THE ANATOMICAL RECORD 208:103-121(1984) Development of Macrophages in the Lungs of Fetal Rabbits, Rats, and Hamsters SERGE1 P. SOROKIN, RICHARD F. HOYT, JR.,AND MARGARET M. GRANT Devartment of Anatomy and Houseman Research Center (S.PS., R.RH.I, a i d Department of Biochemistr-y (M.M. G.), Boston University School of Medicine, Boston, MA 02118 ABSTRACT Fetal rabbits (days 13-32), rats (days 14-22), and hamsters (days 11-15) and selected postnatal animals were examined for pulmonary macrophages or their precursors in 2-,um sections stained by PAS-lead hematoxylin (all species), electron micrographs (rabbit and rat), and cytochemical incubations for acid phosphatase (rabbit and rat), aliesterase, and N-acetyl glucosaminidase (rabbits). All methods revealed macrophages in perinatal specimens. The appearance and distribution of these cells were compared in the different preparations to establish the reliability of PAS-lead hematoxylin for identifying them in less developed fetal lungs, where they are less active for lysosomal enzymes the earlier the stage examined. In the sections, macrophages are seen to possess a round or indented nucleus, a n irregular contour, and a deep purplish-gray cytoplasm containing a variety of pink PAS-stained granules, equated with heterolysosomes by ultrastructural cytochemistry. In less developed lungs, macrophages occur along with putative precursors having a more rounded outline and fewer PAS-stained granules. In pseudoglandular lungs these precursors predominate over rather vacuolated macrophages resembling Hofbauer cells. In all three species both cell types first appear in the stroma during the bronchial bud stage and are frequently seen to divide from that time on. The earliest precursors have a relatively sparse cytoplasm which later increases in daughter cells. Hofbauer-like cells disappear during the canalicular stage of development, replaced by macrophages and transitional forms from the more rounded precursors. In day 21 rabbit lungs, scattered stromal cells are reactive for aliesterase, and, some days later, for acid phosphatase and glucosaminidase. Free mononuclear cells are rare in airways of pseudoglandular lungs but become common later. A day or two before birth in rats, free cells range between rather undifferentiated leukocytes to typical macrophages, but cells with the macrophage’s complete repertory of inclusions are seen only after birth. In the fetus, typical monocytes were not identified in either the pulmonary stroma or the airways. A replicating population of macrophage-like cells therefore resides in fetal lungs. It is established before bone marrow is formed and, in rats, before monocytes have appeared in the circulation. It has proven difficult to give a simple answer to the question, “Where do pulmonary alveolar macrophages come from?” because the topic is complex. Without doubt these cells resemble the amoeboid cells called macrophages that generally are derived from embryonic or extraembryonic mesoderm, scavenge the tissues of nearly all Metozoans, 0 1984 ALAN R. LISS,INC. and usually function in association with the hematopoietic system, when phylogenetically this is present (Metchnikoff, 1893). In the course of evolution this ancient cell type has become adapted to work in concert with developing serum-borne immunological deReceived March 24,1983; accepted June 6, 1983. 104 S.P. SOROKIN, R.F. HOYT, JR.,AND M.M. GRANT fense mechanisms, but it retains its character as a relatively unspecialized cell with some capacity for further transformation (Sorokin, 1977). In the lungs of healthy mammals, the macrophage population is distinctly heterogeneous in the appearance and functional activity of individual cells. If all these cells still comfortably fit within a broad definition of “macrophage” it is possible that at any moment important components of the population may have had different past histories, some for example having become transformed from newly arrived monocytes (Bluss6 and van Furth, 1979)and others having been resident for some time in the pulmonary interstitium before migrating onto the alveolar surface (Bowden and Adamson, 1972). Furthermore, lavaged alveolar cells are able to form colonies when cultured in semisolid, conditioned medium, and this capacity is enhanced if the cells are obtained from lungs subjected to mild ozone injury (Boorman et al., 1979a,b). Granting that these rather undifferentiated cells are able to divide and yet are often undisposed to do so, one might well understand why a comprehensive answer to the question of pulmonary macrophage origin has eluded investigators despite considerable effort to address it (Brain et al., 1977). This study was undertaken to provide information on the development of macrophages in fetal lungs of three mammalian species. Certain morphological and cytochemical observations are presented in a n attempt 1)to identify macrophage precursors and macrophages in fetal lungs; 2) to note the time of their first appearance in the lungs and the location(s) in which they are first found 3) to trace the subsequent appearance of these cells in other pulmonary compartments, such as the airway and respiratory surfaces; 4) to look for evidence of macrophage replication andor maturation in situ and to place these events in the lungs a t various gestational ages. Finally, using our own data and those from the literature, we comment briefly on the timing of some of the preceding events in comparison with timings of the first appearance and subsequent maturation of macrophages or their allies in yolk sac, liver, and bone marrow of the same species. MATERIALS AND METHODS New Zealand White Rabbits (Associated Rabbit Industries, Easton, MA) Light microscopy Littermates of fetal and newly born rabbits provided a series of developing lungs a t the following gestational ages: 13, 15, 17, 19, 21, 23,25,26,27,29,31, and 32 days (birth); and 12 hr, 1,and 2 days postnatal. Prenatal stages were obtained by caesarean section. Specimens taken a t 13 and 15 days were whole fetuses; those obtained a t 17-21 days included both neck and thorax segments and dissected lung pairs; and older stages were represented by lungs only. Tissues were fixed in 2% wlv formaldehyde (freshly prepared from paraformaldehyde)-2% glutaraldehyde buffered with 0.1 M cacodylate (pH 7.2) for 24 hr, after which they were rinsed in buffer, dehydrated in graded ethanols, embedded in glycol methacrylate, and sectioned a t 2 pm. Staining was by PAS-lead hematoxylin (Sorokin and Hoyt, 1978) and Giemsa. Electron microscopy Electron micrographs were made from blocks of fetal lungs aged 25,27, and 30 days and from macrophages lavaged from lungs of adult animals. Lung blocks were fixed by immersion for 3-12 hr in the aldehyde mixture just described, then rinsed in 0.1 M cacodylate buffer, postfixed in 1%w/v osmium tetroxide with 15 mglml potassium ferrocyanide for 2 hr a t 4°C (Karnovsky, 1971), dehydrated, and embedded in Epon. Grids were stained with uranyl acetate and lead citrate. Pulmonary macrophages from adult rabbits were obtained by multiple lung lavage (10 washes of 50 ml each using Dulbecco’s calcium- and magnesium-free balanced salt solution prewarmed to 37“C), as described by Grant et al. (1979). Cells were centrifuged, and pellets were prepared for electron microscopy as described for lung blocks. Lysosomal enzyme histochemistry Neck and thorax segments were obtained from all previously listed developmental stages older than the 19th fetal day. These were fixed briefly in 4% wlv formaldehyde and frozen on dry ice. Cryostat sections cut a t 10 pm were then reacted either for aliesterase, using the method of Pearse (19721, in which alpha-naphthyl acetate acts a s substrate and fast blue RR as the coupling agent, or for acid phosphatase, using the method of Burstone as modified by Barka and Anderson (1962), in which AS-TR phosphate is the substrate and hexazotized pararosaniline is the coupling agent. Incubations for esterase were carried out a t room temperature (20°C) for periods from 5-30 min, and those for acid phosphatase at pH 5.0 and room temperature for periods from 20-90 min. DEVELOPMENT OF PULMONARY MACROPHAGES Frozen sections cut from additional specimens fixed in 4% formaldehyde containing 1.5%w/v calcium chloride were reacted for Nacetyl-beta-glucosaminidase (Reed and Ben nett, 1975)using a-naphthol AS-BI N-acetyl 0glucosaminide as substrate and fast garnet GBC as coupling agent. Incubations were for 20 min a t room temperature. Sections were counterstained with 0.1% aqueous methyl green, air dried, and studied uncovered or with immersion oil mounted coverslips. Photographs were taken within a day of incubation. For each histochemical reaction, control incubations lacked substrate and were negative in all cases. CD-1 Rats (Charles River Laboratories, Wilmington, MA) Light microscopy Litters were delivered by caesarean section from timed pregnant mothers a t 14, 15, 16, 17, 18, 19, 20, 21, and 22 days (term) of gestation; additional pups were killed 3 and 4 days postnatally. Fourteen-day specimens were whole fetuses; dissected lung pairs were obtained at other stages. Tissues were fixed in 0.1 M phosphate buffered 2% formaldehyde-2% glutaraldehyde (pH 7.2-7.3) containing 0.01% picric acid (Ito and Karnovsky, 1968), embedded in glycol methacrylate, and sectioned at 2 pm. Staining was by PAS-lead hematoxylin and 1% aqueous toluidine blue. Electron microscopy Blocks of 21-day fetal and 3- and 4-day postnatal lungs were fixed in phosphate buffered 4% formaldehyde-1%glutaraldehyde (pH 7.2) having a final buffer osmolarity of 176 mOsm (McDowell and Trump, 1976) and postfixed in osmium-ferrocyanide before being embedded in Epon. Grids were stained with uranyl acetate and lead citrate. Ultrastructural acid phosphatase cytochemistry Pieces of minced 22-day fetal lung (0.5 mm3) were fixed for 90 min a t 21°C in Karnovsky’s (1965) glutaraldehyde-formaldehyde solution (ph 7.2), washed in 0.02 M Trismaleate buffer (pH 5.0), and incubated a6 that pH for 1.5 h r in a medium containing 0.08% wlv lead nitrate and 2.5 mg/ml 0-glycerophosphate (Novikoff et al., 1971).After washing in 0.1 M cacodylate buffer, tissues were postfixed with osmium-ferrocyanide and embedded in Epon. Grids were stained with uranyl acetate and lead citrate. 105 Syrian Golden Hamsters (Charles River Laboratories, Wilmington, MA) Litters were delivered by caesarean section from timed pregnant females on days 11, 12, 13, 14, and 15 of gestation (term 16-17 days). Whole fetuses were taken on day 11, head and thorax on day 12, and lung pairs on days 13-15. Tissues were fixed in buffered formaldehyde-glutaraldehyde as specified either by Karnovsky (1965) or McDowell and Trump (1976). They were embedded in glycol methacrylate, cut at 2 pm, and stained by PASlead hematoxylin. RESULTS Illustrations for this paper are grouped by species and in order of increasing gestational age, as follows: Rabbit, figures 1-11, light microscopy; Figures 20-25, electron microscopy. Rat, Figures 12-17, light microscopy; Figures 26-30 electron microscopy. Hamster, Figures 18 and 19, light microscopy. (Figures 5-19 are color illustrations that appear in the Color Figure Section elsewhere in this issue. Figures 5-11 appear on pp. 8485; Figures 12-19 appear on pp. 86-87.) Criteria Used to Identify Macrophages or Macrophage Precursors in the Sections Histochemistry Mononuclear leukocytes can be identified in the connective tissue stroma of pseudoglandular rabbit lungs after a rather long incubation for aliesterase (30 min, compared to 1-5 min in adult lungs). The esterase-positive cells are few in number and stand out in 21-day fetal lungs because few pulmonary tissues are then reactive (Fig. 1). They are not identifiable by acid phosphatase or Nacetyl glucosaminidase cytochemistry at this time, although these lysosomal enzyme markers are useful for identifying alveolar macrophages in late fetal, postnatal, and adult lungs. Acid phosphatase is first localizable to specific cells in fetal lungs at about the 26th day of gestation in rabbits, when only a few cells are active in the stroma and a n occasional, more reactive cell appears in the lumen of canalicular intrapulmonary airways (Fig. 2). Thereafter, mononuclear phagocytes can be identified in both the pulmonary stroma and the air spaces with increasing frequency .using cytochemical methods for esterase, acid phosphatase, or glucosaminidase and incubation times suitable for demonstrating alveolar macrophages in postnatal lungs. At 31 days of gestation, aliesterase positive cells are demonstrable in the lungs after 5 min- 106 S.P. SOROKIN, R.F. HOYT, JR.,AND M.M. GRANT Utes of incubation. These cells are particularly prone to aggregate in the loose connective tissue adventitia surrounding sizeable pulmonary arteries and veins (Fig. 3), but they also occur in the interstices between the terminal sacs and on the primitive respiratory surface. At one day after birth, such cells are found in increased number in the interstitium and on the surfaces of the primitive alveoli, as revealed by their reactivity either for esterase, acid phosphatase, or glucosaminidase (Fig. 4). able degree), a n indented nucleus, a cytoplasm replete with free ribosomes but poor in granular endoplasmic reticulum, and a small Golgi complex. Most of these cells contain a few dense bodies about 0.1 pm in diameter, but in some these range up to 0.6 pm and resemble heterolysosomes (Fig. 21). Others contain recognizably ingested material as well, but this is never as characteristic a feature of the cytoplasm of mononuclear phagocytes in fetal lungs as it becomes after birth. As shown in a macrophage from a 22day fetal rat, the dense bodies are reactive for acid phosphatase, which codirms their identification as lysosomes (Fig. 261.l Correspondence between acid phosphatase reactive cells of late fetal lungs and interstitial or free mononuclear cells stainable by PAS-lead hematoxylin Development of Macrophages in Fetal Rabbit Lungs (Gestation 32 Days) In 30-day fetal rabbit lungs, a n accumulation of reactive mononuclear cells is demonAs seen in glycol methacrylate sections of strable in the perivascular adventitia as well 13-day fetuses, a very few PAS-lead hematoxby incubations for acid phosphatase (Fig. 11) ylin-positive mononuclear cells occur in the a s for aliesterase (Fig. 3). In PAS-lead hematoxylin-stained glycol methacrylate sections of lungs of the same gestational age, the position of these esterase or acid phosphatase po'Without phagosomes or a complement of heterolysosomes in cytoplasm, putative macrophage precursors resemble other sitive cells is occupied by mononuclear cells their blast-like hematopoietic cells, including those functionally in a made conspicuous by the dark purplish-gray lymphocytic line of development. Therefore, based on techniques staining of a cytoplasm that contains a num- used in ths study, attribution of a cell like that in Figure 20 be made to the mononuclear phagocyte line with absolute ber of medium pink, PAS-positive granules cannot certainty. This same difficulty also hampers attempts to identify (Fig. 10). These do not contain glycogen but pluripotential stem cells of the hematopoletic system, as evifrom a paper by Dicke et al. (1973) and the discussion correspond to lysosomal granules seen in the denced following, because few functional markers are expressed by such cytochemical preparations. protoplasts which, according to Weiss (1983),structurally are lymphocytes. In the setting of the developing lung, however, and Free cells in the airways of late fetal lungs amidst a spectrum of morphologically related cells, some posare also reactive for acid phosphatase and sessing macrophage markers, it appears likely that the destiny accentuated by PAS-lead hematoxylin (Figs. of these undifferentiated cells is correctly stated. 2, 9). In this location their identification as macrophages is certifiable on morphological grounds alone: possession of a n often irregular outline, a rounded to indented nucleus, Fig. 1. Aliesterase in 21-day fetal rabbit lung, which one or more nucleoli, and PAS-positive gran- is minimally reactive after a long incubation (30 min, ules that are variable in number and size 20°C). Scattered mononuclear cells in the pulmonary stroma (arrows) and differentiating myoblasts beneath and become more so as term approaches. the bronchus (right center) are the main sites of activity. Once studied as they occur in the air spaces, The stroma has a distorted, netlike appearance in this the cells become easier to find in the pulmo- frozen section. x 100. nary interstitium (Fig. 7) or a s they emerge Fig. 2. Acid phosphatase in 26-day fetal rabbit lung. onto the surface (Fig. 8). Tissues are rather unreactive overall in both the epitheCorrespondence between the cells highlighted by PAS-lead hematoxylin and mononuclear phagocytes identified by electron microscopy As seen in electron micrographs of fetal rabbit lungs aged 25-30 days of gestation (Figs. 20-231, the cells corresponding in location and number to those made prominent in 25-30 day fetal lungs by PAS-lead hematoxylin are mononuclear leukocytes frequently of a n undifferentiated appearance, characterized by surface projections (present to a vari- lium (lower left) and stroma of this canalicular lung, except for an occasional free mononuclear phagocyte (upper right). (Incubation 45 min, 20°C). Frozen section; x290. Fig. 3. Aliesterase in 31-day fetal rabbit lung after a short incubation (5 min, 20°C). A large number of reactive mononuclear phagocytes are gathered in the adventitia outside the pulmonary artery, which runs alongside the bronchus (upper left.) Frozen section; X550. Fig. 4. N-acetyl glucosaminidase activity in newborn rabbit lung (20 min, 20°C). The most reactive cells are macrophages in the alveolar interstitium and on the surface (arrow). Esterase has a similar localization. Frozen section; ~ 4 0 0 . DEVELOPMENT OF PULMONARY MACROPHAGES 107 108 S.P. SOROKIN, R.F. HOYT, JR.,AND M.M. GRANT undifferentiated connective tissue surrounding the bronchial buds just a s they do in the connective tissue of the body wall. Such cells are more prevalent in the stroma of pseudoglandular lungs at 15 days. They are smaller than the macrophages of late fetal lungs and are more difficult to identify because they usually have few PAS-positive granules. A minority of these putative macrophage precursors are rather larger and more vacuolated than the remainder, and some of the vacuoles are faintly stainable by PAS. The latter cells resemble macrophage-like elements, termed Hofiauer cells, that occur in the placenta as well as in other fetal organs. The pulmonary stroma has a mesenchymatous appearance a t this time and for several days thereafter. It consists largely of stellate cells that extend throughout a matrix containing few extracellular solids and stain more lightly purplish than the macrophage precursors. The tissue is supplied by a loose plexus of sinusoidal capillaries. Examination of serial Giemsa-stained sections confirms that formation of blood cells is under way in the livers of 13- and 15-day specimens but that the bone marow is not yet established. In the liver the most prominent blood cells undergoing formation are megakaryocytes and those of the erythrocytic series. Some basiphil mononuclear cells of a decidedly undifferentiated appearance occur among them, but few precursors of granulocytes are recognizable. Mitotic figures are observed in blood cells located both in the hepatic cords and sinusoids and to a more limited extent in the lumens of distant blood vessels, including those of the forming lungs. In these preparations the basiphil mononuclear cells can also be observed sparsely inhabiting the pulmonary stroma. They are generally smaller than the mesodermal cells of the developing lungs and compare in size to the PAS-lead hematoxylinpositive cells, to which they evidently correspond. During the 17th, 19th, and 21st days of gestation, fetal rabbit lungs contain increased numbers of PAS-lead hematoxylinstained precursors of macrophages, and mitotic figures are not infrequently observed in them (Fig. 5). Like typical mononuclear cells found in the lungs a t 15 days, those present from the 17th to 21st days possess a high nucleuslcytoplasm ratio and comparatively few PAS-positive granules, although these are more conspicuous than at 15 days. Most of the cells remain well away from the developing bronchial tree within the pulmonary stroma. Nonetheless, a n occasional mononuclear cell may occur in the lumens of both large and small airways a t any time during this interval. Such cells have little activity for acid phosphatase or glucosaminidase and share morphological characteristics with the macrophage precursors located in the stroma (Fig. 6). Subsequent to 25 days, macrophage-like cells become more frequently observable along the internal surface of the lungs, by then a t a canalicular stage of development. Most of these are better provided with PASpositive granules than their counterparts in the stroma, many of which are stainable only about the centrosome (Figs. 7,8). This difference is in keeping with the greater lysosomal activity surface cells tend to exhibit, compared with those in the tissues. Once gained in later fetal life, the preeminence of free macrophages in granule content and lysosoma1 enzyme activity is maintained thereafter. Development of Macrophages in Lungs of Fetal Rats (Gestation 22 Days) and Hamsters (Gestation 16 Days) The pattern of macrophage development in fetal lungs of rats and hamsters is similar to that just described in rabbits. Rats A very few PAS-positive mononuclear cells are observable in the stroma of the rat lung a t 14 days of gestation, when the main bronchus and buds for the lobar bronchi have appeared on each side and pulmonary development is a t a late lung bud stage (Fig. 12). These cells are seen more frequently in subsequent days. Their increase results in part from mitotic activity in situ, as shown in pseudoglandular lungs at 16 days (Fig. 13)as well as during the terminal sac stage at 21 days (Fig. 27). As illustrated in the field from 16 day lungs, the fetal macrophage population includes vacuolated cells of the Hofbauer type (Fig. 13); nonetheless, the impression gained from examination of our whole fetal series is that the stromal macrophage population of rat lungs consists mainly of cells with comparatively few PAS-positive granules, still fewer than in typical stromal macrophages of fetal rabbits. Undifferentiated mononuclear cells are first seen in very small number on the airway surface a t 16 DEVELOPMENT OF PULMONARY MACROPHAGES 109 Ultrastructural Appearance of Maturing Macrophages in Fetal and Postnatal Lungs Rabbits As seen in 25-day fetal lungs, the greater number of mononuclear leukocytes occurring in the pulmonary stroma outside of the blood vessels are rounded cells with an undifferentiated appearance as already described (Fig. 20). A smaller number of this population are polymorphous (Fig. 21). These cells usually contain a variety of cytoplasmic inclusions, including heterolysosomes of different sizes and a few pinosomes or phagosomes. They also extend protoplasmic leaflets into the surrounding matrix. In contrast, the rounded cells rarely contain more $han a few dense bodies in the 1000-1500 A size range. The rounded and polymorphous cells resemble each other in the appearance of the nucleus, which is indented and slightly heterochromatic, and both stand out from their connective tissue neighbors by their greater overall electron density. Neither cell type forms junctions with any of the neighboring cells. At 27 days of gestation, many of the presumptive macrophage precursors and macrophages are identical in appearance with those seen at 25 days, but among this group an increase in the amount of ingested material is seen in some, while in others the granular endoplasmic reticulum has undergone expansion (Fig. 22). At 30 days, apart from the preceding, additional cells appear whose relationship to mature rabbit alveolar macrophages is unmistakable (Fig. 23). In them the Golgi apparatus is enlarged and a number of electron dense rod- or dumbbell-shaped bodies have Hamsters begun to gather in the cytoplasm (Fig. 23). Macrophages and their precursors were These appear t o be nascent lysosomes, and it sought in fetal hamster lungs using only is evident that in adult rabbits a great numPAS-lead hematoxylin-stained preparations. ber of similar structures are combined to proIn hamsters these cells stain like their hom- duce a smaller number of large, rounded ologues in fetal rabbit and rat lungs. They heterolysosomes such as macrophages of are present in the mesodermal tissue sur- many species possess (Figs. 24, 25). rounding the bronchial buds at 11 days of gestation and become increasingly prevalent Rats In macrophages present in the stroma of in 12-day and 13-daypseudoglandular lungs, when they occur throughout the stroma just prenatal rat lungs, lysosomal enzyme prooutside the pulmonary blood vessels as well duction is at a relatively modest level, and as near the terminal epithelial buds (Fig. 18). acid phosphatase activity is mainly confined By 15 days the rapidly forming hamster to a few large lysosomes (Fig. 26). Among lungs have begun to construct terminal sacs, cells found undergoing mitosis, a few share and macrophages with PAS-positive gran- enough characteristics with other macroules are then frequently seen in the air spaces phages t o make their identification fairly certain (Fig. 27). These cells have only a mod(Fig. 19). days. As fetal development proceeds past the pseudoglandular stage, the mononuclear cells increasingly gain the appearance of typical macrophages and increasingly appear in the air spaces of the conducting airways or terminal sacs (Figs. 14, 15). Macrophages versus mast cells in fetal rat lungs. In light microscopic preparations, the macrophage precursors of older fetal lungs have to be differentiated from emergent mast cells, which also occur in the stroma at that time. These are rounded cells that also contain PAS-positive granules, only the granules tend to develop more nearly together and to reach a more uniform size (ca. 0.5-0.8 pm) than the granules of macrophages. The two ceII types can be distinguished easily if adjacent sections are stained by toluidine blue, thereby revealing the metachromasia of the PAS-positivegranules in the mast cells (Figs. 16, 17) and its absence from the granules of the macrophages (Figs. 14,15).By this method the mast cells of fetal rat lungs are first discernible at 16 days as a very few sparsely granulated cells with relatively scanty cytoplasm located alongside the lobar pulmonary arteries. They become a little more abundant at 17 days, when they are found as well in the condensed connective tissue just beneath the endodermal epithelium and in the looser tissue farther away. The granules accumulate in the cytoplasm during subsequent days, and the cells achieve an adult pattern of distribution beneath the airways, blood vessels, and pleura by the 20th prenatal day, although the cells do not reach their full size until well after birth. Fig. 20. Putative macrophage precursor in connective tissue of 25-day fetal rabbit lung. The cell is rounded with inconspicuous cytoplasmic projections. The chromatin pattern of the nucleus is comparable to that of the macrophage in Figure 21. The cytoplasm has many ribosomes, a large number in polysomal configurations, and a small Golgi complex, located above the nuclear notch. A few small dense bodies occur along the margin on the left side of the cell. ~ 8 , 1 0 0 . Fig. 21. “Fetal” macrophage in a 25-day fetal rabbit lung. The cell shares many features with the macro- phage precursor of Figure 20, except that the cell outline is irregular and adorned with many lamellipodia (left side and lower margin), and the cytoplasm contains dense bodies in a larger range of sizes as well as clear vacuoles. The former are considered lysosomal equivalents of the PAS-stained granules seen in light microscopic sections; the latter are thought to be macropinosomes. The cell therefore is a macrophage, although not typical of those in postnatal lungs. It closely resembles Hofbauer cells of the term human placenta (Fig. 9 in Enders and King, 1970). ~8,100. Fig. 22. Developing macrophage in a 27-day fetal rabbit lung. The cell occupies a niche in the stroma framed by fibroblasts sporting an endoplasmic reticulum with dilated cisternae, whose activity in the synthesis of extracellular fibers nonetheless still largely lies ahead. In the macrophage the granular reticulum has undergone expansion compared to cells shown at 25 days (Figs. 20, 21). A few lysosomal granules are present above the nucleus next to the Golgi region and to each side; they are ovoid or dumbbell shaped. X 11,200. Fig. 23. Stromal macrophage in a 30-day fetal rabbit lung. The cell appears stimulated and resembles alveolar macrophages from adult animals more than the macrophage-like cells seen earlier in gestation, possessing a nucleus with numerous nuclear pores, a cytoplasm with moderate development of granular reticulum, an expanded Golgi complex associated with small electron dense vesicles (above the nucleus), and a number of rod-, club-, or dumbbell-shaped dense bodies, considered to represent nascent lysosomes in this species. X9,OOO. Figs. 24.25. Alveolar macrophages obtained by intrapulmonary lavage from adult rabbits. The upper cell (Fig. 24) has many small lysosomes, most of which are elongated, as well a s some larger ones, which are ring shaped or spheroidal. The lower cell (Fig. 25) has fewer lysosomes, but most of them are large and spheroidal heterolysosomes, and only a minority are slender and rodlike, as seen in Figures 23 and 24. Macrophages with this kind of lysosomal development usually are seen only in postnatal lungs, especially where there is a history of exposure to airborne contaminants (Pratt et al., 1971). ; 25, ~ 7 , 1 0 0 . Fig. 24, ~ 6 , 5 0 0Fig. 114 S.P. SOROKIN, R.F. HOYT, JR.,AND M.M. GRANT DEVELOPMENT OF PULMONARY MACROPHAGES erately expanded granular endoplasmic reticulum typically without dilated cisternae, a few small, dense lysosomal bodies, an electron dense cytoplasm containing numerous polysomes, and peripheral processes that occasionally contact but do not make junctions with neighboring cells (Fig. 27). These neighbors are mostly fibroblasts; at this time they are distinguishable from macrophages by cytoplasmic texture as well as by cell shape, because their endoplasmic reticulum is dilated by cell products of an intermediate density, and because the cells store glycogen and droplets of triglyceride. Free cells in terminal air sacs of late fetal lungs range in appearance from rather undifferentiated (Fig. 28) to rather macrophagelike (Fig. 29). In spite of its scanty cytoplasm, the undifferentiated cell can be considered a possible precursor of macrophages, because its endoplasmic reticulum is more extensive than that of lymphocytes, and because a number of elongated lysosomal bodies are present. Other free cells are intermediate in cytological characteristics between this cell and typical alveolar macrophages. As judged ultrastructurally, many of the latter contain a greater variety of phagolysosomes than their interstitial contemporaries and so are considered more active in phagocytosis (Fig. 29). By three days after birth, a large majority of the cells on the respiratory surface have begun to acquire the look of experienced alveolar macrophages, given them by a richly varied content of phagolysosomes and other inclusions (Fig. 30), and an even stronger impression of the same kind is made by examination of lungs a day older. In this respect the postnatal alveolar macrophage Fig. 26. Interstitial macrophage in a 22-day fetal rat lung. The cell is relatively immature in appearance and lies interposed between blood vessels (above and below the field shown) and near a great alveolar cell, one of whose lamellar bodies is shown (lower left). Acid phosphatase activity is present in some of the dense bodies in the macrophage cytoplasm. (Glycerophosphate, pH 5.0, 90 min); ~ 9 , 2 0 0 . Fig. 27. Dividing macrophage precursor or macrophage in the stroma of a 21-day fetal rat lung. The cell shares cytological features with immature rabbit and rat macrophages (Figs. 20, 21, 26); it is unattached to and differs in appearance from interphase and dividing fibroblasts, myoblasts, and endothelial cells. One pair of centrioles is shown below the metaphase figure, and clusters of small lysosomal bodies are distributed on either side. ~8,600. 115 population is easily distinguished from any obtainable from normal lungs before term in either rabbits or rats. DISCUSSION Using a variety of morphological and functional markers, we have shown that cells evidently in a macrophage line of development are present in fetal lungs of rabbits, rats, and hamsters for a considerable time before birth, although if one starts with newborn lungs and works back to earlier stages of development, the cells in question progressively lose their markers. They become poorer in lysosome-likegranules stainable by PAS, and in rabbits where this was investigated, progressively less active for lysosomal enzymes acid phosphatase and N-acetyl glucosaminidase. In this species, the cytochemical method for aliesterase is more useful than the preceding for identifying macrophagelike cells in early developmental stages because few other cells of the pulmonary stroma or epithelium are active for the enzyme during that period (Sorokin and Hoyt, 1982). As we did not intervene with the normally developing animals in any experimental way, we had to rely on our light and electron microscopic observations for evidence of endocytotic activity by the pulmonary macrophages or their precursors. This could be taken for granted in postnatal macrophages; it also appeared evident from examination of electron micrographs of late prenatal lungs of both rabbits and rats that the cells are active in phagocytosis, especially when they are on the respiratory surface. In this respect our findings are complementary to those of Zeligs et al. (19771, who examined lavaged cells from rabbits aged thirty days of gestation t o four weeks postnatal and found that free macrophages in late fetal lungs contain ingested cellular debris and surfactant-related phospholipids, but that ingestions of these lipids increased dramatically shortly after birth. In our material the macrophagelike cells found in the stroma had fewer inclusions and were less active in phagocytosis than the surface cells, as their more sheltered position might lead one t o expect. With fewer functional attributes expressed by macrophage-likecells the earlier the stage of development being examined, it was necessary for us t o rely on morphology to find the earliest examples present. We chose to examine the lungs using light microscopy so as to survey the developing organ Figs. 28, 29. The range of mononuclear cells present on the respiratory surface in 21-day fetal rat lungs. The upper cell (Fig. 28) is the less mature and yet possesses a greater expansion of the granular reticulum than is present in “uncommitted” mononuclear leukocytes. Slender threadlike bodies (arrows) are not unlike small lysosomes in pulmonary macrophages. The lower cell (Fig. 29) is a more typical alveolar macrophage, as seen in late fetal lungs. It exhibits moderate development of cytoplasmic organelles and includes lysosomes of various sizes including large ones (***I that would be conspicuous in PAS-lead hematoxylin stained preparations. Fig. 28, x 11,200;Fig. 29, ~ 1 0 , 3 0 0 . DEVELOPMENT OF PULMONARY MACROPHAGES Fig. 30. Alveolar macrophage present in 3-day postnatal rat lung. Compared to cells in prenatal lungs, postnatal ones are more highly charged with lysosomes 117 of greatly varied configurations: micron-sized and submicronic, electron dense and lucent, solid and lamellated. x 12,500. 118 S.P. SOROKIN, R.F. HOYT, JR.,AND M.M. GRANT with sufficient thoroughness to find the rnacrophage precursors even if only a few were present at a given stage, and we employed 2pm plastic-embedded sections stained by compatible dyes in order to obtain optimum tissue resolution to aid our search The PASlead hematoxylin-stained cells we singled out as macrophage-like were also present in more developed fetal lungs, where their staining pattern was recognizably similar as seen in cells on the respiratory surface and in the interstitium, and where their correspondence was shown to mononuclear cells having elevated lysosomal enzyme activity and to macrophage-like cells identified in electron micrographs. Using PAS-lead hematoxylin we first found macrophage precursors in the lungs of rabbits, rats, and hamsters very early in development at a bronchial bud stage. The cells were located in the stroma without being preferentially sited nearer the developing epithelial bronchial tree, the pleura, or the blood vessels. They appeared detached from other cells in this compartment and stained differently from the largely undifferentiated mesenchymal cells that occupy most of it in early lungs. Cells intermediate between stromal cells and macrophages were not observed in our preparations, just as they were not seen in fetal rat lungs by Collet and Des Biens (19751, although we obtained no better evidence than this to rule against such a derivation from pulmonary mesoderm. Nevertheless, because a circulation is established in the lung from the earliest time, it may seem likely that the stroma is seeded, perhaps repeatedly, from the blood stream. The cells considered to be precursors of macrophages in fetal lungs are rounded mononuclear cells of undifferentiated appearance. In the earlier stages of development they occur side by side with more obvious macrophages having a n irregular outline and a vacuolated cytoplasm (Figs. 13, 21) that we have likened to Hofbauer cells (Enders and King, 1970). As development proceeds, the rounded leukocytes are gradually supplanted by forms transitional between them and typical macrophages of periand postnatal lungs, whereas the vacuolated cells tend to disjappear. Hofbauer-like cells were also found in fetal rat lungs by Collet and Des Biens (1975),who saw them as transient cells of mesodermal origin in evidence only during days 16 and 17 of gestation and apparently unrelated to macrophages, which were not seen until days 20-21. In our interpretation, these cells are but one manifestation of the presence of rnacrophages and related forms that is continuous from the inception of pulmonary development. As studied by Enders and King (1970) in placentas of bats, guinea pigs, and human beings, Hofbauer cells were found to be more vacuolated during early pregnancy when they occur in tissue poorly furnished with extracellular fibers and demonstrable ground substance. They become less vacuolated in late pregnancy when most are indistinguishable from macrophages of adult tissues, and their recognition by light microscopy is made more difficult by their inclusion in a stroma of increased density. These authors considered the vacuolated appearance to reflect active engagement in macropinocytosis. More recently, mononuclear cells from human placentas have been shown to possess receptors for the Fc segment of immunoglobulin G and complement factor C3, to bind antibodies cytophilically on the cell surface, and to ingest bacteria Coke et al., 1982);they therefore satisfy functional criteria for macrophages. In fetal lungs the early appearance and later apparent fading away of Hofbauer-like cells may result from changing environmental conditions that parallel events described in the placenta. The rounded leukocytes observed in the developing lungs evidently are precursors of the pulmonary macrophages because no other suitable candidates were present, but it remains to see if they compare with any of the cell types present in the maturation sequence from monoblast to macrophage proposed for bone marrow-derived macrophages in adult mammals. Participants in this sequence are all mononuclear leukocytes possessing a variety of characteristics in common (esterase and peroxidase activity, receptors for Fc and C3, and capacity for both macro- and microendocytosis) but differing in mitotic capability, cytological features (nucleuskytoplasm ratio, nuclear chromatin pattern, ribosome content of cytoplasm, extent of development of endoplasmic reticulum and Golgi apparatus, etc.), and the intracellular distribution pattern of certain enzymic activities. For practical purposes, ultrastructural morphology and the intracellular distribution of peroxidase activity have been considered sufficient to distinguish among monoblast, promonocyte, monocyte, and macrophage (van der Meer et al., 1982). DEVELOPMENT OF PULMONARY MACROPHAGES Furthermore, in mouse and in man, the essentially undifferentiated monoblast and immature promonocytes are capable of division, whereas the more mature monocyte and macrophages are considered nondividing cells (van Furth et al., 1979). Judged by light microscopic and ultrastructural criteria, the rounded lekocytes we observed in developing lungs are proximal to monocytes in this developmental sequence (Figs. 5, 6, 20, 22, 27, 28).Those present inbronchialbudandpseudoglandular lungs tend to have a larger nucleuskytoplasm ratio and fewer granules than those formed later on (Figs. 6, 20), and so at first are nearer monoblasts and later promonocytes. None were observed to meet standard criteria for identification as monocytes (Weiss, 1983). Indeed, because in all three species these macrophage precursors were frequently observed in division (Figs. 5, 271, they would not meet van Furth’s criteria for monocytes, either. At the same time, transitional forms were observed between the precursors and macrophages, which therefore must have arisen without passing through a monocyte stage. If the pulmonary macrophage population is seeded one or more times from the blood stream, then for much of the prenatal period the seeding cells are proximal to monocytes and likely derived from the yolk sac or liver, hematopoietic centers proximal to the bone marrow. By our estimates the marrow becomes active in developing hamsters during the 13th day, two days after macrophages were detected in fetal lungs. In rabbits, marrow production commences with erythropoiesis on the 23rd to 25th day of gestation, whereas other blood cell types begin to appear there a few days later (King and Ackerman, 1967), some two weeks after macrophage precursors are first found in the lungs. In rats, marrow forms on the 19th day of gestation (Pino and Bankston, 1979), five days after we detected macrophage precursors in the lungs. Had we somehow failed to notice monocytes sequestered in rat lungs, then the error could apply only to the period after the 17th day of gestation, the earliest date monocytes are detectable in the circulation (Deimann and Fahimi, 1978). Thus, on embryological grounds, the first seeding of mononuclear phagocytes in the lungs could not have been from monocytes nor from cells located in the bone marrow. The active replication in fetal lungs not only of monoblastor promonocyte-like cells but also of Hof- 119 bauer-like and other macrophages (Fig. 13)is unequivocal evidence that a resident, selfreplicating population of macrophages becomes established a t the outset of lung formation and is continued in evidence throughout much of the fetal period. The macrophages we found concentrated beneath large pulmonary blood vessels in late fetal life do not necessarily represent a new emigaration of cells from the blood stream but equally may have become activated by serum exudates containing antigens or factors produced by a n awakening immunological system. Exudation into periarterial tissues would be favored by hemodynamic arrangements during fetal life that produce a blood pressure in the pulmonary artery and its main branches equal to that of the aorta (Comroe, 1974). Whether the initial population subsequently becomes diluted out to a vanishing point by later seedings of hematopoietic cells, including those by monocytes, is a question for further investigation. During embryonic development, macrophages are about the first leukocytes to appear with substantially the same characteristics they possess in adult mammals. In mouse embryos they have been traced to extraembryonic precursors located in the yolk sac prior to the time this organ becomes connected to the fetal circulation (day 10 of gestation). These precursors resemble neither promonocytes nor monocytes; less than 0.1% possess functional Fc receptors or adhere to glass, properties their descendants acquire after some days’ growth in agar culture (Cline and Moore, 1972). Fetal mouse liver also serves as a source of macrophage colonies but only after it has gained access to yolk sac cells. These authors nevertheless obtained “rare” glass-adherent cells from 79 day whole fetuses, so that a n embryonic origin of the macrophages cannot be excluded, however unlikely this may seem from the other evidence presented. Furthermore, because macrophages are the earliest differentiated blood cells to appear in developing human liver and initially are more concentrated in the sinusoids than elsewhere in the fetus or placenta, it is possible that they arise from pluripotent hepatic endothelium (Kelemen and Janossa, 1980).Whether embryonic or extraembryonic in origin, fetal rat macrophages (Kupffer cells) appear in the liver soon after the formation of the hepatic diverticulum, engage in phagocytosis, and form a replicating population well before monocytes or 120 S.P. SOROKIN, R.F. HOYT, JR.,AND M.M. GRANT the bone marrow are present (Deimann and Fahimi, 1978; Pino and Bankston, 1979). As we have shown, a similar situation exists in fetal lungs. In both cases, during adult life mitotic activity of the resident macrophage population is less apparent than during fetal life, but it is demonstrable nonetheless (Widmann and Fahimi, 1975; Sorokin, 1983). If a centralized system based in the bone marrow eventually takes over the task of furnishing tissues with macrophages, and does so by supplying them only with end-stage (nondividing) monocytes, then either the monocytes and the macrophages they become are not end-stage cells or the dividing alveolar macrophages observed in adult lungs are descended from another pool, such as the fetal macrophage population we have described. ACKNOWLEDGMENTS The authors wish to thank Messrs. Gregory Anthon, Victor Carabba, Peter Lauren, and Stephen Sarikas for technical assistance contributed to this study. The work was supported by Research Grants HL-19379, ES00583, and HL-A129175 from the National Institutes of Health, USPH, supplemented by funding from The Gorilla Foundation (Boston). LITERATURE CITED Barka, T., and P.J. Anderson (1962)Histochemical methods using hexazonium pararosaniline as coupler. J. Histochem. Cytochem., lOt741-753. Blusse, A., and R. van Furth (1979) Origin, kinetics, and characteristics of pulmonary macrophages in the normal steady state. J. Exp. Med., 149t1504-1518. Boorman, G.A., L.W. Schwartz, and F.D. Wilson (1979a) Formation of macrophage colonies in vitro by free lung cells. J. Reticuloendothel. SOC.,26t855-866. Boorman, G.A., L.W. Schwartz, and F.D. Wilson (197913) In vitro macrophage colony formation by free lung cells during pulmonary injury. J. Reticuloendothel. SOC.,26:867-872. Bowden, D.H., and I.Y.R. Adamson (1972) The pulmonary interstitial cell as immediate precursor of the alveolar macrophage. Am. J. Pathol., 68.521-528. Brain, J.D., J.J. Godleski, and S.P. Sorokin (1977)Quantification, origin and fate of pulmonary macrophages. In: Respiratory Defense Mechanisms. J.D. Brain, D.F. Proctor, L.M. Reid, eds. Marcel Dekker, New York, pp. 849-892. Cline, M.J., and M.A.S. Moore (1972) Embryonic origin of the mouse macrophage. Blood, 39.2342-849. Collet, A. J., and G. Des Biens (1975) Evolution of mesenchymal cells in fetal rat lung. Anat. Embryol., 147t273-292. Comroe, J. H. (1974) Physiology of Respiration, 2nd ed. Year Book Medical Publishers, Chicago. Deimann, W., and H.D. Fahimi (1978)Peroxidase cytochemistry and ultrastructure of resident macrophages in fetal rat liver. A developmental study. Dev. Biol., 66:43-56. Dicke, K.A., M.J. van Noord, B. Maat, U.W. Schaefer, and D.W. van Bekkum 11973) Attempts at morphological identification of the hemopoietic stem cell in primates and rodents. In: Hemopoietic Stem Cells. (No ed. given) Elsevier Science Publ. Co., New York, pp. 47-71. Enders, A.C., and B.F. King (1970) The cytology of Hofbauer cells. Anat. Rec., 167t231-252. Grant, M.M., S.P. Sorokin, and J.D. Brain (1979) Lysosoma1 enzyme activities in pulmonary macrophages from rabbits breathing iron oxide. Am. Rev. Resp. Dis., 120: 1003-1012. Ito, S., and M.J. Karnovsky (1968)Formaldehyde-glutaraldehyde fixatives containing trinitro compounds. J. Cell Biol., 39t168a-169a. Karnovsky, M.J. (1965) A formaldehyde-glutaraldehyde fixative of high osmolarity for use in electron microscopy. J. Cell Biol., 27t137 (abstract). Karnovsky, M.J. (1971) Use of ferrocyanide reduced os' mium tetroxide in electron microscopy. Eleventh Annual Meeting of the American Society of Cell Biology, New Orleans, LA, p. 146 (abstract). Kelemen, E., and M. Janossa (1980) Macrophages are the first differentiated blood cells formed in human embryonic liver. Exp. Hematol., 8t996-1000. King, J.E., and G.A. Ackerman (1967) Erythropoiesis in the bone marrow of the fetal rabbit. Anat. Rec., 157t589-606. Loke, Y.W., 0. Eremin, J. Ashby, and S. Day (1982) Characterization of the phagocytic cells isolated from the human placenta. J. Reticuloendothel. SOC., 31r317324. McDowell, E.M., and B.F. Trump (1976) Histologic fixatives suitable for diagnostic light and electron microscopy. Arch. Pathol. Lab. Med., I00t405-414. Metchnikoff, I. (1893) Lectures on the Comparative Pathology of Inflammation. Kegan Paul, Trench, Trubner, London. Reprinted by Dover, New York, 1968. Novikoff, P.M., A.B. Novikoff, N. Quintana, J.J. Hauw (1971) Golgi apparatus, GERL, and lysosomes of neurons in rat dorsal root ganglia, studied by thick section and thin section cytochemistry. J. Cell Biol., 5Ot859-886. Pearse, A.G.E. (1972) Histochemistry, Theoretical and Applied, 3rd Ed. Williams and Wilkins, Baltimore. Pino, R.M., and P.W. Bankston (1979) The development of the sinusoids of fetal rat liver: Localization of endogenous peroxidase in fetal Kupffer cells. J. Histochem. Cytochem., 27t643-652. Pratt, S.A., M.H. Smith, A.J. Ladman, T.N. Finley 11971) The ultrastructure of alveolar macrophages from human cigarette smokers and nonsmokers. Lab. Invest., 24t331-338. Reed, C.E., and J.M. Bennett (1975) N-acetyl-8-glucosaminidase activity in normal and malignant leukocytes. J. Histochem. Cytochem., 23t752-757. Borokin, S.P. (1977) Phagocytes in the lungs. In: Respiratory Defense Mechanisms. J.D. Brain, D.F. Proctor, L.M. Reid, eds. Marcel Dekker, New York, pp. 711-848. Sorokin, S.P. (1983) Dynamics of lysosomal elements in pulmonary alveolar macrophages: I. The postactivation lysosomal cycle. Anat. Rec., 206:117-143. Sorokin, S.P., and R.F. Hoyt, Jr. (1978) PAS-lead hematoxylin as a stain for small-granule endocrine cell populations in the lungs, other pharyngeal derivatives and the gut. Anat. Rec., I92:245-260. Sorokin, S.P., and R.F. Hoyt, Jr. (1982) Development of neuroepithelial bodies and solitary endocrine cells in DEVELOPMENT OF PULMONARY MACROPHAGES fetal rabbit lungs. 11. Nonspecific esterase as an indicator of early maturation. Exp. Lung Res., 3.261-272. van der Meer, J.W.M., J.S. van de Gevel, R.H.J. Beelen, D. Fluitsma, E.C.M. Hoefsmit, and R. van Furth (1982) Culture of human bone marrow in the teflon culture bag: Identification of the human monoblast. J. Reticuloendothel. SOC.,32:355-369. van Furth, R., J.A. Raeburn, and T.L. van Zwet (1979) Characteristics of human mononuclear phagocytes. Blood, 54:485-500. Weiss, L., ed (1983) Histology: Cell and Tissue Biology, 121 5th Ed. Elsevier Science Publishing Co., New York, chapters 4, 11, 12; pp. 139-177,447-497. Widmann, J.J.,and H.D. Fahimi (1975) Proliferation of mononuclear phagocytes (Kupffer cells) and endothelial cells in regenerating rat liver. Am. J. Pathol., 80t349-366. Zeligs, B.J., L.S. Nerurkar, J.A. Bellanti, and J.D. Zeligs (1977) Maturation of the rabbit alveolar macrophage during animal development. I. Perinatal influx into alveoli and ultrastructural differentiation. Pediatr. Res., 11:197-208.