Introduction

Top Abstract Introduction Materials and Methods Results Discussion References

Functional nerves to the airway smooth muscle have been demonstrated in the first trimester in fetal pig lung, where functionally mature airway smooth muscle contracts strongly to a variety of stimuli (1). Strong narrowing responses of the airways throughout the bronchial tree are evoked by electric field stimulation as well as by agonists such as acetylcholine and histamine. The responses to electric field stimulation are blocked by atropine, indicating that the airway smooth muscle is innervated by cholinergic nerves (2, 3) and by tetrodotoxin, indicating that action potentials are generated. Moreover, spontaneous narrowing of the airways has been reported in freshly excised human and pig lung from the first trimester onward (3, 4), and in cultures of lung explants from chick embryos (5), fetal guinea pig (6), and fetal mouse (7). Thus, although there is evidence of strong neurogenic control of fetal airway smooth muscle, there is little information regarding morphologic correlates responsible for it.

The neural tissue in the developing lung originates from neural crest cells that migrate through the mesenchyme to populate the future trachea, before its separation from the gut (8, 9). However, there is little information regarding the development of the innervation of the fetal lung itself. One study in the late first trimester of fetal pigs (10) has mapped the nerves to the airway smooth muscle immunohistochemically using antibodies to the pan-neuronal nerve markers, protein gene product (PGP) 9.5 and synaptic vesicle protein 2 (SV2). Both PGP 9.5 and SV2 immunoreactivity showed an abundance of neural tissue comprising forming ganglia and nerve trunks. The latter gave rise to a complex plexus of fiber bundles and fine processes that were observed lying over the smooth muscle of the airways. The antibody to PGP 9.5 revealed many ganglia at the junctions of the nerve trunks, particularly at the airway bifurcations, with multiple connections between the nerve trunks and the ganglia. SV2 revealed a varicosed nerve plexus to the muscle that extended to the growing tips of the airways.

In mammalian lung development, the following stages are commonly described (11, 12), namely the pseudoglandular, canalicular, and saccular stages. The aim of the present work was to characterize the progressive development of the nerves and ganglia in the lungs of pig fetuses during these stages. In the pig, the early stages of the anatomical development of the bronchial tree and pulmonary beds have been well documented (13) from the formation of the lung bud, an outgrowth of the foregut, to the completion of the bronchial tree. The primordial lung comprises an endodermally derived epithelial tube that is destined to become the future trachea, with two branches that arise from the first division and become the main stem bronchi. In common with many other mammals, these main bronchi extend along the length of the main lobes giving rise initially to lateral branches, in contrast to the human, where dichotomous branching occurs. In the pig there are five laterals, with the first lateral on the right branching off the trachea (12). At the outset, the forming airways appear as fine, translucent epithelial tubules with blind ends, often referred to as terminal sacs (14), that subsequently divide into future generations.

The nerve trunks in the late first trimester of fetal pig lungs contained unstained cell profiles (10). We therefore attempted to characterize these profiles by immunohistologic double and triple staining using the nerve markers PGP 9.5 and SV2, and the glial cell markers glial fibrillary acidic protein (GFAP) and S-100, in conjunction with the nucleic acid stain ethidium bromide. S-100 has been found useful as a means of demonstrating the supporting glial cells of pulmonary ganglia and the Schwann cells of peripheral nerves in the adult respiratory tract (15) and GFAP has been widely used as a glial cell marker in the peripheral nervous system (16).

In this study, whole mounts of the bronchial tree from fetal pigs in the early pseudoglandular phase (3.5 wk gestation, 0.6 g body weight) up to the postnatal stage (4 wk, 10 to 15 kg body weight) were stained with fluorescent antibodies and imaged using a confocal microscope. We report here an abundance of ganglia precursors, covering the airways of the early fetal bronchial tree, that develop progressively into mature ganglia in the postnatal lung.

	Materials and Methods

Top Abstract Introduction Materials and Methods Results Discussion References

Animals

Fetal pigs from Large White/Landrace hybrid sows were obtained from a local abattoir. The fetuses were removed from the uterus approximately 20 min after death of the sow, packed on ice, and transported back to the laboratory. The body weight and crown-to-rump length were measured before removal of the trachea and lungs. Fetuses weighed from 0.6 to 600 g with a crown-rump length from 14 to 240 mm. The gestation in the pig is ~ 116 d. Because the fetuses were obtained from the abattoir, the exact gestational age was not known but estimated according to published data (17, 18), relating age to fetus length and weight. The variation of crown-rump length and body weight within a litter increases with gestation (Figure 1a), however there is little variation until 7 wk of gestation (Figure 1b). The average fetus weight within a litter was used to estimate the gestational age according to Pomeroy (17), using the equation:

View larger version (12K): [in this window] [in a new window]   Figure 1.   (a) Variation in crown-rump length and body weight within litters of fetal pigs from the first trimester up to near-term. The error bars show the standard deviation of the variation within each litter. The variation in fetal weight increases with the stage of pregnancy. (b) There is less variation in crown-rump length and body weight within litters of fetal pigs during the pseudoglandular stage.

where W = fetal body weight in grams and t = gestational age in days. The numbers 0.1135 and 16.59 are constants. For each time point presented in the results, at least four fetal pigs from not less than two individual litters were studied.

Four postnatal pigs were obtained from Medina Agricultural Farm (Medina, Western Australia) and killed with a captive bolt in the animal house of the University of Western Australia. The postnatal pigs were 4 wk old and varied in body weight from 10 to 15 kg. This research was approved by the Animal Experimentation Ethics Committee of the University of Western Australia.

Immunohistochemistry

Antibodies (monoclonal, raised in mouse, and polyclonal, raised in rabbit) to PGP 9.5, a protein found exclusively in all nervous tissue (19), were obtained from UltraClone (Isle of Wight, UK). A monoclonal antibody (mAb) to SV2, a component of secretary vesicle membranes (20), was a kind gift from Dr. Kathleen Buckley (Harvard Medical School, Boston, MA). Polyclonal antibodies to GFAP and S-100 were obtained from Dako (Botany, NSW, Australia). A mAb to the 68-kD subunit of neurofilament was obtained from Amersham (Buckinghamshire, UK). The secondary antibodies (antimouse and antirabbit) were conjugated to fluorescein isothiocyanate (FITC) (Silenus, Parkville, Australia), Rhodamine Red (RITC) (Cappel, Organon Teknika Corp., West Chester, PA), or Cy5 (Amersham, Australia). A mAb to -actin (Sigma Chemical Co., St. Louis, MO) was used to identify smooth muscle, which gives identical staining compared with an antibody specific for smooth-muscle myosin (3, 10). A polyclonal antibody to choline acetyltransferase (ChAT) was a kind gift of Prof. M. Schemann (School of Veterinary Medicine, Hannover, Germany) (21). Control experiments to test for nonspecific staining due to autofluorescence or incomplete washout of the secondary antibody were carried out using nonimmune rabbit and mouse sera.

Excised fetal pig lungs were kept immersed in cold Krebs solution gassed with 95% O₂/5% CO₂. The parenchyma and vasculature were carefully teased away from the airways under a dissecting microscope, leaving the bronchial tree largely intact. In near-term and postnatal lungs, the first lateral tracheal branch was primarily studied. Whole mounts of fetal and postnatal pig lung were fixed overnight in 4% paraformaldehyde. The specimens were then cleared by washing in dimethyl sulfoxide for 3 × 10 min to permeabilize the membranes. After washing in phosphate-buffered saline (PBS), pH 7.4, for 2 × 10 min, nonspecific binding was blocked by an additional washing step in PBS that contained 1% bovine serum albumin. Then the primary antibodies (dilutions: PGP 9.5, 1/100; SV2, 1/40; GFAP, 1/500; S-100, 1/200; neurofilament, 1/100; actin, 1/250) were applied and the preparation was incubated overnight in a humidified chamber at 4°C. The samples were then washed several times with PBS over a 4-h period to remove unbound primary antibodies, and then reacted for 12 h at room temperature with the fluorochrome-labeled secondary antibodies FITC (diluted 1/100), RITC (diluted 1/50), and Cy5 (diluted 1/5). Ethidium bromide staining was performed before mounting by bathing the tissues for 2 min in a 0.1 µg/ml ethidium bromide. After further washing with PBS, the preparations were mounted in 90% glycerol containing p-phenylethylenediamine to reduce bleaching of the fluorochromes. Each specimen was mounted on a separate glass slide with individual branches spread out to prevent overlap. The coverslips were raised with custom-made Teflon rings in order to minimize compression of the airways.

Confocal Microscopy

Fluorescent images of the nerves and smooth muscle in the double-stained (FITC/RITC) whole mounts were obtained using a confocal laser scanning microscope (MRC 1000; Bio-Rad, Hemel Hempstead, UK) with COMOS software (version 7.0; Bio-Rad). The excitation wavelengths of the krypton/argon laser were 488 nm for FITC, 568 nm for RITC, and 645 nm for Cy5. The whole mounts were optically sectioned by scanning at increasing depths of focus (typically, in steps of 1 µm) to follow the path of the nerves in relation to the smooth muscle. Many fields of each whole mount were scanned, and multiple areas were selected for detailed study of the ganglia. Representative images of the ganglia at various gestational ages are presented in this manuscript. The maximum intensity of the corresponding pixels in each optical section was used to generate a single image (2-D projection) from a stack of images obtained at varying depths. The FITC image of the nerves was merged with that of the smooth muscle (RITC) to form a composite nerve/muscle image. Image processing (merging and montaging fields) was done with Adobe Photoshop 4.0 software. In multistaining experiments, single fields were scanned for each marker, then colorized and superimposed. The reconstruction of the primordial lung shown in Figure 2b was performed by scanning at high intensity to enable tracing the outline of the airways. Measurements of cell sizes were made from single optical sections. Typically, a ×60 objective was used that, in conjunction with the corresponding settings for laser power, iris width, and photomultiplier gain, allowed sections of less than 1-µm thickness to be obtained. All airway diameters refer to the external diameter of the smooth-muscle layer unless stated otherwise. Cell diameters refer to the largest diameter.

View larger version (67K): [in this window] [in a new window]   Figure 2.   (a) Primordial lung from a GW 3.5 fetal pig (0.6 g body weight) stained for smooth muscle with -actin (red). Actin stained not only the airway smooth muscle but also the pulmonary artery that ran along the side of the trachea, and divided to run adjacent to the main stem bronchi. The tissue was double stained with PGP 9.5, which did not reveal neural tissue but instead delineated the overall shape of the lung, due to trapping of the stain in the pleura. (b) A reconstruction of the bronchial tree shown in (a), showing the two main stem bronchi with the lateral branches labeled from 1 to 5. L1 branches off the right side of the trachea. The pulmonary artery runs down the length of each stem bronchus with branches off to the laterals. (c) Lung from a GW 3.5 fetal pig stained for nerves with SV2 (yellow) and counter-stained with PGP 9.5 (blue) to reveal the mesenchymal cap. The vagus nerve runs along the esophagus (both severed distally) and gives rise to an extensively branching network of fibers extending toward the terminal sacs of the developing airways.

	Results

Top Abstract Introduction Materials and Methods Results Discussion References

The development of the innervation in the fetal lung of the pig is presented using the traditional histologic stages of lung development, namely the pseudoglandular, canalicular, and saccular stages. These were determined using wax-embedded sections stained with hematoxylin and eosin as described previously (12, 14). The definition of these stages of lung development is based on the appearance of the airways and the developing air spaces in the parenchyma, and allows for considerable overlap between the stages. In the pig, they correspond very approximately to the first, second, and third trimesters of gestation. The most pronounced morphologic changes in the lung innervation occur within the pseudoglandular stage, which is the main focus of this developmental study.

The Pseudoglandular Stage (Gestational Wk 3-8)

The pseudoglandular stage (named because of the lung's glandular appearance) is the period of branching morphogenesis of the airways. In the pig, it lasts approximately until gestational week (GW) 8, at which time the fetuses have reached a body weight of ~ 50 g. The youngest fetuses that were available for investigation were only 0.6 g in body weight, and were estimated to be at GW ~ 3.5, which corresponds to the early pseudoglandular stage. The antibody to -actin revealed the smooth muscle surrounding the branching epithelial tubules and also the vascular smooth muscle of the pulmonary vessel. The overall shape of the lung was demonstrated by concurrently using an antibody to PGP 9.5 that highlighted the outline of the mesenchymal cap (Figure 2a). The two stem bronchi with the beginnings of the lateral branches were reconstructed by tracing their outline (Figure 2b) and have been named according to the nomenclature of Flint (13). At this time the primordial bronchial tree exhibits all five lateral airways branching from the two stem bronchi, with the first and second laterals dividing dichotomously. The pulmonary artery branches along with the airways. To stain neural tissue, an antibody to SV2 was used rather than PGP 9.5, which did not show neural structures at this stage of gestation. SV2 revealed branches originating from the vagus that divide to form a network of fine nerves to the epithelial tubules with fibers extending to the base of the terminal sacs (Figure 2c).

A video micrograph of a complete lung at GW 3.8 is shown in Figure 3a. The lung exhibits red patches, presumably due to hemorrhages from the pulmonary artery. Removal of the mesenchymal cap revealed the growing tips of the airways (the terminal sacs), which comprise a single layer of epithelial cells (Figure 3b). To see whether smooth muscle and nerves were present at this gestational age, the lung was stained for smooth muscle with -actin and for nerves with PGP 9.5. Figure 3c shows the distal end of the right stem bronchus covered in airway smooth muscle (red) accompanied by a pulmonary artery, with a branching nerve (green) running along the length of the epithelial tubule (Figure 3d). The smooth muscle extends to base of the terminal sacs. PGP 9.5 stained not only the nerves but also the epithelial tissue, thereby revealing the epithelial buds, which were not covered with airway smooth muscle (Figure 3e). As reported previously (10) and also observed in fetal human lung tissue (22), PGP 9.5 immunoreactivity is found in the undifferentiated epithelial cells of the terminal sacs.

View larger version (80K): [in this window] [in a new window]   Figure 3.   (a) Video micrograph of the lung of a GW 3.8 (1.5 g body weight) fetal pig in the early pseudoglandular phase. Hemorrhages of pulmonary vessels caused the red coloration of the tissue. (b) Video micrograph of the terminal sacs of the epithelial tubules after removing the mesenchymal cap. The width of these epithelial buds is ~ 100 µm. (c-e) A terminal airway of a GW 3.8 (1.5 g body weight) fetal pig stained with -actin for smooth muscle (red, c) and with PGP 9.5 for nerves (green, d). The composite image is shown in e. A nerve bundle runs along the length of the entire airway. The -actin also stained a pulmonary vessel that runs along the airway. The epithelial buds are immunoreactive to PGP 9.5 and are not yet surrounded by smooth muscle. (f, g) The right stem bronchus of a GW 4 (2.8 g body weight) fetal pig stained with PGP 9.5 for nerves and -actin for smooth muscle. ( f ) Two main nerves (arrows) run along the airway and branch onto the newly forming side branches. (g) The epithelial buds have a collar of smooth-muscle cells at their base, and divide again to form another generation of epithelial buds. -Actin stained the pulmonary artery, which gives off branches to the airway laterals. (h) A single optical section through a lobe of the lung of a GW 4 fetal pig stained with PGP 9.5 (blue) and SV2 (yellow). SV2-positive nerves run along the airways and terminate at the base of the epithelial buds that stain brightly with PGP 9.5. (i) Confocal image of the trachea and the first lateral of the bronchial tree of a GW 4 fetal pig with the nerves stained with SV2. The trachea (proximal end on right top corner) is covered by a layer of neural tissue that consists of flat ganglia-like cell accumulations and nerve trunks that interconnect these ganglionic precursors. A part of the large vagus nerve (severed) is visible at the top of the image, with nerves diverging from the vagus nerve onto the trachea. (j) Higher-power view of trachea of the GW 4 fetal pig shown in h. The nerves run along the patches of interconnected ganglionic precursors and give rise to a dense network of varicose fibers that cover the surface of the airway wall. (k) Triple-stained, high-power micrograph showing a single optical section through the cellular thickenings of the nerve trunks that cover the adventitia of a GW 4 fetal pig. The nerves are stained with PGP 9.5 (blue) and SV2 (yellow). Nuclei of all the cell types present were stained with ethidium bromide (red). The SV2 staining reveals fibers both in the large nerve trunks (NT) and within the ganglia-like cell accumulations (C). Some large nerve trunks are free of cells (NT). (l) Montage of a near-complete bronchial tree of a GW 4.5 (4 g body weight) fetal pig stained for nerves with PGP 9.5 (blue). The antibody reveals the nerves running from the trachea to the periphery and also the ganglia, which are largest in the trachea. The epithelial cells in the terminal sacs are also stained. The airway smooth muscle, stained for -actin (red) is present from the trachea to the distal airways, terminating at the base of the terminal sacs. The -actin also stained the smooth muscle of some vessels of the pulmonary vasculature that was not completely removed during dissection.

At a slightly more advanced stage (GW 4), the stem bronchi and laterals have lengthened and are supplied by pulmonary artery branches. On some of the growing laterals, buds have formed that will become secondary airways. Two main nerves run along the stem bronchi with bundles branching to the laterals (Figures 3f and 3g). SV2 was used to show detail of the nerves that supply the branching epithelial tubules (Figure 3h). Two nerve bundles were present that followed the contour of the tubules at a distance of ~ 30 µm from the wall, giving rise to a fine network of fibers descending toward the wall. In double-staining experiments using PGP 9.5 and SV2, the nerves stained by PGP 9.5 were obscured by the stronger staining of SV2.

A prominent feature of the more mature parts of the bronchial tree at this early stage of development (GW 4) is the presence of forming ganglia in the neural network that covers the proximal epithelial tubules. In the trachea, the first lateral, and the proximal parts of both main stem bronchi, ganglionic precursors are found (Figures 3i and 3j). These comprise patches of cells as large as 300 µm in diameter and 30 µm in thickness that are immunoreactive to PGP 9.5 and SV2. The trachea is supplied with nerves that branch from the vagus (often severed during dissection). The ganglionic precursors that cover the trachea stain less strongly with SV2 than do the nerve trunks that interconnect them. Ethidium bromide was used concurrently with PGP 9.5 and SV2 to visualize the nuclei in the cells of the ganglionic precursors. The nuclei within the PGP 9.5-positive areas were 8.1 ± 0.3 µm (n = 30) in diameter and are likely to belong to immature neurons. SV2 revealed fibers around the cell profiles and within the nerve trunks (Figure 3k).

At GW 4.5, further morphologic differentiation of ganglia and nerve trunks occurs concurrently with elongation of the main stem bronchi and the formation of new generations of branching epithelial tubules. An abundance of epithelial buds was revealed by PGP 9.5 (Figure 3l). The length of the bronchial tree, measured from the first lateral branch L1 to the most distal end of the right stem bronchus, is 5 mm, compared with 2.5 mm for the same distance in GW 3.5 fetal lung. The smooth-muscle layer is well developed, with a single layer of muscle cells arranged cylindrically around the epithelial tubules. The nerves are seen within the transparent extracellular matrix that surrounds the airways, and lie at a distance of up to 300 µm from the airway smooth muscle at the trachea. This distance is probably an overestimation resulting from the flattening of this relatively thick tissue when it was mounted on a slide. The network of nervous tissue covering the airways contains many developing ganglia (diameter > 200 µm). These rather flat structures (thickness 20 to 30 µm) cover the trachea and the large bronchi. The typical spindle-shaped cell profiles indicative of Schwann cells were not yet present in the nerve trunks. Staining with S-100 to identify Schwann cells was not feasible due to high background fluorescence. However, chondrocytes in the developing tracheal cartilage plates of a GW 4.5 fetal pig were strongly stained by S-100 (not shown), a marker that has been shown to stain chondrocytes in adult human tissues (23).

At GW 5.5, ganglia had condensed and were abundant at nerve trunk intersections (Figure 4a). They contained numerous spherical neurons of 8.9 ± 0.2 (n = 20) µm diameter, whereas nerve trunks exhibited cellular profiles that resembled spindle-shaped Schwann cells of 16 to 20 µm in length (Figure 4b). In GW 7 pigs, S-100 revealed, albeit weakly, elongated Schwann cell profiles in the large nerve trunks and some in the ganglia (Figure 5). In fetal pigs younger than GW 7, the use of S-100 to identify Schwann cells was hindered by the background fluorescence. The glial marker GFAP stained the developing sheath surrounding the nerve fibers and ganglia. GFAP also stained chondrocytes (not shown), as previously reported in human respiratory tract cartilages (24). Figure 6 is a single optical section cut through a ganglion of a GW 7 fetal pig, stained with PGP 9.5 and GFAP. PGP 9.5 stains predominantly the cytoplasmic rims of the neurons while the nuclear regions appear dark (unstained). Glial processes within the ganglia form sheaths around some but not all neurons.

View larger version (142K): [in this window] [in a new window]   View larger version (135K): [in this window] [in a new window]   Figure 4.   (a) The adventitia of a GW 5.5 (15 g body weight) fetal pig, stained with PGP 9.5. The plethora of interconnected nerve trunks and maturing ganglia gives rise to a dense plexus of fine fibers covering the airways. Arrow identifies area shown in b. (b) A magnified view of the area marked by the arrow in a, showing neural tissue stained with SV2. The cellular appearance of the forming ganglia is evident, with spherically shaped cells in the ganglia precursors and larger, more spindle-shaped cell profiles within the large nerve trunk (arrowheads).

View larger version (179K): [in this window] [in a new window]   Figure 5.   Schwann cells in the large nerve trunks and ganglia in the airways of a fetal pig at GW 7, stained with S-100 and PGP 9.5. Neither S-100 nor PGP 9.5 yielded clear images because of high background fluorescence; however, the staining is of sufficient quality to allow Schwann cells (arrows) to be identified, both in the large nerve trunks (A) and the ganglia (B).

View larger version (167K): [in this window] [in a new window]   Figure 6.   Single optical section through a ganglion from a GW 6.8 (43 g body weight) fetal pig stained with PGP 9.5 and GFAP. PGP 9.5 immunoreactivity is shown in the gray midtones; GFAP immunoreactivity is shown in white. PGP 9.5 stained the cytoplasmic rims of the neurons (~ 10 µm) but not the nuclear region (~ 8 µm). The GFAP-positive processes are found in the nerve trunk that is partly visible in the top left corner and within the ganglion, enveloping some of the neurons.

Thus, the fetal lung in early gestation (pseudoglandular stage) has an extensive innervation at the outset of its development. Ganglionic precursors cover the proximal airways and progressively condense into ganglia predominantly located at nerve trunk intersections. The ganglia are supplied with nerves branching from the vagus, giving rise to a dense neural network throughout the length of the bronchial tree.

The Canalicular Stage (GW 7-13)

In the canalicular stage there is a marked increase in the vascularization, with a multitude of capillaries "canalizing" the future lung parenchyma. In the pig, this stage ranges from GW 7 to 13, with body weights of 40 to 600 g. The most marked morphologic change of the innervation is the increasing spatial separation of the ganglia. The ganglia, previously appearing like nodes at nerve interconnections, gradually become attached laterally to the trunks, especially in the case of the large nerve trunks that run longitudinally along the airways. Often the ganglia are spherically shaped and separated by a thin stem to the trunk, and give rise to a number of smaller nerve trunks.

At GW 8, ganglia were present particularly at the junctions between nerve trunks and at the airway bifurcations (Figure 7a). Trunks radiated from the ganglia, dividing to form a network of smaller bundles covering the smooth muscle (Figure 7b). The -actin stained the airway smooth muscle and the bronchial vasculature. The prominent staining of bronchial blood vessels is typical for the canalicular stage and was not seen in the pseudoglandular stage. Bronchial arterioles are present that run next to the nerve trunks and branch extensively around the ganglia (Figure 7c). The large nerve trunks with associated ganglia comprise many bundles of fibers that are separated from each other and appear folded. Figure 7d shows a single optical section through a ganglion that is separated from a large nerve trunk by a small stem (not seen in this optical section). The neurons exhibit up to three nucleoli that appear as round, unstained holes when stained with PGP 9.5. Dark profiles were seen between the nerve bundles of the large trunk and are most likely Schwann cells.

View larger version (100K): [in this window] [in a new window]   Figure 7.   (a) Montage of the left distal stem bronchus from a GW 8 (90 g body weight) fetal pig stained with PGP 9.5 for nerves (green) and -actin for smooth muscle (red). The numerous ganglia are now more spherical in shape. (b) Magnified view of a field of the bronchus shown in a (circled area). A large nerve trunk that is accompanied by blood vessels runs longitudinally along the airway. A ganglion is situated to the right of the trunk, with nerve bundles radiating out over the smooth muscle. These nerve bundles give rise to a dense network of thinner fibers that descend to the airway smooth muscle. (c) High-power view of a single optical section through the ganglion shown in the center of b. The trunk shows dark cellular profiles that are most likely Schwann cells. The vasculature exhibits extensive branching around the ganglion. The neurons on the perimeter of the ganglion show cytoplasmic and nuclear PGP 9.5 immunoreactivity, whereas the cells in the center of the ganglion exhibit staining of a cytoplasmic rim only, with little or no nuclear staining. (d) High-power view of a single optical section through a ganglion from the distal stem bronchus from a GW 8.5 (110 g body weight) fetal pig, stained with PGP 9.5. The cell bodies are, on average, 11 µm in diameter. Nucleoli can be identified as dark holes of 2 to 3 µm diameter within the nuclear regions of the neurons. The inside of the ganglion shows many dark exclusions that are most likely profiles of blood vessels and satellite cells. The adjacent nerve trunk comprises many folded bundles of fibers that are separated from each other. (e) The adventitia from the proximal stem bronchus of a midterm pig (GW 8.5), stained with PGP 9.5 for nerves (green) and -actin for smooth muscle (red). A network of nerve trunks with large ganglia at their nodes overlies the smooth muscle. Some bronchial vessels are seen running next to the larger nerve trunks. On the right, a main nerve trunk runs longitudinally along the airway axis with many associated ganglia (often connected by short stems) that give rise to many smaller nerve bundles. ( f ) A large ganglion from the center of the field of the proximal stem bronchus shown in e. The ganglion, stained with PGP 9.5, shows multiple nerve-trunk connections. Cell bodies show unstained nucleoli. (g) Triple-stained micrograph of a ganglion from a GW 9.3 (170 g body weight) fetal pig, stained with PGP 9.5 (blue) and GFAP (yellow). The vasculature (red) was revealed by strong autofluorescence of residual red blood cells. This single section demonstrates an abundance of GFAP-positive processes that form sheaths around the nerve fibers in the trunks and partly envelop neurons within the ganglion. The nerves are accompanied by blood vessels with a network of vessels around the ganglion.

The central airways (inner diameter 2 mm) at GW 8.5 contain many ganglia that form nodes at nerve junctions. These ganglia vary in size and shape and are spaced 200 to 500 µm apart (Figure 7e). The large ganglia (diameter 100 µm) comprise more than 100 neurons. A high-power view of one of these ganglia is presented in Figure 7f. The ganglion is approximately 120 µm in length and is estimated to contain in excess of 200 neurons, with average diameters of 11 µm.

The microcirculation of the ganglia was often difficult to image satisfactorily with -actin because the much stronger signal from the airway smooth muscle masked that from the vascular smooth muscle. However, small blood vessels that retained red blood cells in the lumen could be imaged effectively due to their autofluorescence. Blood vessels were observed to run along nerve trunks and form a complex capillary network around ganglia. Figure 7g shows a ganglion and nerve trunks from a GW 9.3 fetal pig stained with PGP 9.5 and GFAP, demonstrating the vascularization of the ganglia and the progressive glial ensheathing of axons and neurons within the ganglia.

The majority of neurons in the ganglia present in the adventitial wall of the bronchial tree of midterm pigs were cholinergic. At GW 10, all neurons in the ganglia observed in both trachea and distal airways appeared immunoreactive with an antibody to ChAT, which is a specific marker for cholinergic nerves (21). This antibody has been used to examine the occurrence of cholinergic neurons in the ferret tracheal plexus (25). We found that the intensity of ChAT staining varied widely, making it difficult to gauge whether 100% of the neurons were cholinergic. ChAT staining was similar in distribution to that of PGP 9.5, sharply outlining the nerve fibers in the trunks and the neurons in the ganglia (Figure 8). However, many of the fine nerves in close proximity to the smooth muscle were not revealed by ChAT (Figure 9).

View larger version (143K): [in this window] [in a new window]   Figure 8.   Cholinergic nerve trunks and ganglia in a fetal pig at GW 10, stained with an antibody to ChAT. The smooth muscle is concurrently stained with -actin. (Inset) Single optical section through the ganglion at higher power. ChAT reveals the cytoplasm, including the axon hillock but not the nucleus.

View larger version (131K): [in this window] [in a new window]   Figure 9.   A distal airway from GW 10 fetal pig, stained with ChAT and SV2. A ganglion on a nerve trunk radiates smaller bundles. (a) ChAT does not stain the fine varicosed fibers distributed over the airway that stain brightly with SV2 (in b). (b) SV2 stains many finer nerve bundles branching off the large nerve trunk, forming a network of varicosed fibers. The center of the ganglion also shows strong SV2 immunoreactivity.

Thus, by midterm, when the lung development is in the canalicular stage, the adventitia of the airway wall contains a population of maturing ganglia that overlie the airway smooth muscle and are accompanied by a complex bronchial vascular network.

The Saccular Stage (GW 12-16.5)

The saccular stage derives its name from the typical terminal clusters of widened air spaces (called saccules) formed by the peripheral airways. In pigs, this stage commences approximately around GW 12 and lasts until birth (GW 16.5), with corresponding body weights of 450 g and 1.1 kg, respectively. In these more mature animals, most ganglia studied were from the first lateral tracheal branch because the greatly increased size of the airways limited the use of the whole mount technique for the more central airways. The airways were stained for nerves with antibodies to PGP 9.5 and neurofilament. In younger fetuses, neurofilament stains only a subpopulation of nerves (3), which limits its usefulness. However, in the advanced fetal stages, the staining pattern of neurofilament reveals more detail (sharper cellular outlines) than PGP 9.5. Figure 10a shows a proximal ganglion from a GW 13 fetal pig, which measures ~ 250 µm in diameter and reveals both perikarya and axons after staining for neurofilament. A smaller ganglion at a bifurcation of a nerve trunk, observed in a more distal region, is presented in Figure 10b. The perikarya appear to be located mainly in the periphery of the ganglia that contain mainly neuritic structures in the center (Figure 10c). Examination of consecutive confocal sections of the ganglion presented in Figure 10c showed the neurons (25 to 30 µm diameter) to possess one major axon, and therefore to correspond to Dogiel type-I neurons. They contain nuclei that appear as spherical, unstained holes of ~ 13 µm diameter. A single section through a smaller, distal ganglion (Figure 10d) indicates that some fibers in the nerve trunk connect directly to neurons in the ganglion. Progressive sectioning through the ganglia along the z-axis revealed that some neurons show strong PGP 9.5 staining of the nucleus, whereas some neurons exhibit only a faint homogenous staining throughout the perikaryon (Figure 11).

View larger version (100K): [in this window] [in a new window]   Figure 10.   Ganglia from the first lateral tracheal branch of a GW 13 (600 g body weight) fetal pig, stained with an antibody to neurofilament. (a) A ganglion of ~ 250 µm diameter, located on a proximal branch of the first lateral. (b) A ganglion of ~ 150 µm diameter, located on a distal branch of the first lateral. (c) High-power view of the ganglion shown in a. The neurofilament is not present in the nuclei of the neurons. The axon hillocks are stained strongly and reveal that the inside of the ganglion consists mostly of nerve fibers. (d) High-power view of the ganglion shown in b. The image was taken at 15-µm depth from the surface of the ganglion. The nerve trunk consists of many distinct bundles. Some fibers pass into the ganglion.

View larger version (87K): [in this window] [in a new window]   Figure 11.   A single optical section through a ganglion from a GW 13 (600 g body weight) fetal pig stained for PGP 9.5, GFAP, and ethidium bromide. The ganglion is 120 µm in its long axis, and the neurons measure up to 40 µm in diameter. (a) Three neurons in the top left exhibit very bright PGP 9.5 staining that is restricted to the nuclear region. Another adjacent neuron shows PGP 9.5-immunoreactivity throughout the neuron. (b) Many GFAP-positive glial cells are present with processes completely enveloping some neurons. Other cells within the ganglion are not GFAP-positive. (c) Ethidium bromide stained the nuclei, with bright nucleoli that appear as unstained holes in a and b. Many nuclei of smaller supporting cells are present.

The Postnatal Lung

The innervation of postnatal pig lungs has been described previously (10). In the distal airways of 4-wk-old suckling pigs, ganglia were not observed. PGP 9.5 revealed two major nerve trunks that run along the length of the airways, giving rise to a very dense network of interconnected thin fibers covering the surface of the smooth muscle. In this study, ChAT revealed nerves that overlie the smooth-muscle layer in a manner similar to that of PGP 9.5 and SV2 in the earlier study (10), but did not detect the fine varicosed fibers in the distal airways (Figure 12). Ganglia were examined from the first lateral tracheal branch (L1), primarily from airways of 700 to 1,500 µm inner diameter. The tightly packed ganglia contain neurons of variable shape which are between 30 and 40 µm at their greatest width. Some neurons exhibit large nucleoli (diameter up to 5 µm) (Figure 13a). The extent of GFAP-positive sheaths around the neurons and around bundles of fibers in the nerve trunks is greatly increased compared with the prenatal stages (Figure 13b).

View larger version (150K): [in this window] [in a new window]   Figure 12.   A distal airway of a postnatal pig (4 wk) stained for ChAT and -actin. ChAT-positive nerves run along the airway, giving off smaller branches that are predominantly oriented parallel with the smooth-muscle bundles.

View larger version (92K): [in this window] [in a new window]   Figure 13.   (a) A ganglion from the first lateral tracheal branch of a postnatal pig (stained for PGP 9.5) contains many tightly packed neurons and measures 300 µm in its longest axis. Some neurons contain large nucleoli of ~ 5 µm diameter. (b) Another ganglion, stained for PGP 9.5 (gray) and GFAP (white). GFAP-positive processes and sheaths are found in the nerve trunk and around the neurons of the ganglion. The nuclei of the neurons appear as round, unstained discs of ~ 17 µm diameter. Satellite cells of < 10 µm diameter are present.

	Discussion

Top Abstract Introduction Materials and Methods Results Discussion References

The present study describes the development of ganglia in the fetal pig lung. In the youngest lungs, the ganglia start out as large, flat patches of neural tissue that presumably originate from neural crest cells (8). These cover the primordial trachea and proximal regions of the bronchial tree. Nerve trunks are seen between these patches that appear to have originated from the vagus. These nerves exhibit many side branches that distribute numerous fibers over the trachea and major bronchi. The distance between the ganglionic precursors progressively increases with the radial and longitudinal growth of the airways. During the first trimester (pseudoglandular stage), the ganglia condense at the interconnections of nerve trunks; later, by midterm (canalicular stage), they gradually become attached laterally to the large nerve trunks, some becoming separated by a short thin stem from the trunks. The shape of the ganglia becomes more spherical, and toward term (saccular stage), the ganglia exhibit a micelle-like arrangement of the neurons, with the soma pointing outward while the axons point to the center of the ganglion. The formation of tightly packed ganglia progresses into the postnatal period.

The nerves and ganglia were immunohistochemically stained with several nerve markers. The pan-neuronal nerve marker PGP 9.5 was used successfully to stain neural tissue in all but the youngest tissue of the pseudoglandular stage. The earliest lungs examined (GW 3.6, 0.6 g body weight) were imaged as whole mounts with an intact mesenchymal cap. Here, the PGP 9.5 antibody appeared unable to penetrate, but the trapped fluorophor highlighted the overall shape of the lung (Figures 2a and 2c). After dissecting the epithelial tubules from the surrounding pleura, PGP 9.5 also stained the epithelial cells, a feature that was used to image the outline of distal tubules and terminal sacs (Figures 3d and 3f). In more mature ganglia, including the postnatal period, PGP 9.5 staining of the neurons revealed variable staining of the nuclear region, with some nuclei showing brighter fluorescence than the perikaryon and vice versa (Figure 11). The varying levels of PGP 9.5 immunoreactivity may indicate different types of neurons.

The antibody to SV2 stained the nerve trunks and varicose fibers throughout the fetal development. SV2 proved especially useful in mapping nerves early in gestation, as the presence of SV2 in this period is not limited to varicosities but is found throughout the length of the fibers, most likely because of this protein's high levels of axonal transport. SV2 also showed the outlines of cells due to staining of SV2-immunoreactive neuritic structures that surround the neurons in the ganglia. Postnatally, SV2 was chiefly restricted to fine nerve fibers, as has been reported (10). Within near-term and postnatal ganglia, the SV2 immunoreactivity was strongest in the center of the ganglia, suggesting the presence of many synapses. An antibody to neurofilament has previously been used in the early stages of development (3) and revealed a subset of nerves that were sufficiently mature to stain positively for neurofilament. By midterm the neurofilament protein had become abundant in the perikaryon, axon hillock, and axons, making neurofilament a useful anatomical marker to identify the structure and arrangement of neurons in ganglia.

In the pseudoglandular stage, the antibody to ChAT did not show positive staining. From midterm onward, ChAT exhibited a staining pattern of the nerves that was similar to PGP 9.5; however, as double staining experiments with PGP 9.5 and SV2 revealed, it did not stain the fine varicosed fibers. This may be due to limited amounts of the enzyme that are below detection in these fibers. In adult guinea pigs, ChAT-immunoreactive nerve fibers were localized to the smooth muscle throughout the conducting airways. All nerve cell bodies in the ganglia intrinsic to the trachea and bronchi of the adult guinea pig displayed a cholinergic phenotype (26). In the fetal pig, the ChAT staining of neurons within the ganglia varied from threshold to bright, which may indicate varying concentrations of the enzyme in the cells. This variability and low sensitivity in thin nerve fibers limits its suitability to conclusively stain all cholinergic structures. Recently, vesicular acetylcholine transporter protein has been used successfully as a cholinergic marker (27, 28) but as yet there is no antibody available for use in pigs.

An abundance of Schwann cells in the nerve trunks was demonstrated by staining with antibodies to GFAP and S-100. GFAP in the autonomic nervous system is specific for non-myelin producing Schwann cells (29). In fetal and postnatal pig lungs, the GFAP immunofluorescence characteristically surrounded the unstained neuronal cell bodies, as has been shown similarly in the enteric nervous system (16). The GFAP-positive Schwann cells within the nerve trunks had tapering processes that formed extensive thin membranous sheaths. From the first trimester (pseudoglandular stage), these formed around bundles of nerve fibers in trunks and also enveloped the ganglia. In postnatal airways, GFAP sheaths were present around most thin nerve fibers (data not shown). From the end of the pseudoglandular stage onward (GW 7), the Schwann cell marker S-100 similarly revealed Schwann cells in the nerve trunks, staining the nucleus and cytoplasm of the cell and short processes of the forming sheath extending along the axons.

The size and the shape of ganglia and their neurons change during fetal development. Initially, in the pseudoglandular stage, the large nucleus occupies most of the cell, leaving only a small cytoplasmic rim of ~ 8 µm diameter. From midterm onward, the soma assume a tadpole-like shape and are assembled in the ganglia with their hillocks directing the axons inward. The neurons observed in the airway ganglia appear to belong to the Dogiel type-I class, as they generally exhibit one prominent axon that stains strongly for PGP 9.5 and neurofilament. The ganglia present in the near-term and postnatal pigs comprise neurons that range from 30 to 40 µm in diameter. These ganglia have been shown to be functionally mature at birth by directly stimulating the nicotinic receptors of the cholinergic intramural ganglia with dimethyl-phenyl-piperazinium (30). In the adult ferret trachea, two different sizes of neurons were identified using histochemical staining for acetylcholinesterase (31). Neurons in the longitudinal trunk ganglia averaged 34 µm in diameter, whereas neurons in the superficial plexus averaged 24 µm. In other species, more uniformly sized neurons are reported; e.g., neurons in the trachea of young mice form only a single size population of 14 to 19 µm in diameter that increased with age to average 24 µm (32).

We conclude that the primordial lung is invested with an abundance of neural tissue that persists into postnatal life. The ganglia in early gestation are likely to have their origin from neural crest cells, which migrated into the foregut before lung bud formation, where they are presumably scattered among the epithelial cells that comprise the endodermal-derived tubule. It appears that the youngest fetal lungs in the current experiments were in a developmental stage that delineates the end of migration and the beginning of differentiation of the neural precursors. In future experiments, the use of specific markers for these neural crest cells should assist in determining the stages of their lineage. HNK-1 has been used for this purpose (33) and more recently ret proto-oncogene product (34, 35) and Mash-1 (36).