Activation of NK cells by target cells leads to cytotoxicity as well as production of various cytokines including IFN-γ. MHC class I molecules on target cells regulate NK cytotoxicity. However, little is known about the regulation of IFN-γ production by NK cells. We examined the production of IFN-γ in individual murine NK cells stimulated with tumor cell lines by flow cytometric analysis of intracellular IFN-γ. Among several tumor lines tested, the rat basophilic leukemia line RBL-1 induced particularly high level of IFN-γ production in IL-2-activated NK cells, whereas other lines, including the prototypic NK target YAC-1, induced very low or no IFN-γ production. Transfection of murine classical MHC class I molecules into RBL-1 cells substantially inhibited IFN-γ production. This inhibition of IFN-γ production by MHC class I was independent of Ly-49 or CD94/NKG2A expression on NK cells. These results indicate that some target cells directly stimulate IL-2-activated NK cells and induce IFN-γ production, but the requirements for the induction of IFN-γ production seem different from those for NK cytotoxicity. Furthermore, similar to NK cytotoxicity, induction of IFN-γ production is inhibited by MHC class I on stimulating cells. However, the MHC class I-specific receptors inhibiting IFN-γ production are different from those for NK cytotoxicity.

Natural killer cells play a major role in early defense against viruses, intracellular bacteria, and parasites and tumor cells by displaying direct cell-mediated cytotoxicity as well as by producing various cytokines (1). Among the cytokines produced by NK cells, IFN-γ is thought to be particularly important as it activates macrophages to kill phagocytosed pathogens and to secrete various cytokines. IFN-γ also promotes Th1 immune responses (2). Therefore, IFN-γ production by NK cells seems to be one of the important links between innate and adaptive immune responses.

NK cytotoxicity is regulated by a balance between receptor-induced stimulatory and inhibitory signals (3). Whereas the identities of stimulatory NK receptors remain unclear, recent studies have identified multiple NK inhibitory receptors that recognize class I MHC on target cells. In mice, these inhibitory receptors belong to two families of C-type lectins, namely the Ly-49 family (4) and CD94/NKG2 heterodimers (5, 6). The former recognizes specific classical class I MHC, whereas the latter recognize the nonclassical MHC class I Qa-1b (5). It is thought that NK cells are activated by a wide range of target cells but are maintained to be self-tolerant due to the inhibitory receptors that recognize self-MHC on normal cells (7, 8).

NK cells can be activated to produce cytokines by monocyte-derived cytokines. Although IL-12, IL-15, or IL-18 alone does not induce significant cytokine production in resting human NK cells, combinations of two of these cytokines are potent inducers of IFN-γ production (9). IFN-γ production by NK cells can also be induced by some tumor cells (10, 11). However, the regulatory mechanisms controlling NK cell IFN-γ production are largely unknown.

In this study, we examined the production of IFN-γ by IL-2-activated NK cells stimulated with tumor cell lines. Our results show that activation of IFN-γ production and cytotoxicity are differentially regulated in NK cells. Furthermore, classical MHC class I molecules on stimulating cells inhibit IFN-γ production. However, Ly-49 or CD94/NKG2A does not seem to be responsible for the inhibition, suggesting that murine NK cells express novel inhibitory receptors for MHC class I that inhibit IFN-γ production.

C57BL/6 mice were purchased from The Jackson Laboratory (Bar Harbor, ME). Tumor cell lines were obtained from American Type Culture Collection (Manassas, VA). The rat basophilic leukemia RBL-1 lines transfected with mouse MHC class I have been described (12).

The hybridomas PK136 (anti-NK1.1), 20-8-4S (anti-Db, Kb, Kd), 28-8-6S (anti-Db), 3-83P (anti-Kk), and 34-5-8S (anti-Dd) were obtained from the American Type Culture Collection (Manassas, VA). The mAbs YE1/48 (anti-Ly-49A) and 5E6 (anti-Ly-49C and -I) have been described (13). For the staining of RBL-1 cells transfected with murine MHC class I, cells were first incubated with anti-MHC class I mAbs at 4°C for 30 min. After two washes, the cells were incubated with appropriate FITC-conjugated secondary Abs (Jackson ImmunoResearch, West Grove, PA) for 30 min at 4°C and analyzed by FACSCalibur (Becton Dickinson, San Jose, CA).

Adherent IL-2-activated NK cells were generated as described (14). In brief, spleen cells were passed through a nylon wool column to remove B cells. The nylon wool-nonadherent cells were incubated on tissue culture plates for 1 h to remove adherent cells. Nonadherent cells were cultured at a density of 106 cells/ml in RPMI 1640 medium supplemented with 10% FCS, 5 × 10−5 M 2-ME, and 1000 U/ml murine IL-2. After 3 days of culture, nonadherent cells were removed, and the adherent cells were then cultured in the presence of IL-2 for an additional 4 days.

For specific lysis of RBL-1 target cells, which were poorly labeled with 51Cr, we used nonradioactive cytotoxicity assay kit (CytoTox 96, Promega, Madison, WI) in which lactate dehydrogenase released from lysed target cells was quantitatively measured according to the manufacturer’s protocols. For other target cells, specific lysis was determined by using a standard 4-h 51Cr release assay as previously described (14). Both assays well correlated with each other when YAC-1 cells were used as target cells. For the isolation of Ly-49A+ NK cells, IL-2-activated NK cells were enriched by using the miniMACS (Miltenyi Biotec, Auburn, CA) immunomagnetic-based separation method. In this method, cells were first incubated with 1 μg/ml of biotinylated YE1/48 mAb followed by streptavidin microbeads. The labeled cells were then passed through the RS+ separation column. Ly-49A+ NK cells retained in the column were removed by several rinses with saline. The purified NK cells were subsequently cultured overnight in 1000 U/ml of IL-2 before used in cytotoxicity assays. For mAb blocking assay, effector cells were preincubated for 30 min at room temperature with 50 μg/ml mAb before the addition of target cells.

NK cells (2 × 105) and stimulator cells were washed and mixed at a ratio of 2:1 in 200 μl of RPMI 1640 + 10% FCS in sterile 96-well plates, centrifuged at 1000 rpm for 1 min, and incubated at 37°C for 11 to 12 h. Brefeldin A (10 μg/ml, Sigma, St. Louis, MO) was added during the final 6 h to inhibit cytokine secretion. Intracellular IFN-γ was determined by flow cytometric analysis as described by Prussin et al. (15) with slight modifications. Briefly, cells were washed and incubated for 30 min with PK136-FITC on ice in the dark. After washing, cells were fixed with 4% paraformaldehyde for 5 min at room temperature. On the following day, cells were treated for 30 min with cell permeation buffer (0.1% saponin, 1 mM CaCl2, 1 mM MgSO4, 0.05% NaN3, 1% BSA, 10 mM HEPES), and then incubated for 30 min with PE-labeled IFN-γ-specific mAb (PharMingen, San Diego, CA) or PE-labeled rat IgG1 (PharMingen) in the permeation buffer (2.5 μg/ml). Cells were washed with the permeation buffer, resuspended in PBS, and analyzed by FACSCalibur. Tumor cells were gated out based on their high forward scatter. In some experiments, NK cells and RBL-1 cells were incubated in 24-well flat-bottom plates with transwell inserts with 3-μm pore size membranes (Becton Dickinson). NK cells (1 × 106) were incubated in the lower wells, whereas RBL-1 cells (5 × 105) were incubated in the upper wells in 1 ml media. In control wells, NK cells were incubated without RBL-1 cells or were incubated together with RBL-1 cells in the lower wells.

The Qa-1b tetramer has been described (16). Briefly, a tetrameric complex of biotinylated Qa-1b, human β2-microglobulin, and peptide (AMAPRTLL) was formed using Streptavidin Red 670 (Life Technologies, Gaithersburg, MD) at a 4:1 molar ratio. For cell staining, Qa-1b tetramer was used at a total protein concentration of 7 μg/ml in PBS, and cells were incubated for 30 min on ice.

Murine NK cells were stimulated with various NK-sensitive tumor cell lines, and intracellular IFN-γ content in individual NK cells (NK1.1+) was determined by two-color flow cytometry. When freshly isolated NK cells were used, the tumor cell lines tested in this study did not induce a significant amount of IFN-γ production (data not shown). However, adherent IL-2-activated NK cells that were cultured for 7 days in the presence of IL-2 produced variable levels of IFN-γ on incubation with tumor cells (Fig. 1). The intracellular IFN-γ accumulated relatively slowly as compared with NK cytotoxicity and peaked at 12 h of incubation with tumor cells (Fig. 2,A). This was dependent on the stimulation with tumor cells, in that IL-2-activated NK cells incubated with media alone did not produce IFN-γ (Fig. 2,B). Although most cell lines tested in this study were sensitive to NK cell cytotoxicity, the ability to stimulate IFN-γ production was quite variable. The percentages of NK1.1+ cells positively stained with anti-IFN-γ mAb as well as the intensities of the staining varied significantly depending on the cell lines used as stimulators (Table I). The prototypic NK target YAC-1 was highly sensitive to NK cytotoxicity, but stimulation of IFN-γ production was almost negligible. Among the cell lines tested, the rat basophilic leukemia line RBL-1 was the most effective stimulator of IFN-γ production. More than 70% of IL-2-activated NK cells produced IFN-γ in response to this cell line. Another rat cell line, YB2/0, was a much poorer stimulator of IFN-γ production. The induction of IFN-γ production by RBL-1 did not seem to be due to production of soluble factors by RBL-1, because culture supernatant of RBL-1 cells had no effect (Fig. 2,B). We also tested whether direct cell-cell contact between NK and RBL-1 cells is required for the stimulation of IFN-γ production using transwells. When NK cells were separated from RBL-1 cells by a porous membrane, they did not produce IFN-γ, whereas NK cell incubated in close contact with RBL-1 cells in the same wells produced significant amount of IFN-γ (Fig. 2 C). Therefore, direct cell-cell contact between NK and RBL-1 cells is required for the induction of IFN-γ production.

FIGURE 1.

Intracellular IFN-γ in NK cells stimulated with tumor cell lines. IL-2-activated NK cells were incubated without tumor cells and stained with anti-IFN-γ (A), incubated with RBL-1 and stained with IgG1 isotype control (B), incubated with YAC-1 and stained with anti- IFN-γ (C), or incubated with RBL-1 and stained with anti-IFN-γ (D). In all cases, cells were also stained with anti-NK1.1. Intracellular IFN-γ and surface NK1.1 were analyzed by flow cytometer as described in Materials and Methods. Tumor cells were larger than NK cells and were gated out by the forward scatter. The results are representative of 12 independent experiments.

FIGURE 1.

Intracellular IFN-γ in NK cells stimulated with tumor cell lines. IL-2-activated NK cells were incubated without tumor cells and stained with anti-IFN-γ (A), incubated with RBL-1 and stained with IgG1 isotype control (B), incubated with YAC-1 and stained with anti- IFN-γ (C), or incubated with RBL-1 and stained with anti-IFN-γ (D). In all cases, cells were also stained with anti-NK1.1. Intracellular IFN-γ and surface NK1.1 were analyzed by flow cytometer as described in Materials and Methods. Tumor cells were larger than NK cells and were gated out by the forward scatter. The results are representative of 12 independent experiments.

Close modal
FIGURE 2.

IFN-γ production by IL-2-activated NK cells stimulated with RBL-1 cells. A, IL-2-activated NK cells were incubated with RBL-1 cells at a ratio of 2:1 for 0, 6, 12, or 24 h, and intracellular IFN-γ in NK cells was analyzed. Percentages of NK1.1+ cells stained with anti-IFN-γ as well as the mean fluorescence intensities (MFI) of the staining are shown. B, NK cells were cultured alone (NK), with supernatant from RBL-1 cells (sRBL-1), or with RBL-1 cells at various ratios as indicated, and the percentages of NK1.1+ cells stained for intracellular IFN-γ were determined by flow cytometry. C, NK cells and RBL-1 cells were incubated in transwells at a ratio of 2:1. NK cells were either incubated without RBL-1 cells (NK alone), incubated with RBL-1 in the same wells allowing physical contact (together), or separated by porous membranes (separate), and the percentages of intracellular IFN-γ-positive cells were determined. The results are representative of two independent experiments.

FIGURE 2.

IFN-γ production by IL-2-activated NK cells stimulated with RBL-1 cells. A, IL-2-activated NK cells were incubated with RBL-1 cells at a ratio of 2:1 for 0, 6, 12, or 24 h, and intracellular IFN-γ in NK cells was analyzed. Percentages of NK1.1+ cells stained with anti-IFN-γ as well as the mean fluorescence intensities (MFI) of the staining are shown. B, NK cells were cultured alone (NK), with supernatant from RBL-1 cells (sRBL-1), or with RBL-1 cells at various ratios as indicated, and the percentages of NK1.1+ cells stained for intracellular IFN-γ were determined by flow cytometry. C, NK cells and RBL-1 cells were incubated in transwells at a ratio of 2:1. NK cells were either incubated without RBL-1 cells (NK alone), incubated with RBL-1 in the same wells allowing physical contact (together), or separated by porous membranes (separate), and the percentages of intracellular IFN-γ-positive cells were determined. The results are representative of two independent experiments.

Close modal
Table I.

IFN-γ production and cytotoxicity in response to various tumor cells

Cell LineTumor Type% IFN-γ+a% Specific Cytotoxicityb
20:110:15:12.5:1
YAC-1 Mouse T lymphoma 6 ± 1 79 ± 1 76 ± 4 68 ± 1 49 ± 5 
C1498 Mouse T lymphoma 17 ± 10 52 ± 6 33 ± 5 21 ± 4 11 ± 4 
EL4 Mouse T lymphoma 3 ± 1 54 ± 2 43 ± 3 32 ± 2 18 ± 2 
A20 Mouse B lymphoma 4 ± 1 80 ± 3 79 ± 3 71 ± 2 56 ± 3 
R1.1 Mouse T lymphoma 3 ± 1 56 ± 4 39 ± 1 31 ± 5 23 ± 1 
IC21 Mouse macrophage 4 ± 1 55 ± 4 44 ± 1 33 ± 2 20 ± 1 
GM979 Mouse erythroleukemia 2 ± 1 43 ± 1 34 ± 1 20 ± 3 12 ± 3 
RBL-1 Rat basophilic leukemia 71 ± 7 55 ± 5 48 ± 2 44 ± 2 28 ± 1 
YB2/0 Rat myeloma 10 ± 2 52 ± 5 43 ± 3 29 ± 3 18 ± 2 
Cell LineTumor Type% IFN-γ+a% Specific Cytotoxicityb
20:110:15:12.5:1
YAC-1 Mouse T lymphoma 6 ± 1 79 ± 1 76 ± 4 68 ± 1 49 ± 5 
C1498 Mouse T lymphoma 17 ± 10 52 ± 6 33 ± 5 21 ± 4 11 ± 4 
EL4 Mouse T lymphoma 3 ± 1 54 ± 2 43 ± 3 32 ± 2 18 ± 2 
A20 Mouse B lymphoma 4 ± 1 80 ± 3 79 ± 3 71 ± 2 56 ± 3 
R1.1 Mouse T lymphoma 3 ± 1 56 ± 4 39 ± 1 31 ± 5 23 ± 1 
IC21 Mouse macrophage 4 ± 1 55 ± 4 44 ± 1 33 ± 2 20 ± 1 
GM979 Mouse erythroleukemia 2 ± 1 43 ± 1 34 ± 1 20 ± 3 12 ± 3 
RBL-1 Rat basophilic leukemia 71 ± 7 55 ± 5 48 ± 2 44 ± 2 28 ± 1 
YB2/0 Rat myeloma 10 ± 2 52 ± 5 43 ± 3 29 ± 3 18 ± 2 
a

IL-2-activated NK cells were incubated with indicated cell lines. NK1.1-positive cells were gated, and the percentages of NK1.1 + cells that express intracellular IFN-γ were calculated. The results are the means ± SD of three independent experiments.

b

Percent specific cytotoxicity was examined at indicated E:T ratios, as described in Materials and Methods. The results are the means ± SD of two to five independent experiments, each done in triplicate.

To investigate whether expression of MHC class I on stimulator cells affects NK cell IFN-γ production, RBL-1 cells transfected with murine MHC class I were used as stimulators. The transfected RBL-1 cells expressed comparable levels of surface MHC class I molecules (Fig. 3). When incubated with MHC class I -transfected RBL-1, IFN-γ production by NK cells was significantly lower than the controls in which NK cells were incubated with either the parental or the vector-transfected RBL-1 cells (Fig. 4). Db and Kb, which are self-MHC for C57BL/6 NK cells, as well as foreign MHC class I Dd and Kk significantly inhibited IFN-γ production.

FIGURE 3.

Flow cytometric analysis of murine MHC class I expression on transfected RBL-1 cells. Untransfected control RBL-1 cells (open histograms) and MHC class I-transfected RBL-1 (filled histograms) were stained with appropriate anti-MHC class I mAbs and FITC-conjugated secondary Ab. The results are representative of three independent experiments.

FIGURE 3.

Flow cytometric analysis of murine MHC class I expression on transfected RBL-1 cells. Untransfected control RBL-1 cells (open histograms) and MHC class I-transfected RBL-1 (filled histograms) were stained with appropriate anti-MHC class I mAbs and FITC-conjugated secondary Ab. The results are representative of three independent experiments.

Close modal
FIGURE 4.

Inhibition of IFN-γ production by MHC class I. IL-2-activated NK cells were incubated with RBL-1 cells transfected with the indicated murine MHC class I. Parental RBL-1 (control) and RBL-1 transfected with vector alone (vector) were used as controls. Cells were stained with anti-IFN-γ and anti-NK1.1 mAbs and analyzed by flow cytometry. Intracellular IFN-γ staining of NK1.1+ cells (filled histograms) and staining with isotype control Ab (open histograms) are shown. The results are representative of eight independent experiments.

FIGURE 4.

Inhibition of IFN-γ production by MHC class I. IL-2-activated NK cells were incubated with RBL-1 cells transfected with the indicated murine MHC class I. Parental RBL-1 (control) and RBL-1 transfected with vector alone (vector) were used as controls. Cells were stained with anti-IFN-γ and anti-NK1.1 mAbs and analyzed by flow cytometry. Intracellular IFN-γ staining of NK1.1+ cells (filled histograms) and staining with isotype control Ab (open histograms) are shown. The results are representative of eight independent experiments.

Close modal

To determine whether the inhibition of IFN-γ production by MHC class I observed above is mediated by the binding of MHC class I to the Ly-49 family of NK-inhibitory receptors, we first examined the effect of MHC class I molecules on the cytotoxicity of Ly-49-positive NK cells. As shown in Fig. 5, Dd expressed on RBL-1 cells efficiently inhibited the cytotoxicity of Ly-49A-positive NK cells, and this inhibition was reversed by anti-Ly-49A Ab. This result showed that MHC class I molecules on the transfected RBL-1 cells are properly expressed and sufficient to inhibit NK cytotoxicity in a Ly-49-dependent manner. To determine whether the Ly-49 family of inhibitory receptors on NK cells is also responsible for the inhibition of IFN-γ production by class I MHC on the transfected RBL-1 cells, we performed multicolor flow cytometric analysis to simultaneously detect cytoplasmic IFN-γ and cell surface Ly-49 on NK1.1+ cells. The YE1/48 mAb detects Ly-49A that recognizes Dd (13), whereas the 5E6 mAb reacts with Ly-49C that recognizes all the class I MHC tested in this study as well as Ly-49I (17). Interestingly, production of IFN-γ in both YE1/48+ and YE1/48 NK subsets was significantly reduced by Dd expression on the stimulator RBL-1 cells (Fig. 6,A). Dd also inhibited IFN-γ production in 5E6+ and 5E6 NK cells. For all MHC class I molecules tested in this study, no significant difference in the degree of inhibition was observed between Ly-49+ and Ly-49 population (Fig. 6,B). Furthermore, preincubation of NK cells with a mixture of anti-Ly-49 mAbs, namely, YE1/48 (anti-Ly-49A), 5E6 (anti-Ly-49C and -I), and 4D11 (anti-Ly-49G), had no significant effect on IFN-γ production (data not shown). These results indicate that Ly-49 is not responsible for the inhibition of IFN-γ production by MHC class I. To determine whether CD94/NKG2A is involved in the inhibition of IFN-γ production, Qa-1b tetramer that binds to CD94/NKG2A (5, 16) was used to identify CD94/NKG2A+ NK cells. Flow cytometric analysis of NK cells stimulated with Kb- or Db-transfected RBL-1 showed that IFN-γ production in NK cells stained with Qa-1b tetramer as well as those not stained by Qa-1b were equally inhibited by these MHC class I molecules (Fig. 7). Therefore, CD94/NKG2 does not seem to be responsible for the inhibition of IFN-γ production by MHC class I.

FIGURE 5.

Ly-49-dependent inhibition of NK cytotoxicity by MHC class I. Ly-49A+ NK cells were incubated with RBL-1 cells or Dd-transfected RBL-1 cells with or without the anti-Ly-49A mAb YE1/48 at the indicated E:T ratios, and NK cytotoxicity was determined by the release of lactate dehydrogenase from RBL-1 cells as described in Materials and Methods. The results are representative of three independent experiments.

FIGURE 5.

Ly-49-dependent inhibition of NK cytotoxicity by MHC class I. Ly-49A+ NK cells were incubated with RBL-1 cells or Dd-transfected RBL-1 cells with or without the anti-Ly-49A mAb YE1/48 at the indicated E:T ratios, and NK cytotoxicity was determined by the release of lactate dehydrogenase from RBL-1 cells as described in Materials and Methods. The results are representative of three independent experiments.

Close modal
FIGURE 6.

Ly-49 expression and inhibition of IFN-γ production by MHC class I. A, IL-2-activated NK cells were incubated with either vector-transfected RBL-1 cells or Dd-transfected RBL-1 cells; stained with anti-IFN-γ, anti-Ly-49A, (YE1/48), and anti-NK1.1; and analyzed by three-color flow cytometry. The dot plots in the upper panel show intracellular IFN-γ vs cell surface Ly-49A expression of NK1.1+ cells. In the lower panel, intracellular IFN-γ in Ly-49A+ and Ly-49A NK cells (filled histograms) and control staining with isotype-matched Ab (open histograms) are shown. Results are representative of three independent experiments. B, IL-2-activated NK cells were incubated with MHC class I-transfected RBL-1 or control vector-transfected RBL-1 and stained with anti-IFN-γ and YE1/48 (anti-Ly-49A) or 5E6 (anti-Ly-49C/I). The percentages of NK1.1+ cells positively stained with anti-IFN-γ are shown for each NK subset. Results are means ± SD of three independent experiments.

FIGURE 6.

Ly-49 expression and inhibition of IFN-γ production by MHC class I. A, IL-2-activated NK cells were incubated with either vector-transfected RBL-1 cells or Dd-transfected RBL-1 cells; stained with anti-IFN-γ, anti-Ly-49A, (YE1/48), and anti-NK1.1; and analyzed by three-color flow cytometry. The dot plots in the upper panel show intracellular IFN-γ vs cell surface Ly-49A expression of NK1.1+ cells. In the lower panel, intracellular IFN-γ in Ly-49A+ and Ly-49A NK cells (filled histograms) and control staining with isotype-matched Ab (open histograms) are shown. Results are representative of three independent experiments. B, IL-2-activated NK cells were incubated with MHC class I-transfected RBL-1 or control vector-transfected RBL-1 and stained with anti-IFN-γ and YE1/48 (anti-Ly-49A) or 5E6 (anti-Ly-49C/I). The percentages of NK1.1+ cells positively stained with anti-IFN-γ are shown for each NK subset. Results are means ± SD of three independent experiments.

Close modal
FIGURE 7.

Inhibition of IFN-γ production by MHC class I and binding of Qa-1b tetramer to NK cells. IL-2-activated NK cells were stimulated with RBL-1 cells transfected with control vector (A and B) or those transfected with Db (C) or Kb (D), stained with anti-IFN-γ alone (A) or anti-IFN-γ and Qa-1b tetramer (B–D), and analyzed by three-color flow cytometry. Dot plots show the staining of intracellular IFN-γ vs Qa-1b binding of NK1.1+ cells. Results are representative of three independent experiments.

FIGURE 7.

Inhibition of IFN-γ production by MHC class I and binding of Qa-1b tetramer to NK cells. IL-2-activated NK cells were stimulated with RBL-1 cells transfected with control vector (A and B) or those transfected with Db (C) or Kb (D), stained with anti-IFN-γ alone (A) or anti-IFN-γ and Qa-1b tetramer (B–D), and analyzed by three-color flow cytometry. Dot plots show the staining of intracellular IFN-γ vs Qa-1b binding of NK1.1+ cells. Results are representative of three independent experiments.

Close modal

The results in this study demonstrated that: 1) some tumor cells directly activate NK cell production of IFN-γ, 2) MHC class I molecules on tumor cells inhibit this activation of IFN-γ production, and 3) inhibitory receptors on NK cells recognizing classical MHC class I responsible for this inhibition seem different from known murine NK receptors for MHC class I. The results suggest that murine NK cells express novel MHC class I-specific inhibitory receptors that regulate IFN-γ production. They further suggest that activation signals required for NK cytotoxicity and IFN-γ production may be different. Many tumor cell lines, including the standard NK target YAC-1, are highly sensitive to NK cytotoxicity, and yet they are very poor inducers of IFN-γ production. NK cytotoxicity and IFN-γ production also differ in their time courses. The former is rapidly triggered on contact with susceptible target cells (18) and can be detected within 3 h, whereas IFN-γ production is a slower process and takes 12 h incubation to reach the maximum level (Fig. 2 A). This difference is likely due to the fact that NK cytotoxicity is mediated by exocytosis of preformed perforin and granzymes on activation of NK cells by targets, whereas IFN-γ production involves de novo gene transcription (19, 20, 21). The differences also suggest that ligands on tumor cells triggering NK cytotoxicity do not seem to be able to provide activation signals required for the induction of IFN-γ production. It is also possible that additional stimulatory signals may be required for IFN-γ production. The differential regulation of NK cytotoxicity and cytokine production has also been suggested by previous studies. Some human NK cell clones killed K562 and 721.221 targets equally well, but they secreted more IFN-γ in response to K562 than to 721.221 cells (10).

The possible difference in stimulatory signals for NK cytotoxicity and IFN-γ production may explain why Ly-49 or CD94/NKG2A does not seem to provide inhibitory signals for IFN-γ production. It has been established that Ly-49A recognizes Dd (22, 23) whereas Ly-49C recognizes a wide range of classical MHC class I molecules (13). Consistent with these findings, we have previously shown that MHC class I-transfected RBL-1 cells bind to Ly-49-transfected COS cells (12). Furthermore, we showed in this study that Ly-49 on IL-2-activated NK cells are able to recognize murine MHC class I molecules on the transfected RBL-1 cells and provide inhibitory signals for NK cytotoxicity. Although the same MHC class I molecules on RBL-1 also inhibited IFN-γ production of NK cells, this inhibition was Ly-49 independent. Two-color flow cytometric analysis showed that expression of Dd on RBL-1 inhibits IFN-γ production in not only Ly-49A+ but also Ly-49A NK cell subsets. Similarly, there was no correlation between inhibition of IFN-γ production by any of the MHC class I molecules tested and the expression of Ly-49C/I on NK cells detected by the 5E6 mAb. Furthermore, anti-Ly-49 mAb had no effects on the inhibition of IFN-γ production by MHC class I (data not shown), whereas the same Ab reversed the inhibition of NK cytotoxicity by MHC class I on target cells. These results indicate that Ly-49 is not responsible for the inhibition of IFN-γ production. It is also unlikely that CD94/NKG2A is responsible for the inhibition. CD94/NKG2A recognizes the nonclassical MHC class I Qa-1b (5) that presents peptides derived from the leader peptides of certain classical MHC class I molecules (24, 25). It is possible that transfection of murine classical MHC class I into RBL-1 cells may enhance expression of the rat Qa-1b homologue. However, transfection of different MHC class I, regardless of whether they contain the Qa-1 determinant modifier motif (25) in the leader sequence, inhibited IFN-γ production. Moreover, the inhibition of IFN-γ production did not correlate with the expression of CD94/NKG2 on NK cells, as determined by the binding of tetrameric Qa-1b.

Taken together, these results strongly suggest that murine NK cells express novel MHC class I-specific receptors that inhibit IFN-γ production, the identity of which is currently unknown. It is also not known whether a single receptor recognizes different MHC class I molecules or another family of receptors, each recognizing specific MHC class I, mediates the inhibition of IFN-γ production. In humans, killer cell-inhibitory receptors that belong to the Ig superfamily have been reported to inhibit both cytotoxicity and cytokine production (10). In addition, another type of MHC class I-specific inhibitory receptors belonging to the Ig superfamily, leukocyte immunoglobulin-like receptor (26), has been identified on human lymphocytes. However, murine homologues of killer cell-inhibitory receptors and leukocyte immunoglobulin-like receptor have not been identified. Another receptor, termed gp49, that also belongs to the Ig superfamily has been identified on murine NK cells and mast cells (27, 28), but the ligands for gp49 are not known. It remains to be determined whether any of these receptors are involved in the inhibition of IFN-γ production by MHC class I.

Direct cell-mediated cytotoxicity and cytokine production are thought to be two important functions of NK cells. Although recent studies have revealed the mechanisms by which NK cytotoxicity is regulated by MHC class I on target cells, little is known about the regulation of IFN-γ production by NK cells. Our study has shown, for the first time, that MHC class I molecules on tumor cells inhibit NK cell IFN-γ production by a mechanism distinct from that responsible for the inhibition of NK cytotoxicity.

We thank Dr. W. Jefferies for the Dk gene and Dr. D. Mager for critical reading of the manuscript.

1

This work was supported by a grant from the National Cancer Institute of Canada with core support provided by the British Columbia Cancer Agency. A.K. is a recipient of a Leukemia Research Fund of Canada fellowship, and S.L. is a recipient of a Deutsche Forschungsgemeinschaft fellowship.

1
Trinchieri, G..
1995
. Natural killer cells wear different hats: effector cells of innate resistance and regulatory cells of adaptive immunity and of hematopoiesis.
Semin. Immunol.
7
:
83
2
Gajewski, T. F., S. R. Schell, G. Nau, F. W. Fitch.
1989
. Regulation of T-cell activation: differences among T-cell subsets.
Immunol. Rev.
111
:
79
3
Lanier, L. L., B. Corliss, J. H. Phillips.
1997
. Arousal and inhibition of human NK cells.
Immunol. Rev.
155
:
145
4
Takei, F., J. Brennan, D. L. Mager.
1997
. The Ly-49 family: genes, proteins and recognition of class I MHC.
Immunol. Rev.
155
:
67
5
Vance, R. E., J. R. Kraft, J. D. Altman, P. E. Jensen, D. H. Raulet.
1998
. Mouse CD94/NKG2A is a natural killer cell receptor for the nonclassical major histocompatibility complex (MHC) class I molecule Qa-1(b).
J. Exp. Med.
188
:
1841
6
Lohwasser, S., P. Hande, D. L. Mager, F. Takei.
1999
. Cloning of murine NKG2A, B and C: second family of C-type lectin receptors on murine NK cells.
Eur. J. Immunol.
29
:
755
7
Raulet, D. H., I. Correa, L. Corral, J. Dorfman, M. F. Wu.
1995
. Inhibitory effects of class I molecules on murine NK cells: speculations on function, specificity and self-tolerance.
Semin. Immunol.
7
:
103
8
Salcedo, M., M. Andersson, S. Lemieux, L. Van Kaer, B. J. Chambers, H. G. Ljunggren.
1998
. Fine tuning of natural killer cell specificity and maintenance of self tolerance in MHC class I-deficient mice.
Eur. J. Immunol.
28
:
1315
9
Fehniger, T. A., M. H. Shah, M. J. Turner, J. B. VanDeusen, S. P. Whitman, M. A. Cooper, K. Suzuki, M. Wechser, F. Goodsaid, M. A. Caligiuri.
1999
. Differential cytokine and chemokine gene expression by human NK cells following activation with IL-18 or IL-15 in combination with IL-12: implications for the innate immune response.
J. Immunol.
162
:
4511
10
Kurago, Z. B., C. T. Lutz, K. D. Smith, M. Colonna.
1998
. NK cell natural cytotoxicity and IFN-gamma production are not always coordinately regulated: engagement of DX9 KIR+ NK cells by HLA-B7 variants and target cells.
J. Immunol.
160
:
1573
11
Koh, C. Y., D. Yuan.
1997
. The effect of NK cell activation by tumor cells on antigen-specific antibody responses.
J. Immunol.
159
:
4745
12
Lian, R. H., Y. Li, S. Kubota, D. L. Mager, F. Takei.
1999
. Recognition of class I MHC by NK receptor Ly-49C: identification of critical residues.
J. Immunol.
162
:
7271
13
Brennan, J., G. Mahon, D. L. Mager, W. A. Jefferies, F. Takei.
1996
. Recognition of class I major histocompatibility complex molecules by Ly-49: specificities and domain interactions.
J. Exp. Med.
183
:
1553
14
Kubota, A., S. Kubota, H. E. Farrell, N. Davis-Poynter, F. Takei.
1999
. Inhibition of NK cells by murine CMV-encoded class I MHC homologue m144.
Cell. Immunol.
191
:
145
15
Prussin, C., P. J. Openshaw.
1998
. Detection of intracellular cytokines by flow cytometry. J. E. Coligan, and A. M. Kruisbeek, and D. H. Margulies, and E. M. Shevach, and W. Strober, eds.
Current Protocols in Immunology
6241
Wiley, New York.
16
Salcedo, M., P. Bousso, H. G. Ljunggren, P. Kourilsky, J. P. Abastado.
1998
. The Qa-1b molecule binds to a large subpopulation of murine NK cells.
Eur. J. Immunol.
28
:
4356
17
Brennan, J., S. Lemieux, J. D. Freeman, D. L. Mager, F. Takei.
1996
. Heterogeneity among Ly-49C natural killer (NK) cells: characterization of highly related receptors with differing functions and expression patterns.
J. Exp. Med.
184
:
2085
18
Trinchieri, G..
1989
. Biology of natural killer cells.
Adv. Immunol.
47
:
187
19
Farrar, W. L., M. C. Birchenall-Sparks, H. B. Young.
1986
. Interleukin 2 induction of interferon-γ mRNA synthesis.
J. Immunol.
137
:
3836
20
Anegon, I., M. C. Cuturi, G. Trinchieri, B. Perussia.
1988
. Interaction of Fc receptor (CD16) ligands induces transcription of interleukin 2 receptor (CD25) and lymphokine genes and expression of their products in human natural killer cells.
J. Exp. Med.
167
:
452
21
Farrar, M. A., R. D. Schreiber.
1993
. The molecular cell biology of interferon-γ and its receptor.
Annu. Rev. Immunol.
11
:
571
22
Daniels, B. F., F. M. Karlhofer, W. E. Seaman, W. M. Yokoyama.
1994
. A natural killer cell receptor specific for a major histocompatibility complex class I molecule.
J. Exp. Med.
180
:
687
23
Kane, K. P..
1994
. Ly-49 mediates EL4 lymphoma adhesion to isolated class I major histocompatibility complex molecules.
J. Exp. Med.
179
:
1011
24
Kurepa, Z., C. A. Hasemann, J. Forman.
1998
. Qa-1b binds conserved class I leader peptides derived from several mammalian species.
J. Exp. Med.
188
:
973
25
Aldrich, C. J., A. DeCloux, A. S. Woods, R. J. Cotter, M. J. Soloski, J. Forman.
1994
. Identification of a Tap-dependent leader peptide recognized by alloreactive T cells specific for a class Ib antigen.
Cell
79
:
649
26
Cosman, D., N. Fanger, L. Borges, M. Kubin, W. Chin, L. Peterson, M. L. Hsu.
1997
. A novel immunoglobulin superfamily receptor for cellular and viral MHC class I molecules.
Immunity
7
:
273
27
Rojo, S., D. N. Burshtyn, E. O. Long, N. Wagtmann.
1997
. Type I transmembrane receptor with inhibitory function in mouse mast cells and NK cells.
J. Immunol.
158
:
9
28
Wang, L. L., I. K. Mehta, P. A. LeBlanc, W. M. Yokoyama.
1997
. Mouse natural killer cells express gp49B1, a structural homologue of human killer inhibitory receptors.
J. Immunol.
158
:
13