Small-granule APUD cells in relation to airway branching and growthA quantitative cartographic study in syrian golden hamsters.код для вставкиСкачать
THE ANATOMICAL RECORD 213:410-420 (1985) Small-Granule APUD Cells in Relation to Airway Branching and Growth: A Quantitative, Cartographic Study in Syrian Golden Hamsters STEPHEN N. SARIKAS, RICHARD F. HOYT, JR.,AND SERGE1 P. SOROKIN Department of Anatomy, Boston University School of Medicine, Boston MA 02118 ABSTRACT Small-granule APUD cell clusters and their clear-cell precursors were mapped in serial 2-pm glycol methacrylate-embedded, periodic acid-Schiff(PAS)lead hematoxylin-stained sections of 13-, 14-, and 15-day fetal hamster lungs. Every sixth section was drawn from a camera lucida projection on tracing paper. Each tracing included the profiles of nonalveolated air passages and the locations of smallgranule cell clusters and solitary clear cells. Airways containing ciliated cells and those surrounded by condensed mesoderm were also labeled. Single clear cells were rare in fetal hamster lung. Of 2,368 endocrine cell loci identified in the three fetal age groups examined, only 14 were single clear cells. A preliminary survey of the entire left and right lungs showed that the pattern of airway and small-granule cell development in the infracardiac lobe was similar to that occurring in the remainder of the lung; this lobe was accordingly considered a model for the whole lung, and the ontogeny of its small-granule cell population was quantitated and compared with results of similar quantitative mapping of this lobe in a n adult animal (Hoyt et al., 1982a,b). Along the lobar bronchus of the 13-day infracardiac lobe and proximal portions of its main branches, small-granule cell clusters occurred most often near airway intersections. As the number and density increased in subsequent fetal stages, small-granule cell clusters became conspicuous along internodal bronchial segments. In distributing bronchioles, the population density of small-granule cell clusters decreased between 13 and 14 days but more than doubled by day 15. Unlike human lungs, where centrifugally developing small-granule cell clusters are firmly established in terminal bronchioles well before birth, most peripheral bronchioles in fetal hamster were devoid of small-granule cell clusters, even a t 15 days, one day before birth. Comparison of numerical population densities in this lobe of fetal and adult lungs indicates that small-granule cell clusters continue to form past day 15 and suggests that they are considerably more numerous in adult than fetal lung. To date, very little quantitative information has been published on the number, density, and distribution of pulmonary neuroepithelial bodies and small-granule cell clusters in any species during fetal development, although the pattern of their differentiation is now known in several. Where these data can be compared with counts of similar cells in adult lungs, they not only extend our understanding of this putative pulmonary endocrine or paracrine system but may help to indicate a t what periods of life it is likely to be active. Currently, the only fully quantitative estimate of small-granule cell populations in adult lungs is that of Hoyt et al. (1982a,b),derived from a serial survey of the infracardiac lobe from one adult hamster lung a t 3-pm intervals in glycol methacrylate sections stained by PASlead hematoxylin. Dimensions of the bronchial tree, locations of endocrine cell loci, and population densities of clustered and solitary endocrine cells were closely examined along all levels of the conducting airways. The present study has been undertaken to obtain similar quantitative data describing the population density 0 1985 ALAN R. LISS, INC. and distribution of small-granule cell clusters in the developing fetal hamster infracardiac lobe, built upon a general understanding of small-granule cell ontogeny in this species (Sarikas et al., 1985)and demonstrating that in this respect the lobe serves as a n adequate model for remaining parts of the lungs. The data provide the basis for comparisons made between the developing population of small-granule cell clusters in fetal lungs and that in the stable, mature organ. MATERIALS AND METHODS Preparation of Tissue Specimens The quantitative results to be described were based upon study of one fetal hamster for each developmental Received February 25, 1985; accepted May 29, 1985. Dr. Sarikas’s present address is: Department of Biochemistry, University of Massachusetts Medical Center, 55 Lake Avenue North, Worcester, MA 01605. Address reprint requests t o Dr. Hoyt. 411 HAMSTER SMALL-GRANULE CELLS AND AIRWAY GROWTH Fig. 1. Representative line drawing of a transverse 2-pm section of plastic-embeddedlung, 14day fetal hamster. Profiles of nonalveolated airways are marked to show small-granule cell clusters (solid black spots) and investment by condensing mesoderddeveloping smooth muscle (dashes). Ciliated cells were absent. L = left lung; RU, RM, RL = right upper, middle, lower lobes. stage a t 13, 14, and 15 days of gestation. Each specimen was selected as fully representative of its age from a much larger sampling of fetuses used for the previous qualitative study (Sarikas et al., 1985). The thoraxes were dissected out, fixed by immersion for 24-48 hr, and processed for light microscopy as described in that report. The entire lung of each fetus was serially sectioned in situ a t 2 pm on a Jl3-4 microtome. Accurate cutting records were made of lost sections and knife changes. Two sections were collected on each precleaned glass slide and every third slide was stained with PAS-lead hematoxylin (Sorokin and Hoyt, 1978). --To_ -1_ - Serial Tracings of Fetal Hamster Lungs Lung profiles of every sixth section (ca. 12-pm inter~ white vals) were drawn a t a magnification of 1 2 5 on tracing paper using a Zeiss binocular microscope and drawing tube. Each tracing included the outlines of all conducting airways and the location of small-granule cell clusters and their “clear cell” precursors (Fig. 1)as confirmed by microscopic examination a t 1,000 x . Airways surrounded by condensed mesoderm as well as those containing ciliated cells were also marked appropriately in the drawings. A total of 619 tracings were made: 152 from the 13-day, 205 from the 14-day, and 262 from the 15-day fetal lung. Bronchi and bronchioles were subdivided into unit airways (Fig. 2), which were defined as the distance between successive branch points (Hoyt et al., 1982b). Then, for each age group, the length of unit airways containing small-granule cell clusters in all pulmonary lobes was calculated and recorded in millimeters as follows: Using the serial tracings, the lung profiles were superimposed by best overall fit of the major airways in all the lobes. The horizontal displacement (Fig. 3) between the centers of airway profiles representing the beginning and end of a given unit airway was determined by measuring the distance between each center fig. 2. In this diagram, each daughter airway joins its parent along a ring of junction (dashed circle) including the carinal point (black spot). Successive branch points define a unit airway (A). point and dividing that value by the magnification. The corresponding vertical displacement (Fig. 3) was approximated by multiplying the section thickness (2 pm) by the number of sections cut between the appropriate airway profiles determined from the cutting records. The lateral and vertical displacements formed two sides of a right triangle. The hypotenuse of this triangle approximated the axial length of the unit airway and was determined by application of the Pythagorean theorem (a2 b2 = c2). The total length of any given airway (e.g., the lobar bronchus) could be calculated by adding together all the units comprising that airway (Fig. 3). + 412 S.N. SARIKAS, R.F. HOYT, JR.,AND S.P.SOROKIN A \a" I 413 414 S.N. SARIKAS, R.F. HOYT, JR.,AND S.P. SOROKIN basis for the studies of Hoyt et al. (1982a,b). Measurements of airway length, number of small-granule cell clusters, and their population densities were compared with corresponding data tabulated from the adult lobe in these same two studies. RESULTS w 0 ' 1 i; 1 c/ In hamsters, the left lung consists of a single large lobe; the right comprises the right upper, right middle, right lower, and infracardiac lobes. Each lobe is aerated by a lobar bronchus that forms a central longitudinal axis and gives rise to lateral airways called divisional distributing bronchioles by Hoyt et al. (1982b). These latter branch directly or indirectly to terminal bronchioles leading into alveolar ducts. Formation of the Hamster Airway -jb a Fig. 3. Approximation of airway length. Serial tracings (viewed from above at top, edge-on at bottom) were superimposed for best fit for all main airways in both lungs. Lateral displacement (a), measured between center points (black spots) of profiles marking the ends of a unit airway, was corrected for magnification. Vertical displacement (b) was determined by multiplying section thickness (2 pm) by the number of sections cut between each profile. Airway length (c) was calculated by the Pythagorean theorem (a2 + b2 = c2). Airway Maps By carefully superimposing the serial tracings and allowing for the distance between them, we could reconstruct the developing lobar airway and represent it in two dimensions. The position of small-granule cell clusters, the airways encircled by condensed mesoderm, and in some lobes those containing ciliated cells were plotted on the resultant airways maps (Figs. 4 and 51, one of which was made for every lung lobe a t each developmental stage. It was then possible to count the number of small-granule cell clusters and their clear-cell precursors and assign them to their appropriate unit airways. Detailed Analysis of the tnfracardiac Lobe Preliminary examination of airway maps and cell cluster counts revealed a similar developmental pattern in the left lung and all four lobes of the right lung. Therefore, the infracardiac lobe, being the smallest and also the subject of exhaustive study in the adult, was chosen for more detailed quantitative analysis as follows: 1)The length of each unit airway in the 13-, 14-, and 15-day fetal infracardiac lobes was measured as described above, and the airway maps were redrawn to scale. 2) The population density of small-granule cell clusters was determined in the lobar bronchus, divisional distributing bronchioles, and more peripheral airways by dividing the number of clusters a t each level by the total length of the appropriate unit airways. 3 ) Microscopic slides, serial tracings, and airway maps were reexamined to determine how many small-granule cell clusters were associated with the rings of intersection of parent and daughter airways (see Fig. 2) and how many were located at bronchioloalveolar junctions, which were not defined until day 15 of gestation. Airway maps prepared from the three fetal infracar~ diac lobes were compared directly with the 7 0 cardboard reconstruction of the adult lobe that served as the The reconstructed adult infracardiac lobar bronchus consists of 15 unit airways and gives rise to 14 welldefined lateral divisional bronchioles, each the origin of a branching subsystem ending in terminal bronchioles. Of these, the proximal are larger and more highly arborized than the distal, and a 15th, most distal lateral branch of the lobar bronchus is simply a terminal bronchiole. The final segment of the lobar bronchus ends in a spray of terminal bronchioles and can itself be considered a peripheral airway subsystem. Examination of Figure 4 shows that on fetal day 13, the infracardiac lobar bronchus has 15 lateral branches and ends in a spray of fine airways. An organized layer of mesoderm, harbinger of smooth muscle differentiation, invests virtually the entire length of the lobar bronchus and the proximal reaches of its lateral branches. At this stage, only eight unit airways can be distinguished along the lobar bronchial stem owing to the fact that the lateral airways tend to arise in pairs a t what appear to be airway trifurcations. As the lobar bronchus elongates over the next 48 hr, however, the origins of its lateral branches are gradually separated, so that by day 15 each is distinct, as in the adult. Arborization of nonalveolated airways is at its height on day 14, by which time it is clear that proximal lateral airway systems are much more highly branched than those arising further out along the bronchial stem. A complete, organized layer of mesoderm now invests even the terminal ramification of the lobar bronchus and has spread peripherally along several additional orders of branches in each of the lateral, divisional bronchiolar systems. The organized mesodermal sleeve makes only a slight peripheral advance during the next 24 hr of gestation. Distal to the sleeve, the previously nonalveolated reaches of the airway now begin to transform into respiratory saccules. This high water mark of substantial mesodermal investment at fetal day 15 corresponds with very little alteration to the extent of the adult conduct- Fig. 5. Rough maps of nonalveolated airways in 13-day (top), 14-day (middle), and 15-day (bottom) fetal hamster left lungs (not to scale), showing spread of small-granule cell clusters (black dots) and peripheral progression of developing smooth muscle (dashes). Positions of ciliated cells are not marked on this figure. Developmental events resemble those in the infracardiac lobe (Fig. 4).In the 15-day sketch, the bronchial axis has been compressed to fit the page. HAMSTER SMALL-GRANULE CELLS AND AIRWAY GROWTH 415 416 S.N. SARIKAS, R.F. HOYT, JR.,AND S.P. SOROKIN ing airway, considered as numbers of specific unit airways and not as airway length. The only apparent differences between the pattern seen in the 15-day airway map and the adult reconstruction can be accounted for by perinatal or postnatal alveolarization of scattered unit airways a t the periphery of the lateral branching systems and in the terminal segment of the lobar bronchus. Thus it appears that although a few peripheral airways destined to become terminal bronchioles may have been laid down by day 13, these cannot reliably be distinguished from presumptive distributing bronchioles, because not all of the latter have yet acquired an investment by organizing myoblasts. In 14- and 15day lobes, however, unit airways in the lateral branching systems can be divided into three main groups: 1) those corresponding to the hierarchy of distributing bronchioles in the adult, which are well invested by smooth muscle; 2) those corresponding to terminal bronchioles, which are also invested by smooth muscle; and 3) those likely to be converted into alveolated airways of the respiratory zone, without a definite mesodermal investment. At 14 days, it is still difficult to discriminate accurately whether a few peripheral airways are to become distributing or terminal bronchioles, but by 15 days the distinction is a simple one. Number and Distribution of Endocrine Cell Loci in the Fetal Hamster Lung The basic centrifugal patterns of airway formation, mesodermal condensation associated with smooth muscle differentiation, and small-granule cell cluster formation were similar in all major subdivisions of the lung pair, as exemplified in the drawn-to-scale airway diagrams of the infracardiac lobe (Fig. 4) and the roughly scaled maps of the left lung (Fig. 5). Small-granule APUD cells and their clear-cell precursors first appeared in lobar bronchi and divisional distributing bronchioles on fetal day 13, and they subsequently developed in a centrifugal wave, spreading into more peripheral unit airways of the lateral and terminal bronchial branching subsystems. Almost without exception, the leading edge of their advance did not exceed that of the organizing myoblasts investing the airway. Both were coterminous on day 13, but smallgranule cell clusters trailed the muscle by 3-4 generations on day 14 only to become coterminous again a t day 15 (Figs. 4,551, when a few clusters had reached bronchioloalveolar junctions. Small-granule cell differentiation preceded that of ciliated epithelial cells by many generations of branches and a full 24-48 hrs (Fig. 4). Counts of small-granulecell loci in fetal lungs (Table 1) In the three age groups studied, small-granule cell clusters and single cells were identified at a total of 2,368 loci. Of this total, only 14 were single clear cells, representing 1.7% of all loci in the 13-day, 1.0% in the lCday, and 0.2% in the 15-day fetal lung. Over half (54%) of all loci occurring in the the 13-day fetal lung pair consisted of cell clusters found along the lobar bronchi. Although the actual number of clusters in lobar bronchi had increased by 139 at day 14 and by another 130 a t day 15, the rise was exceeded by the differentiation of clusters in more recently established peripheral airway generations. Thus, the lobar bronchi contained 40% of all endocrine cell loci on day 14, and only 26% on day 15. This pattern of small-granule cell development was common to all major subdivisions of the lung pair. Close examination of Table 1reveals a steady increase in the number of small-granule cell clusters in each individual lobar bronchus, followed and ultimately exceeded by a TABLE 1. Total number of small-granulecell clusters and solitary clear cells in 13-, 14-,and 15-day fetal hamster lungs No. of No. of Fetal small-granule cell clusters solitary clear cells Total No. of Lobar Peripheral Lobar Peripheral clusters and age (days) Sample bronchus bronchioles Total bronchus bronchioles Total clear cells 3 82 1 2 34 79 45 13 Left lung 1 58 0 16 57 1 41 Right lower lobe 0 26 0 0 9 26 17 Right middle lobe 0 18 0 0 I1 18 Right upper lobe 7 0 41 0 0 29 41 12 Infracardiac lobe 4 225 1 99 221 3 122 Total (1.7) (100) (0.4) (1.3) (54) (44) (98) (%) 3 194 2 1 191 115 76 14 Left lung 2 2 17 0 2 137 215 78 Right lower lobe 0 99 0 0 99 52 47 Right middle lobe 0 65 0 0 65 23 42 Right upper lobe 2 74 1 1 72 32 40 Infracardiac lobe 7 649 5 642 2 261 381 Total (100) (0.7) (1.0) (59) (99) (0.3) (40) (%) 496 0 1 1 497 120 376 Left lung 15 477 1 1 2 479 93 384 Right lower lobe 0 183 0 0 183 71 112 Right middle lobe 151 0 0 0 151 37 114 Right upper lobe 0 184 0 184 0 70 114 Infracardiac lobe 3 1,494 2 1 1,491 391 1,100 Total (0.07) (0.13) (99.8) (0.20) (100) (74) (26) (%I HAMSTER SMALL-GRANULE CELLS AND AIRWAY GROWTH rise in the number of clusters in more peripheral airways as these younger generations are populated through formation and maturation of clear-cell clusters. This sequence of events is seen somewhat earlier in the larger left lung (Fig. 5) and right lower lobe and somewhat later in the smaller right middle lobe. In the right upper and infracardiac (Fig. 4) lobes, which are smaller still, the lobar bronchi are short and contain few smallgranule cell clusters. In these two lobes, on day 13 the majority of endocrine cell loci in the wall of the lobar bronchus are found close to branch points rather than in the internodes, but the pattern of development is the same as in the larger lobes: An increase in loci in the lobar bronchus during days 13-15 is exceeded by the formation of cell clusters in younger peripheral airways. From examination of the airway diagrams and Table 1, it appears that any single lobe would serve as a reasonable model for further quantitative studies of small-granule cell formation in fetal hamster lung. The infracardiac lobe has been chosen in the present instance because comparison can be made with data already available regarding the number and distribution of small-granule cells in the infracardiac lobe of the adult hamster lung published previously by Hoyt et al. (1982a,b). In the following description, all references to the adult hamster pertain to these two studies. 417 As a consequence of these differential growth patterns during fetal and postnatal life, the relative composition of the bronchiaI tree changes. Of total conducting airway length, the lobar bronchus accounted for 23% on day 13,13% on days 14 and 15, and only 7% in the adult. Similarly, the proportion provided by distributing bronchioles decreased continually as the lobe matured, from 77% at 13 days, to only 48% at 15 days and 46% in the adult. Conversely, terminal bronchioles, which could not be identified reliably on day 13, accounted for 33% of total airway length a t 14 days, 39% at 15 days, and finally 47% in the reconstructed adult lobe. Population density and distribution of small-granule cell clusters in the infracardiac lobe (Table 3) Twelve small-granule cell clusters were located in the lobar bronchus of the 13-day infracardiac lobe examined. By day 14 and again by day 15, the number of clusters had more than doubled to 32 and 70, respectively, but there was only a slight further increase noted betwen 15 days and the adult. This was reflected by a steady increase in the density of clusters (clusters per millimeter along the airway long axis) during fetal development, when the lobar bronchus was growing slowly. The population density was lower in the adult, however, owing to perinatal and postnatal lengthening of the unit airways in the lobar bronchus. In distributing bronchioles, there was a reduction in Quantitative Investigation of Small-Granule Cell Cluster the population density of small-granule cell clusters beFormation in the Fetal Hamster lnfracardiac Lobe tween day 13 and day 14 of gestation. This could be Dimensions of the conducting airways in the infracardiac attributed to a n increase of 45 new unit airways (Table lobe (Table 2) 2) but only six clusters during this period. By day 15, The conducting airway system of the infracardiac lobe with a reduction of 12 unit airways (Table 2) and a increased in total length from 4.5 mm on day 13, to 12.1 nearIy threefold increase in the number of small-granmm on day 14, and 18.3 mm on day 15, as compared ule cell clusters, the population density had more than with 59.5 mm in the adult. This growth may be divided doubled. The population of small-granule cell clusters in into two phases. During the first phase, the number of the adult distributing bronchioles was more than 2.5 identifiable conducting airways increases substantially times that a t 15 days, but the density remained nearly as terminal buds divide to form additional generations constant because of the threefold increase in total length of branches and a s the sleeve of developing muscle ad- of these airways (Table 2). vances peripherally. There is relatively little elongation Cell clusters were not common in terminal bronchioles of, say, proximal elements among distributing bron- of the developing infracardiac lobe. Only five were found chioles, and that which does occur is offset by addition in the 14-day and nine in the 15-day fetuses. In the of younger and shorter peripheral generations; conse- adult, however, 256 clusters, or 41% of the total populaquently, mean length of unit airways remains nearly tion, occurred in terminal bronchioles. constant. Small-granule cell clusters were widely distributed This process comes to a n end between day 14 and day throughout the fetal infracardiac lobar bronchus. In the 15 with the appearance of primitive respiratory saccules 13-day lung, 75% of the lobar bronchial unit airways beyond the leading edge of the muscle coat, signaling contained such clusters; a t 14 and 15 days, the figure the beginning of alveolarization in the distalmost divi- increased to about 90%. In contrast, most peripheral sions of the bronchial tree. bronchioles were devoid of clusters. In the 15-day lung During the second, perinatal and postnatal phase, only 53% of distributing bronchiolar unit airways and growth of the conducting airway system is accomplished 8%of terminal bronchioles contained cell clusters, comexclusively through elongation of units previously laid pared with 85% and 83%, respectively, in the adult lobe. down. This phase overlaps the first to some extent and begins in older, proximal airways. For example, the Relationship of small-granule cell clusters to airway increase in lobar bronchial unit airways during days intersections and bronchioloalveolar junctions (Table 4) 13-15 seems due largely to relative separation of the Examination of the serial sections and maps revealed origins of previously formed lateral branches brought that over 25% of all clusters were Iocated on junctional about a s a result of elongation of the bronchus itself. rings (Fig. 2) a t airway intersections in the 13-day and Furthermore, by day 15 the lobar bronchus has attained 14-day infracardiac lobes. However, by 15 days, the numroughly half of its adult length, compared with one-third ber of clusters had increased mainly along regions of the in the case of the more peripheral distributing bron- larger airways between branch points (Fig. 4), so that chioles and only one-quarter for terminal bronchioles, only 11%of all clusters appeared to be located at airway each of which consists of a single unit airway. intersections. In the adult, where loci were marked on a Adult' 13 Day Distributing bronchioles 14 15 Day Day Adult' ~ - - 13 Day 0.05 0.02 3.9 33 76 0.06 0.04 7.2 39 120 55 13 Day 180 219 All airways 14 15 Day Day 21 32 44 92 12 12 29 75 32 14 100 38 90 71 87 70 29 19 8 13 Day 29 'Data from Hoyt et al. (1982a,b). 'Measured along the airway long axis. 31ncludes six cell clusters located in the respiratory zone. Small-granule cell clusters/mm2 Total number of clusters % of all clusters in the lobe Unit airwavs with clusters (% total) 13 Day Lobar bronchus 14 15 Day Day Adult' 23 49 35 5 44 57 85 272 105 53 10 12 Distributing bronchioles 14 15 Day Day Adult' - 13 Day 5 7 8 9 5 7 1 1 83 41 256 9 Terminal bronchioles 14 15 Day Day Adult' 38 100 41 3 13 Day 21 100 72 2 31 100 184 11 85 100 6213 10 All airways 14 15 Day Day Adult' TABLE 3. Distribution and density of small-granule cell clusters in the infracardiac lobe of 13-, 14-, and 15-day fetal and adult hamster lungs 209 Adult' 0.25 0.08 0.07 0.08 0.28 0.19 0.04 0.04 0.06 0.17 4.5 28.0 12.1 18.3 59.5 100 100 100 100 47 112 Terminal bronchioles 14 15 Day Day Adult' 'Data from Hoyt et al. (1982b). 'Of 15 unit airways making up the lobar bronchus in the adult, the two most proximal were not included in the reconstruction. 47 92 80 84 No. of 8 12 19 15' unit airways Length (mm) of unit airways 0.34 0.07 0.07 0.11 0.33 0.13 0.13 0.14 Mean 0.18 0.04 0.04 0.06 f SD 0.05 0.05 0.07 0.14 8.7 27.0 3.5 6.6 1.6 2.4 4.5 Total 1.0 77 54 48 46 13 7 %Total of 23 13 all airways 13 Day Lobar bronchus 14 15 Day Day TABLE 2. Dimensions of the airways in the infracardiac lobe of 13-, 14-, and 15-day fetal and adult hamster lungs. 419 HAMSTER SMALL-GRANULE CELLS AND AIRWAY GROWTH TABLE 4. Number of small-granule cell clusters at airway intersections and bronchioloalveolar junctions in the infracardiac lobe of 13-, 14-, and 15-day fetal and adult hamster lungs Small-granulecell clusters, actual number (9% of total) Position Airway intersections Junctional rings Carinal points Total Bronchioloalveolar junctions Neither intersections nor junctions Total 13 day 14 day 15 day Adult' 6 (15) 5 (12) 11(27) 11(15) 8 (11) 19 (26) - 8 (4) 12 (7) 20 (11) 7 (4) 99 (16) 25 (4) 124 (20) 160 (26) 30 (73) 53 (74) 157 (85) 337 (54) 41 (100) 72 (100) 184 (100) 621 (100) - 'Data from Hoyt et al. (1982a). Patterns of Development in the Pulmonary Airways scale-model reconstruction of this lobe, 20% were found on junctional rings a t intersections. Very few clusters By 13 days of gestation in hamsters, the main scaffold populated the small distributing and terminal bron- of the bronchial tree is already present with all the main chioles of the 15-day infracardiac lobe (Fig. 41, and only subdivisions of the lobes defined. Each lobar bronchus 4% were located at the recently formed bronchioloalveo- forms a much more distinct axis than is the case in lar junctions, compared with 26% in the adult. human beings (Wells and Boyden, 19541, and proximal branches are more extensively subdivided than distal DISCUSSION ones, so that the lobes are approximately pyramidal in The results of this quantitative, cartographic study on configuration. Divisional distributing bronchioles furthe developing small-granule cell system of hamster nish a less distinct axis for their subdivisions and are lungs, based on limited sampling, bears out the main less pyramidal overall. These configurations are reconclusions reached in the associated morphological tained in subsequent developmental stages to maturity. study (Sarikas et al., 1985), based on extensive samAll major segments of the future conducting airways pling. Both studies indicate that the system is not fully can be identified in all lobes of 14-day lungs by the established by late fetal life, although it undergoes rapid presence of a n investment of differentiating smooth development during the last quarter of gestation. muscle and connective tissue cells about them. By 15 Specific maps necessarily are based on individual ani- days these airways have been completely defined as a mal specimens, and because of the time required to result of further advance of the muscle sleeve to cover prepare them accurately, it is a practical necessity to terminal bronchioles. At the same time, alveolarization limit their number. In answering anticipated criticism begins in earnest. Further development of the purely that the value of the quantitative study is diminished conducting passages is mainly in length and diameter because measurements were taken from only one ex- of the unit airways and in the differentiation of their ample for each of the three fetal stages, we would re- constituent cells. spond as follows: 1)The single examples were selected Taking the infracardiac lobe as a model, and after as representative from a wealth of material used in the comparing 15-day prenatal with adult stages, it appears morphological work and were completely rather than that very few muscle-invested airways are subjected to statistically studied by serial sections passing through retrograde alveolarization as a mechanism to increase the entire lung pair a t each stage, so that neither airway the respiratory surface, since only 10 unit airways are branches nor small-granule cell clusters were over- lost in the interval (Table 2). In hamsters, unlike dogs looked. 2) When maps drawn a t 13, 14, and 15 days of or human beings (Boyden and Tompsett, 1961, 1965; gestation for any one of the five lung lobes were laid Boyden, 1967), alveolarization appears virtually conside by side for comparison, it was at once seen that a fined t o present and future branches of the 15-day fetal characteristic branching pattern for the particular lobe lung distal to the muscle coat, and this may help to was recognizable at all three stages, and when arranged explain why transitional, alveolated airways like respiin ascending order of gestational age they formed a ratory bronchioles are either short or absent in this coherent developmental sequence for the airways of that species (Tyler, 1983). lobe. Had any stage been grossly aberrant, it would have confuted the series. 3) In the case of the infracardiac Population Densities of Small-Granule Cells in Developing lobe, it was further evident from comparison with our and Adult Lungs scale model of this lobe in adult lung that the branching As seen in data from the infracardiac lobe (Table 3, pattern seen in fetal life is similar to the pattern present at maturity and that the four stages illustrate the devel- Fig. 4), the overall population density of small-granule cell clusters is very low on days 13 andl4 (ca. 3 clusters opmental history of this lobe. 420 S X . SARIKAS, R.F. HOYT, JR., AND S.P. SOROKIN per millimeter airway length) but rises sharply to 11per millimeter on day 15 and remains about the same (10 per millimeter) in the adult. This could be taken to mean that the population density attained in late fetal life is maintained a t this level by uniform addition of new cell clusters during perinatal and postnatal elongation of airways laid down by day 15, but examination of the breakdown by airway classes shows this to be incorrect. Inasmuch as the lobar bronchus contains 80% of its adult complement of cell clusters by day 15, the distributing bronchioles 39%, and the terminal bronchioles only 3.5%, the data uphold the conclusion of a proximalto-distal progression of small-granule cell development drawn in the morphological part of this study (Sarikas et al., 1985). Between days 13 and 14 the small-granule cell population increases by a factor of 1.8; between days 14 and 15, it increases by 2.4 times; and between day 15 and adulthood, it rises by 3.4 times. A growth spurt in the last day before birth conceivably could account for much of the latter, more than threefold increase in smallgranule cell clusters, but we have no counts for the day of birth (day 16) to determine if this is so. Nevertheless, our finding of a mitotic figure in a neonatal small-granule cell cluster, and the evident immaturity of many small-granule cells in early postnatal lungs as reported in the accompanying paper, both argue that neoformation of clusters continues on into postnatal life. Our data do show that the increase in the population between day 15 and adult life is largely accounted for by the appearance of clusters in previously unpopulated unit airways and only in a minor way by formation of new clusters to fill in between older established clusters in elongating airways. In the lobar bronchus, all but two, or 90%, of the unit airways contained cell clusters at day 15, whereas all possessed them in the adult. In distributing bronchioles, 53% of unit airways had clusters on day 15, compared to 85% in the adult, and in terminal bronchioles, 8% had clusters on day 15 compared to 83% in the adult. Thus, the greater part of small-granule cell neoformation after 15 days occurs in unit airways of the more distal muscle-invested portions of the bronchial tree. Relationship Between Small-Granule Cell and Other Differentiations in the Lungs From the inspection of the developmental maps (Figs. 4, 51, spatial and temporal association Lippears to exist mainly between the appearance of new clusters of smallgranule cells and the peripheral advance of the sleeve of differentiating smooth muscle along the airway. In contrast, branch points have already been selected and branching has occurred before clear-cell precursors of the small-granule cells become manifest at these locations, and other cellular differentiations in the pulmonary endoderm such as ciliation lag many generations of branching behind the wave of small-granule cell formation. During fetal development, precursors of small-granule cells almost without exception do not appear ahead of the muscle coat but either advance with its leading edge of lag behind it by a few generations of branching. Indeed, in the adult infracardiac lobe only 6 of 621 clusters were located in alveolar ducts. bevond the muscle coat. Whether a causal, perhaps inductive interaction occurs between special elements in the investment and the epithelium to initiate formation of small-granule cells awaits further investigation. Because formation of the respiratory zone is already well under way before small-granule cells significantly populate terminal bronchioles, no case can be made in hamsters that the clusters stimulate subepithelial capillary growth in preparation for alveolarization, as recently proposed by Stahlman et al. (19851, based on observations in human lungs. Although most of the clusters occupying the junction between the conducting airway and respiratory zone in hamsters are unusual in being closely invested by capillaries (Hoyt et al., 1982a,b), they appear too late in development to have served such a trophic purpose. Once established in the terminal bronchioles, if these cell clusters exert any influence over formation of the respiratory zone, it must be a n inhibitory one, possibly acting to prevent excessive retrograde alveolarization. CONCLUSIONS Taken together, the morphological and quantitative studies on the development of small-granule cells in hamsters show that prenatal stages are compressed into a shorter period, both in terms of days and proportion of the gestational period (4 days, 25%) than is true of either the rat (8 days, 36%), rabbit (14 days, 44%), or man (30 weeks, 79%, according to Stahlman and Gray, 19841, yet development has not been compressed sufficiently to provide a system largely populated by mature cells a t the time of birth. ACKNOWLEDGMENTS This study was supported in part by NIH Research Grant HL-19379; a major portion of this work was submitted by Dr. Sarikas in partial fulfillment of the requirements for the Ph.D. degree a t Boston University. LITERATURE CITED Boyden, E.A. (1967) Notes on the development of the lung in infancy and childhood. Am. J. Anat., 121:749-762. Boyden, E.A., and D.H.Tompsett (1961) The postnatal growth of the lung in the dog. Acta Anat., 47t185-215. Boyden, E.A., and D.H.Tompsett (1965) The changing patterns in the developing lungs of infants. Acta Anat., 61:164-192. Hoyt, R.F., Jr., H.Feldman, and S.P. Sorokin (1982a) Neuroepithelial bodies (NEB) and solitary endocrine cells in the hamster lung. Exp. Lung Res., 3.299-311. Hoyt, R.F., Jr., S.P. Sorokin, and H. Feldman (1982b) Small-granule (neuro) endocrine cells in the infracardiac lobe of a hamster lung. Number, subtypes and distribution. Exp. Lung Res. 3:273-298. Sarikas, S.N., R.F. Hoyt, Jr., and S.P. Sorokin (1985) Ontogeny of small-granule APUD cells in hamster lung. A morphological study. Anat. Rec., 213t396-409. 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., 192:245-260. Stahlman, M.T., and M.E. Gray (1984) Ontogeny of neuroendocrine cells in human fetal lung. I. An electron microscopic study. Lab. Invest., 51:449-463. Stahlman, M.T., A.G. Kasselberg, D.N. Orth, and M.E. Gray (1985) Ontogeny of neuroendocrine-cells in human fetal lung.- 11. An immunohistochemical study. Lab. Invest., 5252-60. Tyler, W.S. (1983) Comparative subgross anatomy of lungs. Pleura, interlobular septa, and distal airways. Am. Rev. Resp. Dis., 128:S32-S36. Wells, L.J., and E.A. Boyden (1954) The development of the bronchopulmonary segments in human embryos of horizons XVII to XIX. Am. J. Anat.,??5:163-201.