BCG-activated Macrophages Research Articles

Cytolytic interactions between murine macrophages, tumor cells, and monoclonal antibodies: characterization of lytic conditions and requirements for effector activation.

Because of recent successes in inducing the effective rejection of neoplasms in vivo by administration of monoclonal antibodies (MAb), we analyzed lytic interactions in vitro that occur between macrophages and several combinations of tumor targets and MAb that can lead to such successful immunotherapy. Murine macrophages, interacting with MAb of the IgG1, IgG2a, IgG2b, and IgG3 isotypes directed against SW-1116 carcinoma cells, destroyed the tumor targets efficiently over 24 to 48 hr in vitro. Lysis was dependent on both concentration of the MAb and density of the macrophages. Binding and lysis of the targets in the presence of MAb of the IgG2a isotype was dependent on intact Fc gamma 2aR on the macrophages; target binding was necessary but not sufficient for subsequent lysis. The lytic step appeared to have an oxidative basis, at least in part, as shown by inhibition of lysis with a nonspecific scavenger of H2O2 or under either anaerobic or glucose-deprived conditions. TG-elicited or pyran-elicited macrophages, which are incompletely activated for antibody independent kill of tumor cells, were effective in mediating ADCC. By contrast, macrophages fully activated for direct cytolysis by administration of BCG or Propionibacterium acnes in vivo or by MAF and LPS in vitro, had diminished capacity for ADCC. A spectrum of five other tumor cells and antibodies, four of which are also involved in successful models of immunotherapy in vivo, were also killed over 48 hr more effectively by thioglycolate-elicited than by BCG-activated macrophages. Taken together, the data indicate that macrophages can lyse tumor cells in an ADCC reaction that has application to some models of the destruction of tumors in vivo, but that the lysis is slow and requires the macrophages to be activated in a specific way(s).

Interaction of West Nile virus with primary murine macrophages: role of cell activation and receptors for antibody and complement.

We have measured growth of West Nile virus in mouse primary peritoneal macrophages (resident, thioglycolate elicited, and Mycobacterium bovis BCG activated) and in macrophagelike (P338D1) and nonmacrophage (L929, PS clone D) cell lines infected in the absence or presence of specific antibodies (immunoglobulin G ([IgG], IgM), and complement. Monoclonal antibodies directed against Fc receptors (IgG1/2b, 2.4G2) and type 3 complement receptors (Mac-1) were used to define the role of each receptor. Virus yield depended on a balance between enhancement and neutralization and was influenced by the physiologic state of the macrophage, the receptor pathway of viral entry, the mouse strain and age of donor. BCG-activated macrophages displayed a greater ability to restrict West Nile virus than nonactivated cells only in the presence of antiviral IgM, with or without complement; the Fc receptors for various classes of IgG mediated striking enhancement. These studies identify some of the complex innate and acquired factors that determine the interaction between West Nile virus and primary macrophages in vitro.

Activation of mouse peritoneal macrophages alters the structure and surface expression of protein-bound lactosaminoglycans.

We have begun to analyze and compare the surface carbohydrates present on populations of resident and activated mouse peritoneal macrophages. The activated macrophage populations studied include TG-elicited macrophages, BCG-activated macrophages, and resident macrophages cultured for 24 hr in the presence of lymphokines or heterologous serum. Analysis of glycopeptides generated by pronase digestion of surface glycoproteins labeled by the neuraminidase/galactose oxidase/NaB3H4 method indicates that the macrophage surface contains a class of high m.w. carbohydrates susceptible to degradation by endo-beta-galactosidase, lactosaminoglycans. These lactosaminoglycans are sialylated type 2 carbohydrates containing the repeating lactosamine disaccharide Gal beta 1-4GlcNAc as well as fucose residues. Macrophage activation was observed to markedly alter surface lactosaminoglycans. The alterations observed include 1) an increase in surface expression as determined by both an increase in neuraminidase/galactose oxidase/NaB3H4 labeling and by the ability of activated but not resident macrophages to bind I antibodies as assayed by indirect immunofluorescent surface staining, 2) the addition of alpha-galactose residues, and 3) an increase in GlcNAc beta 1-6Gal branching as indicated by an increased resistance to endo-beta-galactosidase degradation and by the ability of activated macrophages to bind I antibodies. These observations demonstrate that macrophage activation results in specific and substantial alterations in protein-bound surface carbohydrates.

Cell surface receptors for lactate dehydrogenase-elevating virus on subpopulation of macrophages.

We examined the binding and internalization of unlabeled and 125I-labeled, purified lactate dehydrogenase-elevating virus (LDV) by peritoneal macrophages cultured in vitro. Upon incubation of the cells at 4 degrees C with greater than 100 ID50/cell, the virus was surface-bound on a small subpopulation of macrophages (about 5% of the total cells) as determined by electron microscopy, fluorescent antibody staining, and autoradiography of cells incubated with 125I-labeled LDV. At 37 degrees C, LDV particles were seen in intracellular endocytic vesicles also in about 5% of the cells, and the proportion of cells with virus-containing vesicles correlated with the proportion of cells which became productively infected with LDV as assessed by determining LDV RNA synthesis in individual cells and by fluorescent antibody staining. Pretreatment of the resident peritoneal macrophages with trypsin inhibited the binding of 125I-labeled LDV and the productive infection of the cells with the virus. After removal of the trypsin and incubation in complete medium, permissiveness for LDV reappeared after an 8-12 h lag, whereas Fc and C3 receptors reappeared more rapidly after trypsin treatment. Populations of resident peritoneal macrophages, starch-elicited peritoneal macrophages, splenic macrophages, and bone marrow macrophages contained a similar proportion of cells that could be productively infected with LDV. Little, if any, LDV replication was detected in cultures of lung, liver and peripheral blood macrophages as well as in thioglycollate-elicited and BCG-activated macrophages. We conclude that the permissiveness for LDV of resident peritoneal macrophages correlates with the presence of a trypsin-sensitive receptor present on a subpopulation of these cells. The identity of the receptor has not been definitively established. Treatment of macrophages with neuraminidase or various sugars had no significant effect on LDV replication. Lysis of I-A-positive macrophages with a monoclonal antibody and complement reduced the number of macrophages which could be productively infected by 50%, which suggests that macrophages lacking surface Ia can be productively infected with LDV in vitro.

Glycolipids of the mouse peritoneal macrophage. Alterations in amount and surface exposure of specific glycolipid species occur in response to inflammation and tumoricidal activation.

We have characterized the major glycolipid constituents of the mouse peritoneal macrophage, and have demonstrated that alterations in the amount and in the accessibility of specific glycolipid species to galactose oxidase/NaB3H4 labeling, an indicator of glycolipid surface exposure, occur in response to inflammation and as a consequence of activation to a tumoricidal state. The key findings are: (a) Asialo GM1, a major neutral glycolipid constituent of all macrophage populations examined, is accessible to galactose oxidase/NaB3H4 labeling on the surface of TG-elicited and BCG-activated macrophages but not on resident macrophages; (b) GM1 is the predominant ganglioside constituent of the mouse macrophage. Resident macrophages contain two distinct GM1 species, as determined by cholera toxin binding, while TG-elicited and BCG-activated macrophages contain an additional GM1 species. Differences in the relative amounts of these GM1 species, as well as in their accessibility to galactose oxidase/NaB3H4 labeling, exist among the macrophage populations. These observations suggest that both a chemical and spatial reorganization of surface glycolipids occurs in response to inflammation and tumoricidal activation.

Killing of human malaria parasites by macrophage secretory products.

The susceptibility of the human malaria parasite, Plasmodium falciparum, to killing in vitro by macrophage secretory products was investigated. The effect of O2 radicals and tumor necrosis factor on parasite viability was assessed both morphologically and by following the uptake of [3H]hypoxanthine. H2O2 produced by the interaction of glucose and glucose oxidase was found to reduce viability; this effect was reversed by the addition of exogenous catalase. Further studies indicated that the catalase level within the erythrocyte was not altered upon parasite invasion. O2 radicals produced during the xanthine-xanthine oxidase interaction also killed P. falciparum. The addition of various O2 radical scavengers (including catalase) did not reverse this effect; therefore, it was not possible to determine which of the O2 radicals were involved in the killing process. Samples from three different sources containing tumor necrosis factor, a nonspecific soluble mediator derived from Mycobacterium bovis BCG-activated macrophages treated with endotoxin, also killed the parasite. There was no evidence that tumor necrosis factor or the products of the xanthine-xanthine oxidase interaction caused damage to the erythrocyte membrane that could be implicated as an important aspect of the killing process. These findings all strongly suggest that such macrophage products play an important role in immunity to malaria.

Modulation of apoprotein E secretion in response to receptor-mediated endocytosis in resident and inflammatory macrophages.

We have determined the effect of various endocytic ligands on the secretion of ApoE by macrophages. ApoE was a major secreted protein of resident macrophages, but BCG-activated macrophages secreted little ApoE and periodate-elicited macrophages secreted intermediate amounts of ApoE. Resident, periodate-elicited, and BCG-activated mouse peritoneal macrophages were incubated with AcLDL, EIgG, EIgMC, dextran sulfate, latex, or zymosan, and the resulting protein secretion patterns were analyzed by [35S]methionine labeling and SDS-polyacrylamide gel electrophoresis. AcLDL increased total [35S]methionine incorporation into secreted proteins. Although AcLDL increased the secretion of ApoE by resident macrophages less than or equal to fivefold in a dose-dependent manner, with maximal stimulation at 4.8 micrograms/ml, it decreased the secretion of ApoE by periodate-elicited macrophages to almost nothing and did not affect the low rate of secretion of ApoE by BCG-activated macrophages. However, EIgG, which increases cellular cholesterol content of macrophages as AcLDL does, did not increase ApoE secretion, and dextran sulfate, which is recognized by the same receptor as AcLDL, also did not increase ApoE secretion. The binding and uptake of EIgG, dextran sulfate, zymosan, latex, and EIgMC all decreased the secretion of ApoE. These endocytic ligands also altered the pattern of secreted and cellular proteins other than ApoE. The pattern of response was ligand-specific. However, increased secretion of polypeptides of Mr 62,000 and 68,000 was common to many stimuli. We conclude that receptor-mediated endocytosis modulates the secretion of ApoE and other proteins pleiotypically in resident, inflammatory, and activated macrophages.

The binding of tumor cells by murine mononuclear phagocytes can be divided into two qualitatively distinct types.

Peritoneal macrophages from C57BL/6N mice infected with BCG have a greater capacity for tumor cell binding than do inflammatory macrophages elicited by several agents. To determine if these quantitatively distinct types of binding were qualitatively distinct as well, we compared them in several ways. Scanning electron microscopy of macrophage-tumor cell interactions reveals clustering of 2 to 5 tumor targets on the central portion of BCG-activated macrophages, in contrast to the very few tumor cells found at the periphery of inflammatory macrophages. Tumor cell binding to BCG-activated macrophages required divalent cations and trypsin-sensitive surface structures on the macrophages, but the low-level binding to inflammatory macrophages required neither. The ability of trypsinized BCG-activated macrophages to bind P815 tumor targets was recovered with time in culture. BCG-activated macrophages required intact energy metabolism, microfilaments, and microtubules to complete augmented tumor cell binding, whereas inflammatory macrophages did not. Targets bound to BCG macrophages could be competitively removed by addition of large excesses of unlabeled tumor cells, but targets bound to inflammatory macrophages could not be so removed. Tumor cells labeled with 111Indium oxine were used to estimate the specific (i.e., competitively inhibitable) and nonspecific (noncompetitively inhibitable) components of binding. Binding of P815 tumor cells to BCG-activated macrophages was predominantly specific, whereas their binding to inflammatory macrophages was predominantly nonspecific. Binding by activated or inflammatory macrophages of lymphocytes or by transformed 3T3 cells of neoplastic or non-neoplastic targets was also nonspecific. Binding by BCG-elicited macrophages from A/J mice, previously demonstrated to be quantitatively deficient in binding ability, was qualitatively similar to that to inflammatory macrophages and was not competitively inhibitable. The data suggest that tumor cell binding by activated macrophages is qualitatively as well as quantitatively distinct from binding to inflammatory macrophages.

Liposome induction or enhancement of macrophage-mediated cancer cell lysis.

Liposomes of different composition have been used to modify macrophage-mediated destruction of syngeneic cancer cells through a modulation of membrane lipid content of macrophages and/or tumor cells. Dipalmitoylphosphatidylcholine (DPPC)1 liposomes induce cancer cell lysis by normal, non-tumoricidal, peritoneal macrophages and enhance tumor cell destruction by BCG-activated macrophages. This effect was produced by large and small unilamellar liposomes, which are in the 25,000 g supernatant of sonicated preparations. Addition of cholesterol or negative charges carried by dicetylphosphate supressed the effect of DPPC liposomes on macrophage-mediated cytolysis. Enhancement of macrophage-mediated tumor cell lysis was observed when both cancer cells and macrophages were incubated with DPPC liposomes, but not when macrophages and/or tumor cells were preincubated with liposomes prior to their coincubation. Liposomes did not promote the binding of the cancer cells to the macrophages. Liposomes could promote formation of phospholipid domains within the plasma membrane of both tumor cells and macrophages leading to the destruction of cancer cells through a temporary fusion with the macrophages.

Mechanism of BCG-activated macrophage-induced tumor cell cytotoxicity: evidence for both oxygen-dependent and independent mechanisms.

Activated peritoneal macrophages have been shown to be cytotoxic to cancer cells. Bacillus Calmette-Guerin (BCG)-activated rat peritoneal macrophages have a basal cytolytic potential for 3H-thymidine-labelled Walker 256 cancer cells in vitro that can be markedly enhanced by digitonin. This stimulation of cytotoxicity can be partially inhibited by catalase and the combination of superoxide dismutase plus catalase. This suggests that digitonin stimulates activated macrophages to produce superoxide, hydrogen peroxide and possibly other free radicals which can augment macrophage-induced tumor cell cytotoxicity. After a 2-hour incubation with digitonin, macrophages are no longer stimulated by digitonin. However, after a 2-hour drug preincubation period, inhibitors of serine protease activity (DFP, TLCK, SBTI) and inhibitors of protein synthesis (cycloheximide) are potent inhibitors of basal macrophage-induced tumor cytotoxicity. We suggest that BCG-activated macrophages have two mechanisms for destroying cancer cells: one mediated by proteolytic activity, and a second mechanism dependent on the generation of oxygen-derived free radicals.

The relationship between competence for secretion of H2O2 and completion of tumor cytotoxicity by BCG-elicited murine macrophages.

BCG-elicited macrophages from C3H/HeJ and A/J mice secrete H2O2 comparably to BCG-elicited macrophages from control C3H/FeJ mice in response to the pharmacologic stimulant phorbol myristate acetate. By contrast, BCG-elicited macrophages from C3H/HeJ and A/J mice could not effect tumor cytotoxicity, whereas macrophages from the control C3H/FeJ mice could. When BCG-activated macrophages from C57BL/6J mice were held in culture overnight, their ability to secrete H2O2 in response to the pharmacologic trigger declined 40 to 60%; the presence of endotoxin in the overnight cultures did not alter competence for secretion of H2O2. By contrast, the cultured macrophages lost all of their cytolytic competence, but this competence could be fully maintained by the presence of endotoxin in the cultures. In three distinct circumstances, competence for secretion of H2O23 thus did not correlate with competence for completion of tumor cytotoxicity. The data imply that competence for release of H2O2 is not sufficient for completion of tumor cytotoxicity.

Macrophage activation for tumor cytotoxicity: reactivity of peritoneal and bone marrow macrophages.

Peritoneal macrophages recovered from mice infected with intracellular parasites such as Mycobacterium bovis, strain BCG,Toxoplasma gondii or Listeria monocitogenes exert non specific cytotoxic activity against tumor cells in vitro (1 ,2).Macrophage cytotoxicity is selectively expressed against neoplastic cells.Tumor cells,derived from viral or chemical carcinogen-induced or spontaneous syngeneic and non syngeneic tumors,were found to be quantitatively more susceptible to macrophage cytophatic effects than were non neoplastic cells.Tumor cell killing requires cell to cell contact;soluble cytotoxic factors released by macrophages or by macrophage-tumor cell interactions have not been described.A cinemicrographic analysis revealed that macrophage-tumor cell interactions are characterized by repeated and relatively short (2 hr.) contacts inducing permanent tumor cytostasis and degenerative morphological changes (3).The development of a cytotoxicity assay based on radioisotope release facilitated quantitative analysis of macrophage tumoricidal activity (4).Tumor cell death was evaluated as release of tritiated thymidine from prelabeled target cells into supernatant culture fluids BCG-activated macrophages induced significant release of 3HTdR from tumor cells by 24 hours;however,optimal measurement of tumor cytotoxicity (50% total counts) was observed after 48–72 hours.Macrophages from uninfected mice or from oil-, starch- or thioglycollate induced peritoneal exudates had little effect on labeled tumor mono-layers.

The binding of BCG-activated macrophages to tumor targets stimulates secretion of cytolytic factor.

The binding of neoplastic targets and the secretion of a potent cytolytic protease (CF) by BCG-activated macrophages have previously been shown to be independent functions, both of which are necessary for completion of macrophage-mediated cytolysis. The present studies demonstrate that secretion of CF is triggered by the binding of neoplastic targets to BCG-activated macrophages. The binding of tumor targets, but not of normal lymphocytes, resulted in enhanced secretion of CF from BCG-activated macrophages, although not from macrophages elicited by thioglycolate broth. Dead or metabolically inactive tumor targets, as well as membrane preparations of tumor targets, induced secretion of CF from BCG-activated macrophages. The blocking of macrophage-target binding with a porous filter prevented augmented secretion of CF. Appreciable secretion of CF occurred in as little as 30 min after addition of tumor targets to BCG macrophages. Binding did not induce a generalized increase in secretion of neutral proteases by BCG macrophages, since secretion of plasminogen activator was actually decreased after the binding of P815 targets. The data suggest the selective binding of BCG-activated murine macrophages to neoplastic targets triggers the secretion of a potent CF.

The role of transmethylation reactions in regulating the binding of BCG-activated murine macrophages to neoplastic target cells.

The ability of BCG-activated macrophages from C57BL/6J mice to lyse neoplastic targets was depressed by inhibitors of methyltransferase reactions (10(-4) M adenosine, 10(-5) M EHNA, and 10(-4) M L-homocysteine or 10(-5) M DZA). Binding of P815 mastocytoma targets to BCG-activated macrophages, which has been shown to be a necessary event in cytolysis of those targets, was also inhibited by adenosine, EHNA, and L-homocysteine or by DZA at the above concentrations. Inhibition of binding was obtained when macrophages were pretreated with the inhibitors, whereas pretreatment of targets with the inhibitors did not alter binding. The inhibitors were not toxic to the macrophages, as judged by morphology and viability of the macrophage cultures as well as by ability of macrophages to bind antibody-coated P815 targets or to secrete plasminogen activator. The inhibitors, at concentrations that inhibited cytolysis and binding, also depressed one type of S-adenosyl-L-methionine-mediated methylation reaction (protein carboxy-O-methylation) in BCG macrophages. The data suggest that transmethylation reactions are essential for the ability of BCG activated murine macrophages to bind and, hence, to destroy P815 tumor cells.

Binding of bacillus Calmette-Guérin-activated macrophages to tumor targets. Selective inhibition by membrane preparations from homologous and heterologous neoplastic cells.

The binding of tumor cells by activated macrophages is an initial and necessary event in the cytolysis of these targets. The data here indicate that membrane preparations from RL sigma 1 leukemia targets, EL-4 lymphoma targets, and P815 mastocytoma targets each inhibited binding of its homologous target to bacillus Calmette-Guérin (BCG)-activated murine macrophages in a dose-dependent fashion. Similar amounts of membrane from lymphocytes did not alter binding of the three neoplastic target to BCG-macrophages. Membranes of the three targets also inhibited binding of the heterologous neoplastic targets. Inhibitory activity of membrane preparations from P815, EL-4, and RL sigma 1 targets could be adsorbed by incubation of limiting concentrations of the membrane preparations with BCG-activated macrophages but not with thioglycollate broth-elicited macrophages. Exposure of BCG macrophages to membrane preparations from RL sigma 1, FL-4, or P815 targets inhibited subsequent cytolysis of the three targets. Inhibitory activity was increased in preparations enriched for plasma membrane. The data suggest that binding of three murine, nonadherent neoplastic targets to BCG-activated murine macrophages is mediated, in part, by recognition structures present within the plasma membranes of the three targets.

Differential effects of prostaglandins on the antitumor activity of normal and BCG-activated macrophages

The effect of exogeneously added prostaglandins E 1, E 2, and F 2α, on the capacity of normal and BCG-activated mouse peritoneal macrophages to destroy tumor cells, inhibit tumor cell proliferation, and phagocytize sheep red blood cells (SRBC) was assessed. BCG-activated macrophages destroyed the B16 melanoma cells as assessed by the release of tritiated thymidine ([ 3H]TdR). The addition of prostaglandins did not effect the tumoricidal activity of the macrophage. BCG-activated macrophages also inhibited the tumor cells from incorporating [ 3H]TdR while normal macrophages had no effect. The addition of prostaglandins to BCG-activated macrophages reduced their cytostatic effect on the tumor cells. In contrast, the growth inhibitory capacity of the normal macrophage was enhanced in the presence of prostaglandins. The effect of various pharmacological agents on the ability of the BCG-activated macrophage to reduce tumor cell growth was assessed. Indomethacin potentiated the prostaglandin-mediated reduction in tumoristatic activity. Dibutyryl cyclic AMP did not effect the antitumor activity of the macrophage. Prostaglandins of the E series and dibutyryl cyclic AMP inhibited the phagocytosis of SRBC by BCG-activated macrophages. The opposing effects of prostaglandins on the tumoristatic activity of the BCG-activated and normal macrophage suggests that prostaglandins may play a role in regulating macrophage function based on their level of activation.

Characterization of genetic defects in macrophage tumoricidal capacity: identification of murine strains with abnormalities in secretion of cytolytic factors and ability to bind neoplastic targets.

Peritoneal macrophages from BCG-infected C3H/HeJ or A/J mice, in contrast to BCG-activated macrophages from C3H/HeN mice, were ineffective at lysing adherent 1023 sarcoma targets, nonadherent P815 mastocytoma targets, or nonadherent EL-4 lymphoma targets. The ability of macrophages from BCG-infected C3H/HeJ mice to secret cytolytic factor (CF) into the supernatant medium was markedly impaired. This secretory deficit, however, did not extend to plasminogen activator, secretion of which was augmented. In contrast, the ability of BCG macrophages from A/J mice to secrete CF was comparable to or even slightly higher than that of macrophages from C3H/HeN mice. The ability of BCG-elicited macrophages from A/J mice to bind either of 2 neoplastic targets (the P815 mastocytoma and the EL-4 lymphoma), however, was greatly reduced. The ability of BCG-elicited macrophages from C3H/HeJ mice to bind these targets was comparable to that of macrophages from C3H/HeN mice. The data suggest that the phenotypically-similar deficits in tumoricidal capacity of BCG-elicited macrophages from C3H/HeJ and A/J mice are mediated by mechanistically different defects in macrophage-tumor cell interactions.

Evidence for a multistep mechanism of cytolysis by BCG-activated macrophages: the interrelationship between the capacity for cytolysis, target binding, and secretion of cytolytic factor.

Evidence for a multistep mechanism of cytolysis by BCG-activated macrophages: the interrelationship between the capacity for cytolysis, target binding, and secretion of cytolytic factor.

The ADCC capacity of macrophages from C3H/HeJ and A/J mice can be augmented by BCG.

Macrophages elicited from the peritoneal cavity of C3H/HejJ and A/J mice by bacillus Calmette-Guérin (BCG), although ineffective in mediating tumor cytotoxicity, had augmented capacity to mediate antibody-dependent cellular cytotoxicity (ADCC) against both erythroid and neoplastic target cells. The level of augmented ADCC was comparable to that expressed by BCG-activated macrophages from control C3H/HeN mice. The data suggest that the defective potential for cytolytic activation in 2 genetically deficient strains of inbred mice does not extend to augmentation of capacity for mediating ADCC and that full expression of cytolytic activation is not necessary for expression of augmented cytolysis of antibody-coated targets.

Effects of four agents that modify microtubules and microfilaments (vinblastine, colchicine, lidocaine, and cytochalasin B) on macrophage-mediated cytotoxicity to tumor cells

The effects of vinblastine, colchicine, lidocaine, and cytochalasin B on tumor cell killing by BCG-activated macrophages were examined. These four drugs were selected for their action on membrane-associated cytoskeletal components, microtubules, and microfilaments. Colchicine and vinblastine, which block microtubular synthesis, inhibit macrophage-mediated tumor-cell cytotoxicity at a concentration of 10−6 M. Cytochalasin B, which disrupts microfilaments, enhances tumor cell lysis and stasis due to activated macrophages at a concentration of 10−7 M. Lidocaine, which may induce the disappearance of both microtubules and microfilaments, has the same inhibiting effect as vinblastine at a concentration of 5×10−7 M. Whereas vinblastine and lidocaine seem to act on the macrophage itself, cytochalasin B exerts its effect predominantly on the tumor cell. These results suggest that microtubules and microfilaments play a role in the destruction of tumor cells by activated macrophages.