Lectin-dependent neutrophil-mediated cytotoxicity. II. Possible mechanisms.

The mechanisms whereby neutrophils become cytotoxic to chicken erythrocyte (CRBC) target cells were investigated in a system of lectin-dependent neutrophil-mediated cytotoxicity (LDNMC). Through the use of drugs and specific metabolic inhibitors, LDNMC was found to be dependent on energy supplied by anaerobic glycolysis and on other active metabolic functions of the neutrophil. 2-Iodoacetamide, 2-deoxy-D-glucose, di-isopropyl-fluorophosphate, colchicine, cytochalasin B, and dibutyryl cyclic AMP all caused dose-dependent inhibition of cytotoxicity, while inhibitors of protein and nucleic acid synthesis were without effect. Cell surface membrane-active agents, such as chloroquine, hydrocortisone and chlorpromazine inhibited cytotoxicity, while vitamin A caused enhancement. Lectins which agglutinated neutrophils, but not necessarily CRBC, such as phytohaemagglutinin (PHA-P), concanavalin A (Con A), soybean agglutinin (SBA), and Ricinus communis agglutinin (RCA), mediated cytotoxicity while lectins which did not cause agglutination, such as pokeweed mitogen (PWM), did not mediate cytotoxicity. Preincubation of neutrophils, but not CRBC with PHA-P, resulted in time-dependent enhancement of cytotoxicity, while pre-incubation with Con A yielded progressive inhibition of cytotoxicity. These studies suggest that lectin binding to the cell surface causes alterations of the membrane, that LDNMC requires cell to cell surface contact, and that cytotoxicity depends on active metabolic processes.     

Lectin-dependent cell-mediated cytotoxicity and blastogenesis by large granular-enriched and depleted lymphocytes.

The role of larger granular-enriched and depleted lymphocytes was studied in lectin-dependent cell-mediated cytotoxicity (LDCC) against adherent HEp-2 human epipharynx carcinoma target cells. LDCC was evaluated by detachment from the monolayer of 3H-thymidine-prelabelled HEp-2 cells in a 24 h assay at effector-target cell ratios of 25:1 and 50:1 in the presence of 25 micrograms/ml concanavalin A (Con A). Under the aforementioned conditions but in the absence of Con A natural cell-mediated cytotoxicity (NCMC) was not found. However, cytotoxicity was significantly augmented in the presence of Con A (= LDCC) using human peripheral blood mononuclear cells (PBMC) as effectors. Large granular lymphocytes (LGL), which show high natural killer (NK) activity to K 562 target cells, failed to be cytotoxic against HEp-2 targets similar to large granular depleted lymphocytes (LGL-DL). On the other hand, LGL caused only a slight LDCC; whilst LGL-DL induced strong LDCC activity towards HEp-2 targets. In comparison to LDCC using LGL-DL as effector cells, LGL and LGL-DL mixed at a ratio of 1:2, and added to target cells, had no major effect on LDCC, while a lower level of LDCC was observed at LGL/LGL-DL ratios of 1:1, and 2:1, suggesting the dilution of LGL-DL, potential effectors of LDCC to HEp-2 cells, rather than a specific regulatory role of LGL in LDCC. In parallel studies, the proliferation of LGL-DL in response to Con A was less than that observed with PBMC or LGL. The response could be restored by replacing half of LGL-DL per culture with an equal number of LGL, or by the addition of 10% monocytes. Significant functional differences between LGL and LGL-DL in LDCC as well as in Con A-induced blastogenesis are suggested.     

Evidence for the role of superoxide radicals in neutrophil-mediated cytotoxicity.

Human peripheral neutrophils became cytotoxic to chicken red blood cells (CRBC) in the presence of lectins as assessed by release of 51chromium from labelled target cells. Phytohaemagglutinin (PHA) and concanavalin A (Con A), which caused time-dependent and dose-dependent cytotoxicity over a concentration range of 25--400 microgram/ml, also caused significant generation of superoxide radicals as measured by ferricytochrome C reduction. Pokeweed mitogen, which does not induce cytotoxicity over the same concentration range, was unable to promote superoxide generation by neutrophils. PHA-induced generation of superoxide paralleled and appeared to precede PHA-dependent cytotoxicity. Superoxide dismutase (SOD), which enzymatically destroys superoxide, caused moderate inhibition of PHA-dependent cytotoxicity over the concentration range of 100--500 microgram/ml whereas catalytically inactive enzyme had no effect. Incubation under oxygen-depleted conditions caused a marked decrease in both PHA-induced superoxide generation and cytotoxicity relative to that obtained with neutrophils incubated aerobically. These findings suggest a central role for superoxide radicals in causing target cell damage in this model of neutrophil-mediated cytotoxicity.     

Complement-dependent, polymorphonuclear neutrophil-mediated cytotoxicity of herpesvirus-infected cells: possible mechanism(s) of cytotoxicity.

Highly enriched bovine polymorphonuclear neutrophils (PMN) were able to destroy herpesvirus-infected cells in the presence of complement. Our results suggest that some viral antigens are capable of activating the alternate complement pathway directly which results in cytotoxicity however, the addition of PMN plus complement resulted in high levels of cytotoxicity. Chelation of Ca2+ or Mg2+ indicated that Ca2+ was involved in PMN--target cell interactions but Mg2+ was involved in activation of the alternate complement pathway. Cytotoxicity in the presence of PMNs was shown to be under bidirectional control of cyclic nucleotides. Thus agents that elevated cAMP reduced cytotoxicity whereas agents that elevated GMP increased it. If the PMN is secreting some product during the cytotoxic reaction it must be performed since chemical shown to inhibit DNA, RNA and protein synthesis had no influence on CDNC. Our results are discussed in terms of the mechanism of CDNC lysis as well as in the vivo significance of such a defence mechanism.     

Inhibition of neutrophil-mediated cytotoxicity by exogenous adenosine 5'-triphosphate

Human polymorphonuclear leukocytes (PMNs) obtained from normal donors kill human tumor cells in vitro. However, if adenosine 5'-triphosphate (ATP) is added to the neutrophil tumor cell suspensions in micromolar concentrations (10-100 microM), there is marked inhibition of neutrophil-mediated cytotoxicity. The inhibitory activity resulted from an effect of ATP on both the effector cells as well as the target cells. When either the effector cells or target cells were preincubated with ATP they became resistant to the effects of the cytotoxic neutrophils. In addition, inhibitory activity was specific to ATP; as it was not demonstrated with GTP, UTP, or CTP. However, when the other adenosine compounds (AMP and ADP) were tested, both AMP and ADP had some inhibitory activity. Cytotoxicity was also inhibited when 100 microM of ATP were added to the neutrophil monolayers either at the time of addition of the tumor cells or 15-60 min after addition of the tumor cells whereas no inhibition of cytotoxicity occurred when ATP was added more than 1 hr after the initiation of the cytotoxic reaction.     

Lectin-dependent cell-mediated cytotoxicity in an invertebrate model: Con A does not act as a bridge.

The plant lectin concanavalin A (Con A) has been used in an invertebrate model of lectin-dependent cell-mediated cytotoxicity (LDCC). Macrophage-like cells from the susceptible host snail Biomphalaria glabrata become cytotoxic effectors when they encounter sporocysts of the parasitic trematode Schistosoma mansoni that have been treated with Con A. The sugar alpha-methyl mannoside and rabbit anti-Con A antibodies fail to block this LDCC. Con A is effective only when the target, not the effector cell, has been exposed to it. These results constitute evidence against the molecular bridging hypothesis and support the notion that surface modulation of the target may be the stimulus that provokes cytotoxicity. Results from this invertebrate model are discussed in the context of murine T lymphocyte LDCC.     

Interferon-gamma is unable to increase monocyte and neutrophil-mediated nonspecific cytotoxicity induced by immune complexes.

We have previously demonstrated that normal human neutrophils and monocytes triggered by immune complexes (IC) are able to destroy non-sensitized target cells through the activation of a nonspecific cytotoxic mechanism (NSC), that is dependent on the generation of reactive oxygen intermediates (IRO). In the present study, we analyze the ability of interferon-gamma (IFN-gamma) to modulate NSC. Our results indicate that, despite the ability of IFN-gamma to increase both the generation of superoxide anion and hydrogen peroxide by phagocytic cells and the expression of the high-affinity 72-kDa Fc gamma RI, it is completely unable to increase NSC mediated either by neutrophils or monocytes. These data suggest that there is no correlation between cytotoxicity and the ability of phagocytic cells to release superoxide anion and/or hydrogen peroxide. They also indicate that Fc gamma RI is not involved in the induction of NSC. To further analyze this point, we studied the ability of two monoclonal antibodies (mAb), specific for different epitopes of Fc gamma RI, to inhibit NSC. These mAb strongly inhibited ADCC mediated by untreated monocytes or IFN-gamma treated monocytes and neutrophils. On the other hand, they were completely unable to inhibit NSC mediated by untreated or IFN-gamma treated cells.     

Virus-induced complement activation and neutrophil-mediated cytotoxicity against respiratory syncytial virus (RSV).

Complement-dependent neutrophil-mediated cytotoxicity (CDNC) was determined by specific release of 51-chromium (51Cr) from respiratory syncytial virus infected HEp2 cells in a microcytotoxicity assay. There was significant release of 51Cr from RSV infected cells as compared to uninfected cells in the presence of complement (C) and neutrophils (PMN). The degree of cytotoxicity was dependent upon the concentration of C used in the assay. Such cytotoxicity was effectively abolished after heat-inactivation of complement. Complement deficient in C4 did not induce cytotoxicity. Similarly, inhibitors of C1 or C3 blocked CDNC. The maximal CDNC was observed at 37 degrees C with little or no response at 4 degrees C. Lymphocytes and monocytes mediated complement-dependent cytotoxicity very poorly in comparison to PMN. Evidence of complement activation by infected cells was demonstrated by the detection of C3 fixed to RSV infected cells by indirect immunofluorescence. Treatment of C with EDTA or heat prevented subsequent attachment of C3 to the infected cells. These in vitro observations suggest an initial activation of complement by RSV infected cells and subsequent lysis by PMN. It is proposed that this process may play a role in the elimination of virus in the early phase of infection in the absence of specific antibody or sensitized lymphocytes.     

Lectin-dependent cellular cytotoxicity by human lymphocytes is blocked by monoclonal antibodies to the sheep red cell receptor

Investigating the functional effects of anti-T cell monoclonal antibodies (obtained through the Second International Workshop on Human Leucocyte Differentiation Antigens), antibodies to the sheep red cell receptor (CD2) were found to block lectin-dependent cytotoxicity. Blocking was seen with all anti-CD2 reagents tested regardless of epitope specificity, immunoglobulin subclass or other factors. Antibodies to other T cell differentiation antigens had no effect.     

Direct cytotoxicity against chicken erythrocytes in mice. I. Fundamental nature of T cell-mediated cytotoxicity.

Immune responses were examined after immunization with chicken erythrocytes (CRBC) in mice. Cytotoxicity of spleen cells was assessed by the release of 51Cr from labelled target cells. (1) At early stages (day 4-7) after primary intraperitoneal immunization, direct cytotoxicity of spleen cells was raised efficiently in C57BL/6 and AKR mice, but not in C3H/He, SL and DDD mice. Delayed hypersensitivity and antibody production were raised to almost the same extent in all the strains at such periods. (2) Effector cells in direct cytotoxicity were theta-positive and IgG-positive, and glass-nonadherent and Nylon wool column-adherent. Effector cells in antibody-dependent cell-mediated cytotoxicity in the presence of antibody to CRBC were eliminated by treatment with anti-IgG serum, but not by treatment with anti-theta serum. (3) Cytotoxicity and antibody production were raised efficiently after intraperitoneal or intravenous immunization, but not after footpad immunization. On the other hand, delayed hypersensitivity developed most efficiently after footpad immunization.     

The effect of inhibitors of macromolecular synthesis on neutrophil-mediated antibody-dependent cellular cytotoxicity: exacerbation by mitomycin C

Previous work correlating 51chromium release with aggregate formation between neutrophils and target cells had indicated that unlike killing by lymphocytes, neutrophil-mediated antibody-dependent cellular cytotoxicity (NADCC) in the guinea-pig against hepatoma cells required that three or more neutrophils were simultaneously attached to the target cells. There was virtually no lysis when only one or two neutrophils were attached. This raised the possibility that the target cells could repair lesions inflicted by one or two cells and this was investigated using inhibitors of macromolecular synthesis. Actinomycin D, cycloheximide, puromycin and adriamycin had no effect--militating against a repair mechanism. Surprisingly, mitomycin C led to an exacerbation of NADCC. Chlorambucil, which like mitomycin C is an alkylating agent, had a similar effect though less marked. One explanation of this phenomenon which seemed likely was that the drugs were affecting the cell cycle of the target cells, i.e. that being alkylating agents they might be holding the target cells in a susceptible phase of the cycle. Experiments in which the target cells alone were preincubated with mitomycin C for varying periods of time (2-18 h) before the start of the assay, did not confirm this. However a thirty minute preincubation with mitomycin C of either the neutrophils or neutrophils plus the target cells resulted in increased cytolysis. The results indicate that the effect of mitomycin C in exacerbating NADCC is primarily on the neutrophil.     

Resistance of Enterococcus faecium to neutrophil-mediated phagocytosis.

During a previous study of the opsonic requirements for neutrophil (polymorphonuclear leukocyte [PMN])-mediated killing of enterococci, we identified two strains of Enterococcus faecium (TX0015 and TX0016) that were resistant to PMN-mediated killing. To better define the mechanism of this resistance, we examined phagocytosis with a fluorescence assay and found that TX0016 was completely resistant to phagocytosis by PMNs; this finding was confirmed by electron microscopy. Examination of multiple strains of enterococci revealed that all 20 strains of Enterococcus faecalis tested were readily phagocytosed (mean, 18 intracellular organisms per PMN; range, 7 to 28). In contrast, only 13 (50%) of 26 strains of E. faecium tested were susceptible to phagocytosis (> or = 7 organisms per PMN); the other 13 strains showed < or = 3 organisms per PMN. Enterococcus casseliflavus ATCC 25788 and one strain of Enterococcus hirae were also resistant to phagocytosis, while two strains of Enterococcus durans, Enterococcus mundtii ATCC 43186, and one strain each of Enterococcus raffinosus and Enterococcus solitarius were readily phagocytosed. Exposure of E. faecium TX0016 to sodium periodate, but not to the protease trypsin or pronase or to phospholipase C, eliminated resistance to phagocytosis. Sialic acid, a common periodate-sensitive structure used by microorganisms to resist opsonization, could not be demonstrated in E. faecium TX0016 by the thiobarbituric acid method, nor was phagocytosis of TX0016 altered by neuraminidase treatment. This study suggests that there is a difference in susceptibility to phagocytosis by PMNs between different species of enterococci and that a carbohydrate-containing moiety which is not sialic acid may be involved in the resistance of E. faecium TX0016 to phagocytosis.     

T lymphocyte induction of non-T cell-mediated nonspecific cytotoxicity. I. Introduction mechanisms.

Mononuclear cells (MNC) from normal humans consistently failed to give nonspecific cytotoxic responses. However, after removal of T cells by sheep erythrocyte (E) rosetting, the remaining non-RFC (rosette-forming cells) now gave significant nonspecific cytotoxic responses against both autologous and allogeneic target cells. Reconstitution experiments with T cell subpopulations failed to suppress these nonspecific non-E-RFC-mediated cytotoxic responses. There was also no evidence to indicate the involvement of antibody in this nonspecific cytotoxicity. The cytotoxic cells were characterized as non-E-rosetting, non-phagocytic, and glass adherent lymphocytes; no evidence of monocyte-macrophage participation was found. The inductive trigger of non-E-RFC-mediated cytotoxicity was found to be soluble factors released by T cells during E-rosette formation at 4 degrees C. Incubation of MNC with horse, marmoset and human erythrocytes under identical conditions failed to trigger cytotoxicity. The incubation of quiescent MNC with E-rosetting supernatants (ERS) induced nonspecific cytotoxic responses equivalent to those mediated by separated non-E-RFC. ERS-activated MNC destroyed both autologous and allogeneic target cells. The ERS supernatants themselves were not cytolytic. These findings suggested that cell separation procedures, and possibly in vivo events, which activate T cells may also induce non-T cell-mediated nonspecific cytotoxicity.     

Cell-mediated cytotoxicity against ectromelia virus-infected target cells. I. Specificity and kinetics

Spleen cells from mice immunized intravenously with attenuated ectromelia virus were cytotoxic for target cells infected with virulent ectromelia virus in the absence of exogenous complement; cytotoxicity was measured by ⁵¹Cr release from target cells. L929 cells were the most sensitive target cells, but the effect could also be demonstrated with P-815 mastocytoma cells and chick embryo cells; mouse embryo cells were unsatisfactory. Hyperimmune anti-ectromelia serum and complement was not significantly cytotoxic although fluorescein-conjugated antiserum stained virus-infected L929 cells. Release of ⁵¹Cr label from ectromelia-immune spleen cells (at a spleen cell: target cell ratio of 100: 1) increased in a linear manner with time, until a plateau was reached at 20–24 hours. Cytotoxicity appeared to be specific in that ectromelia-immune spleen cells killed ectromelia-infected L929 cells, but not uninfected L929 cells, whereas spleen cells from mice immunized with Listeria monocytogenes (a potent stimulus for cell-mediated immunity) or from normal mice, did not kill ectromelia-infected L929 cells. Spleen cells from mice immunized with vaccinia virus (a poxvirus closely related to ectromelia) were cytotoxic for ectromelia-infected L929 cells. Spleen cells from mice immunized with ectromelia virus had acquired cytotoxic activity by 2 days after immunization. Cytotoxic potency reached a peak at 6 days and had declined to a low level by day 10. The potential significance of the phenomenon in relation to control of viral infection is discussed.     

Prevention of neutrophil-mediated injury to endothelial cells by perfluorochemical.

Myocardial salvage after reperfusion may be limited by neutrophil-mediated microvascular damage. The effect of the perfluorochemical, Fluosol-DA, and its various components on neutrophil adherence, cytotoxicity, and proteolytic enzyme release was examined on sheep large and small vessel endothelial cells in vitro. Cells were studied under normoxic (N) and anoxic conditions (A). Various concentrations of Fluosol (10%, 25%, and 50%) significantly reduced neutrophil adherence under both experimental conditions [mean 22 +/- 3.25% versus 7 +/- 0.8% (N) and 20 +/- 3.2% versus 7.5 +/- 0.9% (A); P less than 0.01]. The perfluorocarbons, perfluorodecalin (PFD), and perfluoro-tripropylamine (PFTP) in a 50 volume/percent concentration exhibited profound effects on adherence, particularly on cells subjected to anoxia (51% and 69% reduction in adherence, respectively; P less than 0.01). No effect on adherence was observed with other components, including the detergent, pluronic F68. A 25% reduction (P less than 0.02) in endothelial cytotoxicity was noted when neutrophils were preincubated with Fluosol. However, pretreatment of endothelial cells with Fluosol did not inhibit neutrophil adherence. Neutrophils stimulated with cytochalasin B and FMLP showed a significant reduction in lysozyme release after incubation with Fluosol (28 +/- 5% versus 17 +/- 4%; P less than 0.01). This study demonstrates that Fluosol significantly attenuates neutrophil adherence, cytotoxicity, and enzyme release in an in vitro model of microvascular injury. It also suggests that prevention of neutrophil-mediated microvascular damage may be an important mechanism whereby Fluosol enhances myocardial salvage after ischemia and reperfusion.     

Characteristics of the effector cells mediating cytotoxicity against antibody-coated target cells. I. Phagocytic and non-phagocytic effector cell activity against erythrocyte and tumour target cells in a 51Cr release cytotoxicity assay and [125I]IUdR growth inhibition assay.

Both phagocytic and non-phagocytic effector cells were able to kill rabbit antibody-coated chicken erythrocytes (CRBC) while only non-phagocytic effector cells were active against alloantibody-coated SL2 lymphoma. In addition to the variation in susceptibility of erythrocyte and tumour target cells to various effector cell populations, it was found that different tumour cells can vary markedly in their ability to be killed by non-immune spleen cells in the presence of antibody. It is postulated that both the type of antibody and certain characteristics of the cell membrane are important in determining whether target cells are susceptible to antibody-dependent cell-mediated cytotoxicity detected by the 51Cr release assay. It was also demonstrated that alloantibody-coated P-815-Y mastocytoma, which showed very little evidence of cytotoxicity in the 51Cr release assay, was markedly inhibited in its ability to incorporate [125I]IUdR after incubation with antiserum and non-immune spleen cells. This growth inhibition in the absence of cytotoxicity, or cytostasis, is discussed in relation to the potential mechanisms of target cell damage, and in the light of recent observations (Plata, Gomard, LeClerc and Levy, 1974; Newlands and Roitt, 1975) that cytotoxicity and growth inhibition assays detect different effector cell populations in tumour-bearing animals.