American Journal of Respiratory Cell and Molecular Biology

Recent studies on the morphogenesis of the pulmonary arteries have focused on nonhuman species such as the chick and the mouse. Using immunohistochemical techniques, we have studied 16 lungs from human embryos and fetuses from 28 d of gestation to newborn, using serial sections stained with a panel of antibodies specific for endothelium, smooth muscle, and extracellular matrix proteins. Cell replication was also assessed. Serial reconstruction showed a continuity of circulation between the heart and the capillary plexus from at least 38 d of gestation. The intrapulmonary arteries appeared to be derived from a continuous expansion of the primary capillary plexus that is from within the mesenchyme, by vasculogenesis. The arteries formed by continuous coalescence of endothelial tubes alongside the newly formed airway. Findings were consistent with the pulmonary arterial smooth muscle cells being derived from three sites in a temporally distinct sequence: the earliest from the bronchial smooth muscle, later from the mesenchyme surrounding the arteries, and last from the endothelial cells. Despite their different origins, all smooth muscle cells followed the same sequence of expression of smooth muscle–specific cytoskeletal proteins with increasing age. The order of appearance of these maturing proteins was from the subendothelial cells outward across the vessel wall and from hilum to periphery. The airways would seem to act as a template for pulmonary artery development. This study provides a framework for studying the signaling mechanisms controlling the various aspects of lung development.

The past few years have seen advances in our understanding of pulmonary endothelial and smooth muscle cell phenotype and function, and yet our understanding of the origins of the human pulmonary arteries is still based on the reconstructions of embryos described by Congdon in 1922 (1). Later studies showed that intrapulmonary preacinar airway and arterial branching is complete by 16 wk of gestation (2), but the origins of the capillary plexus and its connection with the pulmonary arteries have remained unclear. A recent study has shown that in the mouse, the pulmonary circulation appears to derive from two sources (3). In that study, using casts and electron microscopy, segmental arteries were shown to develop by angiogenesis from the extrapulmonary circulation, and simultaneously, the peripheral capillary plexus formed by vasculogenesis from the pulmonary mesenchyme. The communication between the two was completed at 13 to 14 d of gestation (3). In the present study, we have studied the development of the capillary plexus in human embryonic and fetal lungs on serial sections, having immunostained the endothelial cells for expression of cluster of differentiation number 31 (CD31) and von Willebrand factor.

Newly formed endothelial tubes in the systemic and pulmonary circulation become invested by smooth muscle cells, but the origins of these cells are not certain. In vitro studies indicate that smooth muscle cells may be recruited by the endothelium from the surrounding mesenchyme or arise by transdifferentiation from the endothelium (4). In the systemic circulation of the developing quail embryo, mesenchymal cells associating with the endothelium express smooth muscle–specific α-actin (α-SM-actin) and these committed cells then express more smooth muscle cell markers as they mature (5). The phenotype of vascular smooth muscle cells varies, with adult bovine and porcine pulmonary arteries containing several different cell phenotypes, differing in their cytoskeletal composition (6, 7). In the present study, we have explored the hypothesis that the different muscle phenotypes have different developmental origins. We have examined the temporal and spatial pattern of pulmonary arterial smooth muscle maturation by studying the expression of smooth muscle–specific, contractile, and cytoskeletal proteins using immunohistochemical labeling in the human fetal lung.

Cell migration during tissue morphogenesis is regulated in part by the composition of the tissue matrix. Fibronectin is a chemoattractant for mesenchymal cells (8), and in vitro migration of smooth muscle cells is dependent on fibronectin (9). Expression of tenascin mediates cell migration and proliferation (10-12). Versican is expressed at sites of tissue interaction in the developing fetal mouse where it regulates paths of cell migration (13, 14). Therefore, we also examined the temporal and spatial distribution of these three proteins in relation to the development of the pulmonary arteries.

Ten human fetuses, aged 28, 34 (two specimens), 38, 44, 47, 56, 84 (two specimens), and 98 d, were obtained from the Medical Research Council Normal Human Embryo Collection (Institute of Child Health, London, United Kingdom). The fetuses were staged by limb bud and facial features (15). In addition, three lungs from normal fetuses, aged 105, 112, and 140 d, were obtained from the Anatomy Department of the University of Leiden (Leiden, The Netherlands), their collection and use were approved by the Medical Ethics Committee of the Leiden University Medical Centre. The whole fetuses or lungs were fixed in 4% paraformaldehyde for 24 h at 4°C, then rinsed in several changes of 70% alcohol before being processed for wax histology. In addition, we examined diagnostic postmortem lung tissue from three neonates who died with normal lungs. This tissue was taken within 24 h of death and fixed in 10% formaldehyde for a minimum of 24 h, before processing for routine histology. Serial 5-μm sections were cut from the thoracic region of each fetus or tissue sample. In all fetuses, the plane of sectioning was transverse. In addition to routine hematoxylin and eosin and Miller's elastic van Gieson stain, the sections were used for immunohistochemical staining with antibodies specific for either vascular endothelial cells or smooth muscle cells or for matrix proteins associated with cell migration. Endothelial cells were identified by expression of von Willebrand factor and CD31, which is specific for endothelial cells. Smooth muscle cells were identified by α-SM-actin expression and with antibodies for γ-SM-actin and the 204-kD, smooth muscle–specific myosin heavy chain isoform SM1. Expression of the actomyosin regulatory proteins calponin and caldesmon and the smooth muscle– specific intermediate filament protein desmin were also examined together with expression of the matrix proteins fibronectin, tenascin, and versican. In addition, to locate dividing cells, sections were incubated with Kiel University–raised antibody number 67 (Ki-67), which specifically locates cells during replication (Table 1).

Table 1. Immunohistochemical markers

AntibodyCloneDilutionSource
Endothelial-specific proteins
 Monoclonal antihuman CD31JC/70APredilutedDako, Ely, Cambs, UK
 Rabbit antihuman von Willebrand factorA 00821:200Dako, Ely, Cambs, UK
Smooth muscle–specific proteins
 Monoclonal antismooth muscle–specific α-actin1A41:3,000Sigma, Poole, Dorset, UK
 Monoclonal antismooth muscle–specific γ-actinB41:100Dr. J. Lessard, University of
 Cincinnati, Cincinnati, OH
 Monoclonal antismooth muscle MHC SM1MCA1:1,000Yamasa Corp., Tokyo, Japan
MH-011:1,000Sigma, Poole, Dorset, UK
 Monoclonal anticalponinhCP
 Monoclonal anticaldesmonhHCD1:400Sigma, Poole, Dorset, UK
 Monoclonal antidesminZC18PredilutedZymed Labs Inc., San
 Francisco, CA
Matrix proteins
 Rabbit antifibronectin1:500Sigma, Poole, Dorset, UK
 Monoclonal antitenascinBC-241:5,000Sigma, Poole, Dorset, UK
 Polyclonal antiversican1:100Dr. T. Shinomura, Nagakute,  Japan
Cell replication marker
 Monoclonal anti-Ki-677B11PredilutedZymed Labs Inc., San
 Francisco, CA

Before incubation with the primary antibody, the sections were first dewaxed and antigen unmasking was carried out by autoclaving at 121°C for 14 min in citric acid buffer (2.1 g citric acid in 1 liter distilled water, pH adjusted to 6.0). Endogenous peroxidase was then blocked by incubation in 0.3% hydrogen peroxide in methanol and nonspecific binding blocked by incubation with serum-free protein block (Dako Ltd., Cambridge, UK). Sections were incubated in 100 μl of primary antibody diluted in phosphate buffered saline with 0.02% bovine serum albumin for 1 h at room temperature. Antibody binding was visualized using StrepABComplex horseradish peroxidase (Dako Ltd.) and diaminobenzidine. In control sections, incubation with the primary antibody was omitted. Sections were lightly counterstained with hematoxylin before examination.

Initially, sections were examined without knowledge of the age of the embryo. Antibody stain was consistent over the entire section, and sections at different levels in the serial showed similar distribution of each antibody. The diameter of the terminal buds, the height of the epithelium, and the distance between adjacent buds were measured on sections stained with hematoxylin and eosin using an eyepiece graticule. In addition to conventional microscopic study, three-dimensional reconstructions were made from images of serial sections acquired using a Zeiss Axioskop microscope (Carl Zeiss Jena, Welwyn Garden City, UK), ×10 objective, Hamamatsu camera (Hamamatsu Enfield, UK), and the 3-D rendering module of the OpenLab Programme (Improvision, Warwick, UK).

The development of the pulmonary arteries is inextricably linked with that of the airways. By studying serial sections, we found that the airways formed by continuous centrifugal branching within the lung mesenchyme as previously reported (Figure 1A). After each new branch formed, the pulmonary artery formed in situ alongside it from the mesenchyme. In all fetuses throughout the gestational period studied, the appearance of the peripheral airways was similar as represented in Figure 1B and illustrated in Figures 2A through 2D. The number of generations separating the terminal buds from the hilum could be counted on the sections and increased with age (Figures 1A and 2A–2D). There were two generations at 38 d (Figures 1C and 2A) to at least the eight generations that we could identify on a single section at 98 d of gestation. Previous reconstruction studies showed that the adult number of generations, up to 23 in the lower lobe segments, is present by 16 wk of gestation (16). The wall structure of the more proximal, and therefore older, airway and arterial generations differentiated as they increased in size with age. Smooth muscle cell differentiation in both airways and pulmonary arteries could be compared between fetuses of different gestational ages, and within a single fetus between proximal and distal generations.

Pattern of Airway Development

In the 28-d-old fetus, we found a single airway branching from the foregut within the splanchnic mesenchyme surrounding it (Figure 1A). By 34 d of gestation, there were two branches within the mesenchymal sheath that was still continuous with that surrounding the foregut, the prospective left and right bronchi (Figure 2A). By 38 d of gestation, the lung bud (airways and mesenchyme) was separated from the foregut mesenchyme, but there was no structural demarcation between the mesenchymal swelling constituting the lung and the extrapulmonary airways and vasculature (Figure 1C is a drawing derived from the serial reconstruction and Figure 2B, a photomicrograph). Thus, both the extrapulmonary bronchus and the first bifurcation of each bronchus, the prospective lobar bronchi, lay within the mesenchyme of the lung bud. Distal to the lobar branch, there was one further airway generation. By 56 d of gestation, the undifferentiated mesenchyme only surrounded the prospective intrapulmonary airways. With increasing gestational age, there was a progressive reduction in the proportion of the mesenchymal matrix surrounding an increasing number of airways (Figures 2B– 2F). The external diameter of the terminal buds remained constant in the six cases aged 38 to 84 d gestation (∼ 95 μm; Figures 2B–2D) as did the height of the epithelium (32 μm). In the older fetuses (98, 105, 112, and 140 d gestation), there was a progressive reduction in external diameter of terminal buds with increase in age (Figure 2E), and by 140 d, the mean diameter was 43 μm and the epithelium was thinned to 12 μm. The epithelium of these terminal buds was still continuous and was not thinned by underlying capillaries as it is in the later, canalicular stage (17).

In all lungs up to 98 d gestation, an incomplete layer of thin, flattened mesenchymal cells adhered to the base of the epithelial cells of the terminal bud. Some of these cells at the dividing point of the terminal bud expressed α-SM-actin. In the airway giving rise to the terminal bud (Figure 2G, A) and in the generation proximal to this (Figure 2G, B), all the cells in this subepithelial cell layer expressed α-SM-actin and could be considered to be precursor smooth muscle cells. The number of smooth muscle cell layers increased from one layer in level A airways to four layers in level B airways (Figures 1B, 2G, and 2H). The innermost two layers were tightly packed and brick-shaped, whereas the two outer layers were loosely packed and had processes extending into the mesenchyme (arrowheads). Proximal to level B, there were two layers of brick-shaped, tightly packed muscle cells (Figures 1B and 2H). Closer to the hilum, these layers became increasingly fragmented and discontinuous around the airway (Figure 2D). In lungs of older fetuses of 105 to 140 d, instead of there being four layers of smooth muscle cells surrounding level B airways, there was only one layer (Figures 1B and 2I). From level C, this smooth muscle cell layer was fragmented, an appearance seen in all preacinar airways of the newborn lung (Figure 2F).

Cell multiplication, studied by Ki-67 labeling, was greatest in the epithelial cells of the most peripheral airways at all ages. There was little multiplication of bronchial smooth muscle cells except in the outer two smooth muscle cell layers of level B airways at 38 to 98 d gestation (Figure 2J).

Pattern of Arterial Development

At 28 d of gestation, endothelial tubes containing blood cells could be identified in the mesenchyme surrounding the airways and the foregut. At 34 d of gestation, two small vessels could be identified on either side of the prospective trachea. In each lung bud at 38 d gestation, a single pulmonary artery was identified accompanying the prospective extrapulmonary bronchus (Figure 2B). Serial reconstruction showed that these arteries were continuous proximally with pulmonary arteries positioned one on each side of the trachea, which could be traced back to the heart (so-called primitive pulmonary arteries [1]), and peripherally with a plexus of capillaries around the terminal buds. The serial reconstruction is illustrated in Figure 1C. Cells expressing α-SM-actin were present beneath the endothelium of the prospective extrapulmonary arteries (Figure 3A).

A plexus of capillaries labeling with CD31 and von Willebrand factor was present within the mesenchyme of the lung bud in all fetuses studied, from 28 to 140 d gestation. In lungs of the 34-d-old fetus, a basketwork mesh of capillaries could be seen around the airway (Figure 3B compared with Figure 2A). In the 38- to 84-d fetuses, each terminal bud was surrounded by a halo of CD31-labeled capillary endothelial cells about 30 μm from the epithelium (Figure 3C). Many cells in this region, both endothelial and mesenchymal, were Ki-67 positive, whereas there were fewer dividing cells in the general mesenchyme (Figure 3D). Alongside level B airways, the capillaries coalesced to form a progressively larger vessel with a wide lumen. These vessels were lined by flattened (CD31-labeled) endothelial cells (Figure 3E) and were surrounded by an incomplete layer of α-SM-actin positive cells (Figure 3F). The coalescing capillaries were continuous proximally with arteries running parallel with their accompanying airways. The wall of these arteries was made up of a continuous layer of smooth muscle cells, and the endothelial cells now protruded into the narrow lumen (arrows) (Figure 2H). At this level, few of the endothelial cells were replicating.

In the lungs at 98 to 140 d gestation, the relationship of the capillaries to the epithelial cells of the terminal bud altered in that the capillaries at the end of each branching airway were now closely applied to the epithelium (Figure 3G). The endothelial cells of many of these capillaries stained with both CD31 and with α-SM-actin (compare Figures 2I and 3G). These capillaries coalesced to form the arteries alongside the airways proximal to the terminal buds as in the younger cases. Thus, in all the lungs from 28 to 140 d, new arteries appeared to form by vasculogenesis around the terminal airways.

In the newborn lung, there was a complete bed of CD31-positive capillaries throughout the walls of the alveoli; their endothelial cells did not express α-SM-actin (Figure 3H).

Origin of Pulmonary Arterial Smooth Muscle Cells

Our observations suggest that the pulmonary arterial smooth muscle cells may have three origins. First, for a finite period in fetuses of 38 to 98 d of gestation, pulmonary artery smooth muscle cells appeared to derive from the bronchial smooth muscle cells of the adjacent airways in the penultimate generations. Evidence for this was that the two outermost layers of smooth muscle cells surrounding level B airways had cytoplasmic processes extending toward, and making contact with, the coalescing endothelial tubes (Figures 2G, 2H, and 3F), principally on the side of the vessel closest to the airway. Many of these muscle cells were labeled with Ki-67, indicating replication (Figure 2J). Immunostaining with matrix proteins fibronectin, tenascin, and versican showed different distributions through the lung: fibronectin was concentrated around endothelial tubes (Figure 4A), tenascin around the airway and arterial smooth muscle (Figure 4B), and versican throughout the mesenchyme (Figure 4C). However, in the region between the inner smooth muscle cells of the airway and the endothelial tubes, the three matrix proteins colocalized (asterisk), a finding compatible with this being an area of cell movement.

In the more proximal artery generations, one to two layers of brick-shaped, nonreplicating, smooth muscle cells surrounded the endothelium of the newly formed artery (Figure 2H). These inner two layers remained closely packed and brick-shaped, even in the newborn (Figure 4D). Thus, it appeared that the inner one to two smooth muscle layers of most if not all of the preacinar arteries had derived from the bronchial smooth muscle. In the fetuses of 105 to 140 d gestation, there was no longer any connection between the smooth muscle of the airway and arterial walls although capillary tubes were still coalescing at this level (Figure 2I).

The second source of the smooth muscle cells was the lung mesenchyme. By 56 d of gestation, the first intrapulmonary artery at the hilum of the lung bud had two layers of brick-shaped α-SM-actin positive cells that in addition were surrounded by four further layers of fusiform-shaped, loosely packed cells (Figure 4E). The number of layers of fusiform cells increased with age (Figure 4F). These layers initially consisted of mesenchymal cells, circumferentially orientated, containing no α-SM-actin or any other smooth muscle cell–specific protein, but these proteins appeared with age (Figure 4F). Cell multiplication was seen in cells across the whole width of the prospective media. The number of generations having this structure increased with age and was seen over the first six generations from the hilum in the older fetuses. This section of the arterial pathway eventually becomes the elastic intrapulmonary segment of the postnatal lung where at least seven layers of smooth muscle cells and elastic laminae are identified (Figure 4D) (18, 19).

The third possible source for the smooth muscle was from the endothelial cells. In the lungs of fetuses aged 98 to 140 d, many of the peripheral capillary endothelial cells stained with antibodies to α-SM-actin as well as CD31 (Figure 4G). In some of the endothelial tubes, α-SM-actin positive cells were seen wrapped around the endothelial cells (Figure 4H). These singular endothelial cells seemed to be specific for a given time during development. They were not seen before 98 d of gestation nor were they present in capillaries of the neonatal lungs (Figure 3H).

Maturation of Smooth Muscle

With increase in age and with increasing proximity to the hilum, there was progressive maturation of the cytoskeletal proteins in the smooth muscle cells of both airways and arteries (Figures 5 and 6). From 56 to 98 d of gestation, smooth muscle cells of level A airways and the outer two layers of cells surrounding level B airways expressed α-SM-actin and SM1 (Figures 5A and 5B). However, the inner two muscle layers of level B airways also expressed γ-SM-actin and calponin (Figure 5B). More proximally, the airway smooth muscle also expressed the actomyosin regulatory protein caldesmon and the intermediate filament protein desmin (Figure 5C), a marker of mature smooth muscle (20). In the 105- to 140-d fetuses, the staining pattern of the proximal airways was unchanged, but the smooth muscle cells of level A and B airways only expressed α-SM-actin (Figure 5D).

Arterial smooth muscle cells were less mature than the adjacent airway smooth muscle cells in that desmin was not detected in arteries in any fetal or neonatal lung (Figures 5C and 6). For arteries with one to two layers of muscle in their wall, all smooth muscle cells expressed α-SM-actin. At 38 d, only extrapulmonary arteries had smooth muscle cells and these expressed α-SM-actin and SM1 (Figure 5E paired with Figure 3A). In lungs of fetuses age 56 to 84 d gestation, arteries accompanying level B airways and proximally also expressed SM1 (Figure 5A). Level C arterial smooth muscle cells also expressed γ-SM-actin. In the lungs of fetuses of 105-140 d gestation, the smooth muscle cells of level A and B arteries only expressed α-SM-actin although SM1 and γ-SM-actin appeared at level C (Figures 5D and 6). In the neonatal lung, such peripheral arteries, including the respiratory unit arteries, were more mature than those of the 140-d-old fetus and expressed α-SM-actin, SM1, γ-SM-actin, calponin, and caldesmon.

In arteries with more than two muscle cell layers (only seen from age 56 d onward) maturational changes were seen across the width of the wall. From 56 d of gestation onward, the inner two layers of brick-like smooth muscle cells strongly expressed α-SM-actin, SM1, and γ-SM-actin, whereas calponin was weakly expressed and caldesmon became detectable by 84 d of gestation (Figure 5C). At this time and until 112 d gestation, the outer fusiform cells only expressed α-SM-actin and the intensity of stain decreased with increasing distance from the lumen (Figures 4E and 4F). The outer layers, though expressing α-SM-actin, were outside the elastic laminae (Figure 5F). After 112 d, in the largest vessels all fusiform cells showed strong expression of α- and γ-SM-actin, SM1, calponin, and caldesmon, and all were within the elastic laminae of the media. A similar expression to this was seen in the perihilar elastic arteries and in the midlung vessels in the neonatal lung (Figure 4D).

This study has shown that the development of the pulmonary arteries in humans is closely related to that of the airways. Previous studies showed that the branching pathways of the airways and arteries develop together during fetal life, and preacinar branching is complete by 16 wk of gestation (2, 16). The present study on normal human embryonic and fetal lungs suggests that the airways are essential for pulmonary arterial development. They appear to impose organization on randomly formed endothelial tubes in that the tubes line up around the terminal buds of the airways, suggesting an inductive influence from the epithelium. In addition, the airways also appear to be involved in the derivation of at least some of the pulmonary arterial smooth muscle cells. In fetuses of up to 98 d gestation, a cuff of bronchial smooth muscle located just behind the terminal branches appears to give rise to the vascular smooth muscle cells surrounding the adjacent pulmonary arteries as they form from the merging endothelial tubes.

Formation of the Pulmonary Arteries

This study has suggested that the human intrapulmonary arterial branching system, at least from 34 d of gestation, develops by the continuous expansion of the mesenchymal primary capillary plexus. We found by serial reconstruction that there was continuity of circulation between the peripheral capillaries and the heart from at least 38 d of gestation. This is an earlier stage of development than was previously described in the mouse lung where an injection technique showed that the mouse capillary plexus only linked to the hilar vessels at 13 to 14 d gestation, the equivalent of 16 wk gestation in humans (3). In mice, the proximal arteries were thought to develop by angiogenesis. In any species it is generally accepted that the arteries and airway branch from the hilum toward the periphery, accompanied by centrifugal maturation of their walls. The present study confirms that this is true for the human airways but may not be for the pulmonary arteries. It appears that the intrapulmonary arteries are formed by a process of sustained addition of newly formed endothelial tubes at the lung periphery. These arteries increase in size with age and become invested with smooth muscle cells that also mature with age judged by their cytoskeletal protein expression. Thus, the older more proximal arteries have a more mature cytoskeleton than do the peripheral arteries during development, but by birth, all show a similar expression of proteins.

Vascularization during organogenesis proceeds either by vasculogenesis or angiogenesis or both. Vasculogenesis is usually associated with embryonic development (21). We found dividing cells around the terminal buds, and in the same area on adjacent sections, cells were positive for CD31. Serial reconstruction showed that the small groups of CD31-positive cells were continuous with tubes that progressively expanded in size, by definition vasculogenesis (21). This vasculogenesis appeared to continue at least until the end of the glandular period (approximately 15 to 17 wk gestation) when preacinar airway branching is complete (16). There was less mesenchyme surrounding the airways after 16 wk gestation, and the capillaries were more closely associated with the airway walls. Vasculogenesis may only be possible while there is considerable mesenchymal tissue around the airways, in which the endothelial tubes can form. We could not tell whether vasculogenesis continued beyond 17 wk gestation or whether it was superseded by a period of angiogenesis. At this time, the human lung enters the canalicular phase, and although airway branching continues, it is to form the prospective respiratory unit airways (18). Previous experimental studies have shown that there are time-dependent changes in the way in which the mesenchyme and endoderm interact in the developing lung. In the rat lung, by 16 d gestation the endoderm no longer responds to mesenchymal signals to branch (22). In the mouse lung, during the saccular stage apoptotic mesenchymal cells are evident and are not seen in the earlier pseudoglandular, airway branching stage (23).

What may induce the capillary plexus to form around the epithelial bud? Previous experimental graft studies using the embryonic chick and quail lung have shown that the endothelial cells forming the lung vasculature arise from the mesenchyme surrounding the endoderm and are not formed by outgrowths from the host endoderm (24). Pardanaud and coworkers (24) hypothesized that the endoderm induces the emergence of endothelial cells. Using a Millipore filter (Millipore, Bedford, MA) between endoderm and mesenchyme, Taderera (25) showed that vascular tubes only formed if the endodermal, epithelial tube was present, presumably being influenced by a diffusable substance. Vascular endothelial growth factor may perhaps mediate this epithelial-mesenchymal cross talk. This product is detected in the epithelial cells during the glandular stage of the human fetal lung (26), and gene expression has been demonstrated on adult rat and guinea pig epithelia (27, 28).

Endothelial tubes are known to recruit smooth muscle precursor cells that differentiate rapidly around them, within 24 h in the chick (29). In the present study, smooth muscle cells appeared to derive from at least three sites in a temporally distinct sequence. The first (and probably earliest) surprisingly arose from the bronchial smooth muscle cells surrounding the airways just behind the terminal bud. These, a unique population of airway smooth muscle cells, were less mature than those in more proximal airways; they appeared to be dividing and were surrounded by a colocalization of matrix proteins known to be associated with cell migration (11, 14). The stimulus for migration of the airway smooth muscle cells may have been the proximity of the capillary endothelium. In vitro experiments, have shown both a directed migration of arterial smooth muscle cells toward endothelial cells and an increased expression of smooth muscle–specific markers on contact (15, 30). Thus, it appeared that the innermost vascular smooth muscle layers of the pulmonary artery and the bronchial smooth muscle were derived from a common precursor cell type originating from the mesenchyme. These arterial smooth muscle cells appeared different in shape from those of the outer layers, even in the newborn lung.

The second type of pulmonary artery vascular smooth muscle cell appeared when mesenchymal cells were found orientated around the inner two layers of smooth muscle cells. These cells, originally α-SM-actin negative, appeared to develop specific smooth muscle cell markers as gestation advanced. In postnatal life, interstitial fibroblasts are thought to be recruited and transformed into muscle cells during the development of thick-walled vessels seen in hyperoxia-induced pulmonary hypertension in rats (31). Chemotactic factors are probably derived from the endothelium or from existing muscle cells and may include endo-thelin (32), transforming growth factor β (33, 34), and platelet-derived growth factor (35), all of which are known to be chemotactic for smooth muscle cells. These multilayered arteries will become the elastic arteries of the neonatal and adult lungs, and the outer muscle cell layers are likely to produce much of the collagen and elastin in the vessel wall.

In the older fetuses of 98 to 140 d gestation, in the canalicular phase, capillary endothelial tubes coexpressed α-SM-actin and CD31. A similar appearance was described in the canalicular stage of the rat lung (36). Endothelial cells have been reported to give rise to smooth muscle cells in the early muscularization of the chick aorta (4). In vitro, adult bovine aortic endothelial cells can be transformed into spindle-shaped α-SM-actin positive cells by transforming growth factor β1 (37). In the human fetal lung, it is these later formed and more peripheral vessels, ultimately in the alveolar region, that may be formed by angiogenesis (38). However, Burri (39) has proposed that intussesception of the capillary bed may increase alveolar capillary plexuses.

Maturation of the Smooth Muscle Cells

Although each fetus studied presented a snapshot in time, the immunocytochemical studies indicated that both airway and vascular smooth muscle cells matured with age. The more proximal structures, which formed first, were more advanced in their maturation within a lung. Maturation was also faster and therefore more advanced in the airways than in adjacent arteries, as in the developing rat lung (36). Desmin, a marker of differentiated vascular smooth muscle was present in the airways at 56 d, and other investigators have found desmin expressed from 8 wk of age in human tracheal smooth muscle (40). But desmin was not seen in the human pulmonary arteries, even in the newborn cases, confirming earlier studies (41). Nanaev and colleagues (42) found that desmin was only present in certain human systemic vessels during fetal life. Although in the present study the pulmonary artery smooth muscle cells appeared to derive from different sources, they acquired the same cytoskeletal markers in the same order, that is α-SM-actin, SM1, γ-SM-actin, calponin, and then caldesmon, an order showing progressive maturation (43). At birth, all the cells expressed all these proteins. The hilar arteries showed this expression by 105 d gestation, whereas the later-formed arteries in the respiratory region were still immature at 140 d but had caught up by term. Although all the muscle cells expressed the same cytoskeletal proteins at birth, we do not know if the balance between the elements is the same in cells from different origins and they may therefore still have a different phenotype. Further maturational changes are known to occur after birth with the acquisition of more smooth muscle–specific actin, contractile myofilaments, surface dense bodies (44), and after 3 yr of age, desmin (41). The innermost smooth muscle cells retain their distinctive shape and several smooth muscle phenotypes have been described in the postnatal bovine and porcine pulmonary arteries, expressing different cytoskeletal proteins (6, 7).

We can only speculate on the functional implications of our findings. In the fetal airways just proximal to the terminal bud, the smooth muscle cells express α-SM-actin, SM1, γ-SM-actin, and calponin. A confocal microscopic study of a 12-wk-old human fetus (45) also found dense α-SM-actin immunostaining in this region. This fetal airway smooth muscle is capable of spontaneous movement, and in vitro, it can respond to relaxant and contractile agonists (46). In the pulmonary arteries, we found that the lumen was large when the cells contained only α-SM-actin but had the appearance of a contracted vessel having a small lumen with bulging endothelial cells when γ-SM-actin was also present, even though the regulatory proteins calponin and caldesmon were not yet present. The lumen of even the large proximal pulmonary arteries remained narrow until after birth, presumably offering considerable resistance to flow.

In summary, our findings suggest that human preacinar intrapulmonary arteries are formed primarily by vasculogenesis, by coalescence of cells derived from the mesenchyme into endothelial tubes. The pulmonary arterial smooth muscle cells seemed to originate from more than one source, but all appear to follow the same pattern of maturation, expressing the same cytoskeletal proteins sequentially in utero. In this study, we speculate that the airway acts as a template for pulmonary arterial development and as one source of smooth muscle cells. The signaling mechanisms controlling these developmental processes remain to be elucidated.

The authors would like to thank Professor A. C. Gittenberger-de-Groot for providing them with tissue. This work was supported by the British Heart Foundation.

1. Congdon, E. D. 1922. Transformation of the aortic arch system during the development of the human embryo. Carnegie Institution of Contributions to Embryology, Carnegie Institution of Washington. 14:47–110.
2. Hislop A., Reid L.Intrapulmonary arterial development during fetal life : branching pattern and structure. J. Anat.11319723548
3. deMello D., Sawyer D., Galvin N., Reid L. M.Early fetal development of lung vasculature. Am. J. Respir. Cell Mol. Biol.161997568581
4. DeRuiter M. C., Poelmann R. E., VanMunsteren J. C., Mironov V., Markwald R. R., Gittenberger-de-Groot A. C.Embryonic endothelial cells transdifferentiate into mesenchymal cells expressing smooth muscle actins in vivo and in vitro. Circ. Res.801997444451
5. Hungerford J. E., Owens G. K., Argraves W. S., Little C. D.Development of the aortic vessel wall as defined by vascular smooth muscle and extracellular matrix markers. Dev. Biol.1781996375392
6. Frid M. G., Moiseeva E. P., Stenmark K. R.Multiple phenotypically distinct smooth muscle cell populations exist in the adult and developing bovine pulmonary arterial media in vivo. Circ. Res.751994669681
7. Hall S. M., Haworth S. G.Conducting pulmonary arteries: structural adaptation to extrauterine life. Cardiovasc. Res.211987208216
8. Kullman A., Vaillant P., Muller V., Martinet Y., Martinet N.In vitro effects of pentoxifylline on smooth muscle cell migration and blood monocyte production of chemotactic activity for smooth muscle cells: potential therapeutic benefit in the adult respiratory distress syndrome. Am. J. Respir. Cell Mol. Biol.819938388
9. Boudreau N., Turley E., Rabinovitch M.Fibronectin, hyaluronan, and a hyaluronan binding protein contribute to increased ductus arteriosus smooth muscle cell migration. Dev. Biol.1431991235247
10. Hahn A. W., Kern F., Jonas U., John M., Buhler F. R., Resink T. J.Functional aspects of vascular tenascin-C expression. J. Vasc. Res.321995162174
11. Tremble P., Chiquet-Ehrismann R., Werb Z.The extracellular matrix ligands fibronectin and tenascin collaborate in regulating collagenase gene expression in fibroblasts. Mol. Biol. Cell51994439453
12. Mackie E. J.Tenascin in connective tissue development and pathogenesis. Perspect. Dev. Neurobiol.21994125132
13. Bensadoun E. S., Burke A. K., Hogg J. C., Roberts C. R.Proteoglycans in granulomatous lung diseases. Eur. Respir. J.10199727312737
14. Henderson D. J., Copp A. J.Versican expression is associated with chamber specification, septation, and valvulogenesis in the developing mouse heart. Circ. Res.831998523532
15. Larsen, W. J. 1993. Human Embryology. Churchill Livingstone, New York. 281–340.
16. Bucher U., Reid L.Development of the intrasegmental bronchial tree: the pattern of branching and development of cartilage at various stages of intra-uterine life. Thorax161961207218
17. Loosli C. G., Potter E. L.Pre and postnatal development of the respiratory portion of the human lung. Am. Rev. Respir. Dis.801959520
18. Hislop A., Reid L.Pulmonary arterial development during childhood: branching pattern and structure. Thorax281973129135
19. Elliott F. M., Reid L.Some new facts about the pulmonary artery and its branching pattern. Clin. Radiol.161965193198
20. Hungerford J. E., Little C. D.Developmental biology of the vascular smooth muscle cell: building a multilayered vessel wall. J. Vasc. Res.361999227
21. Risau, W. 1991. Vasculogenesis, angiogenesis and endothelial cell differentiation during embryonic development. In The Development of the Vascular System. R. N. Feinberg, G. K. Shereraud, and R. Auerbach, editors. Karger, Basel. 58–68.
22. Shannon, J. M., and R. R. Deterding. 1997. Epithelial-mesenchymal interactions in lung development. In Lung Growth and Development. J. A. McDonald, editor. Marcel Dekker, New York. 81–118.
23. Roman, J. 1997. Cell-cell and cell-matrix interactions in development of the lung vasculature. In Lung Growth and Development. J. A. McDonald, editor. Marcel Dekker, New York. 365–399.
24. Pardanaud L., Yassine F., Dieterlen-Lievre F.Relationship between vasculogenesis, angiogenesis and haemopoiesis during avian ontogeny. Development1051989473485
25. Taderera J. T.Control of lung differentiation in vitro. Dev. Biol.161967489512
26. Shifren J. L., Doldi N., Ferrara N., Mesiano S., Jaffe R. B.In the human fetus, vascular endothelial growth factor is expressed in epithelial cells and myocytes, but not vascular endothelium: implications for mode of action. J. Clin. Endocrinol. Metab.791994316322
27. Monacci W. T., Merril M. J., Oldfield E. H.Expression of vascular permeability factor/vascular endothelial growth factor in normal rat tissues. Am. J. Physiol.2641993C995C1002
28. Berse B., Brown L. F., Van De Water L., Dvorak H. F., Senger D. R.Vascular permeability factor (vascular endothelial growth factor) gene is expressed differentially in normal tissues, macrophages, and tumors. Mol. Biol. Cell31992211220
29. Owens G. K.Regulation of differentiation of vascular smooth muscle cells. Physiol. Rev.751995487517
30. Nicosia R., Villaschi S.Rat aortic smooth muscle cells become pericytes during angiogenesis in vitro. Lab. Invest.731995658665
31. Jones R., Jacobson M., Steudel W.α-smooth muscle actin and microvascular precursor smooth-muscle cells in pulmonary hypertension. Am. J. Respir. Cell Mol. Biol.201999582594
32. Peacock A. J., Dawes K. E., Shock A., Gray A. J., Reeves J. T., Laurent G. J.Endothelin-1 and endothelin-3 induce chemotaxis and replication of pulmonary artery fibroblasts. Am. J. Respir. Cell Mol. Biol.71992492499
33. Postlethwaite A., Keski-Oja J., Moses H., Kang A. H.Stimulation of the chemotactic migration of human fibroblasts by transforming growth factor β. J. Exp. Med.1651987251256
34. Perkett E. A., Pelton R. W., Meyrick B., Gold L. I., Miller D. A.Expression of transforming growth factor-beta mRNAs and proteins in pulmonary vascular remodeling in the sheep air embolization model of pulmonary hypertension. Am. J. Respir. Cell Mol. Biol.1119941624
35. Folkman J., D'Amore P. A.Blood vessel formation: what is its molecular basis? Cell87199611531155
36. Mitchell J. J., Woodcock-Mitchell J., Low R. B., Leslie K. O., Reynolds S. E.Smooth muscle cell markers in developing rat lung.Am. J. Respir. Cell Mol. Biol.31990515523
37. Arciniegas E., Sutton A. B., Allen T. D., Schor A. M.Transforming growth factor beta 1 promotes the differentiation of endothelial cells into smooth muscle-like cells in vitro. J. Cell Sci.1031992521529
38. Short R. H. D.Alveolar epithelium in relation to growth of the lung. Philos. Trans. R. Soc. Lond. B. Biol. Sci.23519503587
39. 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.
40. Leslie K., Mitchell J., Woodcock-Mitchell J., Low R. B.Alpha smooth muscle actin expression in developing and adult human lung. Differentiation441990143149
41. Allen K., Haworth S. G.Cytoskeletal features of immature pulmonary vascular smooth muscle cells and the influence of pulmonary hypertension on normal human development. J. Pathol.1581989311317
42. Nanaev A. K., Shirinsky V. P., Birukov K. G.Immunofluorescent study of heterogeneity in smooth muscle cells of human fetal vessels using antibodies to myosin, desmin, and vimentin. Cell Tissue Res.2661991535540
43. Frid M. G., Shekhonin B. V., Koteliansky V. E., Glukhova M. A.Phenotypic changes of human smooth muscle cells during development: late expression of heavy caldesmon and calponin. Dev. Biol.1531992185193
44. Allen K., Haworth S. G.Human postnatal pulmonary arterial remodelling: ultrastructural studies of smooth muscle cell and connective tissue maturation. Lab. Invest.591988702709
45. Sparrow M. P., Weichselbaum M., McCray P. B.Development of the innervation and airway smooth muscle in human fetal lung. Am. J. Respir. Cell Mol. Biol.201999550560
46. McCray P. B.Spontaneous contractility of human fetal airway smooth muscle. Am. J. Respir. Cell Mol. Biol.81993573580
Address correspondence to: Prof. S. G. Haworth, Unit of Vascular Biology and Pharmacology, Cardiovascular and Respiratory Sciences, Institute of Child Health, University College London, 30 Guilford Street, London WC1N 1EH, UK. E-mail:
Abbreviations: smooth muscle–specific α-actin, α-SM-actin; cluster of differentiation number 31, CD31; smooth muscle–specific γ-actin, γ-SM- actin; Kiel University (Germany)–raised antibody number 67, Ki-67; myosin heavy chain, smooth muscle isoform (204 kD), SM1.

Related

No related items
American Journal of Respiratory Cell and Molecular Biology
23
2

Click to see any corrections or updates and to confirm this is the authentic version of record