Spatial and temporal distribution of nerves, ganglia, and smooth muscle during the early pseudoglandular stage of fetal mouse lung development



Neural tissue and smooth muscle appear early in the developing fetal lung, but little is known of their origin and subsequent distribution. To investigate the spatial and temporal distribution of nerves, ganglia, and airway smooth muscle during the early pseudoglandular stage, fetal mouse lungs at embryonic days (E) 11 to 14 were immunostained as whole-mounts and imaged by confocal microscopy. At E11, the primordial lung consisted of the future trachea and two budding epithelial tubules that were covered in smooth muscle to the base of the growing buds. The vagus and processes entering the lung were positive for the neural markers PGP 9.5 (protein gene product 9.5) and synapsin but no neurons were stained at this stage. An antibody to p75NTR revealed neural crest cells on the future trachea as well as in the vagus and in processes extending from the vagus to the lung. This finding indicates that even though neuronal precursors are already present at this stage, they are still migrating into the lung. By E12, neural tissue was abundant in the proximal part of the lung and nerves followed the smooth muscle-covered tubules to the base of the growing buds. At E13 and E14, a neural network of interconnected ganglia, innervated by the vagus, covered the trachea. The postganglionic nerves mainly followed the smooth muscle-covered tubules, but some extended out into the mesenchyme beyond the epithelial buds. Furthermore, we show in a model of cultured lung explants that neural tissue and smooth muscle persist and continue to grow and differentiate in vitro. By using fluorescent markers and confocal microscopy, we present the developing lung as a dynamic structure with smooth muscle and neural tissue in a prime position to influence growth and development. © 2001 Wiley-Liss, Inc.


The mammalian lung develops from an endodermal outgrowth of the foregut. Two epithelial buds invade the surrounding mesenchyme, form tubules, and continue to elongate and branch until the entire bronchial tree is formed (Spooner and Wessells, 1970; Ten Have-Opbroek, 1981). Mesenchymal-epithelial interactions are known to play a significant role in lung morphogenesis (Slavkin et al., 1984; Warburton et al., 1999), but there is also evidence that other tissues such as smooth muscle and nerves are involved. Recent confocal microscopic studies in the fetal pig and human lung have shown that the developing tubules are surrounded by smooth muscle and ensheathed in a network of newly forming nervous tissue (Sparrow et al., 1999; Weichselbaum and Sparrow, 1999; Weichselbaum et al., 1996). In general, two large nerve trunks lying in the adventitial surface of the airway wall run the length of the bronchial tree. They give rise to a network of bundles, with fine varicose fibers covering the airway smooth muscle (ASM) from the trachea to the growing tips of the airways. Numerous ganglia form at the junction of the nerve trunks in the trachea, and smaller ganglia interconnected by nerve bundles extend distally to the subsegmental airways. The maturation of the nerve trunks and ganglia occurs centrifugally in the bronchial tree, and progressively during gestation based on morphologic criteria, approaching postnatal characteristics toward the end of the canalicular stage. Histologic sections have also shown nerves, ganglia, and neuroendocrine cells in developing lungs of fetal rabbits (Hung, 1980) and rats (Morikawa et al., 1978). However, the early distribution of ASM and nerves has not been investigated and the origin of the neural tissue in the lung is unknown. It has been proposed that the same population of neural crest cells that populate the gastrointestinal tract will also migrate into the lung (Dey and Hung, 1997).

Little is known about the structural relationship between nerves and smooth muscle during early lung development. Sparrow et al. (1995) have shown that neural stimulation elicits contraction of ASM in first-trimester fetal pigs. The ASM is functionally mature because it also responds by contracting to agonists such as acetylcholine, histamine, and substance P, and by relaxing to β-adrenoreceptor agents (Sparrow et al., 1994). Fetal airways also narrow and relax spontaneously in freshly excised first trimester human (McCray, 1993) and pig lung (Sparrow et al., 1995) and in cultures of fetal lung explants from chick embryos (Lewis, 1924), guinea pigs (Schopper, 1935), and mice (Roman, 1995; Schittny et al., 2000). Airway smooth muscle may function as a mechanical prerequisite for lung growth by generating a positive pressure in the liquid-filled tubules. Studies have shown that a maintained pressure in the lung liquid is necessary for normal lung growth. If the pressure is reduced by draining of its liquid contents, the lung becomes hypoplastic (Harding and Hooper, 1996), while tracheal occlusion, which increases the pressure, causes increased lung growth (De Paepe et al., 1998; Kitano et al., 2000). Airway smooth muscle tone and spontaneous contractions could act as a stimulus for lung growth by means of mechanotransduction by inducing the expression of developmentally important genes (Cilley et al., 2000) and the production of growth factors (Liu et al., 1995; Schittny et al., 2000).

The objective of this study was to characterize the emergence of neural tissue and ASM shortly after the first lung buds form and to follow their development during the early pseudoglandular stage. This provides a morphologic basis for understanding the functional observations that have been previously reported (Blewett et al., 1996; Schittny et al., 2000; Sparrow et al., 1995). Also, we wished to see whether neural tissue and ASM continue to grow and mature in cultured lung explants, so that an in vitro model could be developed in which the differentiation of neurons and the interaction between nerves and smooth muscle can be investigated. Airway smooth muscle was detected by using antibodies to the smooth muscle-specific proteins α-actin and calponin, whereas antibodies to PGP 9.5 (protein gene product 9.5) and synapsin (a synaptic vesicle protein) were used to detect neural tissue. PGP 9.5 is a pan-neuronal protein (Day and Thompson, 1990), which is also present in undifferentiated epithelial cells (Haley et al., 1997), making it possible to visualize nerves and epithelial tubules concurrently. p75NTR (the low affinity neurotrophin receptor) was used as a marker of migrating neural crest cells because this receptor has been shown to be present in most or possibly all neural crest-derived cells of the developing enteric nervous system (Gershon, 1999; Young et al., 1999). Fluorescently labeled secondary antibodies were used to localize and follow the targeted tissues daily from embryonic day 11 to 14 (E11–E14) and in the left lobe (E12) after 5 days in culture. The small size of the mouse lung allowed intact, whole lungs to be imaged as whole-mounts so that the structural relationship between epithelial tubules, mesenchyme, smooth muscle, and nerves could be revealed in situ by using the confocal microscope.


Growth and Development of the Fetal Mouse Lung

To provide a basis for normal lung development, we recorded the size of the fetus as well as the developmental maturity of the lung at each day of gestation. Fetal mice at E11 to E15 were weighed and their crown-rump length measured. At E11, the average weight (± standard deviation) was 46 ± 8.3 mg and had increased 11-fold to 523 ± 37.5 mg by E15, whereas the crown rump length increased 2-fold from 6.7 ± 0.62 mm (E11) to 14.6 ± 0.79 mm (E15) (Fig. 1A). The left lobe of the lung, the largest of the five lobes, was chosen for detailed analysis. At the early stages of development the lobe is a flat structure so we used the increase in area as a measure of lung growth and the number of epithelial buds around the periphery as an index of development (Fig. 1B). The area increased from 0.19 ± 0.03 mm2 to 3.58 ± 0.48 mm2, while the number of peripheral buds increased from 2.8 ± 0.6 to 122 ± 18.6 between E11 and E15 (Fig. 1C). The rate of formation of peripheral buds increased daily during the entire period, while the lobe area showed a rate increase only during E11 to E14. Both fetal size and the developmental maturity of the lung changed strikingly from day to day during the early pseudoglandular stage, whereas there was very little variability between animals of the same age.

Figure 1.

Growth parameters for fetal mice (A) and left lung lobes (C). The values are presented as means ± SD. Error bars are not visible if they are smaller than the size of the symbols. B: The left lobe (ventral view) of a fetal mouse lung at embryonic day (E) 13, showing where the area is measured and how the peripheral buds are counted.

Origin of Neural Tissue in the Fetal Mouse Lung

The first two lung buds begin to evaginate from the foregut at E9.5 and by E11 (the earliest stage investigated) the lung consists of the future trachea and two main stem epithelial tubules with buds developing laterally (Fig. 2A). To detect neural tissue, whole-mounts of lungs were immunochemically stained for PGP 9.5, synapsin, and p75NTR. PGP 9.5 has been used throughout this study to reveal neurons and their processes, but at E11, neurons have not yet differentiated so we used an antibody to p75NTR to detect migrating neural crest cells. Figure 2B is an overview showing a confocal projection of the dorsal aspect of an E11 fetal lung, stomach, and intestine, stained for synapsin. The vagal nerves followed the esophagus and gave rise to nerve fibers entering the stomach and lung. Likewise, the antibody to PGP 9.5 stained the vagus and the processes entering the lung (Fig. 2C). The antibody to p75NTR also stained the vagus as well as a plexus of neural crest-derived cells and processes on the trachea (Fig. 2D). An optical section through the cells on the trachea revealed staining of the cell membranes (Fig. 2E, showing enlargement of right hand box in 2D). The vagus also displayed immunoreactivity to p75NTR with axons and cell membranes positive (Fig. 2F, central box in 2D). Nerve trunks originating in the vagus extended into the lung toward the tubules. These processes contained cells positive for p75NTR (Fig. 2G, left hand box in 2D), indicating that neural crest cells are still migrating from the vagus into the lung at E11.

Figure 2.

A: A video photomicrograph showing the ventral view of a fetal lung at embryonic day (E) 11. The budding epithelial tubules are embedded in mesenchyme. t, trachea; br, bronchi; b, buds. B: Confocal projection of the dorsal view of fetal lung, stomach (s) and gut (g) showing the synapsin-positive vagal nerves (v) on each side of the esophagus. The vagus (v) sends out branches to the lobes of the lung and to the stomach. C: Protein gene product 9.5 (PGP 9.5) –positive nerves entering the lung from the vagus (v). D: p75NTR staining in E11 lung. The vagus (v), cells and nerves on the trachea and esophagus (e) as well as nerves entering the lobes are positive for p75NTR. E: Right hand box in D; the membranes of cells on the trachea are positive for p75NTR. F: Central box in D; the vagus contains p75NTR-positive cells and axons. G: Left hand box in D; nerves entering the lobes consist of p75NTR-positive cells. Scale bars = 250 μm in A,B, 100 μm in C, 200 μm in D, 25 μm in E,G, 50 μm in F.

Smooth Muscle and Neural Tissue at E11

Airway smooth muscle was detected by using antibodies to the smooth-muscle–specific markers α-actin (Fig. 3A) and calponin (Fig. 3B). At E11, ASM covered the trachea and epithelial tubules to the base of the growing buds (Fig. 3, the buds are unstained in the picture). The distribution of nerves and smooth muscle was demonstrated in the same preparation by double staining by using antibodies to PGP 9.5 and α-actin. Both neural tissue and undifferentiated epithelial cells showed immunoreactivity to PGP 9.5, enabling us to follow the development of neural tissue and view the structure of the bronchial tree concurrently. It should be noted when viewing projections of confocal images that the entire information in the z-axis (often >100 μm depth) is shown in one plane. When interpreting superimposed images of two markers, the more brightly stained one often appears to be above the less bright one when in reality it may be below. To reveal the true structural relationship, the optical sections have been individually examined and the findings are explained in the text and figure legends. Figure 4A shows a lung at E11, with airway smooth muscle surrounding the tubules to the base of the epithelial buds; the vagal nerves were very prominent. A network of nerve fibers extended from the vagus to the trachea, pulmonary arteries (Fig. 4B, upper box in 4A), and lobar bronchi (Fig. 4C, lower box in 4A). The majority of these nerve bundles were directed toward the smooth muscle-covered tubule, but some penetrated toward the edge of the mesenchymal cap (Fig. 4C).

Figure 3.

A: Embryonic day (E) 11 lung (dorsal view) showing immunoreactivity to α-actin. B: E11 lung (ventral view) showing calponin-immunoreactivity. t, trachea; br, bronchi; e, esophagus. Scale bars = 250 μm.

Figure 4.

Confocal projection of a fetal lung at embryonic day (E) 11 (ventral view) showing the relationship between protein gene product 9.5 (PGP 9.5) –positive nerves (bright green) and α-actin–positive airway smooth muscle (red). PGP 9.5 also stains the undifferentiated epithelium (light green). A: Smooth muscle covers the trachea and bronchi, extending to the base (thin arrows) of the epithelial buds (long arrows). The pulmonary arteries (short arrows) also stain positively for α-actin. The two vagal nerves (arrowheads) lie on the dorsal side of the lung, partly obscured here by the trachea and bronchi. B: Upper boxed area in A. Higher power view of the trachea and the division of the bronchi reveals extensive innervation. The proximal part of the left vagus (arrowhead) gives rise to a network of nerve fibers and ganglionic precursors, with nerves branching to the trachea and bronchi. Some cross to the right pulmonary artery (short arrow). C: Lower boxed area in A. Higher magnification of the left bronchus showing the distal part of the vagus (arrowhead) sending out thin processes to the smooth muscle. Some nerves extend toward the mesenchymal cap. Scale bars = 250 μm in A, 100 μm in B,C.

Smooth Muscle and Neural Tissue at E12

At E12, the five lobes of the fetal mouse lung were well defined (Fig. 5A, with the accessory lobe missing). In the left lobe, the laterals L1 to L5 (marked in figure), arising from the lobar bronchus were present, with the epithelial buds of L1 and L2 in the process of budding to form the next generation of tubules. A higher magnification of the left lobe in Figure 5B reveal airway smooth muscle encircling the epithelial tubules to the base of the growing buds. The vascular smooth muscle of the pulmonary arteries was strongly immunoreactive to α-actin and followed the trachea into the fetal lung (Fig. 5A). Nerve trunks followed the lobar bronchus with smaller nerve bundles branching along the laterals toward the epithelial buds of L1 and L2. Nerve trunks on the lobar bronchi terminated distally in varicose arborizations (Fig. 5C, boxed area in 5B). Some ganglia could be detected with the antibody to PGP 9.5 on the trachea (not shown) and in the proximal part of the lung (Fig. 5B), but at this stage, the staining was diffuse. Figure 5D shows a ganglion situated at the junction of the lobar bronchi of the middle and the accessory lobes. Two nerve trunks emerged from the ganglion and ran on both sides of the airway, sending out fine varicose fibers. By comparing the optical sections (2 μm steps) of the two stacks of nerves and smooth muscle, it was found that nerves were present in the sections above the smooth muscle and in the same sections as the smooth muscle.

Figure 5.

A: A partial confocal projection of the mouse lung at embryonic day (E) 12 (ventral view) shows the lobular organization (the accessory lobe and the vagi have been removed) and the epithelial tubules in longitudinal section. The laterals (L1–L5) of the left lobe are present with the end buds of L1 and L2 in the process of dividing to form the next generation of tubules. The undifferentiated epithelium of the end buds are immunoreactive to protein gene product 9.5 (PGP 9.5; light green). PGP 9.5 diffusely stains ganglionic precursors (long arrows) connected by nerve trunks and fibers (bright green) in the proximal part of the lung, with smaller trunks running along the lobar bronchi, giving off branches to the laterals. α-Actin–positive airway smooth muscle (red) is mostly seen in section, outlining the green tubules. The pulmonary arteries (short arrows) extend into the lobes. B: Higher power view of the left lobe reveals nerve trunks arising from ganglia (arrows) in the left proximal bronchus. They follow the bronchus and laterals, giving off fine varicose fibers (boxed area) to the smooth muscle. The smooth muscle encircles the tubules to the base (thin arrows) of the growing buds. C: Boxed area in B; varicose fibers spread out over the smooth muscle. D: Ganglion (arrow) at the junction of two lobar bronchi covered with smooth muscle. Two nerve trunks emerge from the PGP 9.5–positive ganglion and give off fine fibers to the smooth muscle. Scale bars = 250 μm in A,B, 50 μm inC, 100 μm in D.

Nerves, Ganglia, and Neuronal Precursors at E13 and E14

By E13, the antibody to PGP 9.5 revealed multiple generations of epithelial tubules in each lobe. Figure 6A shows a dorsal view of a lung at E13, with the vagal nerves lying on both sides of the esophagus. The vagus contributed many neural processes to the hilum of the lung where an abundance of ganglionic precursors and nerve trunks were present. Nerve trunks arising from ganglia on the lobar bronchi ran along the length of the tubules with finer varicose nerves lying in the mesenchyme, reaching out toward the periphery of the lung (Fig. 6B, enlargement of boxed area in 6A turned 90 degrees). These arborizations in the periphery were more developed at E14, and the varicose fibers were finer (Fig. 6C). At E14, the increasing thickness of the mesenchyme and the extensive network of tubules made visualization of the bronchial tree within the fetal lung difficult.

Figure 6.

A: Confocal projection, showing immunoreactivity to protein gene product 9.5 (PGP 9.5) in a mouse lung (dorsal view) at embryonic day (E) 13 (upper esophagus and trachea have been removed). Multiple generations of tubular branching are revealed by the PGP 9.5–positive undifferentiated epithelium. The brightly staining vagal nerves (v) on each side of the esophagus (e) send out branches to form a neural plexus comprising nerve trunks and precursor ganglia (intentionally oversaturated in image so that the finer nerves in the rest of the lung could be detected). Nerve trunks run along the primary bronchi, giving off branches, which reach the periphery. B: Boxed area in A; higher power view of the rostral part of the cranial lobe reveals fine varicose fibers extending to the periphery of the lung. C: In the E14 lung, distal nerves form varicose terminal arborizations. Scale bars = 250 μm in A, 100 μm in B,C.

By E13, a network of nerve trunks and ganglia on the trachea were connected to the vagus (Fig. 7A). Analysis by using NIH-image software showed that ganglia and nerves mainly lay on top of, and ran along the smooth muscle layer but some nerves penetrated it (not shown). The ganglia contained many PGP 9.5–positive cell bodies (Fig. 7B,C) that varied in diameter (5–15 μm). The intensity of the staining was also variable, which may reflect their stage of maturity. At E14, the neural network on the trachea was more extensive with larger ganglia connected by nerve trunks that were more numerous and thicker than at E13 (Fig. 7D). The ganglia contained from a few to over a hundred cell bodies. At E14, finer nerves originating from the ganglia spread over the trachea (Fig. 7E, boxed area in 7D). Ganglia were also present along the tubules. Figure 7F is a projection of both nerves and ASM, which shows an E14 ganglion at the junction of the smooth muscle-covered lobar bronchi of the middle and the accessory lobes. Compared with the E12 ganglion in Figure 5D, the E14 ganglion is larger and cellular in appearance and the nerve trunks are thicker. Two large nerve trunks emerged from the ganglion and followed the tubule, sending out fine varicose fibers to the smooth muscle. These fibers cannot be seen in the composite picture because the smooth muscle was as bright as the nerves, but Figure 7G (boxed area in Fig. 7F) shows only the PGP 9.5–positive nerves spreading over the surface of the tubule. A comparison of the optical sections (2-μm steps) of nerves and smooth muscle by using NIH-image software revealed nerves both above and in the same sections as the smooth muscle (not shown).

Figure 7.

A: The embryonic day (E) 13 trachea shows protein gene product 9.5 (PGP 9.5) –positive ganglia on the surface, interconnected by nerve trunks and innervated by the vagus (v). B: Boxed area in A. Projection of a ganglion showing cell bodies. C: Single section through the ganglion in B showing PGP 9.5 immunoreactivity in the cytoplasm and nucleus, leaving the nucleoli unstained. D: The E14 trachea displaying the PGP 9.5–positive vagus and network of ganglia and nerve trunks. Both ganglia and nerve trunks are larger than at E13. E: Boxed area in D. Nerves from ganglia spread over the surface of the trachea. F: E14 ganglion at the junction of two bronchi-composite image of neural tissue and airway smooth muscle. Nerve trunks emerge from the ganglion and run along the length of the tubule. Bundles of smooth muscle encircle the tubules. G: Boxed area in F. The nerve trunk gives rise to varicose nerves lying on the surface. Scale bars = 200 μm in A,D, 25 μm in B,C,G, 100 μm in E,F.

Nerves and Airway Smooth Muscle in Adult Airways

To show that nerves and ASM survive and mature in the airways through to adulthood, the adventitia was carefully dissected away from the airways before immunostaining. Antibodies to α-actin (Fig. 8A) and PGP 9.5 (Fig. 8B) showed that the airways were covered in smooth muscle and ensheathed in a plexus of nerves. The figure shows a representative part of a lobar bronchus where a nerve trunk runs along its length, giving rise to fine varicose fibers that were oriented along the airway smooth muscle bundles. Some of these fine fibers were located in the same optical sections (2 μm steps) as the smooth muscle bundles that were regularly arranged around the airway circumference.

Figure 8.

Confocal projection from an adult mouse lobar bronchus stained for α-actin and protein gene product 9.5 (PGP 9.5). A: Bundles of α-actin–positive smooth muscle are oriented around its circumference. B: A large nerve trunk lying on the airway gives off fine fibers that similarly run in the direction of the smooth muscle bundles. Scale bars = 100 μm.

Neural Tissue and Smooth Muscle Persist and Continue to Grow in Culture

To see whether the growth of neural tissue and ASM could be sustained in vitro, left lobes at E12 were cultured for 5 days. During this time, the lobes enlarged, the laterals lengthened, and further generations of tubules appeared. An antibody to α-actin revealed that the airway smooth muscle was preserved in vitro (Fig. 9A) and that new smooth muscle had been laid down around the circumference of the tubules as they grew. During culture, spontaneous narrowing of the tubules was observed, showing that the smooth muscle was active. At the start of culture (E12), nerves could be detected along the lobar bronchus and clusters of diffusely staining PGP 9.5–positive neurons were restricted to the trachea and the division of the lobar bronchi (Fig. 5). After 5 days in culture, all lobes revealed neurons that had differentiated along the length of the smooth muscle-covered lobar bronchus and matured into a network of nerves and ganglia (Fig. 9B). Figure 9C shows a projection of two ganglia connected by a nerve trunk, from which nerve bundles gave rise to fine fibers. Single optical sections through ganglia revealed PGP 9.5 immunoreactivity in the cytoplasm, whereas the nucleus remained unstained (Fig. 9D). In some ganglia, the nuclei were also stained (not shown), as sometimes also seen in vivo (Fig. 7C).

Figure 9.

Left lobe at embryonic day (E) 12 cultured for 5 days. A: Confocal projection (dorsal view) showing α-actin immunoreactivity, revealing the smooth muscle covering the epithelial tubules. br, lobar bronchus. B: Higher magnification of the lobar bronchus double stained for smooth muscle and neural tissue with α-actin and protein gene product 9.5 (PGP 9.5), respectively. The two images have been superimposed to show how the nerves follow the smooth muscle–covered tubule. The intensity of the smooth muscle has been reduced to differentiate between the two tissues. C: PGP 9.5 immunoreactivity reveals ganglia interconnected by nerve trunks and branching nerve fibers. D: A single optical section through the left ganglion in C showing the large nuclei with a thin rim of PGP 9.5–positive cytoplasm. Scale bars = 500 μm in A, 100 μm in B, 50 μm in C (applies to C,D).


This study documents the structural and spatial organization of nerves and smooth muscle in relation to the branching epithelial tubules of the primordial mouse lung, a feature that hitherto has not been recognized. Here, we report that both neural tissue and ASM are an integral part of the tubules from the onset of lung development. Smooth muscle forms a layer of cells which encircle the circumference of the liquid-filled epithelial tubules to their growing tips and is in a prime position to generate a pressure in the tubules and, hence, influence lung growth. Clusters of neural crest-derived cells form ganglia that are innervated from the vagus. These send out postganglionic fibers that follow the tubules to the base of the epithelial buds. The close association of the nerves with the smooth muscle suggests that trophic interactions are occurring during the early pseudoglandular stage.

Growth of the Mouse Lung In Vivo

We have characterized the structure of the mouse lung at daily intervals between E11 to E14 because of the striking changes that occur during this early pseudoglandular stage. At E11, approximately five to seven epithelial buds arise from the two main tubules. By E12, a rapid transformation has occurred with the five lobes of the mouse lung clearly defined. Primary branches (laterals) extend from the lobar bronchi with the more proximal forming a new generation of branching through divisions of the epithelial buds. By E13, the many divisions of the terminal epithelial buds have greatly increased the number of generations of airways and additional buds have appeared along the length of the laterals, ventrally and dorsally. At E14, the thickness of the lung and the intricate network of tubules made it difficult to see the overall structure of the branching tubules. By recording the area of the left lobe and the number of peripheral buds, we show that the developmental maturity of the left lobe was consistent between animals of the same age, whereas the changes from day to day were striking. These parameters have not been reported to this degree of detail in vivo, but counting of the peripheral buds has been frequently used as a convenient measure of development in vitro where cultured fetal lungs are used to study branching morphogenesis (Souza et al., 1995; Warburton et al., 1992). The data collected are intended to provide a basis for normal growth and can be used to compare conditions that influence the lung growth in utero as well as in vitro.

ASM Is an Integral Component of the Branching Tubules

Airway smooth muscle is an integral part of the structure of the developing bronchial tree. That this is smooth muscle and not a precursor is based on the following evidence. At E11, antibodies to either α-actin or calponin revealed bundles of muscle fibers encircling the circumference of the tubules. Both of these proteins are regarded as specific for smooth muscle in vivo (Gimona et al., 1992; Mitchell et al., 1990). Smooth muscle myosin and metavinculin mRNA (both specific smooth muscle markers) are already present in a sheath of adventitial cells surrounding the tubules at E10.5 (Jostarndt-Fogen et al., 1998). α-Actin gives identical staining compared with smooth muscle myosin throughout the smooth muscle of the first trimester bronchial tree of fetal pigs, including the region at the base of the epithelial buds where the newly differentiated smooth muscle is being laid down (Sparrow et al., 1995; Weichselbaum et al., 1996). The smooth muscle is active from an early age as indicated by the spontaneous narrowing of the tubules observed in vitro, as reported in this study and by others (Roman, 1995; Schittny et al., 2000). In fresh fetal human (McCray, 1993) and pig lungs (Schittny et al., 2000; Sparrow et al., 1994, 1995), the terminal tubules through to the larger airways narrow and relax spontaneously, and in response to agonists. Furthermore, the ASM tone has the potential to generate a positive pressure of 2.3 cm H2O in the liquid-filled lumen of the tubules of fetal rabbit lung in vivo (Schittny et al., 2000). Pressures have also been measured in fetal mouse lungs in vitro. It was found that cultured ligated left lobes with a pressure of 3.1 mm Hg showed more maturity than unligated lobes with a pressure of 0.8 mm Hg (Blewett et al., 1996). There is now strong evidence that the force generated by the ASM across the airway wall and adjacent parenchyma can stimulate lung growth by inducing growth factor production and cell proliferation by means of mechanotransduction (Cilley et al., 2000; Liu et al., 1995; Kitano et al., 2000). In this study, we have provided the morphologic evidence for airway narrowing by showing that smooth muscle is present from the onset of lung development. The orientation of the smooth muscle is already evident at E11 — encircling the circumference of the tubules, an arrangement whereby the ASM tone would have the potential to produce narrowing and relaxation. We also show that further development of the ASM takes place in culture, with the structural arrangement required for airway narrowing being maintained.

Origin of Neural Tissue in Lung

The origin of neurons of the lung has not been documented. Because the neural crest cells that will populate the gut are already present in the foregut at the time the lung buds emerge, it has been postulated that these cells will migrate into the lung as it is forming (Dey and Hung, 1997). We used an antibody to p75NTR to detect neural crest cells in the lung, because it has been shown that all neural crest-derived cells in the gut express this receptor (Young et al., 1999). At E11, p75NTR-positive cells were present on the trachea in a pattern resembling that seen with an antibody to PGP 9.5 on E13. The vagus as well as nerves originating from the vagus extending into the lung also contained cells positive for p75NTR. It seemed that even though neural crest cells were already present at E11 they were still migrating from the vagus to the lung at this stage. Evidence of neuronal precursors in the vagi has previously been reported, and it was found that cells with neuronal properties developed in cultures of explanted vagus nerves (Baetge et al., 1990). Non-neuronal cells, immunoreactive to p75NTR, have also been found in the lung (Wheeler et al., 1998). To ascertain that all the cells seen are neural crest cells, we would need to use antibodies against other early markers such as the transcription factor Phox2b (Pattyn et al., 1999; Young et al., 1999).

Expression of Neural Tissue

At E11, nerves sprouted from the vagus to the lung, with some directed toward the trachea, bronchi, and pulmonary arteries. By E12, ganglionic precursors could be distinguished with anti-PGP 9.5, appearing as flat, diffuse patches, lying on the future trachea and proximal lobar bronchi. Individual cell bodies could not be detected within the neural tissue, except occasionally, a few isolated neurons in the trachea. This pattern of PGP 9.5 staining has been reported in the developing pig lung, where triple staining by using an antibody to synaptic vesicle protein 2 showed nerve fibers within the diffusely stained neural tissue and the nuclei of the cell population was revealed with ethidium bromide (Weichselbaum and Sparrow, 1999). Postganglionic fibers, arising from the proximal neural tissue, ran along the smooth muscle-covered tubules toward the base of the end buds. At this stage, nerves were rarely seen that were not in close association with the tubules so it is possible that these are attracted to the smooth muscle by the secretion of neurotrophic factors. By E13, the nerves still predominantly followed the future airways but some had extended beyond the buds to the edge of the mesenchymal cap where they displayed fine arborized endings, suggesting that a new chemotactic signal had appeared in the mesenchyme. It is not known which neurotrophic factors participate in the innervation of the lung and further studies are needed. Members of the glial-derived neurotrophic factor (GDNF) family could play a significant role, because GDNF is the most important neurotrophic factor in the gut (Chalazonitis et al., 1998; Moore et al., 1996; Sanchez et al., 1996), and because the mRNA of GDNF (Towers et al., 1998), neurturin (Widenfalk et al., 1997), as well as their respective receptors have been detected in the developing lung. In our lung explants, we show that nerves and smooth muscle persist and continue to grow in culture. This model will be used to study the differentiation of neural tissue, and in a preliminary study, we have observed an increase in nerve growth in response to GDNF (unpublished observations).

We have documented the distribution of neural tissue and airway smooth muscle in the fetal lung by using the mouse to obtain a detailed picture of their development during the early pseudoglandular stage. In addition, this information will provide an in vivo basis for comparing the extent to which nerves and airway smooth muscle persist and continue to develop in cultured lung explants. This in vitro model will be used to investigate the factors regulating ASM tone, nerve growth and survival, and to what extent these tissues influence lung development.



Timed-pregnant mice (outbred Quackenbush) at 11, 12, 13, 14, and 15 days of gestation were obtained from the Animal Resources Centre (Canning Vale, Western Australia, Australia). Mice were mated overnight, and the morning of finding the vaginal plug was considered as embryonic day 0 (E0). The mice were killed by inhalation of a 70% CO2/30% O2 gas mixture for 30 sec, followed by cervical dislocation. Embryos were dissected from the uteri, placed on ice, and decapitated. The primordial lungs were excised in phosphate-buffered saline (PBS) by using microdissection tools. Video photomicrographs of fetuses and lungs were captured at ×10 to ×70 with a Sony CCD-Iris/RGB color video camera, connected to a dissecting microscope (Olympus SZH10) and digitized by using the software Apple Video Player on a PowerMac. Digital images of whole fetuses and left lung lobes were analyzed by using PhotoShop® 5; the crown-rump length and the area of the left lobe were measured and the peripheral buds counted. The number of fetuses and lungs were 15 at E11, 27 at E12 and E13, 9 at E14 and E15. Means and standard deviations (SD) are shown. Curves were fitted by using Microsoft Excel 98.

Organ Culture

Left lung lobes (E12) were placed in Falcon 12 well culture inserts with transparent filters (4 μm pore size, Australian Biosearch, Karrinyup, Western Australia, Australia). One or two lobes were placed on each filter. Each well contained 1 ml of media, and the lobes were transferred onto the filter in a drop of media to keep them moist but still exposed to the air. The culture medium consisted of DMEM/F12 with 0.048 mg/ml penicillin, 0.67 mg/ml streptomycin, and 0.205 mg/ml amphoterizin B (Fungizone) and 10% fetal bovine serum. The medium was changed after 2 and 4 days in culture. All culture reagents were obtained from GIBCO (Life Technologies Pty., Ltd, Mulgrave, Victoria, Australia). Lobes were cultured for 5 days in a humidified, 5% CO2 environment at 37°C. A total of 17 lobes in four separate experiments were cultured and stained for nerves and smooth muscle.


Primordial lungs were immunochemically stained as whole-mounts and processed according to Weichselbaum and Sparrow (1999). Nerves were detected by using polyclonal rabbit antibodies to PGP 9.5 (1/200) (Ultraclone, Isle of Wight, UK), synapsin (1/200) (Calbiochem-Novabiochem Corporation, San Diego, CA), and p75NTR (1/200) (Promega Corporation, Madison, WI). Smooth muscle was recognized by a monoclonal mouse antibody to α-actin (1/2,000) (Sigma Chemical Company, St. Louis, MO) and by a polyclonal rabbit antibody to calponin (1/5), a kind gift from Dr. M. Gimona (Institute of Molecular Biology, Salzburg, Austria). Secondary antibodies (anti-rabbit and anti-mouse), conjugated to CY3, CY5 (Zymed Laboratories, San Francisco, CA), and Oregon green 488 (Og 488) (Molecular Probes, Eugene, OR), were used to visualize the labeled tissues with the confocal laser-scanning microscope. As a control, the primary antibody was omitted with no staining above background as a result.

Confocal Microscopy

Fluorescent images of nerves and smooth muscle were obtained by using a confocal laser scanning microscope (MRC 1000; Bio-Rad, Hemel Hempstead, UK) with COMOS software (version 7.0; Bio-Rad). Og 488 and CY3 were detected by a krypton/argon laser with the excitation wavelengths of 488 and 568 nm, respectively. The whole-mounts were optically sectioned by scanning at increasing depths of focus (in steps of 1 to 10 μm, depending on the magnification used). The stacks of images obtained were projected into two-dimensional (2-D) images with the aid of Confocal Assistant, a software program that uses the maximum intensity of the corresponding pixels in each optical section to produce a 2-D image. In some cases, only a partial projection (i.e., a selected sequence of optical sections) was made to show more clearly the structures of interest. With larger whole-mounts, several fields of view were captured and made into a montage representing the complete whole-mount. After double staining, the Og488 and CY3images were captured separately, colorized, and merged to show a composite nerve/smooth muscle image. Image processing (montaging, colorizing, and merging) was done by using Adobe® PhotoShop® 5.0 software. Note that when interpreting the projected confocal images where two markers of different colors have been superimposed, the more intensely stained tissue or structure may appear to overly the less bright image. To see the true structural relationship, the single optical sections must be compared. This is not always obvious in a 2-D image, but the text and figure legends describe the true relationship. To demonstrate cell bodies in ganglia, single optical sections were imaged. Cells were counted and cell sizes measured by using NIH-image software.