JOURNAL OF PATHOLOGY, VOL.
164: 253-259 (1 99 1)
IMMUNOCYTOCHEMICAL OBSERVATIONS ON MACROPHAGE POPULATIONS IN NORMAL FETAL AND ADULT HUMAN LIVER KRZYSZTOF A. BARDADIN", PETER J. SCHEUER, ALICJA PECZEK* AND JAROSLAW WEJMAN"
Department o j Histopathology, Royal Free Hospilal School ojniledicine. London. U.K.; *Department OJ Pathomorphology, Postgraduate Medical Centre Hospital, Warsaw, Poland Received 2 November 1990 Accepted5 February 1991
SUMMARY Phenotypic expression of macrophages was studied immunocytochemicallyin 25 human fetal livers at various stages of development and in 20 normal human adult livers. A panel of commercially available polyclonal and monoclonal antibodies (KP1, Mac387, LN3, CR3/43, and antibodies against muramidase, alpha-1-antitrypsin,and factor XIIIa) was applied to paraffinsections. From the seventh week of gestation macrophages in the fetal liver showed differences in distribution with the various antibodies. Macrophages in adult liver similarly varied in morphology and phenotypic expression. In the light of these results, we conclude that the population of human liver macrophages is heterogeneous from an early stage of fetal development and that this heterogeneity extends into adult life. KEY
woms-Liver histology, human, macrophages, Kupffer cell, immunocytochemistry,embryology.
of adult liver. Several immunocytochemical and electron microscopical studies have been carried out Immunocytochemical and ultrastructural evidence on human material,'&I6 but most of these were indicates that in pathological conditions the macro- principally directed to problems of fetal lymphophage population of the liver is heterogeneo~s,l-~ and erythropoiesis. The results obtained with but the extent to which this heterogeneity reflects macrophages in fetal and normal adult human liver varying phenotypic expression or different cell varied. In this study, we have used a panel of antilineages is unclear. According to the concept of bodies against macrophage-associated antigens to the mononuclear phagocyte ~ y s t e m ,Kupffer ~,~ cells document the development of macrophage popuare fixed macrophages of bone marrow origin. lations in human liver. The enzyme muramidase (lysozyme) has long Opponents of the concept of the mononuclear phagocyte system stress that Kupffer cells are been used as a marker of monocytes, granulocytes, present within the liver before the bone marrow has and macrophages, and can be demonstrated in de~eloped.~.'Electron microscopical studies have paraffin sections;16 l 8 however, not all macrophage - ~ ~ recently, the been carried out on human and animal tis~ues.'.~ populations are p o s i t i ~ e . ' ~More anti-protease alpha-1-antitrypsin has been used as a Immunoelectron microscopical observations in marker of benign and malignant histi~ cytes. *'-~~ mouse yolk sac revealed that primitive macrophages appeared in the vascular lumen and fetal liver sinu- This, however, also gave variable results. Factor soids by the ninth day of gestation, before the XIIIa has been used as a marker of mono- and , ~ ~is useful for the identifiappearance of promonocytes and monocytes.' The megakar yop~ iesis'~and cation of connective tissue histiocytes. As far as we authors concluded that primitive yolk sac macroare aware, it has not so far been used for the study phages might be the precursors of the Kupffer cells of liver macrophages. Among the monoclonal antiAddressee for correspondence: Professor P. J. Scheuer, bodies available for use in paraffin-embedded material, Mac387*' is an antibody raised against Department of Histopathology, Royal Free Hospital, Pond a formalin-resistant monocyte and macrophage Street, London NW3 2QG, U.K. INTRODUCTION
0022-341 719 1/070253-07 $05.00 0 1991 by John Wiley & Sons, Ltd.
254
K. A. BARDADIN ET AL.
Table I-Antibodies, sources, and dilutions used Antigen
Antibody KPI
Mac387 Muramidase (lysozyme) a ,-Antitrypsin LN3
CR3/43 Factor XIIIa
Source
Type*
specificity
Dakopatts Dakopatts Dakopatts Dakopatts Clonab Dakopatts
M M P P M M P
CD68 L1, cystic fibrosis antigen Muramidase a,-Antitrypsin Ia region Bchain, HLA class I1 Factor XIIIa
Behring
Dilution
References
1:40
1 :so
1,3, 17 16, 18-20
1:200 1:2000 1:so
16,21-25 16,2629 30
1:20 1 :200
33,34
31,32
*M = Monoclonal; P = polyclonal.
antigen Ll,28*29and LN3 and CR3/43 probably detect HLA-DR.’&’* The most recent antibody to become available is KPl ,” probably identical to antibody E B M l l and described as a panmacrophage antibody. These two antibodies have been used in studies of macrophage populations in adult human liver.’.3333 Both show many stellate cells within hepatic sinusoids of normal liver, but give different results in pathological conditions. We used a panel consisting of all the above antibodies in an attempt to document the development of the various macrophage populations in adult human liver and to throw light on their lineages.
acid-Schiff method after diastase digestion (DPAS), and by Gordon and Sweets’s method for reticulin. Immunocytochemical staining was performed on paraffin sections with the antibodies listed in Table I. Serial sections 5 pm thick were mounted on poly- I lysine, dried overnight at 37”C, and predigested with 0.1 per cent trypsin for 10 min. Trypsinization was shortened to 5 min for LN3 and omitted for CR3/43. The peroxidase-antiperoxidase (PAP) method was used with polyclonal antibodies, and the avidinbiotin complex (ABC) method with monoclonal antibodies. In negative control slides, the primary antibody was omitted.
MATERIAL A N D METHODS
RESULTS
Twenty-five fetal livers from the files of the Department of Pathomorphology of the Postgraduate Medical Centre Hospital, Warsaw, were studied by immunocytochemical methods. Of these, ten livers were from embryos aged 7-1 5 weeks, obtained at termination of pregnancy. A second group of 1 1 livers was obtained at autopsy from fetuses delivered between 16 and 30 weeks, and a further four livers from premature deliveries between 30 and 39 weeks of gestation. Fetal age was verified from clinical dates. Normal adult livers were obtained by needle biopsy (ten patients) or at laparotomy (ten patients) from patients with no clinical or biochemical evidence of liver dysfunction. Operative wedge biopsies were mainly taken in patients with cholelithiasis. The needle biopsy specimens were from the files of the Department of Histopathology of the Royal Free Hospital. All specimens were fixed in phosphate-buffered formalin. Sections of paraffin-embedded tissue were stained with haematoxylin and eosin (H&E), by the periodic
KP1
The first cells giving a definite positive reaction with this antibody were seen within the liver parenchyma at the 1 1th week of gestation, in the form of solitary stellate cells within sinusoidal lumens and haemopoietic foci. Cellular debris could be identified in their cytoplasm. Their number increased with fetal age, moderate numbers being found in the 30-39 week group. Connective tissue histiocytes, including those in portal tracts, remained negative. In adult liver, numerous stellate macrophages within sinusoidal lumens stained strongly. A few stellate portal histiocytes were also positive, as were granulocytes and monocytes in the sinusoids. Mac387 As early as the seventh week of gestation, a few strongly Mac387-positive cells were identified within the capsule of the liver and as solitary cells lying free in the subcapsular sinuses. The positive cells had a
MACROPHAGES IN NORMAL HUMAN LIVER
Fig. I-Fetus, I0 weeks’ gestation. Macrophages staining with Mac387 are sken in the connective tissue of a portal tract. There are a few isolated positive cells in the sinusoids. lmmunoperoxidase
regular ovoid shape and nuclei resembling those of monocytes. Their number increased with age, and their distribution changed. At 10-15 weeks of gestation, Mac387-positive cells were found mainly within portal tracts, but a few were also seen in hepatic acini as solitary ovoid cells between hepatocytes (Fig. 1). Exceptionally, elongated Mac387positive cells were seen within sinusoidal lumens. Later in fetal life, positive macrophages were numerous at the periphery of portal tracts, whereas in the acini only a few were identified. In none of the material did we observe Mac387-positivecells whose shape and location resembled those of classical Kupffer cells. In adult human livers, Mac387 stained granulocytes and monocytes in portal tracts and sinusoids in addition to the cells described above. Sinusoidal endothelial cells were negative in all specimens. M u ram idase In 7-week fetuses, solitary positive cells were seen in the liver capsule. Isolated positive cells were also seen in the lumens of subcapsular sinuses. The cells generally resembled the Mac387-positive cells
255
Fig. 2-Fetus, 10 weeks’ gestation. Scattered muramidasecontaining cells (arrows) are seen in the hepatic sinusoids. In some, the irregular cell shape suggests that they are macrophages. The smaller deeply-stained cells are haemopoietic. Immunoperoxidase
described in the previous section. In 10-week embryos, there were, in addition to these cells in portal tracts, a few positive cells within the sinusoids of the hepatic acini. These had more irregular outlines and nuclei than those within connective tissues, and were tentatively identified as macrophages (Fig. 2). Further muramidase-positive cells, different from both types described above, were oval in shape, with oval nuclei and deeply-stained granular cytoplasm, lying within sinusoids. These cells resembled myeloblasts and myelocytes. In fetuses between 16 and 30 weeks, numerous muramidasepositive macrophages were observed in connective tissue of portal tracts, their shape and location being similar to those of Mac387-positive cells. In the parenchyma, stellate muramidase-positive macrophages were identified mainly within foci oferythropoiesis, and often contained cellular debris. In addition to macrophages, some myeloblasts were strongly muramidase-positive. In the third trimester, thenumberofpositivemacrophages in sinusoids was relatively high, the cells being observed both within and outside foci of haemopoiesis. Positive sinusoidal cells were often stellate in shape and resembled
256
K. A. BARDADIN E T A L .
Fig. 3-Fetus, 12 weeks' gestation. In addition to blasts staining for alpha-I-antitrypsin (open arrow), positive cells appear to be lining sinusoids (solid arrow). Immunoperoxidase
Fig. &Adult liver. Elongated LN3-positive cells are seen in a portal tract and within sinusoids. In the latter location, many of them have the shape of typical Kupffer cells. Immunoperoxidase
Kupffer cells. Muramidase-positive cells were regularly seen at the periphery of portal tracts. In adult livers, muramidase marked numerous stellate macrophages within sinusoids and a few cells in portal tracts, as well as monocytes and granulocytes free within the lumens of the hepatic sinusoids.
sinusoids and only occasionally bulging into their lumens. Their number varied from liver to liver, but as a rule they were sparse compared with muramidase-positive cells.
Alpha-1 -antitrypsin Staining patterns obtained with antibody to alpha- 1-antitrypsin differed from those described above for Mac387 and muramidase. Positive cells were first observed at 10 weeks. Solitary ovoid blast-like cells were present free within the hepatic sinusoids. In addition, many fusiform or stellate macrophage-like cells were identified within foci of haemopoiesis or lining sinusoids. Some of these positive cells resembled sinusoidal endothelial cells in their location and morphology (Fig. 3). Only a small number of connective tissue cells were positive. There was marked variation in the number of alpha- 1-antitrypsin-positive cells in the liver, independent of the age of the fetus. Alpha-l-antitrypsincontaining cells were also found in normal adult human livers; these were elongated cells, lining
HLA-DR (LN3, CR3/43) In all the fetal tissues studied, only a few solitary cells of irregular shape gave weak positive staining with these antibodies. Positive cells were mainly in portal tracts rather than in sinusoids. By contrast, adult livers contained many LN3-positive stellate cells both in sinusoids and in portal tracts (Fig. 4). CR3/43 mainly stained portal macrophages, and only occasional sinusoidal cells. Granulocytes and monocytes were negative.
Factor XIIIa In all the fetal livers studied, numerous factor XIIIa-positive cells were observed within theconnective tissue of the liver capsule and portal tracts. They had an irregular fusiform or stellate shape and coarse, granular cytoplasm. In addition to these fusiform cells in connective tissue, two other factor XIIIa-expressing cell populations were found in the
MACROPHAGES IN NORMAL HUMAN LIVER
257
mainly been concerned with haemopoietic cells, but muramidase-positive cells were noted in early weeks and LN3-positive cells from the 14th week.15Kamps et aI.l4 found large irregular cells staining with an antibody to HLA-DR (OKIa) and suggested that they might be Kupffer cells. Our own studies clearly show muramidase- and Mac387-positive cells in 7week-old fetuses. These cells, mainly in connective tissue but occasionally within hepatic sinusoids, could represent either a population of primitive monocytes or macrophages, or the earliest Kupffer cells in the sense of cells destined to become the 'fixed' macrophages of the liver. In view of their location, it seems likely that they are the first generation ofprimitive macrophages of yolk sac origin, as suggested by Naito et ai.9and Takahashi et a13' In later fetal life, different staining patterns with the various antibodies became striking. Mac387 consistently labelled large monocytoid cells at the edges of portal tracts, and only a few cells in the sinusoids. Their site is in keeping with that of myelomonocytopoIesis.37 Antibody to muramidase labelled these cells, but additionally marked sinusoidal macrophages. In the third group of fetal Fig. S-Adult liver. Elongated cells staining for factor XIIIa are livers between 30 and 39 weeks of gestation, less muramidase-positive cells were found in connective located in portal connective tissue. Immunoperoxidase tissue, but sinusoidal cells persisted. A possible liver. First, many cells of megakaryocytic differen- interpretation of this change is that monocytotiation were positive. These wereeasily recognized by poiesis had moved to the bone marrow, whereas the their oval shape, fine pale cytoplasm with strongly muramidase-positive cells now represented a popustained peripheral rim, and two or more nuclei. lation of sinusoidal, liver-committed macrophages. Second, there were stellate or fusiform macrophage- Similar stellate muramidase-positive cells were conlike cells within foci of haemopoiesis or lining sinu- sistently found in normal adult liver. These findings soids. Both types of cells were identified at various suggest that muramidase is a consistent marker of times of gestation, but their number decreased with Kupffer cells and that variations in positivity in age. In adult human livers, factor XIIIa-positive disease reflect altered Kupffer-cell physiology rather stellate or elongated macrophages were observed in than the validity of this marker.3 The presence of a few fusiform Mac387-positive portal tracts but not in sinusoids (Fig. 5). In some cells within the sinusoids and portal tracts of fetal livers there was focal granular positivity along the livers probably represents the early stages of transwalls of sinusoids, possibly within endothelial cells. formation from monocytes to macrophages, as Monocytes and granulocytes were consistently suggested by Brandtzaeg and c o - w o r k e r ~ . ' ~ , ~ ' negative. Mac387 detects the leukocyte L1 antigen which is thought to be transferred from monocytes to endothelial cells when the former become adherent.38It is DISCUSSION possible that under pathological conditions in The first cells in man with the ultrastructural which Mac387-positive cells increase in number, characteristics of macrophages appear in the yolk this reflects retention of the molecule by antigensac and mesenchyme in the fourth week of gestation, presenting cells, with reduced transfer to vascular and in the prehaemopoietic liver in the fifth week.3s endothelium. In this way, antigen presentation A few muramidase-positive cells have been ident- might depend on altered monocyteeendothelial cell ified in fetal liver at 7-8 weeks.'* Recent immuno- interaction. Alternatively, Karakucuk et al.16postucytochemical studies in the human fetal liver have late that under pathological conditions such as
258
K. A. B A R D A D I N ET A L .
alcoholic liver disease there might be recruitment of macrophages, seen in part as an increase in Mac387positive cells, with consequent release of cytotoxic mediators. The remaining antibodies gave different results from those obtained for muramidase and with Mac387. Alpha- 1-antitrypsin-containing cells in the hepatic sinusoids were fusiform rather than stellate, and did not resemble classical Kupffer cells. A wide variation in the intensity of alpha- 1-antitrypsin staining was noted, and there appeared to be no correlation with staining for muramidase. Antibody to alpha-1-antitrypsin was the only one to give linear positivity in sinusoidal walls. The gene of the human histocompatibility (HLA) complex consists of at least four subregions (DP, DQ, DX, DR) containing a minimum of one alpha and one beta chain. CR3/43 reacts with the beta chain of all products of the gene subregions DP, DQ, and DR. LN3 recognizes a non-polymorphic antigen of the HLA-DR (Ia) region. The two antibodies give good results in lymph node^.^^.'^ In the present study, little positivity was found in fetal liver, but in the adult large numbers of hepatic macrophages were stained. The results in fetal liver may reflect low expression, only detectable in fresh frozen material,’3.3’and the lack of challenge of the fetal immune system by extrinsic stimuli. Presentation of antigens to lymphoid cells is mainly mediated by macrophages expressing HLA-DR on their surfaces (also known as Ia, or immune antigen). Ia antigens are expressed by murine, rat, and human Kupffer cell^,^^^^' but their role in regulating the immune response remains unclear. HLA-DRexpressing cells have also been demonstrated in portal tracts” and correlated with Steinman and Cohn’s dendritic cells.39 In our study, LN3 and CR3/43-positive cells in portal tracts had the same distribution and characteristics as those of factor XIIIa-positive cells, and should thus be regarded as macrophages. They probably represent a population of connective tissue macrophages, unrelated to Kupffer cells and reacting d i f f e r e n t l ~ .Their ~~ presence in portal tracts is of particular interest in that it might explain trapping of lymphocytes in portal tracts in pathological conditions. Furthermore, these cells might be responsible for the establishment in portal tracts of B- and T-cell zones, as suggested previously.* Factor XIIIa is involved not only in haemostasis, but also in connective tissue organization including proliferation of fibroblast^.^^,^^ This might explain the large number of factor XIIIa-positive cells in
growing fetal tissues, compared with the much smaller number in adult portal tracts. The sinusoidal granules stained by antibody to factor XIIIa have not so far been identified with certainty. They may represent platelets, but it seems more likely from their size, shape, and location that they are the endothelial-cell granules previously described in hepatitis.44 In summary, we have demonstrated marked heterogeneity of phenotypic expression among hepatic macrophages, both in fetal and in adult life. This may reflect variation of phenotypic expression in a single macrophage cell line, or the co-existence in the liver of several distinct populations of macrophages. ACKNOWLEDGEMENTS
This work was supported by the Wellcome Trust. We are grateful to Ms Linda More for introducing AP and JW to technical aspects of immunocytochemistry and for giving valuable help with the project. REFERENCES I . Van den Oord JJ, De Vos R, Facchetti F, Delabie J. De Wolf-Peeters C, Desmet VJ. Distribution of non-lymphoid, inflammatory cells in chronic HBV infection. JPufhol1990; 160 223-230. 2. Bardadin KA, Desmet VJ. Interdigitating and dendritic reticulum cells in chronic active hepatitis. Hisfopa/holog.v 1984; 8: 657-667. 3. Kelly PM, Heryet AR, McGee JO’D. Kupffer cell number is normal, but their lysozyme content is reduced in alcoholic liver disease. J Hepurol 1989; 8: 173-180. 4. Van Furth R, Crofton RW, Diesselhoff-Den Dulk MMC. The bone marrow origin of KupiTer cells. In: Wisse E, Knook DL, eds. Kupffer Cells and Orher Liver Sinusoidal Cells. Amsterdam: Elsevier, 1977; 471480. 5. Gale RP, Sparkes, RS, Golde DW. Bone marrow origin of hepatic macrophages (Kupffer cells) in humans. Science 1978; 201: 937-938. 6. Deimann W, Fahimi HD. Peroxidase cytochemistry and ultrastructure of resident macrophages in fetal rat liver. A developmental study. Dev Biol 1978; 66.43-56. 7 Naito M, Yamamura F, Takeya M, Takahashi K. Ultrastructural analysis of Kupffer cell progenitors. In: Kirn A, Knook DL, Wisse E, eds. Cells of the Hepatic Sinusoid. Vol. 1. Rijswijk: The Kupffer Cell Foundation, 1986; 13-20. 8. Bankston PW, Pino RM. Thedevelopment ofthe sinusoids offetal rat liver: morphology of endothelial cells, Kupffer cells, and the transmural migration of blood cells into the sinusoids. Am J Anaf 1980; 1 5 9 1-15 9. Naito M, Takahashi K., Ohno H . Nishikawa SI. Yolk sac macrophages-a possible Kupffer cell precursor in the fetal mouse liver. In: Wisse E, Knook DL, Decker K, eds. Cells of the Hepatic Sinusoid. Vol. 2 . Rijswjijk: The Kupffer Cell Foundation, 1989; 419420. 10. Fukuda T. Fetal hemopoiesis. 11. Electron microscopic studies on human hepatic hemopoiesis. Virchows Arch B 1974; 1 6 249-270. I I Enzan H, Hara H, Yamashita Y, Ohkita T , Yamane T. Fine structure of hepatic sinusoids and their development in human embryos and fetuses. Acra Parhol Jpn 1983; 3 3 447466. 12. Watanabe S. Nakajima T , Shimosato Y, Shimamura K, Sakuma H. T-zone histiocytosis with SlOO protein. Development and distribution in human fetuses. Arru Patliol Jpn 1983; 33: 15-22.
M A C R O P H A G E S I N N O R M A L H U M A N LIVER 13. Janossy G, Botill M , Poulter LW, er a/. Separate ontogeny of two macrophage-like accessory cell populations in the human fetus. J Immunol1986; 1 3 6 43544361. 14. Kamps WA, Timens W, De Boer GJ, Spanjer H H , Poppema S. In siru study of haemopoiesis in human fetal liver. ScandJImmunol1989: 30: 399408. 15. Timens W, Kamps WA, Rozeboom-Uiterwijk T, Poppema S. Haemopoiesis in human fetal and embryonic liver. Immunohistochemical determination in B5-fixed, paraffin-embedded tissues. Virchows Arch A 1990; 4 1 6 429436. 16. Karakucuk I, Dilly SA, Maxwell J D . Portal tract macrophages are increased in alcoholic liver disease. Hi.stuparhology 1989: 1 4 245-253. 17. Pulford KAF, Rigney EM, Micklem KJ, er a/. KPI : a new monoclonal antibody that detects a monocytejmacrophage associated antigen in routinely processed tissue sections. J Clin Parhol1989: 42: 414421 18. Flavell DJ, Jones DB. Wright D H . Identification of tissue histiocytes on paraffin sections by a new monoclonal antibody. J Hisrochem Cyrodiem 1987;35 1217-1226. 19. Brandtzaeg P, Jones DB, Flavell DJ, Fagerhol M K . Mac387 antibody and detection offormalin resistant myelomonocytic LI antigen. JClin Prrfhol1988; 41: 963-970. 20. Brandtzaeg P. The new monoclonal antibody (Mac387) that reacts with macrophages on paraffin sections detects the well-known leukocyte LI antigen. JHisrochem Cyrochem 1988: 36 1203-1206. 21. Mason DY,TaylorCR.Thedistributionofmuramidase(lysozyme)in human tissues. J Clin Pathol 1975; 28: 126132. 22. Klockars M , Reitamo S. Tissue distribution of lysozyme in man. J Hi.~roc./iemCyrochem 1975; 23: 932-940. 23. Pinkus GS, Said JW. Profile of intracytoplasmic lysozyme in normal tissues, myeloproliferative disorders, hairy cell leukemia, and other pathological processes. An immunoperoxidase study of paraffin sections and smears. Am J Pathol 1977; 8 9 351-366. 24. Mendelsohn G, Eggleston JC, Mann RB. Relationship of lysozyme (muramidase) to histiocytic differentiation in malignant histiocytosis. An immunohistochemical study. Cancer 1980; 4 5 273-279. 25. Kerdel FA, Morgan EW, Holden CA, MacDonald D M . Demonstration of alpha I-antitrypsin and alpha I-antichymotrypsin in cutaneous histiocytic infiltrates and a comparison with intracellular lysozyme. J Am AcatlDermatol 1982; 7: 177-182. 26. Wilson GB, Walker J H Jr, Watkins J H Jr. Wolgroch D. Determination of subpopulations of leukocytes involved in the synthesis of alpha-I-antitrypsin In virro. Proc Soc E , p Biol Med 1980; 164: 105-1 14. 27. Isaacson PG, Jones DB, Millward-Sadler G H , Judd MA, Payne S . Alpha-I-antitrypsin in human macrophages. J Clin Parhol 1981; 34: 982-990. 28. Van Furth R, Kramps JA, Diesselhoff-Den Dulk MMC. Synthesis of a,-anti-trypsin by human monocytes. Clin E x p Immunol 1983; 51: 551-557.
259
29. Perlemutter DH, Sessions Cole F, Kilbridge P, Rossing T H , Colten HR. Expression of the a,-proteinase inhibitor gene in human monocytes and macrophages. Proc Narl Arad Sci USA 1985: 82: 795-799. 30. Marder RJ, Variakojis D , Silver J, Epstein AL. Immunohistochemical analysis of human lymphomas with monoclonal antibodies to B cell and la antigens reactive in paraffin sections. Lab Invest 1985; 52: 497-504. 31. Barbatis C, Kelly P, Greveson J, Heryet A, McGee J O D . Immunocytochemical analysis of H L A class 11 (DR) antigens in liver disease in man. J Clin Parhol1987; 4 0 879-884. 32. Smith MEF, Holgate CS, Williamson JMS, Grigor I, Quirke P, Bird CC. Major histocompatibility complex class TI antigen expression in B and T cell non-Hodgkin's lymphoma. J Clin Pathol 1987; 4 0 3Ml.
33. Nemes Z, Thomazy V, Adiny R, Muszbek L. Identification ofhistiocytic reticulum cells by the immunohistochemical demonstration of factor XIIJ (F-XIlla) in human lymph nodes. J Pathol 1986: 149: 121-132. 34. Adany R, Kiss A, Muszbek L. Factor XIII: a marker of mono- and megakaryocytopoiesis. Br J Haemarol1987: 67: 167-172. 35. Kelemen E, Calvo W, Fliedner TM. Atlas of Human Hemopoietic Development. Berlin: Springer-Verlag, 1979; 83-98. 36. Takahashi K, Yamamura F, Naito M. Differentiation, maturation, and proliferation of macrophages in the mouse yolk sac: a lightmicroscopic, enzyme-cytochemical,immunohistocbemical, and ultrastructural study. J Leukoc B i d 1989: 4 5 87-96. 37. Kamps WA, Cooper M D . Microenvironmental studies of pre-B cell and B cell development in human and mouse fetuses. J ImmunolI982; 1 2 9 52&531 38. Hogg N, Allen C, Edgeworth J. Monoclonal antibody 5.5 reacts with p8,14, a myeloid molecule associated with some vascular endothelium. Eur J Immunol1989: 1 9 1053-1061. 39. Steinman R M , Cohn ZA. Identification of a novel cell type in peripheral lymphoid organs of mice. 1. Morphology, quantitation, tissue distribution. J Exp Med 1973; 137: 1142-1 162. 40. Shiratori Y, Okano K, Matsumoto K, Murao S. Antigen presentation by Kupffer cells in the rat. Scand J Gasrroenterol1984; 19: 733-739. 41. Ramadori G, Dienes HP, Burger R, Meuer S, Rieder H, Meyer zum Buschenfelde K H . Expression of la-antigens o n guinea pig Kupffer cells. Studies with monoclonal antibodies. JHeparol 1986;2: 208-2 17. 42. Beck E, Duckert F , Ernst M . The influence of fibrin stabilizing factor in the growth of fibroblasts in vitro and wound healing. Thromb Haemosr 1961: 6 485491. 43. Bruhn H D , Pohl J. Growth regulation of fibroblasts by thrombin, factor XI11 and tibronectin. Klin Wochenschr 1981: 5 9 145-146. 44. Bardadin KA, Scheuer PJ. Endothelial cell changes in acute hepatitis. A light and electron microscopic study. J Parhol 1984: 144: 21 3-220.
164: 253-259 (1 99 1)
IMMUNOCYTOCHEMICAL OBSERVATIONS ON MACROPHAGE POPULATIONS IN NORMAL FETAL AND ADULT HUMAN LIVER KRZYSZTOF A. BARDADIN", PETER J. SCHEUER, ALICJA PECZEK* AND JAROSLAW WEJMAN"
Department o j Histopathology, Royal Free Hospilal School ojniledicine. London. U.K.; *Department OJ Pathomorphology, Postgraduate Medical Centre Hospital, Warsaw, Poland Received 2 November 1990 Accepted5 February 1991
SUMMARY Phenotypic expression of macrophages was studied immunocytochemicallyin 25 human fetal livers at various stages of development and in 20 normal human adult livers. A panel of commercially available polyclonal and monoclonal antibodies (KP1, Mac387, LN3, CR3/43, and antibodies against muramidase, alpha-1-antitrypsin,and factor XIIIa) was applied to paraffinsections. From the seventh week of gestation macrophages in the fetal liver showed differences in distribution with the various antibodies. Macrophages in adult liver similarly varied in morphology and phenotypic expression. In the light of these results, we conclude that the population of human liver macrophages is heterogeneous from an early stage of fetal development and that this heterogeneity extends into adult life. KEY
woms-Liver histology, human, macrophages, Kupffer cell, immunocytochemistry,embryology.
of adult liver. Several immunocytochemical and electron microscopical studies have been carried out Immunocytochemical and ultrastructural evidence on human material,'&I6 but most of these were indicates that in pathological conditions the macro- principally directed to problems of fetal lymphophage population of the liver is heterogeneo~s,l-~ and erythropoiesis. The results obtained with but the extent to which this heterogeneity reflects macrophages in fetal and normal adult human liver varying phenotypic expression or different cell varied. In this study, we have used a panel of antilineages is unclear. According to the concept of bodies against macrophage-associated antigens to the mononuclear phagocyte ~ y s t e m ,Kupffer ~,~ cells document the development of macrophage popuare fixed macrophages of bone marrow origin. lations in human liver. The enzyme muramidase (lysozyme) has long Opponents of the concept of the mononuclear phagocyte system stress that Kupffer cells are been used as a marker of monocytes, granulocytes, present within the liver before the bone marrow has and macrophages, and can be demonstrated in de~eloped.~.'Electron microscopical studies have paraffin sections;16 l 8 however, not all macrophage - ~ ~ recently, the been carried out on human and animal tis~ues.'.~ populations are p o s i t i ~ e . ' ~More anti-protease alpha-1-antitrypsin has been used as a Immunoelectron microscopical observations in marker of benign and malignant histi~ cytes. *'-~~ mouse yolk sac revealed that primitive macrophages appeared in the vascular lumen and fetal liver sinu- This, however, also gave variable results. Factor soids by the ninth day of gestation, before the XIIIa has been used as a marker of mono- and , ~ ~is useful for the identifiappearance of promonocytes and monocytes.' The megakar yop~ iesis'~and cation of connective tissue histiocytes. As far as we authors concluded that primitive yolk sac macroare aware, it has not so far been used for the study phages might be the precursors of the Kupffer cells of liver macrophages. Among the monoclonal antiAddressee for correspondence: Professor P. J. Scheuer, bodies available for use in paraffin-embedded material, Mac387*' is an antibody raised against Department of Histopathology, Royal Free Hospital, Pond a formalin-resistant monocyte and macrophage Street, London NW3 2QG, U.K. INTRODUCTION
0022-341 719 1/070253-07 $05.00 0 1991 by John Wiley & Sons, Ltd.
254
K. A. BARDADIN ET AL.
Table I-Antibodies, sources, and dilutions used Antigen
Antibody KPI
Mac387 Muramidase (lysozyme) a ,-Antitrypsin LN3
CR3/43 Factor XIIIa
Source
Type*
specificity
Dakopatts Dakopatts Dakopatts Dakopatts Clonab Dakopatts
M M P P M M P
CD68 L1, cystic fibrosis antigen Muramidase a,-Antitrypsin Ia region Bchain, HLA class I1 Factor XIIIa
Behring
Dilution
References
1:40
1 :so
1,3, 17 16, 18-20
1:200 1:2000 1:so
16,21-25 16,2629 30
1:20 1 :200
33,34
31,32
*M = Monoclonal; P = polyclonal.
antigen Ll,28*29and LN3 and CR3/43 probably detect HLA-DR.’&’* The most recent antibody to become available is KPl ,” probably identical to antibody E B M l l and described as a panmacrophage antibody. These two antibodies have been used in studies of macrophage populations in adult human liver.’.3333 Both show many stellate cells within hepatic sinusoids of normal liver, but give different results in pathological conditions. We used a panel consisting of all the above antibodies in an attempt to document the development of the various macrophage populations in adult human liver and to throw light on their lineages.
acid-Schiff method after diastase digestion (DPAS), and by Gordon and Sweets’s method for reticulin. Immunocytochemical staining was performed on paraffin sections with the antibodies listed in Table I. Serial sections 5 pm thick were mounted on poly- I lysine, dried overnight at 37”C, and predigested with 0.1 per cent trypsin for 10 min. Trypsinization was shortened to 5 min for LN3 and omitted for CR3/43. The peroxidase-antiperoxidase (PAP) method was used with polyclonal antibodies, and the avidinbiotin complex (ABC) method with monoclonal antibodies. In negative control slides, the primary antibody was omitted.
MATERIAL A N D METHODS
RESULTS
Twenty-five fetal livers from the files of the Department of Pathomorphology of the Postgraduate Medical Centre Hospital, Warsaw, were studied by immunocytochemical methods. Of these, ten livers were from embryos aged 7-1 5 weeks, obtained at termination of pregnancy. A second group of 1 1 livers was obtained at autopsy from fetuses delivered between 16 and 30 weeks, and a further four livers from premature deliveries between 30 and 39 weeks of gestation. Fetal age was verified from clinical dates. Normal adult livers were obtained by needle biopsy (ten patients) or at laparotomy (ten patients) from patients with no clinical or biochemical evidence of liver dysfunction. Operative wedge biopsies were mainly taken in patients with cholelithiasis. The needle biopsy specimens were from the files of the Department of Histopathology of the Royal Free Hospital. All specimens were fixed in phosphate-buffered formalin. Sections of paraffin-embedded tissue were stained with haematoxylin and eosin (H&E), by the periodic
KP1
The first cells giving a definite positive reaction with this antibody were seen within the liver parenchyma at the 1 1th week of gestation, in the form of solitary stellate cells within sinusoidal lumens and haemopoietic foci. Cellular debris could be identified in their cytoplasm. Their number increased with fetal age, moderate numbers being found in the 30-39 week group. Connective tissue histiocytes, including those in portal tracts, remained negative. In adult liver, numerous stellate macrophages within sinusoidal lumens stained strongly. A few stellate portal histiocytes were also positive, as were granulocytes and monocytes in the sinusoids. Mac387 As early as the seventh week of gestation, a few strongly Mac387-positive cells were identified within the capsule of the liver and as solitary cells lying free in the subcapsular sinuses. The positive cells had a
MACROPHAGES IN NORMAL HUMAN LIVER
Fig. I-Fetus, I0 weeks’ gestation. Macrophages staining with Mac387 are sken in the connective tissue of a portal tract. There are a few isolated positive cells in the sinusoids. lmmunoperoxidase
regular ovoid shape and nuclei resembling those of monocytes. Their number increased with age, and their distribution changed. At 10-15 weeks of gestation, Mac387-positive cells were found mainly within portal tracts, but a few were also seen in hepatic acini as solitary ovoid cells between hepatocytes (Fig. 1). Exceptionally, elongated Mac387positive cells were seen within sinusoidal lumens. Later in fetal life, positive macrophages were numerous at the periphery of portal tracts, whereas in the acini only a few were identified. In none of the material did we observe Mac387-positivecells whose shape and location resembled those of classical Kupffer cells. In adult human livers, Mac387 stained granulocytes and monocytes in portal tracts and sinusoids in addition to the cells described above. Sinusoidal endothelial cells were negative in all specimens. M u ram idase In 7-week fetuses, solitary positive cells were seen in the liver capsule. Isolated positive cells were also seen in the lumens of subcapsular sinuses. The cells generally resembled the Mac387-positive cells
255
Fig. 2-Fetus, 10 weeks’ gestation. Scattered muramidasecontaining cells (arrows) are seen in the hepatic sinusoids. In some, the irregular cell shape suggests that they are macrophages. The smaller deeply-stained cells are haemopoietic. Immunoperoxidase
described in the previous section. In 10-week embryos, there were, in addition to these cells in portal tracts, a few positive cells within the sinusoids of the hepatic acini. These had more irregular outlines and nuclei than those within connective tissues, and were tentatively identified as macrophages (Fig. 2). Further muramidase-positive cells, different from both types described above, were oval in shape, with oval nuclei and deeply-stained granular cytoplasm, lying within sinusoids. These cells resembled myeloblasts and myelocytes. In fetuses between 16 and 30 weeks, numerous muramidasepositive macrophages were observed in connective tissue of portal tracts, their shape and location being similar to those of Mac387-positive cells. In the parenchyma, stellate muramidase-positive macrophages were identified mainly within foci oferythropoiesis, and often contained cellular debris. In addition to macrophages, some myeloblasts were strongly muramidase-positive. In the third trimester, thenumberofpositivemacrophages in sinusoids was relatively high, the cells being observed both within and outside foci of haemopoiesis. Positive sinusoidal cells were often stellate in shape and resembled
256
K. A. BARDADIN E T A L .
Fig. 3-Fetus, 12 weeks' gestation. In addition to blasts staining for alpha-I-antitrypsin (open arrow), positive cells appear to be lining sinusoids (solid arrow). Immunoperoxidase
Fig. &Adult liver. Elongated LN3-positive cells are seen in a portal tract and within sinusoids. In the latter location, many of them have the shape of typical Kupffer cells. Immunoperoxidase
Kupffer cells. Muramidase-positive cells were regularly seen at the periphery of portal tracts. In adult livers, muramidase marked numerous stellate macrophages within sinusoids and a few cells in portal tracts, as well as monocytes and granulocytes free within the lumens of the hepatic sinusoids.
sinusoids and only occasionally bulging into their lumens. Their number varied from liver to liver, but as a rule they were sparse compared with muramidase-positive cells.
Alpha-1 -antitrypsin Staining patterns obtained with antibody to alpha- 1-antitrypsin differed from those described above for Mac387 and muramidase. Positive cells were first observed at 10 weeks. Solitary ovoid blast-like cells were present free within the hepatic sinusoids. In addition, many fusiform or stellate macrophage-like cells were identified within foci of haemopoiesis or lining sinusoids. Some of these positive cells resembled sinusoidal endothelial cells in their location and morphology (Fig. 3). Only a small number of connective tissue cells were positive. There was marked variation in the number of alpha- 1-antitrypsin-positive cells in the liver, independent of the age of the fetus. Alpha-l-antitrypsincontaining cells were also found in normal adult human livers; these were elongated cells, lining
HLA-DR (LN3, CR3/43) In all the fetal tissues studied, only a few solitary cells of irregular shape gave weak positive staining with these antibodies. Positive cells were mainly in portal tracts rather than in sinusoids. By contrast, adult livers contained many LN3-positive stellate cells both in sinusoids and in portal tracts (Fig. 4). CR3/43 mainly stained portal macrophages, and only occasional sinusoidal cells. Granulocytes and monocytes were negative.
Factor XIIIa In all the fetal livers studied, numerous factor XIIIa-positive cells were observed within theconnective tissue of the liver capsule and portal tracts. They had an irregular fusiform or stellate shape and coarse, granular cytoplasm. In addition to these fusiform cells in connective tissue, two other factor XIIIa-expressing cell populations were found in the
MACROPHAGES IN NORMAL HUMAN LIVER
257
mainly been concerned with haemopoietic cells, but muramidase-positive cells were noted in early weeks and LN3-positive cells from the 14th week.15Kamps et aI.l4 found large irregular cells staining with an antibody to HLA-DR (OKIa) and suggested that they might be Kupffer cells. Our own studies clearly show muramidase- and Mac387-positive cells in 7week-old fetuses. These cells, mainly in connective tissue but occasionally within hepatic sinusoids, could represent either a population of primitive monocytes or macrophages, or the earliest Kupffer cells in the sense of cells destined to become the 'fixed' macrophages of the liver. In view of their location, it seems likely that they are the first generation ofprimitive macrophages of yolk sac origin, as suggested by Naito et ai.9and Takahashi et a13' In later fetal life, different staining patterns with the various antibodies became striking. Mac387 consistently labelled large monocytoid cells at the edges of portal tracts, and only a few cells in the sinusoids. Their site is in keeping with that of myelomonocytopoIesis.37 Antibody to muramidase labelled these cells, but additionally marked sinusoidal macrophages. In the third group of fetal Fig. S-Adult liver. Elongated cells staining for factor XIIIa are livers between 30 and 39 weeks of gestation, less muramidase-positive cells were found in connective located in portal connective tissue. Immunoperoxidase tissue, but sinusoidal cells persisted. A possible liver. First, many cells of megakaryocytic differen- interpretation of this change is that monocytotiation were positive. These wereeasily recognized by poiesis had moved to the bone marrow, whereas the their oval shape, fine pale cytoplasm with strongly muramidase-positive cells now represented a popustained peripheral rim, and two or more nuclei. lation of sinusoidal, liver-committed macrophages. Second, there were stellate or fusiform macrophage- Similar stellate muramidase-positive cells were conlike cells within foci of haemopoiesis or lining sinu- sistently found in normal adult liver. These findings soids. Both types of cells were identified at various suggest that muramidase is a consistent marker of times of gestation, but their number decreased with Kupffer cells and that variations in positivity in age. In adult human livers, factor XIIIa-positive disease reflect altered Kupffer-cell physiology rather stellate or elongated macrophages were observed in than the validity of this marker.3 The presence of a few fusiform Mac387-positive portal tracts but not in sinusoids (Fig. 5). In some cells within the sinusoids and portal tracts of fetal livers there was focal granular positivity along the livers probably represents the early stages of transwalls of sinusoids, possibly within endothelial cells. formation from monocytes to macrophages, as Monocytes and granulocytes were consistently suggested by Brandtzaeg and c o - w o r k e r ~ . ' ~ , ~ ' negative. Mac387 detects the leukocyte L1 antigen which is thought to be transferred from monocytes to endothelial cells when the former become adherent.38It is DISCUSSION possible that under pathological conditions in The first cells in man with the ultrastructural which Mac387-positive cells increase in number, characteristics of macrophages appear in the yolk this reflects retention of the molecule by antigensac and mesenchyme in the fourth week of gestation, presenting cells, with reduced transfer to vascular and in the prehaemopoietic liver in the fifth week.3s endothelium. In this way, antigen presentation A few muramidase-positive cells have been ident- might depend on altered monocyteeendothelial cell ified in fetal liver at 7-8 weeks.'* Recent immuno- interaction. Alternatively, Karakucuk et al.16postucytochemical studies in the human fetal liver have late that under pathological conditions such as
258
K. A. B A R D A D I N ET A L .
alcoholic liver disease there might be recruitment of macrophages, seen in part as an increase in Mac387positive cells, with consequent release of cytotoxic mediators. The remaining antibodies gave different results from those obtained for muramidase and with Mac387. Alpha- 1-antitrypsin-containing cells in the hepatic sinusoids were fusiform rather than stellate, and did not resemble classical Kupffer cells. A wide variation in the intensity of alpha- 1-antitrypsin staining was noted, and there appeared to be no correlation with staining for muramidase. Antibody to alpha-1-antitrypsin was the only one to give linear positivity in sinusoidal walls. The gene of the human histocompatibility (HLA) complex consists of at least four subregions (DP, DQ, DX, DR) containing a minimum of one alpha and one beta chain. CR3/43 reacts with the beta chain of all products of the gene subregions DP, DQ, and DR. LN3 recognizes a non-polymorphic antigen of the HLA-DR (Ia) region. The two antibodies give good results in lymph node^.^^.'^ In the present study, little positivity was found in fetal liver, but in the adult large numbers of hepatic macrophages were stained. The results in fetal liver may reflect low expression, only detectable in fresh frozen material,’3.3’and the lack of challenge of the fetal immune system by extrinsic stimuli. Presentation of antigens to lymphoid cells is mainly mediated by macrophages expressing HLA-DR on their surfaces (also known as Ia, or immune antigen). Ia antigens are expressed by murine, rat, and human Kupffer cell^,^^^^' but their role in regulating the immune response remains unclear. HLA-DRexpressing cells have also been demonstrated in portal tracts” and correlated with Steinman and Cohn’s dendritic cells.39 In our study, LN3 and CR3/43-positive cells in portal tracts had the same distribution and characteristics as those of factor XIIIa-positive cells, and should thus be regarded as macrophages. They probably represent a population of connective tissue macrophages, unrelated to Kupffer cells and reacting d i f f e r e n t l ~ .Their ~~ presence in portal tracts is of particular interest in that it might explain trapping of lymphocytes in portal tracts in pathological conditions. Furthermore, these cells might be responsible for the establishment in portal tracts of B- and T-cell zones, as suggested previously.* Factor XIIIa is involved not only in haemostasis, but also in connective tissue organization including proliferation of fibroblast^.^^,^^ This might explain the large number of factor XIIIa-positive cells in
growing fetal tissues, compared with the much smaller number in adult portal tracts. The sinusoidal granules stained by antibody to factor XIIIa have not so far been identified with certainty. They may represent platelets, but it seems more likely from their size, shape, and location that they are the endothelial-cell granules previously described in hepatitis.44 In summary, we have demonstrated marked heterogeneity of phenotypic expression among hepatic macrophages, both in fetal and in adult life. This may reflect variation of phenotypic expression in a single macrophage cell line, or the co-existence in the liver of several distinct populations of macrophages. ACKNOWLEDGEMENTS
This work was supported by the Wellcome Trust. We are grateful to Ms Linda More for introducing AP and JW to technical aspects of immunocytochemistry and for giving valuable help with the project. REFERENCES I . Van den Oord JJ, De Vos R, Facchetti F, Delabie J. De Wolf-Peeters C, Desmet VJ. Distribution of non-lymphoid, inflammatory cells in chronic HBV infection. JPufhol1990; 160 223-230. 2. Bardadin KA, Desmet VJ. Interdigitating and dendritic reticulum cells in chronic active hepatitis. Hisfopa/holog.v 1984; 8: 657-667. 3. Kelly PM, Heryet AR, McGee JO’D. Kupffer cell number is normal, but their lysozyme content is reduced in alcoholic liver disease. J Hepurol 1989; 8: 173-180. 4. Van Furth R, Crofton RW, Diesselhoff-Den Dulk MMC. The bone marrow origin of KupiTer cells. In: Wisse E, Knook DL, eds. Kupffer Cells and Orher Liver Sinusoidal Cells. Amsterdam: Elsevier, 1977; 471480. 5. Gale RP, Sparkes, RS, Golde DW. Bone marrow origin of hepatic macrophages (Kupffer cells) in humans. Science 1978; 201: 937-938. 6. Deimann W, Fahimi HD. Peroxidase cytochemistry and ultrastructure of resident macrophages in fetal rat liver. A developmental study. Dev Biol 1978; 66.43-56. 7 Naito M, Yamamura F, Takeya M, Takahashi K. Ultrastructural analysis of Kupffer cell progenitors. In: Kirn A, Knook DL, Wisse E, eds. Cells of the Hepatic Sinusoid. Vol. 1. Rijswijk: The Kupffer Cell Foundation, 1986; 13-20. 8. Bankston PW, Pino RM. Thedevelopment ofthe sinusoids offetal rat liver: morphology of endothelial cells, Kupffer cells, and the transmural migration of blood cells into the sinusoids. Am J Anaf 1980; 1 5 9 1-15 9. Naito M, Takahashi K., Ohno H . Nishikawa SI. Yolk sac macrophages-a possible Kupffer cell precursor in the fetal mouse liver. In: Wisse E, Knook DL, Decker K, eds. Cells of the Hepatic Sinusoid. Vol. 2 . Rijswjijk: The Kupffer Cell Foundation, 1989; 419420. 10. Fukuda T. Fetal hemopoiesis. 11. Electron microscopic studies on human hepatic hemopoiesis. Virchows Arch B 1974; 1 6 249-270. I I Enzan H, Hara H, Yamashita Y, Ohkita T , Yamane T. Fine structure of hepatic sinusoids and their development in human embryos and fetuses. Acra Parhol Jpn 1983; 3 3 447466. 12. Watanabe S. Nakajima T , Shimosato Y, Shimamura K, Sakuma H. T-zone histiocytosis with SlOO protein. Development and distribution in human fetuses. Arru Patliol Jpn 1983; 33: 15-22.
M A C R O P H A G E S I N N O R M A L H U M A N LIVER 13. Janossy G, Botill M , Poulter LW, er a/. Separate ontogeny of two macrophage-like accessory cell populations in the human fetus. J Immunol1986; 1 3 6 43544361. 14. Kamps WA, Timens W, De Boer GJ, Spanjer H H , Poppema S. In siru study of haemopoiesis in human fetal liver. ScandJImmunol1989: 30: 399408. 15. Timens W, Kamps WA, Rozeboom-Uiterwijk T, Poppema S. Haemopoiesis in human fetal and embryonic liver. Immunohistochemical determination in B5-fixed, paraffin-embedded tissues. Virchows Arch A 1990; 4 1 6 429436. 16. Karakucuk I, Dilly SA, Maxwell J D . Portal tract macrophages are increased in alcoholic liver disease. Hi.stuparhology 1989: 1 4 245-253. 17. Pulford KAF, Rigney EM, Micklem KJ, er a/. KPI : a new monoclonal antibody that detects a monocytejmacrophage associated antigen in routinely processed tissue sections. J Clin Parhol1989: 42: 414421 18. Flavell DJ, Jones DB. Wright D H . Identification of tissue histiocytes on paraffin sections by a new monoclonal antibody. J Hisrochem Cyrodiem 1987;35 1217-1226. 19. Brandtzaeg P, Jones DB, Flavell DJ, Fagerhol M K . Mac387 antibody and detection offormalin resistant myelomonocytic LI antigen. JClin Prrfhol1988; 41: 963-970. 20. Brandtzaeg P. The new monoclonal antibody (Mac387) that reacts with macrophages on paraffin sections detects the well-known leukocyte LI antigen. JHisrochem Cyrochem 1988: 36 1203-1206. 21. Mason DY,TaylorCR.Thedistributionofmuramidase(lysozyme)in human tissues. J Clin Pathol 1975; 28: 126132. 22. Klockars M , Reitamo S. Tissue distribution of lysozyme in man. J Hi.~roc./iemCyrochem 1975; 23: 932-940. 23. Pinkus GS, Said JW. Profile of intracytoplasmic lysozyme in normal tissues, myeloproliferative disorders, hairy cell leukemia, and other pathological processes. An immunoperoxidase study of paraffin sections and smears. Am J Pathol 1977; 8 9 351-366. 24. Mendelsohn G, Eggleston JC, Mann RB. Relationship of lysozyme (muramidase) to histiocytic differentiation in malignant histiocytosis. An immunohistochemical study. Cancer 1980; 4 5 273-279. 25. Kerdel FA, Morgan EW, Holden CA, MacDonald D M . Demonstration of alpha I-antitrypsin and alpha I-antichymotrypsin in cutaneous histiocytic infiltrates and a comparison with intracellular lysozyme. J Am AcatlDermatol 1982; 7: 177-182. 26. Wilson GB, Walker J H Jr, Watkins J H Jr. Wolgroch D. Determination of subpopulations of leukocytes involved in the synthesis of alpha-I-antitrypsin In virro. Proc Soc E , p Biol Med 1980; 164: 105-1 14. 27. Isaacson PG, Jones DB, Millward-Sadler G H , Judd MA, Payne S . Alpha-I-antitrypsin in human macrophages. J Clin Parhol 1981; 34: 982-990. 28. Van Furth R, Kramps JA, Diesselhoff-Den Dulk MMC. Synthesis of a,-anti-trypsin by human monocytes. Clin E x p Immunol 1983; 51: 551-557.
259
29. Perlemutter DH, Sessions Cole F, Kilbridge P, Rossing T H , Colten HR. Expression of the a,-proteinase inhibitor gene in human monocytes and macrophages. Proc Narl Arad Sci USA 1985: 82: 795-799. 30. Marder RJ, Variakojis D , Silver J, Epstein AL. Immunohistochemical analysis of human lymphomas with monoclonal antibodies to B cell and la antigens reactive in paraffin sections. Lab Invest 1985; 52: 497-504. 31. Barbatis C, Kelly P, Greveson J, Heryet A, McGee J O D . Immunocytochemical analysis of H L A class 11 (DR) antigens in liver disease in man. J Clin Parhol1987; 4 0 879-884. 32. Smith MEF, Holgate CS, Williamson JMS, Grigor I, Quirke P, Bird CC. Major histocompatibility complex class TI antigen expression in B and T cell non-Hodgkin's lymphoma. J Clin Pathol 1987; 4 0 3Ml.
33. Nemes Z, Thomazy V, Adiny R, Muszbek L. Identification ofhistiocytic reticulum cells by the immunohistochemical demonstration of factor XIIJ (F-XIlla) in human lymph nodes. J Pathol 1986: 149: 121-132. 34. Adany R, Kiss A, Muszbek L. Factor XIII: a marker of mono- and megakaryocytopoiesis. Br J Haemarol1987: 67: 167-172. 35. Kelemen E, Calvo W, Fliedner TM. Atlas of Human Hemopoietic Development. Berlin: Springer-Verlag, 1979; 83-98. 36. Takahashi K, Yamamura F, Naito M. Differentiation, maturation, and proliferation of macrophages in the mouse yolk sac: a lightmicroscopic, enzyme-cytochemical,immunohistocbemical, and ultrastructural study. J Leukoc B i d 1989: 4 5 87-96. 37. Kamps WA, Cooper M D . Microenvironmental studies of pre-B cell and B cell development in human and mouse fetuses. J ImmunolI982; 1 2 9 52&531 38. Hogg N, Allen C, Edgeworth J. Monoclonal antibody 5.5 reacts with p8,14, a myeloid molecule associated with some vascular endothelium. Eur J Immunol1989: 1 9 1053-1061. 39. Steinman R M , Cohn ZA. Identification of a novel cell type in peripheral lymphoid organs of mice. 1. Morphology, quantitation, tissue distribution. J Exp Med 1973; 137: 1142-1 162. 40. Shiratori Y, Okano K, Matsumoto K, Murao S. Antigen presentation by Kupffer cells in the rat. Scand J Gasrroenterol1984; 19: 733-739. 41. Ramadori G, Dienes HP, Burger R, Meuer S, Rieder H, Meyer zum Buschenfelde K H . Expression of la-antigens o n guinea pig Kupffer cells. Studies with monoclonal antibodies. JHeparol 1986;2: 208-2 17. 42. Beck E, Duckert F , Ernst M . The influence of fibrin stabilizing factor in the growth of fibroblasts in vitro and wound healing. Thromb Haemosr 1961: 6 485491. 43. Bruhn H D , Pohl J. Growth regulation of fibroblasts by thrombin, factor XI11 and tibronectin. Klin Wochenschr 1981: 5 9 145-146. 44. Bardadin KA, Scheuer PJ. Endothelial cell changes in acute hepatitis. A light and electron microscopic study. J Parhol 1984: 144: 21 3-220.
