are closely associated with variations in their functionality and cytotoxic capacity. CD56 expression enables the categorization of human NK cells based on their function [26–29]. The level of CD56 expression signifies the developmental transition from immature to mature NK cells. Accordingly, the NK cells studied were categorized into two main groups: CD56^dim (mature) and CD56^bright (immature).

In the present study, we investigated the expression of TIM-3, LAG-3, TIGIT, and Siglec-7 on circulating NK cells from 35 healthy donors (1) as baseline without stimulation, (2) after in vitro stimulation by HLA class I-deficient K562 cancer cells (n = 11), and (3) in comparison to peripheral NK cells from gastric cancer patients (n = 21). Additionally, we also compared IC expression on CD56^dim (mature) and CD56^bright (immature) NK cells.

TIGIT

TIGIT is an inhibitory checkpoint receptor expressed on T lymphocytes and NK cells [30]. In healthy individuals, TIGIT expression on NK cells shows considerable variability and can directly inhibit NK-cell cytotoxicity [31, 32]. In our study, we clearly detected TIGIT in healthy donors (1), but also observed variable expression levels among different individuals. Notably, mature CD56^dim NK cells exhibited approximately 60% higher mean fluorescence intensity (MFI) of TIGIT compared to CD56^bright NK cells. The majority of NK cells in the peripheral blood are mature CD56^dim NK cells (∼90%), which likely explains these observations [33]. Interestingly, we identified lower MFI of TIGIT in TIGIT⁺CD56^dim NK cells than in TIGIT⁺CD56^bright NK cells, suggesting a higher expression density of TIGIT in CD56^bright NK cells. These findings may be related to those of Wang et al. [32] who classified NK cells based on TIGIT expression into low-level (30–50%), middle-level (50–70%), and high-level (70–90%) groups. They found that TIGIT expression is inversely correlated with NK-cell cytotoxic potential. While CD56^bright NK cells are potent producers of inflammatory cytokines, CD56^dim NK cells exhibit enhanced cytotoxicity [26, 29]. Collectively, CD56^dim NK cells with low TIGIT expression appear to represent a highly cytotoxic NK-cell population. Previous studies have also indicated that downregulation of CD56 during maturation is associated with anti-tumor cytotoxicity [23, 26]. Furthermore, Imai et al. [34] observed in a long-term study that low NK-cell activity was associated with an increased risk of cancer. However, the role of TIGIT, particularly in dysfunctional NK cells within tumors, remains poorly understood. It warrants further investigation to determine whether TIGIT^lowCD56^dim NK cells might correlate with better outcomes in cancer disease, indicating that TIGIT in combination with CD56 could have the potential to function as dual prognostic marker in cancer diseases.

Upon co-incubation of NK cells from healthy donors (2) with HLA class I-deficient K562 cells, the MFI of TIGIT increased by approximately 56%. However, no correlations were found with either the MFI of CD69 (as an activation marker) or the percentage of dead K562 cells. Similar to unstimulated CD56^dim NK cells, TIGIT expression on K562-stimulated CD56^dim NK cells increased at a substantially higher rate compared to CD56^bright NK cells. These findings align with those of Wang et al. [32] who observed significantly higher percentages of CD107a⁺NK cells in the TIGIT^low CD56^dim NK-cell group than in the TIGIT^high CD56^dim NK-cell group following LPS stimulation and co-culture with K562 cells. Collectively, these characteristics position TIGIT as a potential target in immunotherapies.

However, when comparing TIGIT expression on peripheral NK cells from healthy donors and gastric cancer (3) patients, no significant differences were observed between the two groups, neither in CD56^dim nor in CD56^bright NK cells. Moreover, TIGIT expression has been frequently reported in various peritumoral lymphocytes, but TIGIT expression levels are highly variable across different cancer types [35, 36]. Our data also emphasize the variability of TIGIT expression across different individuals and cancer types, highlighting the necessity for further research on TIGIT as a potential biomarker in conjunction with other NK-cell receptors, such as CD56.

LAG-3

We analyzed LAG-3 expression on (1) unstimulated circulating NK cells from healthy donors and found nearly undetectable levels on both CD56^dim and CD56^bright NK cells. Additionally, LAG-3 expression was also not affected by exposure to K562 cells (2). However, we observed an approximately 20% higher LAG-3 MFI on CD56^dim circulating NK cells in cancer patients (3), while expression on the CD56^bright subset remained unchanged.

Narayanan et al. [37] stimulated PBMCs with IFN-α and IL-2 and observed increased LAG-3 expression on mature CD56^dim NK cells. These LAG-3⁺CD56^dim NK cells expressed higher levels of CD57, NKG2C, CD158, and TIM-3, indicating a mature and activated NK-cell subset. As mentioned before, LAG-3 was also characterized as a negative regulator of cytokine production in mature NK cells. If this is the case in circulating NK cells from gastric cancer patients, these NK cells would not only be inherently inhibited but might also indirectly inhibit the activation of adaptive immune cells such as T cells and dendritic cells. To establish LAG-3 as a potential NK-cell exhaustion marker in gastric cancer patients, the next step could involve analyzing their ability to express CD107a or produce IFN-γ in response to in vitro stimulation. Maruhashi et al. [38] concluded in their review that LAG-3, like PD-1 and CTLA-4, could play a significant role in immune escape mechanisms in various solid tumors. One might speculate that the role of LAG-3 in NK cells might be similar to its role in T cells. This is interesting because LAG-3 expression has been reported on exhausted tumor-infiltrating T cells, which lose their cytokine production, and on peripheral regulatory T cells of patients with various solid cancers [39–41]. LAG-3 expression has also been detected in leukemia patients and is considered as a potential target, alongside PD-1 and CTLA-4, in checkpoint immunotherapy [42]. There is evidence of blocking the LAG-3 pathway in tumor patients, although without observed effects on cytotoxicity [40, 43]. Therefore, dual blockade of LAG-3 and PD-1 has shown promise as a potential new strategy for anti-tumor immunotherapy [44, 45]. Regarding gastric cancer, the RELATIVITY-060 trial studies the effectiveness of relatlimab (LAG-3 inhibitor) in combination with nivolumab and chemotherapy in participants with gastric and gastroesophageal junction cancers. However, this phase 2 study did not meet its primary endpoint [46]. Altogether, LAG-3 represents a highly promising next-generation IC molecule and warrants further investigation, including in the context of gastric cancer.

TIM-3

TIM-3 has several cognate ligands, including high mobility group protein 1 (HMGB1), phosphatidylserine (PS), carcinoembryonic antigen-related cell adhesion molecule 1 (Ceacam1), and galectin-9 (Gal-9) [47–49]. Various studies have described TIM-3, an inhibitory checkpoint receptor, as a marker of NK-cell activation or maturation, and targeting TIM-3 could inhibit NK cell-mediated cytotoxicity [50–52].

We detected TIM-3 on nearly a quarter of all circulating NK cells from healthy donors (1), with notable inter-individual variability in expression levels. Upon exposure to K562 cells (2), we observed a significant increase in TIM-3 expression, which strongly correlated with NK-cell activation and the proportion of dead K562 cells. Both stimulated and unstimulated CD56^dim NK cells exhibited higher TIM-3 MFI than their CD56^bright counterparts. These observations confirm that TIM-3 is also expressed on resting NK cells and upregulated after NK-cell stimulation with K562 cells. The upregulation of TIM-3 on NK cells can also be induced by other stimuli, including Fc receptor engagement and exposure to cytokines such as IL-2, IL-15, and IL-18. Some studies suggest that NK cells may harbor intracellular stores of TIM-3, allowing for rapid upregulation upon ligand binding [50, 53]. It was assumed that the TIM-3-galectin-9 pathway is involved in immune escape mechanisms in different cancer types and observed enhanced IFN-γ production in response to TIM-3-galectin-9 interaction [54, 55]. Blocking this pathway to inhibit immune escape could be an approach to develop novel anti-cancer immunotherapeutics.

Several studies described an upregulated expression of TIM-3 in patients with various solid tumor types, including lung cancer, renal cancer, colorectal cancer, bladder urothelial carcinoma et cetera [56–60]. For instance, Wu et al. [61] observed an increased expression level of TIM-3 on cells in the peripheral blood of patients with ovarian cancer. However (3), we did not find any differences in the expression of TIM-3 on circulating NK cells when comparing healthy and adenocarcinoma cohorts, neither on all NK cells nor in a separate comparison of CD56^dim and CD56^bright NK cells subsets. One might assume that TIM-3 expression on NK cell is dependent on cancer types and may also be influenced by other factors. Interestingly, Da Silva et al. [52] compared the expression and function of TIM-3 on circulating NK cells from healthy donors and patients with metastatic melanoma. They observed increased expression levels of TIM-3 on NK cells in melanoma patients and the NK cells were exhausted. Similar, So et al. [50] demonstrated functionally diverse TIM-3⁺ NK cells activated by different stimuli and therefore suggest being careful when using TIM-3 as an independent exhaustion marker.

Furthermore, Liu et al. [62] recently concluded in their review that TIM-3 is differentially expressed on both innate and adaptive immune cells, and that TIM-3 could have distinct effects on the functions of these cell types. Therefore, further characterization is needed to clarify if or in which cases TIM-3 has potential as an exhaustion marker and if blocking of TIM-3 could be a therapeutic option. There are also studies suggesting that TIM-3 may function as prognostic marker in autoimmune disease such as multiple sclerosis, rheumatoid arthritis, and systemic lupus erythematosus (SLE) [63–65].

Siglec-7

Siglec-7 is expressed on NK cells, CD8+ T cells, and at lower levels on monocytes and granulocytes [66, 67]. While various ligands for the inhibitory receptor Siglec-7 have been identified, its binding domain is also well characterized [68–73]. The binding of Siglec-7 to its ligands helps NK cells distinguish self from non-self, leading to reduced chemokine production and an increased cytolytic potential of NK cells [68, 74]. Many tumor cells express Siglec-7 ligands on their surface to evade anti-tumor surveillance [74, 75]. Consequently, targeting the Siglec-7/ligand axis has garnered significant attention in the development of new immunotherapies [76–78].

Consistent with findings from other research groups, we (1) detected Siglec-7 on the vast majority of NK cells, with the MFI of CD56^dim NK cells being approximately twice as high than that on CD56^bright NK cells [67, 79]. Both subsets exhibited an equal proportion of Siglec-7⁺ cells, suggesting a higher density of Siglec-7 on mature CD56^dim NK cells compared to CD56^bright NK cells. After co-incubation (2) of circulating NK cells with HLA class I-deficient K562 cells, we did not observe changes in the expression of Siglec-7. Huang et al. [80] provided evidence that transcription might be the key step in regulating Siglec-7 expression on circulating NK cells as they were able to increase its expression by using a DNA methyltransferase inhibitor (5azaC) or a histone deacetylase inhibitor (butyric acid). This is interesting because such epigenetic therapeutic drugs are in clinical use for the treatment of AML, one of the diseases where a reduction in Siglec-7 expression on circulating NK cells has been reported [80–82]. However, Siglec-7 expression on NK cells has been proposed to indicate maturity and a high degree of functionality [67]. Activation of Siglec-7⁺ NK cells might be easier than activation of Siglec-7⁻ NK cells as multiple studies have shown that Siglec-7⁺ NK cells express activating receptors (e.g., NKp30, NKp46, NKG2D, CD16, 2B4, CD38, DNAM1) at higher levels, while expressing less inhibitory receptors (NKG2A, CD158b) [67, 83, 84].

Previous studies have shown that Siglec-7 expression on circulating NK cells is reduced in both hematologic malignancies and solid tumors as well as in different viral infections [82–85]. In gastric cancer, there are no ongoing clinical trials targeting Siglec-7. But research on other Siglec family members gives us clues on potential strategies. For example, Siglec-15 is expressed in gastric cancer, so targeting Siglec pathways might be relevant here [86]. Also, overexpression of sialylated tumor associated antigens like MUC1 in various cancers including gastric cancer can engage Siglec receptors like Siglec-7 and Siglec-9 and lead to immune suppression. Targeting the Siglec-sialylated MUC1 axis has been proposed as a potential therapeutic strategy [76, 87]. These findings suggest that blocking Siglec-7 may enhance NK-cell anti-tumor activity in gastric cancer. Even though Siglec-7 was present on the vast majority of NK cells in both healthy donors and patients, we report here for the first time that gastric cancer patients show a reduced expression of Siglec-7 on their circulating NK cells. The decrease of around 25% in Siglec-7’s expression was confined to the CD56^dim NK-cell subset. Given the fact that NK cells that express Siglec-7 at lower levels are less functional and limit anti-tumor responses, Siglec-7 may be a potential target in gastric cancer [67]. Our results highlight the significance of Siglec-7 as a targetable IC to enhance the cytotoxic activity of NK cells, thereby promoting the activation of adaptive immune cells in patients with gastric cancer.

In conclusion, the ICs TIGIT, TIM-3, LAG-3, and Siglec-7 exhibit heterogeneous expression on NK cells. TIGIT, TIM-3, and Siglec-7 are broadly expressed, whereas LAG-3 is typically absent under normal conditions but may emerge in specific activated states or subsets. TIGIT expression appears to be influenced by tumor cell stimulation, potentially serving as a marker of tumor progression. The upregulation of TIM-3 and TIGIT following various stimuli suggests their role as indicators of NK-cell exhaustion. The role of LAG-3 in NK cells remains unclear and warrants further study. Meanwhile, reduced Siglec-7 expression in circulating NK cells of gastric cancer patients highlights its potential as a potentially target for further investigation.

Statement of Ethics

Sample collection and processing was approved by the Ethics Committee of Leipzig University (351/17-ek, 255/20-ek) and completed according to the Medical Faculty, University of Leipzig standard operating guidelines and regulations. Written informed consent was obtained from all participating individuals.

Conflict of Interest Statement

The authors have no conflicts of interest to declare.

Funding Sources

This study was not supported by any sponsor or funder.

Author Contributions

Conceptualization: U.S. and A.B. Acquisition of data (acquired and managed participants, provided facilities, flow cytometry, in vitro studies, and so on): S.S., B.S., J.K., and A.B. Analysis and interpretation of data (for example, statistical analysis, biostatistics, and computational analysis): S.S., A.-R.B., B.S., U.S., and A.B. Writing, review and/or revision of the manuscript: S.S., A.-R.B., B.S., J.K., U.K., S.F., U.S., and A.B.

Funding Statement

This study was not supported by any sponsor or funder.

Data Availability Statement

All data generated or analyzed during this study are included in this article and its supplementary material files. Further enquiries can be directed to the corresponding author.