Search

The expression of two surface antigens present on the cell membrane of both human granulocytes and monocytes was studied during the process of myelomonocytic differentiation using two monoclonal antibodies (B9.8.1 and B13.4.1). These surface antigens are not present on immature myeloid cells nor on nonmyeloid hematopoietic cells, but can be detected when the cells are terminally differentiated. Among the bone marrow cells, B13.4.1 binds to metamyelocytes and B9.8.1 to metamyelocytes and a fraction (30%) of myelocytes. HL60 human promyelocytic leukemia cells did not react with such monoclonal antibodies. However, when such cells were induced to differentiate in vitro into mature myeloid elements by treatment with retinoic acid or dimethyl sulfoxide, 70%--90% of the differentiated cells expressed both surface antigens. Cell sorting studies on these treated HL60 cells indicated that myelocytes and metamyelocytes were the most immature cells expressing such markers. Expression of the two surface antigens was also observed when HL60 cells were induced to differentiate into monocyte/macrophage cells by treatment with the tumor promoter 12-O- tetradecanoyl-phorbol-13-acetate. Thus, human promyelocytic leukemia cells induced to differentiate in vitro by treatment with specific chemical agents express membrane antigens in the same pattern as normal bone marrow myeloid cells at the corresponding stage of differentiation.

In this study, we analyzed the effect of tumor necrosis factor (TNF) on retinoic acid (RA)-induced myeloid differentiation of the promyelocytic HL-60 leukemia cell line. We show that low concentrations of the two substances, almost inactive in inducing differentiation when used separately, induce differentiation when added simultaneously to the cell cultures. Cells simultaneously expressing both monocyte/macrophage phenotype (typically induced by TNF) and granulocyte characteristics (typically induced by RA) are induced by a combination of the two factors, indicating that TNF and RA potentiate each other's activity. The results obtained using immune interferon (IFN-gamma) in combination with the two inducers suggest that the mechanism of action of TNF and IFN-gamma are possibly different. The inhibitory effect of RA on the expression of HLA class I antigens and of the high-affinity Fc receptor is potentiated by TNF but completely reversed by rIFN-gamma.

We present evidence that human peripheral blood lymphocytes, free of contaminating monocytes, rapidly produce high levels of tumor necrosis factor (TNF) when stimulated with phorbol diester and calcium ionophore, and lower but significant levels of TNF when stimulated with mitogens. These two types of inducers act preferentially on T cells, both CD4+ and CD8+. NK cells produce TNF only when stimulated with phorbol diester and calcium ionophore, and they do so at a much lower level than T cells. The procedures used in the purification of lymphocytes and the differential ability to respond to various inducers allow us to exclude that monocytes or basophils contaminating the lymphocyte preparation participate in the production of TNF. In particular, LPS, a potent inducer of TNF production from monocytes, is unable to induce significant levels of TNF in the lymphocyte preparations. The TNF produced by lymphocytes has antigenic, physicochemical, and biochemical characteristics identical to those of the TNF produced by myeloid cell lines or monocytes upon stimulation with LPS. LT is also produced by lymphocyte preparations. Production of TNF and LT proteins in response to the different inducers is paralleled by accumulation of cytoplasmic TNF and LT mRNA. Both at mRNA and at protein levels, stimulation of T lymphocytes with phorbol diester and calcium ionophore preferentially induces TNF, whereas mitogen stimulation preferentially induces LT. Our data suggest that the TNF and LT genes, two closely linked genes encoding two partially homologous proteins with almost identical biological functions, are independently regulated in lymphocytes.

The role of HLA-DR+ cells in NK activity against CMV-infected FS4 foreskin fibroblasts and K562 erythroleukemia cells was examined. When nonadherent PBMC were depleted of either HLA-DR+ or Leu-11b+ cells by treatment with mAbs plus C, NK activity against CMV-FS4 target cells was markedly reduced. In contrast, depletion of HLA-DR+ cells had no effect on NK activity against K562 target cells. When HLA-DR-depleted cells were added to Leu-11b-depleted cells, NK activity against CMV-FS4 was restored. Negative selection experiments indicated that the HLA-DR+ cells contributing to NK activity against CMV-FS4 are not B or T cells, while negative and positive selection experiments excluded a role for monocytes. Experiments in which HLA-DR- and Leu-11b- cells were mixed in varying proportions indicated that NK(CMV-FS4) is mediated by Leu-11b+ cells, while HLA-DR+ cells provide an accessory function. Irradiation (50 GY) abolished the NK effector function of Leu-11b+ cells, but not the accessory function of HLA-DR+ cells. The NK activity against CMV-FS4 of HLA-DR- cells was restored by the addition of rIFN-alpha or of cell-free supernatants generated by coculturing PBMC or Leu-11b- cells with CMV-FS4. The ability of these supernatants to restore NK activity of HLA-DR- cells was completely abrogated by the addition of neutralizing amounts of antibody to IFN-alpha. In related experiments, neutralization of IFN-alpha in NK assays had little or no effect on NK activity against CMV-FS4, suggesting that the accessory function of HLA-DR+ cells might be mediated by alternative mechanisms in addition to the secretion of extracellular IFN-alpha.

We show that the cytotoxins tumor necrosis factor (TNF) or lymphotoxin (LT), at concentrations of approximately 10(-11) M induce monocytic differentiation of human myeloid cell lines. After 5 d of culture in the presence of rTNF and LT, a significant proportion of the myeloid cell lines express monocyte differentiation antigens and ANAE activity, and become able to reduce nitroblue tetrazolium (NBT) and mediate low levels of ADCC against tumor target cells. These markers of differentiation, however, are maximally induced when rIFN-gamma, at concentrations as low as 4 U/ml, is present simultaneously with the cytotoxins, and the two classes of cytokines act synergistically to induce terminal differentiation. The appearance of monocytic antigens is accompanied by acquisition of morphology and other functional properties of mature monocytic cells, such as chemiluminescence and phagocytosis, and by expression of FcR for monomeric IgG. A decrease in cell proliferation accompanies induced differentiation, and is not due to the cytotoxic properties of TNF or LT, as indicated in simultaneous analysis of surface phenotype and cell cycle. The lack of cytotoxicity of TNF on the HL-60 cell line is also demonstrated by the enhancing effect of TNF on HL-60 cell growth and nucleoside uptake in the first 2 d of culture. These data show that the cytotoxins TNF and LT mediate complex effects on cells of the myelomonocytic lineage and, in synergy with IFN-gamma, can fully induce immature myeloid cells to differentiate into cells with phenotypic, functional, and proliferative characteristics of terminally differentiated myelomonocytic cells.

Conditioned medium (CM) from lectin-stimulated human leukocytes contains factors that induce human promyelocytic cell lines to differentiate along the monocytic pathway. In this report, we show that human promyelocytic cell lines are also induced to differentiate along this pathway by immune interferon (IFN gamma). Various preparations of IFN alpha tested did not induce this differentiation. In cultures containing IFN gamma, the cells are induced to coordinately express monocyte markers and functions such as monocyte-specific surface antigens, HLA-DR antigens, nonspecific esterase, receptors for the Fc fragment of IgG, and the ability to mediate antibody-dependent cell- mediated cytotoxicity. Our data indicate that differentiation induced by IFN gamma is not secondary to an arrest of growth of promyelocytic cell lines, but rather that a proportion of cells is induced along a programmed pathway of terminal differentiation similar to that of normal monocytes. CM contains IFN gamma, but its ability to induce differentiation is greater than expected on the basis of its content of IFN gamma. Treatments at 56 degrees C or at pH 2.0, which abolish IFN gamma activity, abrogate the differentiation ability of CM. The antiviral activity and the differentiation activity contained in the CM are coeluted from gel filtration and reverse-phase columns. Monoclonal antibodies anti-IFN gamma, which completely abrogate the differentiation ability of IFN gamma and the antiviral activity in the CM, completely suppress the induction of some monocyte markers by CM, but only reduce the expression of others. When IFN gamma is added to CM, promyelocytic cell lines are induced to differentiate to a much greater extent than that induced by either IFN gamma or IFN gamma- depleted CM alone. These results show that the differentiation activity of leukocyte CM is due to the synergistic effect of IFN gamma and other factors not yet identified.

Following stimulation with adenosine diphosphate (ADP), collagen, or arachidonic acid, unstirred human platelet suspensions bind 125I- fibrinogen in a reaction that reaches completion within 30 min. Scatchard analysis of these binding data reveals two sets of binding sites with all 3 agents: a high affinity site (Kd 0.029–0.045 microM) binding 1000–1600 fibrinogen molecules per platelet, and a lower affinity site (Kd 1.2–2.0 microM) binding 46,000–76,000 fibrinogen molecules per platelet. At a concentration of apyrase that inhibited ADP-induced fibrinogen binding by greater than 85%, fibrinogen binding induced by collagen and arachidonic acid was only partially affected. This suggests that fibrinogen binding induced by collagen or arachidonic acid does not require released ADP. We isolated a monoclonal antibody, B59.2, which precipitated the glycoprotein IIb- IIIa complex from solubilized platelet membranes. Binding of labeled antibody to platelets before or after exposure to ADP, collagen, or arachidonic acid showed a single class of approximately 22,000 binding sites with Kd 0.019 microM. Binding of B59.2 was complete within 1 min and was not inhibited by EDTA. Preincubation of platelet suspensions with a 2.1 microM concentration of B59.2 caused inhibition of secretion and aggregation, but not of thromboxane-B2 synthesis, in response to 1 microgram/ml collagen, 40 microM arachidonic acid, or 4 microM ADP, concentrations of aggregating agents that produced complete aggregation and secretion in the absence of B59.2. At this concentration of B59.2, fibrinogen binding to stimulated platelets was inhibited by approximately 45%-55%. These data demonstrate that collagen and arachidonic acid can expose fibrinogen binding sites independently of released ADP; and that the glycoprotein IIb-IIIa complex is involved in secretion, aggregation, and fibrinogen binding, but not in thromboxane synthesis occurring in response to collagen, arachidonic acid, or ADP.

We describe here the production and characterization of several new monoclonal antibodies that recognize differentiation antigens present on human cells of the myelomonocytic lineage. The lineage and the stage specificities of our reagents (myeloid-, monocytic-, and myelomonocytic- specific) were determined on the basis of their reactivity with human cell lines and with human peripheral blood and bone marrow cells. Cross- competition experiments demonstrated that some of the antibodies react with the same or closely associated antigenic determinants. Five antigens have been identified in this way: one present on myeloid, one on monocytic, and three on both myeloid and monocytic cells. The possible relationship of our antibodies with other established monoclonal antibodies is discussed, in addition to their use in the in vitro study of the differentiation pathways of human hemopoietic cells and in the characterization of leukemias.

We have identified and purified a novel cytokine, NK cell stimulatory factor (NKSF), from the cell-free supernatant fluid of the phorbol diester-induced EBV-transformed human B lymphoblastoid cell line RPMI 8866. NKSF activity is mostly associated to a 70-kD anionic glycoprotein. The purified 70-kD protein, isolated from an SDS-PAGE gel, yields upon reduction two small species of molecular masses of 40 and 35 kD, suggesting that this cytokine is a heterodimer. When added to human PBL, purified NKSF preparations induce IFN-gamma production and synergize with rIL-2 in this activity, augment the NK cell-mediated cytotoxicity of PBL preparations against both NK-sensitive and NK-resistant target cell lines, and enhance the mitogenic response of T cells to mitogenic lectins and phorbol diesters. The three activities remain associated through different purification steps resulting in a 9,200-fold purification, and purified NKSF mediates the three biological activities at concentrations in the range of 0.1-10 pM. These data strongly suggest that the same molecule mediates these three activities, although the presence of traces of contaminant peptides even in the most purified NKSF preparations does not allow us to exclude the possibility that distinct biologically active molecules have been co-purified. The absence of other known cytokines in the purified NKSF preparations, the unusual molecular conformation of NKSF, the high specific activity of the purified protein, and the spectrum of biological activities distinguish NKSF from other previously described cytokines.

We show here that human peripheral blood NK cells contain a pore-forming protein (PFP) with an Mr of 70,000-72,000 that assembles structural lesions (with an average internal diameter of 150-170 A) and forms functional channels. The PFP was isolated by affinity chromatography from human NK cells, using a specific anti-C9 antiserum as the immunoadsorbent. The NK cells were isolated from PBL by positive or negative selection by indirect rosetting using a panel of monoclonal antibodies directed against different NK and T cell surface antigens. PFP was identified in NK cells freshly isolated and isolated from cultured PBL, both stimulated with interleukin 2, but not in NK cell-depleted lymphocytes. In planar bilayers, the channels formed by the NK cell-derived PFP are highly voltage resistant, with most channels persisting in the open state once they have inserted into the bilayer. The unit conductances of these channels range 0.3-1 nS in 0.1 M NaCl. The channels show poor selectivity for monovalent and divalent ions. The PFP is also released from human NK cells stimulated with the calcium ionophore A23187, suggesting that this protein, like the one produced by murine CTL lines, may be similarly secreted during cell-mediated killing. Its identification in primary human NK cell cultures indicates that this protein may play an active role in NK cell-mediated killing.

We report here that FcR for human monomeric IgG1 can be induced on cells of myeloid origin cultured in the presence of IFN gamma for 8 h. Supernatant fluids from cultures of lymphocytes infected with a variety of viruses or cocultured with cell lines have the same FcR enhancing effect as IFN gamma. We identify the factor in the supernatant fluid responsible for the induction as immune interferon. Among the different types of IFN, only the gamma type (both purified and recombinant) specifically induces the appearance of FcR for monomeric IgG1 on normal and leukemic myeloid cells but not on cells of lymphoid origin. This effect is also evident on mature PMN. We show that the specificity and the affinity of the receptor induced on HL-60 promyelocytic cells, peripheral blood monocytes, and PMN are identical to those of the receptor spontaneously present on the same cells, except for PMN, which do not spontaneously express this type of receptor. The results of inhibition experiments performed with mouse IgG of and IgG3. These results suggest that the receptor present on human monocytes different isotypes indicate that the receptor can be inhibited by murine IgG2a or immature myeloid cells, selectively inducible by IFN gamma, has a specificity similar to the FcR1 described on mouse macrophages.

We report evidence that FcR(CD16) on human NK cells are signal-transducing molecules that, upon ligand binding, induce transcription of genes encoding surface activation molecules [IL-2-R(CD25)] and cytokines (IFN-gamma and TNF) relevant to NK cell biology and functions. Homogeneous NK and T cell populations purified from short-term bulk cultures of PBMC with irradiated B lymphoblastoid cell lines were cultured in the presence of FcR ligands (particulate immune complexes or immobilized anti-CD16 antibodies) alone or with rIL-2. Upon 18 h of stimulation, NK cells express Tac, TfR, and 4F2 antigens and produce IFN-gamma and TNF; both effects are synergistically enhanced in the presence of rIL-2, which is itself ineffective. Treatment of NK cells with FcR(CD16) ligands induces accumulation of mRNA for IFN-gamma and TNF and, to a lesser extent, IL-2-R with fast kinetics also in the absence of de novo protein synthesis. rIL-2 and FcR(CD16) ligands synergize to induce mRNA accumulation. mRNA accumulation and transcription of TNF and IFN-gamma genes induced by FcR(CD16) ligands are greater than those induced by rIL-2, and the reverse is true for IL-2-R. The two stimuli do not synergize at the transcriptional level. These observations indicate that the mechanisms through which FcR(CD16) ligands and rIL-2 induce NK cell activation are, in part, distinct. Both operate at the transcriptional level, although other mechanisms are probably induced by the FcR ligand stimulus per se or in combination with other lymphokines and synergize at a post-transcriptional or translational level to enhance NK cell activation.

We characterize the natural killer (NK) cell colony-inhibiting activity (CIA) produced in supernatants from cultures of human peripheral blood lymphocytes (PBL) with NK-sensitive target cell lines, and study its relationship with NK cell-derived cytotoxic factor (NKCF). Using monoclonal antibodies (mAb) specific for NK cells or other lymphocyte populations, we unambiguously identify NK cells as the only PBL subset able to produce both NKCF and NK-CIA. We present functional and biochemical data suggesting that NKCF and NK-CIA represent the same molecule: (a) a highly significant positive correlation exists between the quantity of NKCF and NK-CIA in supernatants independently produced by different PBL subsets; (b) both NK-CIA and NKCF are induced by culture of PBL with NK-sensitive, but not with NK-insensitive cell lines, and with HLA-DR+ bone marrow cells; (c) both NKCF and NK-CIA are absorbed on the same cell lines or bone marrow cell types; (d) the two activities coelute in the same gel filtration fractions; (e) D-mannose-6-phosphate blocks both NKCF and NK-CIA activity, and prevents their absorption by K562 cells; and (f) both NKCF and NK-CIA activity are lost after 2 d at 37 degrees C. The NK-CIA-containing preparations are devoid of antiviral activity, and antiinterferon (anti-IFN) antibodies do not block the inhibitor activity of NK-CIA. The effect of NK-CIA on day 14 (early) colony-forming units of granulocytes and macrophages (CFU-GM) is synergistic with that of IFN-gamma, and this synergy is also evident on day 7 (late) CFU-GM growth. A combination of NK-CIA and IFN-gamma suppresses late CFU-GM, at concentrations of the two lymphokines that are completely ineffective when used independently. No synergy between NK-CIA and IFN-alpha or -beta was observed, due to a direct inhibitory effect of these two IFN types on late CFU-GM. Antibodies specific for tumor necrosis factor (TNF), but not those specific for lymphotoxins, inhibit both NK-CIA and NKCF activity in the NK cell-derived supernatant. Recombinant TNF, in the range of concentrations corresponding to that of the cytotoxic activity on L-929 cells present in supernatants, mediated both NKCF and NK-CIA activity.(ABSTRACT TRUNCATED AT 400 WORDS)

The present study shows that recombinant interleukin 2 (IL-2) purified to homogeneity induces a rapid and potent enhancement of spontaneous cytotoxicity of human peripheral blood lymphocytes. The cells mediating cytotoxicity after 18-h treatment with IL-2 have surface markers of natural killer (NK) cells and are generated from the peripheral blood subset containing spontaneous cytotoxic cells. A parallel production of gamma interferon (IFN-gamma) is induced by recombinant IL-2 (rIL-2), and NK cells appear to be the major producer cells, whereas T cells are unable to produce IFN-gamma under these experimental conditions. However, the kinetics of the enhancement of cytotoxicity are faster than those of IFN-gamma production, and monoclonal anti-IFN-gamma antibodies do not suppress this effect, making it unlikely that the IFN-gamma produced is responsible for the enhancement. The enhancement of NK cell activity induced by rIL-2 precedes any proliferative response of the lymphocytes, which is instead observed in longer-term cultures of both NK and T cells.

We report evidence that murine NK cells express a functional Fc gamma RII encoded by the Fc gamma RII alpha gene. Several lines of indirect evidence indicate that freshly obtained NK cells from mice of several strains bear a functional Fc gamma RII: (a) anti-Fc gamma RII antibody 2.4G2 detects a small but significant proportion of sIg- cells and a small proportion of the 2.4G2+ cells are included in the Thy-1+ population; (b) sIg- lymphocytes contain 2.4G2+ and Fc gamma R-bearing cells in similar proportions; (c) binding of particulate immune complexes by sIg- lymphocytes is completely inhibited by 2.4G2; (d) 2.4G2+ cells mediate greater than 50% of the spontaneous cytotoxicity in sIg- splenic lymphocytes. Direct evidence for the presence of Fc gamma RII on murine NK cells is provided by the results of two-color immunofluorescence studies performed on splenic lymphocytes from C57BL/6 mice showing coexpression of NK-1.1 and 2.4G2. Studies of in vitro propagated homogeneous NK cell populations confirm that murine NK cells express only Fc gamma RII and that this Fc gamma R is functional, as shown in experiments of inhibition of ADCC by the anti-Fc gamma RII antibody 2.4G2. The results of studies at the molecular level show that an Fc gamma RII alpha transcript identical to that expressed in macrophages is the only molecule encoding Fc gamma RII in murine NK cells.

Conditioned medium from phytohemagglutinin-stimulated human leukocytes contains a factor that can induce promyelocytic cell lines and certain acute myelogenous leukemia cells to differentiate along the monocytic pathway. In this report, we show that immature myeloid cells from normal bone marrow or the peripheral blood of patients with chronic myelogenous leukemia can be induced to differentiate to monocyte-like cells by immune gamma interferon (IFN gamma). We have identified IFN gamma as the predominant differentiation factor contained in the conditioned medium. Purified or recombinant IFN gamma, but not various preparations of IFN alpha or beta, can induce monocytic differentiation in myeloid cells. In cultures containing conditioned medium, the cells fail to continue myeloid maturation, and are induced to express monocyte markers and functions, such as monocyte-specific surface antigens, HLA-DR antigens, Fc receptors for monomeric immunoglobulins, nonspecific esterase, and the ability to mediate antibody-dependent, cell-mediated cytotoxicity. Even myeloid cells as mature as metamyelocytes or band cells can be induced by IFN gamma to undergo monocyte differentiation, but monocyte-specific or HLA-DR antigens are not induced in mature neutrophils. These findings reveal a previously unknown, specific function of human IFN gamma and offer new insights to the regulation of monocyte recruitment and differentiation during a virus infection or immune response.

Incubation of human peripheral blood lymphocytes with bone marrow cells resulted in significant inhibition of colony formation by committed myeloid and erythroid cells. Using positively selected homogeneous natural killer (NK) cell preparations and lymphocyte subpopulations depleted of or enriched for NK cells, we definitively characterize as NK cells the cells in normal peripheral blood that are responsible for inhibition of bone marrow colony growth. The inhibitory effect of NK cells on hematopoiesis can be mediated by a soluble factor that is produced only by NK cells upon culture with HLA-DR+ hematopoietic cells and with NK-sensitive cell lines. Both NK cells and the NK-produced, colony-inhibiting activity (NK-CIA) are suppressive for allogeneic and autologous bone marrow CFU-GEMM (colony-forming units, granulocyte, erythroid, monocyte, megakaryocyte), CFU-E (CFU, erythroid), and early CFU-GM (CFU, granulocyte, monocyte), but not for either BFU-E (burst-forming units, erythroid) or late CFU-GM. [3H]Thymidine incorporation was inhibited by NK-CIA-containing supernatants in HLA-DR+ but not HLA-DR- bone marrow cell populations stimulated to proliferative by colony-stimulating factor (CSF). These data suggest that the NK cell-mediated inhibitory effect on proliferation and differentiation of hematopoietic precursor cells is mediated in part or completely by the secreted NK-CIA. The concentration of NK-CIA reached in the supernatant of the mixture of NK cell-containing lymphocyte populations with bone marrow cells is sufficient to account for the inhibitory effect mediated by NK cells. Our data support the hypothesis that human NK cells play a major role in the control of hematopoiesis, down-regulating it under conditions in which the NK cells are functionally activated.

A low affinity receptor for IgG immune complexes, Fc gamma RIII(CD16), is expressed on human NK cells as an integral membrane glycoprotein anchored through a transmembrane peptide; on polymorphonuclear neutrophils (PMN) the receptor is anchored through a phosphatidylinositol (PI) linkage. The protein on NK cells has a molecular mass 6-10 kD larger than that on PMN, and, unlike the latter, is resistant to PI-specific phospholipase C (PI-PLC). Fc gamma RIII(CD16) transcripts isolated from PMN and NK cells of single donors revealed multiple single nucleotide differences, one of which converts an in frame UGA termination codon to a CGA codon. The resulting open reading frame encodes a longer cytoplasmic domain for Fc gamma RIII(CD16) in NK cells, contributing to its transmembrane anchor. Two nearly identical, linked genes that encode these transcripts have been cloned for Fc gamma RIII(CD16), one of which (III-1) is allelic for NA-1 and NA-2. The allelic sites have been mapped to two single nucleotides in the extracellular domain. These genes are transcribed in a cell type-specific fashion to generate the alternatively anchored forms of this receptor.

We have analyzed the ability of highly purified preparations of human NK cells to produce CSF. NK cells, purified by negative selection from 10-d cultures of PBMC incubated with irradiated B-lymphoblastoid cell lines, were stimulated with rIL-2, FcR(CD16) ligands (particulate immune complexes or anti-CD16 antibodies bound to Sepharose), a combination of CD16 ligands and rIL-2, or the phorbol diester phorbol dibutyrate (PDBu) together with the Ca2+ ionophore A23187. Both rIL-2 and CD16 ligands induce accumulation of GM-CSF mRNA in NK cells and the combined effect of the two stimuli is synergistic. Maximal accumulation of GM-CSF mRNA is observed after PDBu/A23187 stimulation. The participation of contaminant T cells in the observed expression of the GM-CSF gene is excluded because CD16 ligands do not stimulate T cells and CD3 ligands, powerful stimulators of T cells, are inactive on NK cells. Accumulation of CSF-1 mRNA is observed only in NK cells stimulated with both CD16 ligands and rIL-2, whereas accumulation of IL-3 mRNA is observed only in NK cells stimulated with PDBu/A23187. Transcripts of the G-CSF, IL-1 alpha, and IL-1 beta genes were never detected in NK cells in these experiments. The kinetics of accumulation of GM-CSF and CSF-1 mRNA in NK cells stimulated with CD16 ligands and rIL-2 peaked at 2-4 h and was slower than that of TNF and IFN-gamma mRNA, which peak at 1 h. GM-CSF was precipitated from the supernatant fluids of NK cells stimulated with PDBu/A23187 and its biological activity was demonstrated by the ability of the supernatants to sustain proliferation of the TALL-101 cell line or CML blasts. Biological activity of IL-3 and CSF-1 was demonstrable in supernatant fluids of NK cells stimulated with PDBu/A23187 and CD16 ligands/rIL-2, respectively.

In this study, we present evidence that interaction of Fc gamma R(CD16) with ligands (immune complexes or anti-CD16 antibodies) induces a rapid rise in [Ca2+]i and fast production of both inositol 1,4,5 triphosphate (IP3) and IP4 in homogeneous NK cell preparations. Part of the initial [Ca2+]i rise observed upon stimulation of NK cells with either anti-CD16 antibodies alone or after their crosslinking at the cell membrane depends on Ca2+ mobilization from intracellular stores, but sustained [Ca2+]i levels are maintained, after the initial spike, through influx of extracellular Ca2+. The [Ca2+]i rise is mediated, at least in part, by increases in IP3 after receptor-induced hydrolysis of membrane polyphosphoinositides (PPI). The role of extracellular Ca2+ in Fc gamma R(CD16)-dependent induction of lymphokine gene expression has been tested by evaluating production, mRNA accumulation and transcription of IFN-gamma and TNF in NK cells stimulated with Fc gamma R(CD16) ligands and/or rIL-2 in the presence of EGTA. Under these conditions, accumulation and transcription of both IFN-gamma and TNF mRNA induced by CD16 ligands, but not that induced by rIL-2, is completely abolished and neither cytokine can be detected at significant levels in the supernatant fluids of cells so treated. These data confirm that NK cell activation by specific ligands occurs through mechanisms distinct from those induced by IL-2, and indicate that extracellular Ca2+ represents a stringent requirement for cytokine production induced in NK cells through specific (Fc gamma R) stimulation. Our data also indicate that the [Ca2+]i rise induced upon Fc gamma R(CD16) crosslinking, though necessary, is not sufficient per se to induce activation of lymphokine genes, compatible with the hypothesis that Fc gamma R(CD16) crosslinking generates additional transducing signals that synergize with IL-2 to maximally activate NK cells.

CTL and NK cells resist self-mediated killing and lysis by their own pore-forming protein (PFP; perforin). Perforin, like C, lyses RBC. Efficient C-mediated lysis of RBC occurs when both C and RBC are from different species (homologous species restriction). A protective surface protein (C8-binding protein, homologous restriction factor) has been reported to mediate both homologous species restriction in C-dependent cytolysis and protection of some target cells against perforin-induced lysis. We show here that perforin, unlike C, lyses target cells across a variety of species, including the homologous one, while the same target cell populations resist the attack by homologous C. Perforin-containing extracts of CTL and LAK/NK cells from three species (rat, mouse, and human) and purified mouse perforin were tested against RBC from 10 different species, several nucleated target cell lines, and one primary cell population (thymocytes). While resisting lysis by homologous C, most of these cell types were lysed effectively by perforin without any homologous restriction pattern. CTL and NK cells, like other nucleated targets, are resistant to lysis by homologous but not heterologous C; however, these cell types are resistant to both homologous and heterologous perforin. Together, our results suggest that the protective mechanisms associated with C- and perforin-mediated lysis are distinct.