Development of the Innervation and Airway Smooth Muscle in
Human Fetal Lung
Malcolm P. Sparrow, Markus Weichselbaum, and Paul B. McCray, Jr.
Department of Physiology, University of Western Australia, Nedlands, Australia; and Department of Pediatrics,
College of Medicine, University of Iowa, Iowa City, Iowa
Human and porcine fetal airways have been shown to contract spontaneously from the first trimester, the
latter also contracting in response to neural stimulation. Our object was to map immunohistochemically
the innervation and its relationship to the airway smooth muscle (ASM) in the human fetal lung from early
gestation to the postnatal period. Whole mounts of the bronchial tree were stained with antibodies to the
pan-neuronal marker protein gene product 9.5, the Schwann cell marker S-100, and the ASM contractile
protein a-actin, and imaged using confocal microscopy. By the end of the embryonic period (53 d gestation), the branching epithelial tubules in the primordial lung were covered with ASM to the base of the terminal sacs. An extensive plexus of nerve trunks containing nerve bundles, forming ganglia, and Schwann
cells ensheathed the ASM. By 16 wk (canalicular stage), maturation of the innervation was advanced with
two major nerve trunks running the length of the bronchial tree, giving rise to varicosed fibers lying on the
ASM. An extensive nerve plexus in the mucosa was also present. The distal airways of infants who had
died of Sudden Infant Death Syndrome were also covered with smooth muscle and were well innervated.
Thus, an essentially complete coat of ASM and an abundant neural plexus ensheathing the airways are an integral part of the branching epithelial tubules from early in lung development. Sparrow, M. P., M. Weichselbaum, and P. B. McCray, Jr. 1999. Development of the innervation and airway smooth muscle in
human fetal lung. Am. J. Respir. Cell Mol. Biol. 20:550–560.
Human and porcine fetal airways have been shown to contract spontaneously from the first trimester, the latter also
contracting in response to neural stimulation (1–3). Realtime video imaging of contracting airways revealed that
the epithelial tubules are contractile from the proximal
airways to the base of the terminal sacs. This phenomenon
has been well documented in fetal pig lungs in the first and
second trimesters, where spontaneous narrowing and relaxation of the airways move the lung liquid back and
forth (2). The narrowing of the airways of fetal pigs in response to neural stimulation was blocked by atropine and
by tetrodotoxin, indicating that functional cholinergic
nerves were present (2, 3). The presence of nerves and the
capacity of the airways to contract indicates that innervated, functional smooth muscle is present in the newly
(Received in original form April 1, 1998 and in revised form August 6, 1998)
Address correspondence to: Dr. Malcolm P. Sparrow, Dept. of Physiology,
University of Western Australia, Nedlands, Australia 6907. E-mail: msparrow
@cyllene.uwa.edu.au
Abbreviations: airway smooth muscle, ASM; fluorescein isothiocyanate,
FITC; phosphate-buffered saline, PBS; protein gene product 9.5, PGP 9.5;
rhodamine red, RITC; Sudden Infant Death Syndrome, SIDS.
Am. J. Respir. Cell Mol. Biol. Vol. 20, pp. 550–560, 1999
Internet address: www.atsjournals.org
forming airways early in gestation. The objective of the
present study was to map the development of the nerves
and their relationship to the airway smooth muscle (ASM)
in the human fetal lung.
The human lung develops in the embryonic period as
an outgrowth of the foregut, an epithelial tube, which is
destined to become the trachea. As it elongates, it invades
the surrounding mesenchyme and at 4.5 wk in the human
embryo comprises five tiny saccules, two on the left and
three on the right—the future lobar bronchi and corresponding lung lobes (4). These epithelial tubules, which terminate in blind sacs, have the appearance of glands when examined in section, and this stage of lung development is
termed the pseudoglandular phase. It continues for about
another 12 wk with subsequent dichotomous branching
giving rise to about 20 generations, with two more added
in the subsequent canalicular phase (5). This nearly completes the formation of the bronchial tree, which is 24 generations in the adult, inclusive of the trachea.
Recently Weichselbaum and colleagues (6) showed not
only that the growing airways in first-trimester fetal pig
lungs were covered in a well-formed layer of smooth-muscle cells, but also that an extensive nerve plexus comprising nerve trunks and ganglia invested the airways, with
fine bundles innervating the smooth muscle. To stain the
developing nerves, an antibody to protein gene product
Sparrow, Weichselbaum, and McCray: Development of Nerves and Smooth Muscle in Human Fetal Lung
9.5 (PGP 9.5) was used that stains all nerves and neurons
(7). In first-trimester fetal pigs, PGP 9.5 was also found to
stain the epithelial cells diffusely (6). This appears to be a
property of the undifferentiated epithelial cells, as has
been shown in human fetal lung younger than 16 wk gestation (8). In the present study, this diffuse staining has been
used to image the growing tips of the epithelial tubules,
thereby facilitating the mapping of the innervation of the
developing airways. PGP 9.5 has also been used to stain
551
pulmonary neuroendocrine cells (9). The colocalization of
PGP 9.5 and other neuroendocrine markers has been investigated in cross sections of human fetal lung (8, 10) and
explants of cultured rat lungs (11).
The aim of the present study was to map immunohistochemically the nerves and the smooth muscle of the
bronchial tree in the human fetal lung from early gestation
with antibodies to the nerve marker PGP 9.5 and the
smooth-muscle protein a-actin (12). The latter gives iden-
Figure 1. (a) Videomicrograph of a human fetal lung at 53 d gestation. The left upper lobe has become detached from its bronchus. (b) Low-power view of a whole mount of part of the right
upper lobe of this lung taken with a confocal microscope. The
neural tissue is stained with an antibody to the pan-neuronal
marker PGP 9.5 (green), which reveals a network of nerve trunks
and forming ganglia. PGP 9.5 also stains the epithelial cells of the
terminal sacs at this early stage of development. The ASM
stained for a-actin (red) is seen more clearly in the periphery
where the mesenchyme is thinner (arrows). Smooth-muscle bundles can be seen encircling the tubules in places. A few remaining
pulmonary blood vessels are present that also stain with a-actin
(arrowheads). (c) ASM in the trachea of the human fetal lung at
53 d of gestation stained for a-actin (shown in a). The smooth
muscle completely encircles the circumference of the trachea, lying perpendicular to the long axis of the trachea. The muscle fibers are tightly packed into bundles, which can just be distinguished by their slight separation in some places around the
circumference (arrows). Cartilage rings are absent at this stage of
development.
552
AMERICAN JOURNAL OF RESPIRATORY CELL AND MOLECULAR BIOLOGY VOL. 20 1999
Figure 2. Innervation and ASM in the developing airways of a fetal human lung at 58 d of gestation. The field in a through c shows
branching epithelial tubules in the periphery of a lobe; d through f show a single optical section taken at high power that is cut lengthwise through a nerve trunk and ganglion to reveal neural tissue, Schwann cells, and nuclei in the same whole mount as a through c. Left
panels: (a) Nerves and ganglia stained for PGP 9.5 (green) form a network ensheathing the future airways. (b) Same field, showing the
ASM stained for a-actin (red), which is arranged circumferentially around the epithelial tubule and lies essentially perpendicular to
the long axis of the tubule. (c) Nerve and smooth-muscle images are merged to show that the main nerves form a sheath that lies above
the ASM; in some places as far as z 40 mm (arrows). Smaller nerve bundles can be discerned ( arrowheads) that, at higher power, are
seen to descend to the surface of the smooth muscle. Right panels: A single optical section (thickness , 1 mm) cut through a slightly
larger, more proximal nerve than shown in a. Images are progressively merged from d to f. The nerve trunk and the adjoining ganglion
are stained for: (d) Schwann cells with S-100 ( yellow). The S-100 stains the nuclei of the developing Schwann cells as well as the processes extending from their ends. (e) Neural tissue with PGP 9.5 (blue) and Schwann cells with S-100 (yellow). The nuclei of the neurons
occupy most of the cell and exhibit nucleoli. ( f ) Nucleic acid with ethidium bromide (red) show the nuclei present in the tissues in the
composite picture. Two round nuclei of unknown identity stain up brightly in the nerve trunk.
Sparrow, Weichselbaum, and McCray: Development of Nerves and Smooth Muscle in Human Fetal Lung
tical staining compared with a polyclonal antibody specific
for smooth-muscle myosin in fetal human and porcine airways (3, 6). In addition, Schwann cells have been revealed
using S-100, which 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 (13). Whole mounts of the bronchial tree
at 7.5 to 8 wk (early pseudoglandular stage), and 16 to 18 wk
(early canalicular stage) were imaged using confocal laser
scanning microscopy to reveal the neural, ASM, and vascular structures of the airway wall. Projections of optical
sections cut through the thickness of the airway wall have
been used to produce overviews of the innervation along
the length of the bronchial tree, as well as to provide structural detail. The study has been extended into postnatal
life using lungs from infants who died of Sudden Infant
Death Syndrome (SIDS).
Materials and Methods
Lung Tissues
Human fetal lungs from the first (7.5 to 8 wk) and second
(16, 18, and 24 wk) trimesters from therapeutic terminations of pregnancy were obtained from the Department of
Pediatrics, University of Iowa Hospitals and Clinics (Iowa
City, IA). The tissues were received fixed in 4% buffered
formaldehyde. Gestational age was estimated by footlength measurement (14). In the first trimester, lobes from
four lungs at 53, 54, 56, and 58 d gestation were studied.
The only complete lung obtained was at 53 d and contained
the distal half of the trachea. In the second trimester, lobes
from lungs of 16-, 18-, and 24-wk gestation were obtained.
The lobar bronchi of these lobes had been removed before
they were made available for this study. The results of airways in the 24-wk lung lobe were qualitatively similar to
those of the 18-wk lobe and have not been included in the
present work because of inferior image quality. The study
was approved by the Institutional Review Board of the
University of Iowa. The left upper lobes of two infants
aged 5 and 10 mo who had died of SIDS were obtained
from the Victorian Institute of Forensic Medicine (VIFM),
Figure 3. Video image of a segmental bronchus and subsegmental branches from a lobe of a 16-wk-gestation human fetal lung
oriented to show the luminal opening of the bronchus (arrow).
553
Melbourne, Australia. The time that elapsed between the
estimated time of death and the fixing of the lungs in 4%
buffered formaldehyde at autopsy was 22 and 51 h, respectively. A period of about 16 wk had elapsed prior to dissection. This study was approved by the VIFM and by the
Human Ethics Committee of the University of Western
Australia (Nedlands, Australia).
Immunohistochemistry
The parenchyma and vasculature in the lobes from the
first-trimester fetal lungs were carefully teased away from
the airways using a dissecting microscope, isolating most
of the bronchial tree. In the lobes from the second-trimester fetal lungs the segmental bronchus and its branches
were separated from the pulmonary arterial tree, leaving
both largely intact. In the left upper lobes from the SIDS
lungs, distal airways of z 1 mm diameter were dissected
free of pulmonary arteries and parenchyma. It was difficult to separate the airways from the surrounding tissue
without damaging their adventitia, particularly the proximal bronchi of the lobe. The lobe obtained 51 h postmortem could not feasibly be imaged with sufficient quality. In
the lobe obtained 22 h after death, the proximal bronchi
did not image clearly; however, the distal airways could be
imaged satisfactorily. Following dissection, the specimens
from fetal and postnatal lung were cleared in dimethyl sulfoxide three times for 10 min each (15). After washing in
phosphate-buffered saline (PBS), pH 7.2, twice for 10 min
each, nonspecific binding was blocked by an additional
washing step in PBS that contained 1% bovine serum albumin before application of the primary antibodies. To
stain for all nerves, a rabbit polyclonal antibody to PGP
9.5 (Protein Gene Product 9.5; Ultra Clone, Isle of Wight,
UK) was used. Schwann cells were stained with a rabbit
polyclonal S-100 antibody (Dako, Sydney, Australia).
Smooth muscle was identified with mouse monoclonal
anti-a-actin (Sigma Chemical Co., St. Louis, MO). The final dilutions of the respective antibodies from the stock
solutions obtained from the above suppliers were: PGP
9.5, 1:100; S-100, 1:200; a-actin, 1:250.
The preparation was then incubated overnight in a humidified chamber at 48C. After being washed several times
with PBS over a 4-h period, the samples were incubated for
12 h at room temperature with the fluorochrome-labeled
secondary antibodies. The secondary antibodies (antimouse
and antirabbit) were conjugated to either fluorescein isothiocyanate (FITC) (Silenus, Melbourne, Australia), rhodamine red (RITC) (Cappel; Organon Teknika Corp., West
Chester, PA), or Cy5 (Amersham, Sydney, Australia). Cy5
was used in Figures 2e and 2f in conjunction with FITC
and ethidium bromide, allowing for registration of nonoverlapping fluorescence emission. Control experiments
to test for autofluorescence and nonspecific staining due to
incomplete washout of the secondary antibody were carried out using nonimmune rabbit and mouse sera. Ethidium bromide staining was performed before mounting by
exposing the tissues for 2 min to a 0.1 mg/ml ethidium bromide solution. After further washing with PBS, the preparations were mounted in 90% glycerol containing p-phenylethylenediamine (1 mg/ml) to reduce bleaching of the
fluorochromes. Each specimen was mounted on a separate
554
AMERICAN JOURNAL OF RESPIRATORY CELL AND MOLECULAR BIOLOGY VOL. 20 1999
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 for FITC and ethidium bromide
were 488 nm. RITC and Cy5 were excited at 568 and 645
nm, respectively. Green emission by FITC was recorded
with a 522-nm bandpass filter. Red emission by RITC and
ethidium bromide was registered with a 585-nm long-pass
filter. Far-red emission by Cy5 was recorded with a 685-nm
long-pass filter. The whole mounts were optically sectioned
by scanning at increasing depths of focus (typically, in
steps of 1 mm) to follow the path of the nerves in relation
to the smooth muscle. The maximum intensity of the corresponding pixels in each optical section was used to generate a single image (two-dimensional 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 (Adobe Systems, Inc., San
Jose, CA). In multistaining experiments, single fields were
scanned for each marker, then colorized and superimposed. Measurements of cell sizes in ganglia were made
from single optical sections through ganglia. Typically, a
360 objective was used which, in conjunction with the corresponding settings for laser power, iris width, and photomultiplier gain, allowed sections of less than 1 mm thickness to be obtained. All airway diameters refer to the
external diameter of the smooth-muscle layer unless stated
otherwise. Cell diameters refer to their largest diameter.
The distance separating the nerve plexus from the ASM in
the airways of the 58-d gestation fetal pig (Figure 2c) was
determined from the number of confocal sections separating the nerve trunks and the surface of the ASM. Data are
shown as means 6 SEM.
Results
Fetal Lung, 7.5 to 8 wk Gestation
Four lungs from 53 to 58 d of gestation were studied. Figure 1a (color plate) shows a lung at 53 d of gestation with
well-developed lobes composed of epithelial tubules that
appear as delicate, translucent, tubular branching structures, with each tubule terminating in a budlike blind end
(Figure 1b) and supported in a loose extracellular matrix
of mesenchyme. The growing tips, which comprise a single
layer of epithelial cells, are referred to here as terminal
sacs. The pan-neuronal marker PGP 9.5 stained nerves in
all tissues examined. Control experiments conducted with
nonimmune rabbit sera demonstrated no staining. PGP 9.5
revealed a network of nerve trunks overlying the adventitial surface of the tubules, with finer nerves reaching toward the terminal sacs. Many forming ganglia were present
at the divisions of nerve bundles.
ASM, stained for a-actin, forms an essentially continuous layer over these future airways extending to the base
of the terminal sacs. These sacs are stained by the PGP 9.5
antibody (6, 8), thus delineating the collar of smooth muscle terminating at their base. The ASM is most clearly revealed in the periphery (Figure 1b, arrow), where the overlying mesenchyme is less dense. Pulmonary vessels that have
not been removed during dissection stained with a-actin
and are seen to follow the distal tubules closely (Figure 1b,
arrowheads). ASM completely surrounds the circumference
of the trachea (Figure 1c), lying perpendicular to its long
axis. The continuity of the smooth muscle was confirmed
by also imaging the other side of the trachea. The arrangement of smooth-muscle fibers into bundles (z 50 mm wide)
is typical for mature smooth muscle seen postnatally (M. P.
Sparrow and H. W. Mitchell, unpublished data). The cartilage rings are absent at this stage of development.
The relationship of the neural plexus to the ASM in the
distal tubules in a 58-d fetal lung is shown in Figures 2a
through 2c. A network of nerve trunks with forming ganglia at their intersections covers the future airways (Figure
2a) with a continuous layer of smooth-muscle cells arranged
cylindrically around them (Figure 2b). The pitch of the fibers is essentially at right angles to the long axis of the airways, contrasting with data by Ebina and associates (16),
whose electron microscopic study indicated a pitch of approximately 30 degrees in adult human airways. The neural tissue appears significantly separated from the airways
because of the unstained mesenchymal tissue surrounding
the airways that supports the neural sheath (Figure 2c).
An estimate of the average distance (along the z-axis; see
MATERIALS AND METHODS) of the nerve trunks to the smooth
muscle was 19 6 1.0 mm (n 5 27) in three fields of this
58-d fetal lung. In some places this separation appears considerably greater, as much as 40 mm (Figure 2c, arrows),
possibly because of compression of the circular airway
during mounting. This field is representative of the innervation and ASM in similar preparations from 53-, 54-, and
56-d fetal lungs. Small nerve bundles were observed descending from the trunks toward the smooth-muscle layer
(Figure 2c, arrowheads), and at higher power (not shown)
these gave rise to some fine varicosed fibers lying in close
proximity (, 1 mm) to the smooth-muscle cells.
A developing Schwann cell population was present in
the nerve trunks, and to a lesser extent in the ganglia, revealed using the Schwann cell marker S-100 that stains
their nuclei and sheaths. Figure 2d shows a nerve trunk
and ganglion cut lengthwise by the confocal microscope.
This single optical section shows the Schwann cell nuclei
with short processes extending from their ends. Some
Schwann cell nuclei and fibrillar material are seen in the
adjacent ganglion where dark, rounded nuclei of nerve
bodies are present. Concurrent staining with the PGP 9.5
antibody reveals the neural cytoplasm in the nerve trunk
and in the nerve bodies of the ganglion (Figure 2e). The
dark spots within the nuclei appear to be nucleoli when examined at higher power. Staining with ethidium bromide
(Figure 2f) shows the nuclei of the nerve bodies in the ganglion, the Schwann cell nuclei, and a host of nuclei of the
cells present in the surrounding mesenchyme. The nuclei
of the Schwann cells appear orange-yellow because of the
Sparrow, Weichselbaum, and McCray: Development of Nerves and Smooth Muscle in Human Fetal Lung
555
Figure 4. The wall of a proximal part of the segmental bronchus (see Figure 3) stained for nerves with PGP 9.5 and smooth muscle with
a-actin, imaged from both the adventitial (a) and the mucosal (b) surfaces. (a) A representative single field of the adventitial surface
showing large nerve trunks with ganglia (green), bronchial vessels (red), and mucous glands (red). These glands contain myoepithelial
cells that stain for a-actin and are innervated by fine nerve fibers (arrows). Some ASM bundles can be seen but most are obscured by
the wall structures and adhering parenchyma. (b) A view of the mucosa from the luminal surface where bundles of intersecting nerves
run parallel with the long axis of the airway, with bundles of the ASM lying underneath. A prominent bronchial vessel is visible that is
located on the adventitial side of the smooth muscle (localized by confocal sectioning).
color mixing in the composite image of this triple-stained
tissue.
Thus, by about the middle of the first trimester the airways are covered with smooth muscle and enveloped by a
rich network of developing ganglia and nerve trunks.
These trunks contain a population of Schwann cells.
Fetal Lung, 16 to 18 wk Gestation
The innervation and ASM were examined in the segmental airways of lobes from fetal lungs at 16 to 18 wk gestation. The airways were gently separated from the pulmonary
vessels and parenchyma. Figure 3 shows the configuration
of these airways from a lobe at 16 wk gestation. At the
proximal end (arrow), the airway was opened and the
bronchial wall imaged from both the adventitial and mucosal surface. Figure 4a shows a representative area of the
adventitia where large nerve trunks (45- to 72-mm diameter), bronchial vessels (chiefly arterioles), and mucous
glands are present. The a-actin stains not only the airway
and vascular smooth muscle but also the myoepithelial
cells of mucous glands (17). Fine nerves to the acini of the
mucous glands can just be detected ( arrows) and appear
varicosed when viewed at higher power. This projection
(Figure 4a) is 120 mm in depth and comprises 31 optical
sections. The bundles of ASM, which can be seen faintly,
lie at the bottom of the projected volumetric data set. The
specimen was mounted on a custom-made double-sided
slide and imaged from both the adventitial and luminal aspects. Figure 4b was imaged from the luminal surface to a
depth of 145 mm (35 optical sections), where bundles of
nerves, separated by z 200 mm, run parallel with the long
axis of the airway. This distance corresponds to the separation of the mucosal folds that could be seen under the dissecting microscope.
A large montage of the nerve tracts in the segmental
and subsegmental airways of a lobe from an 18-wk fetal
lung is shown in Figure 5a. It measures 11 mm in overall
length and 2.2 mm in external diameter at the proximal
end, reducing to a mean diameter of 200 6 11 mm (n 5 15)
at the distal ends, spanning seven branchings. On the adventitial surface of the more distal airways, smaller trunks,
nerve bundles, and interconnected ganglia were present,
with a network of finer nerves and small ganglia lying over
the ASM. Compared with the 53- to 58-d gestation lungs,
the nerve trunks and associated nerve bundles are now
more defined and compact, with larger ganglia located
proximally in the more mature airways. In general, at least
one and more often two major trunks run the length of
these airways, with branches giving rise to a plexus of
smaller nerves connected by small ganglia lying closer to
the ASM. The continuity of the main nerves could be followed to the extremities of the bronchial tree. A large continuous nerve trunk in this segment measured z 45 mm at
the proximal end, progressively decreasing to z 20 mm diameter at the airway extremities. Similar observations
were made in the lung of the 16-wk-old fetus described
previously. At higher power (not shown), the network of
small-diameter fibers that branched from the larger nerves
to overlie the smooth muscle was similar to that seen in fetal pigs at the end of the first trimester (6). The thin, varicosed fibers lying on the smooth muscle are not yet aligned
with the orientation of the smooth-muscle cells, but connect
between larger nerve bundles without an apparent orientation.
Figure 5b shows a higher-power view of the straight
segment of the right branch of the montage (Figure 5a).
The bundles of ASM cells are regularly arranged around
the circumference of the wall and lie essentially parallel to
one another and perpendicular to the airway. The density
of the ganglia averaged 70/mm 2 in this region and their
556
AMERICAN JOURNAL OF RESPIRATORY CELL AND MOLECULAR BIOLOGY VOL. 20 1999
Figure 5. (a) A composite montage showing the innervation of the adventitial surface of the bronchial tree of a segmental bronchus and
its branches from a human fetal lung, 18 wk gestation. The tissue was stained for PGP 9.5 and a-actin to show both nerves and smooth
muscle. Nerve trunks extend to the most distal airways. Ganglia are present along the trunks and at the divisions of nerve bundles. The
inset shows a higher-power projection of a ganglion at the junction of several nerve trunks (arrow). (b) A higher-power view of the
straight region on the right-hand side of the montage in a showing the disposition of the nerves, ganglia, ASM, and bronchial arteries.
PGP 9.5 (green) stained a plexus of fine nerves containing many small ganglia. a-Actin (red) stained both ASM and arterioles of the
bronchial circulation that accompany the larger nerve trunks, with some less-distinct vessels branching off to overlie the ASM.
Sparrow, Weichselbaum, and McCray: Development of Nerves and Smooth Muscle in Human Fetal Lung
Figure 6. A single optical section of the ganglion shown in Figure
5 cut at a depth of 10.5 mm from its surface showing the neurons,
some with prominent nucleoli. Some neurons show higher immunoreactivity for PGP 9.5 in the nuclear region. Between the neurons, unstained regions are observed that correspond in size and
shape to satellite cells often seen in ganglia. The nerve trunks exhibit unstained cell profiles of Schwann cells (see Figure 2d).
557
size appeared to be proportional to the diameter of the
nerves interconnecting with them. The diameter of the
large, spherically shaped ganglia, which lay 20 to 30 mm
from the smooth muscle, did not exceed 100 mm. There
were many smaller ganglia lying closer to the muscle, some
of which were less than 20 mm in diameter. Brightly staining bronchial arterioles ran along the path of the nerve
trunks and frequently followed them as they divided.
Smaller arterioles branched off to supply the ganglia.
Other vessels, less distinct against the brightly staining
ASM, branched off to overlie the ASM bundles. Blood
vessels were more difficult to detect when the airway diameter decreased below 400 mm diameter because the fluorescent signal from these small vessels was much less intense than that of the large expanse of ASM. However,
they could be revealed by examining only those optical
sections lying above the ASM, where they appeared as
fine, branching arterioles ranging from 10 to 30 mm in diameter.
The large ganglion near the right margin of the main
airway was imaged at higher power (Figure 5a, arrow and
inset), and optical sections at increasing depth were examined. Figure 6 shows an optical section through this ganglion and individual neurons with cell diameters of z 22
mm. Prominent nucleoli were present and appeared as
Figure 7. (a) Videomicrograph of a pulmonary arterial tree from
a lobe of a 16-wk human fetal lung. The large-diameter central
region of the artery is located adjacent to the segmental bronchus
shown in Figure 3. (b) A low-power confocal image of a distal region of the pulmonary vascular bed stained for nerves with PGP
9.5 and for the vascular smooth muscle with a-actin. The fine
nerve bundles are supported by an extracellular matrix. Small
ganglia are present along the length of the nerves and at branching points. (c) A higher-power view of the ganglia with a few fine
fibers distributing to the arterial wall.
558
AMERICAN JOURNAL OF RESPIRATORY CELL AND MOLECULAR BIOLOGY VOL. 20 1999
black holes of 2 to 3 mm diameter within the nuclear region of the neurons. The ganglion was estimated to contain approximately 50 nerve bodies. The unstained cell
profiles are attributed to glial cells and blood vessels.
Schwann cells were seen as unstained, spindle-shaped cell
profiles in the nerve trunks (see also Figures 2d and 2e).
A well-developed pulmonary arterial bed (Figure 7a)
was present in the lobes of the lung at 16 wk gestation.
These pulmonary arteries accompanied the subsegmental
airways (see Figure 3) and comprised a branching structure of smooth-walled vessels. After separating the pulmonary arterial tree from the airways, it was used as a whole
mount, stained for vascular smooth muscle and nerves,
and imaged with the confocal microscope. Figure 7b shows
the distal region of this arterial tree. The outer region of
the vessel walls was strongly stained by the a-actin, and
in the most distal vessels a thin wall with hollow lumen was
seen. A brightly staining plexus of fine nerves (, 17-mm
diameter), with small ganglia-like swellings present at intervals along the length of the nerve bundles and at their
divisions, was observed. In the larger vessels (z 250-mm
diameter), these nerves lay in the surrounding extracellular matrix at a distance of up to z 75 mm from the wall. In
the most distal vessels (, 60-mm diameter) the nerve fibers lay closer, at a distance of z 16 mm. The small ganglia
consisted of less than 10 neurons and gave rise to thin bundles of fibers, some of which could be followed to the vascular smooth muscle (Figure 7c).
The Postnatal Lung
The airways of postnatal lungs were also examined to allow for a more complete view of the development of the
airway innervation. Lungs from two infants who died of
SIDS, at ages 5 and 10 mo, respectively, were studied. Figure 8 shows a representative field of a distal airway (1-mm
diameter) from the 5-mo-old infant, stained for nerves (Figure 8a) and for smooth muscle (Figure 8b). The field shown
was chosen because the rich vascular supply that obscured
most of the airway wall was sparse enough here to allow
the ASM to be imaged. The composite image showing both
nerves and smooth muscle is presented in Figure 8c. Two
nerve trunks, showing numerous folds, run along the longitudinal axis of the airway, with smaller nerve bundles and
varicosed fibers aligned with the smooth-muscle cells. The
loss of sharpness of the fine nerve fibers and the poor definition of the varicosities suggests that some autolysis had
occurred during the 22 h that elapsed between death and
fixation.
Discussion
We have shown here, from early in gestation in the human
fetal lung, that ASM covers the branching epithelial tubules that are destined to become the future bronchial
tree. In turn, the smooth-muscle layer is ensheathed in a
network of nerves and ganglia, and these two tissues persist as an integral part of the airways into postnatal life.
This study commenced at 7.5 wk gestation, which corresponds to the end of the embryonic period and, in the
lung, the early pseudoglandular stage. The airways of the
Figure 8. Nerves and smooth muscle in a whole mount of a distal
airway (1-mm diameter) from the left upper lung lobe of an infant that died of SIDS. The tissue was stained for nerves with
PGP 9.5 (a) and ASM with a-actin (b). c shows the composite image of nerves and smooth muscle. (a) Two nerve trunks (only
part of the lower one can be seen at the bottom of the field) that
show regular folding give rise to smaller nerve bundles and varicosed fibers. (b) Bundles of ASM cells that exhibit partial separation. A large blood vessel with several smaller branches crosses
the field from the left-hand side. (c) This composite image shows
that the general direction of the varicosed nerve fibers is along
the long axis of the ASM bundles.
Sparrow, Weichselbaum, and McCray: Development of Nerves and Smooth Muscle in Human Fetal Lung
developing bronchial tree are conventionally described as
a layer of epithelial cells that are cuboidal at the growing
tips, becoming columnar as the airways enlarge and the
wall thickens; that is, maturation occurs centripetally (4,
18). Our findings show that the airway wall is a more complex structure than is generally appreciated at this early
stage of development.
The ASM runs continuously from the trachea to the
growing tips, where it ends at the base of the terminal sac,
the locus of smooth-muscle differentiation from mesenchymal cells that occurs via epithelial–mesenchymal interaction (19, 20). The muscle cells form a cylindrical layer
around the airway wall, pitched perpendicularly to its long
axis; this orientation is maintained throughout the length
of the airways to the base of the terminal sac, and throughout gestation into postnatal life. At the trachea the ASM is
closely packed, is multilayered, and forms bundles typically
seen in postnatal airways (M. P. Sparrow and H. W. Mitchell, unpublished data), indicative of its greater state of maturity; whereas more distally, the layer of ASM is one to
two cells deep (2). The spontaneous narrowing and relaxation of these epithelial tubules in the first trimester in human (1) and in pig (2, 3) fetal lungs indicates that the
smooth muscle is functional. It was not possible to ascertain whether smooth muscle is present around the epithelial tubules earlier than 53 d of gestation because of the lack
of younger human tissue. However, in embryonic pig lungs
of comparatively earlier gestational age, ASM is present in
the trachea (M. Weichselbaum and M. P. Sparrow, unpublished data). Postnatally, in the distal airways of SIDS
lungs, the ASM was present as bundles, many of which
were partially separated from each other, a feature frequently seen in excised lungs inflated to 25 cm H2O before
fixation. This separation could be caused in part by lengthening of the airways due to inflation, and by the dissection
procedure needed to separate the airways from the surrounding structures. The varicosed nerve fibers overlying
the ASM were typically oriented along the muscle bundles, as seen in distal porcine airways (6).
Neural tissue invested the epithelial tubules, forming a
rich network of ganglia interconnected by nerve trunks at
7.5 to 8 wk gestation. These ganglia (or, rather, ganglia
precursors) appeared as thickenings or swellings of the
nerve trunks. The neural plexus covering the airway wall
gave rise to smaller bundles that descended toward the
ASM (Figure 2c). The large nerve trunks stained with
PGP 9.5 did not appear smooth but exhibited many cellular profiles that were shown to be due to the Schwann
cells, as revealed using the Schwann cell marker S-100.
Some Schwann cell processes were also present in the ganglia. The neurons of the ganglia appeared spherical in
shape, with a diameter of z 8.5 mm, and contained one or
more nucleoli in their nuclei.
By 16 to 18 wk the nerve trunks were more compact
and severalfold larger in diameter. Ganglia also increased
in size and became more spherical in shape. Smaller nerve
bundles divided into fine varicosed fibers overlying the
surface of the smooth-muscle layer, as seen in the fetal pig
lung (6). Direct contact of these fibers with the smoothmuscle cells cannot be established by means of confocal
microscopy, but requires the resolving power of electron
559
microscopy. With ongoing maturation, the orientation of
the thin, varicosed fibers was increasingly aligned to the
smooth-muscle cells. The arrangement of the nerve cells
in the ganglia, with their axons pointing toward the center of the ganglion, gave the appearance of spherical neurons at the surface of the ganglion (Figure 5a, inset). However, in cross-sections (Figure 6), the less regular shape of
the neurons became apparent and axon hillocks were identifiable.
In adult airways, ganglia are considered to be relatively
few and primarily associated with the extrapulmonary airways, with occasional small ganglia present at bifurcations
of intrapulmonary bronchi (21). We found a large number
of ganglia in the segmental and subsegmental airways at
the canalicular stage (16 wk onward), and we predict that
this complement of ganglia persists into postnatal life and
adulthood. The reported infrequency with which ganglia
are detected in postnatal airways is likely to be methodological. The chance of finding them in cross-sections by
conventional histology is low because of the great increase
in airway surface area during growth that results in a substantial separation of the ganglia.
The function of the smooth muscle in the fetal lung in
early gestation is proposed to provide tone to maintain the
lung liquid at a positive pressure to stimulate lung growth
(2, 3). This pressure has been estimated to be about 2.4
to 2.7 cm H2O in third-trimester fetal sheep (22), 1.4 to
2.9 cm H2O in fetal rabbits (M. P. Sparrow and G. Miserocchi, unpublished data), and 1.4 to 4.2 cm H2O in cultured
fetal mouse lungs (23). This force, generated by the ASM
acting across the airway wall and adjacent parenchyma,
may stimulate lung growth by mechanically induced expression of growth factors from early in gestation. Increasing the intraluminal pressure by obstructing the trachea in
the late-term fetal sheep increases the production of insulin-like growth factor II and causes lung hyperplasia (24).
The spontaneous narrowing of the airway wall of the fetal
lung that has been demonstrated in isolated human and
pig fetal lung (1, 2) would confer the additional advantage
of a pulsatile stimulus that is more effective than a static
one in stimulating lung growth in vitro (25, 26).
The function of the neural tissue, particularly in early
gestation, is unlikely to be in neurotransmission but may
contribute trophically to the development of the lung.
Nerves are currently viewed as secreting trophic factors that
influence the growth and survival of their developing target
organs (27)—ASM and mucous glands, in this instance. In
turn, the target organs secrete neurotrophins that maintain
survival of the nerves and may even influence the phenotypic expression of the neurotransmitters secreted during
development (28). The close association of the neural plexus
and the ASM throughout development of the lung makes a
trophic role for the nerves seem possible.
Acknowledgments: This research was supported by a grant from the National
Health and Medical Research Council of Australia to one author (M.P.S.). One
author (P.B.M.) is the recipient of a Career Investigator Award from the American Lung Association. The authors thank the Central Laboratory for Human
Embryology at the University of Washington for providing fetal tissues; Dr. Peter Campbell, Victorian Institute of Forensic Medicine, Melbourne, Australia,
for providing the SIDS lungs; and Ms. Andreia Schineanu for expert technical
assistance and confocal microscopy. The assistance of Astra Australia in meeting the publication costs of the color pages is gratefully acknowledged.
560
AMERICAN JOURNAL OF RESPIRATORY CELL AND MOLECULAR BIOLOGY VOL. 20 1999
References
1. McCray, P. B., Jr. 1993. Spontaneous contractility of human fetal airway
smooth muscle. Am. J. Respir. Cell Mol. Biol. 8:573–580.
2. Sparrow, M. P., S. P. Warwick, and H. W. Mitchell. 1994. Foetal airway motor tone in prenatal lung development of the pig. Eur. Respir. J. 7:1416–
1424.
3. Sparrow, M. P., S. P. Warwick, and A. W. Everett. 1995. Innervation and
function of the distal airways in the developing bronchial tree of fetal pig
lung. Am. J. Respir. Cell Mol. Biol. 13:518–525.
4. Burri, P. H. 1997. Structural aspects of prenatal and postnatal development
and growth of the lung. In Lung Growth and Development. J. A. McDonald, editor. Marcel Dekker, New York. 1–35.
5. Kitaoka, H., P. H. Burri, and E. R. Weibel. 1996. Development of the human fetal airway tree—analysis of the numerical density of airway endtips.
Anat. Rec. 244:207–213.
6. Weichselbaum, M., M. P. Sparrow, and A. W. Everett. 1996. Mapping the
innervation of the bronchial tree in fetal and postnatal pig lung using antibodies to PGP 9.5 and SV2. Am. J. Respir. Cell Mol. Biol. 15:703–710.
7. Thompson, R. J., J. F. Doran, P. Jackson, A. P. Dhillon, and J. Rode. 1983.
PGP 9.5—a new marker for vertebrate neurons and neuroendocrine cells.
Brain Res. 278:224–228.
8. Haley, K. J., J. M. Drazen, R. Osathanondh, and M. E. Sunday. 1997. Comparison of the ontogeny of protein gene product 9.5, chromogranin A and
proliferating cell nuclear antigen in developing human lung. Microsc. Res.
Tech. 37:62–68.
9. Lauweryns, J. M., and L. Van Ranst. 1988. Protein gene product 9.5 expression in the lungs of humans and other mammals: immunocytochemical detection in neuroepithelial bodies, neuroendocrine cells and nerves. Neurosci. Lett. 85:311–316.
10. Bhatnagar, M., D. R. Springall, M. A. Ghatei, P. W. Burnet, Q. Hamid, A.
Giaid, N. B. Ibrahim, F. Cuttitta, E. R. Spindel, and R. Penketh. 1988. Localisation of mRNA and co-expression and molecular forms of GRP gene
products in endocrine cells of fetal human lung. Histochemistry 90:299–
307.
11. Sorokin, S. P., M. Ebina, and R. F. Hoyt, Jr. 1993. Development of PGP 9.5and calcitonin gene-related peptide-like immunoreactivity in organ cultured fetal rat lungs. Anat. Rec. 236:213–225.
12. Mitchell, J. J., S. E. Reynolds, K. O. Leslie, R. B. Low, and J. WoodcockMitchell. 1990. Smooth muscle cell markers in developing rat lung. Am. J.
Respir. Cell Mol. Biol. 3:515–523.
13. Sheppard, M. N., S. S. Kurian, S. C. Henzen-Logmans, F. Michetti, D. Cocchia, P. Cole, R. A. Rush, P. J. Marangos, S. R. Bloom, and J. M. Polak.
1983. Neurone-specific enolase and S-100: new markers for delineating the
innervation of the respiratory tract in man and other mammals. Thorax
38:333–340.
14. Mercer, B. M., S. Sklar, A. Shariatmadar, M. S. Gillieson, and M. E. D’Alton. 1987. Fetal foot length as a predictor of gestational age. Am. J. Obstet.
Gynecol. 156:350–355.
15. Costa, M., J. B. Furness, S. Pompolo, S. J. Brookes, J. C. Bornstein, D. S.
Bredt, and S. H. Snyder. 1992. Projections and chemical coding of neurons
with immunoreactivity for nitric oxide synthase in the guinea-pig small intestine. Neurosci. Lett. 148:121–125.
16. Ebina, M., H. Yaegashi, T. Takashi, M. Motomiya, and M. Tanemura. 1990.
Distribution of smooth muscles along the bronchial tree: a morphometric
study of ordinary autopsy lungs. Am. Rev. Respir. Dis. 141:1322–1326.
17. Junqueira, L. C., J. Carneiro, and R. O. Kelley. 1992. Basic Histology, 7th
ed. Prentice-Hall International, Inc., London.
18. Drews, U. 1995. Color Atlas of Embryology. Thieme Medical Publishers,
Inc., New York.
19. Minoo, P., and R. J. King. 1994. Epithelial mesenchymal interactions in lung
development. Annu. Rev. Physiol. 56:13–45.
20. Spooner, B. S., and N. K. Wessels. 1970. Mammalian lung development: interactions in primordium formation and bronchial morphogenesis. J. Exp.
Zool. 175:445–454.
21. Canning, B. J., and B. J. Undem. 1994. Parasympathetic innervation of airways smooth muscle. In Airways Smooth Muscle: Structure, Innervation
and Neurotransmission. D. Raeburn and M. A. Giembycz, editors.
Birkhäuser Verlag, Basel. 65–77.
22. Vilos, G. A., and G. C. Liggins. 1982. Intrathoracic pressures in fetal sheep.
J. Devel. Physiol. 4:247–256.
23. Blewett, C. J., S. E. Zgleszewski, M. R. Chinoy, T. M. Krummel, and R. E.
Cilley. 1996. Bronchial ligation enhances murine fetal lung development in
whole-organ culture. J. Paediatr. Surg. 31:869–877.
24. Hooper, S. B., V. K. M. Han, and R. Harding. 1993. Changes in lung expansion alter pulmonary DNA synthesis and IGF-II gene expression in fetal
sheep. Am. J. Physiol. 265:L403–L409.
25. Liu, M., S. J. Skinner, J. Xu, R. N. Han, A. K. Tanswell, and M. Post. 1992.
Stimulation of fetal rat lung cell proliferation in vitro by mechanical
stretch. Am. J. Physiol. 263:L376–L383.
26. Liu, M., J. Liu, S. Buch, A. K. Tanswell, and M. Post. 1995. Antisense oligonucleotides for PDGF-b and its receptor inhibit mechanical strain-induced
fetal lung cell growth. Am. J. Physiol. 269:L178–L184.
27. Milner, P., and G. Burnstock. 1994. Trophic factors and the control of
smooth muscle development and innervation. In Airways Smooth Muscle
Development and Regulation of Contractility. D. Raeburn and M. A.
Giembycz, editors. Birkhäuser Verlag, Basel. 1–39.
28. Habecker, H. A., and S. C. Landis. 1994. Noradrenergic regulation of cholinergic differentiation. Science 264:1602–1604.