0022-202X/ 80/ 7405-030 I$ 02.00/ 0 THE JOUHNAL OF I N VESTIGATIVE DEHMATOLOGY, 74:30 1-306, 1980 Vol. 74 , No. 5 Copyright © 1980 by The Williams & Wilkins Co. Printed in U. S.A. Immunologic and Nonimmunologic Activation of

J. L. TuRK, M.D., D.Sc. Department of Pathology, Royal Co llege of Surgeons of E ngland, Lincoln's Inn Fields, London

Cells of the mononuclear system take on tem. However epithelioid cell are generally found differ ent morphological features and appear to have only when host resistance is high and are also a pal'ticular different functions in immunological and nonimmuno­ featlll'e of a s ituation where there is a strong hypersensitivity logical granulomas. In imm•mological granulomas the reaction. This type of is typical of the chronic appearance is of epithelioid cells. Epithelioid cells are of inflammatory reaction associated with other immunological 2 types. In borderline tuberculoid granulomas, reactions, pal'ticularly to metals and infecting organisms such the appearance is of activated macrophages. Similar as schistosomes. activation can be induced by lymphokine which in­ In contrast the granulomas of lepromatous leprosy, which creases respiratory enzyme activity. Other types of epi­ occUl' in patients with low host resistance, consist of undiffer­ thelioid cell are found in experimental zirconium gran­ entiated macrophages that have taken up large numbers of ulomas. These cells contained rough endoplasmic retic­ mycobacteria which they are unable to eliminate. These gran­ ulum, indicating that they may play a secretory role. It ulomas are n.ot associated with a lymphocytic infiltration and is suggested that these cells could play a part in stimu­ do not have the regular appearance of the tuberculoid lesion. lating fibrosis. Other substances that produce nonim­ Granulomas of this type which are typically not associated with munological granulomas, such as aluminum containing an immunological reaction, are also found in lipid storage compounds, are directly toxic to macrophages in vitro diseases and following the deposition of foreign material that resulting in the rapid release of cytoplasmic enzymes. does not induce a specific immunological reaction. Similar These compounds also activate complement, causing C3 granulomas may also be found following the deposition of conversion and anaphylatoxin production. C3 conver­ certain nonsensitizing metal compounds such as aluminum sion may be through pathways other than the classical hydroxide [2,3]. However, with other metal compounds that and alternative pathways a nd does not occur in the have a sensitizing capacity the pattern is for epithelioid cell absence of plasminogen. Mycobacteria can activate com­ granulomas with an associated immunological background. Ep­ plen ent through the alternative pathway, suggesting a ithelioid cell granulomas are found particularly in individuals mechanism for granuloma formation in lepromatous lep­ and experimental animals sensitized to sodium zil'conium lac­ rosy. Loss ofC3 membrane receptors from macrophages tate. in lepromatous leprosy could be reproduced by feeding peritoneal macrophages with mycobacteria in WHAT IS AN EPITHELIOID CELL? vitro. This was not just due to , as a similar The term epithelioid cell is derived from observations using receptor loss was not obtained when the cells were fed the light microscope. Classically these al'e large polygonal cells latex particles or zymosan. Epithelioid cells in tubercu­ with pale oval nuclei and abundant eosinophilic cytoplasm loid leprosy and lose Fe m embran e receptors whose borders blend imperceptibly with those of their neigh­ but retain C3 receptors. Thus epithelioid cells can be bors [ 4]. The origin of these cells from other cells of the MPS readily distinguish ed from other cells of the mononu­ was fu·st recognized by Metchnikoff [1]. Cells of this type may clear phagocyte series. be found in BCG granulomas in experimental mice [ 4] or granulomas produc.ed by the intradermal injection of sodium zirconium lactate in sensitized guinea pigs [3,5]. Epithelioid A granuloma may be defined as a chronic inflammatory cells a r e typical of immunological granulomas although they reaction containing a predominance of cells of the mononuclear may be produced nonimmunologically by cellophane implants phagocyte series (MPS). Granulomas may take on 2 types of in mice. Electron microscopical examination of epithelioid cell histological appearance depending on the nature of the stimu­ granulomas may reveal cells with a distinctive ultrastructure. lus. In immunologically stimulated granulomas such as those The nucleus of these cells is round or oval with finely mru·gin­ that can occm in certain chronic infectious diseases such as ated chromatin and large, spherical reticulated nucleoli. There or in tuberculoid leprosy, cells of the MPS are is frequently a lru·ge amount of endoplasmic reticulum and converted to take o n a morphological appearance described as mitochondTia are well developed, sometimes swollen. The cy­ "epithelioid cells" [1]. These cells may also aggregate to form toplasm contains a number of vesicles and vacuoles. Phagocy­ giant cells. The immunological origin of these granulomas is tosis occms but need not be a reg.ulru· feature. Typically the ge numbers of emphasized by the presence of lal' cell-membrane is fimbriated and junctional interdigitations which sunound the central epithelioid cell area. These granu­ with neighboring cells may be seen (Fig 1) . and may be the site of lomas contain numerous When one compares the ultrastructural appearance of these considerable fibrosis. Epithelioid cells are poorly phagocytic cells with other cells of the MPS, one notices a definite devel­ ononucleal' phagocyte sys- compared with other cells of the m opmental sequence from circulating , through tissue undifferentiated , through activated macro­ Reprint requests to: J. L. Turk, Department of Pathology, Royal phage, to the epithelioid cell seen in these lesions. These c ells College of Surgeons of England, Lincoln's Inn Fields, London WC2A can be classified according to the amount of rough endoplasmic 3PN, V.K. reticulum, lysosomes, mitochondria, Golgi apparatus, phago­ Abbreviations: somes, fimb1·iated cell periphery and microftlaments. It can be ACH: aluminum chlorhydrate AlOH: aluminum hydroxide seen that there is a progressive increase in the amount of these EA: sheep erythrocytes and antibody elements as one passes from the c ir culating monocyte to the EAC: sheep eryth1·ocytes antibody a nd complement epithelioid cell (Table I). The only exception is in the degree of MPS: mononuclear phagocyte series ' phagocytosis, as epithelioid cells ru·e poorly phagocytic. ZAG: zirconium aluminum glycine The question therefore arises as to the evolution of the

301 302 TURK Vol. 74, No. 5

\1 \ II It \ 1 [11'1. 111 li.J.J 'o 111 J ILl \ 111'1.0 \1 I 1.1 \ It 1'11\1 I I\\ I I '>\'I l \1

\u•w•u 1 ~- ~---- ~~'~; '1 - L y ouphhC)'h' .. -•• .11.

M••rHorytr - \L ;o c r n!>hal!t'

It•·•·

FIG 2. Maturation of cells of the mononuclear phagocyte se ries.

FIG 1. Epithelioid cell with fenestrated cetl periphery, fimbriated and interdigitating cetl membrane, relatively abundant endoplasmic reticulum and polyribosomes. X 10,000. (5)

TABLE l. Ultra~;tructure of the mononuclear phagocyte system Circu- Macro- Acti- Epithe- lating phage vated lioid mono- macro- cell cyte phage RER" + ++ ++ ++++ Lysosomes + ++ +++ ++++ Mitochondria + + ++ ++++ Golgi apparatus + ++ +++ ++++ Phagosomes + ++++ + Fimbriated ce tl periphery + ++ ++ ++++ 0 Control fraction added Micro filaments + ++ +++ ++++ ~ Lymphokine " "RER rough endoplasmic reticulum.

epithelioid cell and how it arises in immunological· granulomas. .P< 0·001 In clinical situations such as tuberculoid leprosy, epithelioid P<0·001 100 cell granuloma formation in the skin can be directly associated with the development of delayed hypersensitivity. One of the possible ways in which changes might be induced could there­

FIG 7. Epithelioid cell fi·om Mitsuda reaction containing abundant rough endoplasmic reticulum and ingested heat killed bacilli. X 9,000.

intradermal injection of dead M. leprae as part of the Mitsuda skin test reaction in the same patient. The mononuclear phag­ ocytes in the lesions are characterized by the presence of many membrane bound vacuoles which contained material of varying electron density (Fig 6). The appearance was of activated macrophages. No obvious bacteria could be found intracellu­ FIG 5. Peritoneal exudate macrophage cultured for 14 days. x 9,000. larly and there was no endoplasmic reticulum. Thus these cells ofthe MPS in classically accepted epithelioid cell granulomas turned out to be typical activated macrophages without any secretory capacity. None of the cells previously found to be typical of epithelioid cell granulomas with rough endoplasmic reticulum and fimbriated cell margin was present in these lesions. It was interesting however, that phagocytic cells with abundant endoplasmic reticulum were seen in the granuloma of the Mitsuda skin reaction (Fig 7) in one of the patients in whom the clinical lesions contained cells of the MPS devoid of rough endoplasmic reticulum. Consideration might also be given to the role of epithelioid cells in immunological granulomas. Granulomas containing ep­ ithelioid cells with rough endoplasmic reticulum are frequently also found to contain large numbers of fibroblasts and be the centres of a considerable amount of flbrogenesis. Fibrosis is an important end-point of the natural history of epithelioid cell granulomas. The question therefore arises as to whether epi­ thelioid cells play an important role in fibrosis. Particularly, could those epithelioid cells with an endoplasmic reticulum be secreting a stinmlating factor. Such a factor might stimulate fibroblast proliferation or increase the synthesis of collagen by these cells. This however does not clarify the difference between the 2 types of epithelioid cell granulomas into those containing activated macrophages and those contain­ ing cells with a strong secretory capacity.

CHANGES IN MACROPHAGES IN NON­ IMMUNOLOGICAL GRANULOMAS Recently it has been found that a number of aluminum and zirconium containing compounds could cause foreign body gran­ FIG 6. Epithelioid cell fi·om borderline tuberculoid lesion containing ulomas when injected intradermally into guinea pigs [8]. These large numbers of intracellulru· vacuoles. X 10,750. granulomas contained either undifferentiated macrophages or 304 TURK Vol. 74, No.5 presented a pattern in which th ere was basophilic degeneration 100 of collagen associated with a p leomorphic g iant cell infiltrate. The effect of a number of th ese compounds, aluminum hydrox­ ide (AlOH), aluminum chlorhydrate (ACH), zirconium alumi­ num glycine (ZAG) and zirconium hydroxide, were compared 80 on macrophages and fibroblasts in culture [9]. Cell death was AI(OHl3 measured by the release of lactate dehydrogenase. Al(OH)a, ACH and ZAG at a concentration of 1 mg/ml were found to have a s imilar toxicity to silica at 200 J.L g/ml in culture (Fig 8). l.f) 60 ic. Zr(OH),., a nongranuloma producing compound, was not tox l.f) The toxic effect on fibroblasts was similar. Electron microscopic >- _j ACH examination revealed that the cells had taken up the gel into 0 large phagosomes (Fig 9). In some cells t h e gel appeared not to :::E L.O be confined to membrane limited vacuoles and was found w apparently lying free in the cytoplasm. These compounds were <{ I ZAGS '

"-"::::::;;....J--.....,.-==-t===-=9 Z r ( 0 H) L. 80 1mg / ml 200 600 1000 CONCENTRATION (J,Jg/ml) 0·5mg/ml w FIG 10. Haemolysis induced b y AI and Zr compounds. Guinea pig 2: 60 were incubated with gels for 4 hr at >- erythrocyte suspensions (1% v/ v) N 37° in medium 199. (9) z 0 · 25mg / ml w w Table II. Anaphylatoxin production by various compounds in the _J presence or absence of 10 mM EDTA CD 0 ·1 mg/ml ::J Contraction of guinea pig ileum as a percentage of contraction _J 0 induced by 0.1 l'g " V1 CONTROL Compou nd added Chelator lo serum 0~ mgml Nil EDTA 10 mM Zymosan 2 78.8 ± 5.3 0 Inulin 20 88.3 ± 10.4 0 L.O 50 10 20 30 Kaolin 50 78.3 ± 21.3 88.3 ± 10.7 TIME IN HOURS Aluminium chlorhydrate 10 106.1 ± 17.2 75.3 ± 10.8 FIG 8. LDH release from macrophages cul tured in the presence of Aluminium hydroxide 15 90.7 ± 21.6 106.7 ± 15.0 Al(OH):. or silica (200 l'g/ ml). (9) Sodium zirconium lactate 5 80.7 ± 14.4 77.3 ± 11.0 Zirconium a luminium glycin- 10 93.5 ± 11.9 92.8 ± 13.1 ate Zirconium hydroxide 10 0 0 " Mean ± standard error of fo ur replicates.

found to act directly on cell membranes and were directly h aemolytic when mixed with guinea pig erythrocytes (Fig 10). Using nontoxic doses of the compounds t here was no greater selective release of lysosomal enzymes (N-acetyl {3-D glucosa­ minidase, {3-glycerophosphatase and cathepsin D) than could be detected in control cultures. A parallel has been shown to exist between the induction of chronic by s uch compounds as carageenan and their ability to release lysosomal enzymes selectively from macrophages and activate the alternative complement pathway [10]. A number of aluminum and zirconium compounds which had been fo und in previous studies to cause chronic inflamma­ tion were therefore studied for their ability to activate C3 in vitro. It was found that kaolin, Al(OHb, ACH, ZAG and sodium zirconium lactate would cause the release of anaphylatoxin from normal guinea pig serum (Table II) [11]. Zr(OH). that produced only minimal inflammation did not activate C3. All the compounds that produced anaphylatoxin were shown to produce C3b as demonstrated by 2-dimensional electrophoresis (Fig 11). Both anaphylatoxin production and C3 conversion were shown not to be Ca++ and Mg++ dependent and therefore could not have taken place through the classical or alternative FIG 9. Macrophage c ultured with AI( OH):. for 24 hours. Gel does pathways. Removal of plasminogen using a lysine-Sepharose not appear to be confined to membrane li mited vacuoles. X 6,000. (9) column [1 2] showed that C3 conversion by some of these metal May 1980 ACTIVATION OF MACROPHAGES 305

containing compounds (Al(OHh ACH and ZAG) was depen­ Peritoneal macrophages from guinea pigs were cultured for dent on plasminogen. 7 days in the presence of the mycobacteria of BCG vaccine. An It would therefore appear that a pathway exists for nonanti­ 80% decrease in the number of EAC receptors was seen at 24 h. genic compounds to induce chronic inflammation through a C3 100% receptor activity returned after 7 days of culture. Neither conversion pathway. The likelihood exists that other nonim­ zymosan nor latex caused a change in EA or EAC receptor over munological granulomas such as those of lepromatous leprosy the 7-day period. Thus the effect was not due to phagocytosis complement activation. Experiments in progress using BCG alone. No loss of EA receptors was found after treatmen t with vaccine and Mycobacterium lepraemurium indicate that my­ lymphokine, which did however partially decrease EAC recep­ cobacteria are potent activators of C3, both producing anaphy­ tors. Thus the loss of C3 receptors in lepromatous leprosy could latoxin and producing C3b as demonstrated by 2-dimensional be associated with the ingestion of mycobacteria. No clue, electrophoresis. however, could be obtained as to the mechanisms of Fe receptor loss from epithelioid cells [16]. MEMBRANE RECEPTORS ON CELLS OF THE MONONUCLEAR PHAGOCYTE SERIES REFERENCES Distinct membrane receptor sites fo r the Fe portion of IgG 1. MetchnikoffE: Lectures on The Comparative Pathology oflnflam­ mation. Kegan, Paul, Trench, Trubner. London, 1893, pp 159-166 and for C3 were discovered on human by Huber, et 2. Gaafar SM, Tmk JL: Granuloma formation in lymph nodes. J al [13]. In experimental animals Mariano, Nikitin and Malucelli Pathol 100:9- 20, 1973 [14] found that macrophages on implanted coverslips lost their 3. Turk JL, Parker D: Granuloma formation in normal guinea pigs Fe but not their C3 receptors when they transformed into injected intradermally wi th aluminum and zirconium compounds. J Invest D ermatol 68:336-340, 1977 epithelioid cells. As part of our study on the changes that occur 4. Papadimitriou JM, Spector WG: The origin, properties and fate of in cells of the mononuclear phagocyte series in granulomas, we epit helioid cells. J Pathol 105:187-284, 1971 examined sections of skin of patients across the clinical spec­ 5. Tmk JL, Badenoch-Janes P, Parker D: Ultrastructural observation trum of leprosy for Fe and C3 receptors. Cryostat sections were on epithelioid cell granulomas induced by zirconium in the guinea pig. J Pathol 124:45-49, 1978 layered with sheep erythrocytes and antibody (EA) for Fe 6. Nath I, Poulter LW, T urk JL: Effect of mediators on receptors and sheep erythrocytes antibody and complement macrophages in vitro. A correlation of morphological and cyto­ (EAC) for C3. EAC but not EA adherence was seen at the BT chemical changes. Clin Exp Immunol 13:455-466, 1973 end of the leprosy spectrum and was particularly dense around 7. Poulter LW, Turk JL: Studies on t he effect of soluble lymphocyte products (lymphokines) on macrophage physiology. II. Cyto­ th e epithelioid cells. EA adherence was seen over the active chemical changes associated with activation. Cell Immunoi 20: macrophages of BL or LL leprosy. In LL leprosy, EAC adher­ 25-32, 1975 ence was poor or absent. The phenomenon seemed to be related 8. Tmk JL, Parker D: Sensitization with Cr, Ni and Zr salts and to bacterial density. Thus 2 patterns of membrane receptors allergic type granuloma fo rmation in the guinea pig. J Invest Dermatol 68:341-345, 1977 could be discerned. In epithelioid cell granulomas including 9. Badenoch-Janes P, Turk JL, Parker D: The effects of some alu­ sarcoidosis there was a loss of Fe receptors, whereas in nonim­ minum and zirconium compounds on guinea pig peritoneal mac­ munological granulomas of lepromatous leprosy there was a rophages and skin fibroblasts in cultme. J Pathol 124:51-62, 1978 loss of C3 receptors [15] (Table III). 10. Schorlemmer HU, Bitteleer-Suermann D, Allison AC: Complement activation by the alternative pathway and macrophage enzyme in the pathogenesis of chronic inflammation. Immunol­ ogy 32:929- 940, 1977 11. Ramanathan VD, Badenoch-Janes P, Turk JL: Complement acti­ vation by aluminum and zirconium compounds. , 37: SERUM + ZYMOSAN 881- 888, 1979 SERUM + BUFFER 12. Deutsch DG, Mertz ET: Plasminogen: Pmifi cation from human plasma by affinity chromatography. Science 170:1095-1096, 1970 13. Huber H , Polley MJ, Linscott WD, Miiller-Eberhru·d HJ: Human monocytes: D istinct receptor sites for the third component of complement and for immunoglobulin G. Science 162:1281-1283 1968 ' 14. Mariano M, Nikitin T , Malucelli BE: Immunological and nonim­ muno\ogical phagocytosis by inflammatory macrophages, epithe­ liOid cells and macrophage polykru·yons from foreign body gran- .r ulomata. J Patho\ 120:151- 160, 1976 · 15. R idley MJ, Ridley DS, Turk JL: Surface mru·kers on lymphocytes and cells of the mononucleru· phagocyte series in skin sections in leprosy. J Pathol 125:91-98, 1978 16. Ridley M, Turk JL, Badenoch-Janes P: In. vitro modification of membrane receptors on cells of the mononucleru· phagocyte SERUM-+ ACH SERUM+ ZAG system. J Patho\ 127:173-184, 1979

DISCUSSION GIGLI: I do not think that one should be surprised that all these compounds activate complement. It has been known for a long time that they can activate complement by mechanisms --...- .. "' - independent of immunologic activation. We are aware that there are a variety of very simple chemicals that are capable of FIG 11. Two-dimensional electrophoresis demonstrating the forma­ interacting with complement, either cleaving some of the com­ tion of C3b in serum treated with zymosan, aluminum chlorhydrate (ACH) and zirconium aluminum glycine. (11) ponents or altering their behavior. Potassium iodide, and por­ phyrins are good examples. TURK: We are not saying that we are original in showing that TABLE III. M embrane receptors on mononuclear phagocytes in these compounds interact with complements, but what we think leprosy is of interest is the concept that the plasminogen pathway is Test system TT BT BB BL LL important for granuloma formation. DvoRAK: Do you have any idea what your epitheloid cells EA ± ++ +++ ++ are secreting? EAC +++ +++ ++ ++ ± TURK: At the moment we're culturing granuloma and com- 306 TURK Vol. 74, No. 5

paring the effects of the supernatant of these cultures with deficient serum, you can see whether your metals work in that lymphokines. system. BYERS: Do you think this is a typical delayed type hypersen­ EPSTEIN: It seems to me that we have come to the same sitivity reaction, and if so can you transfer it to another guinea endpoint. The monocyte differentiates to an activated macro­ pig? Secondly, do you think that the delayed hypersensitivity phage and this cell can go on to further differentiation to a reaction is a prerequisite for the granuloma formation. secretory epitheloid cell. TURK: It takes a long time for the granulomas to form and CoHN: I wanted to comment on the epitheloid cell. A number they occur sporadically therefore we have not been able to get of years ago when we were looking at sarcoids, w e examined enough animals at one time to do transfer studies. The histo­ the progression of the formation of epitheloid cells, by mixing logic appearance of the lesion in 2 days is t hat of a delayed electron dense markers with Kveim antigen and injecting these hypersensitivity lesion. However, the presence of polymorph into the skin. Initially the marker was in the newly immigrated nuclear leukocytes in the granuloma would indicate to me that monocyte, then in the activated macrophage with many pha­ this could be the result of both antibody production and cell gosomes, and eventually all the epitheloid cells were labeled in mediated immunity. this case with colloidal gold. I think it just corroborates your AUSTEN: You started out commenting on the properdin sys­ statements. I think that one does not necessarily have to tem and I think it is important for all of us to recall that the distinguish between an activated macrophage and a secretory properdin system establishes beyond the question of a doubt cell because in many instances these properties are caJTied out that the activation of complement is not always dependent on by the same population of cells. In the case of the e pitheloid the host immune response. This pathway may well have pre­ cells, it seemed to us however, that with the passage of time ceeded adaptive immunity so that I think it is probably a these cells became highly interdigitated and closely apposed so mistake to imply that complements should be linked irrevoca­ that perhaps the availability of phagositizable particles had bly to immunologic mechanisms. With regard to your very decreased and therefore they were now expressing this endo­ interesting experience in vitro with the metals, in which you plasmic reticulum function to a greater extent. The second showed that the lysine-affinity chromatography was inhibitory comment relates to the modulation of receptors on the mem­ while EDT A was not, raised the issue of the plasminogen­ brane. I wanted to point out that if you put macrophages down pathway. One still has to address how t he metals for example, on immune complexes, you can completely remove convert plasminogen to plasmin. I think that by using reagents the expression of the Fe receptor on the upper surface. If you deficient in Hageman factor or Kallikrein you could define the put them down on an I gMC3 complex, you selectively take the pathway more clearly. complement receptor off but leave the Fe receptor. Probably TURK: Is there enough Hageman factor in serum for these the mechanism is migration of the receptor in the plane of the metals to activate the pathway through it. membrane. That seems to make most sense of some of the AUSTEN: Since the one reagent you can get is Hageman factor more current data.