NK lytic-associated molecule, involved in NK cytotoxic function, is an E3 ligase

NK lytic-associated molecule (NKLAM) is a protein involved in the cytolytic function of NK cells and CTLs. It has been localized to the cytolytic granules in NK cells and is up-regulated when cells are exposed to cytokines IL-2 or IFN-beta. We report in this study that NKLAM contains a cysteine-rich really interesting new gene (RING) in between RING-RING domain, and that this domain possesses strong homology to the RING domain of the known E3 ubiquitin ligase, Dorfin. To determine whether NKLAM functions as an E3 ligase, we performed coimmunoprecipitation binding assays with ubiquitin conjugates (Ubcs) UbcH7, UbcH8, and UbcH10. We demonstrated that both UbcH7 and UbcH8 bind to full-length NKLAM. We then performed a similar binding assay using endogenous NKLAM and UbcH8 expressed by human NK clone NK3.3 to show that the protein interaction occurs in vivo. Using the yeast two-hybrid system, we identified uridine kinase like-1 (URKL-1) protein as a substrate for NKLAM. We confirmed that NKLAM and URKL-1 interact in mammalian cells by using both immunoprecipitation and confocal microscopy. We demonstrated decreased protein expression and enhanced ubiquitination of URKL-1 in the presence of NKLAM. These data indicate that NKLAM is a RING finger protein that binds Ubcs and has as one of its substrates, URKL-1, thus defining this cytolytic protein as an E3 ubiquitin ligase.     

Downregulation of uridine-cytidine kinase like-1 decreases proliferation and enhances tumor susceptibility to lysis by apoptotic agents and natural killer cells

Natural killer (NK) cells target and kill tumor cells by direct anti-tumor cytotoxicity. NK lytic-associated molecule (NKLAM) is a protein involved in this cytolytic function. Acting as an E3 ubiquitin ligase, NKLAM binds to and ubiquitinates a novel protein, uridine-cytidine kinase like-1 (UCKL-1), targeting it for degradation. However, UCKL-1’s function in tumor cell survival and NK cell cytotoxicity is unknown. UCKL-1’s homology to uridine kinases and over expression in tumor cells suggests a role for UCKL-1 in tumor growth and/or survival. We propose that NKLAM and UCKL-1 interact in the tumor cell, where degradation of UCKL-1 leads to increased tumor cell apoptosis. Here we use RNA interference to downregulate UCKL-1 expression in K562 erythroleukemia cells. It was seen that downregulation of UCKL-1 initiated apoptosis and slowed the cell cycle, resulting in lower growth in the small interfering UCKL-1 RNA treated K562 cell culture. In addition, the chemotherapeutic agent staurosporine was seen to be more effective in inducing cell death by apoptosis in UCKL-1 depleted K562 cells compared with controls. We also found that UCKL-1 depleted K562 cells were more susceptible to NK mediated cytolysis than controls. These results indicate a role for UCKL-1 in tumor cell survival and suggest possible therapeutic potential of UCKL-1 inhibitors in cancer treatment.     

Chemotactic effect of human recombinant interleukin 2 on mouse activated large granular lymphocytes

Human recombinant interleukin 2 (hrIL-2) was demonstrated in vitro to be chemotactic for mouse large granular lymphocytes (LGL) activated in vivo by virus infection. Peritoneal exudate cells harvested from virus-infected mice were used as a source of LGL. LGL collected from mouse hepatitis virus-infected mice at 3 days postinfection were a source for NK 1.1 positive natural killer (NK)/LGL. LGL collected from mice treated with antiserum to gangliotetraosylceramide and infected with lymphocytic choriomeningitis virus for 7 days were used as a source for Lyt-2 positive cytotoxic T lymphocytes (CTL)/LGL. Both NK/LGL and CTL/LGL responded chemotactically to hrIL-2, purified IFN-beta, and to crude cell-free washout fluids collected from the peritoneal cavity of virus-infected mice. hrIL-2 had chemotactic activity for virus-elicited granular and agranular lymphocytes but did not attract the contaminating macrophages, in contrast to IFN-beta, which displayed chemotactic activity for virus-elicited granular and agranular lymphocytes as well as macrophages. The migration to hrIL-2 was inhibited by a monoclonal antibody (7D4) to the IL-2 receptor, but treatment with 7D4 did not affect migration in response to IFN-beta. Microscopic examination of Wright's-Giemsa-stained migrated NK/LGL and CTL/LGL revealed that the majority of migrated LGL in either LGL population had a blast cell morphology (enlarged cells with rich basophilic cytoplasm). The frequency of cells bearing the LGL morphology within the virus-elicited nonadherent peritoneal exudate cell population was on incubation in vitro, stabilized by either hrIL-2 or IFN-beta. These data suggest that another important immunomodulating function of IL-2 may be to attract activated NK/LGL and CTL/LGL to sites of inflammation.     

Induction of alloreactive cytotoxic T cells by acute virus infection of mice

Alloreactive cytotoxic T lymphocytes (CTL) distinct from virus-specific CTL and activated natural killer (NK) cells were generated during acute lymphocytic choriomeningitis virus (LCMV) infection of C57BL/6J mice. The alloreactive CTL shared similar antigenic markers (Thy-1.2+, Lyt-2.2+, and asialo GM1-) with the virus-specific CTL that appeared at the same time 7 days postinfection, but had different target specificities. These alloreactive CTL lysed allogeneic but not syngeneic or xenogeneic targets. These were distinct from activated NK cells, which lysed all target cell types, peaked 3 days postinfection, and had a phenotype of asialo GM1+, Thy-1 +/-, Lyt-2.2-. Cold target competition studies indicated that there were several subsets of alloreactive T cells with distinct specificities, and that these alloreactive T cells were not subsets of the virus-specific T cells. Similar types of alloreactive CTL were induced at much lower levels in C3H/St mice. This may indicate that the generation of this "aberrant" T cell activity is under genetic control. Hence, the LCMV infection of C57BL/6J mice induces several cytotoxic effector populations including alloreactive CTL, activated NK cells, and virus-specific CTL. Virus infections therefore have the ability not only to polyclonally stimulate B cells, as previously described, but also to stimulate CTL.     

Regulation of the expression on mouse T lymphocytes of the epitope identified by monoclonal antibody 3A35

Monoclonal antibody 3A35 (MA 3A35) has previously been shown to be an activation marker of macrophages and T lymphocytes. It immunoprecipitated from macrophages a 200-kDa molecule belonging to the T200 family and from T cells a 85-kDa antigen. In the present work, the factors controlling the expression of the epitope identified by MA 3A35 on polyclonal activated T cells and T-cell clones, as well as the ability of 3A35 alone or together with complement to interfere with T-cell functions, were investigated. Corticoresistant thymocytes unreactive with MA 3A35 became fully reactive after 2 days of in vitro stimulation by PMA and IL-2 and the level of reactivity per cell declined to a low level thereafter. In helper and cytolytic T-cell clones, the expression of the epitope defined by MA 3A35 was also maximal soon after antigenic stimulation then declined. In helper-T-cell clones, the epitope remained detectable during the entire culture period, whereas in cytolytic clones its expression was markedly reduced at the end of the culture. The lineage of cytotoxic T lymphocytes (CTL) as studied in a bulk culture of spleen cells primed in vivo against a syngeneic tumor exhibited similar regulation by antigenic stimulation. The CTL precursors were resistant to lysis by MA 3A35 plus complement; after 3 days of culture with the stimulatory antigen, they became highly sensitive but their sensitivity then diminished and mature CTL were completely resistant. MA 3A35 plus complement also killed the activated T cells which responded to macrophage-presented antigens and were thought to be mainly Lyt-1+. Therefore, the epitope identified by MA 3A35 was expressed predominantly at an early stage of T-cell activation. At a late stage, it persisted almost exclusively on helper and Lyt-1+ cells. In addition, MA 3A35 plus complement lysed NK cells, AK cells, and their precursors present in normal spleen. In the absence of complement, MA 3A35 had no detectable effect on T-cell functions.     

A homogeneous population of lymphokine-activated killer (LAK) cells is incapable of killing virus-, bacteria-, or parasite-infected macrophages

Previous reports have suggested a role for natural killer (NK) cells in directly lysing host cells infected with bacteria and other intracellular microorganisms. Here, we determined the inability of a highly homogeneous population of lymphokine activated killer (LAK) cells to kill macrophages infected with the following intracellular parasites: Mycobacterium avium, Listeria monocytogenes, Legionella pneumophila, Toxoplasma gondii, and Trypanosoma cruzi. In parallel cytotoxicity assays, LAK cells lysed the tumor targets YAC-1 and P815 effectively. Furthermore, we were able to demonstrate that influenza-specific cytotoxic T lymphocytes (CTL), but not LAK cells, were efficient killers of influenza virus-infected macrophages.     

NK cells play a critical protective role in host defense against acute extracellular Staphylococcus aureus bacterial infection in the lung

Staphylococcus aureus remains a common cause of nosocomial bacterial infections and are often antibiotic resistant. The role of NK cells and IL-15 and their relationship in host defense against extracellular bacterial pathogens including S. aureus remain unclear. We have undertaken several approaches to address this issue using wild type (WT), IL-15 gene knock-out (KO), and NK cell-depleted mouse models. Upon pulmonary staphylococcal infection WT mice had markedly increased activated NK cells, but not NKT or gammadelta T cells, in the airway lumen that correlated with IL-15 production in the airway and with alveolar macrophages. In vitro exposure to staphylococcal products and/or coculture with lung macrophages directly activated NK cells. In contrast, lung macrophages better phagocytosed S. aureus in the presence of NK cells. In sharp contrast to WT controls, IL-15 KO mice deficient in NK cells were found to be highly susceptible to pulmonary staphylococcal infection despite markedly increased neutrophils and macrophages in the lung. In further support of these findings, WT mice depleted of NK cells were similarly susceptible to staphylococcal infection while they remained fully capable of IL-15 production in the lung at levels similar to those of NK-competent WT hosts. Our study thus identifies a critical role for NK cells in host defense against pulmonary extracellular bacterial infection and suggests that IL-15 is involved in this process via its indispensable effect on NK cells, but not other innate cells. These findings hold implication for the development of therapeutics in treating antibiotic-resistant S. aureus infection.     

Cutting edge: IFN-gamma signaling to macrophages is required for optimal Valpha14i NK T/NK cell cross-talk

Activated NK T cells are known to rapidly stimulate NK cells and, subsequently, CD8(+) T cells and B cells. In this report, we first demonstrate that the downstream effects induced by alpha-galactosylceramide activated NK T cells on NK cells are mainly dependent on IFN-gamma. We found that NK T cell activation of NK cells requires a functional IFN-gamma signaling in macrophages and dendritic cells but not in B cells, NK cells, or NK T cells. NK T cell activation is dendritic cell-dependent whereas NK T cell activation of NK cells is indirect and in part mediated by macrophages. Interestingly, in this context, macrophage participation in the CD1d Ag presentation of alpha-galactosylceramide to NK T cells is not necessary. These data indicate that NK T cell-dependent activation of macrophages is required for optimal NK T cell-induced stimulation of NK cells.     

Enhanced lytic susceptibility of Ha-ras transformants after oncogene induction is specific to activated NK cells

C3H 10T1/2 mouse fibroblasts were transfected with a plasmid vector composed of EJ, the mutated c-Ha-ras, and a metallothionein promotor that induced amplified ras expression when activated by culture in the presence of zinc. Experiments were conducted to compare the effect of induction on killing by activated natural killer (NK) cells, cytotoxic T lymphocytes, activated macrophages, and antibody plus complement. The only effector that recognized increased ras expression and exhibited high-inducible cytolysis was an activated NK cell. The effectors from spleen were poly I.C. boostable, Lyt-1.1 negative, NK 1.2 positive, and asialo GM1 positive. Spleen cells from T cell-deficient nude mice, but not NK-deficient beige mice, exhibited high levels of killing activity, and experiments with NK cell clones demonstrated that these lines were also highly cytolytic and killed Ha-ras transfectants in parallel to YAC. Transfection of the same fibroblast line with c-myc did not alter the level of activated NK sensitivity. Cold target competition experiments revealed that Ha-ras-transfected and non-transfected 10T1/2 fibroblasts competed equally for lysis of either YAC or Ha-ras transfectants. Rat-1 fibroblasts did not compete, but gained this capacity when transformed with the v-Ki-ras oncogene but not v-fps. These data suggest that Ha-ras acts in target cells at a post-binding step, whereas Ki-ras may affect expression of target-effector binding structures. The findings that activated NK cell lysis may be specifically influenced by ras expression support a role for NK cells in host surveillance against early neoplastic changes.     

CD11b (Mac-1): a marker for CD8+ cytotoxic T cell activation and memory in virus infection.

We have found that CD11b, a cell surface integrin of macrophages, granulocytes, and NK cells, is expressed by a subset of CD8+ T cells that include both the active virus-specific CTL and the virus-specific memory CTL populations. CD8+CD11b+ cells comprise less than 3% of naive mouse splenocytes, but after lymphocytic choriomeningitis virus (LCMV) infection increase by 9- to 12-fold by the peak (day 8) of the virus-specific CTL response. Depletion of day-8 splenocytes with anti-Mac-1 and C' or enrichment by sorting for CD11b+ or CD8+CD11b+ spleen cells demonstrated that LCMV-specific CTL are CD11b+. The CD11b+ subpopulation also contained the bulk of the IL-2-responsive CD8+ cells. MEL-14, a homing marker down-regulated on activated T cells, was down-regulated on the majority of CD8+ cells that became CD11b+. Less than 1% of LCMV-immune splenic lymphocytes expressed CD11b. Antibody and C' depletion of this population severely impaired the ability of immune splenocytes to respond to in vitro secondary stimulation with LCMV-infected peritoneal macrophages, but did not affect the generation of a primary allospecific CTL response in MLC. Mixing of CD8-depleted and CD11b-depleted LCMV-immune splenocytes failed to restore the ability of these cells to mount a virus-specific memory CTL response, indicating that a cell coexpressing CD8 and CD11b is essential for this response. As determined by limiting dilution analysis, the precursors for the LCMV-specific memory CTL response were enriched in the CD11b+ population of LCMV-immune splenocytes. CD11b stained far fewer CD8+ splenocytes from naive mice than did CD44 (Pgp-1), and among immune splenocytes it identified a small subpopulation of CD44hi cells, indicating that CD11b may be the best single marker available for discriminating between naive and memory CD8+ T cells.     

Alteration of the metastatic potential of line 1 lung carcinoma cells: opposite effects of class I antigen induction by interferons versus DMSO or gene transfection.

Class I antigens are necessary for the recognition of tumor cells by cytotoxic T lymphocytes (CTL). The line 1 lung carcinoma is a spontaneous murine tumor deficient in class I antigen expression. Consistent with this, line 1 cells are highly metastatic in vivo. We investigated whether increasing class I antigen expression on line 1 cells could alter the metastatic potential of these tumor cells using an in vivo lung metastasis model. We used three methods to induce class I antigen expression on line 1 cells: gene transfection, treatment with dimethyl sulfoxide (DMSO), or treatment with interferon (IFN)-beta or -gamma. We found that line 1 cells expressing a transfected class I gene were significantly less metastatic than parental line 1 cells. DMSO-treated line 1 cells also formed significantly fewer metastases than parental line 1 cells. These results indicate that increased class I antigen expression decreases the metastatic potential of line 1 cells in vivo. However, we did not observe a significant decrease in the number of lung metastases in mice receiving line 1 cells treated with IFN-beta or -gamma, despite high levels of class I antigen expression. Thus, increasing class I antigen expression with IFN has an opposite effect on metastasis from class I antigen expression induced by transfection or DMSO. These results show that the method used to increase class I antigen expression is critical in terms of the in vivo effect observed. To investigate a possible mechanism for the differences observed in vivo between these class I expressing cells, we tested whether IFN alters or blocks susceptibility of line 1 cells to immune effector cells. We found IFN treatment increased the ability of line 1 cells to be recognized by CTL but concomitantly decreased the susceptibility of line 1 cells to NK cell lysis by a non-class I antigen-related mechanism. In contrast, transfected or DMSO-treated line 1 cells which were less metastatic in vivo were susceptible to both CTL and NK-mediated lysis. Taken together, these results suggest that immune intervention against metastasizing line 1 cells may involve NK cells and CTL.     

The expression,regulation and adhesion function of a novel CD molecule,CD226, on human endothelial cells

CD226 is a 67 kDa type I transmembrane glycoprotein mainly expressed on activated T cells, NK cells and platelets, and involved in the differentiation of cytotoxic T lymphocytes (CTL) and NK, as well as platelet activation and aggregation. Here we found that the expression of CD226 protein and CD226mRNA were very weak in resting HUVEC and ECV304 cells, whereas high level expression could be observed when these cells were stimulated. The binding activities between activated endothelial cells and activated Jurkat cells could be partly blocked by CD226/Ig fusion protein. Similarly, CD226/Ig could also partly block the adhesion between activated endothelial cells and some leukocytes or colo205 cells. These data provided the evidence that activated endothelial cells could express high level of CD226, and CD226 was involved in the endothelial cells' adhesion. The above findings suggested that CD226 is a novel inducible adhesion molecule on human endothelial cells.     

Perioperative IFN-alpha to avoid surgically induced immune suppression in colorectal cancer patients

Surgical treatment of colorectal cancer is associated with postoperative immunosuppression, which might facilitate dissemination of tumor cells and outgrowth of minimal residual disease/(micro) metastases. Minimal residual disease has been shown to be of prognostic relevance in colorectal cancer. Therefore, stimulation of (anti-tumor) immune responses may be beneficial in the prevention of metastases formation. Important anti-tumor effector cells, which serve this function, are natural killer (NK) cells, CD8+ lymphocytes (CTL), dendritic cells (DC) and macrophages. In this review the immunomodulating properties of IFN-alpha are discussed, with a particular focus on perioperative stimulation of immune function in cancer patients. IFN-alpha is known to enhance innate immune functions such as stimulation of NK cells, transition from innate to adaptive responses (activation of DC) and regulating of CD8+ CTL activity and memory. Moreover, it exerts direct antitumor effects by regulating apoptosis and cell cycle. In several clinical trials, perioperative administration of IFN-alpha has indeed been shown to improve T cell responsiveness, prevent impairment of NK cell cytotoxicity and increase expression of activation markers on NK, T and NKT cells. In a clinical pilot study we showed in colorectal cancer patients that received perioperative IFN-alpha enhanced activation markers on T cells and NK cells, combined with better-preserved T cell function as indicated by phytohemaggluttinin skin tests. In the liver of these patients significantly more CD8+ T cells were found. In conclusion, IFN-alpha provides an effective adjuvant in several forms of cancer and improves several postoperative immune functions in perioperative administration. However, larger clinical trials are necessary to investigate effects on disease-free and overall survival.     

An early event in murine cytomegalovirus replication inhibits presentation of cellular antigens to cytotoxic T lymphocytes.

Cytomegalovirus (CMV) infection of simian virus 40 (SV40)-immune mice inhibits priming of SV40-specific helper and cytotoxic T lymphocytes (CTL) in vivo (A. E. Campbell, J. S. Slater, and W. S. Futch, Virology 173:268-275, 1989; J. S. Slater, W. S. Futch, V. J. Cavanaugh and A. E. Campbell, Virology 185:132-139, 1991). We now demonstrate that murine CMV (MCMV) infection of SV40-transformed macrophages and fibroblasts prevents presentation of SV40 T antigen to SV40-specific CTL. MCMV-infected macrophages failed to stimulate SV40-immune CTL precursors in vitro. In addition, MCMV-infected, SV40-transformed macrophage and fibroblast target cells lost their susceptibility to lysis by major histocompatibility complex class I-restricted, SV40-specific CTL clones. MCMV infection did not alter the synthesis of SV40 T antigen in the target cells. MCMV early gene expression was required for inhibition of SV40 T-antigen presentation; immediate-early gene expression was insufficient for this effect. Early viral gene expression also resulted in significant reduction of H-2K and H-2D molecules on the surface of MCMV-infected fibroblasts. However, this reduction occurred independently from suppression of antigen presentation to CTL. The same target cells which were resistant to lysis by SV40 CTL were susceptible to lysis by MCMV-specific CTL. MCMV early gene products therefore interfere with the processing and/or presentation of SV40 T-antigen determinants to CTL independent of alterations in the major histocompatibility complex.     

Localization of Fas ligand in cytoplasmic granules of CD8+ cytotoxic T lymphocytes and natural killer cells: participation of Fas ligand in granule exocytosis model of cytotoxicity

Fas ligand (FasL) has been implicated in cytotoxic T lymphocyte (CTL)- and natural killer (NK) cell-mediated cytotoxicity. In the present study, we investigated the localization of FasL in murine CTL and NK cells. Immunocytochemical staining showed that FasL was stored in cytoplasmic granules of CD8+ CTL clones and in vivo activated CTL and NK cells, where perforin and granzyme A also resided. Immunoelectron microscopy revealed that FasL was localized on outer membrane of the cytoplasmic granules, while perforin was localized in internal vesicles. Western blot analysis showed that the membrane-type FasL of 40 kDa was stored in CD8+ CTL clones but not in CD4+ CTL clones. By utilizing a granule exocytosis inhibitor (TN16), we demonstrated that FasL translocated onto cell surface upon degranulation of anti-CD3-stimulated CD8+ CTL clones. Moreover, TN16 markedly inhibited the FasL-mediated cytotoxicity by CD8+ T cell clones and NK cells. These results suggested a substantial contribution of FasL to granule exocytosis-mediated target cell lysis by CD8+ CTL and NK cells.     

Role of tumor necrosis factor-alpha in E1A oncogene-induced susceptibility of neoplastic cells to lysis by natural killer cells and activated macrophages

NIH-3T3 cells transfected with adenovirus E1A oncogene cDNA were found to exhibit cytolytic susceptibility to murine NK cells and activated macrophages associated with a threshold level of oncogene product expression exceeding that required for morphological transformation. A similar correlation was observed between threshold levels of E1A gene product expression and target cell susceptibility to direct cytotoxicity by rTNF. Inhibition of splenic NK cell and peritoneal macrophage cytolysis by antisera specific for murine rTNF confirmed the importance of E1A-induced TNF susceptibility as one determinant of target cell cytolytic susceptibility. Anti-TNF antibody was, however, unable to block killing of E1A-expressing targets by the NK cell line, NKB61A2. These results suggest a direct link between the functions of E1A oncogene products and cellular mechanisms of action of TNF elaborated by host effector cells and indicate that E1A expression also affects target cell susceptibility to TNF-independent cytolytic mechanisms.     

Intrahepatic lymphocyte expression of dipeptidyl peptidase I-processed granzyme B and perforin induces hepatocyte expression of serine proteinase inhibitor 6 (Serpinb9/SPI-6)

Human proteinase inhibitor 9 (PI-9/serpinB9) and the murine ortholog, serine proteinase inhibitor 6 (SPI-6/serpinb9) are members of a family of intracellular serine proteinase inhibitors (serpins). PI-9 and SPI-6 expression in immune-privileged cells, APCs, and CTLs protects these cells against the actions of granzyme B, and when expressed in tumor cells or virally infected hepatocytes, confers resistance to killing by CTL and NK cells. The present studies were designed to assess the existence of any correlation between granzyme B activity in intrahepatic lymphocytes and induction of hepatic SPI-6 expression. To this end, SPI-6, PI-9, and serpinB9 homolog expression was examined in response to IFN-alpha treatment and during in vivo adenoviral infection of the liver. SPI-6 mRNA expression increased 10- to 100-fold in the liver after IFN-alpha stimulation and during the course of viral infection, whereas no significant up-regulation of SPI-8 and <5-fold increases in other PI-9/serpinB9 homolog mRNAs was observed. Increased SPI-6 gene expression during viral infection correlated with influxes of NK cells and CTL. Moreover, IFN-alpha-induced up-regulation of hepatocyte SPI-6 mRNA expression was not observed in NK cell-depleted mice. Additional experiments using genetically altered mice either deficient in perforin or unable to process or express granzyme B indicated that SPI-6 is selectively up-regulated in hepatocytes in response to infiltration of the liver by NK cells that express perforin and enzymatically active granzyme B.     

Macrophages help NK cells to attack tumor cells by stimulatory NKG2D ligand but protect themselves from NK killing by inhibitory ligand Qa-1

Natural killer (NK) cells and their crosstalk with other immune cells are important for innate immunity against tumor. To explore the role of the interaction between NK cells and macrophages in the regulation of anti-tumor activities of NK cells, we here demonstrate that poly I:C-treated macrophages increased NK cell-mediated cytotoxicity against target tumor cells in NKG2D-dependent manner. In addition, IL-15, IL-18, and IFN-β secreted by poly I:C-treated macrophages are also involved in NKG2D expression and NK cell activation. Interestingly, the increase in expression of NKG2D ligands on macrophages induced a highly NK cell-mediated cytotoxicity against tumor cells, but not against macrophages themselves. Notably, a high expression level of Qa-1, a NKG2A ligand, on macrophages may contribute to such protection of macrophages from NK cell-mediated killing. Furthermore, Qa-1 or NKG2A knockdown and Qa-1 antibody blockade caused the macrophages to be sensitive to NK cytolysis. These results suggested that macrophages may activate NK cells to attack tumor by NKG2D recognition whereas macrophages protect themselves from NK lysis via preferential expression of Qa-1.