Differential in vitro modulation of suppressor and antitumor functions of mouse macrophages by lymphokines and/or endotoxin.

Peritoneal macrophages of normal mice exhibited natural suppressor activity, as indicated by their ability to inhibit the proliferation of spleen cells in response to stimulation with phytohemagglutinin (PHA) or concanavalin A (Con A). Their suppressor function could be modulated in vitro with a variety of treatment regimens. High-dose lipopolysaccharide (LPS) (LPSH; 10 micrograms/ml) or lymphokines (supernatant from Con A-stimulated spleen cells) plus low-dose LPS (LPSL; 10 ng/ml) caused a reduction in the suppressor activity of adherent peritoneal macrophages. In contrast, these same treatments induced the macrophages to become tumoricidal and cytostatic for tumor cells, indicating a major dissociation between the regulation of suppressor and cytotoxic activities of macrophages. The lack of correlation between these activities was further demonstrated by macrophages that had been activated in vitro by Corynebacterium parvum: these cells expressed high tumoricidal and cytostatic activities, and also strong suppressor activity. The suppressor function could be selectively downregulated by in vitro pretreatment with LPSH.

Reactivity of anti-asialo GM1 serum with tumoricidal and non-tumoricidal mouse macrophages.

All peritoneal macrophage (pM phi) populations studied exhibited some binding of the anti-asGM1 serum as assessed by flow cytometry. The levels of reactivity varied quantitatively among populations, depending on the combination of eliciting and activating agents employed prior to the harvest of pM phi. Resident pM phi contained a very small percentage (4%) of cells that were strongly asGM1+. Any treatment of these cells that induced them to become stimulated or activated increased the percentage of highly asGM1+ cells. Treatments that enhanced anti-asGM1 binding including eliciting pM phi with proteose peptone (16% asGM1+) or Brewer's thioglycollate medium (66% asGM1+), treatment with the activating biological response modifiers (BRMs) MVE-2 (12% asGM1+) and P acnes (18% asGM1+), or treatment with both peptone + MVE-2 (37% asGM1+) or peptone + poly IC/LC (33%). Increased expression of anti-asGM1 was accompanied by some increase in the reactivity of the various pM phi populations to treatment with anti-asGM1 serum. This conclusion was based on the reduced viabilities of cells treated with both an eliciting agent and an activating agent prior to in vitro treatment with anti-asGM1 + C, as well as by reductions in cytolytic activity of pM phi elicited with peptone and activated by MVE-2, following anti-asGM1 treatment in vitro or administration in vivo. Conversely, the cytolytic activity of resident pM phi activated in vivo by MVE-2 or heat-killed P acnes, agents that induced relatively small increases in the percentage of asGM1+ cells, was resistant to the effects of in vivo and/or in vitro treatment with doses of anti-asGM1 serum that inhibit NK activity. These results indicate that stimulation of pM phi by eliciting or activating agents can increase the level of expression of asGM1. This increased expression of asGM1 may be a useful marker for some aspects of macrophage heterogeneity, but increased expression is not necessarily directly related to expression of tumoricidal activity. In fact, the results of this study demonstrate that anti-asGM1 serum can be used for specific depletion of NK activity in vivo in normal mice and in mice treated with at least some BRMs. However, the results also demonstrate that the use of eliciting agents, particularly thioglycollate, or eliciting agents in conjunction with activating agents can cause pM phi to become reactive with anti-asGM1 serum.

Distribution of peritoneal macrophage populations after intravenous injection in mice: differential effects of eliciting and activating agents.

Murine peritoneal macrophages (pM phi) elicited in vivo by intraperitoneal (IP) inoculation of various agents were tested for their homing/distribution patterns after intravenous (IV) adoptive transfer to syngeneic C57BL/6 recipients. Resident pM phi (RpM phi) obtained from normal mice and pM phi elicited by proteose peptone (PpM phi) or thioglycollate broth (TpM phi) exhibited similar homing patterns following IV transfer. After initial arrest in the lungs, these cells rapidly disseminated to liver and spleen, with minimal or no detectable migration to peripheral lymph nodes, intestine, peritoneum, kidney, heart, or retention in the blood. The pattern of results reflected the properties of pM phi themselves, since highly enriched pM phi populations obtained by treatment of crude peritoneal exudate cells with anti-Thy 1.2 + C, or by fractionation on Percoll density gradients, gave similar results. The distribution of pM phi elicited by Brewer's thioglycollate medium (BTpM phi) was markedly different from other pM phi tested. BTpM phi homed rapidly to the lungs and many remained localized there for at least 72 hr with very little migration to the spleen. The distribution of PpM phi could be altered by activation of these cells in vivo through the IP injection of the pyran copolymer, MVE-2, prior to adoptive IV transfer. Activated PpM phi contained a population of highly differentiated, low density pM phi, separable on density gradients, which arrested in the lungs for appreciably longer periods of time than did PpM phi. These cells exhibited reduced ability for migration to the spleen. Macrophage-like (M phi-like) cell lines did not exhibit migration capability, but rather were rapidly cleared from the circulation in a manner similar to other types of tumor cells.

Suppression of in vitro maintenance and interferon-mediated augmentation of natural killer cell activity by adherent peritoneal cells from normal mice.

The ability of adherent peritoneal cells (APC) to inhibit murine natural killer (NK) cell activity was examined. Nylon wool-nonadherent splenic effector cells were incubated overnight with or without different numbers of APC. NK activity was then measured against YAC-1 in a 4-hr 51Cr-release cytotoxicity assay. Proteose peptone-elicited or unstimulated resident APC from normal mice markedly suppressed NK activity of splenic effector cells in the presence or absence of exogenously added interferon. The suppression was dependent on the number of APC added with 10% APC, relative to the number of effector cells, resulting in a greater than 65% inhibition of cytotoxicity. The effector phase of cytotoxicity was not the target of the suppressor cells, because APC did not suppress NK activity when they were present only during the cytotoxicity assay. The addition of APC to alloimmune cytotoxic T cells under similar conditions resulted in no inhibition of cytotoxicity. Both syngeneic and allogeneic APC suppressed NK activity, but several murine macrophage-like cell lines lacked this property. In contrast to APC, incubation of effector cells with adherent spleen cells from normal mice resulted in no inhibition of NK activity. APC from mice injected with C. parvum were less inhibitory for NK activity than normal resident APC. In contrast, C. parvum APC suppressed concanavalin A-induced lymphoproliferation and were directly cytotoxic to tumor target cells in vitro, whereas normal APC lacked these properties. The results indicate that the peritoneum of untreated mice contains suppressor cells that can inhibit the in vitro maintenance and IFN-mediated augmentation of NK activity. In addition, these results indicate a broader spectrum of immune reactivities regulated by APC and suggest that, depending on their level of activation, APC can preferentially inhibit different immune functions.

Selective inhibition by monosaccharides of tumor cell cytotoxicity mediated by mouse macrophages, macrophage-like cell lines, and natural killer cells.

A series of monosaccharides were tested for their ability to inhibit the effector phase of macrophage-mediated cytolysis against two susceptible murine tumor target cells, L5178Y and RL male I. Two monosaccharides, D-mannose and N-acetyl-D-galactosamine, were found to decrease cytotoxicity consistently in a dose-dependent manner. However, D-mannose preferentially inhibited lysis of RL male I target cells with little effect on lysis of L5178Y target cells, while the reverse was found with N-acetyl-D-galactosamine. Neither monosaccharide interfered with the activation of macrophages by polyinosinic:polycytidylic acid. Natural killer cell activity was decreased by a 25 mM concentration of D-mannose but not by N-acetyl-D-galactosamine, although increasing concentrations of N-acetyl-D-galactosamine were inhibitory. Neither monosaccharide affected cytotoxicity by alloimmune T cells. Cytotoxicity of macrophage-like cell lines against tumor target cells was also decreased by monosaccharides but the pattern of inhibition was different from that seen with activated macrophage effector cells. Both D-mannose and N-acetyl-D-galactosamine inhibited glucose oxidation by activated macrophages but only D-mannose significantly decreased protein synthesis of activated macrophages. These results indicate that monosaccharides can inhibit macrophage-mediated cytotoxicity in a selective manner with the pattern dependent on the tumor target cell used in the assay.

Assessment of in vivo natural antitumor resistance and lymphocyte. Migration in mice: comparison of 125I-iododeoxyuridine with 111indium-oxine and 51chromium as cell labels.

Clearance of IV-injected tumor cells has been correlated with levels of natural killer (NK) cell activity in recipient animals. Studies of in vivo tumor cell clearance strongly suggest a relationship between levels of NK cell activity and antitumor or antimetastatic effector function. This study outlines the applicability of three radiolabels, [125I]iododeoxyuridine, ( [125I]dUrd), indium-111-oxine chelate ( [111In]Ox), and chromium-51 (51Cr), to studies of tumor cell clearance in vivo. The suitability of these labels for analysis of the in vivo migration patterns of normal lymphocytes or thymus-derived T cells cultivated in vitro (CTC) is also discussed. The results indicate that [111In]Ox and 51Cr compare favorably with the more widely used [125]dUrd as radiolabels for the assessment of IV-injected tumor cell clearance from the lungs of mice. The rates of clearance of both [111In]Ox and 51Cr, like that for [125I]dUrd, correlate closely with levels of NK-cell activity of the host. Further studies with [111In]Ox reveal that treatment of recipients with anti-asialo GM1 serum, a regimen known to suppress NK-cell activity, demonstrates the appropriate reduction in isotope clearance from the lungs after NK suppression. However, clearance data obtained by monitoring levels of radioactivity in the liver after IV injection must be viewed cautiously, since the same cells labeled with [111In]Ox and [125I]dUrd had a different pattern of clearance from the liver. The same inconsistencies in clearance were observed when [111In]Ox and [125I]dUrd were injected intrafootpad (i.f.p.). Similar effects were observed when [111In]Ox or 51Cr was applied to studies of CTC migration. Levels of [111In]Ox and 51Cr remained high in the liver after IV injection, while [125I]dUrd was rapidly cleared. Normal spleen or thymic lymphocytes exhibited the expected homing to the spleen after labeling with [111In]Ox, indicating a suitability of this label for migration studies, except possibly in the liver. These results with CTC and normal lymphocytes should be considered during the formulation of immunotherapy protocols based on cell migration data, since the choice of radiolabel can result in widely divergent levels of radioactivity accumulated in some organs, and may not provide an accurate representation of the presence of viable, intact, or functional cells.

Macrophage cell lines behave as activated macrophages in the production and regulation of plasminogen activator.

Seven macrophagelike cell lines, WEHI-3, J774, RAW264, FC-1, P388D-1, PU5-R, and PU5-1.8, commonly used as in vitro models of macrophage function, were studied for their fibrinolytic activity in the presence and absence of purified plasminogen. All cell lines, as well as the freshly harvested peritoneal macrophages, produced plasminogen activator, as indicated by the levels of fibrinolytic activity in the presence of plasminogen. However, fibrinolytic proteolysis was most efficiently measured by direct attachment to 125I-fibrin plates, a method which measures both secreted and cell-membrane-associated plasminogen activator. Other non-plasminogen-dependent fibrinolytic activities were also produced by all cells studied and secreted by WEHI-3, FC-1, RAW-264, and thioglycollate-elicited macrophages. Activation of the thioglycollate-elicited macrophages with macrophage activation factor (MAF) induced threefold higher levels of plasminogen activator production but MAF treatment had no effect on the level of activity produced by macrophagelike cell lines. Indeed, the macrophagelike cell lines produced plasminogen activator at levels seen with MAF-activated thioglycollate-elicited macrophages. We conclude that these cell lines constitutively produce activators of plasminogen and, in their failure to respond to MAF, behave as activated macrophages.

Variation in selectivity of tumor cell cytolysis by murine macrophages, macrophage-like cell lines and NK cells.

Several murine tumor-cell lines were tested by isotope release assays for their susceptibility to lysis by either activated peritoneal macrophages (apMPh), macrophage-like (MPh-like) cell lines, or natural killer (NK) cells. The qualitative selectivity of tumor-cell lysis by these different effector cells was quite disparate. The rank order of target cell susceptibility to lysis by apMPh in 24 h assay was L5178Y greater than P815 approximately equal to RL male greater than YAC-1 approximately equal to MBL-2. This was seen regardless of whether peritoneal MPh (pMPh) were activated by LPS or poly I:C. Two MPh-like cell lines, PU-5R and FC-1, had a pattern of cytotoxic activity against these target cells that closely paralleled that associated with apMPh, although levels of reactivity differed quantitatively among the effector cells. In contrast, the MPh-like cell line RAW-264 expressed a qualitatively different pattern of target-cell selectivity, preferentially lysing MBL-2, which was relatively refractory to lysis by other MPh-like cell lines or apMPh in the 24 h cytolytic assay. When spontaneous or interferon (IFN)-augmented NK activity was measured against the same panel of target cells, the pattern of selectivity was qualitatively different from that observed for apMPh. The consistent rank order of susceptibility to lysis by NK cells was YAC-1 greater than RL male 1 greater than P815 approximately equal to L5178Y approximately equal to MBL-2. The characteristic patterns of target-cell selectivity for apMPh or NK cells were the same for all of the strains of mice tested. From the different selectivity patterns of apMPh and NK cells, it is concluded that lysis of target cells is not based solely on inherent sensitivity to cytolysis. Instead, selectivity of lysis is probably due to variations in expression of target-cell structures recognized by each type of effector cell, and/or in susceptibility to the lytic mechanism(s) of the various effector populations.

Natural killer cell activity in the rat. Analysis of effector cell morphology and effects of interferon on natural killer cell function in the athymic (nude) rat.

Athymic (nude) rats were found to have increased levels of natural killer (NK) activity, 3- to 5-fold higher than in euthymic rats. Studies were performed to determine the nature of the NK cells in these animals and the basis for their increased cytotoxic reactivity. Large granular lymphocytes (LGL), which were previously shown to be the NK cells in euthymic rats, were increased 2- to 7-fold in the peripheral blood and spleen of nude rats. The LGL, enriched by centrifugation on discontinuous Percoll density gradients, were shown to have augmented NK activity similar to that seen with LGL-enriched fractions from euthymic rats. These results indicate that the NK cells in euthymic and athymic rats are morphologically and functionally similar, and that the higher NK activity in nude rats appears to be mainly attributable to an increased proportion of effector cells. In a single-cell cytotoxicity assay, interferon pretreatment of LGL was shown to increase: (a) the percentage of LGL which form conjugates with target cells; (b) the percentage of conjugate-forming cells which kill; and (c) the kinetics of lysis. Different effects were seen depending on the target cell tested.

Augmentation of metastasis formation by thioglycollate-elicited macrophages.

Inoculation of thioglycollate-elicited peritoneal exudate cells (PEC) into C57BL/6 mice reduced the rate of lung clearance of several intravenously (i.v.) injected murine tumor cells, and increased by up to 100-fold the number of artificially-induced metastatic lung nodules produced by the i.v. injection of B16 melanoma or Lewis lung carcinoma (3LL) tumor cells. Maximum effects were observed when PEC were injected either before, or shortly after, tumor cells. Modulation of lung clearance or metastasis formation was observed only with PEC and not with a variety of other cells, such as splenocytes, thymocytes P815 mastocytoma cells, or several macrophage-like cell lines (PU5-1.8 and IC-21). Lysates of PEC were as efficient in reducing lung clearance and augmenting metastasis formation as were intact viable PEC. Lysates of other cell types, including P815 and the macrophage-like cell lines, were unable to produce these effects. PEC populations, enriched for macrophages by adherence to plastic or by percoll density gradient sedimentation, also increased the number of B16-induced artificial metastasis, implicating the macrophage as the cell responsible for these observations.

Anti-tumor immunity in B-lymphocyte-deprived mice. III. Immunity to primary Moloney sarcoma virus-induced tumors.

Tumor induction and immunity to tumors were studied following the injection of Moloney sarcoma virus (MSV) into mice whose B-lymphocyte functions had been suppressed by the chronic administration of anti-IgM antibodies. Two preparations of MSV were used; one which gives rise to tumors which uniformly regress in normal adult mice, and another which elicits progressively growing tumors in the majority of recipients. The tumor incidence, mean tumor size, and tempo of regression were not modified by treatment with anti-IgM. However, whereas tumors induced by the regressor virus were all rejected in 19 NRG-treated and 29 untreated recipients, continued growth was obtained in 2 of 23 B-lymphocyte-deprived mice. Furthermore, in 9 additional mice from this group, apparent rejection was followed by tumor recurrence at the site of the initial tumor. Continued growth was accompanied by widespread metastasis. These tumors were freely transplantable to normal syngeneic recipients. Metastasis and transplantability were also detected in 7 of 24 anti-IgM-treated mice given progressor virus, but were not seen in the control animals. Recurrence and metastasis was obtained despite the presence of high levels of specific cytotoxic T lymphocytes in the spleen. It is concluded that B lymphocytes or their products play an essential role in host protection against MSV-induced tumors.

In vitro induction of tumoricidal and suppressor macrophages by lymphokines: possible feedback regulation.

Proteose-peptone-induced macrophages (pM phi) from C57BL/6 (B6) mice were treated in vitro with lymphokine, and 2 functions were evaluated: the ability to suppress lymphokine production and the cytotoxic capacity against tumor target cells. When 10% to 20% pM phi treated with lymphokine were added to a lymphokine-producing system, strong inhibition of MIF and MAF production occurred. The suppression was not specific, since pM phi inhibited the production of lymphokine obtained by either mitogenic stimulation of normal spleen cells (NSC) or alloantigen stimulation of immune spleen cells (ISC). The suppressor cells were adherent, phagocytic, Thy 1.2 negative, with monocyte-macrophage-like morphology. Lymphokine-treated pM phi were also tumoricidal when tested in 18 hr microcytotoxicity assay against RL male 1 lymphoma target cells. However, using endotoxin-free reagents (less than or equal to 0.01 ng/ml of endotoxin by the LAL assay), we found that small amounts of lipopolysaccharide (LPS) were required, in addition to lymphokine, to induce tumoricidal activity in pM phi. In contrast, noncytotoxic pM phi stimulated with lymphokine alone were also able to inhibit lymphokine production, although not as effectively as pM phi stimulated by lymphokine plus LPS. Therefore, we concluded that lymphokine treatment itself is sufficient to induce M phi to become suppressor cells, whereas an additional signal is necessary to induce cytolytic activity. We suggest that there is a negative feedback mechanism of control on the lymphokine-producing cells through the activation of suppressor M phi by lymphokine.

Role of macrophage in in vitro augmentation of rat, mouse, and human natural killer activities.

Requirement for macrophages in in vitro augmentation, by interferon (IFN), polyinosinic-polycytidylic acid (poly I:C), or Corynebacterium parvum, of rat, mouse, and human natural killer (NK) activities was examined. Several differences were seen among the species. Mouse NK activity demonstrated some lability at 37 degrees C and a strict macrophage requirement for in vitro production of IFN, and augmentation of NK activity was demonstrated by either poly I:C or C. parvum. In contrast, human peripheral blood leukocytes (PBL) depleted of monocytes by adherence on nylon wool demonstrated NK activity, which was not labile but rather increased substantially upon overnight culture at 37 degrees C alone or with poly I:C or C. parvum. Monocyte-depleted human PBL also produced IFN in these cultures. The pattern of reactivity seen with rat spleen cell cultures was different from either that of mouse and human cells. This pattern of reactivity had no lability at 37 degrees C and had a macrophage requirement for IFN production and NK cell augmentation upon culture alone or with poly I:C but not with C. parvum. These results indicated some major differences among species in the regulation of NK activity in vitro and the requirement for macrophages for the in vitro production of IFN. A better understanding of these differences will be helpful in choosing appropriate models for in vitro and in vivo studies of NK cell activity.

The distribution of Helix pomatia lectin receptors on mouse lymphoid cells and other tissues.

The distribution of receptors for the Helix pomatia lectin on mouse lymphoid cells and other tissues was investigated. Using a sensitive rosetting assay combined with immunofluorescence, lectin receptors were found on the membrane of approximately 90% of peripheral T lymphocytes, 75% of thymocytes, 30% of bone marrow cells, 20% of nude spleen cells, 15-50% of peritoneal exudate macrophages, and a subpopulation of peritoneal exudate mast cells. The Thy-1-positive nude spleen cells were predominantly Helix lectin receptor-negative. Approximately 5% of B lymphocytes were weakly positive, and neutrophils were negative. Receptors were present also on a subpopulation of cells of a fibroblast cell line and in acetone powder from the liver and, at a lower level, from the kidney and brain. Membrane receptors on all cell types were partially detectable without neuraminidase treatment of the cells. Two methods of fractionating Helix lextin-positive cells were employed, which gave significantly different results. By rosetting and depletion using density fractionation, T cell mitogen responses were abolished, while B cell mitogen responses, T cell cytotoxicity, and natural killer cytotoxicity were only slightly affected, if at all, Helix lectin-agarose column fractionation seemed more sensitive, in that essentially all natural kill cells bound to the column, as well as considerable number of B lymphocytes. Cytotoxic T cells were heterogeneous; roughly half were not bound, but the remainder were bound and eluted.

Antibody-induced augmentation of murine natural killer cell activity.

Antibodies reactive with effector cells were shown to augment the cytotoxicity of spleen cells from athymic nude and euthymic mice. The addition of alloantibody to the assay or pretreatment of the effector cells with alloantibody resulted in increased cytotoxicity against the human cell K562, a relatively poor target for spontaneous mouse NK activity. When monoclonal antibodies were tested, cytotoxicity was markedly increased by some antibodies, such as anti-H-2, anti-la, and anti-Thy 1.2, while others had no effect. The degree of augmentation of cytotoxicity was dependent on the concentration of antibody added. Nylon-wool-nonadherent nu/nu splenic effector cells mediated the antibody-induced cytotoxicity and anti-asialo GMI plus complement abolished activity, indicating that the cells mediating the cytotoxicity were NK cells and not mature T cells, B cells or macrophages. When spleen cells from mice having different levels of NK activity were evaluated in this system, the magnitude of augmentation by antibody correlated with the level of spontaneous NK activity and no increased cytotoxicity was found with cell populations that had low spontaneous NK activity. Testing a panel of target cells, showed that certain human and mouse cell lines, with low to moderate susceptibility to spontaneous NK activity, were sensitive to antibody-induced NK-cell-mediated cytotoxicity whereas others were completely resistant. Both Fc-IgG receptor-positive and -negative cell lines were susceptible target cells. These results indicate that antibodies reactive with murine NK cells can increase their cytolytic activity.

Measurement of macrophage-mediated cytotoxicity against adherent and non-adherent target cells by release of 11 indium-oxine.

This report describes the utilization of 111 indium-oxine chelate ([111In]Ox) for studies of macrophage-mediated cytotoxicity. [111In]Ox efficiently labeled both non-adherent and adherent tumor targets with no decrease in cell viability. Spontaneous release of intracellularly incorporated [111In]Ox was very slow (0.25-0.50%/h) from most targets, making isotope-release assays of at least 48 h feasible. In addition, released [111In]Ox was not reutilized. In contrast to its low spontaneous release from intact cells, incorporated [111In]Ox was rapidly released from tumor targets after interaction with activated macrophages. Levels of [111In]Ox released in response to cytolytic macrophages correlated well with those observed for the 51Cr and [3H]TdR radiolabels. Therefore, [111In]Ox can be utilized for relatively short-term (less than 20h) assays with lymphoma targets, as well as for longer-term assays with adherent cells. This should facilitate the testing, with the same radioisotope-release assay, of a wide range of tumor targets for susceptibility to macrophage-mediated cytotoxicity.