- Email: [email protected]
Modulation of Macrophage Functions by Lymphokines
Immunobiol., vol. 161, pp. 352-360 (1982) Abteilung fUr Experimentelle Dermatologie, Universitats-Hautklinik, Miinster, Federal Republic of Germany
Modulation of Macrophage Functions by Lymphokines C. SORG
Abstract Macrophage migration inhibitory factors (MIFs) of mouse and guinea pig have been thoroughly characterized with regard to molecular weight and isoelectric points. Several molecular weight species have been identified. In a comparative study with purified MIFs it was found that these molecules were distinct from a series of other lymphokines, particularly so from macrophage activating activities. Investigations on the molecular weight heterogeneity of MIF have led us to a transglutaminase-like activity which was found to be expressed in certain subsets of macrophages. The question whether low molecular weight factors are polymerized by this enzyme to oligomers is further investigated. Studies on the induction by lymphokines of interferon and plasminogen activator revealed a great heterogeneity of responding macrophages. In studies on the biological basis of the functional heterogeneity of macrophages, the question was investigated whether the heterogeneity was due to different macrophage subpopulations or to intermediate relatively stable phenotypes on their way to maturity and senescence. To approach this question, the bone marrow liquid culture system was used as a developing system. Our data are summarized in a unifying model which takes into account the different constitutive and inducible functions during the cell cycle. Accord ingly, lymphokines may act either as differentiation signals, as mitogens or activating signals.
Introduction Lymphocytes and macrophages are the main constituents of acute or chronic inflammatory cell infiltrates. In the past fifteen years it has been increasingly recognized that lymphocyte activation productions (lym phokines) may induce or modulate macrophage functions. The first lym phokine described was the macrophage migration inhibitory factor (MIF) (1). Since then, many other factors have been described, and the list of lymphokines currently numbers well beyond one hundred. In the years following its discovery, a great number of macrophage activating properties has been attributed to MIF.
Characterization of MIF Our laboratory has worked in the past years on the chemical and functional characterization of MIF from three different species (guinea pig, mouse, man) (2-8). Essentially from work on guinea pig MIF we have
Macrophages and Lymphokines· 353
established the following characteristics: MIF activity is associated with several molecular weight species. These have been identified at molecular weights of 60,000 a), 45,000 b), 30,000 c), and 15,000 d) daltons, all of which have an isoelectric point of about 5.2. The buoyant density in CsCI gradient was found at 1.278 for all four molecules. No proteolytic activity was found to be associated with either one of the molecules, nor did treatment with highly purified neuraminidase destroy the biological activity. Reduction and alkylation or treatment with EDTA did not change the molecular weights of a, b, or c. Because of the similarities and their chemical and biological properties, it was assumed that the three molecules are oligomers of a common subunit of the molecular weight of 15,000 and an isoelectric point of 5.2. Similar results have been obtained with murine and human MIF (unpublished data). In another study we investigated the cellular source of migration inhibi tory factor (9), since it had been reported earlier that MIF was also produced by nonlymphoid cells. Using an antilymphokine serum and radio labeled supernatants from various cells, migration inhibitory activity reacting with the antilymphokine serum and properties identical to those determined on MIF from lymphocytes was found in the following cell types: purified B-cells activated with Staphylococcus aureus, L 2C leukemia (strain 2), and growing embryonic fibroblasts. No activity was found in supernatants of peritoneal macrophages. The results indicate that MIF with similar molecular properties may be produced by a variety of cells under appropriate stimulation. This situation is reminiscent of the production of interferons by a diversity of cell types under different conditions. In order to determine further biological effects of MIF on macrophages, activities from Concanavalin A (Con A)-stimulated murine spleen cell supernatants were fractionated by Sephadex chromatography and isoelec tric focusing. Aliquots of these preparations were sent to a number of laboratories, each performing a particular test on macrophages. The func tional profile of the various MIF preparations showed that they were contaminated to a certain degree with other lymphokine activities, e.g. as expected T-cell replacing factor (TRF; tested by Dr. Schimpl) was a contaminant in preparation (c) of molecular weight of 28,000. Induction of leishmania kill (performed by Dr. Mauel) was only observed with prepara tion a) and not with preparations b), c), and d). Tumor cytotoxicity (performed by Dr. Kniep) could be induced with all four preparations, however, we would not conclude that MIF is identical with a macrophage cytotoxic factor, since our preparations were not endotoxin-free, and the observed effect could have been caused by a cooperative effect of endotoxin and MIF. The experiment further proved quite clearly that MIF has no chemotactic activity and is not related to interferon. MIF is also not an inducer of interferon. From these experiments we would conclude that MIF is not identical with most macrophage activating activities which was the widely accepted assumption of recent years.
354· C. SORG
Functional heterogeneity of macrophages During work on the purification and chemical characterization of mac rophage migration inhibitory factor of guinea pig and mouse, we observed that inhibitory activity was associated with several distinct molecules which seemed to be oligomers of a small subunit of approximately 14,000 to 15,000 daltons. Since formation of aggregates and disulfide bridges could be excluded, the involvement of a polymerizing enzyme was suspected. In fact, low amounts of transglutaminase-like activity could be detected in crude, concentrated supernatants of activated lymphocytes, which was not associ ated with MIF activity. It was found that erythrocytes, granulocytes, thymocytes or lymphocytes with or without Concanavalin A stimulation contained no enzyme activity (10). Normal peritoneal wash-out mac rophages in peritoneal exudate cells stained positive for transglutaminase to varying degrees, whereas macrophages derived from bone marrow liquid cultures were 20 % positive at day 3 and 100 % at day 17. Promonocytes and monocytes were negative. Positively stained cells, as detected by incorporation of dansylcadaverine also phagocytosed opsonized sheep ery throcytes. The degree of staining varied considerably in the macrophage like cell lines IC 21 (100 %), J 774.2 (77 %), P 388 Dl (50 %). This result and those from autoradiography studies indicate that expression of trans glutaminase is not associated with the S-phase of the cell cycle. The enzyme is CaH dependent, and it appears neither to be on the outer cell surface nor being released into the culture medium. Circumstantial evidence indicates that it is also not compartmentalized in cytoplasmic vesicles. The question whether this enzyme is responsible for the polymerization of low molecular weight MIF has not been resolved yet. Since trans glutaminase seems not to be released into the culture supernatant, it is possible that polymerization of proteins is an artefact due to the release of the enzyme from dead cells in culture. While the induction and modulation of enzyme expression is still under study, it is concluded that trans glutaminase is a newly described marker for macrophages of a certain differentiation or activation state. Macrophages after exposure to lymphokines alter their behaviour in a sense that they show either enhanced, depressed, or newly induced func tions. Our interest was directed at the induction and modulation of plasminogen activator production by murine macrophages (11). Mac rophage monolayers obtained by 24 hour culture of proteose peptone elicited murine exudate cells were incubated with lymphocyte culture of supernatants. After 48 hours, the supernatants were replaced by serum-free medium, and the macrophages were incubated for another 24 to 48 hours. These supernatants were assayed for plasminogen activator using the lysis of 125I-Iabeled fibrin. It was found that supernatants of antigen or mitogen stimulated spleen cells induced the plasminogen activator secretion, whereas control supernatants were ineffective. The same was found with supernatants of mitogen stimulated lymph node cells. Lymphokine induced
Macrophages and Lymphokines· 355
plasminogen activator secretion by macrophages was enhanced after phagocytosis of latex beads. In the course of these studies on the lymphokine induced production of plasminogen activator it was observed that some macrophage culture supernatants not only lacked plasminogen activator but were even inhibi tory to plasmin dependent fibrinolysis (12). This observation and the scarcity of reports on protease inhibitor production by macrophages prompted us to investigate this subject more systematically. Macrophages from peritoneal exudates which had been induced by various irritants were cultured, and the supernatants were tested for inhibitors of fibrinolysis using the plasmin dependent lysis of 125I-Iabeled fibrinogen as assay. Wash out macrophages and casein or proteose peptone elicited macrophages were found to release fibrinolysis inhibitors in contrast to lipopolysaccharide or thioglycollate induced macrophages. The molecular weight of inhibitors was determined at 60,000, 45,000 and 15,000 daltons by Sephadex chromatography. Whereas the inhibitors at 15,000 and 45,000 daltons could be detected in all experiments, inhibitor at a molecular weight of 60,000 daltons was not present in all preparations. The isoelectric point of inhibitors was determined at 4.15. The proteolytic and esterolytic activity of trypsin and chymotrypsin were both inhibited by each of the two inhibitors. On the other hand, only the proteolytic activity of plasmin could be inhibited. Evidence for the active synthesis of inhibitors by macrophages came from several experiments. Macrophages in serum-free cultures continued to release inhibitors for at least 48 hours; normal mouse serum did not contain inhibitors of the same molecular size, and 3H-Ieucine was incorporated into the inhibitors which were specifically detected by the absorption of plasmin inhibitor complexes to lysyl-sepharose. Since inhibitors and plasminogen activator could not be detected in the same macrophage culture supernatants, it appears that the production of inhibitors and plasminogen activator by the same macrophage population is mutually exclusive. In a subsequent study we investigated whether the production of fi brinolysis inhibitors is modulated under the influence of lymphokines (13). Supernatants from mitogen stimulated spleen cells were fractionated on Sephadex G-75, and the fractions were incubated for 24-48 hours with various types of peritoneal exudate cells. It was found that proteose peptone elicited cells were easily converted from an inhibitor to a plasminogen activator producing state, whereas casein elicited cells stopped producing fibrinolysis inhibitors but could not be converted to the plasminogen activator producing state. Thioglycollate induced cells which already pro duced plasminogen activator were enhanced by the lymphokines. Lipopolysaccharide elicited macrophages which did not produce fibrinoly sis inhibitors were refractory to the modulating effects of lymphokines. Lymphokine activity was found in a molecular weight range of about 65 to 20,000 daltons.
356 .
C.
SORG
Another line of interest was developed from the question what role macrophages play in the production of interferon in combined T lympho cyte-macrophage cultures. It was found that unfractionated murine spleen cells produce immune interferon upon stimulating with antigen or mitogen (14, 15). When spleen cells were passed over glass bead columns, interferon production decreased, while the response to the stimulants drastically increased. When these cells were further purified over nylon wool columns, interferon production was totally abolished, whereas thymidine incorpora tion in stimulated cultures was invariably high. Interferon production by nylon wool column purified lymphocytes could be restored with mac rophages grown from bone marrow cultures or spleen cells, but not with macrophages from proteose peptone induced peritoneal exudate cells. It was also found that pure macrophage cultures from the spleens of BCG (Bacille Calmette Guerin) immunized mice consistently produced interfe ron activity without any further stimulation. When culture supernatants of activated T lymphocytes which did not contain any interferon activity were transferred to macrophage cultures from different sources and incubated for 45 hours, interferon activity could be detected in supernatants of mac rophage cultures from bone marrow and spleen, but not in those from proteose peptone induced peritoneal exudate cells. It is concluded that certain macrophage populations can be induced to produce interferon activity, whereas others are refractory to this induction which appears to be linked to their differentiation state. As described above, macrophages display a great heterogeneity in response to lymphokines. Functional heterogeneity of macrophages has been documented before in many cases. The question what biological principle determines macrophage heterogeneity has not been answered yet. One assumption is that we are dealing with true subpopulations as it is known for the T cell/B cell lineages (16). Another possibility would be that macrophages which are bone marrow derived pass through different ma turation stages or intermediate relatively stable phenotypes on their way to maturity and senescence, thus expressing characteristic functions in each maturation stage. In order to approach this problem, we adopted the bone marrow liquid culture system (17). Bone marrow cells from mice were cultured on teflon membranes in the presence of a colony-stimulating factor produced by L cells. Parallel cultures were harvested daily, and a series of functions was recorded over 12 days. The growth curve was declining rapidly after onset of culture until days 2-3 due to the death of granulocytes and erythrocyte precursors. At days 4-5, the yield of cells increased, reaching a maximum at days 6-8 and then declined again. By Pappenheim staining, the first appearance of macrophages could be seen at day 2 which increased linearly and reached a plateau at day 6. When the cells were stained for unspecific esterase (a-naphthyl acetate), weakly stained cells peaked at days 2 and 3 and again at days 5 and 6, whereas heavily stained cells appeared for the first time at
Macrophages and Lymphokines' 357
day 3 and then again at day 6 and later, reaching a maximum of nearly 90 % at day 9. The phagocytosis of latex beads was observed very early in culture, reaching a plateau at days 7-8. Phagocytosis of immunoglobulin coated sheep red blood cells was also performed with adherent cells. Phagocytic activity was apparent at day 2, decreasing by day 3 and reaching again a maximum at days 5 and 6. From day 7 on, only few cells phago cytosed 3 or more red blood cells. In another series of experiments, the response to three different stimuli was investigated. The chemotactic response against zymosan activated serum was measured in Boyden chambers using cellulose nitrate filters. Cells which penetrated the filter were differentiated for granulocytes and mononuclear cells. Results showed a sharp maximum at days 3-4 for the chemotactic response of mononuclear cells. No chemotactic response was observed after day 5. When the response to MIF containing supernatants of Concanavalin A-stimulated spleen cells was measured, again a sharp maxi mum at days 4 and 5 was found, whereas no response could be detected before and after this period, even though the cells were vigorously migrat ing in the earlier days. Since interferon production by lymphokine induced macrophages requires an incubation time of 24-48 hours and since this time would be too long in order to draw any conclusion on the momentous differentiation state of a cell culture, it was decided to use the LPS-induced production of interferon which -like migration inhibition and chemotaxis requires only three to four hours of incubation. When tested daily, interfe ron induction by LPS was successful only at days 5 and 6. As a further function, the plasminogen activator production was recorded. Plasminogen activator was detected with a fibrinolysis assay which allows the simultaneous detection of fibrinolysis inhibitors (12). No plasminogen activator or significant amounts of fibrinolysis inhibitors were found during the first three days of culture. Beginning at day 4, plasmino gen activator was readily produced reaching a maximum at days 6 and 7 and then decreased rapidly. At day 8, the cells had ceased to produce plasmino gen activator, and at day 9 they began to release fibrinolysis inhibitors for the rest of the culture period. When we surveyed our data, we noticed that those macrophages either derived from bone marrow liquid cultures or from peritoneal exudates which did not produce plasminogen activator could also not be induced to release interferon neither by LPS nor by lymphokines. Further, it was observed that those macrophages that responded best to LPS in terms of interferon production were close to or at the plasminogen activator produc ing state (18, 19). We therefore asked whether plasminogen activator production might be a constitutive trait of those macrophages which can be induced to interferon production. In a series of experiments we found an almost perfect correlation between these two parameters, i.e., macrophage populations which produced plasminogen activator within 25 hours of culture could be induced to release interferon. From these observations we
358 . C. SORG
would have to conclude that certain macrophages have the ability to produce plasminogen activator and are inducible to interferon production at the same time. Whether we are dealing with subpopulations or pheno types of different maturation stages we still cannot decide. Macrophages may be induced to plasminogen activator production by the tumor promotor phorbol myristate acetate (PMA), which is also mitogenic to macrophages. The following question was asked: are mac rophages pushed by PMA and similar signals such as lymphokines into a certain phase of the cell cycle in which plasminogen activator production is expressed? And further, what role plays the cell cycle in macrophage heterogeneity? The following experiments were designed to shed some light on this question (20). Macrophages produce substantial amC'unts of plasminogen activator 2-22 hours after exposure to PMA. The production of plasminogen activator then steadily decreases. In contrast, the number of cells in S-phase peaks later at 26-46 hours. This indicates that macrophages after exposure to PMA are pushed towards the S-phase of the cell cycle. On their way to S, probably in late G!, macrophages express plasminogen activator as a constitutive trait and also become sensitive to activation by lipopolysac charides. Based on our sllmmarized data and those from different laboratories, we are able to describe the following picture: first, what is likely to happen in the bone marrow liquid culture system? Macrophages differentiate from precursors in the presence of colony-stimulating factor. After a phase of intensive proliferation of precursors and differentiation into macrophages, the young macrophages still keep on cycling, thus accumulating cells in late G!, which are characterized by plasminogen activator production, induci bility of interferon, and their response to MIF and certain chemotactic factors. When proliferation is fading away, the cells differentiate gradually in G! loosing a series of constitutive functions and pass on to a Go-like state characterized by production of fibrinolysis inhibitors. In this model, the functional state of normal resident macrophages should be compatible with Go, that of proteose peptone elicited macrophages with «early» and thiogly collate induced macrophages with «late» GI" In kinetic studies with PMA, the time needed to induce normal resident macrophages to plasminogen activator production was considerably longer than that needed for proteose peptone elicited macrophages. The model proposes further that mac rophage differentiation in G! is reversible and that macrophages of various functional states have to go through the bottle neck of «late» G! before entering S. What then is an activated macrophage, and how can the action of lymphokines be fitted in? As shown first by MACKANESS (21) and subse quently by many others, activated macrophages express special functions, in particular bactericidal and tumoricidal properties. Several investigations have demonstrated that macrophages can be rapidly activated by lipopoly-
Macrophages and Lymphokines· 359
saccharides to kill tumor cells (22). However, not all macrophages respond the same. Nonresponding macrophages can be made responsive, however, by preincubation with lymphokines which serve as preparatory signals. In our model this would mean that macrophages are pushed first from Go or «early» G 1 into «late» G 1, where macrophages become inducible to kill tumor cells. These findings are paralleled by ours on the induction of interferon by LPS. According to the proposed model, one now could divide lymphokines and cytokines into differentiogenic, mitogenic, and activating signals. Until present, only cumulative effects on macrophages of a cocktail of lym phokines have been studied. Using purified lymphokines in macrophage populations which are functionally characterized according to our model, we should be able to define the precise nature of a lymphokine signal and the nature of the cellular response to it. While it is still a widespread belief that MIF and macrophage activating factor are different manifestations of the same molecule, our data described above suggest that they are different molecules. On the basis of our model we would predict that differentiation, activation, and homeostasis of the mononuclear phagocytic system 1S regulated by a complex network of lymphokines and other factors.
References 1. BLOOM, B. R., and B. BENNETI. 1966. Mechanism of a reaction in vitro associated with delayed-type hypersensitivity. Science, 153: 80. 2. SORG, c., and B. R. BLOOM. 1973. Products of activated lymphocytes. 1. The use of radiolabelling techniques in the characterization and panial purification of the migration inhibitory factor of the guinea pig. J. Exp. Med. 137: 148. 3. SORG, C. 1975. Radiolabelling and characterization of the products of activated mouse lymphocytes. Eur. J. Biochem. 55: 423. 4. SORG, C. 1978. Use of a radioactive double labeling technique in the chemical analysis of the mediators of cellular immunity. J. Immunol. Methods 19: 173. 5. SORG, C., and W. KLINKERT. 1978. Chemical characterization of products of activated lymphocytes. Fed. Proc. 37: 2748. 6. KLINKERT, W., and C. SORGo 1980. Characterization of four lymphocyte activation products of guinea pig associated with macrophage migration inhibitory activity. Molec. Immunol. 17: 555. 7. SORG, C. 1980. Characterization of murine macrophage migration inhibitory activities (MIF) released by concanavalin A stimulated thymus or spleen cells. Molec. Immunol. 17: 565. 8. SORG, C. 1979. The biochemistry and in vitro activity of soluble factors of activated lymphocytes. Molec. Cell. Biochem. 28: 149. 9. SORG, c., and C. L. GECZY. 1978. Antibodies to guinea pig lymphokines VII. Reactivity with products of lymphoid and nonlymphoid cells. J. Immunol. 121: 1199. 10. SCHROFF, G., CH. NEUMANN, and C. SORGo 1981. Transglutaminase as a marker for subsets of murine macrophages. Eur. J. Immunol. 11: 637. 11. KLIMETZEK, V., and C. SORGo 1977. Lymphokine induced production of plasminogen activator by macrophages. Eur. J. Immunol. 7: 185. 12. KLIMETZEK, V., and C. SORGo 1979. The production of fibrinolysis inhibitors as a parameter of the activation state in murine macrophages. Eur. J. Immunol. 9: 613.
360 . C. SORG 13. SORG, c., CH. NEUMANN, V. KLIMETZEK, and D. HANNICH. 1980. Lymphokine induced modulation of macrophage functions. In: Mononuclear Phagocytes - Functional Aspects (R. van Furth, ed.). Martinus Nijhoff Publishers, The Hague, p. 539. 14. NEUMANN, CH., and C. SORGo 1977. Immune interferon. I. Production by lymphokine activated murine macrophages. Eur. J. Immunol. 7: 719. 15. NEUMANN, CH., and C. SORGo 1978. Immune interferon. II. Different cellular sites for the production of murine macrophage migration inhibitory factor and interferon. Eur. J. Immunol. 8: 582. 16. SORG, c., and CH. NEUMANN. 1981. A developmental concept for the heterogeneity of macrophages in response to lymphokines and other signals. In: Lymphokines (E. Pick, ed.), Vol. 3, Academic Press, New York, p. 85. 17. NEUMANN, CH., and C. SORGo 1980. Sequential expression of functions during mac rophage differentiation in murine bone marrow liquid cultures. Eur. J. Immunol. 10: 834. 18. NEUMANN, CH., and C. SORGo 1981. Heterogeneity of murine macrophages in response to interferon inducers. Immunobiology 158: 320. 19. NEUMANN, CH., and C. SORGo 1981. Independent induction of plasminogen activator and interferon in murine macrophages. J. Reticuloendothel. Soc. 30: 79. 20. NEUMANN, CH., and C. SORGo 1980. Production of plasminogen activator and interferon by various murine macrophages in relation to cell proliferation. Immunobiology 157: 258. 21. MACKANESS, G. B. 1969. The influence of immunologically committed lymphoid cells on macrophage activity in vivo. J. Exp. Med. 129: 973. 22. HIBBS, J. B. Jr., R. R. TAINTOR, H. A. CHAPMAN Jr., and J. B. WEINBERG. 1977. Macrophage tumor killing: influence of the local environment. Science. 197: 279. Dr. C. SORG, Department of Experimental Dermatology, Universitats-Hautklinik, 4400 Munster, Federal Republic of Germany
Modulation of Macrophage Functions by Lymphokines C. SORG
Abstract Macrophage migration inhibitory factors (MIFs) of mouse and guinea pig have been thoroughly characterized with regard to molecular weight and isoelectric points. Several molecular weight species have been identified. In a comparative study with purified MIFs it was found that these molecules were distinct from a series of other lymphokines, particularly so from macrophage activating activities. Investigations on the molecular weight heterogeneity of MIF have led us to a transglutaminase-like activity which was found to be expressed in certain subsets of macrophages. The question whether low molecular weight factors are polymerized by this enzyme to oligomers is further investigated. Studies on the induction by lymphokines of interferon and plasminogen activator revealed a great heterogeneity of responding macrophages. In studies on the biological basis of the functional heterogeneity of macrophages, the question was investigated whether the heterogeneity was due to different macrophage subpopulations or to intermediate relatively stable phenotypes on their way to maturity and senescence. To approach this question, the bone marrow liquid culture system was used as a developing system. Our data are summarized in a unifying model which takes into account the different constitutive and inducible functions during the cell cycle. Accord ingly, lymphokines may act either as differentiation signals, as mitogens or activating signals.
Introduction Lymphocytes and macrophages are the main constituents of acute or chronic inflammatory cell infiltrates. In the past fifteen years it has been increasingly recognized that lymphocyte activation productions (lym phokines) may induce or modulate macrophage functions. The first lym phokine described was the macrophage migration inhibitory factor (MIF) (1). Since then, many other factors have been described, and the list of lymphokines currently numbers well beyond one hundred. In the years following its discovery, a great number of macrophage activating properties has been attributed to MIF.
Characterization of MIF Our laboratory has worked in the past years on the chemical and functional characterization of MIF from three different species (guinea pig, mouse, man) (2-8). Essentially from work on guinea pig MIF we have
Macrophages and Lymphokines· 353
established the following characteristics: MIF activity is associated with several molecular weight species. These have been identified at molecular weights of 60,000 a), 45,000 b), 30,000 c), and 15,000 d) daltons, all of which have an isoelectric point of about 5.2. The buoyant density in CsCI gradient was found at 1.278 for all four molecules. No proteolytic activity was found to be associated with either one of the molecules, nor did treatment with highly purified neuraminidase destroy the biological activity. Reduction and alkylation or treatment with EDTA did not change the molecular weights of a, b, or c. Because of the similarities and their chemical and biological properties, it was assumed that the three molecules are oligomers of a common subunit of the molecular weight of 15,000 and an isoelectric point of 5.2. Similar results have been obtained with murine and human MIF (unpublished data). In another study we investigated the cellular source of migration inhibi tory factor (9), since it had been reported earlier that MIF was also produced by nonlymphoid cells. Using an antilymphokine serum and radio labeled supernatants from various cells, migration inhibitory activity reacting with the antilymphokine serum and properties identical to those determined on MIF from lymphocytes was found in the following cell types: purified B-cells activated with Staphylococcus aureus, L 2C leukemia (strain 2), and growing embryonic fibroblasts. No activity was found in supernatants of peritoneal macrophages. The results indicate that MIF with similar molecular properties may be produced by a variety of cells under appropriate stimulation. This situation is reminiscent of the production of interferons by a diversity of cell types under different conditions. In order to determine further biological effects of MIF on macrophages, activities from Concanavalin A (Con A)-stimulated murine spleen cell supernatants were fractionated by Sephadex chromatography and isoelec tric focusing. Aliquots of these preparations were sent to a number of laboratories, each performing a particular test on macrophages. The func tional profile of the various MIF preparations showed that they were contaminated to a certain degree with other lymphokine activities, e.g. as expected T-cell replacing factor (TRF; tested by Dr. Schimpl) was a contaminant in preparation (c) of molecular weight of 28,000. Induction of leishmania kill (performed by Dr. Mauel) was only observed with prepara tion a) and not with preparations b), c), and d). Tumor cytotoxicity (performed by Dr. Kniep) could be induced with all four preparations, however, we would not conclude that MIF is identical with a macrophage cytotoxic factor, since our preparations were not endotoxin-free, and the observed effect could have been caused by a cooperative effect of endotoxin and MIF. The experiment further proved quite clearly that MIF has no chemotactic activity and is not related to interferon. MIF is also not an inducer of interferon. From these experiments we would conclude that MIF is not identical with most macrophage activating activities which was the widely accepted assumption of recent years.
354· C. SORG
Functional heterogeneity of macrophages During work on the purification and chemical characterization of mac rophage migration inhibitory factor of guinea pig and mouse, we observed that inhibitory activity was associated with several distinct molecules which seemed to be oligomers of a small subunit of approximately 14,000 to 15,000 daltons. Since formation of aggregates and disulfide bridges could be excluded, the involvement of a polymerizing enzyme was suspected. In fact, low amounts of transglutaminase-like activity could be detected in crude, concentrated supernatants of activated lymphocytes, which was not associ ated with MIF activity. It was found that erythrocytes, granulocytes, thymocytes or lymphocytes with or without Concanavalin A stimulation contained no enzyme activity (10). Normal peritoneal wash-out mac rophages in peritoneal exudate cells stained positive for transglutaminase to varying degrees, whereas macrophages derived from bone marrow liquid cultures were 20 % positive at day 3 and 100 % at day 17. Promonocytes and monocytes were negative. Positively stained cells, as detected by incorporation of dansylcadaverine also phagocytosed opsonized sheep ery throcytes. The degree of staining varied considerably in the macrophage like cell lines IC 21 (100 %), J 774.2 (77 %), P 388 Dl (50 %). This result and those from autoradiography studies indicate that expression of trans glutaminase is not associated with the S-phase of the cell cycle. The enzyme is CaH dependent, and it appears neither to be on the outer cell surface nor being released into the culture medium. Circumstantial evidence indicates that it is also not compartmentalized in cytoplasmic vesicles. The question whether this enzyme is responsible for the polymerization of low molecular weight MIF has not been resolved yet. Since trans glutaminase seems not to be released into the culture supernatant, it is possible that polymerization of proteins is an artefact due to the release of the enzyme from dead cells in culture. While the induction and modulation of enzyme expression is still under study, it is concluded that trans glutaminase is a newly described marker for macrophages of a certain differentiation or activation state. Macrophages after exposure to lymphokines alter their behaviour in a sense that they show either enhanced, depressed, or newly induced func tions. Our interest was directed at the induction and modulation of plasminogen activator production by murine macrophages (11). Mac rophage monolayers obtained by 24 hour culture of proteose peptone elicited murine exudate cells were incubated with lymphocyte culture of supernatants. After 48 hours, the supernatants were replaced by serum-free medium, and the macrophages were incubated for another 24 to 48 hours. These supernatants were assayed for plasminogen activator using the lysis of 125I-Iabeled fibrin. It was found that supernatants of antigen or mitogen stimulated spleen cells induced the plasminogen activator secretion, whereas control supernatants were ineffective. The same was found with supernatants of mitogen stimulated lymph node cells. Lymphokine induced
Macrophages and Lymphokines· 355
plasminogen activator secretion by macrophages was enhanced after phagocytosis of latex beads. In the course of these studies on the lymphokine induced production of plasminogen activator it was observed that some macrophage culture supernatants not only lacked plasminogen activator but were even inhibi tory to plasmin dependent fibrinolysis (12). This observation and the scarcity of reports on protease inhibitor production by macrophages prompted us to investigate this subject more systematically. Macrophages from peritoneal exudates which had been induced by various irritants were cultured, and the supernatants were tested for inhibitors of fibrinolysis using the plasmin dependent lysis of 125I-Iabeled fibrinogen as assay. Wash out macrophages and casein or proteose peptone elicited macrophages were found to release fibrinolysis inhibitors in contrast to lipopolysaccharide or thioglycollate induced macrophages. The molecular weight of inhibitors was determined at 60,000, 45,000 and 15,000 daltons by Sephadex chromatography. Whereas the inhibitors at 15,000 and 45,000 daltons could be detected in all experiments, inhibitor at a molecular weight of 60,000 daltons was not present in all preparations. The isoelectric point of inhibitors was determined at 4.15. The proteolytic and esterolytic activity of trypsin and chymotrypsin were both inhibited by each of the two inhibitors. On the other hand, only the proteolytic activity of plasmin could be inhibited. Evidence for the active synthesis of inhibitors by macrophages came from several experiments. Macrophages in serum-free cultures continued to release inhibitors for at least 48 hours; normal mouse serum did not contain inhibitors of the same molecular size, and 3H-Ieucine was incorporated into the inhibitors which were specifically detected by the absorption of plasmin inhibitor complexes to lysyl-sepharose. Since inhibitors and plasminogen activator could not be detected in the same macrophage culture supernatants, it appears that the production of inhibitors and plasminogen activator by the same macrophage population is mutually exclusive. In a subsequent study we investigated whether the production of fi brinolysis inhibitors is modulated under the influence of lymphokines (13). Supernatants from mitogen stimulated spleen cells were fractionated on Sephadex G-75, and the fractions were incubated for 24-48 hours with various types of peritoneal exudate cells. It was found that proteose peptone elicited cells were easily converted from an inhibitor to a plasminogen activator producing state, whereas casein elicited cells stopped producing fibrinolysis inhibitors but could not be converted to the plasminogen activator producing state. Thioglycollate induced cells which already pro duced plasminogen activator were enhanced by the lymphokines. Lipopolysaccharide elicited macrophages which did not produce fibrinoly sis inhibitors were refractory to the modulating effects of lymphokines. Lymphokine activity was found in a molecular weight range of about 65 to 20,000 daltons.
356 .
C.
SORG
Another line of interest was developed from the question what role macrophages play in the production of interferon in combined T lympho cyte-macrophage cultures. It was found that unfractionated murine spleen cells produce immune interferon upon stimulating with antigen or mitogen (14, 15). When spleen cells were passed over glass bead columns, interferon production decreased, while the response to the stimulants drastically increased. When these cells were further purified over nylon wool columns, interferon production was totally abolished, whereas thymidine incorpora tion in stimulated cultures was invariably high. Interferon production by nylon wool column purified lymphocytes could be restored with mac rophages grown from bone marrow cultures or spleen cells, but not with macrophages from proteose peptone induced peritoneal exudate cells. It was also found that pure macrophage cultures from the spleens of BCG (Bacille Calmette Guerin) immunized mice consistently produced interfe ron activity without any further stimulation. When culture supernatants of activated T lymphocytes which did not contain any interferon activity were transferred to macrophage cultures from different sources and incubated for 45 hours, interferon activity could be detected in supernatants of mac rophage cultures from bone marrow and spleen, but not in those from proteose peptone induced peritoneal exudate cells. It is concluded that certain macrophage populations can be induced to produce interferon activity, whereas others are refractory to this induction which appears to be linked to their differentiation state. As described above, macrophages display a great heterogeneity in response to lymphokines. Functional heterogeneity of macrophages has been documented before in many cases. The question what biological principle determines macrophage heterogeneity has not been answered yet. One assumption is that we are dealing with true subpopulations as it is known for the T cell/B cell lineages (16). Another possibility would be that macrophages which are bone marrow derived pass through different ma turation stages or intermediate relatively stable phenotypes on their way to maturity and senescence, thus expressing characteristic functions in each maturation stage. In order to approach this problem, we adopted the bone marrow liquid culture system (17). Bone marrow cells from mice were cultured on teflon membranes in the presence of a colony-stimulating factor produced by L cells. Parallel cultures were harvested daily, and a series of functions was recorded over 12 days. The growth curve was declining rapidly after onset of culture until days 2-3 due to the death of granulocytes and erythrocyte precursors. At days 4-5, the yield of cells increased, reaching a maximum at days 6-8 and then declined again. By Pappenheim staining, the first appearance of macrophages could be seen at day 2 which increased linearly and reached a plateau at day 6. When the cells were stained for unspecific esterase (a-naphthyl acetate), weakly stained cells peaked at days 2 and 3 and again at days 5 and 6, whereas heavily stained cells appeared for the first time at
Macrophages and Lymphokines' 357
day 3 and then again at day 6 and later, reaching a maximum of nearly 90 % at day 9. The phagocytosis of latex beads was observed very early in culture, reaching a plateau at days 7-8. Phagocytosis of immunoglobulin coated sheep red blood cells was also performed with adherent cells. Phagocytic activity was apparent at day 2, decreasing by day 3 and reaching again a maximum at days 5 and 6. From day 7 on, only few cells phago cytosed 3 or more red blood cells. In another series of experiments, the response to three different stimuli was investigated. The chemotactic response against zymosan activated serum was measured in Boyden chambers using cellulose nitrate filters. Cells which penetrated the filter were differentiated for granulocytes and mononuclear cells. Results showed a sharp maximum at days 3-4 for the chemotactic response of mononuclear cells. No chemotactic response was observed after day 5. When the response to MIF containing supernatants of Concanavalin A-stimulated spleen cells was measured, again a sharp maxi mum at days 4 and 5 was found, whereas no response could be detected before and after this period, even though the cells were vigorously migrat ing in the earlier days. Since interferon production by lymphokine induced macrophages requires an incubation time of 24-48 hours and since this time would be too long in order to draw any conclusion on the momentous differentiation state of a cell culture, it was decided to use the LPS-induced production of interferon which -like migration inhibition and chemotaxis requires only three to four hours of incubation. When tested daily, interfe ron induction by LPS was successful only at days 5 and 6. As a further function, the plasminogen activator production was recorded. Plasminogen activator was detected with a fibrinolysis assay which allows the simultaneous detection of fibrinolysis inhibitors (12). No plasminogen activator or significant amounts of fibrinolysis inhibitors were found during the first three days of culture. Beginning at day 4, plasmino gen activator was readily produced reaching a maximum at days 6 and 7 and then decreased rapidly. At day 8, the cells had ceased to produce plasmino gen activator, and at day 9 they began to release fibrinolysis inhibitors for the rest of the culture period. When we surveyed our data, we noticed that those macrophages either derived from bone marrow liquid cultures or from peritoneal exudates which did not produce plasminogen activator could also not be induced to release interferon neither by LPS nor by lymphokines. Further, it was observed that those macrophages that responded best to LPS in terms of interferon production were close to or at the plasminogen activator produc ing state (18, 19). We therefore asked whether plasminogen activator production might be a constitutive trait of those macrophages which can be induced to interferon production. In a series of experiments we found an almost perfect correlation between these two parameters, i.e., macrophage populations which produced plasminogen activator within 25 hours of culture could be induced to release interferon. From these observations we
358 . C. SORG
would have to conclude that certain macrophages have the ability to produce plasminogen activator and are inducible to interferon production at the same time. Whether we are dealing with subpopulations or pheno types of different maturation stages we still cannot decide. Macrophages may be induced to plasminogen activator production by the tumor promotor phorbol myristate acetate (PMA), which is also mitogenic to macrophages. The following question was asked: are mac rophages pushed by PMA and similar signals such as lymphokines into a certain phase of the cell cycle in which plasminogen activator production is expressed? And further, what role plays the cell cycle in macrophage heterogeneity? The following experiments were designed to shed some light on this question (20). Macrophages produce substantial amC'unts of plasminogen activator 2-22 hours after exposure to PMA. The production of plasminogen activator then steadily decreases. In contrast, the number of cells in S-phase peaks later at 26-46 hours. This indicates that macrophages after exposure to PMA are pushed towards the S-phase of the cell cycle. On their way to S, probably in late G!, macrophages express plasminogen activator as a constitutive trait and also become sensitive to activation by lipopolysac charides. Based on our sllmmarized data and those from different laboratories, we are able to describe the following picture: first, what is likely to happen in the bone marrow liquid culture system? Macrophages differentiate from precursors in the presence of colony-stimulating factor. After a phase of intensive proliferation of precursors and differentiation into macrophages, the young macrophages still keep on cycling, thus accumulating cells in late G!, which are characterized by plasminogen activator production, induci bility of interferon, and their response to MIF and certain chemotactic factors. When proliferation is fading away, the cells differentiate gradually in G! loosing a series of constitutive functions and pass on to a Go-like state characterized by production of fibrinolysis inhibitors. In this model, the functional state of normal resident macrophages should be compatible with Go, that of proteose peptone elicited macrophages with «early» and thiogly collate induced macrophages with «late» GI" In kinetic studies with PMA, the time needed to induce normal resident macrophages to plasminogen activator production was considerably longer than that needed for proteose peptone elicited macrophages. The model proposes further that mac rophage differentiation in G! is reversible and that macrophages of various functional states have to go through the bottle neck of «late» G! before entering S. What then is an activated macrophage, and how can the action of lymphokines be fitted in? As shown first by MACKANESS (21) and subse quently by many others, activated macrophages express special functions, in particular bactericidal and tumoricidal properties. Several investigations have demonstrated that macrophages can be rapidly activated by lipopoly-
Macrophages and Lymphokines· 359
saccharides to kill tumor cells (22). However, not all macrophages respond the same. Nonresponding macrophages can be made responsive, however, by preincubation with lymphokines which serve as preparatory signals. In our model this would mean that macrophages are pushed first from Go or «early» G 1 into «late» G 1, where macrophages become inducible to kill tumor cells. These findings are paralleled by ours on the induction of interferon by LPS. According to the proposed model, one now could divide lymphokines and cytokines into differentiogenic, mitogenic, and activating signals. Until present, only cumulative effects on macrophages of a cocktail of lym phokines have been studied. Using purified lymphokines in macrophage populations which are functionally characterized according to our model, we should be able to define the precise nature of a lymphokine signal and the nature of the cellular response to it. While it is still a widespread belief that MIF and macrophage activating factor are different manifestations of the same molecule, our data described above suggest that they are different molecules. On the basis of our model we would predict that differentiation, activation, and homeostasis of the mononuclear phagocytic system 1S regulated by a complex network of lymphokines and other factors.
References 1. BLOOM, B. R., and B. BENNETI. 1966. Mechanism of a reaction in vitro associated with delayed-type hypersensitivity. Science, 153: 80. 2. SORG, c., and B. R. BLOOM. 1973. Products of activated lymphocytes. 1. The use of radiolabelling techniques in the characterization and panial purification of the migration inhibitory factor of the guinea pig. J. Exp. Med. 137: 148. 3. SORG, C. 1975. Radiolabelling and characterization of the products of activated mouse lymphocytes. Eur. J. Biochem. 55: 423. 4. SORG, C. 1978. Use of a radioactive double labeling technique in the chemical analysis of the mediators of cellular immunity. J. Immunol. Methods 19: 173. 5. SORG, C., and W. KLINKERT. 1978. Chemical characterization of products of activated lymphocytes. Fed. Proc. 37: 2748. 6. KLINKERT, W., and C. SORGo 1980. Characterization of four lymphocyte activation products of guinea pig associated with macrophage migration inhibitory activity. Molec. Immunol. 17: 555. 7. SORG, C. 1980. Characterization of murine macrophage migration inhibitory activities (MIF) released by concanavalin A stimulated thymus or spleen cells. Molec. Immunol. 17: 565. 8. SORG, C. 1979. The biochemistry and in vitro activity of soluble factors of activated lymphocytes. Molec. Cell. Biochem. 28: 149. 9. SORG, c., and C. L. GECZY. 1978. Antibodies to guinea pig lymphokines VII. Reactivity with products of lymphoid and nonlymphoid cells. J. Immunol. 121: 1199. 10. SCHROFF, G., CH. NEUMANN, and C. SORGo 1981. Transglutaminase as a marker for subsets of murine macrophages. Eur. J. Immunol. 11: 637. 11. KLIMETZEK, V., and C. SORGo 1977. Lymphokine induced production of plasminogen activator by macrophages. Eur. J. Immunol. 7: 185. 12. KLIMETZEK, V., and C. SORGo 1979. The production of fibrinolysis inhibitors as a parameter of the activation state in murine macrophages. Eur. J. Immunol. 9: 613.
360 . C. SORG 13. SORG, c., CH. NEUMANN, V. KLIMETZEK, and D. HANNICH. 1980. Lymphokine induced modulation of macrophage functions. In: Mononuclear Phagocytes - Functional Aspects (R. van Furth, ed.). Martinus Nijhoff Publishers, The Hague, p. 539. 14. NEUMANN, CH., and C. SORGo 1977. Immune interferon. I. Production by lymphokine activated murine macrophages. Eur. J. Immunol. 7: 719. 15. NEUMANN, CH., and C. SORGo 1978. Immune interferon. II. Different cellular sites for the production of murine macrophage migration inhibitory factor and interferon. Eur. J. Immunol. 8: 582. 16. SORG, c., and CH. NEUMANN. 1981. A developmental concept for the heterogeneity of macrophages in response to lymphokines and other signals. In: Lymphokines (E. Pick, ed.), Vol. 3, Academic Press, New York, p. 85. 17. NEUMANN, CH., and C. SORGo 1980. Sequential expression of functions during mac rophage differentiation in murine bone marrow liquid cultures. Eur. J. Immunol. 10: 834. 18. NEUMANN, CH., and C. SORGo 1981. Heterogeneity of murine macrophages in response to interferon inducers. Immunobiology 158: 320. 19. NEUMANN, CH., and C. SORGo 1981. Independent induction of plasminogen activator and interferon in murine macrophages. J. Reticuloendothel. Soc. 30: 79. 20. NEUMANN, CH., and C. SORGo 1980. Production of plasminogen activator and interferon by various murine macrophages in relation to cell proliferation. Immunobiology 157: 258. 21. MACKANESS, G. B. 1969. The influence of immunologically committed lymphoid cells on macrophage activity in vivo. J. Exp. Med. 129: 973. 22. HIBBS, J. B. Jr., R. R. TAINTOR, H. A. CHAPMAN Jr., and J. B. WEINBERG. 1977. Macrophage tumor killing: influence of the local environment. Science. 197: 279. Dr. C. SORG, Department of Experimental Dermatology, Universitats-Hautklinik, 4400 Munster, Federal Republic of Germany