The differentiation antigens of macrophages in human fetal liver

CLINICAL

IMMUNOLOGY

AND

41, 184- 192 (1986)

IMMUNOPATHOLOGY

The Differentiation Antigens of Macrophages Human Fetal Liver’ LERTLAKANA BHOOPAT,~ Department

of Pathology,

CLIVE

University

R. TAYLOR, of Southern

AND FLORENCE

California,

Los Angeles,

in

M. HOFMAN~ California

90033

Macrophage populations from human fetal liver were examined for the sequential appearance of different antigenic determinant during maturation. Frozen sections of liver, from 12 to 21 weeks gestation were analyzed using a series of four monoclonal antibodies with known specificity. The macrophage monoclonal antibodies used were OKM-I, which defines monocytes, macrophages. and granulocytes; Leu M-3 and MO-2, which identify monocytes and macrophages; and 6B8. a new macrophage monoclonal antibody which binds to tissue macrophages. The staining pattern described by each of these monoclonal reagents was compared with the distribution of morphologically distinguishable tissue macrophages in fetal liver, based on the expression of surface and/or cytoplasmic antigens. The data indicate that the antigens defined by OKM-I and 6B8 are present on large numbers of cells as early as 12 weeks gestation. In contrast, the antigenie determinants identified by Leu M-3 and MO-2 are present only on cells in I5 to 21 weeks of gestation; thus these antigens are mature differentiation antigens. Furthermore, double-staining studies confirmed that with the increase in fetal age unique macrophage populations can be identified based on the matrix of antigenic determinants. Thus, macrophage heterogeneity in the fetal liver may be a function of maturation. D 1986 Academic

Press. Inc.

INTRODUCTION

Heterogeneity of the macrophage population has been recognized and studied on the basis of biochemical (3-5) and morphological characteristics (6-9), tissue distribution and functional properties (10). The reasons for this apparent heterogeneity have yet to be understood. We propose that this macrophage heterogeneity is a reflection of cells at different stages of differentiation. In order to test this hypothesis, we studied macrophage markers at different ages of gestation. The fetal liver was chosen as the model organ analysis since various stages of differentiation would be found in this developing hematopoietic organ (1, 2). The results presented here show that there is indeed heterogeneity based on the expression of antigenic determinant, and these determinants make their appearance at specific stages of development. Thus, the ontogeny of distinct macrophage populations can be determined using monoclonal antibodies to specific antigenic determinants. I This work is supported in part by Grant CA34313 from the United States Public Health Service and Grant RG1678-A-1 from the Multiple Sclerosis Society. 2 Permanent address: Department of Pathology, Faculty of Medicine, Chiang Mai University, Chiang Mai, Thailand. 3 To whom correspondence should be addressed: Department of Pathology, University of Southern California, 2025 Zonal Avenue, Los Angeles, Calif. 90033. 184 0090- 1229186 $1.50 Copyright 0 1986 by Academic Pres,. Inc. A11 rights of reproduction in any form reserved.

MACROPHAGE

DIFFERENTIATION

MATERIALS

IN FETAL

18.5

LIVER

AND METHODS

Tissue rind fixation. Fetal tissues were obtained after clinical termination of pregnancies by suction-assisted mechanical extraction. Fetal age as determined by measurement of foot length (12) ranged from 12 to 21 weeks gestation. Organ specimens approximately 3-5 mm3 were overlaid with OCT embedding medium (Lab-Tek, Naperville, Ill.) and wrapped in aluminum foil. The specimens were snap-frozen in liquid nitrogen and stored at -80°C. Cryostat sections 5-7 km thick were cut and allowed to air dry overnight. The slides were fixed in acetone (reagent grade) for 10 min at 25°C just before staining. Monoclonal antibodies. The specificities of the monoclonal antibodies used are described in Table 1. Each of the reagents are routinely used in our laboratory to study adult tissues, and are known to react with cell types corresponding to their prescribed specificity. 6B8 is a new monoclonal antibody kindly provided by M. Witner and R. Steinman, Rockefeller University, New York. 6B8 was recognized as a distinctive reagent that stains cultured monocytes very actively. 6B8 does not react with isolated lymphocytes, platelets, or dendritic cells. The reagent is an IgM and was used as a culture supernatant. All of the antibodies were used at optimal concentrations as determined by previous titration experiments. Immunoperoxidase staining procedure. This method has been previously described extensively (13). Briefly, the frozen sections were incubated with modified phosphate-buffered saline (PBS) (pH 7.4) for 10 min. The tissue sections were then incubated with the primary antibody for 30 min at 25°C in a humidified chamber. Subsequently, the slides were washed in modified PBS for 10 min. The peroxidase-labeled secondary antibody used depended on the isotype of the primary antibody, either peroxidase-conjugated goat anti-mouse IgG, or IgM (Tago, Inc., Burlingame, Calif.). The reagent was applied to the section for 30 min; the slides were then washed for 10 min in PBS. After this wash, the colored substrate aminoethyl carbozole (AEC) was applied to the tissue and incubated for 10 min. The slides were then rinsed in tap water for 10 mitt, stained with Mayer’s hematoxylin for 3 min, then rinsed in tap water for 10 min and mounted with Aquamount (Lerner Labs, New Haven, Conn.). Controls included the use of an ascites fluid from a mouse injected with a nonsecreting myeloma, in place of the primary antibody. Negligible background TABLE

I

MONOCLONAL ANTIBODIES USED IN STUDY Antibody OKM-

I

Leu M-3 6B8 MO-2 Leu M-4 Anti-mu

Cell specificity Monocytes/macrophages Monocytes/macrophages Tissue macrophages Monocyteslmacrophages Granulocytes Mu heavy chain

* Polyclonal antibody.

granulocytes

Isotype

Source

Ortho Diagnostics Becton-Dickinson Dr. R. Steinman Coulter Electronics Becton-Dickinson Tago

186

BHOOPAT,

TAYLOR,

AND

HOFMAN

staining was observed. A second control included the omission of the primary antibody to determine whether granulocytes or other cells with endogenous peroxidase activity were present. Very few positive cells were observed in the fetal tissues examined. Double-staining procedure. The double-staining procedure has been previously described (14). Briefly, biotinylated horse anti-mouse antibody (Vector, Burlingame, Calif.) was used to couple the first primary monoclonal antibody with avidin-biotin-peroxidase complex (Vector) followed by incubation with AEC to produce a red stain. After washing and application of the second primary monoclonal antibody, the slides were incubated with alkaline phosphatase-conjugated goat anti-mouse IgG (Sigma, St. Louis, MO.) for 30 min. The slides were then washed in NaHCO, (pH 8.3) for 10 min, followed by a 40-min incubation with the alkaline phosphatase substrate composed of 5 mg fast BBN salt, 100 ~1 dimethylformamide, 1000 pl naphthol ASMX alkaline phosphate solution (Sigma), 5 p,l levamisole (1.0 M), and 4.9 ml Tris buffer (pH 8.2) (0.1 M). The slides were then washed in tap water for 10 min and mounted as previously described. A major consideration with double-staining experiments was to ensure that during the second staining procedure, the alkaline phosphatase-conjugated goat anti-mouse immunoglobulin did not bind to the first stage mouse monoclonal antibody. As control for this unwanted binding, the entire procedure was performed, omitting the second primary mouse monoclonal antibody, followed by the alkaline phosphatase-conjugated goat anti-mouse and substrate. Control slides showed no background alkaline phosphatase staining. Other controls included the substitution of the first and second mouse monoclonal antibodies with mouse myeloma serum; no staining was detected. Analysis of staining pattern. The frequency of single-staining cells was determined by counting the number of stained cells per square millimeter of tissue in three randomly selected fields. The distribution of positive cells was designated relative to small blood vessels; cells were not counted near the portal blood vessels because of possible contamination of circulating peripheral blood cells. The double-staining data were expressed in percentage as single-staining cells per 100 cells counted and double-staining cells per 100 cells counted. RESULTS

Twelve to Fourteen

Weeks Gestation

During this early period of gestation, the monoclonal antibodies OKM-1 and 6B8 appeared to stain in great numbers (200-300 cells/mm2) (Table 2). This high frequency of staining was maintained throughout gestation. Morphologically, OKM-l-positive cells appeared to be pleomorphic in size and shape (Fig. I). This monoclonal antibody reagent stained large cells with cytoplasmic extensions, typically identified as tissue macrophages. These cells were scattered individually throughout the tissue, not localized to regions of blood vessels. Also noted were small cells without cytoplasmic processes; these cells appear individually in the region of blood vessels and throughout the tissue. Although the frequency of cells reacting to the 6B8 antibody was similar to that

MACROPHAGE

DIFFERENTIATION

FREQUENCYOFPOSITIVECELLS 12-14 weeks” OKM- I 6B8 Leu M-3 MO-2 Mu Leu M-4

+++++b +++++ + + + e

IN FETAL

187

LIVER

TABLE 2 (/mm2) IN FETALLIVERDURINGGESTATION 15- 17 weeks +++++ +++++ ++ ++ ++ ?

19-21 weeks +++++ +++++ +++ +++ +++ +

a Four samples were tested in each age group. b Represents the number of positive cells present per mm2 of tissue specimen. Range: 0 = -, O-l = +,I-lO= +.lO-30= ++,30-90= +++,90-200= ++++,200-300= +++++.

for OKM-1 (Table 2). the morphology of the 6B8 population was different. The vast majority of the 6B8-positive cells were large with cytoplasmic extensions, while the OKM-l-positive cells were both large and small (Fig. 2). The relationship between these two positive cell populations was made clear in doublestaining experiments. Table 3 demonstrated that the majority of positive ceils expressed both OKM-I and 6B8 antigens (.56%), while a portion of the cells stained with OKM-1 only (32%) and a smaller portion stained with 6B8 only (12%). Thus there are cells which express both OKM-1 and 6B8 simultaneously, while other cells express either OKM-I or 6B8. Leu M-3 and MO-2 stained few cells at 12 weeks gestation (Table 2). The frequency of these populations (both Leu M-3 and MO-2 are ~10 cells/mm2) was similar to that observed for mu-positive B cells (
FIG. I. The 20-week fetal liver stained with OKM-I

shows pleomorphic cells. ( x 500)

188

BHOOPAT,

TAYLOR,

AND HOFMAN

FIG. 2. The 20-week fetal liver stained with 6B8 shows large cells with cytoplasmic e ( x 500)

Fifteen to Seventeen Weeks Gestation With increase in fetal age, the frequency of Leu M-3- and MO-2-positive cells increased dramatically (Table 2). Cells bearing the Leu M-3 determinant were morphologically small and round, without cytoplasmic extensions (Fig. 3); these cells were distributed individually throughout the fetal liver tissue. Leu M-3-positive cells represented a subpopulation of OKM-l-bearing cells, as demonstrated by double-staining studies (Table 3). OKM-1 and Leu M-3 double staining was 25% of the cells counted, while cells positive with Leu M-3 alone counted 3%. In contrast to this pattern, the antigenic determinants identified by Leu M-3 and 6B8 appeared on different cell populations (Table 4): the majority of the cells stained either for Leu M-3 (29%) or 6B8 (66%), while very few cells double stained for both of the determinants on the same cells (5%). Cells expressing the MO-2 an-

FIG. 3. The 20-week fetal liver stained with Leu M-3 shows individually cells. ( x 500)

distributed small, round

MACROPHAGE

COMPARISON

OF PERCENTAGE

DIFFERENTIATION

OF OKM-I DOUBLE

TABLE POSITIVE STAINING Percentage

Antibodies

Single

IN FETAL 3 CELLS WITH OTHER TECHNIQUE positive

189

LIVER

ANTIBODIES

USING

THE

cells per sample”sb

stain

Double

stain

OKM-I 6B8

32 12

56

OKM-1 Leu M-3

72 3

2s

OKM-I MO-2

79 I

20

a Percentage as single and double staining cells per 100 cells counted. b Four samples from 15 to 17 weeks gestation were tested.

tigen were small, round to oval, and were scattered individually throughout the fetal liver tissue (Fig. 4). These cells steadily increased in frequency, with substantial numbers present at 15- 17 weeks gestation (Table 2). MO-2-positive cells were shown to be a subpopulation of OKM-I-bearing cefls, with 20% double staining versus 1% single staining (Table 3). MO-2 was a subpopulation of 6B8bearing cells, as documented in Table 4, with only 5% single staining with MO-2 antibody versus 20% double staining with both MO-2 and 6B8 antibody reagents. The relationships between Leu M-3 and MO-2 using double staining showed that MO-2 was on a population distinct from Leu M-3 (3% double staining, while Leu M-3 alone is 60% and MO-2 alone is 36%). Nineteen to Twenty-One Weeks Gestation At this stage of maturation, the frequency of the different cell populations increased (Table 2). The relationship among the different cells remained approximately the same as compared with staining results at 15- 17 weeks gestation. The morphology and distribution of the cells staining with the panel of monoclonal

COMPARISON

OF PERCENTAGE

TABLE 4 OF 6Bg POSITIVE CELLS WITH OTHER DOUBLE STAINING TECHNIQUE Percentage

Antibodies

Single

positive

ANTIBODIES

USING

cells per sample”,b

stain

Double

stain

6B8 Leu M-3

66 29

5

688 MO-2

15 5

20

n Percentage as single and double staining cells per I00 cells counted. b Four samples from 15 to 17 week gestation were tested.

THE

190

BHOOPAT,

TAYLOR,

AND HOFMAN

FIG. 4. The IS-week liver stained with MO-2 shows small, oval cells with cleaved nuclei in single distribution. ( x 500)

antibodies did not noticeably alter with the increase in gestational age. However, the fetal liver did become more dense, and the number of hepatocytes increased. DISCUSSION

The human fetal liver is the site of early, pre-bone marrow, hematopoiesis (1, 2). We had previously demonstrated that fetal liver contains substantial numbers of B cells, few T cells, and high numbers of large cytoplasmic cells with intracellular IgG (13); these large cells with cytoplasmic extensions appeared to be macrophages by morphologic criteria. Using a panel of monoclonal antibodies to monocyte/macrophage antibodies, we studied the mononuclear cell populations present in the developing fetal liver at different ages of gestation. Monoclonal antibodies have already been used to show heterogeneity of the macrophage populations in adult tissue (14). We demonstrated that the staining pattern of monoclonal antibody reagents, which identify monocyte-macrophage determinants, describes unique population distributions in lymphoid tissue. While the OKM-1 antibody binds to histiocytes that are widely distributed throughout the lymph node, the Leu M-3 reagents stains predominantly histiocytes and dendritic reticular cells (DRC). The different macrophage cell populations, defined by phenotype, were then correlated with the supposed function of the corresponding cells. Histiocytes, cells lining the sinuses and present in the germinal centers, exhibit phagocytic function (IO, 12, 15-17); these are OKM-1 reactive. The macrophages in the germinal center, particularly those underneath the mantle area of the follicle (the DRC), are thought to be involved in presentation of antigen to the T and B cells of the germinal center (18-20); these cells are Leu M-3 positive. Thus the phenotypic markers do appear to correlate with different functional and anatomical subgroups of cells within the macrophage family. Our results show that as early as 12 weeks gestation there are relatively high numbers of OKM-1 and 6B8-positive cells (200-300 positive cells/mm*), compared with the other markers evaluated. Double-staining studies further demonstrated that while many of the cells expressed both OKM-1 and 6B8 deter-

MACROPHAGE

DIFFERENTIATION

IN FETAL

12 weeks

15-20

OKM-1

191

LIVER

weeks

OKM-1 o-

OKM-1 686

OKM-1 LeuM-3

BBB

OKM-1 685 MO-2

c-

BBB

FIG.

5. Schematic

diagram

of phenotypic

markers

of macrophages

during

ontogeny.

minants. there were cells which expressed OKM-1 (30%) or 6B8 (16%) alone (Table 3). The morphological characteristics of OKM-1 cells span the range from small round cells, found in clusters within the liver parenchyma, to large cells with cytoplasmic extensions present as single cells scattered throughout the tissue. In contrast to OKM-1, 6B8 appears to recognize a rather homogeneous, large cell population with cytoplasmic extensions. The reagent 6B8 was shown to stain Kupffer cells. liver macrophages, in the adult liver (data not shown), which would indicate that this antibody may be staining long-term resident macrophages of the liver. The OKM-1 antibody did not stain the adult liver (data not shown). Thus the 6B8-positive cells in the fetal liver may be the precursors of actual Kupffer cells and distinct from OKM-1 positive cells. As Van Furth (11) proposed, tissue macrophages may be characteristically distinct populations, e.g., circulating monocytes (recent immigrants) and long-term residents of the organ: both populations, however, are originally derived from the bone marrow during ontogeny. By 15 weeks gestation, there are four populations with distinct groups of antigenie determinants (Fig. 5). The double-staining data, using the panel of described monoclonal antibodies, distinguished these populations from the 12 week macrophages by the appearance and distribution of two new antigenic determinants, as identified by Leu M-3 and MO-2. These antigenic determinants appear later in fetal development and therefore determine more mature macrophage populations. The antigens, defined by Leu M-3 and MO-2 can thus be considered macrophage differentiation antigens. Another, perhaps more conclusive approach to the study of the sequential expression of differentiation antigens would be using cells maintained in culture, which is under study. The stepwise appearance of antigenic markers on macrophages may provide an important tool in dissecting the different stages of macrophage maturation and therefore lead to an understanding of aberrations that may occur in the disease process. ACKNOWLEDGMENTS The authors thank Dr. R. Steinman for his criticism of this manuscript. staff of lnglewood Hospital for providing tissue specimens. and Bradley pertise in processing the tissue.

Dr. Morton Barke and the Lyons for his technical ex-

192

BHOOPAT,

TAYLOR,

AND HOFMAN

REFERENCES 1. Metcalf, D., and Moore, M. A. S., In “Haemopoietic Cells” A. Newberger and E. L. Tatum, Eds.), pp. 34-35, North-Holland, Amsterdam/London, 1971. 2. Hayward, A. R., Immunol. Rev. 57, 39, 1981. 3. Cohn, Z. A., and Benson, B., J. Exp. Med. 121, 153, 1965. 4. Fishman, M., and Weinberg, D. S., Cell. Immunol. 45, 437, 1979. 5. Gordon, S., and Cohn, Z. A,, lnt. Rev. Cvtol. 36, 171, 1974. 6. Hirsch, S., and Gordon, S., Int. Rev. Exp. Pathol. 25, 51, 1983. 7. Van Furth, R., Hirsch, .I. G., and Fedorko, M. E., J. Exp. Med. 132, 794, 1970. 8. Sutton, J. S., and Weiss, L., J. Cell Biol. 28, 303, 1966. 9. Kaplan, G., and Gaudernack, G., J. Exp. Med. 156, 1101, 1982. 10. Muller-Hermelink, H. K., and Kaiserling, E., In “Malignant Lymphoproliferative Diseases” (J. G. Van den Tweel et al., Eds.), p. 57. Martinus Nighoff, The Hague/Boston/London, 1980. 11. Van Furth, R., and Diesselhoff H. M. C., J. Exp. Med. 160, 1273, 1984. 12. Lee, K. C., Mol. CeII. Biochem. 30, 39, 1980. 13. Hofman, F. M., Danilovs, J., Husmann, L., and Taylor, C. R., J. Immunol. 133(3), 1187, 1984. 14. Hofman, E M., Lopez, D., Husmann, L., Meyer, P. R.. and Taylor, C. R., Cell. Immune/. g&61, 1984. 15. Van Furth, R., In “Mononuclear Phagocytes, Functional Aspects” (R. Van Furth, Ed.), p. 1, Martinus Nighoff, Hague, 1980. 16. Bloom, W., and Fawcett, D. W., In “A Textbook of Histology,” 10th ed. p. 7, Saunders, Philadelphia/London/Toronto, 1975. 17. Metcalf, D., In “Macrophages and Natural Killer Cells: Regulation and Function” (S. J. Normann and E. Sorkin, Eds.), p. 33, Plenum, New York, 1982. 18. Klaus, G. G. B., Humphrey, J. H., Kunkel, A., and Dongworth, D. W.. fmtmlnol. Rev. 53, 1, 1980. 19. Mandel, T. E., Phipps, R. P., Abbot, A., and Tew, J. G., Immunol. Rev. 53, 29, 1980. 20. Szakal, A. K., Homes, K. L., and Tew, J. G., J. Immunol. 131, 1714, 1983. Received January 24, 1986; accepted with revision May 6, 1986