Natural killer (NK) cells are enriched within the liver. Apart from conventional NK (cNK) cells, recent studies identified a liver-resident NK (LrNK) subset, which constitutes about half of hepatic NK cells and exhibits distinct developmental, phenotypic, and functional features. However, it remains unclear whether and how LrNK cells, as well as cNK cells, participate in the development of hepatocellular carcinoma (HCC) individually. Here, we report that both LrNK and cNK cells are significantly decreased in HCC. The T-cell immunoglobulin and mucin domain-containing protein 3 (Tim-3) was significantly upregulated in both tumor-infiltrating LrNK and cNK cells and suppressed their cytokine secretion and cytotoxic activity. Mechanistically, phosphatidylserine (PtdSer) engagement promoted phosphorylation of Tim-3, which then competed with PI3K p110 to bind p85, inhibiting downstream Akt/mTORC1 signaling and resulting in malfunctioning of both NK-cell subsets. Tim-3 blockade retarded HCC growth in a NK-cell–dependent manner. These studies for the first time report the presence and dysfunction of LrNK cells in HCC and show that Tim-3–mediated PI3K/mTORC1 interference is responsible for the dysfunction of both tumor-infiltrating cNK and LrNK cells, providing a new strategy for immune checkpoint-based targeting.

Significance:

Tim-3 enhances hepatocellular carcinoma growth by blocking natural killer cell function.

Hepatocellular carcinoma (HCC), the fifth most common malignancy and the third leading cause of cancer-related death worldwide, has been regarded as an inflammation-related tumor (1, 2). Ninety percent of HCC develops due to liver inflammation of either chronic viral infection (such as hepatitis B virus/HBV and hepatitis C virus/HCV) or noninfectious liver injury. Therefore, the inflamed liver microenvironment plays vital roles in hepatocellular carcinogenesis (3). Natural killer (NK) cells represent approximately 20%–30% of the total hepatic lymphocytes, a percentage five times higher than that seen in the spleen and peripheral blood (4). The unusual high frequency of hepatic NK cells coincides with the critical roles of NK cells during liver tumorigenesis (5). A dramatic reduction of NK frequency and the impaired cytotoxicity and cytokines production was observed in tumor tissues compared with nontumor tissues in patients with HCC and correlated with liver tumor stages and HCC patients' survival (6). Indeed, various strategies aimed to rescue NK-cell dysfunction either by NK-cell transfer or by improvement of NK-cell cytotoxicity have been reported (7, 8).

Notably, over recent years, the heterogeneity of liver NK cells has attracted great interest. Apart from conventional NK (cNK) cells originating in the bone marrow and moving through the circulation, mouse liver tissues contain a subset of resident NK cells, referred as LrNK, which exhibit significant differences in phenotype, development, and functions with cNK, but share some features with type I innate lymphoid cells (ILC1; ref. 9). Phenotypically, the mutually exclusive expression of CD49a and DX5/CD49b is usually used for distinguishing these two NK subsets in murine, that is, CD49aDX5+ cNK cells and CD49a+DX5 LrNK cells (10, 11). In human liver tissues, CD56bright NK cells or CXCR6+ NK cells display some similarities with mouse LrNK cells (9, 12). Functionally, compared with cNK cells, liver-resident NK cells produce high level of IFNγ, TNFα, and GM-CSF (10), constitutively express TRAIL and play a dominant role in contact hypersensitivity after challenge (11). In viral infection, they negatively regulate hepatic T-cell responses against LCMV and adenovirus via the PD-1/PD-L1 axis (13). Most recently, LrNK cells were found to be depleted from colorectal liver metastasis (14). Although the roles of bulk liver NK cells in HCC development have been reported previously, the involvement of LrNK and cNK and the underlying molecular mechanisms needs to be further determined.

Elevating immune checkpoints in tumor microenvironment is a trick exploited by many types of tumors to evade immunosurveillance (15). Thus, therapies targeting checkpoint molecules have become a prominent strategy to boost antitumor immune responses and have achieved promising therapeutic effects in some patients with advanced cancers (16). Although exhausted T cells infiltrated in tumor tissues are well-known targets for these checkpoint inhibitors, high expression of immune checkpoints, such as programmed cell death-1 (PD-1), were also detected in peripheral or tumor-infiltrated NK cells of patients with cancer, and were associated with their exhausted phenotype (17, 18). Moreover, functional reversal of NK cells has been proven to contribute to the therapeutic efficacy of checkpoint inhibitors (19, 20). However, the expression pattern and the exact regulation of immune checkpoints on NK in liver tumor microenvironment require further investigation.

In this study, we showed that Tim-3 was the most abundantly expressed immune checkpoint receptor on tumoral NK cells, that is, cNK and LrNK, in HCC, and was associated with an inhibitory phenotype. Genetic ablation, antibody-based functional blockade, and lentivirus-mediated interference of Tim-3 inhibited growth of orthotopic liver cancer by restoring cytokine secretion and cytotoxicity of cNK and LrNK cells. Furthermore, Phosphatidylserine (PtdSer), as an endogenous ligand, induced Tim-3 phosphorylation, which in turn interfered with PI3K/mTORC1/p-S6 signaling pathway and led to the dysregulation of NK cells. Thus, our study reveals a crucial role of Tim-3 in regulating the functional status of both cNK and LrNK in HCC microenvironment. Our findings elicit a novel mechanism of cancer immunity and provide a new intervention site of immune checkpoint-based tumor therapeutic strategy.

Patient samples

Surgically resected fresh liver cancer specimen and tissue microarrays were involved in flow cytometry analysis or multiplex IHC staining and overall survival analysis. This study was approved by the Ethics Committee of Shandong University School of Basic Medical Sciences (Jinan, China), and all patients provided informed written consent. All the data of the human subjects are summarized in online Supplementary Table S1.

Mice, tumor models, and treatment

Six- to 8-week-old male mice were used for all experiments. Wild-type (WT) BALB/c or C57BL/6 mice, BALB/c nude mice, Tim-3mut C57BL/6 mice, or Tim-3 knockout (Tim-3−/−) C57BL/6 mice were included for establishing subcutaneously transplanted tumor model and orthotopic liver tumor model. All animal experimental procedures were approved by the Animal Care and Use Committee of Shandong University (Jinan, China).

Cell lines

Human NK cell line NK92 cells were cultured in medium (αMEM containing 12.5% FBS, 12.5% horse serum, 0.2 mmol/L inositol, 0.1 mmol/L β-mercaptoethanol, 0.02 mmol/L folic acid) containing 100 U/mL recombinant human IL2. Human NK cell line NKL cells were cultured in RPMI1640 medium with 10% FBS containing 100 U/mL human IL2. B16 and K562 cells were cultured in RPMI1640 medium plus 10%FBS. NK92, B16, and K562 cells were purchased from ATCC. NKL cell line was kindly gifted by Professor B.Q. Jin (Department of Immunology, Fourth Military Medical University, Xi'an, China). The cell lines were recently authenticated in 2017 and Mycoplasma contamination was routinely tested.

Multiplex IHC staining

Multiplex IHC staining for CD3, CD56, CXCR6, and Tim-3 was performed using an Opal Multiplex IHC system. Images were scanned and analyzed using a Vectra 3.0 Automated Quantitative Pathology Imaging System (Perkin Elmer).

Flow cytometry

Isolated intrahepatic mononuclear cells, purified NK cells, or NK cell lines were subjected to surface markers or functional analysis by flow cytometry. Sample acquisition was performed using a Cytoflex S flow cytometer, and data were analyzed with CytExpert Programme (both from Beckman Coulter Inc.).

NK-cell purification and Tim-3 interference or blockade

NK cells from WT or Tim-3−/− mice spleen or tumor tissues were purified using a NK Cell Isolation Kit II (Miltenyi Biotec) according to the supporting protocol. For Tim-3 interference, freshly purified NK cells from WT BALB/c mice were infected with lentivirus expressing Tim-3 shRNA (LV-shTim-3) or negative control (LV-NC). For Tim-3–blocking assay, purified NK cells or isolated tumor infiltrated lymphocytes (TIL) were preincubated with anti-human/mouse Tim-3 antibody or control IgG. These cells were subsequently stimulated with IL12 and IL15 for evaluating cytokine secretion and granule release.

Cytotoxicity assay

B16 cells stably infected firefly luciferase–expressing lentivirus were cocultured with cNK or LrNK cells sorted from WT or Tim-3−/− mice with IL2 (100 UI/mL) at a ratio of 1:1 at 37 °C in a 5% CO2 incubator for 6 hours. The luciferase activity of coculture supernatant were detected by Luciferase Reporter Assay (Promega).

Statistical analysis

Statistical analyses were performed using the GraphPad Prism 5 software. Statistical significance was determined by a P < 0.05. Comparisons between two groups were analyzed using the paired or unpaired Student t test. To analyze patient and mouse survival curves, a log-rank (Mantel–Cox) test was used.

Key reagents or resources used in this study were listed in Supplementary Table S2.

Reduced cNK and LrNK cells in HCC tissues show dysfunctional features

To evaluate the functional features of the two subsets of NK cells in liver tumor microenvironment, a murine model of orthotopical HCC was created, and NK cells were harvested for flow cytometry analyses (Fig. 1A). We found that compared with nontumor liver tissues, CD3NKp46+ NK cells, including CD49a+ LrNK and CD49b+ cNK cells, were significantly reduced in H22 orthotopic homograft, while the relative ratio of LrNK cells in total CD3NKp46+ NK cells increased in tumoral tissues (Fig. 1B). Both tumoral cNK and LrNK cells also demonstrated greatly decreased expression of the activation marker CD69 and activating receptor NKG2D, displayed as reductions in both percentages and expression intensities of the positive cells (Fig. 1C). Moreover, both CD49a+ LrNK and CD49b+ cNK cells had markedly reduced production of the effector molecules IFNγ and CD107a (Fig. 1C; Supplementary Fig. S1A).

Figure 1.

Reduced LrNK and cNK in HCC tissues has an exhausted phenotype. A, Schemes of experimental design and flow cytometric gating strategy of cNK and LrNK cells. B and C, Gating on CD3NKp46+ cells, CD49a+ LrNK, and CD49b+ cNK cells was analyzed by flow cytometry on aspects of cell count, percentage in lymphocyte population or total NK cells (B) and the expression of CD69, NKG2D, and IFNγ (C) in orthotopically transplanted H22 tumors. Each symbol represents data from an individual mouse; mean bars and SEM are presented, n = 5. Two independent experiments were repeated. D, CD3CD56bright and CD3CD56dim cells in human HCC tumoral or paratumoral tissues were detected and analyzed by flow cytometry. Each symbol represents data from an individual subject. E and F, Representative multiplex fluorescent IHC staining images (left), summary plots for all individuals analyzed in each group (right) showing cell number and percentage of CXCR6+ and CXCR6 NK cells (E) and Spearman correlation of NK cell number with patient survival (F). *, P < 0.05; **, P < 0.01; ***, P < 0.001; ****, P < 0.0001. See also Supplementary Fig. S1.

Figure 1.

Reduced LrNK and cNK in HCC tissues has an exhausted phenotype. A, Schemes of experimental design and flow cytometric gating strategy of cNK and LrNK cells. B and C, Gating on CD3NKp46+ cells, CD49a+ LrNK, and CD49b+ cNK cells was analyzed by flow cytometry on aspects of cell count, percentage in lymphocyte population or total NK cells (B) and the expression of CD69, NKG2D, and IFNγ (C) in orthotopically transplanted H22 tumors. Each symbol represents data from an individual mouse; mean bars and SEM are presented, n = 5. Two independent experiments were repeated. D, CD3CD56bright and CD3CD56dim cells in human HCC tumoral or paratumoral tissues were detected and analyzed by flow cytometry. Each symbol represents data from an individual subject. E and F, Representative multiplex fluorescent IHC staining images (left), summary plots for all individuals analyzed in each group (right) showing cell number and percentage of CXCR6+ and CXCR6 NK cells (E) and Spearman correlation of NK cell number with patient survival (F). *, P < 0.05; **, P < 0.01; ***, P < 0.001; ****, P < 0.0001. See also Supplementary Fig. S1.

Close modal

To verify the findings with the murine orthotopical HCC model, human HCC-infiltrated mononuclear cells were isolated and phenotyped by flow cytometry. Confirming our above findings, the number of CD3CD56+ NK cells greatly decreased in human HCC tissues (Supplementary Fig. S1B and S1C). Specifically, the number of both CD3CD56bright cells and CD3CD56low cells, which were reported to resemble with LrNK and cNK cells in mice (9), was reduced (Fig. 1D). Consistently, multiplex fluorescent IHC with human HCC tissues further confirmed the reduced number of CXCR6+ NK and CXCR6 NK cells, which are considered to be the majority of human LrNK and cNK cells, respectively (Fig. 1E; Supplementary Fig. S1D; refs. 9, 21). More importantly, high number of CXCR6 NK cells showed a positive correlation with HCC patient survival (Fig. 1F). Collectively, these results show that both LrNK and cNK cells might be affected during HCC development.

Tim-3 expression is elevated in tumor-infiltrated cNK and LrNK cells

Given the critical roles of checkpoint receptors, such as PD-1, Tim-3, and TIGIT [T-cell immunoreceptor with Ig and immunoreceptor tyrosine–based inhibition motifs (ITIM) domains], in regulating antitumor immunity (22–24), we hypothesized that checkpoint receptors might involve in the dysregulation of NK cells in HCC. Thus, we analyzed the expression of immune checkpoint molecules on CD3CD56+ cells in paratumor and tumor regions of human HCC tissues. Flow cytometric analyses showed that expression of Tim-3, TIGIT, PD-1, and LAG-3 (lymphocyte activation gene 3) on CD3CD56+ cells in HCC tumor tissues was significantly higher than those in paratumor tissues. Among them, Tim-3 was the most abundantly expressed molecule on tumor-infiltrated NK cells (Fig. 2A and B; Supplementary Fig. S2A). Furthermore, the upregulation of Tim-3 was found in both CD3CD56bright and CD3CD56low cells. The augmented Tim-3 expression was also verified with multiplex fluorescent IHC in a human HCC tissue array (Fig. 2C; Supplementary Fig. S2B). Percentage and intensity of Tim-3 were significantly upregulated in both CXCR6 and CXCR6+ NK cells in human tumor tissues (Fig. 2D). Moreover, Tim-3–positive percentages of either CXCR6 or CXCR6+ CD3CD56+ cells showed a tendency of negative correlation with patient survival (Fig. 2E). Consistently, in mouse orthotopical HCC tissues, both tumor-infiltrated CD49a+ LrNK and CD49b+ cNK cells showed higher Tim-3 expression than those cells in normal mouse liver tissues (Fig. 2F; Supplementary Fig. S2C). Together, HCC-infiltrated cNK and LrNK cells in mice or cNK/LrNK–like cells in human had higher Tim-3 expression, suggesting its potential involvement in the exhaustion of NK cells in tumor microenvironment.

Figure 2.

Elevated Tim-3 in tumor-infiltrated LrNK and cNK. A, Representative pseudocolor flow cytometry plots from individual paratumoral and tumoral tissues (top) and summary boxplots for all individuals analyzed in each group (bottom) showing expression of Tim-3, TIGIT, PD-1, and LAG-3 on NK cells gated in CD45+CD3CD56+ cells. Each symbol represents data from an individual subject. B, Tim-3 expression on CD3CD56dim (top) and CD3CD56bright (bottom) cells in paratumoral and tumoral HCC tissues was detected by flow cytometry. Each symbol represents data from an individual subject. C–E, Representative multiplex IHC images (C), summary plots for all individuals analyzed in each group (D) showing expression percentage and intensity of Tim-3 on CXCR6+ and CXCR6 NK, and Spearman correlation of Tim-3 expression with patient survival (E). Each symbol represents data from an individual subject. F, Tim-3 expression on CD49b+ cNK and CD49a+ LrNK cells from orthotopically transplanted liver tumors was detected by flow cytometry. Each symbol represents data from an individual mouse and error bars represent SEM per group in one experiment. Two independent experiments were repeated. *, P < 0.05; **, P < 0.01; ***, P < 0.001; ****, P < 0.0001. See also Supplementary Fig. S2.

Figure 2.

Elevated Tim-3 in tumor-infiltrated LrNK and cNK. A, Representative pseudocolor flow cytometry plots from individual paratumoral and tumoral tissues (top) and summary boxplots for all individuals analyzed in each group (bottom) showing expression of Tim-3, TIGIT, PD-1, and LAG-3 on NK cells gated in CD45+CD3CD56+ cells. Each symbol represents data from an individual subject. B, Tim-3 expression on CD3CD56dim (top) and CD3CD56bright (bottom) cells in paratumoral and tumoral HCC tissues was detected by flow cytometry. Each symbol represents data from an individual subject. C–E, Representative multiplex IHC images (C), summary plots for all individuals analyzed in each group (D) showing expression percentage and intensity of Tim-3 on CXCR6+ and CXCR6 NK, and Spearman correlation of Tim-3 expression with patient survival (E). Each symbol represents data from an individual subject. F, Tim-3 expression on CD49b+ cNK and CD49a+ LrNK cells from orthotopically transplanted liver tumors was detected by flow cytometry. Each symbol represents data from an individual mouse and error bars represent SEM per group in one experiment. Two independent experiments were repeated. *, P < 0.05; **, P < 0.01; ***, P < 0.001; ****, P < 0.0001. See also Supplementary Fig. S2.

Close modal

Tim-3 interference improves the function of both LrNK and cNK cells

Augmented Tim-3 expression in tumor-infiltrated cNK and LrNK cells negatively correlates to HCC patient survival, which strongly motivated us to explore the effects of Tim-3 on cNK and LrNK cells. Thus, we established Tim-3 knockout (Tim-3−/−) mice by TALEN-mediated gene-targeting strategy. DNA sequencing verified that the loss of two base pairs (CC) caused framework mutation at 38th aa, and stopped codon occurred at 69th aa of Tim-3 (Supplementary Fig. S3A and S3B). Tim-3 deficiency markedly enhanced IFNγ and TNFα production and surface CD107a expression of liver CD3NK1.1+ cells as compared with those in WT mice (Fig. 3A), and these were also seen in both CD49b+ cNK and CD49a+ LrNK cells (Fig. 3B). Importantly, sorted CD49b+ cNK and CD49a+ LrNK cells from Tim-3−/− mice showed higher ex vivo–killing activity against B16 melanoma cells than those from WT mice (Fig. 3C). In accordance, the percentage of CD11b+CD27 subset in CD3NK1.1+ cells, which represents the mature population with a high cytotoxicity (25), was significantly higher in the liver tissues of Tim-3−/− mice than that in WT mice (Fig. 3D). Furthermore, shRNA lentivirus–mediated Tim-3 knockdown in CD3Dx5+ cells had enhanced capacity of cytokine production, surface CD107a expression, as well as an increased ratio of mature cell population when subcutaneously injected with the murine liver tumor cells H22 into BALB/c mice (Supplementary Fig. S4A—S4C).

Figure 3.

Tim-3 interference promotes the functions of cNK and LrNK cells. A and B, Expression of IFNγ, TNFα, and CD107a in intrahepatic CD3NK1.1+ cells (A) and CD49b+ cNK or CD49a+ LrNK cells (B) from WT or Tim-3 knockout (Tim-3−/−) mice was detected by flow cytometry. C, Cytotoxic activity of sorted CD49b+ cNK or CD49a+ LrNK cells from WT or Tim-3 knockout (Tim-3−/−) mice was analyzed by luciferase reporter assay after coculturing with B16 melanoma cells expressing firefly luciferase. D, Expression of CD27 and CD11b on intrahepatic CD3NK1.1+ cells from WT or Tim-3 knockout (Tim-3−/−) mice was analyzed by flow cytometry. E and F, Expression of IFNγ, TNFα, and CD107a was analyzed by flow cytometry in tumor-infiltrated CD3NK1.1+ cells (E) and CD49b+ cNK and CD49a+LrNK cells (F) pretreated with anti–Tim-3–neutralizing antibodies (αTim-3; 5 μg/mL for 6 hours) or isotype IgG. Each symbol represents data from an individual mouse. Two independent experiments were repeated. G, Flow cytometry analysis of IFNγ and CD107a expression in human HCC-infiltrated NK cells gated on CD45+CD3CD56bright and CD3CD56low cells with or without treatment of anti-Tim-3 antibody after stimulation with PMA plus ionomycin. Each symbol represents data from an individual patient. ns, nonsignificant; *, P < 0.05; **, P < 0.01; ***, P < 0.001. See also Supplementary Fig. S3 and S4.

Figure 3.

Tim-3 interference promotes the functions of cNK and LrNK cells. A and B, Expression of IFNγ, TNFα, and CD107a in intrahepatic CD3NK1.1+ cells (A) and CD49b+ cNK or CD49a+ LrNK cells (B) from WT or Tim-3 knockout (Tim-3−/−) mice was detected by flow cytometry. C, Cytotoxic activity of sorted CD49b+ cNK or CD49a+ LrNK cells from WT or Tim-3 knockout (Tim-3−/−) mice was analyzed by luciferase reporter assay after coculturing with B16 melanoma cells expressing firefly luciferase. D, Expression of CD27 and CD11b on intrahepatic CD3NK1.1+ cells from WT or Tim-3 knockout (Tim-3−/−) mice was analyzed by flow cytometry. E and F, Expression of IFNγ, TNFα, and CD107a was analyzed by flow cytometry in tumor-infiltrated CD3NK1.1+ cells (E) and CD49b+ cNK and CD49a+LrNK cells (F) pretreated with anti–Tim-3–neutralizing antibodies (αTim-3; 5 μg/mL for 6 hours) or isotype IgG. Each symbol represents data from an individual mouse. Two independent experiments were repeated. G, Flow cytometry analysis of IFNγ and CD107a expression in human HCC-infiltrated NK cells gated on CD45+CD3CD56bright and CD3CD56low cells with or without treatment of anti-Tim-3 antibody after stimulation with PMA plus ionomycin. Each symbol represents data from an individual patient. ns, nonsignificant; *, P < 0.05; **, P < 0.01; ***, P < 0.001. See also Supplementary Fig. S3 and S4.

Close modal

To further validate the regulation of Tim-3 in HCC-infiltrated NK cells, CD3NK1.1+ cells were isolated from murine orthotopical HCC tissues and treated with Tim-3–neutralizing antibody or control IgG ex vivo. As expected, Tim-3 neutralization significantly enhanced the production of IFNγ and TNFα and surface CD107a expression in the total tumor-infiltrated NK cells (Fig. 3E). The enhancement was seen in both CD49b+ cNK and CD49a+ LrNK cells (Fig. 3F). Similarly, Tim-3 neutralization significantly increased IFNγ production and CD107a mobilization of tumor-infiltrated CD3CD56+ cells isolated from human HCC tissues (Fig. 3G). In accordance, Tim-3 blockade greatly restored the cytotoxic activity of mouse HCC-infiltrated NK cells, as evidenced by a marked increase in apoptotic/necrotic population of target cells (Supplementary Fig. S4D). Interestingly, we found that Tim-3 blockade also enhanced the proliferation of both LrNK and cNK cells from normal mouse liver tissues or HCC tissues (Supplementary Fig. S4E). In addition, we analyzed the functional differences between Tim-3+ and Tim-3 NK cells. As shown in Supplementary Fig. S4F, Tim-3+ NK cells showed weaker cytokine secretion, cytotoxicity activity, and less activation than Tim-3 NK cells from orthotopic mouse liver tumor. Together, these data strongly supported the notion that Tim-3 exerts an inhibitory regulation on the function of HCC-infiltrated NK cells, including cNK and LrNK cells.

Phosphatidylserine engagement–induced Tim-3 phosphorylation contributes to the dysregulation of hepatic NK cells including LrNK and cNK cells

Our ex vivo and in vivo data highlighted the critical role of Tim-3 in negative regulation of tumor-infiltrated NK cells. Tim-3 is an immune receptor with several different ligands (galectin 9, phosphatidylserine/PtdSer, HMGB1, CEACAM1, etc.), by which the ligation leads to phosphorylation of its cytoplasmic domain and subsequently triggers the downstream signaling pathway (24). Hence, we next attempted to identify the known ligands involved in Tim-3 regulation on tumor-infiltrated NK cells. We first detected the presence of galectin 9, PtdSer, HMGB1, and CEACAM1 in liver tumor tissues. Flow cytometric analyses showed that albeit all tested ligands could be detected in mouse orthotopic HCC tissues, PtdSer was the one with substantial exposure on both tumor cells and HCC-infiltrated cNK and LrNK cells (Fig. 4A). PtdSer exposure was elevated in IL15-stimulated NKL cells, a human NK cell line (Fig. 4B), indicating activation-dependent exposure of this Tim-3 ligand. Moreover, treatment of exogenous PtdSer significantly enhanced phosphorylation of immunoprecipitated Tim-3 of NKL cells, as monitored by Western blot (Fig. 4C) and immunofluorescence (Supplementary Fig. S5A). PtdSer treatment ex vivo also suppressed TNFα production and CD107a mobilization of cNK and LrNK cells from WT mice, but had no effects upon Tim-3 knockout (Fig. 4D). In accordance, PtdSer treatment significantly inhibited cytokine secretion by NK-92 cells (Supplementary Fig. S5B). Moreover, both CD49a+ LrNK and CD49b+ cNK cells of mutant Tim-3 knock-in mice (with a deleted terminal portion of cytoplasmic domain; ref. 26) showed elevations of TNFα and IFNγ production and CD107a mobilization, as compared with WT mice (Supplementary Fig. S5C–S5E). Thus, these results demonstrated that Tim-3 is induced to be phosphorylated by PtdSer engagement and is responsible for PtdSer-mediated regulation in cNK and LrNK cells.

Figure 4.

Tim-3 phosphorylation induced by PtdSer leads to the dysfunction of NK cells. A, Expression of PtdSer, CEACAM1, HMGB1, and galectin-9 on tumor cells, CD49b+ cNK, or CD49a+ LrNK cells from orthotopically transplanted liver tumors in mice was analyzed by flow cytometry. B, Intensity of Annexin V on NKL cells stimulated with IL15 (20 ng/mL for 6 hours) or PBS was analyzed by flow cytometry. C, The phosphorylation of Tim-3 in NKL cells stimulated with PtdSer or PBS was detected by co-IP assay. D, Expression of TNFα or CD107a in total liver CD3NK1.1+, CD49b+ cNK, and CD49a+ LrNK cells from Tim-3−/− or WT mice was analyzed by flow cytometry. Each symbol represents data from an individual mouse; n = 3, mean ± SEM; error bars, SEM per group in one experiment. Two independent experiments were repeated. *, P < 0.05; **, P < 0.01; ***, P < 0.001; ****, P < 0.01. See also Supplementary Fig. S5.

Figure 4.

Tim-3 phosphorylation induced by PtdSer leads to the dysfunction of NK cells. A, Expression of PtdSer, CEACAM1, HMGB1, and galectin-9 on tumor cells, CD49b+ cNK, or CD49a+ LrNK cells from orthotopically transplanted liver tumors in mice was analyzed by flow cytometry. B, Intensity of Annexin V on NKL cells stimulated with IL15 (20 ng/mL for 6 hours) or PBS was analyzed by flow cytometry. C, The phosphorylation of Tim-3 in NKL cells stimulated with PtdSer or PBS was detected by co-IP assay. D, Expression of TNFα or CD107a in total liver CD3NK1.1+, CD49b+ cNK, and CD49a+ LrNK cells from Tim-3−/− or WT mice was analyzed by flow cytometry. Each symbol represents data from an individual mouse; n = 3, mean ± SEM; error bars, SEM per group in one experiment. Two independent experiments were repeated. *, P < 0.05; **, P < 0.01; ***, P < 0.001; ****, P < 0.01. See also Supplementary Fig. S5.

Close modal

Tim-3–driven PI3K/AKT/mTORC1 signaling mediates its inhibition in NK cells

Tim-3 can bind the PI3K adaptor p85 in T cells to trigger either stimulatory or inhibitory TCR-mediated signaling (27, 28). Because PI3K signaling pathway is critical for NK cell biology, including development, maturation, priming, and cytotoxicity (29), we thus hypothesized that Tim-3 might interact with and regulate PI3K pathway in NK cells. We first evaluated the interaction between p85 and Tim-3 in NK92 cells. As shown in Fig. 5A, immunoprecipitation (IP) with either p85 or Tim-3 antibody showed the interaction of endogenous p85 and Tim-3. Immunofluorescence staining confirmed the colocalization of p85 and Tim-3 in NKL cells, which was further enhanced by IL15 stimulation (Fig. 5B). Because interaction of p85 with p110, the catalytic subunit of PI3K, is critical for the stability and activation of PI3K (30), we next estimated the effect of Tim-3 on p85–p110 interaction. IP results showed that overexpression of Tim-3 impaired interaction of p85 and p110 in HEK293 cells (Fig. 5C). Conversely, anti-Tim-3–blocking antibody enhanced the colocalization of p85 to p110 (Fig. 5D). All these results suggested that Tim-3 potentially regulates PI3K pathway by competitively binding with p85.

Figure 5.

Tim-3 competitively binds with p85 and interrupts PI3K/AKT/mTORC1 pathway in NK cells. A and B, Interaction of Tim-3 with PI3K p85 in NK-92 cells was analyzed by co-IP (A) and immunofluorescence staining (B). C and D, Interaction of PI3K p85 with p110 in HEK-293T cells transfected with pcTim-3 or pcDNA3.0 was analyzed by co-IP (C) and immunofluorescence staining (10 imaging cells were statistically analyzed; D). E, The content of PIP3 and PIP2 in NKL cells treated with PMA plus ionomycin or IL15 was quantified by ELISA kit and the relative ratio of PIP3 to PIP2 was calculated. F, Immunoblot of phosphorylated or total Akt, mTOR, Raptor, and S6 in NKL cells treated with αTim-3 or PBS plus IL15 stimulation. G, The relative band intensity of p-mTOR, p-AKT, p-Raptor, and p-S6 to the corresponding total protein in F was analyzed by ImageJ. H, Immunoblot of Akt or S6 in NKL cells treated with αTim-3 or PBS, or in the presence of rapamycin or A66(10 mmol/L for 6 hours). I, The relative band intensity of p-AKT and p-S6 to the corresponding total protein in G was analyzed by ImageJ. Data are representative of three independent experiments. *, P < 0.05; **, P < 0.01; n.s., no significance.

Figure 5.

Tim-3 competitively binds with p85 and interrupts PI3K/AKT/mTORC1 pathway in NK cells. A and B, Interaction of Tim-3 with PI3K p85 in NK-92 cells was analyzed by co-IP (A) and immunofluorescence staining (B). C and D, Interaction of PI3K p85 with p110 in HEK-293T cells transfected with pcTim-3 or pcDNA3.0 was analyzed by co-IP (C) and immunofluorescence staining (10 imaging cells were statistically analyzed; D). E, The content of PIP3 and PIP2 in NKL cells treated with PMA plus ionomycin or IL15 was quantified by ELISA kit and the relative ratio of PIP3 to PIP2 was calculated. F, Immunoblot of phosphorylated or total Akt, mTOR, Raptor, and S6 in NKL cells treated with αTim-3 or PBS plus IL15 stimulation. G, The relative band intensity of p-mTOR, p-AKT, p-Raptor, and p-S6 to the corresponding total protein in F was analyzed by ImageJ. H, Immunoblot of Akt or S6 in NKL cells treated with αTim-3 or PBS, or in the presence of rapamycin or A66(10 mmol/L for 6 hours). I, The relative band intensity of p-AKT and p-S6 to the corresponding total protein in G was analyzed by ImageJ. Data are representative of three independent experiments. *, P < 0.05; **, P < 0.01; n.s., no significance.

Close modal

We further evaluated the modulation of Tim-3 on the downstream signaling effectors of PI3K. We first detected the ratio of phosphatidylinositol-3,4,5- triphosphate (PIP3) to phosphatidylinositol-4,5-bisphosphate (PIP2), which is directly regulated by PI3K, in NKL cells with or without pretreatment of Tim-3–blocking antibody. Figure 5E shows that Tim-3 blockade upregulated the ratio of PIP3 to PIP2 in unstimulated, PMA plus ionomycin-stimulated, and IL15-stimulated NKL cells. Accordingly, phosphorylation of downstream signaling molecules (Akt, mTOR, Raptor, and S6) was significantly higher in anti-Tim-3 antibody–pretreated NKL cells than that of control cells with or without IL15 stimulation (Fig. 5F and G). Moreover, A66, a PI3K p110α isoform–selective inhibitor, or rapamycin, an mTOR inhibitor, reversed the enhancement of Akt or S6 phosphorylation in NKL cells by Tim-3 blockade (Fig. 5H and I), further validating the cascade of Tim-3–mediated signaling pathway.

Furthermore, the importance of PI3K/Akt/mTOR pathway in NK-cell function was evidenced in human NK cell lines. Pretreatment with the PI3K inhibitor A66 almost abrogated the enhancement of Tim-3 blockade on perforin, granzyme, and TNFα production in NKL cells (Fig. 6A), while rapamycin treatment ameliorated the elevation of TNFα, IFNγ, and granzyme levels in NK92 cells mediated by Tim-3 blockade (Supplementary Fig. S6). Thus, above data suggested that by competitively binding with p85,Tim-3 inhibits NK-cell function via interfering PI3K/Akt/mTOR pathway.

Figure 6.

PI3K/AKT/mTORC1 pathway is responsible for Tim-3 inhibition on cNK and LrNK cells. A, Flow cytometry analysis of perforin, granzyme B, and TNFα expression in NKL cells treated with αTim-3 in the presence of A66. B, Interaction of PI3K p85 with Tim-3 in HEK-293T cells transfected with pcTim-3 or its mutant or truncated constructs was analyzed by co-IP. C, Flow cytometry analysis of p-AKT or p-mTOR in intrahepatic CD49a+ LrNK or CD49b+ cNK cells from Tim-3−/− mice transfected with WT or mutated Tim3 constructs. Each symbol represents data from one batch of transfected cells or an individual mouse; error bars, SEM per group in one experiment. Two independent experiments were repeated. *, P < 0.05; **, P < 0.01; ***, P < 0.001; ns, no significance. See also Supplementary Fig. S6. Fig. 6B republished with permission of American Society for Microbiology, from Phosphotyrosine-dependent coupling of Tim-3 to T-cell receptor signaling pathways, Lee, et al., 2011;31(19).

Figure 6.

PI3K/AKT/mTORC1 pathway is responsible for Tim-3 inhibition on cNK and LrNK cells. A, Flow cytometry analysis of perforin, granzyme B, and TNFα expression in NKL cells treated with αTim-3 in the presence of A66. B, Interaction of PI3K p85 with Tim-3 in HEK-293T cells transfected with pcTim-3 or its mutant or truncated constructs was analyzed by co-IP. C, Flow cytometry analysis of p-AKT or p-mTOR in intrahepatic CD49a+ LrNK or CD49b+ cNK cells from Tim-3−/− mice transfected with WT or mutated Tim3 constructs. Each symbol represents data from one batch of transfected cells or an individual mouse; error bars, SEM per group in one experiment. Two independent experiments were repeated. *, P < 0.05; **, P < 0.01; ***, P < 0.001; ns, no significance. See also Supplementary Fig. S6. Fig. 6B republished with permission of American Society for Microbiology, from Phosphotyrosine-dependent coupling of Tim-3 to T-cell receptor signaling pathways, Lee, et al., 2011;31(19).

Close modal

Our in vitro data showed that PtdSer triggered Tim-3 phosphorylation, leading to downregulation of TNFα and CD107a in cNK and LrNK cells (Fig. 4C and D). We then asked whether the PtdSer binding site (R111A) and cytoplasmic phosphorylation sites (Y265F, Y272F, and Y265/272F) of Tim-3 are involved in the regulation of PI3K signaling. Therefore, coimmunoprecipitation (co-IP) assay was performed using cell lysates of HEK293 cells transfected with a serial of HA-tagged Tim-3 mutants. As shown in Fig. 6B, mutation in PtdSer binding site (R111A) and in cytoplasmic phosphorylation sites (Y265F, Y272F, and Y265/272F), markedly impaired the binding of HA-tagged Tim-3 with p85, indicating the importance of PtdSer binding and cytoplasmic phosphorylation in Tim-3–mediated PI3K interference. To further validate this involvement in NK regulation, we transfected WT or different mutant Tim-3 constructs into hepatic mononuclear cells from Tim-3−/− mice. Thus, WT Tim-3 significantly inhibited phosphorylation of Akt and mTOR in both CD49a+ LrNK and CD49b+ cNK cells, while mutant Tim-3 constructs in R111A, Y265/272F, or with deleted whole cytoplasmic region, partially lost the inhibitory effects (Fig. 6C). Together, these results suggested that, indeed, Tim-3 phosphorylation potently modulates PI3K/Akt/mTOR pathway in both cNK and LrNK cells.

Blockade of Tim-3 signaling in NK cells suppresses tumor growth in vivo

Our findings clearly demonstrated that Tim-3 negatively regulated the function and maturation of NK cells, which are closely related to tumor immune evasion. To elucidate the impacts of Tim-3–regulated NK cells in tumor development, we performed a series of in vivo experiments using different intervention strategies, including Tim-3 knockout, Tim-3 knockdown, and Tim-3 blockade. First, Tim-3 knockdown was performed in CD3Dx5+ cells by infection of Tim-3 shRNA lentivirus (LV-shTim-3). H22 homografts were created with Tim-3-knockdown CD3Dx5+ cells together with H22 cells (Supplementary Fig. S3A). Similar to Tim-3 knockout, Tim-3 knockdown in CD3Dx5+ cells suppressed the growth of H22 homografts (Fig. 7A). Next, we intraperitoneally injected Tim-3 blocking antibody into H22 tumor–bearing, T-cell–deficient nude mice (Fig. 7B). Compared with IgG isotype group, anti-Tim-3 antibody retarded tumor growth as indicted by marked reductions of tumor volume and weight (Fig. 7C). In contrast, anti-asialo GM1 antibody treatment, which depleted both cNK cells and LrNK cells(Supplementary Fig. S7), almost abrogated the tumor-inhibitory effects of anti-Tim-3 antibody (Fig. 7D and E). More importantly, anti-Tim-3 antibody prolonged the overall survival of orthotopical liver tumor-bearing mice, while anti-asialo GM1 antibody abrogated the effect (Fig. 7F and G). Taken together, our results suggested that Tim-3 signaling hinders the function of NK cells, and facilitates tumor progression. Our findings imply that Tim-3 may be a novel intervention target for cancer immunotherapy.

Figure 7.

Tim-3 interference on NK cells suppressed tumor progression. A, Tumor growth and weight in Balb/c mice subcutaneously transplanted with H22 cells plus purified CD3Dx5+ cells pretreated with LV-shTim-3 or LV-NC. Experiment scheme (B and D) and tumor growth and weight (C and E) of Balb/c nude mice subcutaneously transplanted with H22 cells and every other day intraperitoneally injected with anti-Tim-3 or IgG (C), or in the presence of anti-asialo GM1 antibody (E). Experiment scheme (F) and survival curves (G) of Balb/c mice orthotopically transplanted with H22 grafts in liver and treated with αTim-3, IgG or αTim-3 plus anti-asialo GM1. Each symbol represents data from an individual mouse; error bars, SEM per group in one experiment. *, P < 0.05; **, P < 0.01; ***, P < 0.001. See also Supplementary Fig. S7.

Figure 7.

Tim-3 interference on NK cells suppressed tumor progression. A, Tumor growth and weight in Balb/c mice subcutaneously transplanted with H22 cells plus purified CD3Dx5+ cells pretreated with LV-shTim-3 or LV-NC. Experiment scheme (B and D) and tumor growth and weight (C and E) of Balb/c nude mice subcutaneously transplanted with H22 cells and every other day intraperitoneally injected with anti-Tim-3 or IgG (C), or in the presence of anti-asialo GM1 antibody (E). Experiment scheme (F) and survival curves (G) of Balb/c mice orthotopically transplanted with H22 grafts in liver and treated with αTim-3, IgG or αTim-3 plus anti-asialo GM1. Each symbol represents data from an individual mouse; error bars, SEM per group in one experiment. *, P < 0.05; **, P < 0.01; ***, P < 0.001. See also Supplementary Fig. S7.

Close modal

In this study, we have demonstrated that HCC tissues have reduced number of cNK and LrNK cells. We also found that tumor-infiltrated cNK and LrNK cells show elevated Tim-3 expression, which in turn suppresses their cytokine secretion and cytotoxicity. Tim-3 competitively binds with PI3K p85 and disrupts the downstream Akt/mTORC1 signaling in NK cells. Importantly, Tim-3 blockade retarded HCC growth in a NK-dependent manner, which provides an effective pathway for immune checkpoint–based cancer therapy.

The role of NK cells in antitumor immunity is well established and has been extensively studied. A number of studies have demonstrated decreased frequency and functional deficiency of NK cells in tumor tissues. However, most of those data were accumulated before the identification of cNK and tissue-resident NK. There is paucity in knowledge with regard to the roles of tissue-resident NK in tumor development. Most recently, loss of LrNK cells are found in liver metastatic colorectal tumor tissues, which results from lactate-induced mitochondrial stress and apoptosis of LrNK cells (14). Whether this depletion also occurs in HCC and its clinical significance is unclear. Herein, we demonstrate for the first time, to our best knowledge, that the numbers of both LrNK and cNK cells were decreased in HCC tissues, and that these tumor-infiltrated NK cells displayed functional exhaustion, that is, suppressed cytokine production and cytotoxic activities. Importantly, the number of cNK population positively correlated with patient survival.

NK-cell activities are controlled by a repertoire of stimulatory or inhibitory receptors (31). Among them, immune checkpoint receptors have attracted great attention owing to the unprecedented success of immunotherapies targeting these receptors. Tim-3, one of the well-known coinhibitory receptors, is expressed on a number of immune cells, and is closely involved in many inflammatory diseases, including chronic viral infection and tumor (32–35). In this study, we clarified that, among several immune checkpoint receptors, HCC-infiltrated CD3CD56+ NK cells had the highest level of Tim-3. Moreover, Tim-3 expression on tumor-infiltrated cNK and LrNK cells was significantly higher than that in paratumor tissues. Although we only found the tendency of negative correlation between Tim-3 expression and patient survival due to the limited tissue section diameter and sample size, these results still imply the tumorigenesis role of Tim-3 on NK cells and were further supported by several lines of in vivo evidence. First, both Tim-3–deficient CD3NK1.1+ cells (including both cNK and LrNK cells) and Tim-3 knockdown NK cells significantly retarded tumor growth in the mouse models. Moreover, anti-Tim-3 antibody treatment hampered HCC growth in nude mice and prolonged the survival of HCC-homografted mice in NK-dependent manner. All these results pinpoint Tim-3 as a potential intervention site for NK-targeted immunotherapy. Given the concurrent inhibitory regulation of Tim-3 on tumor-infiltrated T cells (36–38), Tim-3 targeting therapy should have the advantage in simultaneously reinvigorating the antitumor function of both T cells and NK cells.

Tim-3 has been shown to execute versatile regulation on different immune cells in tumor microenvironment, such as suppressing effector T cells, enhancing Treg immunosuppression, and polarizing macrophage into M2 phenotype (38, 39). Here, we found that Tim-3 modulated NK cells on two functional aspects. First, Tim-3 inhibited cytokine production and cytotoxic activity of both cNK and LrNK cells. Tim-3 upregulation of tumor-infiltrated NK cells was associated with their dysfunction in both human HCC and mouse models. In line with these observations, Tim-3 knockout, knockdown, and functional blockade promoted cytokine production and cytotoxic activity by cNK cells and LrNK cells. Second, Tim-3 regulated the maturation of NK cells. Tim-3 knockout or knockdown increased the ratio of CD11b+CD27 population with more mature phenotype, which may partially explain their restored functions. Together, our findings support a notion that Tim-3 intervention can be a promising strategy to enhance the functional status of NK cells in tumor microenvironment and subsequently hamper tumor progression.

Tim-3 binds several known ligands, including galectin 9, CEACAM1, PtdSer, and HMGB1, and ignites intracellular signaling (24, 38). Here, we found that, among four known Tim-3 ligands, PtdSer was most abundantly expressed on both tumor cells and NK cells, and that PtdSer treatment inhibited the cytotoxicity and cytokine production by NK cells. Our finding is supported by an earlier report showing that PtdSer exposed to apoptotic cells leads to decreased IFNγ production of NK cells. Furthermore, we demonstrated that Tim-3 coexisted with p85 subunit of PI3K, and that NK-cell activation enhanced Tim-3-p85 colocalization. PtdSer triggered the phosphorylation of Tim-3, which in turn competitively interacted with the p85 subunit of PI3K and impeded the interaction between p85 and p110 as well as the activation of PI3K/Akt/mTOR pathway in both cNK and LrNK cells. Thus, PI3K and mTOR inhibitors abrogated Tim-3 effects on NK cells. Together, our work has, for the first time, demonstrated the interruption of PI3K complex by PtdSer-induced Tim-3 phosphorylation, apart from confirming the effects of Tim-3 on Akt/mTOR pathway in immune cells, for example, short-lived effector T cells (36), effector memory T cells (37), and macrophages (40). However, it should be noted that, unlike most of other checkpoint receptors (e.g., PD-1), Tim-3 does not possess any ITIM or immunoreceptor tyrosine–based switch motifs in the cytoplasmic tail (30, 41). It remains to be addressed how Tim-3 interacts with and interrupts PI3K complex. Moreover, because anticancer drugs targeting PI3K pathway have been tested in clinical trials (30), this study warns that the combination of Tim-3 blockade with those PI3K inhibitors should be avoided. This is because the combination may weaken the restoration of Tim-3 blockade on antitumor immune response by NK cells.

Conclusions

Tim-3 upregulation hampers cytotoxicities of tumor-infiltrated cNK and LrNK cells in a PI3K/Akt/mTOR–dependent manner. Deficiency or blockade of Tim-3 boosts antitumor immunity in a nonredundant NK-dependent manner. Our work has shed light on the novel anticancer strategy using a Tim-3–targeting immunotherapy.

No potential conflicts of interest were disclosed.

Conception and design: X. Liang, C. Ma

Development of methodology: Y. Xu, Z. Wu

Acquisition of data (provided animals, acquired and managed patients, provided facilities, etc.): S. Tan, Y. Xu, Z. Wang, T. Wang, X. Song, X. Guo, J. Peng, J. Zhang, Y. Liang, X. Liang

Analysis and interpretation of data (e.g., statistical analysis, biostatistics, computational analysis): Y. Xu, X. Du, N. Li, L. Gao, C. Ma

Writing, review, and/or revision of the manuscript: N. Li, X. Liang, C. Ma

Administrative, technical, or material support (i.e., reporting or organizing data, constructing databases): Y. Xu, J. Lu, J. Peng, C. Gao, C. Li

Study supervision: X. Liang, C. Ma

The authors thank the Collaborative Innovation Center of Technology and Equipment for Biological Diagnosis and Therapy in Universities of Shandong (Jinan, China) for their support. FCM and immunofluorensence images were performed in Advanced Medical Research Institute, Shandong University (Jinan, China). This work was supported by grants from the National Science Foundation of China (No. 81830017), the National Key Research and Development Program (No. 2016YFE0127000, 2018YFE0126500), the National Natural Science Fund for Outstanding Youth Fund (No. 81425012), the National Science Foundation of China (No. 81672425, 31600714), Key Research & Development Plan of Shandong Province (2016ZDJS07A17, 2018YFJH0503, 2017GSF18185).

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

1.
Siegel
R
,
Naishadham
D
,
Jemal
A
. 
Cancer statistics, 2012
.
CA Cancer J Clin
2012
;
62
:
10
29
.
2.
Greten
TF
,
Wang
XW
,
Korangy
F
. 
Current concepts of immune based treatments for patients with HCC: from basic science to novel treatment approaches
.
Gut
2015
;
64
:
842
8
.
3.
Papatheodoridis
GV
,
Chan
HL
,
Hansen
BE
,
Janssen
HL
,
Lampertico
P
. 
Risk of hepatocellular carcinoma in chronic hepatitis B: assessment and modification with current antiviral therapy
.
J Hepatol
2015
;
62
:
956
67
.
4.
Koo
J
,
Wu
X
,
Mao
Z
,
Khuri
FR
,
Sun
SY
. 
Rictor undergoes glycogen synthase kinase 3 (GSK3)-dependent, FBXW7-mediated ubiquitination and proteasomal degradation.
J Biol Chem
2015
;
290
:
14120
9
.
5.
Male
V
,
Stegmann
KA
,
Easom
NJ
,
Maini
MK
. 
Natural killer cells in liver disease
.
Semin Liver Dis
2017
;
37
:
198
209
.
6.
Sun
C
,
Sun
H
,
Zhang
C
,
Tian
Z
. 
NK cell receptor imbalance and NK cell dysfunction in HBV infection and hepatocellular carcinoma
.
Cell Mol Immunol
2015
;
12
:
292
302
.
7.
Kamiya
T
,
Chang
YH
,
Campana
D
. 
Expanded and activated natural killer cells for immunotherapy of hepatocellular carcinoma
.
Cancer Immunol Res
2016
;
4
:
574
81
.
8.
Childs
RW
,
Carlsten
M
. 
Therapeutic approaches to enhance natural killer cell cytotoxicity against cancer: the force awakens
.
Nat Rev Drug Discov
2015
;
14
:
487
98
.
9.
Klose
CSN
,
Flach
M
,
Mohle
L
,
Rogell
L
,
Hoyler
T
,
Ebert
K
, et al
Differentiation of type 1 ILCs from a common progenitor to all helper-like innate lymphoid cell lineages
.
Cell
2014
;
157
:
340
56
.
10.
Sojka
DK
,
Plougastel-Douglas
B
,
Yang
L
,
Pak-Wittel
MA
,
Artyomov
MN
,
Ivanova
Y
, et al
Tissue-resident natural killer (NK) cells are cell lineages distinct from thymic and conventional splenic NK cells
.
eLife
2014
;
3
:
e01659
.
11.
Peng
H
,
Jiang
X
,
Chen
Y
,
Sojka
DK
,
Wei
H
,
Gao
X
, et al
Liver-resident NK cells confer adaptive immunity in skin-contact inflammation
.
J Clin Invest
2013
;
123
:
1444
56
.
12.
Stegmann
KA
,
Robertson
F
,
Hansi
N
,
Gill
U
,
Pallant
C
,
Christophides
T
, et al
CXCR6 marks a novel subset of T-bet(lo)Eomes(hi) natural killer cells residing in human liver
.
Sci Rep
2016
;
6
:
26157
.
13.
Zhou
J
,
Peng
H
,
Li
K
,
Qu
K
,
Wang
B
,
Wu
Y
, et al
Liver-resident NK cells control antiviral activity of hepatic T cells via the PD-1-PD-L1 axis
.
Immunity
2019
;
50
:
403
17
.
14.
Harmon
C
,
Robinson
MW
,
Hand
F
,
Almuaili
D
,
Mentor
K
,
Houlihan
DD
, et al
Lactate-mediated acidification of tumor microenvironment induces apoptosis of liver-resident NK cells in colorectal liver metastasis
.
Cancer Immunol Res
2019
;
7
:
335
46
.
15.
Fuertes Marraco
SA
,
Neubert
NJ
,
Verdeil
G
,
Speiser
DE
. 
Inhibitory receptors beyond T cell exhaustion
.
Front Immunol
2015
;
6
:
310
.
16.
Ribas
A
,
Wolchok
JD
. 
Cancer immunotherapy using checkpoint blockade
.
Science
2018
;
359
:
1350
5
.
17.
Benson
DM
 Jr.
,
Bakan
CE
,
Mishra
A
,
Hofmeister
CC
,
Efebera
Y
,
Becknell
B
, et al
The PD-1/PD-L1 axis modulates the natural killer cell versus multiple myeloma effect: a therapeutic target for CT-011, a novel monoclonal anti-PD-1 antibody
.
Blood
2010
;
116
:
2286
94
.
18.
Liu
Y
,
Cheng
Y
,
Xu
Y
,
Wang
Z
,
Du
X
,
Li
C
, et al
Increased expression of programmed cell death protein 1 on NK cells inhibits NK-cell-mediated anti-tumor function and indicates poor prognosis in digestive cancers
.
Oncogene
2017
;
36
:
6143
53
.
19.
Hsu
J
,
Hodgins
JJ
,
Marathe
M
,
Nicolai
CJ
,
Bourgeois-Daigneault
MC
,
Trevino
TN
, et al
Contribution of NK cells to immunotherapy mediated by PD-1/PD-L1 blockade
.
J Clin Invest
2018
;
128
:
4654
68
.
20.
Seo
H
,
Kim
BS
,
Bae
EA
,
Min
BS
,
Han
YD
,
Shin
SJ
, et al
IL21 Therapy combined with PD-1 and Tim-3 blockade provides enhanced NK cell antitumor activity against MHC class I-deficient tumors
.
Cancer Immunol Res
2018
;
6
:
685
95
.
21.
Nikzad
R
,
Angelo
LS
,
Aviles-Padilla
K
,
Le
DT
,
Singh
VK
,
Bimler
L
, et al
Human natural killer cells mediate adaptive immunity to viral antigens
.
Sci Immunol
2019
;
4
.
DOI: 10.1126/sciimmunol.aat8116
.
22.
Zhang
Q
,
Bi
J
,
Zheng
X
,
Chen
Y
,
Wang
H
,
Wu
W
, et al
Blockade of the checkpoint receptor TIGIT prevents NK cell exhaustion and elicits potent anti-tumor immunity
.
Nat Immunol
2018
;
19
:
723
32
.
23.
Ohs
I
,
Ducimetiere
L
,
Marinho
J
,
Kulig
P
,
Becher
B
,
Tugues
S
. 
Restoration of natural killer cell antimetastatic activity by IL12 and checkpoint blockade
.
Cancer Res
2017
;
77
:
7059
71
.
24.
Das
M
,
Zhu
C
,
Kuchroo
VK
. 
Tim-3 and its role in regulating anti-tumor immunity
.
Immunol Rev
2017
;
276
:
97
111
.
25.
Chiossone
L
,
Chaix
J
,
Fuseri
N
,
Roth
C
,
Vivier
E
,
Walzer
T
. 
Maturation of mouse NK cells is a 4-stage developmental program
.
Blood
2009
;
113
:
5488
96
.
26.
Cho
JL
,
Roche
MI
,
Sandall
B
,
Brass
AL
,
Seed
B
,
Xavier
RJ
, et al
Enhanced Tim3 activity improves survival after influenza infection
.
J Immunol
2012
;
189
:
2879
89
.
27.
Lee
J
,
Su
EW
,
Zhu
C
,
Hainline
S
,
Phuah
J
,
Moroco
JA
, et al
Phosphotyrosine-dependent coupling of Tim-3 to T-cell receptor signaling pathways
.
Mol Cell Biol
2011
;
31
:
3963
74
.
28.
Tomkowicz
B
,
Walsh
E
,
Cotty
A
,
Verona
R
,
Sabins
N
,
Kaplan
F
, et al
TIM-3 suppresses anti-CD3/CD28-Induced TCR activation and IL-2 expression through the NFAT signaling pathway
.
PLoS One
2015
;
10
:
e0140694
.
29.
Ali
AK
,
Nandagopal
N
,
Lee
SH
. 
IL-15-PI3K-AKT-mTOR: a critical pathway in the life journey of natural killer cells
.
Front Immunol
2015
;
6
:
355
.
30.
Fruman
DA
,
Chiu
H
,
Hopkins
BD
,
Bagrodia
S
,
Cantley
LC
,
Abraham
RT
. 
The PI3K pathway in human disease
.
Cell
2017
;
170
:
605
35
.
31.
Raulet
DH
,
Marcus
A
,
Coscoy
L
. 
Dysregulated cellular functions and cell stress pathways provide critical cues for activating and targeting natural killer cells to transformed and infected cells
.
Immunol Rev
2017
;
280
:
93
101
.
32.
Ju
Y
,
Hou
N
,
Meng
J
,
Wang
X
,
Zhang
X
,
Zhao
D
, et al
T cell immunoglobulin- and mucin-domain-containing molecule-3 (Tim-3) mediates natural killer cell suppression in chronic hepatitis B
.
J Hepatol
2010
;
52
:
322
9
.
33.
Ju
Y
,
Hou
N
,
Zhang
XN
,
Zhao
D
,
Liu
Y
,
Wang
JJ
, et al
Blockade of Tim-3 pathway ameliorates interferon-gamma production from hepatic CD8+ T cells in a mouse model of hepatitis B virus infection
.
Cell Mol immunol
2009
;
6
:
35
43
.
34.
Li
H
,
Wu
K
,
Tao
K
,
Chen
L
,
Zheng
Q
,
Lu
X
, et al
Tim-3/galectin-9 signaling pathway mediates T-cell dysfunction and predicts poor prognosis in patients with hepatitis B virus-associated hepatocellular carcinoma
.
Hepatology
2012
;
56
:
1342
51
.
35.
Anderson
AC
,
Joller
N
,
Kuchroo
VK
. 
Lag-3, Tim-3, and TIGIT: co-inhibitory receptors with specialized functions in immune regulation
.
Immunity
2016
;
44
:
989
1004
.
36.
Avery
L
,
Filderman
J
,
Szymczak-Workman
AL
,
Kane
LP
. 
Tim-3 co-stimulation promotes short-lived effector T cells, restricts memory precursors, and is dispensable for T cell exhaustion
.
Proc Natl Acad Sci U S A
2018
;
115
:
2455
60
.
37.
Sabins
NC
,
Chornoguz
O
,
Leander
K
,
Kaplan
F
,
Carter
R
,
Kinder
M
, et al
TIM-3 engagement promotes effector memory T cell differentiation of human antigen-specific CD8 T cells by activating mTORC1
.
J Immunol
2017
;
199
:
4091
102
.
38.
Banerjee
H
,
Kane
LP
. 
Immune regulation by Tim-3
.
F1000Research
2018
;
7
:
316
.
39.
Yan
W
,
Liu
X
,
Ma
H
,
Zhang
H
,
Song
X
,
Gao
L
, et al
Tim-3 fosters HCC development by enhancing TGF-beta-mediated alternative activation of macrophages
.
Gut
2015
;
64
:
1593
604
.
40.
Yang
X
,
Jiang
X
,
Chen
G
,
Xiao
Y
,
Geng
S
,
Kang
C
, et al
T cell Ig mucin-3 promotes homeostasis of sepsis by negatively regulating the TLR response
.
J Immunol
2013
;
190
:
2068
79
.
41.
Anderson
AC
. 
Tim-3: an emerging target in the cancer immunotherapy landscape
.
Cancer Immunol Res
2014
;
2
:
393
8
.