Binding of cathepsin D to the mannose receptor on rat peritoneal macrophages

Adherent cultures of rat peritoneal macrophages secrete lysozyme and the lysosomal marker enzymes beta-glucuronidase, beta-N-acetylglucosaminidase and acid phosphatase; the levels of secreted lysosomal cathepsin D, however, were found to be insignificant. Incubation of the cells at 4 degrees C for 15 min with yeast mannan or with 50 mM mannose, methyl alpha-glucopyranoside, or N-acetylglucosamine caused the concentration of cathepsin D in the culture medium to increase 30-40-fold; mannose-6-phosphate had no effect. 125I-labeled cathepsin D was prepared and the binding constant to the macrophage cell surface was determined to be KD = 27 nM. The data suggest that cathepsin D binds to the mannose receptor of macrophages and that binding to this receptor is not in equilibrium with the bulk medium.     

Isolation of the receptor for Amaranthus Leucocarpus lectin from murine peritoneal macrophages

The receptor for Amaranthus leucocarpus lectin from CD-1 resident macrophages was purified with affinity chromatography with biotin labeled A. leucocarpus lectin and using avidin-agarose as affinity matrix. The receptor is a glycoprotein of 70 kDa that contains 18% of sugar by weight; it is mainly composed of galactose and N-acetyl D galactosamine in its saccharidic portion, and lacks sialic acid; the protein is rich in glycine, serine and alanine and lacks cysteine residues. The amino terminus of the receptor is blocked. By ionic strength chromatography on a mono P column in anionic form we purified three isoforms from the affinity purified receptor, each showing quantitative differences in glycosylation. The A. leucocarpus lectin receptor is identified only in resting, not activated, macrophages suggesting that it plays a role in activation mechanisms of macrophages     

Lipopolysaccharide receptor on rabbit peritoneal macrophages. I. Binding characteristics

The Bordetella pertussis endotoxin, labeled with tritium ((3H)-LPS), bound irreversibly and nonspecifically to rabbit lung macrophages, but bound reversibly and specifically to both resident and elicited rabbit peritoneal macrophages. The specific binding capacity of the macrophages was saturated with about 3 X 10(4) LPS molecules per cell. The binding was inhibited with the homologous unlabeled endotoxin, but not at all with endotoxin from Proteus mirabilis, thus assessing ligand specificity. Endotoxins from other bacteria gave intermediate inhibition value. Binding of tritium-labeled pertussis endotoxin was significantly inhibited by one of the two polysaccharides (PS-1) present in this endotoxin, but neither the other polysaccharide (PS-2) nor the Lipid A fragment exhibited such activity. These results strongly suggest the presence of a lectin-like receptor for LPS on the membrane of rabbit peritoneal macrophages.     

Role of the maleyl-albumin receptor in activation of murine peritoneal macrophages in vitro

It has been previously demonstrated that maley-lated-BSA (maleyl-albumin) induces functional activation in murine peritoneal macrophages. Furthermore, maleyl-albumin has been shown to interact with two distinct sites on human monocytes; one site is the scavenger receptor, a 260-kDa oligomeric protein which recognizes modified forms of low density lipoprotein (LDL), and the second is a lower affinity site which has yet to be structurally characterized. In the present study, we wished to quantitatively assess the number and character of maleyl-albumin-binding sites on murine peritoneal macrophages and to determine which site or sites are involved in signaling the macrophage to undergo extensive functional development. Binding studies. demonstrate at least two distinct receptors for maleyl-albumin on murine peritoneal macrophages. Scatchard analyses of the binding isotherms reveal two sites characterized by dissociation constants (Kd) of 17.6 nM and 4.9 microM and maximal binding of 1.2 x 10(5) and 1 x 10(6) sites/cell, respectively. The contribution of the scavenger receptor, determined by binding analyses of malondialdehyde-LDL, is described by two sites with Kd of 39.4 pM and 9.6 nM, and maximal binding of 2.7 x 10(3) and 1.9 x 10(4) sites/cell, respectively. Maleyl-albumin blocks binding of malondialdehyde-LDL, whereas modified LDL fails to inhibit binding of maleyl-albumin. Maleyl-albumin, at concentrations producing lower affinity binding, stimulates tumor cytolysis, expression of mRNA encoding TNF, and suppression of INF-gamma-induced expression of Ia Ag. Malondialdehyde-LDL fails to elicit these responses. We conclude that macrophage activation produced by maleyl-albumin is mediated by interaction with the low affinity, high capacity binding site for maleyl-albumin rather than the scavenger receptor.     

IL-4 and IFN-gamma up-regulate substance P receptor expression in murine peritoneal macrophages

While the ability of macrophages to express authentic substance P receptors (i.e., NK-1 receptors) has been inferred from radioreceptor binding assays and functional assays and, most recently, by identification of NK-1 receptor mRNA expression, we know little about NK-1 expression at the protein level or what host factors might up-regulate expression of this receptor. In the present study we demonstrate that the cytokines IL-4 and IFN-gamma can increase the expression of NK-1 receptors on murine peritoneal macrophages. Specifically, we show that IL-4 and IFN-gamma can elicit increases in the level of mRNA encoding the NK-1 receptor by up to 12- and 13-fold, respectively. Furthermore, these cytokines can significantly increase the expression of the NK-1 receptor protein as measured by Western blot and FACS analysis using specific Abs developed in our laboratory. In addition, we have demonstrated the ability of both IL-4 and IFN-gamma to enhance the ability of macrophages to bind substance P as measured by radiolabeled binding assay. The observation that the level of expression of this receptor protein can be enhanced by cytokines that promote either cell-mediated (Th1) or humoral (Th2) immune responses supports the idea that this receptor can be induced during either type of immune response. As such, these results may point to a more ubiquitous role for substance P in the generation of optimal immune responses than previously appreciated.     

Binding of concanavalin A and wheat germ agglutinin by murine peritoneal macrophages: ultrastructural and cytophotometric studies

Summary Concanavalin A and wheat germ agglutinin were employed in conjunction with the horseradish peroxidase-diaminobenzidine method for the detection of sugar residues on the surface coat of exudate and resident murine peritoneal macrophages. Electron microscopical and cytophotometric techniques were used for the visualization and quantification of the final reaction product on the surface of cells. After incubation with concanavalin A and wheat germ agglutinin, both exudate and resident macrophages showed readily detectable final reaction product indicating the presence of numerous, easily accessible, -methyl-d-mannosyl andN-acetyl-d-glucosaminyl residues on their surface. The binding of concanavalin A was higher with resident than with exudate macrophages. With wheat germ agglutinin, a different pattern of lectin binding was observed: more electron-dense product was deposited on exudate than on resident macrophage surfaces. The binding of concanavalin A and wheat germ agglutinin to macrophages was inhibited by the competing sugars -methyl-d-mannoside andN-acetyl-d-glucosamine, respectively.     

Structural characterization of the disialogangliosides of murine peritoneal macrophages

Sialoglycosphingolipids (gangliosides) have been increasinglyimplicated as regulators of membrane signaling events. Macrophageganglioside patterns dramatically increase in complexity whenmurine peritoneal macrophages are stimulated in vivo with theappearance of the sialidase-sensitive monosialoganglioside GMlb(cisGMl) as a major component Gangliosides from stimulated murineperitoneal macrophages were separated into monosialo and polysialofractions and the polysialo fraction structurally characterizedby enzymatic, chemical, and mass spectra methods. All detectablecomponents of the polysialo fraction were determined to be disialogangliosides.Treatment of the polysialo fraction with Clostridium perfringenssiali-dase produced mostly the sialidase-resistant monosialoganglioside,GMIa, and a minor amount of asiaJoGMI. Perio-date oxidationand mass spectrometry analyses demonstrated the lack of tandemdisialo moieties which indicated the absence of GD1b or GD1c(GDI) entities. The combined data showed the major disialogangliosidesconsisted of GDla entities comprising IV3-NeuAc,II3NeuAc-GgOse4Cer,IV3-NeuGc,II3NeuAc-GgOse4Cer, IV3NeuAc,II3NeuGc-GgOse4Cer, andIV3-NeuGc,II3NeuGc-GgOse4Cer. Minor components consisted ofGDl entities, IV3NeuAc, III6NeuAcGgOse4Cer, IV3NeuGc, III6NeuGcGgOse4Cer,and also positional iso-mer(s) of GDl(NeuAc, NeuGc). These isomericcomponents were identified by collision analysis and tandemmass spectrometry. Consistent with previous analyses, the cer-amideportion of all polysialo (disialo) gangliosides contained solelyC18 sphingosine with C16 and C24 fatty acid moieties. Theseresults, combined with the previous characterization of macrophagemonosialogangliosides, indicate normal murine macrophage gangliosidebiosynthesis proceeds along the "a" ganglioside pathway, e.g.,GM3GM2GMlaGDl, and the proposed asialogan-glioside or "" pathway,asialoGMlGMlbGDl. The presence of totally sialidase-sensitivegangliosides appears to be characteristic of functional murineperitoneal macrophages while they are reduced in geneticallyimpaired cells. ganglioside GDla ganglioside GDl murine macrophages tandem mass spectrometry collision induced dis-association electrospray ionization     

Inactivation of thrombin by murine peritoneal macrophages.

We recently showed that murine peritoneal macrophages cultured in vitro express potent prothrombinase activity (Lindahl, U., Pejler, G., B?gwald, J., and Seljelid, R. (1989) Arch. Biochem. Biophys. 273, 180-188). In the present report, we demonstrate that the macrophages also express anticoagulant activity by inactivating the thrombin that is formed due to the action of the prothrombinase. Addition of exogenous purified thrombin to the macrophage cultures resulted in inactivation of the enzyme at a maximum rate of approximately 5 micrograms/h/10(6) cells. The inactivation appeared to be specific for thrombin, since neither Factor Xa, chymotrypsin, nor trypsin, three serine proteases exhibiting homology with thrombin, were inactivated by the macrophages. Thrombin-inactivating activity was not secreted into the culture medium. Inhibitors of endocytosis did not decrease the rate by which thrombin was inactivated, suggesting that internalization of the coagulation factor was not required. In contrast, the thrombin-inactivating activity was strongly inhibited by the polycation Polybrene. Anion-exchange chromatography of extracts obtained after Triton X-100-solubilization of the macrophages demonstrated that the thrombin-inactivating activity exhibited a high negative charge. Incubation of the thrombin-inactivating activity recovered after anion-exchange chromatography with unlabeled thrombin, followed by sodium dodecyl sulfate-polyacrylamide gel electrophoresis analysis, showed that thrombin was proteolytically cleaved into defined fragments. Similar proteolytic fragments were obtained when 125I-labeled thrombin was added to macrophage cultures. Degradation of thrombin was blocked by phenylmethanesulfonic fluoride, an inhibitor of serine proteases, but not by inhibitors of other classes of proteases. Thrombin that had been chemically modified at its active site was degraded at the same rate by the macrophages as active thrombin. Taken together, these findings indicate that the murine macrophages express surface-bound serine protease activity that specifically inactivates thrombin by proteolytic cleavage. The significance of thrombin-inactivating activity in relation to the involvement of macrophage procoagulant activity in the immune response is discussed.     

Peroxidase-induced enhancement of chemiluminescence by murine peritoneal macrophages

A number of substances have been shown to enhance the respiratory burst (RB) of macrophages. Many of these substances are not normally found in vivo. The present study suggests that a group of enzymes characterized as peroxidases have the ability to significantly enhance the RB and concomitant phagocytosis by murine peritoneal macrophages. Horseradish peroxidase (HRP), lactoperoxidase (LPO), and microperoxidase (MPO) can significantly augment these functions. Both resident and thioglycollate-induced macrophages exhibited enhanced chemiluminescence (CL) upon exposure to HRP, however, the effect was more pronounced with the latter. The increase in CL was correlated with an increase in production of superoxide, which was measured by reduction of cytochrome c. Horseradish peroxidase immobilized on an inert carrier, was capable of enhancing the RB suggesting that it does not have to enter the cell in order to function. Hemin, hematoheme and hematoporphyrin had little effect on macrophage stimulated CL. All of the peroxidases tested caused increased phagocytosis of opsonized zymosan. These studies indicate that peroxidases are capable of stimulating the RB, phagocytosis and possibly other macrophage functions.     

Regulation by CRAMP of the responses of murine peritoneal macrophages to extracellular ATP

Peritoneal macrophages were isolated from wild type (WT) mice and from mice invalidated for the P2X₇ receptor (KO) which had been pretreated with thioglycolate. In cells from WT mice, 1 mM ATP increased the intracellular concentration of calcium ([Ca²⁺]_i), the uptake of ethidium bromide, the production of reactive oxygen species (ROS), the secretion of IL-1β, the release of oleic acid and of lactate dehydrogenase; it decreased the intracellular concentration of potassium ([K⁺]_i). In KO mice, ATP transiently increased the [Ca²⁺]_i confirming that the P2X₇ receptor is a major receptor of peritoneal macrophages. WKYMVm, an agonist of receptors for formylated peptides (FPR) also increased the [Ca²⁺]_i in murine macrophages. The slight increase of the [Ca²⁺]_i was strongly potentiated by ivermectin confirming the expression of functional P2X₄ receptors by murine peritoneal macrophages. CRAMP, the unique antimicrobial peptide derived from cathelin in mouse inhibited all the responses coupled to P2X₇ receptors in macrophages from WT mice. Agonists for FPR had no effect on the increase of the [Ca²⁺]_i in response to ATP. CRAMP had no effect on the increase of the [Ca²⁺]_i evoked by a combination of ATP and ivermectin in macrophages from P2X₇-KO mice.In summary CRAMP inhibits the responses secondary to the activation of the murine P2X₇ receptors expressed by peritoneal macrophages. This inhibition is not mediated by FPR receptors and is specific since CRAMP has no effect on the response coupled to P2X₄ receptors. It can thus be concluded that the interaction between P2X₇ receptors and cathelin-derived antimicrobial peptides is species-specific, in some cases (man) positive in others (mouse) negative.     

Expression of the transferrin receptor in murine peritoneal macrophages is modulated in the different stages of activation

A specific high affinity binding site for the serum glycoprotein transferrin was identified on murine peritoneal macrophages. The binding reached equilibrium within 60 min and was reversible, saturable, and specific for transferrin. Although the presence of this receptor was detected on the cell surface by studies carried out using intact cells, the majority (70 to 90%) of the binding activity was detectable only in detergent extracts of such cells. This suggests that a substantial portion of the binding activity is localized within the macrophage. The association constant (Ka) for binding to intact cells (6 to 9 X 10(8) M-1) was comparable to values reported for transferrin receptors described on other cell types. The expression of transferrin binding activity was examined in macrophages exhibiting qualitatively and quantitatively different degrees of functional activation. Resident peritoneal macrophages, exudate macrophages primed by elicitation with pyran copolymer, and activated macrophages induced by chronic infection of mice with bacillus Calmette-Guerin (BCG) or elicitation with heat killed Propionibacterium acnes (P. acnes) had low numbers of binding sites (1000 to 5000 total sites/cell). Macrophages elicited by sterile inflammatory agents (thioglycollate broth, fetal bovine serum, or casein) all exhibited a greater number of transferrin receptors (15,000 to 20,000 total sites/cell). This modulation did not appear to result from differential shifts between surface and internal loci. Our results suggest that the expression of the transferrin receptor may be a useful marker of the responsive stage of macrophage functional activation and the membrane changes that accompany activation.     

Fc-receptor mediated protein phosphorylation in murine peritoneal macrophages

The effect of Fc receptor engagement on protein phosphorylation in murine peritoneal macrophages has been investigated. Treatment of macrophage cultures with insoluble immune complexes resulted in enhanced phosphorylation of six proteins at 73, 66, 53, 37, 31 and 25 kD. Comparison of the protein phosphorylation patterns induced by immune complexes with those induced by agents which mimic the actions of well known intracellular second messengers (i.e., A23187, dibutyryl cAMP, or phorbol myristate acetate) revealed substantial similarity between Fc receptor induced events and those induced in response to phorbol diesters. There were, however, two phosphorylated proteins which were only seen following stimulation with immune complexes. Thus, more than one kind of protein kinase activity appears to be involved in Fc receptor mediated stimulation of macrophage function.     

Ultrastructural cytochemistry of non-specific esterase in murine peritoneal macrophages

Summary Non-specific esterase (NSE) activity was demonstrated in glutaraldehyde-fixed monolayers of murine peritoneal macrophages. Using 2-naphthylthiol acetate (NTA) as substrate and Fast Blue BB as coupling agent a strong osmiophilic reaction product was obtained. The reaction product was observed as electron-dense dots covering cisterns of the rough endoplasmic reticulum, Golgi saccules and vesicles, or as large aggregates in lysosomes. Using -naphthyl butyrate (ANB) as substrate and hexazotized pararosaniline as coupling agent the osmiophilic reaction product was observed extracellularly on the plasma membrane as an electron-dense continuous layer, whereas intracytoplasmic staining of lysosomes was rare. Substitution of the coupling agents in the respective media resulted in a slight reaction with the ANB medium whereas with the NTA medium reaction product was observed only in lysosomal structures. The substrate specificity of the different types of esterases was confirmed after isoelectric focusing on thin-layer polyacrylamide gels. The results indicate that in murine peritoneal macrophages different types on NSE are detected with NTA and ANB, having distinct ultrastructural localizations.     

Characteristics of the beta-glucan receptor of murine macrophages

Phagocytosis of heat-killed yeast (HK-yeast), zymosan, and glucan particles by thioglycollate-elicited mouse peritoneal macrophages (Tg-macrophages) was inhibited by soluble glucan polymers/oligomers. The inhibitory capacity of soluble glucans decreased steeply with the decrease in the degree of polymerization (DPn); i.e., the concentration at which 50% inhibition of phagocytosis was attained was 0.23 microgram/ml for glucan 1 (DPn 24.8), 0.8 microgram/ml for glucan 2 (DPn 21.9), and greater than 40 micrograms/ml for glucan 3 (DPn 13.8). The glucan polymers were obtained by partial hydrolysis of glucan particles with formic acid (90%, 95 degrees C, 20 min) and fractionation according to solubility in ethanol water mixtures. A short preincubation (5 min, 4 or 37 degrees C) of Tg-macrophages with glucan 1 led to a subsequent inhibition of HK-yeast phagocytosis. Recovery of the phagocytic function was slow (27% in 3 h; 68% in 5 h) and required protein synthesis. beta-Glucan receptor expression was also suppressed by dexamethasone treatment. Mannan exerted at high concentrations (5 mg/ml) a partial inhibitory activity which was totally abrogated by beta-glucanase treatment. Treatment of macrophages with glucan together with mannan did not enhance the inhibitory capacity of glucan beyond the component abrogated by enzyme treatment. Contribution of local opsonization of HK-yeast to the phagocytic response (involvement of complement receptors) was indirectly negated; (a) glucan 1 which inhibits HK-yeast phagocytosis by up to 95% is not an activator of complement and therefore could not compete for the opsonizing proteins; (b) cycloheximide treatment in itself inhibited only partially HK-yeast phagocytosis whereas it inhibited the reexpression of the glucan receptors; (c) glucan 1 did not affect the phagocytosis of serum opsonized HK-yeast. Thus under the experimental conditions described, phagocytosis of HK-yeast by murine macrophages is mediated by and large by the beta-glucan receptors, while the mannose receptors and complement receptors do not contribute to the process.     

Phenotypic and functional analysis of murine resident and induced peritoneal macrophages

Primary macrophages from the peritoneal cavities of mice are commonly used ex vivo to produce inflammatory cytokines and test anti-inflammatory agents. Although approximately 1 million peritoneal macrophages can be obtained from an untreated mouse, more than twice that number can be collected 48 to 72 h after intraperitoneal injection of sterile inducing agents such as Brewer thioglycollate broth, casein, and proteose peptone. However, whether 'induced' macrophages are functionally equivalent to 'resident' peritoneal macrophages has been unclear. Flow cytometric analysis revealed significant phenotypic differences between these 2 macrophage types. Resident and induced peritoneal macrophages also demonstrated markedly different capacities to produce the inflammatory cytokines interleukins 6 and 1beta in response to lipopolysaccharide stimulation in vitro. Increased understanding of the differences between resident and induced peritoneal macrophages likely will help investigators decide which macrophage type is appropriate for their in vitro assay needs.     

Maleylated-BSA suppresses IFN-gamma-mediated Ia expression in murine peritoneal macrophages

Maleylated bovine serum albumin (maleyl-BSA) and other polyanionic polymers that are recognized by cell surface receptors on macrophages have been shown to induce chemotaxis, protease secretion, and tumoricidal function in this cell type. In this paper the effect of maleyl-BSA on Ia antigen expression has been evaluated. In a fashion similar to LPS, maleyl-BSA suppressed IFN-gamma-induced expression of Ia in a time- and dose-dependent manner. Also like LPS, maleyl-BSA stimulated the production and secretion of substantial amounts of PGE2 over a 24-hr period. This did not, however, appear to be the primary mechanism by which expression of Ia was suppressed, because co-treatment of the cells with indomethacin, which totally inhibited the production of PGE2, only minimally affected the suppressive activity. Surprisingly, the suppressive activity of both maleyl-BSA and LPS could be largely abrogated by co-treatment of the cells with cyclohexamide during the time period when Ia expression was sensitive to suppression. This effect was selective in that PGE2- or dibutyryl cyclic AMP-induced suppression of Ia expression was not affected by cyclohexamide treatment. The data support the concept that there are multiple molecular mechanisms involved in the negative regulation of IFN-gamma-induced Ia expression in macrophages. Such mechanisms may include, in addition to the synthesis of PGE2 and consequent elevation in intracellular levels of cyclic AMP, one or more proteins made early after treatment with either maleyl-BSA or LPS. Thus the function of some of these early gene products may be to regulate expression of functional genes such as that encoding Ia antigen.     

Expression of the transferrin receptor on murine peritoneal macrophages is modulated by in vitro treatment with interferon gamma

The expression of transferrin receptors on murine peritoneal macrophages has been shown to be down regulated during functional activation in vivo. This observation suggested that the level of transferrin receptor expression varies in response to discrete extracellular signals known to induce macrophage activation. We have tested this concept directly and have shown that decreased transferrin receptor expression can be reproduced in vitro by treatment of inflammatory macrophages with preparations of interferon gamma derived from a T cell hybridoma supernatant. The ability of this agent to down regulate the expression of the transferrin receptor exhibited dose and time dependencies similar to those required for development of other macrophage functions in vitro. The addition of LPS produced no further decrease in receptor expression. Furthermore, murine gamma interferon, produced by recombinant DNA technology also caused a downshift in transferrin receptor expression at doses similar to those which have been shown previously to induce activation. The changes in receptor activity were the result of altered numbers of binding sites and the receptor:ligand affinity remained unaffected. These results indicate that altered expression of the transferrin receptor is one element of the pleiotypic change which macrophages undergo in response to IFN gamma. This system may, therefore, provide a useful model in which to study the biochemical basis of IFN gamma action in mononuclear phagocytes.     

Complement receptor phenotypes of culture-derived murine macrophages

The distribution of complement receptors CR1 and CR3 among macrophages derived from cultures of bone marrow, blood, and elicited or normal peritoneal cell population was studied. Cells and colonies from the first three sources had a common phenotype, CR1 + 3 ?, whereas those from the noram peritoneal populations had either CR1 + 3 ? or CR1 + 3 +. The former phenotype characterized spindle-shaped as well as epithelial-like macrophages; the latter was essentially restricted to colonies made up of the epithelioid cells. Both morphologic features and the CR phenotypes remained stable throughout the culture period. These phenotypic differences might be explaniend by the presence of at leas two clonally derived types of macrophages.     

Induction of metalloelastase mRNA in murine peritoneal macrophages by diethylmaleate

Macrophage-specific metalloelastase (MME) hydrolyzes elastin and other matrix proteins and plays an important physiological role in tissue remodeling and pathological tissue destruction. We have examined the effects of diethylmaleate (DEM), an electrophilic agent that reacts with sulfhydryls, on the expression of MME mRNA in mouse peritoneal macrophages. Quantification of MME mRNA by Northern blot analysis revealed that basal mRNA levels were quite low in freshly isolated cells, although mRNA levels increased markedly and reached a steady level within 12 h when cells were cultured in a serum-supplemented RPMI 1640 medium. When macrophages were challenged with DEM at 0.05-1.0 mM for 8 h the expression of the MME gene was enhanced further. In the presence of 0.1 mM DEM, the level of the MME mRNA increased 2-fold compared to the control levels after 6-9 h and decreased to control levels in 24 h. Other electrophilic agents, catechol and 1-chloro-2,4-dinitrobenzene, also enhanced MME gene expression. However, oxidative stress agents such as hydrogen peroxide, menadione, paraquat (an O-2 generator), sodium arsenite and cadmium chloride had no effect on MME gene expression. These results indicate that the electrophilic agents selectively enhance the expression of MME mRNA during primary culture of the macrophages.