Lytic Granule Polarization, Rather than Degranulation, Is the Preferred Target of Inhibitory Receptors in NK Cells

Asmita Das and Eric O. Long
J Immunol October 15, 2010, 185 (8) 4698-4704; DOI: https://doi.org/10.4049/jimmunol.1001220

Abstract

Natural cytotoxicity is achieved by polarized release of perforin and granzymes at the NK cell–target cell immunological synapse. Signals for granule polarization and degranulation can be uncoupled in NK cells, which raises the question of their respective sensitivity to inhibitory receptors. Expression of either HLA-C or HLA-E on the human cell line 721.221 blocked granule polarization, degranulation, and CD16-dependent MIP-1α secretion by NK cell clones that expressed inhibitory receptors of matching HLA specificity. To test inhibition of signals for polarization and degranulation separately, Drosophila S2 cells expressing ICAM-1 with either HLA-C or HLA-E were used. CD16-dependent degranulation and MIP-1α secretion were not fully inhibited, suggesting that other receptor–ligand interactions, which occur with 721.221 cells, contribute to inhibition. In contrast, HLA-C or HLA-E on S2 cells coexpressing ICAM-1 or ULBP1 were sufficient to block granule polarization induced by either LFA-1 or NKG2D, even during concomitant CD16-dependent degranulation. Similarly, expression of a ligand for NKR-P1A on S2 cells inhibited granule polarization but not CD16-induced degranulation. Therefore, granule polarization, rather than degranulation, is the preferred target of inhibitory receptors in NK cells.

Natural killer cell cytotoxicity is under negative control by inhibitory receptors, such as human killer cell Ig-like receptors (KIRs) specific for HLA-B and HLA-C, and CD94-NKG2A, which binds to HLA-E (1). Upon binding to MHC class I on target cells, inhibitory receptors become phosphorylated on tyrosine residues within ITIMs and recruit the tyrosine phosphatase SH2-containing protein tyrosine phosphatase-1 (SHP-1) to terminate NK cell activation signals at a very proximal step (1). Dephosphorylation of the guanine nucleotide exchange factor Vav1 by SHP-1 during KIR-mediated inhibition may be sufficient to block NK cell cytotoxicity (1, 2), as Vav1 is essential for actin-dependent receptor clustering, immune synapse formation, and for signaling by the NK cell activation receptor NKG2D-DAP10 (3, 4).

Polarized fusion with the plasma membrane of lytic granules containing granzymes and perforin is required for Fas ligand-independent CTL and NK cell cytotoxicity (5, 6). Signals for granule polarization and degranulation can be uncoupled in NK cells (7). Expression of ICAM-1, a ligand of β2 integrin LFA-1, on Drosophila cells was sufficient to induce polarization of granules but not degranulation. Conversely, engagement of the IgG Fc receptor CD16 by rabbit IgG on Drosophila cells induced degranulation without specific polarization (7). A question raised by these findings is whether inhibitory receptors block a proximal step in a common signaling pathway upstream of polarization and degranulation, thereby blocking both, or selectively inhibit one of those two processes. Selective inhibition of either polarization or degranulation would be sufficient to protect a target cell from lysis by NK cells. Inhibition of granule polarization in KIR+ NK cell clones by HLA-C on target cells has been reported (8). However, it is not clear if the inhibitory receptor–HLA class I interaction is sufficient to confer inhibition, independently of the many other receptor–ligand interactions that occur during NK cell–target cell contact. The Drosophila cell line S2 transfected with various ligands of human NK cell receptors, including HLA class I ligands of inhibitory receptors, is a system well suited to address these questions (9–11). Here we show that granule polarization is highly sensitive to inhibition, whereas degranulation has more complex requirements for inhibition.

Materials and Methods

Cells and Abs

Human NK cells were isolated and cloned as described (12) and screened for expression of CD158a, CD158b, and CD159a by flow cytometry. The HLA class I-negative cell line 721.221 (13), referred to as 221 cells, and 221-Cw3, 221-Cw15, and 221-E (0.221-AEH, a gift of D. Geraghty, Fred Hutchinson Cancer Research Center, Seattle, WA) cells were cultured as described (14). Drosophila S2 cells were transfected with the insect expression vector pAc5.1 (Invitrogen, Carlsbad, CA) along with the vector pNeofly as described (15). Full-length lectin-like transcript 1 (LLT1) cDNA in pAc5.1 was a gift of Y. Bryceson (Laboratory of Immunogenetics, National Institute of Allergy and Infectious Diseases, Rockville, MD). HLA class I on S2 cells were loaded with specific peptides (GAVDPLLAL for HLA-Cw3, QYDDAVYKL for HLA-Cw4, and VMAPRTLIL for HLA-E). Abs used were CD158a (EB6), CD158b (GL183), and CD159a (Z199) (Immunotech, Beckman Coulter, Miami, FL) for screening of NK clones, rabbit anti-S2 cell serum for stimulation through CD16 (7), anti-perforin for detecting perforin polarization (7), mAb F4/326 (anti–HLA-C; a gift of S.Y. Yang, Histogenetics, New York, NY), anti–HLA-E mAbs MEM-E/06 and MEM-E/07 (Exbio Antibodies, Vestec, Czech Republic), anti-CD54 (BD Pharmingen, San Diego, CA), anti-ULBP1 (R&D Systems, Minneapolis, MN), and anti-flag Ab, M2 (Sigma Aldrich, St. Louis, MO), for detecting expression of flag-tagged LLT1. To engage CD16 on NK cells, S2 cells were incubated with rabbit anti-S2 cell serum at a 1 × 10−4 dilution (unless indicated otherwise), and 221 cells were incubated with anti–HLA-DR IgG2a mAb L243 at 1 μg/ml.

Cytotoxicity, granule polarization, ELISPOT, and ELISA

NK cell cytotoxicity toward 221 cells was determined by a europium assay (2). Polarization of perforin-containing granules was determined as described (7). Conjugates with polarized perforin were scored visually from three-dimensional confocal z-stacks. Perforin-containing granules were considered polarized when most of the fluorescence was concentrated in the lower quadrant of the NK cell (i.e., the quadrant that is closest to the target cell). The distinction between polarized and unpolarized granules was obvious in most cases. ELISPOT assays were set up to count the number of NK cells that secreted granzyme B (GrzB) above a set threshold level, as described (16). Effector and target cell suspensions were applied in triplicate at a 1:1 ratio and incubated at 37°C for 3 h. Spots were enumerated with an ELISPOT plate reader (CTL Immunospot, Shaker Heights, OH). The number of ELISPOTs obtained with NK cells alone was subtracted from the experimental values. Cells (2 × 105 NK and target cells; 1:1) were incubated overnight at 37°C. MIP-1α released was determined by ELISA (R&D Systems).

Results

Inhibitory receptors for HLA-C and HLA-E block granule polarization and granzyme release induced by 721.221 cells

A large panel of NK cell clones was assembled from peripheral blood of several individuals. To test for inhibition through KIR or CD94-NKG2A receptors, the HLA class I-negative human cell line 721.221 (referred to as 221 cells) transfected with HLA-Cw3 (221-Cw3), HLA-Cw15 (221-Cw15), or HLA-E (221-E) was used (Fig. 1A). HLA-Cw3 is a ligand for the CD158b inhibitory receptors KIR2DL2 and KIR2DL3. HLA-Cw15 is a ligand for the CD158a inhibitory receptor KIR2DL1. HLA-E is a ligand for the inhibitory receptor CD94-NKG2A. The inhibitory function of receptors on NK cell clones was first established in cytotoxicity assays with transfected 221 cells (Fig. 1B). This was necessary because most Abs do not distinguish between inhibitory KIR2DL1 (CD158a) and activating KIR2DS1 (CD158h) or between inhibitory KIR2DL2/3 (CD158b) and activating KIR2DS2 (CD158j). Also, HLA-E binds to both inhibitory CD94-NKG2A and activating CD94-NKG2C receptors. A typical lysis assay with a clone coexpressing CD158b and NKG2A is shown in Fig. 1B. The observed inhibition of lysis with 221-Cw3 and 221-E but not 221-Cw15 cells indicated that this particular NK clone expressed inhibitory CD158b and NKG2A, consistent with GL183 and Z199 mAb staining (not shown). Every NK clone used in this study underwent such functional screening for expression of inhibitory receptors.

FIGURE 1.
FIGURE 1.

Inhibition of NK cell cytotoxicity and granule polarization by HLA-C and HLA-E on 221 cells. A, The 221-Cw3, 221-Cw15, and 221-E cells were stained with control Ab (light gray line) and with Abs for HLA-C or HLA-E (dark lines). B, Cytotoxicity of a CD158a–b+NKG2A+ NK clone toward 221 (circles), 221-Cw15 (diamonds), 221-E (squares), and 221-Cw3 (triangles) cells. C, Representative confocal images of a CD158a–b+NKG2A+ NK clone conjugated with 221, 221-Cw3, or 221-E cells and stained for perforin with an mAb followed by an Alexa 488-conjugated goat anti-mouse IgG Ab. D, Polarization was scored as polarized (black) and nonpolarized (white). NK cells in contact with Drosophila S2 cells served as control. The number (n) of conjugates scored is listed.

To test whether inhibitory receptors can block granule polarization, conjugates of NK cell clones with 221-Cw3 and 221-E cells were fixed, stained for intracellular perforin, and analyzed by three-dimensional imaging of confocal z-stacks. Perforin-containing granules in a CD158a–b+NKG2A+ NK clone polarized toward 221 target cells but not when mixed with either 221-Cw3 or 221-E resistant cells (Fig. 1C, 1D). As control, polarization was scored for NK cells in contact with the insect cell line S2 (Fig. 1D). In this and in many other experiments, granules appeared to be polarized toward S2 cells in 20–30% of NK cell contacts. As granules are often polarized to one side of NK cells, independently of any particular stimulus, one would expect random contacts with S2 cells to give the appearance of polarization in a fair proportion of NK cells. Nevertheless, the distinction between active granule polarization (>60% of conjugates) and inhibition of granule polarization (apparent polarization in 20–30% of conjugates, which is equivalent to the S2 cell control) was always clear. Inhibition of granule polarization was also observed with other NK clones and with the cell line NK92 transfected with inhibitory receptor KIR2DL1 (data not shown). Therefore, we conclude that inhibitory signals delivered by KIR and by CD94-NKG2A block the movement of lytic granules toward the NK cell immunological synapse.

To determine whether inhibition of target cell lysis was also achieved by a block in the fusion of lytic granules with the plasma membrane, a degranulation assay was developed. The release of GrzB by NK clones was measured by an ELISPOT assay. A large panel of NK cell clones expressing either CD158a, CD158b, or both were tested for their ability to release GrzB during mixing with either 221 cells, 221-Cw3 cells, or 221-Cw15 cells (Fig. 2A). The data are presented as the number of ELISPOTs relative to the number of spots obtained with 221 cells (i.e., in the absence of inhibition), and each data point represents a different NK clone. Degranulation was clearly inhibited whenever a specific KIR–HLA-C receptor–ligand combination was present, although the extent of inhibition varied among clones that shared a KIR phenotype. In addition, partial but significant inhibition was observed with CD158a–b+ clones mixed with 221-Cw15 cells (p = 0.006), consistent with the reported cross-reactivity of KIR2DL2 with group 2 HLA-C allotypes, such as HLA-Cw15 (17, 18). This cross-reactivity is also consistent with the more complete inhibition of double-positive CD158a+b+ clones by 221-Cw15 cells compared with inhibition of CD158a+b– clones (p = 0.02).

FIGURE 2.
FIGURE 2.

Inhibition of GrzB release. A, CD158a+b–, CD158a–b+, and CD158a+b+ NK cell clones, as indicated, were mixed with 221, 221-Cw3, or 221-Cw15 cells. B, NKG2A– and NKG2A+ NK clones mixed with 221 and 221-E cells. Results are expressed as GrzB ELISPOTs relative to those obtained with 221 target cells. Each dot represents a single NK clone. The thick line and bars represent the mean and SD, respectively. The p values by two-tailed t test were: n.s., nonsignificant; *p < 0.05; ***p < 0.0001.

The ability of NKG2A to inhibit GrzB release was tested using 221-E target cells (Fig. 2B). Inhibition was observed with every clone expressing NKG2A. Individual variability in the extent of inhibition was again observed among different clones. NK cell clones that do not express CD94-NKG2A are uncommon. Only four such clones were tested. As expected, no inhibition with 221-E target cells was observed (Fig. 2B). Therefore, engagement of inhibitory receptors by HLA class I on 221 cells resulted in inhibition of both granule polarization and GrzB release.

GrzB and MIP-1α release induced by CD16 is not blocked by inhibitory receptor ligands on insect cells

It is not known how the many receptor–ligand interactions that occur during NK-221 cell contact contribute to the inhibition mediated by HLA class I-specific inhibitory receptors. To address this question, we used the Drosophila cell line S2 transfected with ligands of human NK cell receptors, including HLA class I ligands of inhibitory receptors. Stimulation through CD16 was achieved using a rabbit antiserum to S2 cells. Furthermore, S2 cells can be used to separate signals for polarization, through LFA-1, from signals for degranulation, through CD16 (7). Addition of increasing amounts of rabbit antiserum to S2 cells resulted in a small increase of GrzB release, relative to the level of GrzB released spontaneously by IL-2–activated NK cells (Fig. 3A). However, GrzB release was strongly increased when rabbit antiserum was added to S2 cells expressing ICAM-1 (S2–ICAM-1). Half-maximal GrzB release occurred at an ∼1 × 10−5 dilution of anti-S2 rabbit antiserum (Fig. 3A). Engagement of CD16 by rabbit IgG on S2 cells resulted also in the release of MIP-1α (Fig. 3B). Unlike GrzB release, MIP-1α secretion did not show a pronounced dependence on ICAM-1. MIP-1α release assays were preferred over IFN-γ and TNF-α secretion because IL-2–activated NK cells release a high basal level of IFN-γ and TNF-α.

FIGURE 3.
FIGURE 3.

Coengagement of CD16 and LFA-1 on NK cells induces GrzB and MIP-1α release. A, GrzB released by NK cells mixed with Drosophila S2 cells (circles) and with S2–ICAM-1 cells (triangles) preincubated with no Ab (open symbol) or rabbit anti-S2 serum (filled symbols). The experiment is representative of three independent experiments. B, MIP-1α secretion by NK cells mixed with S2 cells or S2–ICAM-1 cells preincubated with or without rabbit anti-S2 serum (RIgG), as indicated. The experiment is representative of seven independent experiments.

To test for inhibition of GrzB and MIP-1α release, S2 cells were cotransfected with ICAM-1, β2-microglobulin, and either HLA-Cw3 (S2–ICAM-1+Cw3 cells) or HLA-E (S2–ICAM-1+E cells) (Fig. 4A). These cells were either loaded or not loaded with peptides that bound specifically to either HLA-Cw3 or HLA-E to test for inhibition of GrzB and MIP-1α release. As class I is transported “empty” to the surface of Drosophila cells, it cannot be recognized by inhibitory receptors unless loaded with exogenous peptide (10, 19). The inhibition induced by HLA class I on S2 cells was tested using the same cells, either before or after loading with specific peptides. With this approach, S2 cells provide a more rigorous test of HLA class I-dependent inhibition than that by use of independent transfected cells that do or do not express an HLA ligand for inhibitory receptors.

FIGURE 4.
FIGURE 4.

Binding of KIR to HLA-C on S2 cells is not sufficient to inhibit degranulation. A, Peptide-loaded S2–ICAM+Cw3 and S2–ICAM+E cells stained with mAbs for ICAM-1 (dark lines), HLA-C and HLA-E (gray lines), or with secondary Ab alone (hatched). B, GrzB release by 23 CD158b+ NK clones mixed with S2–ICAM-1+Cw3 cells and 9 NKG2A+ NK clones mixed with S2–ICAM-1+E cells. Two of the CD158b+ clones were NKG2A– (open circles), all others were NKG2A+ (filled circles). Three of the NKG2A+ clones were KIR– (open circles), all others expressed at least one of either CD158a or CD158b (filled circles). S2 cells had been preincubated with rabbit anti-S2 serum. Results are expressed as GrzB ELISPOTs relative to those obtained with the same cells in the absence of HLA-C or HLA-E–specific peptide. Statistical significance by two-tailed t test: n.s., nonsignificant; ***p < 0.0001. C, S2–ICAM+Cw3 cells either loaded (solid squares) or not loaded (open squares) with HLA-Cw3–specific peptide were incubated with the indicated dilutions of rabbit anti-S2 serum. The experiment is representative of four independent experiments using different NK clones.

In a panel of 23 NK clones that expressed inhibitory CD158b, most of the clones showed no inhibition of GrzB release during mixing with S2–ICAM-1+Cw3 cells that had been loaded with an HLA-Cw3–specific peptide and incubated with rabbit antiserum at a dilution of 1 × 10−4 (Fig. 4B). Partial inhibition, which was beyond the SD, was observed with only three clones. More pronounced but still partial inhibition was observed with CD94-NKG2A+ NK cells during incubation with S2–ICAM-1+E cells loaded with an HLA-E–specific peptide (Fig. 4B). These results suggested that inhibition of NK cell clones by peptide-loaded HLA class I on insect cells was not as efficient as inhibition by HLA class I on human 221 cells. One possibility we considered was that stimulation of NK cells by IgG Fc on S2–ICAM-1 cells was stronger than stimulation by 221 cells and may thus have overridden inhibitory signals. To compare directly the relative strengths of signals, GrzB release was measured with bulk cultures of NK cells, from three unrelated donors, after mixing with 221 cells, S2–ICAM-1 cells, and S2–ICAM-1 cells coated with rabbit IgG. The number of GrzB ELISPOTs obtained was 675 ± 81 with 221, 117 ± 50 with S2–ICAM-1, and 553 ± 62 with S2–ICAM-1+IgG. Therefore, stimulation of GrzB release was not greater with S2–ICAM-1+IgG Fc than with 221 cells. We also tested whether peptide-loaded HLA-Cw3 on S2 cells would inhibit weaker CD16 activation signals by measuring inhibition over a broad range of anti-S2 cell serum concentrations. No inhibition of degranulation was observed at any of the dilutions tested (Fig. 4C).

As another readout for inhibition, CD16-induced secretion of MIP-1α was evaluated. MIP-1α release induced by rabbit IgG on S2–ICAM-1+Cw3 or S2–ICAM-1+E cells was not inhibited by peptide-loaded HLA-Cw3 or HLA-E, as shown with CD158a–b+NKG2A+ NK clones (Fig. 5A, 5B). We conclude that CD16 signals for chemokine secretion are not inhibited by interaction of inhibitory receptors with peptide-loaded MHC class I molecules on insect cells.

FIGURE 5.
FIGURE 5.

Engagement of inhibitory KIRs or NKG2A by ligands on S2 cells, but not 221 cells, was insufficient to inhibit cytokine secretion. A, MIP-1α release by a CD158a–b+NKG2A+ NK clone mixed with S2 or S2–ICAM-1+Cw3 cells. Rabbit anti-S2 serum (RIgG) or HLA-Cw3–specific peptide were added, as indicated. B, MIP-1α release by a CD158a–b+NKG2A+ NK clone mixed with S2 or S2–ICAM-1+E cells. Rabbit anti-S2 serum (RIgG) or HLA-E–specific peptide were added, as indicated. C, MIP-1α release by a CD158a+CD158b–NKG2A+ NK clone mixed with 221-Cw3 or 221-Cw15 cells with or without 1 μg/ml anti–HLA-DR IgG2a Ab, as indicated. D, MIP-1α release by a CD158a–CD158b+NKG2A+ NK clone mixed with 221-Cw7 or 221-Cw15 cells with or without 1 μg/ml anti–HLA-DR IgG2a Ab, as indicated. The experiment in each case is representative of four independent experiments using different NK clones.

The difference between inhibition by HLA class I on 221 and S2 cells may be due to the different activation signals used. Natural ligands on 221 cells activate natural cytotoxicity, whereas S2 cells coated with rabbit IgG activate NK cells via CD16. To test whether activation of NK cells specifically through CD16 was less sensitive to inhibition than activation through receptor–ligand interactions that promote natural cytotoxicity toward 221 cells, we took advantage of our observation that strong MIP-1α secretion was dependent on CD16 signals, as shown with IgG2a-coated 221 cells (Fig. 5C). Despite this CD16-dependent enhancement, MIP-1α secretion by a CD158a+ NK clone was completely blocked by HLA-Cw15 (Fig. 5C). Similar results were obtained with 221-Cw7 cells and a CD158b+ NK clone (Fig. 5D). Therefore, the reduced inhibition observed with Drosophila S2 cells compared with that of human 221 cells cannot be explained simply by activation through CD16.

Granule polarization toward ICAM-1–expressing insect cells is blocked by NK cell inhibitory receptors

Although GrzB release was not inhibited by HLA class I on S2 cells (Fig. 4), the resistance of S2–HLA-C cells from lysis by KIR+ NK clones (10) suggests that inhibition occurs at some other activation step. Indeed, analysis of NK cell clones bound to peptide-loaded S2–ICAM-1+Cw3 cells showed complete inhibition of granule polarization (Fig. 6A, 6B). Likewise, polarization was completely inhibited in CD94-NKG2A+ NK cells mixed with peptide-loaded S2–ICAM-1+E cells (Fig. 6A, 6B). Inhibition of polarization was also observed with a CD158a+b–NKG2A+ clone incubated with S2–ICAM-1+Cw4 cells and with a CD158a–b+NKG2A+ clone and a CD158a–b–NKG2A+ clone incubated with S2–ICAM-1+E cells (data not shown). These results confirm that peptide-loaded HLA class I on Drosophila S2 cells is competent to induce functional inhibition of NK cells. We wished to test whether inhibition of polarization would still occur in the presence of strong CD16-mediated activation signals. We therefore tested whether granule polarization could still be inhibited while degranulation occurred by analyzing granule polarization in NK cells mixed with S2–ICAM-1+Cw3 or S2–ICAM+E cells that had been coated with rabbit IgG. Polarization was completely inhibited in a CD158a+b+NKG2A+ clone during incubation with IgG-coated S2–ICAM-1 cells expressing either HLA-Cw3 or HLA-E (Fig. 6B). These results demonstrate that granule polarization is preferentially inhibited over degranulation and that degranulation can still occur when polarization of the bulk of lytic granules is inhibited.

FIGURE 6.
FIGURE 6.

Inhibition of LFA-1-induced granule polarization by HLA-C and HLA-E expressed on S2 cells. A, CD158a+b+NKG2A+ NK clone in contact with S2–ICAM-1+Cw3 and S2–ICAM-1+E cells, with (+) or without (No) peptide, stained for perforin with an mAb followed by an Alexa 488-conjugated goat anti-mouse IgG Ab. B, Polarization was scored after mixing with S2 cells or with S2–ICAM-1+Cw3 or S2–ICAM-1+E cells in the presence or absence of RIgG or of peptide, as indicated. The experiment is representative of four independent experiments using different NK clones. The number (n) of conjugates scored is listed.

We then asked whether inhibitory receptors other than KIR and CD94-NKG2A would also block granule polarization. S2–ICAM-1 cells were cotransfected with LLT1, a ligand for inhibitory receptor NKR-P1A (CD161) (20, 21). Because NKR-P1A is expressed on most NK cells, it was not necessary to isolate NK clones to test for inhibition. IL-2–activated NK cells were mixed with S2 cells expressing either ICAM-1 or ICAM-1 and LLT1 (S2–ICAM-1+LLT1). LFA-1–dependent granule polarization was completely inhibited by LLT1 expression (Fig. 7A). Furthermore, as observed with S2 cells expressing ICAM-1 and ligands for inhibitory KIR and CD94-NKG2A, CD16-induced degranulation was not inhibited by ICAM-1 and LLT1 coexpression (Fig. 7B). We conclude that granule polarization is the preferred target for inhibition by different inhibitory receptors.

FIGURE 7.
FIGURE 7.

Expression of the NKR-P1A ligand LLT1 on S2 cells inhibited polarization of lytic granules but not degranulation. A, Polarization was scored after mixing with S2 cells, S2–ICAM-1 cells, or S2–ICAM-1+LLT1 cells. B, IL-2–activated NK cells were mixed with S2 cells, S2–ICAM-1 cells, or S2–ICAM-1+LLT1 cells preincubated with or without anti-S2 rabbit serum (RIgG), and GrzB ELISPOT assays were performed. The experiments are representative of four independent experiments using different NK cell populations.

Expression of ULBP1 on S2 cells is sufficient to induce granule polarization, which is blocked by inhibitory KIR

S2 cells expressing ULBP1, which is a ligand for activation receptor NKG2D, triggered polarization of perforin-containing granules in IL-2–activated NK cells (Fig. 8). We therefore tested whether NKG2D-dependent granule polarization was sensitive to inhibition, just as LFA-1–dependent polarization was. S2 cells coexpressing ULBP1 and HLA-Cw4 (S2-ULBP1+Cw4) were mixed with a CD158a+b–NKG2A+ clone. Addition of an HLA-Cw4–specific peptide completely inhibited polarization (Fig. 8A, 8B). As had been observed with S2–ICAM-1 cells, coengagement of CD16 by addition of rabbit IgG to S2 cells did not overcome inhibition (Fig. 8B). Therefore, inhibitory receptors are not restricted to inhibition of LFA-1 signals but can also prevent NKG2D-dependent signals for polarization.

FIGURE 8.
FIGURE 8.

Polarization induced by NKG2D signals is sensitive to inhibition by KIR2DL1 binding to peptide-loaded HLA-C on S2 cells. A, CD158a+b–NKG2A+ NK clone in contact with S2–ULBP1+Cw4 cells loaded with (+) or without (No) peptide, stained for perforin with an mAb followed by an Alexa 488-conjugated goat anti-mouse IgG Ab. B, Polarization was scored after mixing with S2 cells or S2–ULBP1+Cw4 cells in the presence or absence of RIgG or of peptide, as indicated. The number (n) of conjugates scored is listed. C, CD158a+b–NKG2A+ NK clone mixed with S2 cells or S2–ULBP1+Cw4 cells preincubated with or without anti-S2 rabbit serum (RIgG) or HLA-Cw4–specific peptide, as indicated. GrzB ELISPOT assays were performed. The experiments are representative of three independent experiments.

We observed that ULBP1 expression on S2 cells was also a strong costimulator of CD16-dependent degranulation (Fig. 8C), similar to results obtained with S2–ICAM-1 (Fig. 3). This gave us an opportunity to test inhibition of degranulation induced by the combination of CD16 with NKG2D. A CD158a+b–NKG2A+ NK clone was mixed with S2–ULBP-1+Cw4 cells coated with rabbit antiserum in the presence and absence of HLA-Cw4–specific peptide. No inhibition of degranulation was observed (Fig. 8C). These results confirm, using a different combination of activation signals, that granule polarization is the preferred target for inhibition by inhibitory receptors.

Discussion

We examined the sensitivity of granule polarization and degranulation by NK cells to inhibition by ITIM-containing receptors. Polarization of granules was always inhibited, irrespective of the inhibitory receptor tested or the cell on which the receptor ligand was expressed. In contrast, inhibition of degranulation, as measured by GrzB release, and of MIP-1α secretion was variable and dependent on the target cell. Inhibition was tested with three different HLA class I specificities on human 221 cells. Drosophila S2 cells were tested after expression of two HLA class I specificities and of the non-MHC ligand for inhibitory receptor NKR-P1A. Using S2 cells expressing defined combinations of ligands of NK cell receptors, it was possible to analyze inhibition of polarization and of degranulation separately. To study HLA class I-specific receptors, it was necessary to isolate and characterize NK cell clones because individual NK cells express different combinations of KIR and CD94-NKG2 receptors (22). NK cell clones expressing any combination of CD158a, CD158b, and CD94-NKG2A inhibitory receptors were selected for analysis.

Inhibition of degranulation by HLA-C and HLA-E on 221 cells varied among NK clones from very good to incomplete. However, peptide-loaded HLA-C or HLA-E on S2 cells provided partial to very weak inhibition of degranulation. This difference in the inhibitory capacity of human 221 cells versus insect S2 cells was not due simply to the strength of activation signals. Inhibition of the degranulation triggered by 221 cells was not overcome by additional ADCC signals through CD16. Likewise, inhibition of the polarization induced by S2 cells was not overcome by CD16 signals. These data are consistent with the view that activation-inhibition signals are not simply in balance and that the outcome depends on an integration of signals, in which inhibitory signals dominate (1). It is possible that HLA class I expressed on mammalian cells provides stronger inhibition of NK cells than the inhibition conferred by peptide-loaded HLA-C and HLA-E on S2 cells. However, peptide-loaded HLA-C and HLA-E on S2 cells were sufficient to induce strong inhibition of granule polarization.

Factors that may contribute to the formation of a strong inhibitory NK cell immunological synapse have not been defined yet. The complete inhibition of degranulation induced by HLA class I on 221 cells, but not on S2 cells, suggests that other receptor–ligand interactions between NK cells and 221 cells may contribute to inhibition. For example, inhibition of degranulation may require the additive effect of other ITIM-containing receptors, such as leukocyte-associated Ig-like receptor-1 and NKR-P1A. The human cell line 221 expresses LLT1 (Y. Bryceson, unpublished data), which is a ligand of NKR-P1A. We have found that expression of LLT1 on S2 cells is sufficient to block the granule polarization induced by ICAM-1. The Drosophila S2 cell system is well suited to determine the minimal requirements for inhibition, just as it has been to define requirements for activation of NK cells (7). It is worth noting that inhibitory KIR promotes clustering of activation receptors CD2 and 2B4 at inhibitory NK cell immunological synapses (23) while simultaneously blocking receptor phosphorylation (24). These results suggested that clustering of signaling-incompetent activation receptors at the synapse may serve to strengthen inhibitory signaling. However, coexpression of CD48 with ICAM-1 and HLA-C on S2 cells coated with rabbit IgG did not improve the inhibition of degranulation obtained with ICAM-1 and HLA-C (data not shown). Our data clearly show that granule polarization is more sensitive to inhibition than degranulation is, as it was completely inhibited by expression of either LLT1 or peptide-loaded HLA class I on S2 cells.

As shown with resting NK cells, CD16-induced degranulation can occur in the absence of detectable granule polarization (7). Likewise, we show here that GrzB release can persist during inhibition of granule polarization. Such unpolarized degranulation may involve fusion of lytic granules that are predocked at the plasma membrane. Even in the absence of stimulation, lytic granules in primary, resting NK cells can be observed at the plasma membrane by microscopy with live cells (25). The spontaneous release of granzymes by IL-2–activated NK cells may be another example of unpolarized release. This property of NK cells provides an explanation for earlier observations of the lysis of ICAM-1–expressing S2 cells by IL-2–activated NK cells (10), because the polarization signal transmitted by the ICAM-1–LFA-1 interaction is sufficient to orient the basal granzyme release toward the S2 cells. It is also consistent with the inhibition of lysis of S2–ICAM-1–HLA-C cells by HLA-C–specific KIR+ NK clones (10), because inhibition of polarization alone would be sufficient to protect the target cells. The inhibition of polarization shown here in cells that are still degranulating is a direct demonstration that polarization is a preferred target of inhibitory signaling.

Inhibition of degranulation induced by S2 cells was recently examined with resting NK cells (11). Degranulation, as measured by cell surface LAMP-1 (CD107a), induced by CD16 was partially inhibited by CD94-NKG2A binding to peptide-loaded HLA-E. Stronger inhibition was observed when degranulation was induced by coexpression of ULBP1 and CD48 on S2 cells, which leads to synergistic NK cell activation through receptors NKG2D and 2B4. Inhibition of polarization in resting NK cells could not be investigated, as cloning of NK cells is required to isolate cells with defined inhibitory receptor expression.

The signals that control lytic granule polarization in NK cells are not well known, although specific components of the pathway have been identified (26–28). Several receptors in NK cells can contribute to cytolytic granule polarization and cytotoxicity (10, 29, 30). Using insect cells transfected with individual ligands of NK cell receptors, we have shown that engagement of LFA-1 by ICAM-1 alone is sufficient to induce granule polarization (7, 10). However, in mouse NK cells, ICAM-1 on beads induces actin cytoskeleton remodeling, and coligation with activation receptor NKG2D is required to induce granule polarization (30). We report here that expression of ULBP1 on S2 cells is sufficient to induce granule polarization in IL-2–activated human NK cells. Therefore, neither LFA-1 nor NKG2D is necessary, but each one is sufficient to induce polarization of granules in human NK cells, and in both cases polarization is sensitive to inhibition. It will be interesting to investigate how signaling by these two very different receptors converge to induce granule polarization.

The main conclusion is that inhibitory receptors are better equipped to stop granule polarization than to block GrzB and chemokine release. This was shown definitively by observing sustained GrzB and MIP-1α release by NK cells in which granule polarization was inhibited. The evidence obtained for inhibition of NK cells by HLA class I on Drosophila S2 cells is conclusive, because inhibition is dependent on addition of HLA class I-specific peptides. Persistent degranulation during inhibition of polarization in IL-2–activated NK cells suggests that it may occur in high inflammatory conditions. Prevention of NK cell cytotoxicity would be better achieved through inhibition of degranulation rather than polarization. However, the possibility of releasing the block in degranulation while maintaining inhibition of polarization endows NK cells with the potential to provide bystander killing while still refraining from direct attacks on MHC class I-positive cells.

Acknowledgments

We thank Y. Bryceson, D. Geraghty, C. Gross, M. March, and S.Y. Yang for reagents and advice.

Disclosures The authors have no financial conflicts of interest.

Footnotes

  • This work was supported by the Intramural Research Program of the National Institute of Allergy and Infectious Diseases, National Institutes of Health.

  • Abbreviations used in this paper:

    GrzB
    granzyme B
    KIR
    killer-cell Ig-like receptor
    LLT1
    lectin-like transcript 1
    n
    number
    RIgG
    rabbit anti-S2 serum
    SHP-1
    SH2-containing protein tyrosine phosphatase-1.

  • Received April 15, 2010.
  • Accepted August 11, 2010.

References

  1. ↵
    1. Long E. O.
    2008. Negative signaling by inhibitory receptors: the NK cell paradigm. Immunol. Rev. 224: 70–84.
    OpenUrlCrossRefPubMed
  2. ↵
    1. Stebbins C. C.,
    2. C. Watzl,
    3. D. D. Billadeau,
    4. P. J. Leibson,
    5. D. N. Burshtyn,
    6. E. O. Long
    . 2003. Vav1 dephosphorylation by the tyrosine phosphatase SHP-1 as a mechanism for inhibition of cellular cytotoxicity. Mol. Cell. Biol. 23: 6291–6299.
    OpenUrlAbstract/FREE Full Text
  3. ↵
    1. Tybulewicz V. L.
    2005. Vav-family proteins in T-cell signalling. Curr. Opin. Immunol. 17: 267–274.
    OpenUrlCrossRefPubMed
  4. ↵
    1. Graham D. B.,
    2. M. Cella,
    3. E. Giurisato,
    4. K. Fujikawa,
    5. A. V. Miletic,
    6. T. Kloeppel,
    7. K. Brim,
    8. T. Takai,
    9. A. S. Shaw,
    10. M. Colonna,
    11. W. Swat
    . 2006. Vav1 controls DAP10-mediated natural cytotoxicity by regulating actin and microtubule dynamics. J. Immunol. 177: 2349–2355.
    OpenUrlAbstract/FREE Full Text
  5. ↵
    1. Wei S.,
    2. A. M. Gamero,
    3. J. H. Liu,
    4. A. A. Daulton,
    5. N. I. Valkov,
    6. J. A. Trapani,
    7. A. C. Larner,
    8. M. J. Weber,
    9. J. Y. Djeu
    . 1998. Control of lytic function by mitogen-activated protein kinase/extracellular regulatory kinase 2 (ERK2) in a human natural killer cell line: identification of perforin and granzyme B mobilization by functional ERK2. J. Exp. Med. 187: 1753–1765.
    OpenUrlAbstract/FREE Full Text
  6. ↵
    1. Trambas C. M.,
    2. G. M. Griffiths
    . 2003. Delivering the kiss of death. Nat. Immunol. 4: 399–403.
    OpenUrlCrossRefPubMed
  7. ↵
    1. Bryceson Y. T.,
    2. M. E. March,
    3. D. F. Barber,
    4. H. G. Ljunggren,
    5. E. O. Long
    . 2005. Cytolytic granule polarization and degranulation controlled by different receptors in resting NK cells. J. Exp. Med. 202: 1001–1012.
    OpenUrlAbstract/FREE Full Text
  8. ↵
    1. Vyas Y. M.,
    2. K. M. Mehta,
    3. M. Morgan,
    4. H. Maniar,
    5. L. Butros,
    6. S. Jung,
    7. J. K. Burkhardt,
    8. B. Dupont
    . 2001. Spatial organization of signal transduction molecules in the NK cell immune synapses during MHC class I-regulated noncytolytic and cytolytic interactions. J. Immunol. 167: 4358–4367.
    OpenUrlAbstract/FREE Full Text
  9. ↵
    1. Barber D. F.,
    2. E. O. Long
    . 2003. Coexpression of CD58 or CD48 with intercellular adhesion molecule 1 on target cells enhances adhesion of resting NK cells. J. Immunol. 170: 294–299.
    OpenUrlAbstract/FREE Full Text
  10. ↵
    1. Barber D. F.,
    2. M. Faure,
    3. E. O. Long
    . 2004. LFA-1 contributes an early signal for NK cell cytotoxicity. J. Immunol. 173: 3653–3659.
    OpenUrlAbstract/FREE Full Text
  11. ↵
    1. Bryceson Y. T.,
    2. H. G. Ljunggren,
    3. E. O. Long
    . 2009. Minimal requirement for induction of natural cytotoxicity and intersection of activation signals by inhibitory receptors. Blood 114: 2657–2666.
    OpenUrlAbstract/FREE Full Text
  12. ↵
    1. Malnati M. S.,
    2. P. Lusso,
    3. E. Ciccone,
    4. A. Moretta,
    5. L. Moretta,
    6. E. O. Long
    . 1993. Recognition of virus-infected cells by natural killer cell clones is controlled by polymorphic target cell elements. J. Exp. Med. 178: 961–969.
    OpenUrlAbstract/FREE Full Text
  13. ↵
    1. Shimizu Y.,
    2. D. E. Geraghty,
    3. B. H. Koller,
    4. H. T. Orr,
    5. R. DeMars
    . 1988. Transfer and expression of three cloned human non-HLA-A,B,C class I major histocompatibility complex genes in mutant lymphoblastoid cells. Proc. Natl. Acad. Sci. USA 85: 227–231.
    OpenUrlAbstract/FREE Full Text
  14. ↵
    1. Lee N.,
    2. D. R. Goodlett,
    3. A. Ishitani,
    4. H. Marquardt,
    5. D. E. Geraghty
    . 1998. HLA-E surface expression depends on binding of TAP-dependent peptides derived from certain HLA class I signal sequences. J. Immunol. 160: 4951–4960.
    OpenUrlAbstract/FREE Full Text
  15. ↵
    1. Kim H. S.,
    2. A. Das,
    3. C. C. Gross,
    4. Y. T. Bryceson,
    5. E. O. Long
    . 2010. Synergistic signals for natural cytotoxicity are required to overcome inhibition by c-Cbl ubiquitin ligase. Immunity 32: 175–186.
    OpenUrlCrossRefPubMed
  16. ↵
    1. Gross C.,
    2. W. Koelch,
    3. A. DeMaio,
    4. N. Arispe,
    5. G. Multhoff
    . 2003. Cell surface-bound heat shock protein 70 (Hsp70) mediates perforin-independent apoptosis by specific binding and uptake of granzyme B. J. Biol. Chem. 278: 41173–41181.
    OpenUrlAbstract/FREE Full Text
  17. ↵
    1. Winter C. C.,
    2. J. E. Gumperz,
    3. P. Parham,
    4. E. O. Long,
    5. N. Wagtmann
    . 1998. Direct binding and functional transfer of NK cell inhibitory receptors reveal novel patterns of HLA-C allotype recognition. J. Immunol. 161: 571–577.
    OpenUrlAbstract/FREE Full Text
  18. ↵
    1. Moesta A. K.,
    2. P. J. Norman,
    3. M. Yawata,
    4. N. Yawata,
    5. M. Gleimer,
    6. P. Parham
    . 2008. Synergistic polymorphism at two positions distal to the ligand-binding site makes KIR2DL2 a stronger receptor for HLA-C than KIR2DL3. J. Immunol. 180: 3969–3979.
    OpenUrlAbstract/FREE Full Text
  19. ↵
    1. Jackson M. R.,
    2. E. S. Song,
    3. Y. Yang,
    4. P. A. Peterson
    . 1992. Empty and peptide-containing conformers of class I major histocompatibility complex molecules expressed in Drosophila melanogaster cells. Proc. Natl. Acad. Sci. USA 89: 12117–12121.
    OpenUrlAbstract/FREE Full Text
  20. ↵
    1. Rosen D. B.,
    2. J. Bettadapura,
    3. M. Alsharifi,
    4. P. A. Mathew,
    5. H. S. Warren,
    6. L. L. Lanier
    . 2005. Cutting edge: lectin-like transcript-1 is a ligand for the inhibitory human NKR-P1A receptor. J. Immunol. 175: 7796–7799.
    OpenUrlAbstract/FREE Full Text
  21. ↵
    1. Aldemir H.,
    2. V. Prod’homme,
    3. M. J. Dumaurier,
    4. C. Retiere,
    5. G. Poupon,
    6. J. Cazareth,
    7. F. Bihl,
    8. V. M. Braud
    . 2005. Cutting edge: lectin-like transcript 1 is a ligand for the CD161 receptor. J. Immunol. 175: 7791–7795.
    OpenUrlAbstract/FREE Full Text
  22. ↵
    1. Valiante N. M.,
    2. M. Uhrberg,
    3. H. G. Shilling,
    4. K. Lienert-Weidenbach,
    5. K. L. Arnett,
    6. A. D’Andrea,
    7. J. H. Phillips,
    8. L. L. Lanier,
    9. P. Parham
    . 1997. Functionally and structurally distinct NK cell receptor repertoires in the peripheral blood of two human donors. Immunity 7: 739–751.
    OpenUrlCrossRefPubMed
  23. ↵
    1. Schleinitz N.,
    2. M. E. March,
    3. E. O. Long
    . 2008. Recruitment of activation receptors at inhibitory NK cell immune synapses. PLoS ONE 3: e3278.
    OpenUrlCrossRefPubMed
  24. ↵
    1. Watzl C.,
    2. C. C. Stebbins,
    3. E. O. Long
    . 2000. NK cell inhibitory receptors prevent tyrosine phosphorylation of the activation receptor 2B4 (CD244). J. Immunol. 165: 3545–3548.
    OpenUrlAbstract/FREE Full Text
  25. ↵
    1. Liu D.,
    2. Y. T. Bryceson,
    3. T. Meckel,
    4. G. Vasiliver-Shamis,
    5. M. L. Dustin,
    6. E. O. Long
    . 2009. Integrin-dependent organization and bidirectional vesicular traffic at cytotoxic immune synapses. Immunity 31: 99–109.
    OpenUrlCrossRefPubMed
  26. ↵
    1. Jiang K.,
    2. B. Zhong,
    3. D. L. Gilvary,
    4. B. C. Corliss,
    5. E. Hong-Geller,
    6. S. Wei,
    7. J. Y. Djeu
    . 2000. Pivotal role of phosphoinositide-3 kinase in regulation of cytotoxicity in natural killer cells. Nat. Immunol. 1: 419–425.
    OpenUrlCrossRefPubMed
    1. Banerjee P. P.,
    2. R. Pandey,
    3. R. Zheng,
    4. M. M. Suhoski,
    5. L. Monaco-Shawver,
    6. J. S. Orange
    . 2007. Cdc42-interacting protein-4 functionally links actin and microtubule networks at the cytolytic NK cell immunological synapse. J. Exp. Med. 204: 2305–2320.
    OpenUrlAbstract/FREE Full Text
  28. ↵
    1. Orange J. S.
    2008. Formation and function of the lytic NK-cell immunological synapse. Nat. Rev. Immunol. 8: 713–725.
    OpenUrlCrossRefPubMed
  29. ↵
    1. Chen X.,
    2. P. P. Trivedi,
    3. B. Ge,
    4. K. Krzewski,
    5. J. L. Strominger
    . 2007. Many NK cell receptors activate ERK2 and JNK1 to trigger microtubule organizing center and granule polarization and cytotoxicity. Proc. Natl. Acad. Sci. USA 104: 6329–6334.
    OpenUrlAbstract/FREE Full Text
  30. ↵
    1. Mace E. M.,
    2. S. J. Monkley,
    3. D. R. Critchley,
    4. F. Takei
    . 2009. A dual role for talin in NK cell cytotoxicity: activation of LFA-1-mediated cell adhesion and polarization of NK cells. J. Immunol. 182: 948–956.
    OpenUrlAbstract/FREE Full Text
Back to top

In this issue

Print
Download PDF
Article Alerts
Email Article
Citation Tools

Jump to section