Missing self by heterogeneous natural killer cells

Some murine (YAC, P815 and SP20) and human (Molt4, Raji and HR7) tumour cell lines were (i) treated with IFN-γ for inducing enhanced expression of MHC class I antigen, or (ii) given a brief treatment with citrate buffer (pH 3.0), which resulted in denaturation of class I MHC antigens on these tumour cells. IFL-γ or acid treated tumour cells were used as unlabelled competing targets in cold target inhibition assays. The results indicated that the competing ability of acid-treated tumour cells remained unaltered, whereas IFN-γ treated tumour cells competed with significantly less efficiency. These results have been evaluated in light of the current view of NK cell development and the expression of inhibitory receptors for MHC class I molecules (IRMs), on NK cells. A modified view on NK cell heterogeneity based upon IRM expression has been proposed which reconciles several apparently discordant observations about the activity and role of NK cells. Two classes of NK cells have been proposed. Type I NK ceils have target recognition receptors which do not recognize autologous normal cells, lack IRMs, and may participate in first line of defence against transformed cells in vivo. Type II NK cells have target recognition receptors for autologous normal cells and express at least one self-reactive IRM in order to prevent auto-killing. Type II NK cells participate in killing those transformed cells which down-regulate their MHC class I expression in order to escape cytotoxic T-cell surveillance. It is also postulated that mechanism of inverse correlation of target cell MHC class I expression levels and their susceptibility to NK cells, involves interference model of missing self hypothesis for type I NK cells and inhibitory signal model of missing self hypothesis for type II NE cells. Finally, it is proposed that acid treatment of tumour cells enhances their lysis susceptibility by making them additionally susceptible to type II NK cells, rather than enhancing their killing by type I NK cells. This proposition would explain the lack of effect of acid treatment on the competing ability of tumour cells, when target cells are only lysed by type I NE cells.     

Interferon-gamma upregulates the susceptibility of human neuroblastoma cells to interleukin-2-activated natural killer cells

The influence of interferon-gamma on the susceptibility of neuroblastoma cells in cell-mediated killing was investigated. Neuroblastoma cells were only weakly susceptible targets for peripheral mononuclear cells. However, enrichment of natural killer (NK) cells or activation of NK cells with interleukin-2 resulted in a considerable increase of neuroblastoma cell lysis. Pretreatment of neuroblastoma targets with interferon-gamma additionally increased the susceptibility to enriched NK cells as well as to interleukin-2-activated NK cells. The conjugate formation between enriched NK cells and the neuroblastoma targets was not affected by the pretreatment of the targets with interferon-gamma. Concomitantly, treatment of the neuroblastoma targets with interferon-gamma resulted in a strong induction of otherwise poorly expressed major histocompatibility complex (MHC) class I antigen expression. These results suggest that the increased expression of MHC class I antigens on target cells is not always correlated with decreased sensitivity for NK cells but can also be followed by an increased susceptibility for NK cells.     

A human NK cell activation/inhibition threshold allows small changes in the target cell surface phenotype to dramatically alter susceptibility to NK cells

NK cell activation is negatively regulated by the expression of target cell MHC class I molecules. We show that this relationship is nonlinear due to an NK cell activation/inhibition threshold. Ewing's sarcoma family tumor cell monolayers, which were highly susceptible to NK cells in vitro, developed a highly resistant phenotype when cultured as three-dimensional multicellular tumor spheroid structures. This suggested that tumor architecture is likely to influence the susceptibility to NK cells in vivo. Resistance of the multicellular tumor spheroid was associated with the increased expression of MHC class I molecules and greatly reduced NK cell activation, implying that a threshold of NK cell activation/inhibition had been crossed. Reducing MHC class I expression on Ewing's sarcoma family tumor monolayers did not alter their susceptibility to NK cells, whereas increased expression of MHC class I rendered them resistant and allowed the threshold point to be identified. This threshold, as defined by MHC class I expression, was predictive of the number of NK-resistant target cells within a population. A threshold permits modest changes in the target cell surface phenotype to profoundly alter the susceptibility to NK cells. Whereas this allows for the efficient detection of target cells, it also provides a route for pathogens and tumors to evade NK cell attack.     

Enhanced recognition of human NK receptors after influenza virus infection

The NK cell cytotoxic activity is regulated by both inhibitory and activating NK receptors. Thus, changes in the expression levels and in the affinity or avidity of those receptors will have a major effect on the killing of target cells. In this study, we demonstrate that the binding of NK-inhibitory receptors is enhanced after influenza virus infection. Surprisingly, however, no change in the level of class I MHC protein expression was observed on the surface of the infected cells. The increased binding was general, because it was observed in both the killer cell Ig-like receptor 2 domain long tail 1 and leukocyte Ig-like receptor-1. The increased binding was functional, was not dependent on the interaction with viral hemagglutinin-neuraminidase, was not dependent on the glycosylation site, and was not abolished after mutating the transmembrane or cytosolic portions of the class I MHC proteins. Confocal microscopy experiments showed increased binding of NK receptor-coated beads to infected cells expressing the appropriate class I MHC proteins. In addition, specific cell-free bead aggregates covered with class I MHC proteins were observed only in infected cells. We therefore suggest that the influenza virus use a novel mechanism for the inhibition of NK cell activity. This mechanism probably involves the generation of class I MHC complexes in infected cells that cause increased recognition of NK receptors.     

Selective MHC expression in tumours modulates adaptive and innate antitumour responses

Progress towards developing vaccines that can stimulate an immune response against growing tumours has involved the identification of the protein antigens associated with a given tumour type. Epitope mapping of tumour antigens for HLA class I- and class II-restricted binding motifs followed by immunization with these peptides has induced protective immunity in murine models against cancers expressing the antigen. MHC class I molecules presenting the appropriate peptides are necessary to provide the specific signals for recognition and killing by cytotoxic T cells (CTL). The principle mechanism of tumour escape is the loss, downregulation or alteration of HLA profiles that may render the target cell resistant to CTL lysis, even if the cell expresses the appropriate tumour antigen. In human tumours HLA loss may be as high as 50%, inferring that a reduction in protein levels might offer a survival advantage to the tumour cells. Alternatively, MHC loss may render tumour cells susceptible to natural killer cell-mediated lysis because they are known to act as ligands for killer inhibitory receptors (KIRs). We review the molecular features of MHC class I and class II antigens and discuss how surface MHC expression may be regulated upon cellular transformation. In addition, selective loss of MHC molecules may alter target tumour cell susceptibility to lymphocyte killing. The development of clinical immunotherapy will need to consider not only the expression of relevant CTL target MHC proteins, but also HLA inhibitory to NK and T cells. Received: 20 March 1999 / Accepted: 3 May 1999     

Enhanced MHC I antigen expression on tumour target cells is inversely correlated to lysis by allogenic but not by xenogenic NK cells

Relationship between the levels of MHC class 1 antigen expressed on tumour cells and their susceptibility to allogenic and xenogenic NK cells was investigated. Mouse and human natural killer-resistance inducing factor (NK-RIF) preparations were used for augmenting/inducing MHC 1 antigen expression on murine YAC and human K562 tumour cells, respectively YAC cells with augmented MHC I antigen expression became relatively resistant to lysis by murine NK cells but not to rat NK cells. Similarly, induction of MHC I antigens on K562 cells reduced their susceptibility to human NK cells but not to monkey NK cells. These results indicate that the inverse correlation of MHC I antigen expression and NK susceptibility does not hold true for xenogenic pairs of NK effector and target cells.     

Cell surface molecules involved in NK recognition by cloned cytotoxic T lymphocytes

Short-term treatment of cloned mouse cytotoxic T lymphocytes (CTL) with interferon (IFN) induces lytic activity for natural killer- (NK) sensitive targets. Extended culture of CTL in high concentrations of interleukin 2 induces promiscuous lytic activity in which state both NK-sensitive and NK-resistant target cells are lysed. Cold-target competition analysis showed that the development of NK activity was associated with the acquisition of binding activity for NK-sensitive but not for NK-resistant targets, whereas the development of promiscuous lytic activity was associated with the acquisition of binding activity for both types of target. Antigen-specific cytolysis was inhibited by antibodies to Ly-2, Ly-5, LFA-1 and to the V region of the T cell antigen receptor (TCR), whereas NK and promiscuous lytic activity in the same cells was resistant to inhibition by anti-Ly-2 and anti-TCR. NK activity was expressed normally against a variant NK-sensitive cell line lacking all MHC antigens. These results show that, in contrast to antigen-specific recognition, the NK and promiscuous lytic activities of CTL are expressed without participation of effector cell Ly-2 and TCR molecules or target cell MHC molecules, and are most likely mediated through novel and distinct receptor systems.     

Analysis of the mechanisms involved in NK resistance induced by a new tumor factor NK-RIF

The mechanisms involved in susceptibility or resistance of neoplasic cells to lysis by NK cells are not well known. We have recently described a 12-kDa factor (NK-RIF), produced and released by different tumor cell lines, making K562 resistant to NK lysis without affecting the cytotoxic function of NK effector cells. In this paper we further study the mechanism involved in NK resistance of K562 mediated by NK-RIF and its biological implications. The results show that NK-RIF does not affect the binding capacity of target and effector cells nor the levels of HLA class I antigen expression on the target cells, as a proof that resistance to NK-mediated lysis is not always associated with a defect in target effector binding or with an increased MHC class I antigen expression. However NK-RIF-treated K562 loses its capacity to induce NK cell activation and the subsequent capacity to release NKCF and makes K562 resistant to lysis by NKCF. Therefore our results show that induction of resistance to NK cytotoxicity can be the result of the modulation of target structures responsible for inducing effector cell activation without affecting target/effector binding molecules. This indicates that the structures involved in adherence and activation of NK cells have a different nature and that molecules other than HLA participate in NK resistance.     

Mechanisms involved in NK resistance induced by interferon-gamma.

Human tumor cell lines were treated with interferon-gamma (IFN-gamma) and then used as target cells in NK assays to measure their ability to form conjugates and stimulate the production of NK cytotoxic factors (NKCF) and to determine their susceptibility to NKCF lysis. K562 and cell lines RS1, RS3, RS7, CAC, and CAP2, obtained from solid brain tumors, were used as targets, and peripheral blood lymphocytes (PBL) from normal donors were used as effector cells. IFN-gamma-treated cell lines had a decreased susceptibility to NKCF lysis and a decreased ability to induce the release of these factors without affecting target-effector cell binding. These results were not due to changes in HLA class I antigen expression, given that the level of HLA class I antigens on the tumor cell lines was not affected, the only exception being K562. In an attempt to further clarify the possible influence of HLA class I expression on K562, IFN-gamma-pretreated K562 cells were separated into HLA class I positive and HLA class I negative subsets for the NK assays. The results showed that both populations behaved similarly upon target-effector conjugate formation, whereas the HLA class I positive population showed a reduced susceptibility to lysis by NK cells and NKCF. Thus, these results establish that NK resistance induced by IFN-gamma is mediated by blocking the target cell's ability to activate NK cell triggering and release of NKCF and by blocking its susceptibility to lysis by these factors. This analysis helps to clarify not only the NK process but also the controversial regulatory effect of IFN in NK lysis.     

Distinct effects of interferon-gamma and MHC class I surface antigen levels on resistance of the K562 tumor cell line to natural killer-mediated lysis

Various investigators have examined the relationship between tumor cell susceptibility to natural killer (NK) cell lysis and the expression of HLA class I antigens on the tumor cell. There is controversy as to whether or not an inverse relationship exists, and if so, the basis of the relationship between these two phenomena remains undefined. To address these questions, the genomic clones for two HLA antigens were transfected into the erythroleukemia cell line K562, a cell line that is used as the standard to assess human NK and major histocompatibility complex (MHC) nonrestricted cytolysis. Susceptibility to NK lysis was not affected by the de novo expression of HLA antigens on the K562 after DNA mediated gene transfer. Interferon-gamma (IFN-gamma) treatment of K562 induced levels of MHC class I antigen surface expression comparable to those found on the transfected cells; however, the IFN-gamma-treated cells were resistant to NK lysis. When very high levels of surface HLA antigens were induced on the transfectants, a potential effect of class I MHC expression on K562 lysis could be discerned that was distinct from the resistance to NK lysis induced by IFN-gamma-treatment.     

Target cell expression of MHC antigens is not (always) a turn-off signal to natural killer cells

Recent evidence has demonstrated that the lytic function of natural killer cells might be regulated by potential target cells through the target cells' expression of cell surface components that are able to inhibit the lytic process. Specifically, it has been shown in many target cell systems that the expression of class I MHC proteins by target cells is inversely proportional to their susceptibility to lysis by NK cells. It has been suggested, therefore, that MHC proteins may act as important negative regulatory elements in the ongoing control of NK cell function. Herein, we examined two closely related murine lymphoma cells (ASL1 and ASL1w), both in terms of their susceptibility to lysis by NK cells as well as their expression of both H-2K and H-2D class I MHC proteins. The results of these studies showed that whereas ASL1 and ASL1w cells differed greatly in their susceptibility to NK cell lysis (ASL1 was much more NK resistant than ASL1w), both expressed high levels of H-2K and D proteins. In contrast to what might have been predicted base on reports from other target cell systems, the more NK susceptible ASL1w cells expressed somewhat higher levels of H-2K Ag than did ASL1 cells. These results indicate that expression of H-2 class I proteins by target cells, in and of itself, is not sufficient to inhibit the lytic activity of murine NK cells.     

Susceptibility to natural killer cell-mediated cytolysis is independent of the level of target cell class I HLA expression

To test the hypothesis that susceptibility to NK cell-mediated cytolysis varies inversely with the levels of target cell class I HLA expression, NK-susceptible K562 and MOLT-4 target cells have been transfected via electroporation with cloned human class I HLA-A2 and HLA-B7 genes. Stably transfected cells expressing varying levels of cell-surface class I HLA have been selected by fluorescent activated cell sorting and tested for susceptibility to NK-mediated cytolysis by freshly isolated peripheral blood NK cells from nine normal volunteers as well as by cloned human NK effectors and tumor cells from a patient with an NK cell lymphoproliferative disorder. Expression of class I HLA did not alter the susceptibility of K562 or MOLT-4 target cells to NK-mediated cytolysis by any of the effectors tested. In addition, the class I HLA-expressing transfectant cells were identical to mock transfected cells in their ability to compete for lysis in cold target inhibition assays. Treatment of both mock-transfected and class I HLA-transfected K562 cells with IFN-gamma resulted in decreased susceptibility to NK-mediated cytolysis which was independent of the total level of class I HLA expression. These results demonstrate that the level of target cell class I HLA expression is not sufficient to determine susceptibility or resistance to NK-mediated cytolysis of the classical NK targets K562 and MOLT-4.     

Properties and characterization of a rat spleen cell-derived factor that induces resistance to natural killer cell lysis in YAC lymphoma cells

Supernatants of Con A-stimulated rat spleen cell cultures contain a factor that induces relative resistance to NK cell-mediated cytotoxicity in the YAC cell line, a line that is otherwise highly susceptible to murine NK cell-mediated lysis. This NK-lysis resistance-inducing factor (LRIF) has a Mr of 12,600 Da, as determined by gel filtration chromatography, and an isoelectric pH of 4.8. NK-LRIF is heat labile and is de-activated by treatment with proteolytic enzymes. Unlike immune-IFN (IFN-gamma), NK-LRIF is not inactivated by pH 2 treatment, and antibodies capable of neutralizing IFN-alpha and IFN-gamma do not abrogate the effect of NK-LRIF. Highly purified IL-2 preparations lack NK-LRIF activity. NK-LRIF does not induce a general resistance to lysis in YAC cells, because control and NK-LRIF-treated YAC cells were equally susceptible to alloimmune cytotoxic T cells. YAC cells treated with NK-LRIF showed a marked enhancement (5- to 10-fold) in the expression of class I MHC Ag. This observation supports the proposition that the NK susceptibility of target cells could be inversely related to the expression of class I MHC Ag.     

Differential effects of protein synthesis inhibition on CTL and targets in cell-mediated cytotoxicity

The reactions that lead to target cell lysis by cytotoxic T cells (CTL) are despite intensive investigations poorly understood. To examine the relative roles effectors and targets play in the lytic reaction, protein synthesis in either CTL or targets was inhibited before assay of lysis. We show, in agreement with previous results, that de novo protein synthesis is not necessary in either effectors or targets during the cytolytic reaction. However, activation of CTL requires protein synthesis. Activated CTL respond to protein synthesis inhibitors with a cycling of activity, a result that is interpreted to be consistent with a stimulus secretion mechanism. Treatment of targets with protein synthesis inhibitors prior to incubation with CTL leads to a very rapid and irreversible loss of lytic susceptibility. It is shown that the decrease in lysability is not due to lack of proper CTL target interaction: MHC class I antigens are expressed on drug-treated targets and these cells serve as cold targets in competitive inhibition experiments. Moreover, drug-treated targets trigger transient Ca2+ mobilization and generation of inositol phosphates in CTL. It is therefore concluded that drug-treated targets are able to trigger CTL function but lack a component that is required for their successful lysis.     

Differential recognition of MHC class I molecules of xeno-/allo-endothelial cells by human NK cells

Using human umbilical vein endothelial cells (HUVEC) and porcine aortic endothelial cells (PAEC) as target cells, human peripheral blood NK cells (PBNK) and NK92 cells as effector cells, the differential cytotoxicities of NK cells to allo- and xeno-endothelial cells were studied. The influence of MHC class I molecules on the cytotoxicity of human NK cells was assayed using acid treatment, and blockades of MHC class I antigens, CD94 and KIR (NKB1). The results indicated that the killing of PAEC by the two kinds of NK cells is higher than that of HUVEC. After acid- treatment, the cytotoxicity of the two kinds of NK cells to PAEC and HUVEC is significantly enhanced, but the magnitude of the enhancement is different. The enhancement of NK killing to acid treated HUVEC is much greater than that to PAEC. Blockade of CD94 mAb did not alter the NK cytotoxicity, while blockade of NKB1 mAb enhanced the cytotoxicity of PBNK to HUVEC and PAEC by 95% and 29% respectively. The results above suggested that the differential recognition of MHC I molecules of xeno-endothelial cells by human NK cells could be the major reason for higher NK cytotoxicity to PAEC. KIR might be the primary molecule that transduced inhibitory signals when endothelial cells were injured by NK cells.     

Bacterial superantigens as anti-tumour agents: induction of tumour cytotoxicity in human lymphocytes by staphylococcal enterotoxin A

Summary Activation of lymphocytes by interleukin-2 (IL-2) induces lymphokine-activated killer (LAK) cells that show promising effects on tumour growth in clinical trials. We examined the effect of the superantigen staphylococcal enterotoxin A (SEA) on anti-tumour activity of freshly prepared human lymphocytes. Picomolar amounts of SEA rapidly induced cytotoxic activity against K562 and Raji cells as well as some natural-killer(NK)-resistant tumour cell lines. Cytotoxic activity was not dependent on target cell expression of either major histocompatibility complex (MHC) class I or II antigens as shown using mutated cell lines. Cell-sorting experiments showed that the activity was expressed by NK (CD5^–CD56⁺) as well as T (CD5⁺) cells, although the former contained the majority of cytotoxic activity. NK cells could not be directly activated by SEA. In contrast, SEA activated purified T cells to the same extent as in bulk cultures. It is suggested that SEA activation of NK cells is secondary to that brought about by lymphokines produced by T cells. Activation of LAK cells with SEA was comparable in magnitude as well as target cell spectrum to that of IL-2. In addition to the LAK-like cytotoxic activity induced by SEA, a superimposed cytotoxicity towards target cells expressing MHC class II antigens coated with SEA was observed. This staphylococcal-enterotoxin-dependent cell-mediated cytotoxicity (SDCC) was exclusively mediated by T cells. It is well established that MHC class II antigens function as receptors for staphylococcal enterotoxins on mammalian cells and that the complex between MHC class II antigen and — SEA apparently functions as a target structure for activated T cells with target cell lysis as a consequence. Activation of T lymphocytes with IL-2 also resulted in the capability to mediate SDCC. Staphylococcal enterotoxins represent a novel way of inducing anti-tumour activity in human lymphocytes, which could be of value in therapeutic applications.     

The role of T8 in the cytotoxic activity of cloned cytotoxic T lymphocyte lines specific for class II and class I major histocompatibility complex antigens

It is reported here that most cytotoxic T lymphocytes (CTL), which recognize class I major histocompatibility complex (MHC) loci, express the T cell differentiation antigen T8. However, a minority of T8+ CTL clones was found to recognize class II MHC antigens. To test the hypothesis that T8 is involved only in T cell recognition of class I MHC antigens, we studied the role of T8 in the cytotoxic activity of class II MHC-specific CTL. Monoclonal antibodies specific for T8 blocked the activity of most class I MHC-specific CTL clones but did not affect the activity of class II MHC-specific CTL clones. Moreover, a mild trypsin treatment of the clones, which removed and T8 determinant, affected the activity of class I MHC but not that of class II MHC-specific CTL clones. These findings indicate that the class II-specific MHC CTL clones described here did not require T8 for their cytolytic activity. The activity of one T8+ class I MHC-specific (HLA-B27) CTL clone (HG-61) against the B cell line JY, which was used to raise this CTL clone, was not blocked by trypsin treatment of this clone. However, the activity of CTL clone HG-61 against target cells different from JY but carrying the appropriate HLA specificity was blocked by anti-T8 antibodies and trypsin treatment. The implications of these findings for the hypothesis that T8 is involved only in the activity of CTL with a relatively low avidity for class I MHC antigens are discussed.     

Differential recognition of MHC class I molecules of xeno-/allo-endothelial cells by human NK cells

Using human umbilical vein endothelial cells (HUVEC) and porcine aortic endothelial cells (PAEC) as target cells, human peripheral blood NK cells (PBNK) and NK92 cells as effector cells, the differential cytotoxicities of NK cells to allo- and xeno-endothelial cells were studied. The influence of MHC class I molecules on the cytotoxicity of human NK cells was assayed using acid treatment, and blockades of MHC class I antigens, CD94 and KIR (NKB1). The results indicated that the killing of PAEC by the two kinds of NK cells is higher than that of HUVEC. After acid-treatment, the cytotoxicity of the two kinds of NK cells to PAEC and HUVEC is significantly enhanced, but the magnitude of the enhancement is different. The enhancement of NK killing to acid treated HUVEC is much greater than that to PAEC. Blockade of CD94 mAb did not alter the NK cytotoxicity, while blockade of NKB1 mAb enhanced the cytotoxicity of PBNK to HUVEC and PAEC by 95% and 29% respectively. The results above suggested that the different     

Interaction of monocytes with NK cells upon Toll-like receptor-induced expression of the NKG2D ligand MICA

Reciprocal interactions between NK cells and dendritic cells have been shown to influence activation of NK cells, maturation, or lysis of dendritic cells and subsequent adaptive immune responses. However, little is known about the crosstalk between monocytes and NK cells and the receptors involved in this interaction. We report in this study that human monocytes, upon TLR triggering, up-regulate MHC class I-Related Chain (MIC) A, but not other ligands for the activating immunoreceptor NKG2D like MICB or UL-16 binding proteins 1-3. MICA expression was associated with CD80, MHC class I and MHC class II up-regulation, secretion of proinflammatory cytokines, and apoptosis inhibition, but was not accompanied by release of MIC molecules in soluble form. TLR-induced MICA on the monocyte cell surface was detected by autologous NK cells as revealed by NKG2D down-regulation. Although MICA expression did not render monocytes susceptible for NK cell cytotoxicity, LPS-treated monocytes stimulated IFN-gamma production of activated NK cells which was substantially dependent on MICA-NKG2D interaction. No enhanced NK cell proliferation or cytotoxicity against third-party target cells was observed after stimulation of NK cells with LPS-activated monocytes. Our data indicate that MICA-NKG2D interaction constitutes a mechanism by which monocytes and NK cells as an early source of IFN-gamma may communicate directly during an innate immune response to infections in humans.