INTRODUCTION

TOP ABSTRACT INTRODUCTION FUTURE DIRECTIONS REFERENCES

Keywords: sprouty; fibroblast growth factor; ectodomain shedding

Modeling of the lung during morphogenesis is tightly coordinated by stimulatory and inhibitory signaling mechanisms. In humans, branching morphogenesis of 23 generations of airways is completed between 5 and 25 wk of gestation, while alveoli begin to form at 20 wk and expansion of the alveolar surface to the adult size of 70 m² continues for several years postnatally. The proximal pulmonary vessels develop as angiogenic sprouts from systemic arteries and veins, whereas pulmonary capillary vasculogenesis depends on transdifferentiation of capillary endothelium from mesenchyme. Intimate juxtaposition of the alveolar epithelium and endothelium is clearly essential for efficient gas exchange. Pulmonary lymphatics and nerves also develop, but their origin is less well understood (for reviews see Cardoso [1], Perl and Whitsett [2], and Warburton and coworkers [3]).

While the anatomical homology might not seem to be intuitively obvious, informative evolutionary-developmental (evo-devo) and functional conservation parallels have begun to be drawn between basic mechanisms governing branching morphogenesis of the respiratory organs in flies, mice, and humans. In the respiratory system of the fly embryo, gas is delivered to individual cells by segmental paired networks of tubes termed tracheae (Figure 1A). The terminal tracheal branches deliver oxygen to and remove carbon dioxide from individual cells in a highly reproducible, branched network. Among the 50 or more genes that have emerged from functional screening studies as directing fly respiratory organogenesis, Sonic hedgehog, patched, smoothened, Gli, Fgf, Fgfr, sprouty, Bmp4, Tgf-, and Smad are functionally conserved in mice and by homology based inference, also in humans (1, 4). Among these, fibroblast growth factor (FGF) signaling is absolutely required for the initiation and maintenance of tracheal branching in the fly as demonstrated by the Branchless (FGF ortholog) and Breathless (FGF receptor [FGFR] ortholog) null mutant phenotypes (Figure 1B). In contrast, the Sprouty null phenotype is characterized by exuberant tracheal branching.

View larger version (77K): [in this window] [in a new window]   Figure 1.   (A) Drosophila embryo (stage 15  /  16) tracheal system. Immunohistochemistry with a monoclonal antibody (kindly provided by M. Affolter, Biozentrum, University of Basel, Basel, Switzerland) recognizes the tracheal lumen. Tracheae arise segmentally from the paired respiratory placodes. Magnification bar: 50 µm. (B) Schematic representation of tracheal placode morphogenesis in Drosophila (after Metzger and Krasnow [16]). Primary tracheal branching is induced by Branchless, the Fgf ortholog, which signals through Breathless, the Fgfr ortholog. Signaling through Breathless and Branchless is involved in the directional guidance of the primary, secondary, and tertiary tracheal branches. Null mutation of Branchless or Breathless, results in failure of respiratory placodes to form or branch. Conversely, the Sprouty null mutant phenotype results in the formation of supernumerary secondary branches (Hacohen and coworkers [12]).

In mice, FGF signaling is also clearly essential for lung morphogenesis. Not only is FGF signaling essential for branching morphogenesis of the bronchi, but also for postnatal modeling of alveoli as well, because null mutation of Fgf10 abrogates bronchial modeling distal to the carina (17, 18), while double null mutation of Fgfr3 and Fgfr4 abrogates postnatal alveolar morphogenesis (19). Interestingly, the latter phenotype is associated with abnormal elastin deposition, emphasizing the intimate functional association between correct modeling of the epithelium with the matrix.

Studies of the sprouty family have provided further evidence supporting evolutionary, developmental, and functional conservation of key positive and negative regulatory elements in the FGF signaling pathway. Null mutation of sprouty in flies results in exuberant overgrowth of the distal tracheae, suggesting that Sprouty functionally antagonizes FGF signaling (12, 20). Loss of function studies of early mouse embryonic lung culture using mSpry2 antisense oligonucleotides also produced an increased branching phenotype (21), while transgenic misexpression of mSpry2 using the surfactant apoprotein C (SP-C) promoter produced a decreased branching phenotype (15).

Finely coordinated expression of Fgf10 in murine lung peripheral mesenchyme with mSpry2 in the adjacent epithelium at embryonic day 11  /  12 (E11/12) suggests a regulatory paradigm by which these genes may serve to determine interbud branch length during lung modeling. At the point of initiation of a new lobar branch, Fgf10 is highly expressed in the mesenchyme overlying the new bud (Figure 2). Meanwhile, mSpry2 is expressed at a relatively low level in the adjacent bud epithelium, while Fgfr2 is widely expressed throughout the epithelium. As the bud grows out toward the mesenchymal source of FGF10, epithelial mSpry2 expression increases to high levels in the growing tip. Immediately before the next step of bud bifurcation into two segmental bronchi, the site of Fgf10 expression divides laterally into two (Figure 2). Simultaneously, the epithelial mSpry2 expression locus divides into two, adjacent to the two new Fgf10 foci (Figure 3). Bud bifurcation then occurs and the cycle of elongation and gene expression is repeated many times until branching is complete. During subsequent stages of murine lung modeling (E14-16), Fgf10 is expressed not only in each bud tip but is also expressed in a band along the edge of each developing lobe, coincident with the pattern of increased expression of SP-C (15) (Figure 4). The latter Fgf10 spatial distribution resembles the distribution of Fgf8 in the morphogenetic apical ectodermal ridge in the limb bud; thus inviting speculation that Fgf10 also plays an inductive role in lobar shape modeling, as well as in branching. Somewhat similar counterregulatory inductive interactions between Fgf10 and Bmp4 have been mooted in experiments with isolated embryonic lung epithelium (13, 22, 23). However, addition of either Fgf10 or Bmp4 to intact embryonic lung explants at E11/12 stimulates branching morphogenesis. This raises a number of interesting questions about tissue interactions and signaling mechanisms that remain to be addressed.

View larger version (80K): [in this window] [in a new window]   Figure 2.   Whole mount in situ hybridizations demonstrate specific and complementary temporospatial patterns of distribution of (A) Fgf10, (B) mSpry2, and (C ) Fgfr2 in early embryonic mouse lung prepared at E11.5. Note that Fgf10 is expressed in mesenchyme adjacent to the sites of origin of lobar and segmental branches and that mSpry2 is expressed in the peripheral epithelium at sites adjacent to Fgf10. In addition, note that Fgfr2 is more widely expressed throughout the epithelium. It is also noteworthy that the level of mSpry2 expression appears to be low as buds arise and increases as buds elongate. Magnification bars: 100 µm.

View larger version (149K): [in this window] [in a new window]   Figure 3.   Whole mount in situ hybridization vibratome section demonstrating details of mSpry2 expression in peripheral epithelium of E12.5 lung. In this specimen two buds (arrowheads) have recently separated. Note that the epitopes of mSpry2 expression have also divided laterally, leaving a small area adjacent to the cleft (arrow) where mSpry2 is no longer expressed. We speculate that because mSpry2 is strongly induced by Fgf10, this change in distribution of mSpry2 protein reflects lateral movement of Fgf10 expression (also see Figure 4). Magnification bar: 10 µm.

View larger version (91K): [in this window] [in a new window]   Figure 4.   Whole mount in situ hybridization comparing the temporospatial distribution of (A and B) Fgf10 and (C and D) mSpry2 in E14 mouse embryonic lung. Note that at this stage of development, Fgf10 is most highly expressed in the extreme periphery of the developing lung lobes (arrow in panel B). This distribution is reminiscent of the apical ectodermal ridge distribution of Fgf8 in limb buds. Furthermore, in the periphery, Fgf10 expression surrounds the budding tips of both peripheral epithelium on the lobar edges, as well as distal epithelial buds lying within the lobes. Closer examination shows that Fgf10 actually surrounds the epithelial bud tips as they prepare to branch, and as they branch, Fgf10 is expressed on the lateral aspects of the tip where lateral branches are about to arise. The temporospatial distribution of mSpry2 appears to follow that of Fgf10, so that mSpry2 is always expressed in the peripheral epithelium adjacent to sites where Fgf10 is also expressed (arrowhead in panel D). Lobe gauche = left lobe; lobe accessoire = accessory lobe. Magnification bars: 250 µm.

Several key morphogenetic genes are coexpressed in and around peripheral lung buds. Mesenchymal morphogenetic genes expressed in the vicinity of peripheral lung bud epithelium include Fgf10, mSpry4, patched, smoothened, Wnt, and Hox family members. On the other hand, Bmp4, Shh, mSpry2, and Smad2, -3, and -4 are coexpressed in the adjacent peripheral epithelium. This invites speculation that the long-appreciated inductive effects of peripheral lung mesenchyme on airway branching may be mediated through signaling networks that include these genes. Further, possible functional parallels are suggested between inductive centers that direct modeling of other organs that arise at an epithelial/mesenchymal or ectodermal/mesodermal interface, such as the brain and neural tube, tooth knot, feather and hair follicles, limb bud, kidney, salivary gland, and prostate gland.

Many of the above-mentioned morphogenetic molecules are activated from membrane-bound and/or latent forms by limited proteolysis, a process termed "ectodomain shedding." Conversely, shed extracellular domains of transmembrane receptors may compete for ligand with their membrane-anchored forms, thereby modulating the signal amplitude. For example, we have discovered that null mutation of TACE/ADAM17 (tumor necrosis factor  [TNF-]-converting enzyme/A Disintegrin And Metalloprotease 17) results in a neonatal lethal lung hypoplasia phenotype. Airway branching, SP-C and aquaporin-5 expression as epithelial markers, as well as PECAM-1 (platelet endothelial cell adhesion molecule 1) as a marker of vascularization, are all significantly decreased in the absence of TACE (24).

Premature delivery places the underdeveloped alveoli of the human fetus suddenly into an environment where tissue PO₂ and mean airway pressure at least double compared with intrauterine values. This can have serious long-term adverse consequences on alveolar modeling, as well as inducing an acute inflammatory reaction, followed by interstitial fibrosis and emphysema, a clinical syndrome termed bronchopulmonary dysplasia (BPD). Likewise, in adults, exposure to increased levels of inspired oxygen denudes the alveolar epithelium. In human babies with BPD, increased levels of transforming growth factor 1 (TGF-1) ligand are found in airway lavage samples early in the course of the disease, and particularly high levels are associated with an adverse prognosis (14). Excessive amounts of TGF-1 ligand are well known to inhibit embryonic lung morphogenesis in culture, while adenoviral vectors expressing TGF-1 induce severe, progressive pulmonary fibrosis (25, 26). We have also found that the same TGF-1 adenoviral vectors inhibit alveolar modeling during the neonatal period (our unpublished data). Thus, dysregulation of the TGF- signaling pathway appears to be pivotal in the abnormal repair response to lung injury. Furthermore, we have shown that null mutation of Smad3 in mice renders murine lung refractory to the adverse effects of excess TGF- induced by bleomycin administration (our unpublished data). Thus, we postulate that the TGF- signaling pathway should yield effective therapeutic targets to ameliorate abnormal lung repair.

The alveolar epithelial type 2 cell (AEC2) can be considered as a remodeling toolbox. Two key features of the alveolar remodeling process during recovery from injury are AEC2 migration and proliferation, which rapidly recover and reseal the denuded alveolar surface. AECs isolated from fetal rats migrate rapidly and aggressively immediately on isolation: in fact, they migrate as fast as A549 lung epithelial adenocarcinoma cells. This migratory capacity of fetal AECs is compatible with their role in alveolar modeling. In contrast, adult rat AEC2 migrate sluggishly and only after 48 hr in culture. However, during the recovery phase from acute hyperoxia, adult rat AEC2 regain the capability to migrate rapidly, suggesting that this is an important function for alveolar repair. AEC2 migration is further optimized when cells are cultured on fibronectin and exposed to epidermal growth factor (EGF). Fibronectin is also secreted by AECs after injury, whereas EGF is actively synthesized by AEC2 in adult rats and is released from them after injury (27). Both fetal AECs and adult AEC2 after injury secrete and activate matrix metalloprotease 2 (MMP2) and MMP9. The activity of MMP9 is required for AEC migration as well as for keratinocyte growth factor (KGF) to exert its protective effects against hyperoxia induced apoptosis in injured AEC2 (References 6-9; and our unpublished data).

The question therefore arises as to whether the general population of AEC2 can respond to injury by proliferating or whether it is a small subpopulation of AEC stem cells that do the job. AEC2 are normally quiescent in G₀ of the cell cycle, expressing few or no cell cycle proteins (10). Fetal AECs, on the other hand, do proliferate and express several key cyclins and cyclin-dependent kinases (CDKs) (28). After acute hyperoxic injury, AEC2 transiently regain the capacity to proliferate and express numerous cell cycle-related genes (10). We have recently determined that a small (< 5%) population of putative stem cells does indeed exist within the AEC population (11). These putative stem cells express telomerase and are relatively resistant to hyperoxic apoptosis, suggesting that they may enjoy a competitive advantage over the general population in terms of injury resistance. We are currently pursuing their further characterization.

	FUTURE DIRECTIONS

TOP ABSTRACT INTRODUCTION FUTURE DIRECTIONS REFERENCES

Lung regeneration raises intriguing questions about the role of stem cells. It has long been known that, in juvenile rodent at least, partial pneumonectomy is followed quite rapidly by compensatory hypertrophy of the remaining lung. Whether this increase in lung growth is mediated by local stem cell activation is not known for certain but seems likely to be the case. Moreover, the role of circulating stem cells arising from other sites such as the bone marrow or from fat is as yet unknown, but postulated to be potentially important. In view of the exciting developments in pluripotentiality of human and animal embryonic stem cells, a scenario wherein embryonic stem cells, extracorporeally engineered by nuclear transfer to achieve genetic/immunological host compatibility, swim to the injured/ hypoplastic/diseased lung and repopulate damaged alveolar epithelium/vascular endothelium/smooth muscle cell populations no longer seems so much like a "fantastic voyage." Moreover, clues to many important new gene targets in lung morphogenesis, injury repair, and regeneration will be derived from emerging studies in null mutant flies. It will therefore behoove researchers in this field to check their flies regularly.