Workshop Summary

Basic Mechanisms of Lung Development
Eighth Woods Hole Conference on Lung Cell Biology 2000

Wellington V. Cardoso and Mary C. Williams

Pulmonary Center, Department of Medicine, Boston University, Boston, Massachusetts

	Article

Top References

Studies of the molecular regulation of lung development have reached the log-growth phase. The key questions related to lung development are formidable ones to answer, demanding sophisticated and sensitive molecular tools and the use of many different animal, organ, and cellular experimental models. The eighth Y2K Woods Hole Conference on Lung Cell Biology, organized by faculty of the Pulmonary Center, Boston University School of Medicine, and supported by the Division of Lung Diseases of the National Heart, Lung, and Blood Institute, explored some of the recent research in this basic area of lung cell and molecular biology. Many of the fundamental questions, concepts, and observations about basic developmental processes and lung development have been addressed in recent reviews (1).

Fundamental Questions

The fundamental questions in lung development are interesting but difficult to answer in specific molecular terms. Many, if not all, of these questions apply to the development of other organs and are not unique to lung development per se or to the mammalian lung. For that reason a number of presentations summarized the current molecular understanding of model systems including chick limb, tracheal system of Drosophila, and organogenesis of the mammalian liver. Other presentations described postnatal development because it is well known that certain changes in lung structure and function, such as alveolization, occur mainly after birth.

The key questions discussed at the meeting included these: What directs an area of the foregut to become lung, and what cellular and molecular mechanisms account for initiation of budding to form the lung? What are the positive and negative regulators of branching that result in bronchial tree formation? What signaling pathways are involved, and what target genes are activated or repressed to initiate cell proliferation, differentiation, deposition of matrix, and vasculogenesis of the lung? What are the regulatory events that result in the proximal-to-distal patterns of epithelial differentiation? What genes control airway and alveolar dimensions? How is expression of important lung genes specifically regulated? What are the developmental and adult functions of proteins expressed solely and/or mainly in the lung? How plastic is the phenotype of differentiated lung cells, both in developing and adult lung?

It is clear that we have assembled lists of molecules that appear to influence lung development, and other lists describing changes in cellular, molecular, and organ phenotypes due to manipulation of the amount, site, or timing of expression of these molecules. However, we do not understand wellif at allthe integrated regulatory pathways that direct the complex processes that produce this highly patterned, multicellular organ.

Initiation of Lung Bud Formation

What are the events, signaling pathways, and regulatory molecules that initiate lung bud formation from the foregut in the first place? The answer is not certain, although interfering with expression of several molecules can prevent initiation. Studying the formation of liver and pancreas, also foregut derivatives, provides some insight into foregut specification. Direct contact between cardiac mesoderm and foregut epithelium is required for liver induction. Fibroblast growth factors (FGFs) from the mesoderm, including FGF-1, -2, and -8, interact with the endodermal FGF receptors, resulting in the binding to their cognate cis-elements of factors that transactivate liver genes, such as albumin, -fetoprotein, and transthyretin.

The transcription factor pancreatic-duodenum homeobox (PDX)-1 appears to distinguish between sites of liver and pancreas induction. PDX-1 expression is inhibited by cardiac mesoderm, establishing a presumptive liver field. If cardiac mesoderm is absent, the distal endoderm is PDX-1- positive throughout and the liver fails to form. Distal endoderm cells not juxtaposed to cardiac mesoderm express PDX-1 and become pancreas; thus, formation of the pancreas is the default event. Thyroid transcription factor (TTF)-1 is also expressed at localized sites in the foregut epithelium and may identify the lung and thyroid fields; however, disruption of this gene in mice results in agenesis of the thyroid, but not lung, primordia.

Bronchial Tree Formation

Many presentations addressed regulation of bronchial tree branching or patterning. Many of the principles of budding organogenesis, including lung patterning, are based on the understanding of patterning of the chick limb. In the chick, limbs develop on the lateral body surfaces at four specific sites that can be identified by the presence of apical ridges in the ectoderm. Later, a small bud protrudes from the ridge that goes on to develop into the limb with the appropriate number and anterior-posterior pattern of digits and other structures. Formation of this complex structure is believed to be directed by a signaling center that provides positional information to all of the elements involved in formation of the limb.

Analysis of this system suggests that Hox transcription factors organize the head-tail axis of the embryo, and thus where limbs will form in relation to this axis, whereas ventral-to-dorsal positioning is regulated by other factors, such as fringed and engrailed. Together these specify the site where the ridge forms. Cells in this area then produce a growth factor, in this case, a member of the FGF family, that likely affects the migration and proliferation of cells at the interface of the flat body surface and the outgrowing bud. Budding appears to be directed by bidirectional signals between the epithelium and mesenchyme, including sonic hedgehog (Shh), bone morphogenetic proteins (BMPs, members of the transforming growth factor [TGF]- growth factor superfamily), retinoids, and others. These molecules work in several ways: they may act as morphogens that influence cell behavior over a long distance or as competence factors to make cells able to respond to other signals. Shh also restricts its own influence by increasing expression of molecules such as patched (Ptc) that bind Shh protein but do not signal.

Many of the same molecules (or family members) are implicated in regulation of mammalian lung development. Likewise, there appear to be both positive and negative regulators (or inhibitors of the positive regulators). As in the chick limb, these signal back and forth between the mesenchyme and epithelium. A number of presentations described important roles for FGF 10, BMP-4, Shh, Ptc, and retinoids in regulation of branching patterns. In the lung it is not clear where the primary signaling center is located (if it exists), nor is it clear that the molecules that signal initiation of lung bud formation are the same as those that maintain epithelial branching until the bronchial tree is completed. Perhaps the distal lung tips provide the signaling center(s); BMP, Shh, Ptc, and Wnt 7b are found in epithelial or mesenchymal cells at the distal tip. However, in the chick limb, fate maps show that the same group of mesenchymal cells does not interact with the same group of epithelial cells throughout development. If true in the mammalian lung, this implies a complex regulation of key signaling molecules by molecular gradients and feedback loops, particularly in mesenchyme, to allow expression of positional signals in cells in appropriate locations.

A novel Shh binding protein (hedgehog inhibitory protein, HIP) has recently been described whose expression pattern matches that of Shh, although, like Ptc, it is localized to mesenchymal cells rather than to the epithelium where Shh is expressed. Where Shh goes, so goes HIP. This protein binds all hedgehog ligands with the same affinity as Ptc. When Shh is ectopically expressed, HIP is induced nearby and appears to enhance or repress Shh signaling. Thus, in mammals, there are two pathways that can downregulate Shh signaling, one mediated by Ptc and one mediated by HIP. This suggests that, in lung, Shh levels are controlled by both positive and negative regulation to maintain a critical level of signaling.

Many genes have been shown to influence lung development, particularly branching patterns of the bronchial tree, and many models have been used as experimental test systems. These include null mutant (knockout) mice; mice carrying transgenes driven by the surfactant protein (SP)-C or Clara cell secretory protein promoters to test the effects of ectopic, increased, or untimely gene expression; antisense treatment of cultured fetal lungs, and others. Dramatic lung phenotypes have been observed in mice with various null gene mutations. These phenotypes vary from total or unilateral lung agenesis (e.g., zinc finger transcription factor originally isolated from gliomas [Gli] 2/3, FGF-10, RAR/ double knockout); to loss of lobation on the right (e.g., Shh, HIP); to retarded epithelial maturation (e.g., vascular endothelial growth factor [VEGF]) or small, "baggy" alveoli with poorly developed septa (e.g., tumor necrosis factor- converting enzyme); to absence of left pulmonary artery and lung (e.g., RAR/RAR double mutant); and others.

This diversity of phenotypes suggests that there must be an important time-dependent sequence of key regulators, often collectively called the developmental program, that direct each event in lung formation and that this program continues through postnatal alveolar development. However, factors involved in budding, for example, may not be important in alveolization, and vice versa. This is evident from observations that certain molecules, such as Iroquois (IRX)-1 and -2, are expressed in the very early forming lung but are absent by late in gestation. IRX-1 and -2 are molecules downstream from Gli in the Shh signaling pathway, which is likely, therefore, to be more important in early lung development than in later.

Of considerable interest is the report that a null mutation in the SP-C gene does not affect lung development or surfactant function. Because of its early and highly specific pattern of expression, SP-C has been suspected of directing some aspect of lung morphogenesis, perhaps providing spatial cues for development of the peripheral lung. Based on the knockout mouse, this idea appears to be incorrect inasmuch as the null mutant animals have no observable defect in lung development, function, or structure.

Over- or underexpression of certain regulatory molecules also affect lung development. For example, SP-C promoter-driven expression of BMP-4 results in smaller, cystic lungs. Likewise, SP-C-driven -Sprouty-2 results in smaller lungs, fewer branches, failed septal formation, longer bronchioles, and decreased epithelial proliferation. In contrast, decreased expression in cultured lung of Gremlin, a negative regulator of BMP-4 signaling, increases branching, epithelial proliferation, and SP-C expression.

Retinoids in Lung Development

From studies of both lung and chick limb it is clear that retinoids, via retinoid receptors, play a fundamental role in development as shown with several experimental models. Retinoic acid (RA)-signaling is required for formation of the initial lung bud from the foregut in intact animals; however, inhibition of RA signaling in cultured lung buds during branching increases the number of terminal buds. These seemingly disparate observations point to the stage-dependent requirements of retinoid signaling during lung development. Retinoid signaling is controlled, at least in part, by a number of enzymes, such as retinaldehyde dehydrogenase-2, that produce active retinoid molecules; by expression of the retinoid family of receptors (RARs and RXRs); and by enzymes that metabolize RA. These molecules appear to be expressed at specific (and continually changing) sites in the developing lung and together they tightly control the influence of RA on gene expression and overall development.

Important new information on the regulation of diameter and length of tubules of the Drosophila tracheal system suggests principles of growth that may apply to the mammalian lung. Regulation of the pattern of the fly tracheal system involves at least 50 genes, including FGF- and FGF-receptor-like molecules. In this system the characteristic dimensions of each type of tube are not established as tubular branching proceeds, but occur by later remodeling. Changes in tubule diameter result from alterations in cell size, not cell number. Length, however, increases continually during development and there appear to be separate regulators of diameter and length. Like branching, a large number of genes appear to be involved in size regulation, including cystic, gnarled, varicose, bulbous, and others, although the signaling pathways, target genes, and cellular responses are not yet characterized.

Regulation of Key Lung Genes

The surfactant-associated proteins SP-A, -B, -C, and -D have served as important markers for lung development and for identification of epithelial cell types and their precursors. Current information from knockout animals and in vitro studies indicates that only two of the four proteins play important roles in surfactant metabolism, namely, SP-B and SP-D. Nonetheless, more is known about transcriptional regulation of these genes than about other lung genes; these analyses have identified some key transcription factors that participate in regulation of many lung epithelial genes, including TTF-1, hepatocyte nuclear factor 3, GATA 6, and others. Detailed studies of SP-A transactivation show that these and other transcription factors are involved. Oxygen may also influence SP gene expression, as illustrated by the observation that oxygen has a permissive role in cyclic adenosine monophosphate-stimulated SP-A transcription, perhaps resulting from the interaction with other transcription factors and coactivators. These and other observations raise issues about postnatal developmental changes, namely, whether a change in oxygen concentration at birth might play a role in gene induction or modulation. When placed in explant culture, embryonic lung buds also experience a dramatic change in oxygen tension, but whether this influences behavior of this experimental model is not known.

Lung Cell Plasticity

Questions about the plasticity of cell phenotypes is currently a topic of considerable interest. In developing rodent lung the tracheal epithelium can be induced to form SP-A-, -B-, and -C--expressing supernumerary peripheral lung buds by grafting peripheral lung mesenchyme onto it. If gut mesenchyme is used instead of peripheral lung mesenchyme, no supernumary buds are formed. The plasticity of the tracheal mesenchyme is short-lived (a few days). The converse experiment of grafting tracheal mesenchyme aside peripheral lung buds maintains expression of SP-A and -B but SP-C expression is lost. The effects of lung mesenchyme can be mimicked by a growth "cocktail" containing insulin, epithelial growth factor, alveolar lavage fluid, and certain growth factors. Tests of FGF family members show that a combination of FGFs-2 (bFGF), -1 (aFGF), and -7 (keratinocyte growth factor), but not FGF-10, can induce transdifferentiation of the tracheal epithelium.

In the context of Dolly the cloned sheep and the inevitable interpretation that no gene is permanently silenced under normal conditions, these early graft observations are likely to be followed by many showing that lung cells can change their phenotype. Such observations have already been described for many non-lung cell types. Endothelial cells, or perhaps a subset, from adult animals have been shown to transdifferentiate into cells expressing smooth-muscle actin but not endothelial cell markers. The change in phenotype appears to be TGF--dependent.

This raises the possibility that pulmonary vascular smooth muscle, both in normal development and in pathologic conditions such as neonatal and adult pulmonary hypertension, is derived from multiple sources (mesenchyme, endothelium, and airway smooth muscle). Such observations are relevant to understanding formation of the pulmonary vasculature, a developing area of research. These vessels form by connecting vascular sprouts (producing the proximal vessels by angiogenesis) and localized blood lakes (producing the distal vessels by vasculogenesis) and match the bronchial tree as to pattern and time of appearance. It is likely that a large of group of signaling molecules regulates vessel formation, including VEGF derived from the epithelium. VEGF-deficient mice show reduction in both pulmonary vascular and epithelial development, suggesting interdependence of their development, as would be expected.

The Future

It seems reasonable to anticipate a substantial increase in publications on lung development in the next few years as more genes are studied and more cellular responses are identified, and as multigene/protein screening and analytical methods are used. The Y2K+20 problem then becomes one not of generating more and more data but of integrating the multiple observations into fundamental and interactive molecular pathways that together describe lung development from start to finish. Researchers in this field look forward to this challenge with enthusiasm.

	Footnotes

Address correspondence to: Mary C. Williams, R-3, Pulmonary Center, Boston University School of Medicine, 715 Albany St., Boston, MA 02118. E-mail: mwilliams{at}lung.bumc.bu.edu

(Received in original form February 21, 2001 and in revised form April 10, 2001).

	References

Top References

1. Cardoso, W. V.. 2001. Molecular regulation of lung development. Annu. Rev. Physiol. 63: 471-494 [Medline].

2. Glasser, S. W., T. R. Korfhagen, S. E. Wert, and J. A. Whitsett. 1994. Transgenic models for study of pulmonary development and disease. Am. J. Physiol. 267: L489-L497 [Abstract/Free Full Text].

3. Hogan, B. L.. 1999. Morphogenesis. Cell 96: 225-233 [Medline].

4. Mark, M., N. B. Ghyselinck, O. Wendling, V. Dupe, B. Mascrez, P. Kastner, and P. Chambon. 1999. A genetic dissection of the reinoid signaling pathway in the mouse. Proc. Nutr. Soc. 58: 609-613 [Medline].

5. Mendelson, C. R.. 2000. Role of transcription factors in fetal lung development and surfactant protein gene expression. Annu. Rev. Physiol. 62: 875-915 [Medline].

6. Metzger, R. J., and M. A. Krasnow. 1999. Genetic control of branching morphogenesis. Science 284: 1635-1639 [Abstract/Free Full Text].

7. McMahon, A. P.. 2000. More surprises in the hedgehog signaling pathway. Cell 100: 185-188 [Medline].

8. Tickle, C.. 2000. Limb development: an international model for vertebrate pattern development. Int. J. Dev. Biol. 44: 101-108 [Medline].

9. Warburton, D., M. Schwarz, D. Tefft, G. Flores-Delgado, K. D. Anderson, and W. V. Cardoso. 2000. The molecular basis of lung morphogenesis. Mech. Dev. 92: 55-81 [Medline].

10. Zaret, K. S.. 2000. Liver specification and early morphogenesis. Mech. Dev. 92: 83-88 [Medline].