Despite the classical function of NK cells in the elimination of tumor and of virus-infected cells, evidence for a regulatory role for NK cells has been emerging in different models of autoimmunity, transplantation, and viral infections. However, this role has not been fully explored in the context of a growing tumor. In this article, we show that NK cells can limit spontaneous cross-priming of tumor Ag-specific CD8+ T cells, leading to reduced memory responses. After challenge with MC57 cells transduced to express the model Ag SIY (MC57.SIY), NK cell–depleted mice exhibited a significantly higher frequency of SIY-specific CD8+ T cells, with enhanced IFN-γ production and cytotoxic capability. Depletion of NK cells resulted in a CD8+ T cell population skewed toward an effector memory T phenotype that was associated with enhanced recall responses and delayed tumor growth after a secondary tumor challenge with B16.SIY cells. Dendritic cells (DCs) from NK cell–depleted tumor-bearing mice exhibited a more mature phenotype. Interestingly, tumor-infiltrating and tumor-draining lymph node NK cells displayed an upregulated expression of the inhibitory molecule programmed death ligand 1 that, through interaction with programmed death-1 expressed on DCs, limited DC activation, explaining their reduced ability to induce tumor-specific CD8+ T cell priming. Our results suggest that NK cells can, in certain contexts, have an inhibitory effect on antitumor immunity, a finding with implications for immunotherapy in the clinic.

Natural killer cells are important mediators of the innate immune response against intracellular pathogens and tumors (1, 2). The recognition of target cells is mediated by a complex repertoire of activating and inhibitory receptors allowing for the discrimination between normal and infected/transformed cells. An extra level of regulation of NK cell activation is provided by diverse cytokines (3, 4) and the recognition of pathogen-associated molecular patterns through TLRs (5). NK cell activation results not only in a cytotoxic response toward target cells, but also in the secretion of IFN-γ and other cytokines and chemokines (68), rendering NK cells capable of modulating the activity of other cells of the immune system and the outcome of the immune response (911).

Although it has been widely demonstrated in various model systems that NK cells can play a positive role during immune responses against tumors and infected cells, evidence of an inhibitory role for NK cells is beginning to emerge in diverse models of viral infection (1215), transplantation (16), and autoimmunity (17). It has been reported that NK cells regulate T cell responses through multiple direct and indirect mechanisms, including NK cell–mediated killing of activated CD8+ T cells (14, 15, 1821), CD4+ T cells (18, 19, 22, 23), and also of dendritic cells (DCs) (12, 24, 25). However, little is known about a possible NK cell–mediated inhibitory/regulatory role during antitumor immune responses.

In many instances, spontaneous priming of tumor Ag-specific CD8+ T cells can occur in both human cancer patients and in murine models, through DC-mediated cross-presentation. Moreover, a growing body of evidence suggests that an inflamed tumor microenvironment that includes the presence of tumor-infiltrating CD8+ T cells has a positive prognostic role in multiple cancer types (26). However, multiple regulatory mechanisms that blunt T cell function within the tumor microenvironment arise in tumors (27). Programmed death ligand 1 (PD-L1), an inhibitory molecule frequently upregulated on tumor cells, is one of the major immunological checkpoints contributing to tumor-immune escape (28) through the inhibition of T cell activation and promotion of apoptosis of DCs (29) and CD8+ T cells (30). In several clinical trials, blockade of programmed death-1 (PD-1) or PD-L1 resulted in enhanced T cell function and improved immune-mediated tumor control (31). PD-L1 is expressed not only by tumor cells but also by tumor-infiltrating immune cells (32), and it has been shown that some patients with PD-L1 tumors can also respond to treatment (33). These and other related data have increased the motivation to identify additional regulatory mechanisms that restrain the priming or effector function of tumor-specific T cells.

In this study, we examined the functional role of NK cells during a spontaneous antitumor immune response. We show that NK cells inhibit the expansion of functional tumor-specific CD8+ T cells during the priming phase and control the frequency of effector memory T cells (TEMs) CD8+ cells, leading to a diminished recall response and reduced tumor control after a secondary tumor challenge with B16.SIY cells. The underlying mechanism involved the regulation of DC maturation, through PD-L1hi NK cells that emerged during tumor growth. Accordingly, NK cell–depleted mice showed an increased frequency of PD-1+ DCs. Our results are consistent with a model in which the presence of a growing tumor results in upregulated expression of PD-L1 on NK cells leading to a direct regulation of DC maturation, compromising CD8+ T cell priming and recall responses against tumor-derived Ags.

C57BL/6 mice (8–12 wk) were obtained from the animal facility of the School of Veterinary, University of La Plata (Argentina) and housed at the local animal facility according to National Institutes of Health guidelines. Studies have been approved by the institutional review committee.

Mouse cell lines MC57, B16.F10 (henceforth referred to as B16), and YAC1 were obtained from American Type Culture Collection and transduced to express the model antigenic peptide SIYRYYGL (SIY) that is cross-presented to CD8+ T cells through H2-Kb (34). Human cell lines (ECC-1, MDA-MB-453, PC3, Caco2, K562, and HeLa) were obtained from American Type Culture Collection. Tumor cell lines were cultured by standard procedures.

Mice were s.c. injected on the flank with 2 × 106 tumor cells or PBS (naive mice). After 6 or 11 d, animals were euthanized and peripheral blood (lysed with ACK buffer), spleens, lymph nodes (LNs), and tumors (disrupted with 1 mg/ml collagenase IV [Sigma-Aldrich] in complete DMEM for 30 min at 37°C) were collected. For NK cell transfer experiments, sorted NK cells from naive spleens were inoculated intratumorally 3 d after mice were challenged with MC57.SIY cells, and analyses were performed 6 d after tumor challenge. For recall experiments, on day 110 after primary tumor challenge, mice were s.c. injected with 5 × 106 B16.SIY cells and, the day before and 4 d later, peripheral blood samples were collected. For IFN-γ, CD107a, and tetramer staining, and for the identification/phenotypification of memory T cells, regulatory T cells (Tregs; CD25+Foxp3+CD4+), myeloid-derived suppressor cells (MDSCs; CD11b+Gr-1+), DCs (CD11c+CD3NKp46), and NK cells (NKp46+CD3), samples were collected at the indicated time points, and cell suspensions were prepared and analyzed by flow cytometry. For quantitative PCR (qPCR), tumor-draining LNs (TDLNs) and tumors were analyzed 6 d after tumor challenge. For tumor growth experiments, the longest (l) and shortest (d) diameters were measured three times per week, and the tumor volume was calculated as: (l × d2)/2.

For NK cell depletion, mice were injected i.p. with 100 μg of anti-NK1.1 (PK136; BioXCell) or isotype control (IC; C1.18.4; BioXCell) 1 d before and every 3 d after tumor challenge. For CD8+ T cell depletion, 200 μg of anti-CD8 (YTS 169.4, in-house produced) was injected 1 d before tumor challenge and once per week thereafter. Depletion was confirmed by flow cytometry in blood and tumor samples.

Nonspecific staining was blocked with anti-CD16/32 mAb (2.4G2) for mouse samples or with 10% normal mouse serum for human samples. Cells were labeled according to the experiment with the following fluorochrome-coupled Abs: CD3ε (17A.2; UCHT1), CD4 (RM4-5), NKp46 (29A1.4), CD49b (DX5), CD56 (N901), CD8 (53-6.7), CD11b (M1/70), CD11c (N418), Gr-1 (RB6.8C5), CD45R (B220, RA3-6B2), CD86 (GL-1), CD44 (IM7), CD127 (A7R34), PD-1 (29F.1A12), PD-L1 (10F.9G2; 29E.2A3), CD25 (PC61), NKG2D (CX5), KLRG1 (2F1), Ly6C (HK1.4),CD27 (LG.3A10) c-Kit (ACK2), CTLA-4 (UC10-4F10-11), CCR7 (4B12), Foxp3 (FJK-16s), and CD107a (ID4B) from Biolegend, Tonbo Biosciences, eBioscience, Beckman Coulter, or Immunotools. For tetramer staining, cells were labeled following the manufacturer’s instructions with PE-MHC class I dexamers (Immudex) consisting of murine H-2Kb complexed to SIY peptide and analyzed in the CD8+CD4B220 population. In some experiments, 5000 beads (Spherotech) were added for quantification purposes. Cell viability was determined with Zombie Green (Biolegend). Samples were acquired in a FACSCanto II-plus flow cytometer (BD Biosciences), and data analysis was conducted with FlowJo software (Tree Star). For cell sorting, spleen cells from naive mice and tumor cells were stained with lineage-specific mAbs for NK cells (CD3CD49b+ cells), DCs (CD3CD49bCD11c+B220 cells), or CD8+ T cells (CD3+CD8+ cells) and sorted in a FACSAria II-plus cell sorter (BD Biosciences).

Six days after tumor challenge, 2 × 105 cells were cultured overnight at 37°C in the presence or absence of the SIY peptide (10 μM). During the last 6 h, Golgi-Plug/Golgi-Stop reagents were added; cells were harvested and stained with anti-CD8, anti-CD4, and anti-B220 mAbs, fixed with 1% paraformaldehyde, permeabilized with Perm Buffer II (BD), stained with anti–IFN-γ mAb, and analyzed by flow cytometry.

Eleven days after tumor challenge, 105 cells were cultured for 6 h with anti-CD107a mAb and Golgi-Plug/Golgi-Stop reagents (BD), in the absence or in the presence of 105 MC57.SIY cells. Cells were then harvested, stained with anti-CD8, anti-CD4, and anti-B220 mAbs, and analyzed by flow cytometry.

A total of 1 × 106 splenocytes from naive mice was cultured for 48 h in the absence or in the presence of IL-12 (10 ng/ml; Peprotech), IL-15 (1 ng/ml; Peprotech), and IL-18 (10 ng/ml; MBL), or 1 × 105 MC57 cells, in the presence of blocking mAbs against NKG2D (2.5 μg/ml, 191004; R&D Systems) or IFN-γ (10 μg/ml, XMG1.2; Tonbo) or IC (LTF-2; Tonbo). Cells were harvested, stained with anti-CD3, anti-NKp46, and anti–PD-L1 mAbs, and analyzed by flow cytometry.

PBMCs were isolated from healthy human volunteers (provided by the Service of Transfusion Medicine of the Hospital ChurrucaVisca, Buenos Aires, Argentina) by Ficoll-Paque Plus (GE Life Sciences) gradient centrifugation, and cultured (5 × 105 cells/well) with the different human cell lines (2 × 105 cells/well) for 48 h. Cells were then harvested and stained with anti-CD3, anti-CD56, and anti–PD-L1 mAbs. Studies have been approved by the institutional review committee of the Instituto de Biología y Medicina Experimental.

Total RNA was obtained and cDNA was synthesized as described previously (35). For qPCR, the SYBR Green PCR Master Mix (Applied Biosystems) was used with the CFX96 Real-Time PCR System (Bio-Rad). Primers were as follows: IL-10 (forward: 5′-TGCTAACCGACTCCTTAATGCAGGAC-3′, reverse: 5′-CCTTGATTTCTGGGCCATGCTTCTC-3′), TGF-β (forward: 5′-AATTCCTGGCGTTACCTTGG-3′, reverse: 5′-ATCGAAAGCCCTGTATTCCG-3′), and GAPDH (forward: 5′-CAGAACATCATCCCTGCAT-3′, reverse: 5′-GTTCAGCTCTGGGATGACCTT-3′). For IFN-β, primer and probe sets from TaqMan Gene Expression Assays (Applied Biosystems) and TaqMan-based quantification were used. Results were expressed as 2−ΔCt using GAPDH as endogenous control.

DC and NK cells from naive spleens and NK cells from tumors were sorted. DCs (5 × 104 cells/well) were stimulated with LPS (25 ng/ml; Invivogen) and R848 (10 μM; Invivogen) in the absence or in the presence of 5 × 104 naive NK cells or tumor-infiltrating NK (TINK) cells treated or not with anti–PD-L1 blocking mAb (10F.9G2, 10 μg/ml). After 18 h beads were added and cells were harvested, stained with Zombie Green and for CD11c, CD86, and NKp46, and analyzed by flow cytometry.

NK cells from naive spleens and from tumors were sorted. T cells isolated from naive spleens by magnetic cell sorting with CD90.2 microbeads (Miltenyi Biotec) were stained with cell proliferation dye eFluor 670 (2.5 μM; eBioscience). T cells (1 × 105 cells/well) were stimulated with plate-bound anti-CD3 (145-2C11; 2 μg/ml; BD Pharmingen) and soluble anti-CD28 (37.51, 1 μg/ml; BD Pharmingen) mAbs in the absence or in the presence of 5 × 104 naive NK cells or TINK cells. After 72 h, cells were harvested, stained with Zombie Green and for CD3, CD8, and NKp46, and eFluor 670 dilution was analyzed by flow cytometry.

CD8+ T cells (targets) and NK cells (effectors) from naive spleens and tumors were sorted. Tumor-infiltrating CD8+ T cells (labeled with 0.25 μM eFluor 670) and naive CD8+ T cells (labeled with 2.5 μM eFluor 670) were cocultured 1:1 in the absence or in the presence of naive NK cells or TINK cells at different E:T ratios. After 6 h beads were added and cells were harvested, stained with Zombie Green, and analyzed by flow cytometry to obtain the percentage of Zombie Green low cells. Percent specific cytotoxicity was calculated as (100 − % Zombie Greenlow cells in the presence of NK cells) − (100 − % Zombie Greenlow cells in the absence of NK cells) for each condition.

Differences between data sets were analyzed with the two-sided Student t test, one-way and two-way ANOVA (Tukey or Sidak multiple comparison tests) using GraphPad Prism Software.

To assess the effect of NK cells on the spontaneous priming of tumor Ag-specific CD8+ T cells, we analyzed the frequency and effector responses of endogenous SIY-specific CD8+ T cells after s.c. implantation of MC57.SIY cells in control or NK cell–depleted mice. MC57.SIY cells naturally express the NKG2D ligand Rae1 and can be recognized and lysed by NK cells (data not shown); however, NK cell depletion did not affect the kinetics of tumor growth (Fig. 1A). In contrast, tumor rejection is completely dependent on CD8+ T cells (Fig. 1B). We observed that 6 d after tumor challenge it was possible to detect an expansion of SIY-specific CD8+ T cells in control and NK cell–depleted mice. However, by day 11, mice depleted of NK cells displayed a statistically significant increase in the frequency of SIY-specific CD8+ T cells in blood (Fig. 1C, 1D) and spleens (data not shown) compared with nondepleted mice. Such an increased frequency mirrored a heightened ability of these CD8+ T cells to degranulate in response to restimulation with MC57.SIY cells (Fig. 1E) and was also accompanied by an increased percentage of IFN-γ–producing CD8+ T cells upon restimulation with the SIY Ag (Fig. 1F). These findings were validated using the MHC class I cell line YAC1.SIY where tumor-specific CD8+ T cell responses are entirely dependent on cross-priming and are rejected in an NK cell–dependent manner. Similarly, we observed a statistically significant increase in frequency of SIY-specific CD8+ T cells 11 d after tumor challenge in NK cell–depleted versus control-treated mice (Fig. 1G). As a complementary approach, NK cells were adoptively transferred intratumorally, which resulted in a reduction in the percentage and number of SIY-specific CD8+ T cells (Fig. 1H) and IFN-γ–producing CD8+ T cells upon stimulation with SIY peptide (data not shown). These results demonstrate that NK cells can control the expansion of functional Ag-specific CD8+ T cells during the priming phase of a spontaneous antitumor immune response.

FIGURE 1.

NK cell depletion enhances tumor-specific CD8+ T cell priming. (A) Control (IC) or NK cell–depleted (anti-NK1.1) mice and (B) control or CD8+ T cell–depleted (anti-CD8) mice were inoculated s.c. with MC57.SIY cells on the flank, and tumor size was evaluated every 2–3 d. (CF) IC or anti-NK1.1–treated mice were inoculated with MC57.SIY cells or left unchallenged (no tumor), and blood was collected at the indicated time points after tumor challenge. (C) Percentage of SIY+ CD8+ T cells assessed by flow cytometry using specific tetramers. (D) Representative dot plots of SIY-tetramer staining at day 11. Numbers indicate percent of cells within the indicated gate. (E) Percentage of CD107a+ CD8+ T cells after ex vivo restimulation with MC57.SIY cells or medium. (F) Percentage of IFN-γ–producing CD8+ T cells after restimulation with soluble SIY peptide or medium. (G) Control or NK cell–depleted mice were inoculated with YAC1.SIY cells and blood was collected 11 d later. Percentage of SIY-specific CD8+ T cells assessed as in (C). (H) Mice were challenged with MC57.SIY cells and 3 d later they received either 7.5 × 104 NK cells intratumorally or PBS. Six days after tumor challenge, splenocytes were harvested. Percentage (left panel) and absolute number of SIY-specific CD8+ T cells (right panel) assessed as in (C). Data represent mean ± SEM (n = 4) and correspond to three (C, E, and F) or two (A, B, G, and H) independent experiments. *p < 0.05, **p < 0.01, ***p < 0.001, two-sided Student t test (C, G, and H) and two-way ANOVA and Tukey multiple comparison test (E and F).

FIGURE 1.

NK cell depletion enhances tumor-specific CD8+ T cell priming. (A) Control (IC) or NK cell–depleted (anti-NK1.1) mice and (B) control or CD8+ T cell–depleted (anti-CD8) mice were inoculated s.c. with MC57.SIY cells on the flank, and tumor size was evaluated every 2–3 d. (CF) IC or anti-NK1.1–treated mice were inoculated with MC57.SIY cells or left unchallenged (no tumor), and blood was collected at the indicated time points after tumor challenge. (C) Percentage of SIY+ CD8+ T cells assessed by flow cytometry using specific tetramers. (D) Representative dot plots of SIY-tetramer staining at day 11. Numbers indicate percent of cells within the indicated gate. (E) Percentage of CD107a+ CD8+ T cells after ex vivo restimulation with MC57.SIY cells or medium. (F) Percentage of IFN-γ–producing CD8+ T cells after restimulation with soluble SIY peptide or medium. (G) Control or NK cell–depleted mice were inoculated with YAC1.SIY cells and blood was collected 11 d later. Percentage of SIY-specific CD8+ T cells assessed as in (C). (H) Mice were challenged with MC57.SIY cells and 3 d later they received either 7.5 × 104 NK cells intratumorally or PBS. Six days after tumor challenge, splenocytes were harvested. Percentage (left panel) and absolute number of SIY-specific CD8+ T cells (right panel) assessed as in (C). Data represent mean ± SEM (n = 4) and correspond to three (C, E, and F) or two (A, B, G, and H) independent experiments. *p < 0.05, **p < 0.01, ***p < 0.001, two-sided Student t test (C, G, and H) and two-way ANOVA and Tukey multiple comparison test (E and F).

Close modal

To further study the role of NK cells on the outcome of tumor-specific CD8+ T cell compartment, we characterized the phenotype of effector CD8+ T cells primed in control or NK cell–depleted mice based on the expression of CD44, KLRG1, and CD127 markers (36). Mice primed in the absence of NK cells show an augmented proportion of activated CD8+ T cells (CD44+) including both short-lived effector cells (KLRG1+CD127CD44+CD8+) and memory precursor effector cells (KLRG1CD127+CD44+CD8+) (Fig. 2A). In turn, memory T cells can be divided into two major subtypes: central memory T cells (TCMs) (CD44+CD62Lhi/CCR7hi) that have little or no effector function but readily proliferate and differentiate to effector cells in response to antigenic stimulation, and TEMs (CD44+CD62Llo/CCR7lo) that are widely distributed and display immediate effector functions (37). Although we observed no difference in the TCM CD8+ cell population (Fig. 2B), we found a marked increase in the proportion of TEM CD8+ subset in spleens and blood of NK cell–depleted mice compared with control mice (Fig. 2C), whereas there was no difference in TCMs and TEM CD4+ cells between both groups of mice in blood, spleen, or TDLNs (data not shown). Using the YAC1.SIY model, we recapitulated the results obtained with MC57.SIY (Fig. 2D). These results indicate that NK cell depletion generates a CD8+ T cell population skewed toward a TEM phenotype in tumor-bearing mice.

FIGURE 2.

NK cell depletion induces the expansion of the TEM CD8+ subset. (AC) IC or anti-NK1.1–treated mice were inoculated with MC57.SIY cells; blood, spleen, and LNs were collected. The percentage of CD44+, short-lived effector cells (SLECs) and memory precursor effector cells (MPECs) in the CD8+ T cell population was evaluated in blood (A), and the percentage of TCMs (B) and TEMs (C) was evaluated by flow cytometry in the CD8+ T cell population in blood, spleen, and LNs. (D) IC or anti-NK1.1–treated mice were inoculated with YAC1.SIY cells, blood was collected 11 d later, and the percentage of TEM CD8+ cells was evaluated by flow cytometry. Data represent mean ± SEM (n = 4) and correspond to two independent experiments. *p < 0.05, **p < 0.01, ***p < 0.001, two-sided Student t test. n.s., not significant.

FIGURE 2.

NK cell depletion induces the expansion of the TEM CD8+ subset. (AC) IC or anti-NK1.1–treated mice were inoculated with MC57.SIY cells; blood, spleen, and LNs were collected. The percentage of CD44+, short-lived effector cells (SLECs) and memory precursor effector cells (MPECs) in the CD8+ T cell population was evaluated in blood (A), and the percentage of TCMs (B) and TEMs (C) was evaluated by flow cytometry in the CD8+ T cell population in blood, spleen, and LNs. (D) IC or anti-NK1.1–treated mice were inoculated with YAC1.SIY cells, blood was collected 11 d later, and the percentage of TEM CD8+ cells was evaluated by flow cytometry. Data represent mean ± SEM (n = 4) and correspond to two independent experiments. *p < 0.05, **p < 0.01, ***p < 0.001, two-sided Student t test. n.s., not significant.

Close modal

Given these differences, we examined whether the expanded TEM CD8+ cell population generated in the absence of NK cells can elicit enhanced recall responses. We challenged NK cell–depleted or control mice with MC57.SIY cells as described earlier, maintaining the NK cell–depleted mice only during the priming phase, and 110 d later (when NK cell counts were normal and tumors were rejected in both groups of mice), we rechallenged mice with the unrelated and poorly immunogenic B16 melanoma, also transduced to express the SIY Ag (B16.SIY) to monitor for SIY-specific memory responses (Fig. 3A). As control, a group of naive mice (that had not been previously injected with MC57.SIY cells) was also challenged with B16.SIY cells. One day before and 4 d after rechallenge we analyzed the frequency of SIY+CD8+ T cells in the blood of the three groups of mice. Mice that had been primed in the absence of NK cells retained an expanded TEM CD8+ cell population before the second challenge (Fig. 3B) and showed an augmented proportion of Ag-specific CD8+ T cells during the recall response (Fig. 3C) and an increased proportion of IFN-γ–producing CD8+ T cells when restimulated ex vivo with the SIY Ag (Fig. 3D). Priming with MC57.SIY cells induced a strong memory response in mice rechallenged with B16.SIY cells. In addition, if priming occurred in the absence of NK cells, it resulted in delayed tumor growth among mice that did not reject the tumor (Fig. 3E) and a trend toward enhanced protection against secondary tumor formation (Fig. 3F, tumor development in 11.11% of anti-NK1.1–treated mice versus 31.25% of IC-treated mice). These results indicate that the NK cell–dependent shaping of the memory compartment during the first tumor challenge may result in diminished recall responses and poor tumor control.

FIGURE 3.

NK cell depletion during priming enhances antitumor recall responses. IC or anti-NK1.1–treated mice were inoculated with MC57.SIY cells (first challenge) or left unchallenged (no tumor). After 110 d, mice were challenged with B16.SIY cells (second challenge); 1 d before (pre-B16.SIY) and 4 d later (post-B16.SIY), blood was collected and analyzed. (A) Experimental design. (B) Percentage of TEM CD8+ cells 109 d after challenge with MC57.SIY (pre-B16.SIY). (C) Percentage of SIY-specific CD8+ T cells. (D) Percentage of IFN-γ–producing CD8+ T cells after restimulation with soluble SIY peptide. Data represent mean ± SEM (n = 5) and correspond to two independent experiments. (E) Tumor size (in mice that did not reject the tumor) after challenge with B16.SIY in the three groups of mice and (F) percentage of tumor-free mice (n represents the total number of mice analyzed for each group of mice). Data represent mean ± SEM. *p < 0.05, two-sided Student t test, **p < 0.01, two-way ANOVA and Tukey multiple comparison test.

FIGURE 3.

NK cell depletion during priming enhances antitumor recall responses. IC or anti-NK1.1–treated mice were inoculated with MC57.SIY cells (first challenge) or left unchallenged (no tumor). After 110 d, mice were challenged with B16.SIY cells (second challenge); 1 d before (pre-B16.SIY) and 4 d later (post-B16.SIY), blood was collected and analyzed. (A) Experimental design. (B) Percentage of TEM CD8+ cells 109 d after challenge with MC57.SIY (pre-B16.SIY). (C) Percentage of SIY-specific CD8+ T cells. (D) Percentage of IFN-γ–producing CD8+ T cells after restimulation with soluble SIY peptide. Data represent mean ± SEM (n = 5) and correspond to two independent experiments. (E) Tumor size (in mice that did not reject the tumor) after challenge with B16.SIY in the three groups of mice and (F) percentage of tumor-free mice (n represents the total number of mice analyzed for each group of mice). Data represent mean ± SEM. *p < 0.05, two-sided Student t test, **p < 0.01, two-way ANOVA and Tukey multiple comparison test.

Close modal

NK cells with regulatory functions, capable of dampening Ag-specific T cell responses through the secretion of the immunosuppressive cytokines IL-10 and/or TGF-β, have been described previously (38). However, when we compared the amounts of IL-10 and TGF-β mRNA in tumors and TDLNs, we found no differences between NK cell–depleted and control mice (Fig. 4A, 4B). Although NK cells can recruit Tregs via secretion of the chemokine CCL22, which may contribute to immune suppression (39), we found no differences in the frequency of Tregs in tumor or TDLNs between control and NK cell–depleted mice (Fig. 4C).

FIGURE 4.

NK cell depletion augments DC maturation without affecting other major regulatory populations. IC or anti-NK1.1–treated mice were inoculated with MC57.SIY cells. (A and B) Six days later, TDLNs and tumors were collected and the relative abundance of IL-10 (A) and TGF-β (B) transcripts were evaluated by qPCR. (C and D) Eleven days after tumor inoculation, TDLNs, spleens, and blood were collected and the percentage of Tregs (C) and MDSCs (D) were evaluated by flow cytometry. (EG) Six days after tumor challenge, tumors and TDLNs were collected and the expression of CD86 on DCs (defined as CD3NKp46 B220CD11c+) was evaluated by flow cytometry (E), the relative abundance of IFN-β transcripts was evaluated by qPCR (F), and the frequency of DCs was evaluated by flow cytometry. Data represent mean ± SEM (n = 3) and correspond to two independent experiments. *p < 0.05, ***p < 0.001, two-sided Student t test. n.s., not significant.

FIGURE 4.

NK cell depletion augments DC maturation without affecting other major regulatory populations. IC or anti-NK1.1–treated mice were inoculated with MC57.SIY cells. (A and B) Six days later, TDLNs and tumors were collected and the relative abundance of IL-10 (A) and TGF-β (B) transcripts were evaluated by qPCR. (C and D) Eleven days after tumor inoculation, TDLNs, spleens, and blood were collected and the percentage of Tregs (C) and MDSCs (D) were evaluated by flow cytometry. (EG) Six days after tumor challenge, tumors and TDLNs were collected and the expression of CD86 on DCs (defined as CD3NKp46 B220CD11c+) was evaluated by flow cytometry (E), the relative abundance of IFN-β transcripts was evaluated by qPCR (F), and the frequency of DCs was evaluated by flow cytometry. Data represent mean ± SEM (n = 3) and correspond to two independent experiments. *p < 0.05, ***p < 0.001, two-sided Student t test. n.s., not significant.

Close modal

MDSCs constitute one of the major populations of immune cells capable of regulating antitumor immune responses (40). In addition, it has recently been described that CD11b+CD27+ NK cells could be converted into CD11b+Gr1+ MDSCs in tumor-bearing mice (41). However, when we compared the proportion of MDSCs in blood and spleens of NK cell–depleted or control mice, we found no differences (Fig. 4D).

The striking difference for CD8+ T cell priming observed upon NK cell depletion could point to a defect at the level of DC costimulation and cross-priming of CD8+ T cells. Therefore, we compared the maturation status of DCs in both groups of mice and found an enhanced expression of CD86 on tumor-infiltrating and TDLN DCs in mice lacking NK cells compared with control mice (Fig. 4E). These observations suggest that NK cells can restrict DC maturation. Considering that IFN-β produced by DCs early during an antitumor immune response is critical for priming of CD8+ T cells (42), expression of this cytokine also was examined. In fact, tumors from NK cell–depleted mice showed a 2.33-fold increase in the levels of IFN-β (Fig. 4F). Moreover, we found an elevated frequency of DCs in TDLNs from NK cell–depleted mice compared with control mice (Fig. 4G). Together, these results suggest a regulatory function of NK cells at the level of DCs.

In a search for possible phenotypic changes on NK cells during antitumor immune responses that could account for their regulatory ability, we assessed the expression of several stage-related and activation/inhibitory markers on NK cells. We found that TINK cells are mostly the terminally differentiated CD11b+CD27 subpopulation (43) (Supplemental Fig. 1A). Still, compared with NK cells from naive LN, TINK cells showed an altered phenotype that included upregulation of KLRG1 (capable of inhibiting NK cell effector functions [44]), Ly6C (associated with an inert state [45]), and CD25 (Supplemental Fig. 1B), the downregulation of activating receptors NKG2D and NKp46 (Supplemental Fig. 1C), and no change in the potentially inhibitory molecules PD-1, CTLA-4 (46), and c-Kit (Supplemental Fig. 1D). The PD-1/PD-L1 pathway is one of the most critical checkpoints responsible for mediating tumor-induced immune suppression (33). Notably, 6 d after tumor inoculation, we observed an expansion in the frequency and numbers of PD-L1–expressing NK cells (PD-L1hi NK cells), not only in tumors (Fig. 5A) but also in TDLNs compared with naive or nondraining LNs (NDLNs; Fig. 5B–D). To evaluate whether this phenotypic change could be mediated by interaction with tumor cells, we cocultured splenocytes with MC57 tumor cells in vitro. Indeed, this resulted in a 2-fold expansion of PD-L1hi NK cells (Fig. 5E). This effect was dependent on IFN-γ, and the activating NK cell receptor NKG2D as neutralization of IFN-γ or blockade of NKG2D during the coculture of splenocytes with MC57 cells partially inhibited the upregulation of PD-L1 (Fig. 5E). In contrast with this tumor-induced phenotypic change, no increase in PD-L1–expressing NK cells was observed upon stimulation with the cytokines IL-12, IL-15, and IL-18 (Fig. 5F). In addition, human NK cells also upregulated PD-L1 expression after tumor recognition in vitro (Fig. 5G), indicating that certain tumors can induce the upregulation of PD-L1 on both mouse and human NK cells.

FIGURE 5.

Tumors induce PD-L1 expression on NK cells. (AD) Mice were inoculated with MC57.SIY cells, and 3 and 6 d later, tumors, TDLNs, NDLNs, and naive LNs were collected and the expression of PD-L1 was analyzed by flow cytometry on NK cells (defined as CD3NKp46+ cells). (A) Percentage of PD-L1hi NK cells in naive LNs and tumors. (B) Percentage and (C) number of PD-L1hi NK cells in TDLNs on days 0, 3, and 6 after tumor challenge. (D) Representative histograms of PD-L1 expression on NK cells in tumors, TDLNs, and NDLNs. The filled histogram corresponds to the IC. (A–C) Data represent mean ± SEM (n = 4) and correspond to two independent experiments. (E and F) A total of 106 splenocytes were cultured with 105 MC57.SIY in the presence of IC, anti–IFN-γ neutralizing mAb, or anti-NKG2D blocking mAb (E), or in the presence of IL-12, IL-15, and IL-18 or 105 MC57.SIY (F), and after 48 h, the percentage of PD-L1hi NK cells was analyzed by flow cytometry. (E and F) Data represent mean ± SEM (n = 4) and correspond to two independent experiments. (G) Human PBMCs were cultured with the indicated tumor cell lines, and 48 h later the percentage of PD-L1hi NK cells (defined as CD3CD56+ cells) was evaluated by flow cytometry. Data represent mean ± SEM (n = 4 donors). *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001, two-sided Student t test (A–C) and one-way ANOVA and Tukey multiple comparison test (E–G). MFI, mean fluorescence intensity.

FIGURE 5.

Tumors induce PD-L1 expression on NK cells. (AD) Mice were inoculated with MC57.SIY cells, and 3 and 6 d later, tumors, TDLNs, NDLNs, and naive LNs were collected and the expression of PD-L1 was analyzed by flow cytometry on NK cells (defined as CD3NKp46+ cells). (A) Percentage of PD-L1hi NK cells in naive LNs and tumors. (B) Percentage and (C) number of PD-L1hi NK cells in TDLNs on days 0, 3, and 6 after tumor challenge. (D) Representative histograms of PD-L1 expression on NK cells in tumors, TDLNs, and NDLNs. The filled histogram corresponds to the IC. (A–C) Data represent mean ± SEM (n = 4) and correspond to two independent experiments. (E and F) A total of 106 splenocytes were cultured with 105 MC57.SIY in the presence of IC, anti–IFN-γ neutralizing mAb, or anti-NKG2D blocking mAb (E), or in the presence of IL-12, IL-15, and IL-18 or 105 MC57.SIY (F), and after 48 h, the percentage of PD-L1hi NK cells was analyzed by flow cytometry. (E and F) Data represent mean ± SEM (n = 4) and correspond to two independent experiments. (G) Human PBMCs were cultured with the indicated tumor cell lines, and 48 h later the percentage of PD-L1hi NK cells (defined as CD3CD56+ cells) was evaluated by flow cytometry. Data represent mean ± SEM (n = 4 donors). *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001, two-sided Student t test (A–C) and one-way ANOVA and Tukey multiple comparison test (E–G). MFI, mean fluorescence intensity.

Close modal

We wondered whether increased expression of PD-L1 on NK cells might contribute to the diminished priming of CD8+ T cells observed in tumor-bearing mice, through targeting DCs and/or CD8+ T cells. Therefore, we first analyzed whether NK cells might control DC and CD8+ T cell numbers in a PD-1/PD-L1–dependent manner in vivo. Mice depleted of NK cells exhibited a 1.6-fold increase in the frequency of PD-1hi DCs (Fig. 6A) but showed no difference in PD-1–expressing CD8+ T cells (Fig. 6B) compared with control mice. Accordingly, in vitro DC maturation with LPS and R848 was inhibited in the presence of TINK cells but not in the presence of NK cells from a naive spleen, and this inhibition was partially reverted by blockade of PD-L1 (Fig. 6C). To explore a potential direct effect of PD-L1hi NK cells on CD8+ T cells, we analyzed the proliferation and survival of activated CD8+ T cells in the presence of TINK cells. We found that in vitro proliferation of these CD8+ T cells (that uniformly expressed high levels of PD-1, data not shown) was not altered in the presence of control (isolated from naive spleens) or tumor-derived NK cells (Fig. 6D). In addition, naive and tumor-infiltrating CD8+ T cells were equally resistant to NK cell–mediated lysis when we used either control NK cells or PD-L1hi TINK cells as effectors (Fig. 6E, 6F). These results indicate that TINK cells were unable to directly suppress CD8+ T cell proliferation or to preferentially kill tumor-infiltrating CD8+ T cells, but that they can regulate the maturation status of DCs, in part through PD-1/PD-L1 interactions, and in such manner indirectly affect CD8+ T cell priming.

FIGURE 6.

NK cells negatively regulate DC maturation through PD-1/PD-L1 interactions. (A and B) IC or anti-NK1.1–treated mice were inoculated with MC57.SIY cells, and 6 d later PD-1 expression was evaluated by flow cytometry in DCs (A) and CD8+ T cells (B) from TDLNs. (C) Sorted splenic DCs were stimulated with LPS and R848 (control) in the absence or in the presence of control NK cells (isolated from spleen of naive mice) or TINK cells, in the absence or in the presence of an anti–PD-L1 blocking mAb, and 18 h later the number of viable mature DCs (defined as CD11c+CD86+Zombie Green cells and depicted as percentage of the control) was evaluated by flow cytometry. (D) eFluor 670–labeled CD3+ T cells were stimulated with anti-CD3/anti-CD28 and cultured in the absence or in the presence of control NK cells or TINK cells, and 72 h later the percentage of divided CD8+ T cells was evaluated by flow cytometry. (E and F) eFluor 670–labeled naive (eFluor 670high) and tumor-infiltrating (TI; eFluor 670low) CD8+ T cells were cultured in the absence or in the presence of control NK cells or TINK cells, and 6 h later the percentage of nonviable CD8+ T cells (100-Zombie Green cells) was evaluated by flow cytometry and % specific cytotoxicity (E), and the viable TI/Naive CD8+ T cell ratio (F) was calculated (E:T = 4:1). Data represent mean ± SEM [(A and B) n = 4; (C–F) n = 2)] and correspond to two independent experiments. *p < 0.05, two-sided Student t test (A and B), one-way ANOVA and Tukey (C, D, and F), or Sidak (E) multiple comparison test. n.s., not significant.

FIGURE 6.

NK cells negatively regulate DC maturation through PD-1/PD-L1 interactions. (A and B) IC or anti-NK1.1–treated mice were inoculated with MC57.SIY cells, and 6 d later PD-1 expression was evaluated by flow cytometry in DCs (A) and CD8+ T cells (B) from TDLNs. (C) Sorted splenic DCs were stimulated with LPS and R848 (control) in the absence or in the presence of control NK cells (isolated from spleen of naive mice) or TINK cells, in the absence or in the presence of an anti–PD-L1 blocking mAb, and 18 h later the number of viable mature DCs (defined as CD11c+CD86+Zombie Green cells and depicted as percentage of the control) was evaluated by flow cytometry. (D) eFluor 670–labeled CD3+ T cells were stimulated with anti-CD3/anti-CD28 and cultured in the absence or in the presence of control NK cells or TINK cells, and 72 h later the percentage of divided CD8+ T cells was evaluated by flow cytometry. (E and F) eFluor 670–labeled naive (eFluor 670high) and tumor-infiltrating (TI; eFluor 670low) CD8+ T cells were cultured in the absence or in the presence of control NK cells or TINK cells, and 6 h later the percentage of nonviable CD8+ T cells (100-Zombie Green cells) was evaluated by flow cytometry and % specific cytotoxicity (E), and the viable TI/Naive CD8+ T cell ratio (F) was calculated (E:T = 4:1). Data represent mean ± SEM [(A and B) n = 4; (C–F) n = 2)] and correspond to two independent experiments. *p < 0.05, two-sided Student t test (A and B), one-way ANOVA and Tukey (C, D, and F), or Sidak (E) multiple comparison test. n.s., not significant.

Close modal

Evidence supporting a regulatory role for NK cells in diverse immunopathological conditions such as viral infections, autoimmunity, and transplantation is emerging (1217). However, little is known about a possible regulatory activity during an antitumor immune response. In this study, we took advantage of the SIY model Ag to track the priming of endogenous tumor-specific CD8+ T cells in vivo in the presence or in the absence of NK cells. We show that NK cells can control the spontaneous priming and memory responses of tumor Ag-specific CD8+ T cells, in part through a mechanism involving regulation of DC maturation through PD-1/PD-L1 interaction.

Despite the enhanced priming of tumor-specific CD8+ T cells in NK cell–depleted mice, the primary tumor is equally rejected by control mice. This is due to the high immunogenicity of MC57.SIY cells, which induce substantial amounts of tumor-specific CD8+ T cells, high enough to induce tumor rejection simultaneously or before the regulatory activity of NK cells becomes apparent.

In vivo, the absence of NK cells resulted in enhanced priming of antitumor CD8+ T cells and a memory response skewed toward a TEM phenotype that led to a faster and more robust recall response resulting in delayed tumor growth. These results may have implications for the rational design of cancer immunotherapies given that the induction of durable antitumor T cell responses in cancer patients is a major goal of therapeutic interventions (26).

The regulatory role described in this article for NK cells is in sharp contrast with its well-known role in the control of tumor growth and infectious agents. This apparent contradiction may reflect the complex and yet incompletely understood biology of NK cells, which when faced with different stimuli might generate alternative outcomes at different stages of the immune response. In the tumor context, Schreiber and colleagues (47) showed that, although NK cells could play a critical role during the elimination of tumor cells, during the equilibrium (dormancy phase) of an antitumor immune response, NK cells seem to be expendable. In our model, PD-L1hi NK cells become detectable in TDLNs 6 d after tumor challenge, consistent with the idea that NK cells can be functional and contribute to tumor control early after tumor challenge and become suppressive later on.

Such a regulatory activity of NK cells at the level of the adaptive immune response may have evolved as a mechanism to control the extent of T cell activation in the context of chronic viral infections and other immune pathological conditions, to limit damage that can be associated with widespread T cell responses. Accordingly, some patients with autoimmune disorders display reduced NK cell numbers with impaired cytotoxicity (48).

PD-L1 is a powerful immune modulator differentially expressed during tumor progression on tumor cells, stroma, or immune cells that contributes to immune escape. Blocking agents targeting this pathway are currently being tested with promising results in clinical trials, including FDA approval in both melanoma and lung cancer (49). In this article, we show that, consistent with a regulatory function of NK cells, PD-L1 was overexpressed by NK cells from tumor-bearing mice, and that induction of PD-L1 expression was dependent on NKG2D recognition. PD-L1 was also upregulated on human NK cells upon coculture with some but not all of the human cell lines tested. However, there was no correlation between NKG2D ligand expression and PD-L1 upregulation, suggesting that other activating/inhibiting receptors might be involved. It has been shown that IFN-γ is a positive regulator for PD-L1 expression (50). Accordingly, we observed that blockade of IFN-γ prevented the generation of PD-L1hi NK cells. Although we cannot rule out the possibility that NK cell–derived IFN-γ might also trigger PD-L1 upregulation on stroma or other immune cells that may partially contribute to the regulation of the immune response, we found no difference between NK cell–depleted and control mice when analyzing PD-L1 expression on tumor cells in vivo (data not shown).

The observed PD-L1 expression on NK cells has functional consequences, because tumor-experienced NK cells controlled DC maturation in vivo and in vitro, and PD-L1 blockade partially restored the numbers of mature DCs recovered in vitro. Accordingly, we found an expanded population of PD-1hi DCs in TDLNs of NK cell–depleted mice compared with control mice. Conversely, PD-L1hi NK cells were unable to kill CD8+ T cells or to directly suppress its proliferation in vitro. Moreover, in vivo, the frequency of PD-1hi CD8+ T cells was unchanged in the presence or in the absence of NK cells.

PD-L1 expression on c-Kit+CD11b NK cells during experimental diabetes (17) and metastatic spread of cancer (51) has been reported. However, in our experimental setting, we observed that PD-L1+ NK cells showed no detectable expression of c-Kit and were mostly terminally differentiated, expressing high levels of CD11b.

PD-L1 expression on tumor cells is a suggestive, but inadequate, predictive biomarker of response to immune-checkpoint blockade (31), suggesting that PD-L1 expressed by other cells might also be a relevant target for therapy with these agents. It is conceivable that the blockade of PD-L1 on NK cells could be part of the mechanism of action of these Abs, through disruption of the regulatory interaction between NK cells and DCs, which could contribute to the more robust and efficient antitumor CD8+ T cell response seen with anti–PD-1.

In summary, our results suggest a model in which tumor-induced PD-L1hi NK cells regulate DC activation resulting in a reduced ability to support CD8+ T cell priming. The assessment of PD-L1 expression on NK cells should be investigated as a potential biomarker for the presence of regulatory NK cells and also in association with clinical outcomes with anti–PD-1 mAbs.

We thank Dr. Gabriel A. Rabinovich (Laboratory of Immunopathology, Instituto de Biología y Medicina Experimental) for continuous support and for generously providing unlimited access to the Flow Cytometry and Cell Sorting Facility and Fundación René Barón for providing additional support.

This work was supported by grants from the National Agency for Promotion of Science and Technology from Argentina and the National Research Council of Argentina (CONICET) (to M.B.F. and N.W.Z.). X.L.R.I., R.G.S., N.I.T., A.Z, R.E.A., F.S., and S.Y.N. are fellows of CONICET. J.M.S. is a fellow of the Instituto Nacional del Cancer. C.I.D., M.B.F., and N.W.Z. are members of the Researcher Career of CONICET.

The online version of this article contains supplemental material.

Abbreviations used in this article:

DC

dendritic cell

IC

isotype control

LN

lymph node

MDSC

myeloid-derived suppressor cell

NDLN

nondraining LN

PD-1

programmed death-1

PD-L1

programmed death-ligand 1

qPCR

quantitative PCR

TCM

central memory T cell

TDLN

tumor-draining LN

TEM

effector memory T cell

TINK

tumor-infiltrating NK

Treg

regulatory T cell.

1
Lanier
L. L.
2008
.
Evolutionary struggles between NK cells and viruses.
Nat. Rev. Immunol.
8
:
259
268
.
2
Moretta
L.
,
Bottino
C.
,
Pende
D.
,
Castriconi
R.
,
Mingari
M. C.
,
Moretta
A.
.
2006
.
Surface NK receptors and their ligands on tumor cells.
Semin. Immunol.
18
:
151
158
.
3
Zwirner
N. W.
,
Domaica
C. I.
.
2010
.
Cytokine regulation of natural killer cell effector functions.
Biofactors
36
:
274
288
.
4
Ziblat
A.
,
Domaica
C. I.
,
Spallanzani
R. G.
,
Iraolagoitia
X. L.
,
Rossi
L. E.
,
Avila
D. E.
,
Torres
N. I.
,
Fuertes
M. B.
,
Zwirner
N. W.
.
2015
.
IL-27 stimulates human NK-cell effector functions and primes NK cells for IL-18 responsiveness.
Eur. J. Immunol.
45
:
192
202
.
5
Girart
M. V.
,
Fuertes
M. B.
,
Domaica
C. I.
,
Rossi
L. E.
,
Zwirner
N. W.
.
2007
.
Engagement of TLR3, TLR7, and NKG2D regulate IFN-gamma secretion but not NKG2D-mediated cytotoxicity by human NK cells stimulated with suboptimal doses of IL-12.
J. Immunol.
179
:
3472
3479
.
6
Lanier
L. L.
2008
.
Up on the tightrope: natural killer cell activation and inhibition.
Nat. Immunol.
9
:
495
502
.
7
Vivier
E.
,
Tomasello
E.
,
Baratin
M.
,
Walzer
T.
,
Ugolini
S.
.
2008
.
Functions of natural killer cells.
Nat. Immunol.
9
:
503
510
.
8
Vivier
E.
,
Raulet
D. H.
,
Moretta
A.
,
Caligiuri
M. A.
,
Zitvogel
L.
,
Lanier
L. L.
,
Yokoyama
W. M.
,
Ugolini
S.
.
2011
.
Innate or adaptive immunity? The example of natural killer cells.
Science
331
:
44
49
.
9
Moretta
A.
,
Marcenaro
E.
,
Sivori
S.
,
Della Chiesa
M.
,
Vitale
M.
,
Moretta
L.
.
2005
.
Early liaisons between cells of the innate immune system in inflamed peripheral tissues.
Trends Immunol.
26
:
668
675
.
10
Walzer
T.
,
Dalod
M.
,
Robbins
S. H.
,
Zitvogel
L.
,
Vivier
E.
.
2005
.
Natural-killer cells and dendritic cells: “l’union fait la force”.
Blood
106
:
2252
2258
.
11
Martín-Fontecha
A.
,
Thomsen
L. L.
,
Brett
S.
,
Gerard
C.
,
Lipp
M.
,
Lanzavecchia
A.
,
Sallusto
F.
.
2004
.
Induced recruitment of NK cells to lymph nodes provides IFN-gamma for T(H)1 priming.
Nat. Immunol.
5
:
1260
1265
.
12
Cook
K. D.
,
Whitmire
J. K.
.
2013
.
The depletion of NK cells prevents T cell exhaustion to efficiently control disseminating virus infection.
J. Immunol.
190
:
641
649
.
13
Schuster
I. S.
,
Wikstrom
M. E.
,
Brizard
G.
,
Coudert
J. D.
,
Estcourt
M. J.
,
Manzur
M.
,
O’Reilly
L. A.
,
Smyth
M. J.
,
Trapani
J. A.
,
Hill
G. R.
, et al
.
2014
.
TRAIL+ NK cells control CD4+ T cell responses during chronic viral infection to limit autoimmunity.
Immunity
41
:
646
656
.
14
Lang
P. A.
,
Lang
K. S.
,
Xu
H. C.
,
Grusdat
M.
,
Parish
I. A.
,
Recher
M.
,
Elford
A. R.
,
Dhanji
S.
,
Shaabani
N.
,
Tran
C. W.
, et al
.
2012
.
Natural killer cell activation enhances immune pathology and promotes chronic infection by limiting CD8+ T-cell immunity.
Proc. Natl. Acad. Sci. USA
109
:
1210
1215
.
15
Waggoner
S. N.
,
Taniguchi
R. T.
,
Mathew
P. A.
,
Kumar
V.
,
Welsh
R. M.
.
2010
.
Absence of mouse 2B4 promotes NK cell-mediated killing of activated CD8+ T cells, leading to prolonged viral persistence and altered pathogenesis.
J. Clin. Invest.
120
:
1925
1938
.
16
Laffont
S.
,
Seillet
C.
,
Ortaldo
J.
,
Coudert
J. D.
,
Guéry
J. C.
.
2008
.
Natural killer cells recruited into lymph nodes inhibit alloreactive T-cell activation through perforin-mediated killing of donor allogeneic dendritic cells.
Blood
112
:
661
671
.
17
Ehlers
M.
,
Papewalis
C.
,
Stenzel
W.
,
Jacobs
B.
,
Meyer
K. L.
,
Deenen
R.
,
Willenberg
H. S.
,
Schinner
S.
,
Thiel
A.
,
Scherbaum
W. A.
, et al
.
2012
.
Immunoregulatory natural killer cells suppress autoimmunity by down-regulating antigen-specific CD8+ T cells in mice.
Endocrinology
153
:
4367
4379
.
18
Rabinovich
B. A.
,
Li
J.
,
Shannon
J.
,
Hurren
R.
,
Chalupny
J.
,
Cosman
D.
,
Miller
R. G.
.
2003
.
Activated, but not resting, T cells can be recognized and killed by syngeneic NK cells.
J. Immunol.
170
:
3572
3576
.
19
Cerboni
C.
,
Zingoni
A.
,
Cippitelli
M.
,
Piccoli
M.
,
Frati
L.
,
Santoni
A.
.
2007
.
Antigen-activated human T lymphocytes express cell-surface NKG2D ligands via an ATM/ATR-dependent mechanism and become susceptible to autologous NK- cell lysis.
Blood
110
:
606
615
.
20
Soderquest
K.
,
Walzer
T.
,
Zafirova
B.
,
Klavinskis
L. S.
,
Polić
B.
,
Vivier
E.
,
Lord
G. M.
,
Martín-Fontecha
A.
.
2011
.
Cutting edge: CD8+ T cell priming in the absence of NK cells leads to enhanced memory responses.
J. Immunol.
186
:
3304
3308
.
21
Peppa
D.
,
Gill
U. S.
,
Reynolds
G.
,
Easom
N. J.
,
Pallett
L. J.
,
Schurich
A.
,
Micco
L.
,
Nebbia
G.
,
Singh
H. D.
,
Adams
D. H.
, et al
.
2013
.
Up-regulation of a death receptor renders antiviral T cells susceptible to NK cell-mediated deletion.
J. Exp. Med.
210
:
99
114
.
22
Lu
L.
,
Ikizawa
K.
,
Hu
D.
,
Werneck
M. B.
,
Wucherpfennig
K. W.
,
Cantor
H.
.
2007
.
Regulation of activated CD4+ T cells by NK cells via the Qa-1-NKG2A inhibitory pathway.
Immunity
26
:
593
604
.
23
Waggoner
S. N.
,
Cornberg
M.
,
Selin
L. K.
,
Welsh
R. M.
.
2011
.
Natural killer cells act as rheostats modulating antiviral T cells.
Nature
481
:
394
398
.
24
Andrews
D. M.
,
Estcourt
M. J.
,
Andoniou
C. E.
,
Wikstrom
M. E.
,
Khong
A.
,
Voigt
V.
,
Fleming
P.
,
Tabarias
H.
,
Hill
G. R.
,
van der Most
R. G.
, et al
.
2010
.
Innate immunity defines the capacity of antiviral T cells to limit persistent infection.
J. Exp. Med.
207
:
1333
1343
.
25
Hayakawa
Y.
,
Screpanti
V.
,
Yagita
H.
,
Grandien
A.
,
Ljunggren
H. G.
,
Smyth
M. J.
,
Chambers
B. J.
.
2004
.
NK cell TRAIL eliminates immature dendritic cells in vivo and limits dendritic cell vaccination efficacy.
J. Immunol.
172
:
123
129
.
26
Gajewski
T. F.
,
Schreiber
H.
,
Fu
Y. X.
.
2013
.
Innate and adaptive immune cells in the tumor microenvironment.
Nat. Immunol.
14
:
1014
1022
.
27
Rabinovich
G. A.
,
Gabrilovich
D.
,
Sotomayor
E. M.
.
2007
.
Immunosuppressive strategies that are mediated by tumor cells.
Annu. Rev. Immunol.
25
:
267
296
.
28
Blank
C.
,
Gajewski
T. F.
,
Mackensen
A.
.
2005
.
Interaction of PD-L1 on tumor cells with PD-1 on tumor-specific T cells as a mechanism of immune evasion: implications for tumor immunotherapy.
Cancer Immunol. Immunother.
54
:
307
314
.
29
Park
S. J.
,
Namkoong
H.
,
Doh
J.
,
Choi
J. C.
,
Yang
B. G.
,
Park
Y.
,
Chul Sung
Y.
.
2014
.
Negative role of inducible PD-1 on survival of activated dendritic cells.
J. Leukoc. Biol.
95
:
621
629
.
30
Driessens
G.
,
Kline
J.
,
Gajewski
T. F.
.
2009
.
Costimulatory and coinhibitory receptors in anti-tumor immunity.
Immunol. Rev.
229
:
126
144
.
31
Mahoney, K. M., G. J. Freeman, and D. F. McDermott. 2015. The next immune-checkpoint inhibitors: PD-1/PD-L1 blockade in melanoma. Clin. Ther. 37: 764–782
.
32
Herbst
R. S.
,
Soria
J. C.
,
Kowanetz
M.
,
Fine
G. D.
,
Hamid
O.
,
Gordon
M. S.
,
Sosman
J. A.
,
McDermott
D. F.
,
Powderly
J. D.
,
Gettinger
S. N.
, et al
.
2014
.
Predictive correlates of response to the anti-PD-L1 antibody MPDL3280A in cancer patients.
Nature
515
:
563
567
.
33
Teng
M. W.
,
Ngiow
S. F.
,
Ribas
A.
,
Smyth
M. J.
.
2015
.
Classifying cancers based on T-cell infiltration and PD-L1.
Cancer Res.
75
:
2139
2145
.
34
Zhang
B.
,
Bowerman
N. A.
,
Salama
J. K.
,
Schmidt
H.
,
Spiotto
M. T.
,
Schietinger
A.
,
Yu
P.
,
Fu
Y. X.
,
Weichselbaum
R. R.
,
Rowley
D. A.
, et al
.
2007
.
Induced sensitization of tumor stroma leads to eradication of established cancer by T cells.
J. Exp. Med.
204
:
49
55
.
35
Rossi
L. E.
,
Avila
D. E.
,
Spallanzani
R. G.
,
Ziblat
A.
,
Fuertes
M. B.
,
Lapyckyj
L.
,
Croci
D. O.
,
Rabinovich
G. A.
,
Domaica
C. I.
,
Zwirner
N. W.
.
2012
.
Histone deacetylase inhibitors impair NK cell viability and effector functions through inhibition of activation and receptor expression.
J. Leukoc. Biol.
91
:
321
331
.
36
Joshi
N. S.
,
Cui
W.
,
Chandele
A.
,
Lee
H. K.
,
Urso
D. R.
,
Hagman
J.
,
Gapin
L.
,
Kaech
S. M.
.
2007
.
Inflammation directs memory precursor and short-lived effector CD8(+) T cell fates via the graded expression of T-bet transcription factor.
Immunity
27
:
281
295
.
37
Sallusto
F.
,
Geginat
J.
,
Lanzavecchia
A.
.
2004
.
Central memory and effector memory T cell subsets: function, generation, and maintenance.
Annu. Rev. Immunol.
22
:
745
763
.
38
Deniz
G.
,
Erten
G.
,
Kücüksezer
U. C.
,
Kocacik
D.
,
Karagiannidis
C.
,
Aktas
E.
,
Akdis
C. A.
,
Akdis
M.
.
2008
.
Regulatory NK cells suppress antigen-specific T cell responses.
J. Immunol.
180
:
850
857
.
39
Mailloux
A. W.
,
Young
M. R.
.
2009
.
NK-dependent increases in CCL22 secretion selectively recruits regulatory T cells to the tumor microenvironment.
J. Immunol.
182
:
2753
2765
.
40
Youn
J. I.
,
Gabrilovich
D. I.
.
2010
.
The biology of myeloid-derived suppressor cells: the blessing and the curse of morphological and functional heterogeneity.
Eur. J. Immunol.
40
:
2969
2975
.
41
Park
Y. J.
,
Song
B.
,
Kim
Y. S.
,
Kim
E. K.
,
Lee
J. M.
,
Lee
G. E.
,
Kim
J. O.
,
Kim
Y. J.
,
Chang
W. S.
,
Kang
C. Y.
.
2013
.
Tumor microenvironmental conversion of natural killer cells into myeloid-derived suppressor cells.
Cancer Res.
73
:
5669
5681
.
42
Fuertes
M. B.
,
Kacha
A. K.
,
Kline
J.
,
Woo
S. R.
,
Kranz
D. M.
,
Murphy
K. M.
,
Gajewski
T. F.
.
2011
.
Host type I IFN signals are required for antitumor CD8+ T cell responses through CD8alpha+ dendritic cells.
J. Exp. Med.
208
:
2005
2016
.
43
Chiossone
L.
,
Chaix
J.
,
Fuseri
N.
,
Roth
C.
,
Vivier
E.
,
Walzer
T.
.
2009
.
Maturation of mouse NK cells is a 4-stage developmental program.
Blood
113
:
5488
5496
.
44
Robbins
S. H.
,
Nguyen
K. B.
,
Takahashi
N.
,
Mikayama
T.
,
Biron
C. A.
,
Brossay
L.
.
2002
.
Cutting edge: inhibitory functions of the killer cell lectin-like receptor G1 molecule during the activation of mouse NK cells.
J. Immunol.
168
:
2585
2589
.
45
Omi
A.
,
Enomoto
Y.
,
Kiniwa
T.
,
Miyata
N.
,
Miyajima
A.
.
2014
.
Mature resting Ly6C(high) natural killer cells can be reactivated by IL-15.
Eur. J. Immunol.
44
:
2638
2647
.
46
Stojanovic
A.
,
Fiegler
N.
,
Brunner-Weinzierl
M.
,
Cerwenka
A.
.
2014
.
CTLA-4 is expressed by activated mouse NK cells and inhibits NK Cell IFN-γ production in response to mature dendritic cells.
J. Immunol.
192
:
4184
4191
.
47
Koebel
C. M.
,
Vermi
W.
,
Swann
J. B.
,
Zerafa
N.
,
Rodig
S. J.
,
Old
L. J.
,
Smyth
M. J.
,
Schreiber
R. D.
.
2007
.
Adaptive immunity maintains occult cancer in an equilibrium state.
Nature
450
:
903
907
.
48
Fogel
L. A.
,
Yokoyama
W. M.
,
French
A. R.
.
2013
.
Natural killer cells in human autoimmune disorders.
Arthritis Res. Ther.
15
:
216
.
49
Brahmer
J. R.
,
Tykodi
S. S.
,
Chow
L. Q.
,
Hwu
W. J.
,
Topalian
S. L.
,
Hwu
P.
,
Drake
C. G.
,
Camacho
L. H.
,
Kauh
J.
,
Odunsi
K.
, et al
.
2012
.
Safety and activity of anti-PD-L1 antibody in patients with advanced cancer.
N. Engl. J. Med.
366
:
2455
2465
.
50
Dong
H.
,
Strome
S. E.
,
Salomao
D. R.
,
Tamura
H.
,
Hirano
F.
,
Flies
D. B.
,
Roche
P. C.
,
Lu
J.
,
Zhu
G.
,
Tamada
K.
, et al
.
2002
.
Tumor-associated B7-H1 promotes T-cell apoptosis: a potential mechanism of immune evasion.
Nat. Med.
8
:
793
800
.
51
Terme
M.
,
Ullrich
E.
,
Aymeric
L.
,
Meinhardt
K.
,
Coudert
J. D.
,
Desbois
M.
,
Ghiringhelli
F.
,
Viaud
S.
,
Ryffel
B.
,
Yagita
H.
, et al
.
2012
.
Cancer-induced immunosuppression: IL-18-elicited immunoablative NK cells.
Cancer Res.
72
:
2757
2767
.

The authors have no financial conflicts of interest.

Supplementary data