THE ANATOMICAL RECORD 232:520-526 (1992)
Macrophage Development: I. Rationale for Using Griffonia simplicifolia lsolectin B, as a Marker for the Line SERGE1 P. SOROKIN AND RICHARD F. HOYT, JR. Laboratory of Pulmonary Cell Biology, Department of Anatomy and Neurobiology, Boston University School of Medicine, Boston, Massachusetts
ABSTRACT
The isolectin B, of Griffonia simplicifolia (GSA I-B,) binds to cell membrane glycoconjugates bearing terminal alpha-D-galactose, which macrophages possess. We have investigated the merits of its use as a marker for cells of this lineage when examining the early origin of macrophage populations in r a t embryos, the stages and time scale of transformation from precursor forms to active, matured cells, and the response of precursors and macrophages to colonystimulating blood factors, the last two studies conducted in organ cultures of prenatal lungs. In the present instance, GSA I-B, was used either coupled with fluorescein (FITC) for light microscopy of living and fixed cells, or with peroxidase for light or electron microscopy. Control incubations of lung culture-derived macrophages proved that staining resulted from specific binding to galactosyl units on the cell membrane, since it was competitively inhibited by alpha-D-galactose. The lectin binds to few cells in 14-day prenatal lung explants but to a great many macrophages that subsequently develop in the cultures, indicating that it can be relied on for quantitative studies on population growth; however, it is important to provide reagents with good access to the cells. Apart from macrophages and their precursors, virtually no cells in prenatal lung cultures bind this lectin. Granulocytes of adult blood are GSA positive, but they are not yet present in 14-day prenatal explants and do not develop subsequent to culturing; hence they are not a source of confusion for experimental studies using this system. Precursors of granulocytes begin to appear in rat embryos around day 13 and have GSA-positive cell membranes, but like definitive granulocytes they also have conspicuous peroxidase-positive lysosomal granules which serve to distinguish them from early macrophages, particularly when cells are studied a t a n ultrastructural level. With these objections cleared away, GSA I-B, emerges as a valuable means to mark cells of the macrophage line, mature or immature.
In 1984 we showed that macrophage precursors and macrophage-like cells are present in vivo in developing lungs of rabbits, rats, and hamsters, beginning with the early bronchial bud stage (Sorokin et al., 19841, and postulated that such cells contribute significantly to a resident population of alveolar macrophages in adult lungs, as against the then-prevailing concept that alveolar macrophages are almost exclusively derived from repeated seedings of the lungs by blood-borne monocytes. Later studies showed that organ cultured embryonic rodent lungs consistently give rise to a pure population of mononuclear phagocytes which by manifold criteria (ultrastructure, lysosomal enzyme activity, endocytotic activity, Fc and complement receptors, and macrophage-specific membrane markers) are indistinguishable from typical alveolar macrophages of adult lungs, and which proliferate in the absence of lymphocytes (Sorokin and Hoyt, 1987; Sorokin et al., 1989a,b). Within 24 hours of explantation, numerous macrophages became readily identifiable in the connective tissue stroma, and hints that transformation is already underway were found as early as 6 hours. This paper examines pros and cons for using a selec0 1992 WILEY-LISS, INC.
tive lectin marker adaptable to light and electron microscopy to trace the transformation of precursor forms into macrophages in lung cultures, as well as for following the early history of macrophages in intact embryos prior to that time. In a survey of peroxidase labeled lectins showing affinity for cells of developing rat lungs, Honda et al. (1989) found that the agglutinin I-B, of Griffoniasimplicifolia (GSA I-B,), which recognizes terminal alpha-galactose, binds to macrophagelike cells with sufficient selectivity to qualify as a potential marker for the line. They applied it to paraffin sections of 14-21 day prenatal, early postnatal, and adult lungs, and to ultrastructural specimens from a 1-day postnatal and a n adult lung, finding GSA I-B,labeled mononuclear, “monocytoid,” or macrophage-
Received June 19, 1991; accepted October 2, 1991. Address reprint requests to Professor Sergei P. Sorokin, Pulmonary Cell Biology, Department of Anatomy and Neurobiology, Boston University School of Medicine, 80 East Concord Street, Boston, MA 02118.
521
LECTIN MARKER FOR DEVELOPING MACROPHAGES
like cells present at all developmental stages, and indicating semi-quantitatively that free lectin-binding cells become increasingly prevalent towards birth. As will be shown here, the lectin can indeed be used to follow expansion of macrophage populations during pulmonary development, provided a number of important constraints are observed. The same approach and methodology may well have equal application to other systems where macrophages figure importantly, beginning with hematopoiesis in the yolk sac (Takahashi e t al., 1989) and extending to the reactions of microglia in the adult brain, where GSA I-B, has been successfully used for some time as a marker for this macrophage species (Streit and Kreutzberg, 1987). MATERIALS AND METHODS
Various tissues but principally the lungs from embryonic rats aged 13 or 14 days of gestation provided specimens for study. Fourteen-day lungs were also grown as organ cultures and harvested a t intervals ranging from a n hour to several days thereafter. In addition, smears and buffy coat preparations were made from the blood of adult rats. Fetal lung cultures were either of entire right or left lungs, or the pair from one animal. They were grown in T-flasks at 375°C on the surface of growth medium containing 40% fetal bovine serum, Gey's balanced salt solution, and 0.3% glucose, solidified by the inclusion of 1.5% agar, according to our routine practice (Sorokin and Hoyt, 1987). These cultures were extensively sampled; those furnishing illustrations for this paper were taken after 2 and 4 days in vitro. Most tissues and organ cultures were intended for labeling with peroxidase-coupled GSA I-B, (Sigma Chemical Co., St Louis, MO). These were fixed with 1% glutaraldehyde in 0.1 M phosphate-buffered saline (PBS, pH 7.2, 4"C, 2 hr), and rinsed in two changes of buffer (4"C, overnight). While immersed in buffer, specimens were lightly pricked with a pin or cut into smaller pieces to improve access of the lectin to cells within. They were then incubated in horseradish peroxidase (HRP)-labeled GSA I-B, (1.0 Fgm/ml) made up in buffer, pH 7.2, consisting of 0.1 M PBS with 0.1 mM MnCl,, MgCl,, and CaCl, (4"C, 24 hr). Following another rinse in PBS (1 hr) and a succeeding one in 0.05 M Tris-HC1, pH 7.6 (10-15 min), sites binding the lectin were revealed by incubation after Graham and Karnovsky (1966) in a diaminobenzidine substrate (DAB, 100 mg per 200 ml 0.05 M Tris-HC1, 10 min), followed by one containing both DAB and hydrogen peroxide (1 ml 1% H,O, per 100 ml Tris-DAB solution, 10 min). Tissues were then rinsed once in Tris-HC1 and once in PBS before being postfixed in PBS-buffered 1%osmium tetroxide (4"C, 1 hr), dehydrated in graded ethanols, and embedded in Epon. Specimens were examined as whole mounts, as semi-thin (0.5 wm) sections counterstained with toluidine blue, and as thin sections for electron microscopy, both unstained as well a s after staining with uranyl acetate and lead citrate. Adult rat blood was mixed with 5 units/ml of heparin in 0.1 M phosphate-buffered saline and spun in a clinical centrifuge. The buffy coat was then either siphoned off and fixed in 1% glutaraldehyde in 0.1 M phosphatebuffered saline (PBS, pH 7.2,4"C, 2 hr), or fixed in situ using 1.5% glutaraldehyde with 1%sucrose in 0.1 M
cacodylate-HC1 buffer (pH 7.4, 4"C, 3 hr) to obtain a buffy coat disc for subsequent incubation, processing, and sectioning (Anderson, 1965). Following fixation, suspended leukocytes were spun down to form a pellet and then incubated for GSA I-B, according to the preceding protocol, suspending the cells and repelleting at each step. Cells in the buffy coat disc were processed like solid tissue. Specificity of the peroxidase-coupled GSA I-B, reaction was ascertained by comparison of specimens incubated in the lectin with those incubated in control media. 1)The principal control depended on competitive inhibition of GSA I-B, binding to membrane sites in the presence of 0.1 M alpha-D-galactose, added to the normal lectin staining solution. Other control procedures included 2) incubation of specimens in diaminobenzidine (DAB)/H,O, without prior exposure to the lectin, and 3) reincubation of specimens previously exposed to lectin/galactose and DAB/H,O, in fresh lectin medium without galactose, followed by DAB/H,O, (a "rerun"). No effort was made to suppress the intrinsic peroxidase activity of granules in certain leukocytes, since this criterion could be helpful in discriminating among cell types present. All these control incubations were carried out on 4-day lung cultures, when macrophages were sure to be abundant. Controls 1 and 2 were also applied to buffy coat specimens. Two live 4-day lung cultures having macrophages on the pleural surface as well as within were exposed (22"C, 2 hr) to 10 Fg/ml fluorescein isothiocyanate (FITCI-coupled GSA I-B, (Sigma Chemical Co., St. Louis, MO) in 0.1 M PBS, pH 7.2, containing 0.1 mM MnCl,, MgCl,, and CaCl,, introduced dropwise onto the surface of the medium in the T-flasks. The cultures were thereafter rinsed in PBS and examined under a n ultraviolet microscope equipped with a filter block appropriate for the wave length of fluorescein (B filter, cat number 78946, Nikon, New York, NY). A few airdried smears of adult rat blood were similarly treated. The FITC-coupled reaction was controlled by preincubating a smear in a 1%solution of periodic acid (22"C, 10 min) before immersion in the lectin substrate, in order to destroy potential binding sites. RESULTS
Staining by GSA I-B, was examined by light and electron microscopy after applying the FITC and/or peroxidase coupled lectin to 1) well-developed macrophages in the lung cultures, as well as to 2) peripheral blood smears and buffy coat pellets of adult rat blood; 3) semi-thin and thin sections of 14-day prenatal rat lung; and 4) semi-thin and thin sections of connective tissue from 13- and 14-day rat embryos, In order to confirm our identifications of specific cell types in lectinstained specimens, we also examined Wright's-stained smears of adult blood and Giemsa-stained serial glycol methacrylate sections of whole 14-day rat embryos prepared earlier (Sorokin and Hoyt, 1987). Comparison of Test and Control Preparations
1) Pretreatment of blood smears with periodic acid abolished all selective staining by FITC-coupled GSA I-B,. 2) Control incubations in 14 + 4 day lung cultures satisfied u s that observed peroxidase-coupled GSA I-B, staining of macrophage cell membranes was
522
S.P. SOROKIN AND R.F. HOYT, JR.
attributable to lectin binding of terminal galactose units on membrane components (Fig. 1). In electron micrographs of macrophages otherwise unstained by uranyl acetate or lead citrate, the GSA I-B,-peroxidase reaction outlined each cell completely and continued along membrane invaginations but did not appreciably involve intracytoplasmic membranes (Fig. la). Addition of 0.1 M alpha-galactose to the incubation medium completely prevented binding of lectin-peroxidase to these cells (Fig. lb). Exposure to DAB/H,O, alone resulted in general staining of the cytoplasm of free erythroblasts or of those ingested by the macrophages, but cell membranes remained unstained (Fig. lc). In thin sections of buffy coat leukocytes, a similar control left intracytoplasmic granules stained owing to their content of intrinsic peroxidases, but cell membranes were unstained. Such reactive granules occurred principally among granulocytes or their precursors and much less frequently among cells judged “mononuclear,” where granules were smaller in size and few in number. Occasionally a cell exposed only to DAB/H,02 exhibited peroxidase staining along a short segment of cell membrane, and among sectioned tissues incubated in the GSA test medium, staining of macrophages sometimes became transferred to the plasmalemma of a n adjacent fibroblast around points of mutual contact, leaving the rest of the fibroblast’s cell membrane unstained (see Fig. 4).These observations led to our adoption of the criterion that a cell must display peroxidase staining along virtually the entire circumference to be considered marked by GSA lectin binding. A final control, the rerunning of galactose-inhibited (unstained) specimens in fresh GSA I-B, medium without the hexose, resulted in positive staining of macrophage cell membranes (Fig. Id).
similar way, where cultures were surrounded by a tightly packed corona of macrophages, only those in the peripheral layer were stained, whereas if the cells were more loosely packed, all stained. Thus we attributed occasional absence of staining to inadequate penetration of reagents rather than to the presence of a subpopulation of GSA-negative macrophages. In practice, better staining indeed occurred in specimens that were extensively pricked prior to incubation. GSA 1-6, Binding to Circulating Leukocytes in Adult Rats
As examined by light microscopy of air-dried smears of fresh whole blood, FITC-coupled GSA I-B, binds most strongly to the surface and cytoplasm of large leukocytes with a notched, and often more than simply lobated nuclear profile. After comparing lectin-stained with Wright’s-stained smears of the same blood, we identified most of the positive cells as heterophils and some as monocytes. Red cells were unstained as were many of the small circulating lymphocytes. Electron micrographs of the buffy coat incubated with peroxidase-coupled GSA I-B, indicated that binding was present on cell membranes of granulocytes and on the surfaces of platelets. Among heterophils intrinsic peroxidase was localized to numerous intracytoplasmic granules; in the smaller population of eosinophils intrinsic peroxidase seemed localized primarily to the cortical portion of the large, crystalloid-bearing specific granules. We found no basophils, but our samples included several large (- 20 pm) mast cells with large granules and imperfectly preserved, apparently GSAnegative cell membranes. Among mononuclear cells, typical monocytes stained and many smaller cells resembling lymphocytes did not. More objectively, about three quarters of the larger cells having notable development of granular endoplasmic reticulum bound the Staining of the Pulmonary Macrophage Population lectin, and more than half of the smaller cells having Electron microscopy of lung explants grown two or a n undifferentiated cytoplasm and many polysomes did more days in vitro reveals that virtually all GSA-pos- not. itive cells are morphologically recognizable a s macrophages (Fig. l),whereas light microscopic examination GSA-Reactive Cells in Embryonic Tissues of whole mounts gives one a n overview of the entire Comparatively few cells bound GSA I-B, in 13- or population. When a series of lung cultures harvested periodically between 2-7 days was incubated with GSA 14-day embryonic or extraembryonic tissues, and they I-B, and examined intact, lectin-binding cells within appeared to represent a single population. This was were observed progressively to increase in number made up of rounded cells resembling undifferentiated with time and to change their distribution inside the leukocytes and solitary stellate or angular cells that culture, behaving exactly like cells we had studied ear- except for their GSA staining resembled mesenchymal lier by other techniques and determined to be macro- fibroblasts. Only a small number contained phagocytic phages (Sorokin and Hoyt, 1987; Sorokin et al., 1989b). inclusions. Megakaryocytes as a rule were GSA negaFor example, at 2 days after explantation, numerous tive prior to platelet formation, and red cell memGSA-positive cells were recognized within a cultured branes were almost never stained. lung pair (Fig. 2, peroxidase-coupled lectin), but the The leukocyte-like GSA-positive cells were agranunumber increased markedly by 4 days when many cells lar and occurred both within blood vessels and outside, had emerged onto the pleural surface (Fig. 3, FITC- typically in mesenchyme where the majority of GSAcoupled lectin). GSA I-B, therefore appears to bind to positive cells assumed the angular form. Within this the great majority of macrophages present. tissue it was easy to find some GSA-positive cells in In cases where recognizable macrophages in a thin mitosis (Fig. 4). Because these cells appear to transsection of lung culture were not all stained, the defi- form into typical macrophages after embryonic rudiciency was usually apparent among cells most removed ments are placed in organ culture, we consider them from reagents in the incubation medium. Conversely, early representatives of the macrophage line. A fuller macrophages lying next to a pin prick through the account of their early history is given in the second pleura were often more strongly outlined by GSA-per- paper of this series, and changes after organ culturing oxidase than macrophages at some distance away. In a are presented in the third (Sorokin et al., 1991a,b).
LECTIN MARKER FOR DEVELOPING MACROPHAGES
Fig. 1. GSA I-B,-peroxidase reaction and controls, shown in electron micrographs of macrophages from 4-day (14 4) lung cultures. a: Standard incubation medium. The lectin binds to the cell membrane. Preparation was not counterstained with uranyl acetate/lead citrate (UNPb). x 6300. b Standard medium with 0.1 M alpha-D-galactose. Lectin binding is completely inhibited, although red cell fragments (top) exhibit a cytoplasmic peroxidase-like reaction attributable to
+
523
hemoglobin. Counterstained with UA/Pb. x 4800. c: Incubation in DAB/H,O, alone. A peroxidase-like reaction is present only in red cells, whether free or ingested by the macrophages. UA/Pb. x 2500. d Rerun of galactose-inhibited incubation in fresh GSA medium without galactose. The macrophage cell membrane now binds the lectin. UA/Pb. x 6700.
S.P. SOROKIN AND R.F. HOYT, J R .
524
Fig. 2. GSA I-B,-peroxidase reaction in a whole mount of 14-day fetal rat lungs after 2 days in culture (14 + 2). Numerous macrophages (dark spots) stand out against the left and right bronchial axes (pale areas). Hemoglobin of fetal erythroblasts in engorged blood vessels beside the bronchi exhibits peroxidase-like activity. x 60.
Fig. 3.Living culture of fetal rat lung at 14 + 4 days, after a 2-hr exposure to FITC-coupled GSA I-B,. Macrophages bind the lectin and so are visible under ultraviolet light. The plane of focus shows them emerging onto the pleural surface, and many more cells are present than after 2 days in vitro (see Fig. 2). x 60.
DISCUSSION
were in general GSA-negative until cells began to undergo fragmentation into platelets, which also bound the lectin, as noted in buffy coat preparations. Membranes of nucleated or enucleate erythrocytes almost never became stained, except among cells in certain organ cultures where tissue disruption and cell death were plainly evident. Indeed, erythrocytes are not considered to interact with ligands recognizing galactosyl residues until the cells become aged or damaged and their surface coating undergoes partial desialation to unmask them (Schlepper-Schafer and Kolb-Bachofen, 1988). Our study indicates that galactosyl residues are present on macrophage cell membranes, and the cells therefore can be tagged by GSA I-B, whenever morphologically recognizable as macrophages, whether they occur in prenatal or postnatal tissues, or in lung cultures a few days old. It is equally evident that isolated cells having either a leukocyte-like or mesenchyme-like appearance bind GSA I-B, in embryos; the succeeding two papers in this series provide evidence to show that these are precursors of macrophages. In sum, our experience is that peroxidase-coupled GSA I-B, can be relied on to mark macrophages and their precursors, particularly when used at an electron microscopic level where ultrastructural characteristics as well as the ac-
The appropriateness of relying on GSA I-B, as a marker for cells in the lineage of pulmonary macrophages was evaluated in a series of steps, initially by finding that typical macrophages of lung cultures bound this lectin on their cell membranes and that the staining reactions were well controlled and specific. Second, we sought to determine whether all or only some of the macrophages are stainable and found that virtually the whole population can be tagged, provided the cells have adequate access to the reagents. Thereafter, guided by an assumption that macrophages have hematopoietic forebears, we wished to learn if other kinds of mature leukocytes also bind the lectin and what the binding capacity of the leukocytes in the embryo might be, both circulating and in formation, a t the moment the lung rudiments were removed to culture. The results indicate that for rats, peroxidase-coupled GSA I-B, stains immature mononuclear cells in embryos. In adults certain mononuclear cells bind the lectin, specifically monocytes and perhaps a number of immature related forms present in the circulation, but typical lymphocytes more often than not are unreactive. Mature and immature granulocytes of heterophil and eosinophil lines also bind the lectin. In the megakaryocytic line, those encountered in embryonic tissues
LECTIN MARKER FOR DEVELOPING MACROPHAGES
Fig. 4. Dividing GSA-positive macrophage precursor in the mesenchyme of a 13-day rat embryo. Surrounding mesenchymal fibroblasts are unreactive, but one can see that where macrophage and fibroblast membranes are apposed, some transfer of staining has occurred. In
tivity of intrinsic peroxidases in lysosomes andlor specific granules would prevent one from mistaking granulocytes for macrophages. The same experience would nevertheless make us cautious about depending on it exclusively to identify macrophages in paraffin sections of late fetal or postnatal material, because the more gestation is prolonged beyond the stage represented by the 1Cday prenatal rat, the more severe will be contamination of the growing population of macrophages by equally GSA-positive granulocytes, and cells in both categories would be stained dark brown. In such specimens, therefore, it is a difficult task to obtain a reliable estimate of macrophage population growth based solely on light microscopic counts of GSA I-B,positive cells. In contrast, organ cultured prenatal lungs are cut off from further influx of cells from the developing embryo. Here, where circumstances favor generation of a pure population of macrophages unmixed with other leukocytes, the lectin marker can be very useful at the light microsopic as well as ultrastructural level for studies assessing population growth and related macrophage dynamics.
525
contrast, fibroblast surfaces not apposed to the dividing cell are not stained. This common artifact of GSA I-B, staining poses little diffculty for identifying lectin-positive cells. x 12,000.
ACKNOWLEDGMENTS
The participation of Dana G. Blunt and Nancy A. McNelly is gratefully acknowledged. The work was supported by USPH research grant HL-33070. LITERATURE CITED Anderson, D.R. 1965 A method of preparing peripheral leucocytes for electron microscopy. J. Ultrastr. Res., 13:263-268. Graham, R.C., and M.J. Karnovsky 1966 The early stages of absorption of injected horseradish peroxidase in the proximal tubules of mouse kidney: ultrastructural cytochemistry by a new technique. J . Histochem. Cytochem., I4r291-302. Honda, T., B.A. Schulte, and S.S. Spicer 1989 Glycocongugate with terminal galactose. A selective property of macrophages in developing rat lung. Histochemistry, 91r61-67. Schlepper-Schafer, J., and V. Kolb-Bachofen 1988 Red cell aging results in a change of cell surface carbohydrate epitopes allowing for recognition by galactose-specific receptors of rat liver macrophages. Blood Cells, 14r259-269. Sorokin, S.P., and R.F. Hoyt, J r . 1987 Pure population of nonmonocyte derived macrophages arising in organ cultures of embryonic rat lungs. Anat. Rec., 21 7:35-52. Sorokin, S.P., R.F. Hoyt, Jr., and M.M. Grant 1984 Development of macrophages in the lungs of fetal rabbits, rats, and hamsters. Anat. Rec., 208r103-121.
526
S.P. SOROKIN AND R.F. HOYT, J R .
Sorokin, S.P., R.F. Hoyt, Jr., and N.A. McNelly 1989a Nonimmunemediated phagocytosis by “premedullary” lung macrophages: Effects of concanavalin A, tuftsin, and macrophage-inhibitory peptide. Anat. Rec., 223t55-61. Sorokin, S.P.,L. Kobzik, R.F. Hoyt, Jr., and J. J. Godleski 1989b Development of surface membrane characteristics of “premedullary” macrophages in organ cultures of embryonic rat and hamster lungs. J. Histochem. Cytochem., 37t365-376. Sorokin, S.P., R.F. Hoyt, Jr., D.G. Blunt, and N.A. McNelly 1992a Macrophage development. 11. Early ontogeny of macrophage populations in brain, liver, and lungs of rat embryos as revealed by a lectin marker. Anat. Rec., 232:527-550.
Sorokin, S.P., N.A. McNelly, D.G. Blunt, and R.F. Hoyt, J r . 199213 Macrophage development. 111. Transformation of pulmonary macrophages from precursors in fetal lungs and their later maturation in organ culture. Anat. Rec. 232551-571. Streit, W.J., and G.W. Kreutzberg 1987 Lectin binding by resting and reactive microglia. J. Neurocytol., 16:249-260. Takahashi, K., F. Yamamura, and M. Naito 1989 Differentiation, maturation, and proliferation of macrophages in the mouse yolk sac: A light-microscopic, enzyme-cytochemical, immunohistochemical, and ultrastructural study. J. Leuk. Biol., 45:87-96.
Macrophage Development: I. Rationale for Using Griffonia simplicifolia lsolectin B, as a Marker for the Line SERGE1 P. SOROKIN AND RICHARD F. HOYT, JR. Laboratory of Pulmonary Cell Biology, Department of Anatomy and Neurobiology, Boston University School of Medicine, Boston, Massachusetts
ABSTRACT
The isolectin B, of Griffonia simplicifolia (GSA I-B,) binds to cell membrane glycoconjugates bearing terminal alpha-D-galactose, which macrophages possess. We have investigated the merits of its use as a marker for cells of this lineage when examining the early origin of macrophage populations in r a t embryos, the stages and time scale of transformation from precursor forms to active, matured cells, and the response of precursors and macrophages to colonystimulating blood factors, the last two studies conducted in organ cultures of prenatal lungs. In the present instance, GSA I-B, was used either coupled with fluorescein (FITC) for light microscopy of living and fixed cells, or with peroxidase for light or electron microscopy. Control incubations of lung culture-derived macrophages proved that staining resulted from specific binding to galactosyl units on the cell membrane, since it was competitively inhibited by alpha-D-galactose. The lectin binds to few cells in 14-day prenatal lung explants but to a great many macrophages that subsequently develop in the cultures, indicating that it can be relied on for quantitative studies on population growth; however, it is important to provide reagents with good access to the cells. Apart from macrophages and their precursors, virtually no cells in prenatal lung cultures bind this lectin. Granulocytes of adult blood are GSA positive, but they are not yet present in 14-day prenatal explants and do not develop subsequent to culturing; hence they are not a source of confusion for experimental studies using this system. Precursors of granulocytes begin to appear in rat embryos around day 13 and have GSA-positive cell membranes, but like definitive granulocytes they also have conspicuous peroxidase-positive lysosomal granules which serve to distinguish them from early macrophages, particularly when cells are studied a t a n ultrastructural level. With these objections cleared away, GSA I-B, emerges as a valuable means to mark cells of the macrophage line, mature or immature.
In 1984 we showed that macrophage precursors and macrophage-like cells are present in vivo in developing lungs of rabbits, rats, and hamsters, beginning with the early bronchial bud stage (Sorokin et al., 19841, and postulated that such cells contribute significantly to a resident population of alveolar macrophages in adult lungs, as against the then-prevailing concept that alveolar macrophages are almost exclusively derived from repeated seedings of the lungs by blood-borne monocytes. Later studies showed that organ cultured embryonic rodent lungs consistently give rise to a pure population of mononuclear phagocytes which by manifold criteria (ultrastructure, lysosomal enzyme activity, endocytotic activity, Fc and complement receptors, and macrophage-specific membrane markers) are indistinguishable from typical alveolar macrophages of adult lungs, and which proliferate in the absence of lymphocytes (Sorokin and Hoyt, 1987; Sorokin et al., 1989a,b). Within 24 hours of explantation, numerous macrophages became readily identifiable in the connective tissue stroma, and hints that transformation is already underway were found as early as 6 hours. This paper examines pros and cons for using a selec0 1992 WILEY-LISS, INC.
tive lectin marker adaptable to light and electron microscopy to trace the transformation of precursor forms into macrophages in lung cultures, as well as for following the early history of macrophages in intact embryos prior to that time. In a survey of peroxidase labeled lectins showing affinity for cells of developing rat lungs, Honda et al. (1989) found that the agglutinin I-B, of Griffoniasimplicifolia (GSA I-B,), which recognizes terminal alpha-galactose, binds to macrophagelike cells with sufficient selectivity to qualify as a potential marker for the line. They applied it to paraffin sections of 14-21 day prenatal, early postnatal, and adult lungs, and to ultrastructural specimens from a 1-day postnatal and a n adult lung, finding GSA I-B,labeled mononuclear, “monocytoid,” or macrophage-
Received June 19, 1991; accepted October 2, 1991. Address reprint requests to Professor Sergei P. Sorokin, Pulmonary Cell Biology, Department of Anatomy and Neurobiology, Boston University School of Medicine, 80 East Concord Street, Boston, MA 02118.
521
LECTIN MARKER FOR DEVELOPING MACROPHAGES
like cells present at all developmental stages, and indicating semi-quantitatively that free lectin-binding cells become increasingly prevalent towards birth. As will be shown here, the lectin can indeed be used to follow expansion of macrophage populations during pulmonary development, provided a number of important constraints are observed. The same approach and methodology may well have equal application to other systems where macrophages figure importantly, beginning with hematopoiesis in the yolk sac (Takahashi e t al., 1989) and extending to the reactions of microglia in the adult brain, where GSA I-B, has been successfully used for some time as a marker for this macrophage species (Streit and Kreutzberg, 1987). MATERIALS AND METHODS
Various tissues but principally the lungs from embryonic rats aged 13 or 14 days of gestation provided specimens for study. Fourteen-day lungs were also grown as organ cultures and harvested a t intervals ranging from a n hour to several days thereafter. In addition, smears and buffy coat preparations were made from the blood of adult rats. Fetal lung cultures were either of entire right or left lungs, or the pair from one animal. They were grown in T-flasks at 375°C on the surface of growth medium containing 40% fetal bovine serum, Gey's balanced salt solution, and 0.3% glucose, solidified by the inclusion of 1.5% agar, according to our routine practice (Sorokin and Hoyt, 1987). These cultures were extensively sampled; those furnishing illustrations for this paper were taken after 2 and 4 days in vitro. Most tissues and organ cultures were intended for labeling with peroxidase-coupled GSA I-B, (Sigma Chemical Co., St Louis, MO). These were fixed with 1% glutaraldehyde in 0.1 M phosphate-buffered saline (PBS, pH 7.2, 4"C, 2 hr), and rinsed in two changes of buffer (4"C, overnight). While immersed in buffer, specimens were lightly pricked with a pin or cut into smaller pieces to improve access of the lectin to cells within. They were then incubated in horseradish peroxidase (HRP)-labeled GSA I-B, (1.0 Fgm/ml) made up in buffer, pH 7.2, consisting of 0.1 M PBS with 0.1 mM MnCl,, MgCl,, and CaCl, (4"C, 24 hr). Following another rinse in PBS (1 hr) and a succeeding one in 0.05 M Tris-HC1, pH 7.6 (10-15 min), sites binding the lectin were revealed by incubation after Graham and Karnovsky (1966) in a diaminobenzidine substrate (DAB, 100 mg per 200 ml 0.05 M Tris-HC1, 10 min), followed by one containing both DAB and hydrogen peroxide (1 ml 1% H,O, per 100 ml Tris-DAB solution, 10 min). Tissues were then rinsed once in Tris-HC1 and once in PBS before being postfixed in PBS-buffered 1%osmium tetroxide (4"C, 1 hr), dehydrated in graded ethanols, and embedded in Epon. Specimens were examined as whole mounts, as semi-thin (0.5 wm) sections counterstained with toluidine blue, and as thin sections for electron microscopy, both unstained as well a s after staining with uranyl acetate and lead citrate. Adult rat blood was mixed with 5 units/ml of heparin in 0.1 M phosphate-buffered saline and spun in a clinical centrifuge. The buffy coat was then either siphoned off and fixed in 1% glutaraldehyde in 0.1 M phosphatebuffered saline (PBS, pH 7.2,4"C, 2 hr), or fixed in situ using 1.5% glutaraldehyde with 1%sucrose in 0.1 M
cacodylate-HC1 buffer (pH 7.4, 4"C, 3 hr) to obtain a buffy coat disc for subsequent incubation, processing, and sectioning (Anderson, 1965). Following fixation, suspended leukocytes were spun down to form a pellet and then incubated for GSA I-B, according to the preceding protocol, suspending the cells and repelleting at each step. Cells in the buffy coat disc were processed like solid tissue. Specificity of the peroxidase-coupled GSA I-B, reaction was ascertained by comparison of specimens incubated in the lectin with those incubated in control media. 1)The principal control depended on competitive inhibition of GSA I-B, binding to membrane sites in the presence of 0.1 M alpha-D-galactose, added to the normal lectin staining solution. Other control procedures included 2) incubation of specimens in diaminobenzidine (DAB)/H,O, without prior exposure to the lectin, and 3) reincubation of specimens previously exposed to lectin/galactose and DAB/H,O, in fresh lectin medium without galactose, followed by DAB/H,O, (a "rerun"). No effort was made to suppress the intrinsic peroxidase activity of granules in certain leukocytes, since this criterion could be helpful in discriminating among cell types present. All these control incubations were carried out on 4-day lung cultures, when macrophages were sure to be abundant. Controls 1 and 2 were also applied to buffy coat specimens. Two live 4-day lung cultures having macrophages on the pleural surface as well as within were exposed (22"C, 2 hr) to 10 Fg/ml fluorescein isothiocyanate (FITCI-coupled GSA I-B, (Sigma Chemical Co., St. Louis, MO) in 0.1 M PBS, pH 7.2, containing 0.1 mM MnCl,, MgCl,, and CaCl,, introduced dropwise onto the surface of the medium in the T-flasks. The cultures were thereafter rinsed in PBS and examined under a n ultraviolet microscope equipped with a filter block appropriate for the wave length of fluorescein (B filter, cat number 78946, Nikon, New York, NY). A few airdried smears of adult rat blood were similarly treated. The FITC-coupled reaction was controlled by preincubating a smear in a 1%solution of periodic acid (22"C, 10 min) before immersion in the lectin substrate, in order to destroy potential binding sites. RESULTS
Staining by GSA I-B, was examined by light and electron microscopy after applying the FITC and/or peroxidase coupled lectin to 1) well-developed macrophages in the lung cultures, as well as to 2) peripheral blood smears and buffy coat pellets of adult rat blood; 3) semi-thin and thin sections of 14-day prenatal rat lung; and 4) semi-thin and thin sections of connective tissue from 13- and 14-day rat embryos, In order to confirm our identifications of specific cell types in lectinstained specimens, we also examined Wright's-stained smears of adult blood and Giemsa-stained serial glycol methacrylate sections of whole 14-day rat embryos prepared earlier (Sorokin and Hoyt, 1987). Comparison of Test and Control Preparations
1) Pretreatment of blood smears with periodic acid abolished all selective staining by FITC-coupled GSA I-B,. 2) Control incubations in 14 + 4 day lung cultures satisfied u s that observed peroxidase-coupled GSA I-B, staining of macrophage cell membranes was
522
S.P. SOROKIN AND R.F. HOYT, JR.
attributable to lectin binding of terminal galactose units on membrane components (Fig. 1). In electron micrographs of macrophages otherwise unstained by uranyl acetate or lead citrate, the GSA I-B,-peroxidase reaction outlined each cell completely and continued along membrane invaginations but did not appreciably involve intracytoplasmic membranes (Fig. la). Addition of 0.1 M alpha-galactose to the incubation medium completely prevented binding of lectin-peroxidase to these cells (Fig. lb). Exposure to DAB/H,O, alone resulted in general staining of the cytoplasm of free erythroblasts or of those ingested by the macrophages, but cell membranes remained unstained (Fig. lc). In thin sections of buffy coat leukocytes, a similar control left intracytoplasmic granules stained owing to their content of intrinsic peroxidases, but cell membranes were unstained. Such reactive granules occurred principally among granulocytes or their precursors and much less frequently among cells judged “mononuclear,” where granules were smaller in size and few in number. Occasionally a cell exposed only to DAB/H,02 exhibited peroxidase staining along a short segment of cell membrane, and among sectioned tissues incubated in the GSA test medium, staining of macrophages sometimes became transferred to the plasmalemma of a n adjacent fibroblast around points of mutual contact, leaving the rest of the fibroblast’s cell membrane unstained (see Fig. 4).These observations led to our adoption of the criterion that a cell must display peroxidase staining along virtually the entire circumference to be considered marked by GSA lectin binding. A final control, the rerunning of galactose-inhibited (unstained) specimens in fresh GSA I-B, medium without the hexose, resulted in positive staining of macrophage cell membranes (Fig. Id).
similar way, where cultures were surrounded by a tightly packed corona of macrophages, only those in the peripheral layer were stained, whereas if the cells were more loosely packed, all stained. Thus we attributed occasional absence of staining to inadequate penetration of reagents rather than to the presence of a subpopulation of GSA-negative macrophages. In practice, better staining indeed occurred in specimens that were extensively pricked prior to incubation. GSA 1-6, Binding to Circulating Leukocytes in Adult Rats
As examined by light microscopy of air-dried smears of fresh whole blood, FITC-coupled GSA I-B, binds most strongly to the surface and cytoplasm of large leukocytes with a notched, and often more than simply lobated nuclear profile. After comparing lectin-stained with Wright’s-stained smears of the same blood, we identified most of the positive cells as heterophils and some as monocytes. Red cells were unstained as were many of the small circulating lymphocytes. Electron micrographs of the buffy coat incubated with peroxidase-coupled GSA I-B, indicated that binding was present on cell membranes of granulocytes and on the surfaces of platelets. Among heterophils intrinsic peroxidase was localized to numerous intracytoplasmic granules; in the smaller population of eosinophils intrinsic peroxidase seemed localized primarily to the cortical portion of the large, crystalloid-bearing specific granules. We found no basophils, but our samples included several large (- 20 pm) mast cells with large granules and imperfectly preserved, apparently GSAnegative cell membranes. Among mononuclear cells, typical monocytes stained and many smaller cells resembling lymphocytes did not. More objectively, about three quarters of the larger cells having notable development of granular endoplasmic reticulum bound the Staining of the Pulmonary Macrophage Population lectin, and more than half of the smaller cells having Electron microscopy of lung explants grown two or a n undifferentiated cytoplasm and many polysomes did more days in vitro reveals that virtually all GSA-pos- not. itive cells are morphologically recognizable a s macrophages (Fig. l),whereas light microscopic examination GSA-Reactive Cells in Embryonic Tissues of whole mounts gives one a n overview of the entire Comparatively few cells bound GSA I-B, in 13- or population. When a series of lung cultures harvested periodically between 2-7 days was incubated with GSA 14-day embryonic or extraembryonic tissues, and they I-B, and examined intact, lectin-binding cells within appeared to represent a single population. This was were observed progressively to increase in number made up of rounded cells resembling undifferentiated with time and to change their distribution inside the leukocytes and solitary stellate or angular cells that culture, behaving exactly like cells we had studied ear- except for their GSA staining resembled mesenchymal lier by other techniques and determined to be macro- fibroblasts. Only a small number contained phagocytic phages (Sorokin and Hoyt, 1987; Sorokin et al., 1989b). inclusions. Megakaryocytes as a rule were GSA negaFor example, at 2 days after explantation, numerous tive prior to platelet formation, and red cell memGSA-positive cells were recognized within a cultured branes were almost never stained. lung pair (Fig. 2, peroxidase-coupled lectin), but the The leukocyte-like GSA-positive cells were agranunumber increased markedly by 4 days when many cells lar and occurred both within blood vessels and outside, had emerged onto the pleural surface (Fig. 3, FITC- typically in mesenchyme where the majority of GSAcoupled lectin). GSA I-B, therefore appears to bind to positive cells assumed the angular form. Within this the great majority of macrophages present. tissue it was easy to find some GSA-positive cells in In cases where recognizable macrophages in a thin mitosis (Fig. 4). Because these cells appear to transsection of lung culture were not all stained, the defi- form into typical macrophages after embryonic rudiciency was usually apparent among cells most removed ments are placed in organ culture, we consider them from reagents in the incubation medium. Conversely, early representatives of the macrophage line. A fuller macrophages lying next to a pin prick through the account of their early history is given in the second pleura were often more strongly outlined by GSA-per- paper of this series, and changes after organ culturing oxidase than macrophages at some distance away. In a are presented in the third (Sorokin et al., 1991a,b).
LECTIN MARKER FOR DEVELOPING MACROPHAGES
Fig. 1. GSA I-B,-peroxidase reaction and controls, shown in electron micrographs of macrophages from 4-day (14 4) lung cultures. a: Standard incubation medium. The lectin binds to the cell membrane. Preparation was not counterstained with uranyl acetate/lead citrate (UNPb). x 6300. b Standard medium with 0.1 M alpha-D-galactose. Lectin binding is completely inhibited, although red cell fragments (top) exhibit a cytoplasmic peroxidase-like reaction attributable to
+
523
hemoglobin. Counterstained with UA/Pb. x 4800. c: Incubation in DAB/H,O, alone. A peroxidase-like reaction is present only in red cells, whether free or ingested by the macrophages. UA/Pb. x 2500. d Rerun of galactose-inhibited incubation in fresh GSA medium without galactose. The macrophage cell membrane now binds the lectin. UA/Pb. x 6700.
S.P. SOROKIN AND R.F. HOYT, J R .
524
Fig. 2. GSA I-B,-peroxidase reaction in a whole mount of 14-day fetal rat lungs after 2 days in culture (14 + 2). Numerous macrophages (dark spots) stand out against the left and right bronchial axes (pale areas). Hemoglobin of fetal erythroblasts in engorged blood vessels beside the bronchi exhibits peroxidase-like activity. x 60.
Fig. 3.Living culture of fetal rat lung at 14 + 4 days, after a 2-hr exposure to FITC-coupled GSA I-B,. Macrophages bind the lectin and so are visible under ultraviolet light. The plane of focus shows them emerging onto the pleural surface, and many more cells are present than after 2 days in vitro (see Fig. 2). x 60.
DISCUSSION
were in general GSA-negative until cells began to undergo fragmentation into platelets, which also bound the lectin, as noted in buffy coat preparations. Membranes of nucleated or enucleate erythrocytes almost never became stained, except among cells in certain organ cultures where tissue disruption and cell death were plainly evident. Indeed, erythrocytes are not considered to interact with ligands recognizing galactosyl residues until the cells become aged or damaged and their surface coating undergoes partial desialation to unmask them (Schlepper-Schafer and Kolb-Bachofen, 1988). Our study indicates that galactosyl residues are present on macrophage cell membranes, and the cells therefore can be tagged by GSA I-B, whenever morphologically recognizable as macrophages, whether they occur in prenatal or postnatal tissues, or in lung cultures a few days old. It is equally evident that isolated cells having either a leukocyte-like or mesenchyme-like appearance bind GSA I-B, in embryos; the succeeding two papers in this series provide evidence to show that these are precursors of macrophages. In sum, our experience is that peroxidase-coupled GSA I-B, can be relied on to mark macrophages and their precursors, particularly when used at an electron microscopic level where ultrastructural characteristics as well as the ac-
The appropriateness of relying on GSA I-B, as a marker for cells in the lineage of pulmonary macrophages was evaluated in a series of steps, initially by finding that typical macrophages of lung cultures bound this lectin on their cell membranes and that the staining reactions were well controlled and specific. Second, we sought to determine whether all or only some of the macrophages are stainable and found that virtually the whole population can be tagged, provided the cells have adequate access to the reagents. Thereafter, guided by an assumption that macrophages have hematopoietic forebears, we wished to learn if other kinds of mature leukocytes also bind the lectin and what the binding capacity of the leukocytes in the embryo might be, both circulating and in formation, a t the moment the lung rudiments were removed to culture. The results indicate that for rats, peroxidase-coupled GSA I-B, stains immature mononuclear cells in embryos. In adults certain mononuclear cells bind the lectin, specifically monocytes and perhaps a number of immature related forms present in the circulation, but typical lymphocytes more often than not are unreactive. Mature and immature granulocytes of heterophil and eosinophil lines also bind the lectin. In the megakaryocytic line, those encountered in embryonic tissues
LECTIN MARKER FOR DEVELOPING MACROPHAGES
Fig. 4. Dividing GSA-positive macrophage precursor in the mesenchyme of a 13-day rat embryo. Surrounding mesenchymal fibroblasts are unreactive, but one can see that where macrophage and fibroblast membranes are apposed, some transfer of staining has occurred. In
tivity of intrinsic peroxidases in lysosomes andlor specific granules would prevent one from mistaking granulocytes for macrophages. The same experience would nevertheless make us cautious about depending on it exclusively to identify macrophages in paraffin sections of late fetal or postnatal material, because the more gestation is prolonged beyond the stage represented by the 1Cday prenatal rat, the more severe will be contamination of the growing population of macrophages by equally GSA-positive granulocytes, and cells in both categories would be stained dark brown. In such specimens, therefore, it is a difficult task to obtain a reliable estimate of macrophage population growth based solely on light microscopic counts of GSA I-B,positive cells. In contrast, organ cultured prenatal lungs are cut off from further influx of cells from the developing embryo. Here, where circumstances favor generation of a pure population of macrophages unmixed with other leukocytes, the lectin marker can be very useful at the light microsopic as well as ultrastructural level for studies assessing population growth and related macrophage dynamics.
525
contrast, fibroblast surfaces not apposed to the dividing cell are not stained. This common artifact of GSA I-B, staining poses little diffculty for identifying lectin-positive cells. x 12,000.
ACKNOWLEDGMENTS
The participation of Dana G. Blunt and Nancy A. McNelly is gratefully acknowledged. The work was supported by USPH research grant HL-33070. LITERATURE CITED Anderson, D.R. 1965 A method of preparing peripheral leucocytes for electron microscopy. J. Ultrastr. Res., 13:263-268. Graham, R.C., and M.J. Karnovsky 1966 The early stages of absorption of injected horseradish peroxidase in the proximal tubules of mouse kidney: ultrastructural cytochemistry by a new technique. J . Histochem. Cytochem., I4r291-302. Honda, T., B.A. Schulte, and S.S. Spicer 1989 Glycocongugate with terminal galactose. A selective property of macrophages in developing rat lung. Histochemistry, 91r61-67. Schlepper-Schafer, J., and V. Kolb-Bachofen 1988 Red cell aging results in a change of cell surface carbohydrate epitopes allowing for recognition by galactose-specific receptors of rat liver macrophages. Blood Cells, 14r259-269. Sorokin, S.P., and R.F. Hoyt, J r . 1987 Pure population of nonmonocyte derived macrophages arising in organ cultures of embryonic rat lungs. Anat. Rec., 21 7:35-52. Sorokin, S.P., R.F. Hoyt, Jr., and M.M. Grant 1984 Development of macrophages in the lungs of fetal rabbits, rats, and hamsters. Anat. Rec., 208r103-121.
526
S.P. SOROKIN AND R.F. HOYT, J R .
Sorokin, S.P., R.F. Hoyt, Jr., and N.A. McNelly 1989a Nonimmunemediated phagocytosis by “premedullary” lung macrophages: Effects of concanavalin A, tuftsin, and macrophage-inhibitory peptide. Anat. Rec., 223t55-61. Sorokin, S.P.,L. Kobzik, R.F. Hoyt, Jr., and J. J. Godleski 1989b Development of surface membrane characteristics of “premedullary” macrophages in organ cultures of embryonic rat and hamster lungs. J. Histochem. Cytochem., 37t365-376. Sorokin, S.P., R.F. Hoyt, Jr., D.G. Blunt, and N.A. McNelly 1992a Macrophage development. 11. Early ontogeny of macrophage populations in brain, liver, and lungs of rat embryos as revealed by a lectin marker. Anat. Rec., 232:527-550.
Sorokin, S.P., N.A. McNelly, D.G. Blunt, and R.F. Hoyt, J r . 199213 Macrophage development. 111. Transformation of pulmonary macrophages from precursors in fetal lungs and their later maturation in organ culture. Anat. Rec. 232551-571. Streit, W.J., and G.W. Kreutzberg 1987 Lectin binding by resting and reactive microglia. J. Neurocytol., 16:249-260. Takahashi, K., F. Yamamura, and M. Naito 1989 Differentiation, maturation, and proliferation of macrophages in the mouse yolk sac: A light-microscopic, enzyme-cytochemical, immunohistochemical, and ultrastructural study. J. Leuk. Biol., 45:87-96.
