DEVELOPMENTAL DYNAMICS 217:62–74 (2000) Regulation of the Hoxa4 and Hoxa5 Genes in the Embryonic Mouse Lung by Retinoic Acid and TGF␤1: Implications for Lung Development and Patterning ALAN I. PACKER,1 KARIMI G. MAILUTHA,1 LORETTE A. AMBROZEWICZ,4 AND DEBRA J. WOLGEMUTH1–5* 1 Department of Genetics and Development, Columbia University College of Physicians and Surgeons, New York, New York 2 Department of Obstetrics and Gynecology, Columbia University College of Physicians and Surgeons, New York, New York 3 Center for Reproductive Sciences, Columbia University College of Physicians and Surgeons, New York, New York 4 Institute of Human Nutrition, Columbia University College of Physicians and Surgeons, New York, New York 5 Herbert Irving Comprehensive Cancer Center, Columbia University College of Physicians and Surgeons, New York, New York ABSTRACT We have previously described a 5‘ cis-acting retinoic acid response element that is required for a subset of Hoxa4 expression, including the midgestation mouse lung. As both retinoids and Hox genes have been implicated in lung development and patterning, we have examined Hoxa4 expression in the developing mouse lung and extended our work on its regulation. At E12.5, a Hoxa4/lacZ transgene is expressed in the mesenchymal compartment of the lung. Later in development expression is restricted to the proximal mesenchyme and is also observed in smooth muscle cells, subepithelial fibroblasts, and alveolar cells. We show that both Hoxa4 and Hoxa5 are upregulated when cultured in the presence of all-trans retinoic acid. In addition, retinoic acid extends the domain of Hoxa4 and Hoxa5 expression to the periphery of the explants where the distal epithelia are developing. Interestingly, the effect of retinoic acid on Hoxa5 expression was not observed in a Hoxa4 mutant background. In contrast, TGF␤1 was found to downregulate both Hoxa4 and Hoxa5 expression in cultured lung explants. We also establish that retinoic acid has the effect of proximalizing the mouse lung when cultured in a serum-free medium, as evidenced by reduced expression of the distal marker surfactant protein-C. Lungs from Hoxa4 mutant embryos exhibited a similar response to retinoic acid, suggesting that Hoxa4 alone is not required for the proximalizing effect. Based on their retinoid-dependent expression, we conclude that members of the group 4 and/or group 5 Hox genes are likely to be involved in patterning of the mouse lung. Dev Dyn 2000;217:62–74. © 2000 Wiley-Liss, Inc. of inductive interactions. These interactions, usually involving reciprocal signals between contiguous epithelial and mesenchymal cell layers, are mediated by soluble and membrane-bound polypeptide growth factors, adhesion molecules, and components of the extracellular matrix. In recent years a great deal of progress has been made in identifying such factors through tissue explant studies, in vivo gene targeting, and the analysis of gene expression patterns. In particular, a number of the proteins and other molecules that are required for the development of the mammalian lung have been identified (reviewed in Hogan and Yingling, 1998). In the mouse embryo, lung development begins at embryonic day 9.5 (E9.5) as the ventral foregut is induced to grow caudally to form an epithelial outgrowth (the pulmonary endoderm) surrounded by mesodermallyderived splanchnic mesenchyme. Thereafter, lung morphogenesis is divided into the pseudoglandular (E9.5E16.2), canalicular (E16.6-E17.4) and terminal sac stages (E17.4-P5), during which branching morphogenesis and septation occur (Ten Have-Opbroek, 1991). Among the secreted factors that have been found to play critical roles in mouse lung development are EGF (Miettinen et al., 1997); TGF␤1 (Serra et al., 1994; Zhao et al., 1996; Zhao et al., 1998); FGF-1, 7, and 10 (Nogawa and Ito, 1995; Cardoso et al., 1997; Bellusci et al., 1997b; Park et al., 1998); HGF (Ohmichi et al., 1998); BMP-4 (Bellusci et al., 1996); and sonic hedgehog (Bellusci et al., 1997a; Pepicelli et al., 1998, Litingtung et al., 1998). Transcription factors such as N-myc (Moens et al., 1992); TTF-1 (Minoo et al., 1995); Gli2 and 3 (Motoyama et al., 1998); Hoxa5 (Aubin et al., 1997); and the retinoid receptors RAR␣ and ␤2 (Men- Key words: Hoxa4; Hoxa5; retinoic acid; TGF␤1; embryonic lung patterning Grant sponsor: National Institutes of Health; Grant number: R01 HD18122; Grant sponsor: United States Department of Agriculture; Grant number: 97–35200 – 4291. *Correspondence to: Debra J. Wolgemuth, Ph.D., Department of Genetics and Development, Columbia University College of Physicians and Surgeons, 630 West 168th Street, New York, NY 10032. E-mail: email@example.com Received 16 July 1999; Accepted 6 October 1999 INTRODUCTION The growth, patterning, and differentiation of tissues in metazoans is orchestrated by a complex series © 2000 WILEY-LISS, INC. Hox REGULATION AND LUNG PATTERNING delsohn et al., 1994b) are also required for normal lung development (reviewed in Cardoso, 1995). The retinoid signaling pathway is of particular interest in lung biology. The RAR␣/␤2 double knockout results in agenesis of the left lung and hypoplasia of the right lung (Mendelsohn et al., 1994). In addition, there is evidence to suggest that maternal-fetal transmission of retinol and/or retinoic acid is involved in lung maturation in the human fetus (reviewed in Chytil, 1992; Shenai et al., 1992). Attempts to elucidate the mechanisms of action of retinoic acid in lung development have focused on organ culture of the rat lung as a model. The addition of all-trans retinoic acid to embryonic rat lung explants promotes the development of proximal airways while suppressing the development of distal epithelia (Cardoso et al., 1995). Proximal epithelial cells are relatively slowly proliferating cells surrounded by a layer of smooth muscle cells or cartilage. These epithelia give rise to the bronchial portion of the adult lung. Distal epithelial cells, by contrast, are relatively rapidly proliferating cells that lack surrounding smooth muscle or cartilage, and they are fated to give rise to the gas exchange surface of the lung—the respiratory acini. It has been known for some time that the fate of the various lung epithelia is completely dependent on signals from the adjacent mesenchyme. Tracheal, proximal, and distal mesenchymal cells specifically promote the differentiation of their own cognate epithelium in tissue recombination experiments where they are grafted to a denuded endodermal layer (Wessells, 1970; Shannon et al., 1998). That retinoic acid might be affecting lung epithelia via the mesenchyme was suggested by tissue and organ culture experiments employing the mouse embryonic lung as a model. Retinoic acid was found to stimulate proliferation of both lung mesenchymal and epithelial cells when co-cultured (Schuger et al., 1993). However, when cultured separately, increased proliferation was observed only in the mesenchymal cells, suggesting that the effect of retinoic acid on the epithelial cells was an indirect one (Schuger et al., 1993). Interestingly, retinoic acid promoted increased branching of mouse embryonic lungs in organ culture, but had no apparent effect on proximal/distal patterning of the kind that had been observed in retinoic acid-treated rat lung explants (Schuger et al., 1993; Cardoso et al., 1995). Whether the observed effects of retinoic acid are due to the direct activation of the synthesis of mesenchymal-derived growth factors, such as the EGFs or FGFs, or whether there are intermediary transcription factors that carry out this program, is not yet clear. Among the genes whose expression has been examined in retinoic acid-treated rat lung explants are the Hox genes, some of which had been shown previously to be retinoid-responsive in midgestation mouse embryos (Kessel and Gruss, 1991). Hox genes encode a family of 39 highly conserved transcription factors involved in the patterning of axial and appendicular structures 63 (reviewed in Krumlauf, 1994) and particular organs (Hsieh-Li et al., 1995; Manley and Capecchi, 1995). In the cultured rat lung, Hoxa2 (Cardoso et al., 1996), Hoxa5 (Bogue et al., 1994), and Hoxb6 (Cardoso et al., 1996) were found to be upregulated by retinoic acid. These Hox genes are expressed in the lung mesenchyme and it was proposed that they mediate the effects of retinoic acid on lung patterning (Cardoso et al., 1996). Recently, it was reported that mice lacking the Hoxa5 gene die perinatally due to respiratory distress that arises from specific defects in lung development (Aubin et al., 1997). The mid-late gestation lung in the Hoxa5 null embryo was found to have a disorganized mesenchymal compartment, reduced branching, thickened alveolar walls, and disorganization of the proximal bronchi and distal respiratory pathways (Aubin et al., 1997). These observations confirm that at least one Hox gene is essential for lung development, and suggest that retinoid-dependent Hox gene expression may be involved in lung patterning. We and others have been studying the regulation and expression of the Hoxa4 gene in mouse embryos. An initial examination of Hoxa4 expression in the embryonic lung by in situ hybridization showed expression in the mesenchymal compartment at E12.5 and E15.5 (Galliot et al., 1989). To begin a characterization of Hoxa4 cis-acting elements, a line of transgenic mice was generated in which it was demonstrated that 3.8 kb of 5' flanking sequence in addition to intronic sequences were necessary and sufficient to drive appropriate lacZ reporter gene expression in the embryo (Behringer et al., 1993; Keegan et al., 1997). In the lung, expression was reported at E11.5, specifically in the lower half of the lung (Behringer et al., 1993). Subsequently, it was found that a conserved retinoic acid response element (RARE) in the 5' flanking region (at ⫺2.9 kb) was specifically required for Hoxa4/lacZ transgene expression in the gut, kidney, and lung when expression was examined at E14.5 (Packer et al., 1998). Given the retinoid-dependent expression of Hoxa4 in the developing lung, the reported effects of retinoic acid on the patterning of the rat lung in organ culture, and the requirement for a neighboring gene, Hoxa5, in mouse lung development, we have undertaken an analysis of Hoxa4 expression and regulation in the embryonic lung. We show that Hoxa4 is expressed at a uniformly high level in the lung in mid-late gestation, and that the specification of its expression may be determined by the opposing effects of all-trans retinoic acid and TGF␤1. In addition, we find that Hoxa5, expressed in the lung in a pattern similar to that of Hoxa4, also responds to all-trans retinoic acid and TGF␤1, but that its response to retinoic acid is abolished in the Hoxa4 mutant background. Finally, we report that all-trans retinoic acid can alter the patterning of cultured embryonic mouse lungs from both wild-type and Hoxa4 mutant backgrounds. 64 PACKER ET AL. RESULTS Expression of the Hoxa4/Lacz Transgene in the Embryonic Mouse Lung We have initiated a survey of Hoxa4 expression in the embryonic mouse lung through the use of our Hoxa4/lacZ transgenic line (Behringer et al., 1993). At E9.5 and E10.5, we observed no transgene expression in the developing lung (not shown), suggesting that expression is initiated at or around E11.5. At E12.5, transgene expression was restricted to the mesenchymal compartment in the lower half of the lobes (Fig. 1A). Of particular interest is the modestly elevated expression apparent in the mesenchymal cells that are directly apposed to the developing epithelia (Fig. 1A). After E12.5 it is no longer possible to distinguish reasonably well between epithelial and mesenchymal compartments in the intact lung, but it is clear at E14.5 that expression of the Hoxa4/lacZ transgene persists at a high level and in a somewhat polarized pattern, with expression specifically absent in the distal, peripheral regions of the lung (Fig. 1B). High level expression continued at E16.5 (Fig. 1C) and E18.5 (Fig. 1D) as the domain of transgene expression was expanded. Sections of the stained lungs revealed the particular cell types expressing Hoxa4. As was evident in the whole lung, expression of the transgene was first seen in the mesenchymal cells at E12.5 (Fig. 2A). At E14.5, in addition to the expression in the mesenchyme, staining was also observed in the smooth muscle cells surrounding the proximal epithelial cells of the bronchi (Fig. 2B). At E16.5 expression was also occasionally seen in subepithelial fibroblasts (Fig. 2C). Finally, at E18.5 expression was apparent in cells of the maturing alveoli (Fig. 2D). This dynamic pattern of expression suggests roles both in patterning of the lung via mesenchymal-epithelial signaling, and in the differentiation of particular cell types. Retinoid-Responsiveness of the Endogenous Hoxa4 and Hoxa5 Genes in the Embryonic Lung As indicated in the Introduction, exogenous retinoids have been found to promote the expression of particular Hox genes in rat lung explants (Cardoso et al., 1996). In addition, our previous results indicate that expression of the murine Hoxa4 gene in the lung requires an intact 5' retinoic acid response element (Packer et al., 1998). To extend these results we examined the expression of the endogenous Hoxa4 gene in response to exogenous all-trans-retinoic acid in an organ culture setting. In these experiments we cultured E14.5 lung explants on 0.8 m Millipore filters in serum-free DMEM/F12 medium (see Experimental Procedures for details of the culture protocol). Initially we cultured the lung explants for 4 hr with or without 10⫺5 M all-trans retinoic acid (Cardoso et al., 1996). This was followed by whole-mount in situ hybridization with a digoxigenin-labeled Hoxa4 riboprobe. The pattern of expression of the endogenous Hoxa4 gene in an untreated explant was quite similar to the pattern of Hoxa4/lacZ transgene expression in that there was a distal, peripheral region where expression was not detected (compare Fig. 1B (transgene) and Fig. 3A (endogenous)). In addition, sections of untreated explants revealed that expression was restricted to the mesenchymal compartment, as observed for expression of the transgene, again suggesting that the transgene expression faithfully recapitulates Hoxa4 expression in the lung (compare Fig. 2A (transgene) and Fig. 3C, D (endogenous)). No specific staining was observed with a sense Hoxa4 riboprobe (not shown). Retinoic acid treatment (10⫺5 M) resulted in an expansion of the domain of expression of Hoxa4 in the lung (Fig. 3B). This expansion was also seen in sections of retinoic acid-treated explants where mesenchymal expression was found to extend to peripheral regions surrounding the distal epithelia (Fig. 3D, F). Since Hoxa5 expression had been shown to be responsive to retinoic acid in the rat lung (Bogue et al., 1994), and since mice lacking Hoxa5 exhibit significant lung defects (Aubin et al., 1997), we also examined the effect of retinoic acid on Hoxa5 in our culture system. E14.5 lung explants were cultured for 4 hr as before either in medium alone or in medium containing 10⫺5 M alltrans retinoic acid. Whole-mount in situ hybridization with a digoxgenin-labeled Hoxa5 riboprobe revealed that the effect of retinoic acid on Hoxa5 expression was at least as robust as that observed on Hoxa4. Hoxa5 expression in the untreated lung was found to be restricted to the proximal mesenchyme and was essentially identical to that of Hoxa4 (Fig. 4A). This pattern is similar to that reported by Aubin and colleagues (1997) in that it is restricted to the mesenchyme, although they did not describe the polarized expression that we observe here. The retinoic acid-treated lung clearly exhibited an increased intensity of Hoxa5 expression that extended to the entire region of the explant, including the most distal regions (Fig. 4B). We also examined the response of each gene to a lower concentration of retinoic acid (10⫺6 M). This concentration also resulted in an expansion of Hox gene expression, but it appeared to be less dramatic than that observed in response to 10⫺5 M (data not shown). Given the demonstrated importance of Hoxa5 in lung development (Aubin et al., 1997), the overlap of Hoxa4 and Hoxa5 expression in the lung at E12.5–E14.5, and the similar response of each gene to exogenous retinoic acid, we were interested in determining if Hoxa4 is in fact required for the response of Hoxa5 to retinoic acid. It has already been demonstrated that Hoxa4 is required for Hoxa5 expression in the anterior portion of the prevertebral column (Aubin et al., 1998). When we examined retinoic acid-treated E14.5 lung explants from embryos lacking Hoxa4 we found that there was no increase in Hoxa5 expression by whole-mount in situ hybridization (Fig. 4C, D). We suggest that the effect of retinoic acid on Hoxa5 is mediated by a shared Hoxa4/Hoxa5 cis-acting enhancer whose activity is af- Hox REGULATION AND LUNG PATTERNING fected by the neo insertion used to generate the Hoxa4 mutation (see Discussion). Effect of TGF␤1 on Hoxa4 and Hoxa5 Expression in the Embryonic Lung A recent study examining cultured murine fetal lungs presented evidence that TGF␤1 (2 ng/ml) could reduce the levels of Hoxb5 protein after three days of culture (Chinoy et al., 1998). This prompted us to examine Hoxa4 and Hoxa5 expression in cultured E14.5 murine lungs after 18 hr of exposure to TGF␤1 at the same concentration. Overall, the effects of TGF␤1 that we observed were the opposite of those observed in response to all-trans retinoic acid. By whole-mount in situ hybridization, Hoxa4 (Fig. 5A, B) and Hoxa5 (Fig. 5C, D) expression was modestly reduced. A direct comparison of the effects of retinoic acid and TGF␤1 on Hoxa4 expression is shown in Fig. 5E, highlighting the opposing effects of these two factors in determining the distribution of Hoxa4 mRNA in the lung. Effect of All-Trans Retinoic Acid on Lung Morphology Previous reports have established that all-trans retinoic acid can affect the patterning of the developing rat lung by promoting the formation of proximal airways and suppressing the formation of distal airways (Cardoso et al., 1995, 1996). However, the addition of exogenous retinoic acid to mouse embryonic lung explants was reported to promote branching morphogenesis with no apparent effect on lung patterning (Schuger et al., 1993). The difference in the responses of rat and mouse tissue to retinoic acid might be attributed to differences in the culture protocols. For example, the medium used by Schuger and colleagues contained 2% fetal calf serum to arrest branching and a maximum concentration of 10⫺6 M retinoic acid. In contrast, Cardoso and colleagues used a serum-free medium and up to 10⫺5 M retinoic acid to establish its effect on patterning. To determine if we could demonstrate an effect of retinoic acid on lung patterning in the mouse, we cultured E12.5 explants in our standard, serum-free culture medium for 48 hr. Examples of such cultures are shown in Figures 6A and B. The explant cultured in the presence of 10⫺5 M retinoic acid (Fig. 6B) clearly differed in appearance from the explant cultured in the absence of retinoic acid (Fig. 6A). In particular, the morphology of the epithelia appeared to have been affected by the retinoic acid treatment in that large airways now extended to the periphery of the lung. To determine if the effect of retinoic acid on the mouse lung was similar to that reported for the rat, we carried out the same culture experiment followed by whole-mount in situ hybridization with a digoxigeninlabeled riboprobe for surfactant protein-C (SP-C). Expression of SP-C has been shown to be restricted to the distal epithelia of the lung (Cardoso et al., 1995). Figures 6C and D show that there is a dramatic decrease in SP-C expression in the mouse lung explant treated 65 with retinoic acid, indicating that retinoic acid has affected the differentiation pathway of the epithelia in the cultured mouse explant. Figures 6E and F show sections of these lungs, again demonstrating the loss of SP-C expression. The normal pattern of SP-C expression in untreated lungs also corresponded with the degree of epithelial cell proliferation, as evaluated by bromodeoxyuridine uptake in E14.5 lung explants over a 24-hour period (Fig. 5G). Most of the characteristically rapidly proliferating cells of the distal airways were located at the periphery of the explant, roughly overlapping with the pattern of SP-C expression. The columnar, proximal epithelial cells were located in the interior of the explant and exhibited significantly less proliferation. Effect of Retinoic Acid on Lung Morphology in the Absence of Hoxa4 It is likely that the dramatic effects of retinoic acid on lung patterning are mediated by a group of factors rather than a single downstream target. Such factors must meet several criteria, including i) expression in the mesenchyme, since the development of the epithelia—including the effects of retinoids on epithelia— requires the presence of mesenchymal cells; ii) sensitivity to exogenous retinoids and/or direct activation by retinoids through a cis-acting retinoic acid response element; and iii) a demonstration that the effect of retinoic acid on lung patterning is in some way altered by the absence of the factor. Previous studies in the rat have suggested that Hox genes are good candidates, particularly since particular genes like Hoxa2 and Hoxb6 are expressed in the lung mesenchyme in a polarized manner and are responsive to retinoic acid (Cardoso et al., 1996). However, the rat is not a convenient genetic model wherein the effect of retinoic acid Fig. 1. (Figure on following page.) Expression of the Hoxa4/lacZ transgene in the mouse lung from E12.5 to E18.5. In all figures, “proximal” refers to the central/rostral regions of the lung, i.e., the bronchial region of the lung, while “distal” refers to the more peripheral and caudal regions of the lung. A: Expression at E12.5 is restricted to the mesenchymal compartment in the lower half of the lobes (arrow). Arrowheads indicate slightly elevated expression in mesenchymal cells directly apposed to the developing epithelia. B: Expression at E14.5 is observed in proximal regions of the lung (P), with the most distal, peripheral region of the lung exhibiting no expression of the transgene (D). Rings of expression surrounding developing proximal epithelia are apparent (arrowhead). C: Expression at E16.5 is again observed in more proximal regions of the lung (arrow), now slightly expanded when compared to the expression at E14.5. D: Expression at E18.5 is observed throughout the lung. Fig. 2. (Figure on following page.) Sections of embryonic lungs expressing the Hoxa4/lacZ transgene. A: Sections reveal that expression at E12.5 is restricted to the mesenchymal compartment (M) and excluded from the epithelial cells (E). B: In sections at E14.5, expression is also observed in the smooth muscle cells (arrowheads) surrounding the epithelial cells of a proximal bronchus. Note the tall columnar epithelium characteristic of proximal airways (arrow). C: In sections at E16.5, expression is observed in subepithelial fibroblasts (arrowheads). D: In sections at E18.5, expression is seen in squamous epithelial cells of maturing alveoli (arrowheads). Figure 1. (Legend on preceding page.) Figure 2. (Legend on preceding page.) Hox REGULATION AND LUNG PATTERNING Fig. 3. Effects of all-trans retinoic acid on expression of the endogenous Hoxa4 gene in the embryonic mouse lung. E14.5 lung explants were cultured for 4 hr (see Experimental Procedures) and Hoxa4 expression determined by whole-mount in situ hybridization. A: In a whole lung explant cultured in the absence of retinoic acid, expression of Hoxa4 is observed in proximal regions (P), but not in distal regions (D). B: A whole lung explant cultured in the presence of retinoic acid exhibiting a large expansion of Hoxa4 expression to more distal, peripheral regions (compare with A). C: 10⫻ magnification of a section of a lung explant cultured in the absence of retinoic acid shows mesenchymal Hoxa4 expression in proximal regions of the lung (P) and lack of expression in epithelia. The 67 bracketed regions include the most distal regions (D) containing epithelia that are not in contact with Hoxa4-expressing mesenchyme. D: 10⫻ magnification of a section of a lung explant cultured in the presence of retinoic acid. The area of Hoxa4-expressing mesenchyme now extends to the periphery of the lung. The bracketed areas indicate that the distal (D), non-expressing regions are only one epithelial layer wide. E: 25⫻ magnification of the section shown in C, with bracketed regions again identifying distal epithelia (D) surrounded by Hoxa4-free mesenchyme. F: 25⫻ magnification of the section shown in D, with bracketed regions enclosing the distal-most epithelia (D) in the explant that are nonetheless in contact with Hoxa4-expressing mesenchyme. 68 PACKER ET AL. Fig. 4. Effects of all-trans retinoic acid on expression of the endogenous Hoxa5 gene in the embryonic mouse lung. A: In a whole lung explant at E14.5 cultured in the absence of retinoic acid, expression of Hoxa5 is restricted to proximal regions (P) and excluded from more distal regions (D). B: A whole lung explant cultured in the presence of 10⫺5 M retinoic acid exhibits elevated Hoxa5 expression and an expansion of that expression to include more distal regions of the explant (compare with A). C: Hoxa5 expression in an E14.5 lung explant from a Hoxa4 mutant embryo cultured for 4 hr in the absence of retinoic acid. D: Hoxa5 expression in an E14.5 lung explant from a Hoxa4 mutant embryo cultured for 4 hr in the presence of 10⫺5 M retinoic acid. There is no significant effect of retinoic acid on Hoxa5 expression in this mutant background (compare with C). on lung patterning can be tested on gene-targeted embryos. In order to determine if the absence of Hoxa4 alone can alter the proximalizing effects of retinoic acid on the mouse lung in culture, we carried out an additional explant experiment in which E12.5 lungs from a Hoxa4 mutant background were cultured with or without 10⫺5 M retinoic acid. The 48-hour culture period was followed by whole-mount in situ hybridization with an SP-C riboprobe to examine the patterning of the lung epithelia. As shown in Figure 7, the effect of retinoic acid on SP-C expression is similar to that observed in lungs from a wild-type background (compare to Fig. 6C,D). DISCUSSION Hox Gene Expression in the Embryonic Mouse Lung We have described the expression of a Hoxa4/lacZ transgene in the embryonic and fetal mouse lung from E9.5 to E18.5. At E12.5 expression was restricted to the mesenchymal compartment of the lung, with strongest expression in mesenchymal cells apposed to the epithelia. At E14.5 expression in the mesenchyme persists, particularly in regions surrounding the proximal epithelia, and was also observed in smooth muscle cells. Expression in subepithelial fibroblasts and alveolar cells was seen at E16.5–E18.5. We had demonstrated previously that Hox REGULATION AND LUNG PATTERNING 69 Fig. 5. Effects of TGF␤1 on Hoxa4 and Hoxa5 expression in the embryonic mouse lung. E14.5 lung explants were cultured for 18 hr followed by whole-mount in situ hybridization. A: Hoxa4 expression in an explant cultured in the absence of TGF␤1. B: Hoxa4 expression in an explant cultured in the presence of TGF␤1 (2 ng/ml). Note the reduced level of expression when compared to the untreated explant. C: Hoxa5 expression in an explant cultured in the absence of TGF␤1. D: Hoxa5 expression in an explant cultured for in the presence of TGF␤1 (2 ng/ml). Note the reduction in Hox gene expression in response to TGF␤1 treatment. E: Comparison of Hoxa4 expression in E14.5 lung explants cultured in retinoic acid (10⫺5 M) for 4 hr (left) and cultured in TGF␤1 (2 ng/ml) for 18 hr (right) and processed in parallel. Again, note the differences in the level of expression, particularly in the distal regions of the lung where Hoxa4 transcription has been promoted by retinoic acid. this transgene faithfully recapitulates the pattern of expression of the endogenous Hoxa4 gene in the embryo, although the lung was not examined in detail (Behringer et al., 1993; Packer et al., 1998). This is supported in the present study by whole-mount in situ hybridization experiments at E14.5 showing 70 PACKER ET AL. that, as for the Hoxa4/lacZ transgene, Hoxa4 is expressed in the proximal mesenchyme. Most of the initial reports of Hox function in vertebrate embryos emphasized their role in the specification of the axial and appendicular skeletons, with functions similar to those suggested for the Drosophila homeotic genes in patterning metameric units of the embryo (reviewed in Krumlauf, 1994). More recent studies have established that Hox genes can also function in the patterning of embryonic structures that are polar but not metameric, such as the urogenital tract (Hsieh-Li et al., 1995) and possibly the hindgut (Roberts et al., 1995) and heart (Searcy and Yutzey, 1998). A similar role has been proposed for Hox genes in lung development (Kappen, 1996; Mollard and Dziadek, 1997). Of the 39 Hox genes, 20 are expressed either in the fetal lung, the neonatal lung, or both (reviewed in Cardoso, 1995; Kappen, 1996). Of note is Hoxa5, whose expression in the proximal mesenchyme (Aubin et al., 1997; this study) is required for normal lung development (Aubin et al., 1997). Most similar to the pattern of Hoxa4 expression, Hoxa2 is restricted to the proximal mesenchyme and mesenchymal derivatives (Cardoso et al., 1996), and Hoxb5 is found in mesenchymal cells at E13.5 and later in subepithelial fibroblasts and late fetal epithelial cells (Volpe et al., 1997). Some Hox genes, like Hoxb3 and Hoxb4, are uniformly expressed throughout the lung mesenchyme (Bogue et al., 1996). Others, like Hoxb6, exhibit a different pattern of expression, being restricted to mesenchyme surrounding the distal epithelia (Cardoso et al., 1996). Although Hoxa5 is the only Hox gene for which there is definitive evidence from gene targeting for a role in lung development, the varying patterns of expression of many other Hox genes along the proximal-distal axis suggest that together these genes have significant roles in lung development and patterning. The elevated expression of Hoxa4 at E12.5 in cells directly apposed to the epi- Fig. 6. Effect of all-trans retinoic acid on lung morphology. A: E12.5 lung explant cultured for 48 hr in the absence of retinoic acid. Note the characteristic T-shaped morphology of the distal epithelia at the periphery of the explant (arrowheads). B: E12.5 lung explant cultured for 48 hr in the presence of 10⫺5 M retinoic acid. Note the large airways extending to the periphery of the explant (arrowheads). C E12.5 lung explant cultured for 48 hr in the absence of retinoic acid and processed for whole-mount in situ hybridization with an SP-C riboprobe. Expression is confined to the distal epithelia. D: E12.5 lung explant cultured for 48 hr in the presence of 10⫺5 M retinoic acid and processed for whole-mount in situ hybridization with an SP-C probe. SP-C expression is greatly reduced, suggesting that retinoic acid has altered the differentiation pathway of the epithelial cells to a more proximal fate. Large airways can be seen extending to the periphery of the explant (arrowheads). E: Section of the untreated explant showing SP-C expression in the distal epithelial cells. F: Section of the retinoic acid-treated explant showing reduced SP-C expression. G: Bromodeoxyuridine incorporation into E14.5 explants cultured for 24 hr. Most of the proliferating cells are in the cuboidal epithelia at the edge of the tissue (arrowheads), corresponding to the region of highest SP-C expression (compare with E). Columnar epithelial cells lining the more proximal airways in the interior of the explant exhibited much less proliferation (P). thelia, followed by its expression at E14.5 in smooth muscle cells surrounding developing proximal airways, suggests that Hoxa4 might have a role in the development of the smooth muscle cell lineage. Interestingly, overexpression of Hoxa4 in the developing gut appears to affect the differentiation of smooth muscle cells from Hox REGULATION AND LUNG PATTERNING 71 Fig. 7. Effect of all-trans retinoic acid on lung morphology in the absence of Hoxa4. A: E12.5 lung explant from a Hoxa4 mutant embryo cultured for 48 hr and processed for SP-C whole-mount in situ hybridization. B: E12.5 lung explant from a Hoxa4 mutant embryo cultured for 48 hr in the presence of 10⫺5 M retinoic acid and processed in parallel with the wild-type explant. The reduction in SP-C expression observed upon retinoic acid treatment is comparable to that observed in wild-type explants (see Fig. 6C, D). the enteric mesenchyme (Tennyson et al., 1993), suggesting a possible parallel function. have been identified (Hoxa1, Hoxb1, Hoxa4, Hoxb4, Hoxd4), only the three group 4 paralogs have been identified in the embryonic lung, with Hoxd4 expression lost after E12.5 (Gaunt et al., 1989). We suggest that Hoxa4 might be a central transcription factor in the mesenchymal response to endogenous retinoids during lung patterning and branching morphogenesis. That mice lacking Hoxa4 have no readily apparent lung defects (Horan et al., 1994) might be due to functional redundancy with other Hox genes, particularly the group 4 paralogs Hoxb4 and Hoxd4, both of which are expressed in the embryonic lung (see below). In addition, the dramatic lung phenotype observed in the RAR␣/␤2 knockout mice (Mendelsohn et al., 1994) suggests that other genes expressed very early in lung development must be essential downstream targets for retinoids. The inability of retinoic acid to promote expression of Hoxa5 in the Hoxa4 mutant background is also of interest. Although Hoxa4 is required for expression of Hoxa5 in the anterior prevertebrae, its expression elsewhere in the embryo is apparently Hoxa4-independent (Aubin et al., 1998), suggesting that it is unlikely that Hoxa4 protein is required in trans to promote Hoxa5 expression in the lung. A more attractive possibility is that a retinoid-dependent, lung-specific enhancer is shared by Hoxa4 and Hoxa5. Interestingly, an 11.1-kb genomic fragment of the mouse Hoxa5 locus fails to promote reporter gene expression in the lung, suggesting that a lung-specific element might be shared with a Retinoid-Dependent Hox Gene Expression in the Embryonic Mouse Lung In addition to our previous work showing the dependence of Hoxa4 lung expression on a conserved retinoic acid response element (Packer et al., 1998), we show here that exogenous retinoic acid can promote the expression of both Hoxa4 and Hoxa5 expression in the proximal mesenchyme at E14.5. In addition to upregulating their expression, retinoic acid also expands the domain of expression of both genes into mesenchymal cells that surround distal epithelia at the periphery of the lung at E14.5. This is similar to the effect of retinoic acid on the expression of the Hoxa2 gene in the rat lung (Cardoso et al., 1996). It has been proposed that the proximalizing effect of retinoids on the lung might be due to their effect on downstream target genes in the mesenchyme whose polarized distribution would be altered by elevated levels of retinoic acid (Cardoso et al., 1996). Both the ubiquitously expressed RAR␣ and RAR␤2 and the mesenchyme-specific RAR␥ are the candidate retinoid receptors to mediate retinoid function in the lung mesenchyme (Dollé et al., 1990). Interestingly, of the three Hox genes whose polarized expression in the proximal mesenchyme has been established (Hoxa2, Hoxa4, Hoxa5), a cis-acting retinoic acid response element has been identified only for Hoxa4. Of the five Hox genes for which such elements 72 PACKER ET AL. more 3' Hox gene such as Hoxa4 (Larochelle et al., 1999). Transcriptional interference in cis by the inserted neo cassette in the Hoxa4 mutant might then interfere with the response of Hoxa5 to exogenous retinoic acid. Opposing Effects of Retinoic Acid and TGF␤1 on Hoxa4 and Hoxa5 Expression Since the retinoid receptors and retinoid binding proteins expressed in the embryonic lung mesenchyme do not appear to be restricted along the proximal-distal axis (Dollé et al., 1990), the specification of Hoxa4 expression in the proximal mesenchyme may depend on an inhibitory factor in the distal mesenchyme. There is very little known about tissue-specific inhibition of Hox gene transcription, but, as stated earlier, there was a report of downregulation of Hoxb5 protein expression in the lung in response to TGF␤1 after three days in culture (Chinoy et al., 1998). Our results indicate that TGF␤1 can inhibit Hoxa4 and Hoxa5 expression in E14.5 explants after only 18 hr. TGF␤-dependent signal transduction appears to depend on the Smad family of proteins. A recent study suggests that the Smad3 and Smad4 proteins exhibit sequence-specific DNA binding to TGF␤-inducible elements in the promoter of the human plasminogen activator inhibitor-type 1 gene (Dennler et al., 1998). Two of these “CAGACA” elements can be found in the 5' Hoxa4 regulatory region, at ⫺2.1 kb (AAACAGACA) and ⫺1.8 kb (GCACAGACA), although it is not known whether these are actually functional elements mediating TGF␤1-dependent inhibition of Hoxa4 expression. Mutagenesis of these sequences may determine their roles, if any, in vivo. Perhaps most relevant to this study, the essential Smad signaling protein, Smad4, is expressed most highly in the distal mesenchyme in the lung (Zhao et al., 1998). This leads to a possible model of Hoxa4 regulation in the embryonic lung, whereby retinoids activate Hoxa4 throughout the mesenchyme, followed by TGF␤1/Smad4-dependent inhibition in the distal mesenchyme. Alternatively, other members of the TGF␤ gene superfamily may play a role in inhibiting Hox gene expression in the lung, including BMP4, which is highly expressed in the distal endoderm (Bellusci et al., 1996). The Role of Retinoid-Dependent Hoxa4 Expression in Lung Development and Patterning We have shown that exogenous all-trans retinoic acid has the effect of proximalizing E12.5 mouse lung explants as evidenced by loss of expression of the distal epithelial marker SP-C. The establishment of this effect in the mouse model should allow for the identification of molecules downstream of the retinoids that are required in the mesenchyme for patterning of the lung epithelium. To this end we have tested the effects of retinoic acid on lung development in the absence of Hoxa4. We find that the absence of Hoxa4 by itself does not affect the ability of exogenous retinoic acid to proximalize the lung as measured by SP-C expression. As discussed earlier, given the partial redundancy of Hox gene function observed in many experimental settings, (Horan et al., 1995) it is likely that two or more Hox genes are required to mediate the patterning effects of retinoic acid in the lung. The Hoxb4 gene, expressed in the embryonic and fetal lung, was recently shown to regulated by a retinoic acid response element (Gould et al., 1998), raising the possibility that it has an essential retinoid-dependent function in the lung. Alternatively, retinoids may be affecting lung patterning through the regulation of another class of molecules, as yet unidentified. Further experiments such as these, carried out in varying mutant backgrounds, may allow for the identification of developmental pathways and networks that operate in the embryonic and fetal lung. Moreover, Hox genes are expressed in the mesenchymal compartments of other tissues such as the kidney, gut, and reproductive tract, where retinoid receptors have been identified. This approach may therefore be useful in identifying signaling pathways in those tissues where retinoids and mesenchymal-epithelial interactions are known to be essential for proper development. EXPERIMENTAL PROCEDURES Mice The generation of the Hoxa4/lacZ transgenic and Hoxa4 mutant lines has been described (Behringer et al., 1993; Horan et al., 1994). For timed matings, the day of the detection of the vaginal plug was considered embryonic day (E) 0.5, and the developmental stage was determined according to Theiler (1972). Genotyping of Hoxa4 mutant embryos was carried out by Southern analysis using a Hoxa4 digoxigenin-labeled probe (Horan et al., 1994). ␤-Galactosidase Detection Embryonic tissues carrying a lacZ reporter gene were fixed, processed, stained with X-gal (Bachem, Torrance, CA) and processed for histology as described previously (Behringer et al., 1993). Embryonic Lung Organ Culture Embryonic tissues were dissected, rinsed briefly in phosphate-buffered saline, pH 7.4, and placed on floating 0.8 m nitrocellulose filters (Millipore, Bedford, MA) in a Falcon 24-well plate. Care was taken to place the tissues at the liquid/air interface. The culture medium (1 ml/well) consisted of DMEM/F12 (Gibco, Grand Island, NY) containing 1X insulin/transferrin/selenium (ITS) supplement (Gibco) and 1% penicillin/streptomycin. All cultures were maintained in a humidified 37°C incubator at 5% CO2. All-trans-retinoic acid-treated cultures (Sigma Chemical Company, St. Louis, MO) contained 10⫺5 M all-trans-retinoic acid diluted from the original stock in dimethyl sulfoxide (DMSO) while control cultures contained DMSO alone at an equiva- 73 Hox REGULATION AND LUNG PATTERNING lent concentration. Cultures treated with human TGF␤1 (Genzyme) contained 2 ng/ml diluted in DMEM/F12 culture medium from the original stock. In cultures maintained for 48 hr, the medium was replenished after 24 hr. Whole-Mount In Situ Hybridization At the end of the culture period, tissues were fixed overnight in freshly prepared 4% paraformaldehyde and then processed for whole-mount in situ hybridization according to Wilkinson (1993). Riboprobes were labeled with digoxigenin using the Boehringer Mannheim RNA labeling kit, according to the manufacturer’s instructions. The template for synthesis of the Hoxa4 antisense riboprobe was a 0.8 kb XbaI-HindIII fragment corresponding to the 3' untranslated region. The template for synthesis of the Hoxa5 antisense riboprobe was a 1.2 kb EcoRI-HindIII insert (Watrin and Wolgemuth, 1993). This template was linearized at an internal BglII site to allow synthesis of a 0.8 kb probe that excludes the homeobox. The template for synthesis of the SP-C probe was the 571 base pair rat cDNA described in Cardoso et al. (1995). Whole-mount in situ hybridization was carried out exactly according to Wilkinson (1993), except that the prehybridization and hybridization steps were performed at 70°C. Occasionally, fixed tissues were stored at ⫺20°C in 100% methanol for up to 10 days before in situ hybridization. No significant differences in staining were observed in stored tissues. The alkaline phospatase-coupled antidigoxigenin antibody and NBT/BCIP detection reagents were from Boehringer Mannheim. Following staining, tissues were post-fixed overnight in 4% paraformaldehyde, placed in methanol for 10 min and isopropanol for 15 min, and embedded in paraffin. Sections were cut at 15 m, deparaffinized in Histoclear for 2 min, and mounted with the aqueous mounting medium Crystal/Mount (Biomeda). All photomicrographs were taken using Fujicolor 100 film. Sections were viewed and photographed on a Leitz photomicroscope under bright field optics. Whole tissues were viewed and photographed on a Wild Heerbrugg photomicroscope under dark field optics. Bromodeoxyuridine Labeling and Immunohistochemistry To examine cell proliferation, lung explants were cultured in the presence of 10 M bromodeoxyuridine (Sigma) for 24 hr. Explants were fixed overnight in 4% paraformaldehyde, dehydrated, embedded in paraffin, sectioned at 6 M, and immunohistochemistry was carried out with a monoclonal antibody specific for bromodeoxyuridine (Sigma; 1:1000 dilution) and Vectastain ABC reagents (Vector Labs, Burlingame, CA) as described previously (Packer et al., 1997). ACKNOWLEDGMENTS We are grateful to Chris Marshall and Xiangyuan Wang for excellent technical assistance and to Dr. Wellington Cardoso (Boston University) for providing the SP-C probe. We also thank Dr. Jeanine D’Armiento for a critical review of the manuscript. This work was supported by NIH grant R01 HD18122 to D.J.W., a minority supplement to NIH R01 HD18122 to K.G.M., and USDA grant 97–35200 – 4291 to A.I.P. REFERENCES Aubin J, Lemieux M, Tremblay M, Berard J, Jeannotte L. 1997. Early postnatal lethality in Hoxa-5 mutant mice is attributable to respiratory tract defects. Dev Biol 192:432– 445. Aubin J, Lemieux M, Tremblay M, Behringer RR, Jeannotte L. 1998. Transcriptional interferences at the Hoxa4/Hoxa5 locus: Importance of correct Hoxa5 expression for the proper specification of the axial skeleton. Dev Dyn 212:141–156. Behringer RR, Crotty DA, Tennyson VM, Brinster RL, Palmiter RD, Wolgemuth DJ. 1993. Sequences 5' of the homeobox of the Hox-1.4 gene direct tissue-specific expression of lacZ during mouse development. Development 117:823– 833. Bellusci S, Henderson R, Winnier G, Oikawa T, Hogan BLM. 1996. Evidence from normal expression and targeted misexpression that Bone morphogenetic protein-4 (Bmp-4) plays a role in mouse embryonic lung morphogenesis. Development 122:1693–1702. Bellusci S, Furuta Y, Rush MG, Henderson R, Winnier G, Hogan BLM. 1997a. Involvement of sonic hedgehog (Shh) in mouse embryonic lung growth and morphogenesis. Development 124:53– 63. Bellusci S, Grindley J, Emoto H, Itoh N, Hogan BLM. 1997b. Fibroblast growth factor 10 and branching morphogenesis in the embryonic mouse lung. Development 124:4867– 4878. Bogue CW, Gross I, Vasavada H, Dynia DW, Wilson CM, Jacobs HC. 1994. Identification of Hox genes in newborn lung and effects of gestational age and retinoic acid on their expression. Am J Physiol 266:L448 –L454. Bogue CW, Lou LJ, Vasavada H, Wilson CM, Jacobs HC. 1996. Expression of Hoxb genes in the developing mouse foregut and lung. Am J Respir Cell Mol Biol 15:163–171. Cardoso WV, Williams MC, Mitsialis SA, Joyce-Brady M, Rishi AK, Brody JS. 1995. Retinoic acid induces changes in the pattern of airway branching and alters epithelial cell differentiation in the developing lung in vitro. Am J Respir Cell Mol Biol 12:464 – 476. Cardoso WV. 1995. Transcription factors and pattern formation in the developing lung. Am J Physiol 269:L429 –L442. Cardoso WV, Mitsialis SA, Brody JS, Williams MC. 1996. Retinoic acid alters the expression of pattern-related genes in the developing rat lung. Dev Dyn 207:47–59. Cardoso WV, Itoh A, Nogawa H, Mason I, Brody JS. 1997. FGF-1 and FGF-7 induce distinct patterns of growth and differentiation in embryonic lung epithelium. Dev Dyn 208:398 – 405. Chinoy MR, Volpe MV, Cilley RE, Zgleszewski SE, Vostaka RJ, Martin A, Nielsen HC, Krummel TM. 1998. Regulation of Hoxb5 protein in cultured murine fetal lungs. Am J Physiol 274:L610 –L620. Chytil F. 1992. The lungs and vitamin A. Am J Physiol 262:517–527. Dennler S, Itoh S, Vivien D, Dijke P, Huet S, Gauthier J-M. 1998. Direct binding of Smad3 and Smad4 to critical TGF␤-inducible elements in the promoter of human plasminogen activator inhibitor-type 1 gene. EMBO J 17:3091–3100. Dollé P, Ruberte E, Leroy P, Morriss-Kay G, Chambon P. 1990. Retinoic acid receptors and cellular retinoid binding proteins. I. A systematic study of their differential pattern of transcription during mouse organogenesis. Development 110:1133–1151. Galliot B, Dollé P, Vigneron M, Featherstone M, Baron A, Duboule D. 1989. The mouse Hox-1.4 gene: primary structure, evidence for promoter activity and expression during development. Development 107:343–359. Gaunt SL, Krumlauf R, Duboule D. 1989. Mouse homeo-genes within a subfamily, Hox-1.4, 2.6, and -5.1, display similar anteroposterior domains of expression in the embryo, but show stage- and tissuedependent differences in their regulation. Development 107:131– 141. 74 PACKER ET AL. Gould A, Itasaki N, Krumlauf R. 1998. Initiation of rhombomeric Hoxb4 expression requires induction by somites and a retinoid pathway. Neuron 21:39 –51. Hogan BLM, Yingling JM. 1998. Epithelial/mesenchymal interactions and branching morphogenesis of the lung. Curr Opin Genet Dev 8:481– 486. Horan GSB, Wu K, Wolgemuth DJ, Behringer RR. 1994. Homeotic transformation of cervical vertebrae in Hoxa-4 mutant mice. Proc Natl Acad Sci USA 91:12644 –12648. Horan GSB, Ramirez-Solis R, Featherstone M, Wolgemuth DJ, Bradley A, Behringer RR. 1995. Compound mutants for the paralogous Hoxa-4, Hoxb-4, and Hoxd-4 genes show more complete homeotic transformations and a dose-dependent increase in the number of vertebrae transformed. Genes Dev 9:1667–1677. Hsieh-Li HM, Witte DP, Weinstein M, Branford W, Li H, Small K, Potter SS. 1995. Hoxa-11 structure, extensive antisense transcription, and function in male and female fertility. Development 121: 1373–1385. Kappen C. 1996. Hox genes in the lung. Am J Respir Cell Mol Biol 15:156 –162. Keegan LP, Haerry TE, Crotty DA, Packer AI., Wolgemuth DJ, Gehring WJ. 1997. A sequence conserved in vertebrate Hox gene introns functions as an enhancer regulated by posterior homeotic genes in Drosophila imaginal discs. Mech Dev 63:145–157. Kessel M, Gruss P. 1991. Homeotic transformations of murine prevertebrae and concomitant alteration of Hox codes induced by retinoic acid. Cell 67:89 –104. Krumlauf R. 1994. Hox genes in vertebrate development. Cell 78:191– 201. Larochelle C, Tremblay M, Bernier D, Aubin J, Jeannotte L. 1999. Multiple cis-acting regulatory regions are required for restricted spatio-temporal Hoxa5 gene expression. Dev Dyn 214:127–140. Litingtung Y, Lei L, Westphal H, Chiang C. 1998. Sonic hedgehog is essential to foregut development. Nature Genet 20:58 – 61. Manley NR, Capecchi MR. 1995. The role of Hoxa-3 in mouse thymus and thyroid development. Development 121:1989 –2003. Mendelsohn C, Lohnes D, Decimo D, Lufkin T, LeMeur M, Chambon P, Mark M. 1994. Function of the retinoic acid receptors (RARs) during development. (II) Multiple abnormalities at various stages of organogenesis in RAR double mutants. Development 120:2749 – 2771. Miettinen PJ, Warburton D, Bu D, Zhao J, Berger JE, Minoo P, Koivisto T, Allen L, Dobbs L, Werb Z, Derynck R. 1997. Impaired lung branching morphogenesis in the absence of functional EGF receptor. Dev Biol 186:224 –236. Minoo P, Hamdan H, Bu D, Warburton D, Stepanik P, deLemos R. 1995. TTF-1 regulates lung epithelial morphogenesis. Dev Biol 172: 694 – 698. Moens CB, Auerbach AB, Conlon RA, Joyner AL, Rossant J. 1992. A targeted mutation reveals a role for N-myc in branching morphogenesis in the embryonic mouse lung. Genes Dev 6:691–704. Mollard R, Dziadek M. 1997. Homeobox genes from clusters A and B demonstrate characteristics of temporal colinearity and differential restrictions in spatial expression domains in the branching mouse lung. Int J Dev Biol 41:655– 666. Motoyama J, Liu J, Mo R, Ding Q, Post M, Hui C. 1998. Essential function of Gli2 and Gli3 in the formation of lung, trachea, and oesophagus. Nature Genet 20:54 –57. Nogawa H, Ito T. 1995. Branching morphogenesis of embryonic mouse lung epithelium in mesenchyme-free culture. Development 121: 1015–1022. Ohmichi H, Koshimizu U, Matsumoto K, Nakamura T. 1998. Hepatocyte growth factor (HGF) acts as a mesenchyme-derived morphogenic factor during fetal lung development. Development 125:1315– 1324. Packer AI, Elwell VA., Parnass JD, Knudsen KA, Wolgemuth DJ. 1997. N-cadherin protein distribution in normal embryos and in embryos carrying mutations in the homeobox gene Hoxa-4. Int J Dev Biol 41:459 – 468. Packer AI, Crotty DA, Elwell VA, Wolgemuth DJ. 1998. Expression of the murine Hoxa4 gene requires both autoregulation and a conserved retinoic acid response element. Development 125:1991–1998. Park WY, Miranda B, Lebeche D, Hashimoto G, Cardoso WV. 1998. FGF-10 is a chemotactic factor for distal epithelial buds during lung development. Dev Biol 201:125–134. Pepicelli CV, Lewis PM, McMahon AP. 1998. Sonic hedgehog regulates branching morphogenesis in the mammalian lung. Curr Biol 8:1083–1086. Roberts DJ, Johnson RL, Burke AC, Nelson CE, Morgan BA, Tabin CJ. 1995. Sonic hedgehog is an endodermal signal inducing Bmp-4 and Hox genes during induction and regionalization of the chick hindgut. Development 121:3163–3174. Schuger L, Varani J, Mitra Jr. R, Gilbride K. 1993. Retinoic acid stimulates mouse lung development by a mechanism involving epithelial-mesenchymal interaction and regulation of epidermal growth factor receptors. Dev Biol 159:462– 473. Searcy R, Yutzey K. 1998. Analysis of Hox gene expression during early avian heart development. Dev Dyn 213:82–91. Serra R., Pelton RW, Moses HL. 1994. TGFbeta 1 inhibits branching morphogenesis and N-myc expression in lung bud organ cultures. Development 120:2153–2161. Shannon JM, Nielsen LD, Gebb SA, Randell SH. 1998. Mesenchyme specifies epithelial differentiation in reciprocal recombinants of embryonic lung and trachea. Dev Dyn 212:482– 494. Shenai JP, Rush MG, Stahlman MT, Chytil F. 1992. Vitamin A supplementation and bronchopulmonary dysplasia—revisited. J Pediatr 121:399 – 401. Ten Have-Opbroek AAW. 1991. Lung development in the mouse embryo. Exp Lung Res 17:111–130. Tennyson VM, Gershon MD, Sherman DL, Behringer RR, Raz R, Crotty DA, Wolgemuth DJ. 1993. Structural abnormalities associated with congenital megacolon in transgenic mice that overexpress the Hoxa-4 gene. Dev Dyn 198:28 –53. Theiler K. 1972. The house mouse: development and normal stages from fertilization to 4 weeks of age. New York: Springer-Verlag. p 53–128. Volpe MV, Martin A, Vosatka RJ, Mazzoni CL, Nielsen HC. 1997. Hoxb-5 expression in the developing mouse lung suggests a role in branching morphogenesis and epithelial cell fate. Histochem Cell Biol 108:495–504. Watrin F, Wolgemuth DJ. 1993. Conservation and divergence of patterns of expression and lineage-specific transcripts in orthologues and paralogues of the mouse Hox-1.4 gene. Dev Biol 156:136 –145. Wessells NK. 1970. Mammalian lung development: Interactions in formation and morphogenesis of tracheal buds. J Exp Zool 175:455– 466. Wilkinson DG. 1993. In situ hybridization. In: Stern CD, Holland PWH, editors. Essential developmental biology: a practical approach. New York: Oxford University Press. p 257–274. Zhao J, Bu D, Lee M, Slavkin HC, Hall FL, Warburton D. 1996. Abrogation of transforming growth factor-␤ type II receptor stimulates embryonic mouse lung branching morphogenesis in culture. Dev Biol 180:242–257. Zhao J, Lee M, Smith S, Warburton D. 1998. Abrogation of Smad3 and Smad2 or of Smad4 gene expression positively regulates murine embryonic lung branching morphogenesis in culture. Dev Biol 194:182–195.