Compensatory Roles of Foxa1 and Foxa2 during Lung Morphogenesis*

THE JOURNAL OF BIOLOGICAL CHEMISTRY © 2005 by The American Society for Biochemistry and Molecular Biology, Inc.

Vol. 280, No. 14, Issue of April 8, pp. 13809 –13816, 2005 Printed in U.S.A.

Compensatory Roles of Foxa1 and Foxa2 during S Lung Morphogenesis*□ Received for publication, December 15, 2004, and in revised form, January 14, 2005 Published, JBC Papers in Press, January 24, 2005, DOI 10.1074/jbc.M414122200

Huajing Wan‡, Sharon Dingle‡, Yan Xu‡, Vale´rie Besnard‡, Klaus H. Kaestner§, Siew-Lan Ang ¶, Susan Wert‡, Mildred T. Stahlman储, and Jeffrey A. Whitsett‡** From the ‡Division of Pulmonary Biology, Cincinnati Children’s Hospital Medical Center and the University of Cincinnati College of Medicine, Cincinnati, Ohio 45229-3039, the §Department of Genetics, University of Pennsylvania School of Medicine, Philadelphia, Pennsylvania 19104-6145, the ¶National Institute for Medical Research, Division of Developmental Neurobiology, London NW7 1AA, United Kingdom, and the 储 Department of Pediatrics, Vanderbilt University School of Medicine, Nashville, Tennessee 37232-2370

Foxa1 and Foxa2 are closely related family members of the Foxa group of transcription factors that are coexpressed in subsets of respiratory epithelial cells throughout lung morphogenesis. Shared patterns of expression, conservation of DNA binding, and transcriptional activation domains indicate that they may serve complementary functions in the regulation of gene expression during lung morphogenesis. Whereas branching morphogenesis of the fetal lung occurs normally in the Foxa2⌬/⌬ and Foxa1ⴚ/ⴚ mice, deletion of both Foxa1 and Foxa2 (in Foxa2⌬/⌬, Foxa1ⴚ/ⴚ mice) inhibited cell proliferation, epithelial cell differentiation, and branching. Dilation of terminal lung tubules and decreased branching were observed as early as embryonic day 12.5. Foxa1 and Foxa2 regulated Shh (sonic hedgehog) and Shh-dependent genes in the respiratory epithelial cells that influenced the expression of genes in the pulmonary mesenchyme that are required for branching morphogenesis. Epithelial cell differentiation, as indicated by lack of expression of surfactant protein B, surfactant protein C, the Clara cell secretory protein, and Foxj1, was inhibited. Foxa family members regulate signaling and transcriptional programs required for morphogenesis and cell differentiation during formation of the lung.

Lung morphogenesis begins with a ventral out-pouching of endodermally derived cells from the anterior foregut into the surrounding mesenchyme at E9 –9.5 1 in the mouse. Lung tubules are formed by branching morphogenesis as respiratory epithelial cells proliferate and differentiate to form the conducting airways. Thereafter, terminal airways sacculate and septate to form the alveoli typical of the peripheral lung. The * This work was supported by National Institutes of Health Grants HL56387, HL75770, and HL38859. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. □ S The on-line version of this article (available at http://www.jbc.org) contains supplemental Figs. 1 and 2 showing PECAM immunohistochemistry and phosphohistone H3 staining, respectively, of Foxa1⫺/⫺/Foxa2⌬/⌬. ** To whom correspondence should be addressed: Cincinnati Children’s Hospital Medical Center, Divisions of Neonatology and Pulmonary Biology, 3333 Burnet Ave., Cincinnati, OH 45229-3039. Tel.: 513636-4830; Fax: 513-636-7868; E-mail: [email protected]. 1 The abbreviations used are: E, embryonic day; ␣-SMA, ␣-smooth muscle actin; CCSP, Clara cell secretory protein; PECAM, platelet endothelial cell adhesion molecule; Shh, sonic hedgehog; SP, surfactant protein; proSP-C, SP-C propeptide. This paper is available on line at http://www.jbc.org

ordered process mediating branching morphogenesis and the formation of the alveoli are regulated by the precise temporalspatial expression of many transcription factors, including Gata6, Ttf1, and forkhead transcription factors, including Foxa1, Foxa2, Foxj1, Foxf1, Foxp1, and Foxp2, that regulate gene expression and influence cell differentiation in the lung (1– 6). Foxa (previously termed HNF3) transcription factors comprise a subfamily of forkhead transcription factors that share ⬎90% homology in the winged helix DNA binding domain. Likewise, the transactivation domains in the C-terminal and N-terminal regions of Foxa1 and Foxa2 share structural similarity (7, 8). Foxa proteins are expressed primarily in endodermally derived tissues where they influence embryonic patterning, cell differentiation, and function (9). The Foxa family includes Foxa1, Foxa2, and Foxa3 (previously termed HNF3␣, HNF3␤, and HNF3␥ respectively). Only Foxa1 and Foxa2 are expressed in the lung (10). In mouse, Foxa2 is first expressed in the primitive streak at E6.5, shortly after the onset of gastrulation (11). Thereafter, Foxa2 is present in the notochord, gut endoderm, and ventral midline of the central nervous system. Later in embryonic development, Foxa2 is detected in organs derived from the foregut endoderm and the hindgut endoderm (10 –12), tissues of ectodermal and mesodermal origins, and endodermally derived epithelial cells in the lung (13). Foxa1 mRNA is detected in the endoderm during the late primitive streak stage of development in the mice. Later in development, Foxa1 mRNA is observed in the ventral floor plate, notochord, and gut in a pattern similar to that in Foxa2 (10 –12, 14, 15). In the lung, the temporal-spatial expression patterns of Foxa1 and Foxa2 are similar. Foxa1 and Foxa2 are present in the nuclei of epithelial cells from the onset of lung bud formation and in the respiratory epithelial cells lining conducting and peripheral airways of the mouse and human fetus (13, 16, 17). Null mutation of the Foxa2 gene in the mouse embryo inhibits formation of the notochord and endoderm before the onset of lung morphogenesis (18, 19). Thus, the potential role of Foxa2 in lung formation and function could not be studied. Ectopic expression of Foxa2 in distal respiratory epithelial cells in the lungs of transgenic mice disrupts branching morphogenesis and arrests differentiation of peripheral epithelial cells, suggesting its potential role in lung formation (20). Recent studies in which Foxa2 was conditionally deleted in the lung epithelial cells showed that Foxa2 is required for normal alveolarization and postnatal lung homeostasis but is not required for branching morphogenesis (21). Despite the expression of Foxa1 throughout lung development, alterations in branching morphogenesis were not observed in

13809

13810

Foxa1 and Foxa2 in Pulmonary Morphogenesis

Foxa1⫺/⫺ mice. Foxa1⫺/⫺ mice survive after birth but generally die between postnatal days 2 and 14, exhibiting growth retardation and hypoglycemia (22, 23). The DNA binding domains of Foxa1 and Foxa2 are highly conserved and bind similar cis-acting elements. Because activation domains are also conserved, Foxa1 and Foxa2 may either compete or serve complementary roles at transcriptional targets (24 –26). Because the pattern of their expression in the developing lung is similar, we hypothesized that Foxa1 and Foxa2 may play partially overlapping roles in lung morphogenesis (13, 16, 17). To assess the potential roles of Foxa proteins in lung formation, Foxa1⫺/⫺/Foxa2⌬/⌬ mice were produced in which both genes were deleted in respiratory epithelial cells in the mouse fetus. EXPERIMENTAL PROCEDURES

Transgenic Constructs—Triple transgenic mice, SPC-rtT⫺/tg/ (tetO)7Cre⫺/tg/Foxa2LoxP/LoxP, were generated by crossing SPCrtTA⫺/tg/Foxa2LoxP/LoxP and (tetO)7Cre⫺/tg mice as reported previously (21). Doxycycline food was administered to the pregnant dam to induce recombination in epithelial cells of the fetal lung resulting in deletion of the Foxa2 gene (SPC-rtTA⫺/tg/(tetO)7Cre⫺/tg/Foxa2⌬/⌬) referred to as Foxa2⌬/⌬. Transgenic mice, SPC-rtTA⫺/tg/(tetO)7Cre⫺/tg/Foxa2LoxP/LoxP/ Foxa1⫺/⫺ (termed Foxa1⫺/⫺/Foxa2LoxP/LoxP), were generated by crossing SPC-rtTA⫺/tg/(tetO)7Cre⫺/tg/Foxa2LoxP/LoxP and Foxa1⫺/⫹. After doxycycline food was administered to the pregnant dam, the Foxa2 gene was selectively deleted in subsets of epithelial cells in the Foxa1⫺/⫺ lung. These mice are termed Foxa1⫺/⫺/Foxa2⌬/⌬. The control genotype is Foxa1⫹/⫹/(tetO)7Cre⫺/tg/Foxa2LoxP/LoxP. The Foxa1⫺/⫺ genotype is Foxa1⫺/⫺/(tetO)7Cr⫺/tg/Foxa2LoxP/LoxP. The Foxa2⌬/⌬ genotype is Foxa1⫹/⫹/SPC-rtTA⫺/tg/(tetO)7Cre⫺/tg/Foxa2LoxP/LoxP. The Foxa1⫺/⫺/ Foxa2⌬/⌬ genotype is Foxa1⫺/⫺/SPC-rtTA⫺/tg/ (tetO)7Cre⫺/tg/ Foxa2LoxP/LoxP. These genotypes were verified by PCR (21, 22). Animal Use and Doxycycline Administration—Mice were housed in pathogen-free conditions in accordance with the Institutional Animal Care and Use Committee guidelines. Gestational age was determined by detection of the vaginal plug (as E0.5) and correlated with the length and weight of each pup at the time of sacrifice. Dams bearing transgenic pups were maintained on food containing doxycycline (625 mg/kg; Harlan Teklad, Madison, WI) from E0. Histology, Immunohistochemistry, in Situ Hybridization, and Electron Microscopy—Fetal lungs were prepared for histology as described previously (27). The antibodies used for immunohistochemistry were pro-SP-C (1:1000, rabbit polyclonal; AB3428, Chemicon), CCSP (1:7500, rabbit polyclonal; kindly provided by Dr. Barry Stripp, University of Pittsburgh), mature SP-B (1:1000, rabbit polyclonal; AB3426, Chemicon), thyroid transcription factor-1 (1:1000, rabbit polyclonal; generated in this lab), platelet endothelial cell adhesion molecule-1 (PECAM-1) (1:500, rat polyclonal; clone CD31, BD Pharmingen), ␣-smooth muscle actin (␣-SMA) (1:15,000, mouse monoclonal; clone 1A4, Sigma-Aldrich), Foxa2 (1:5000, rabbit polyclonal anti-rat Foxa2; generated in this laboratory) (13), Foxa1 (1:2000, guinea pig polyclonal antibody; generated in this laboratory), ␣-SMA (1:10,000, mouse monoclonal), and ␣-phosphohistone H3 (1:400, rabbit polyclonal, US Biological). The specificity of these Foxa1 and Foxa2 antibodies was discussed recently (13). All experiments shown are representative of findings from at least two independent dams, generating at least four triple transgenic offspring that were compared with litter mates. In situ hybridization for Shh mRNA was performed by using 35S-UTP and 35S-CTP labeled riboprobes as described previously (28). Electron microscopy was performed on lung tissue obtained from E15.5 and E18.5 Foxa1⫺/⫺/Foxa2⌬/⌬ mice and litter mate controls after fixation in modified Karnovsky’s fixative as described (20). RNA Isolation and Analysis—Lung RNAs from Foxa1⫺/⫺/Foxa2⌬/⌬ and age-matched control mice were prepared using a Rapid Total RNA isolation kit (Eppendorf-5Prime, Inc., Boulder, CO) following the manufacturer‘s specifications. RNA was treated with DNase at 37 °C for 30 min. Lung cRNA was hybridized to the murine genome MOE430A (consisting of ⬃22,000 gene entries; Affymetrix Inc.) chips using the manufacturer’s protocol. Affymetrix MicroArray Suite version 5.0 was used to scan and quantitate the gene chips using default scan settings. Normalization was performed using the Robust Multichip Average model that consists of three steps, namely background adjustment, quartile normalization, and summarization (29, 30). Detection of differ-

ential expression was performed using Welch’s approximate t test for mutant and control groups at a p value of ⱕ0.05. Additional filters for positive candidate selection included a minimum 1.5-fold change in absolute ratio and a minimum of two present calls by the Affymetrix algorithm in three samples. DNase was removed by a DNase inactivation reagent (Ambion, Austin, TX) before cDNA synthesis. DNase-treated total lung RNA (0.5 ␮g) was reverse transcribed and analyzed by real time PCR with the Smart Cycler System (Cepheid, Sunnyvale, CA). The relative concentration of each gene was normalized to the concentration of glyceraldehyde-3-phosphate dehydrogenase in each sample. Glyceraldehyde-3phosphate dehydrogenase mRNA was quantitated using the primers 5⬘-CTTCACCACCATGGAGGAGGC-3⬘ (5⬘) and 5⬘-GGCATGGACTGTGGTCATGAG-3⬘ (3⬘). The Shh primers were 5⬘-AATCTGCAACGGAAGCGAGG-3⬘ (5⬘) and 5⬘-AACAGCCAGGTGCCAATGTG-3⬘ (3⬘). The Wnt7b primers were 5⬘-TCCAGATGGGGACAGAAGAACTCAG-3⬘ and 5⬘-TCAGGGTGGGACATTTCCAGTG-3⬘ (3⬘). The relative abundance of smooth muscle mRNAs was assessed with MOE430A RNA chips by hybridizing to lung mRNAs from control and Foxa1⫺/⫺/ Foxa2⌬/⌬ mice at E14.5 as discussed previously. Statistical Analysis of ␣-Phosphohistone-stained Cell Counting—Sections of lungs from control and Foxa1⫺/⫺/Foxa2⌬/⌬ mice were immunostained for ␣-phosphohistone H3, and the number of mitotic nuclei was assessed (stained cells/␮m2) using Metamorph Imaging software (Universal Imaging Corporation). Excel software was used to carry out t test analysis. RESULTS

Deletion of Both Foxa1 and Foxa2—To assess the efficiency of Cre-mediated Foxa2 gene deletion in the Foxa1⫺/⫺/Foxa2⌬/⌬ mice, Foxa2 immunostaining was assessed in the fetal mouse lung at E12.5, E15.5, and E18.5 (Fig. 1). In control (Fig. 1, A, E, and I) and Foxa1⫺/⫺ mice (Fig. 1, B, F, and J), Foxa2 was detected in the nuclei of most pulmonary epithelial cells, including those lining the airways, and alveolar type II cells in peripheral saccules. In Foxa2⌬/⌬ and Foxa1⫺/⫺/Foxa2⌬/⌬ mice, Foxa2 was absent in subsets of respiratory epithelial cells. As shown previously (21), Foxa2 staining was absent in most epithelial cells in both the conducting airways and the peripheral saccules in Foxa2⌬/⌬ mice at E18.5 (Fig. 1K). The extent of loss of Foxa2 staining in Foxa1⫺/⫺/Foxa2⌬/⌬ mice was less than that observed in the Foxa1⫹/⫹/Foxa2⌬/⌬ mice (21), indicating that the lack of Foxa1 may decrease the efficiency of Foxa2 gene targeting directed by the SP-C-rtTA, (tetO)7Cre system. Foxa1 was detected in nuclei of respiratory epithelial cells in both the proximal and peripheral regions of the lungs of control (Fig. 2A) and Foxa2⌬/⌬ mice (Fig. 2C). As expected, Foxa1 staining was absent in lungs from Foxa1⫺/⫺ (Fig. 2B) and Foxa1⫺/⫺/Foxa2⌬/⌬ mice (Fig. 2D). In Foxa1⫺/⫺ mice, deletion of Foxa1 resulted in increased level of Foxa2 mRNA (1.5-fold, n ⫽ 3, p ⬍ 0.05) at E18.5, suggesting that Foxa2 may compensate for the loss of Foxa1 in lungs from the Foxa1⫺/⫺ mice. Similarly, Foxa1 mRNA was increased after the conditional deletion of Foxa2 in the lung (21). Deletion of Both Foxa1 and Foxa2 Inhibits Branching Morphogenesis of the Lung—At the light microscopic level, lung morphology was similar in Foxa1⫺/⫺ (Fig. 1, D, F, and J), Foxa2⌬/⌬ (Fig. 1, C, G, and K), and control mice (Fig. 1, A, E, and I) at E12.5, E15.5, and E18.5. At E12.5, branching of lung tubules was similar in mice of each genotype (Fig. 1, A–D). However, by E15.5 enlarged tubules were readily observed at sites in which both Foxa1 and Foxa2 were deleted (Fig. 1H). At E18.5, enlarged cystic bronchial tubules were noted in the lungs of Foxa1⫺/⫺/Foxa2⌬/⌬ mice. Small terminal lung saccules typical of the normal lung at E18.5 failed to form in the regions lacking both Foxa1 and Foxa2. Lung morphology was maintained in the regions of the lung parenchyma in which Foxa2 staining was present (Fig. 1, I–L). The sites and extent of lung malformation correlated well with the sites and extent of the deletion of Foxa1 and Foxa2. To assess the effect of deletion of both Foxa1 and Foxa2 on


13811

FIG. 1. Deletion of Foxa1 and Foxa2 disrupts branching morphogenesis. Lung sections of control (A, E, and I), Foxa1⫺/⫺ (B, F, and J), Foxa2⌬/⌬ (C, G, and K), and Foxa1⫺/⫺/Foxa2⌬/⌬ (D, H, and L) mice were prepared on E12.5, E15.5, and E18.5 and stained with Foxa2 polyclonal antibody. Whereas lung architecture was not altered in Foxa1⫺/⫺ and Foxa2⌬/⌬ mice at these different developmental stages, the combined deletion of Foxa1 and Foxa2 disrupted branching after E12.5, resulting in the formation of large lung cysts by E15.5 and thereafter. Scale bar, 300 ␮m.

FIG. 2. Deletion of Foxa1 in the fetal lung. Lung sections of E18.5 mice were prepared and stained with polyclonal Foxa1 antibody. Foxa1 was absent in all the lung cells of Foxa1 ⫺/⫺ (B) and Foxa1⫺/⫺/Foxa2⌬/⌬ (D) mice. In control (A) and Foxa2⌬/⌬ (C) mice, Foxa1 was present in airway epithelial cells and in subsets of alveolar type II cells. Expression of Foxa1 was most prominent in proximal airways. Scale bar, 100 ␮m.

branching morphogenesis, lungs from control and Foxa1⫺/⫺/ Foxa2⌬/⌬ mice were excised and imaged at E12.5, E13.5, E14.5, and E18.5. Differences in the extent of branching were observed in Foxa1⫺/⫺/Foxa2⌬/⌬ mice as early as E12.5 and, thereafter, less branching was detected in the lungs of Foxa1⫺/⫺/ Foxa2⌬/⌬ mice as compared with controls (Fig. 3, A–H). Lobulation was not disturbed in the Foxa1⫺/⫺/Foxa2⌬/⌬ mice. Abnormal dilated peripheral tubules were observed as early as E13.5 (Fig. 3B). Lungs of Foxa1⫺/⫺/Foxa2⌬/⌬ mice were visibly smaller than those from Foxa1⫺/⫺, Foxa2⌬/⌬ (Fig. 3, B, D, F, and H), and controls (Fig. 3, A, C, E, and F). Deletion of Foxa1 and Foxa2 Disrupts Epithelial Cell Differentiation in the Lung—At E15.5 and E18.5 the dilated lung tubules and peripheral lung saccules seen in the lungs of Foxa1⫺/⫺/Foxa2⌬/⌬ mice were lined primarily by immature cuboidal epithelial cells. At the ultrastructural level, epithelial

cells lining the abnormal lung tubules of Foxa1⫺/⫺/Foxa2⌬/⌬ mice at E18.5 were glycogen rich but lacked the microvilli, cilia, and granules typical of various fully differentiated cell types seen in the control mice (Fig. 4). In control mice, squamous type I and cuboidal type II cells (the latter with lamellar bodies) were observed throughout the peripheral regions of the lungs. Ciliated and non-ciliated columnar epithelial cells were observed in the conducting airways of control litter mates. At E15.5, the respiratory epithelium of both control and Foxa1⫺/⫺/ Foxa2⌬/⌬ mice consisted of undifferentiated cuboidal and columnar epithelial cells that lacked secretory granules, cilia, or lamellar bodies (data not shown). Type I and differentiated type II cells were not observed in the abnormal tubules seen in Foxa1⫺/⫺/Foxa2⌬/⌬ mice at E18.5. Cuboidal cells lining the abnormal lung tubules and many of the terminal saccules had the ultrastructural features typical of respiratory epithelial

13812


FIG. 3. Abnormal branching morphogenesis after the deletion of Foxa1 and Foxa2. Fetal lungs were imaged at E12.5 (A and B), E13.5 (C and D), E14.5 (E and F), and E18.5 (G and H). Abnormal branching was observed in Foxa1⫺/⫺/Foxa2⌬/⌬ as early as E12.5 (B) compared with controls (A). At E13.5 and E14.5 dilated tubules were observed in lungs from Foxa1⫺/⫺/Foxa2⌬/⌬ mice (D and F). At E18.5 cystic malformations were observed in lungs from Foxa1⫺/⫺/ Foxa2⌬/⌬ mice (H). Scale bar, 1 mm.

cells seen in normal mice at E15.5 (Fig. 4). The pulmonary vascular bed was intact, and there was no evidence of hemorrhage. PECAM staining demonstrated that the lungs of Foxa1⫺/⫺/Foxa2⌬/⌬ mice were well vascularized (see supplemental Fig. 1 in the on-line version of this article). To assess whether deletion of Foxa1 and Foxa2 altered expression of differentiation markers, serial sections were immunostained for proSP-C, SP-B, and TTF1 at E18.5 and correlated with sites of Foxa2 deletion. Although TTF1 staining was not altered following deletion of both Foxa1 and Foxa2 (Fig. 5, E and F), SP-B and pro-SPC staining was absent. Cells lacking SP-B and proSP-C were correlated with the absence of Foxa2 staining in the Foxa1⫺/⫺/Foxa2⌬/⌬ mice (Fig. 5, C and D), supporting a cell autonomous role for the Foxa proteins in epithelial cell differentiation. Likewise, staining for CCSP and Foxj1 was absent in conducting airway epithelial cells lacking both Foxa1 and Foxa2 (Fig. 6). Thus, Foxa1 and Foxa2 are required for the expression of epithelial cell differentiation markers typical of normal peripheral and conducting airways.

Deletion of Foxa1 and Foxa2 Affects Cell Proliferation— Lungs of Foxa1⫺/⫺/Foxa2⌬/⌬ mice were visibly smaller than those of control mice. To discern whether the deletion of Foxa1 and Foxa2 altered cell proliferation, serial sections of lung tissue were stained for ␣-phosphohistone H3 and Foxa2 (see supplemental Fig. 2 in the on-line version of this article) at E15.5. The number of cells stained for ␣-phosphohistone H3 was significantly decreased in the lungs of Foxa1⫺/⫺/Foxa2⌬/⌬ mice as compared with controls (p ⬍ 0.05). In serial sections, ␣-phosphohistone H3 staining was observed in epithelial cells that lacked both Foxa1 and Foxa2 staining, indicating their mitotic activity. In addition, there was no evidence of cell death, nuclear fragmentation, or abnormal mitotic activity in the respiratory epithelium after the deletion of Foxa1 and Foxa2 as assessed by light and electron microscopy at E15.5 and E18.5 (Fig. 4 and data not shown). Foxa Activity Is Required for Expression of Shh—Shh is expressed by epithelial cells of the developing lung, providing a paracrine signal that is required for specification of smooth


13813

FIG. 4. Lung ultrastructure in Foxa1ⴚ/ⴚ/Foxa2⌬/⌬ mice. Electron microscopy was performed on lungs from Foxa1⫺/⫺/Foxa2⌬/⌬ mice (B) and control litter mates (A) observed at E18.5. Squamous type I cells (thick arrow) and cuboidal type II cells (arrowhead) containing lamellar bodies, apical microvilli, and highly organized rosette glycogen were observed in lungs of control mice (A). The densely stained cell below the type I cell is a capillary endothelial cell creating the gas exchange area (A). Type I cells were not observed in lung of Foxa2⌬/⌬ mice (B). The pulmonary vascular bed in Foxa1⫺/⫺/Foxa2⌬/⌬ mice was intact, but in the absence of type I cells the thin walled alveolar-capillary interface seen in the control mice was not formed. Type II cells were immature. Lamellar bodies were absent and apical microvilli were sparse in Foxa1⫺/⫺/Foxa2⌬/⌬ mice (B).

muscle in the lung mesenchyme and branching morphogenesis (31, 32). To determine whether the deletion of Foxa1 and Foxa2 altered Shh signaling, in situ hybridization was used on lung sections prepared at E15.5. Shh mRNA was decreased or absent in the dilated tubules of the Foxa1⫺/⫺/ Foxa2⌬/⌬ mice (Fig. 7B) compared with that in litter mate controls (Fig. 7A), indicating that the normal pattern of Shh expression was dependent on the presence of Foxa family members. Likewise, the content of Shh mRNA was significantly decreased at E14.5 (Fig. 7C). ␣-SMA, a smooth muscle cell marker and known Shh target, is normally expressed in the smooth muscle cells around the bronchi and bronchioles of the developing lung. Consistent with previous data showing that ␣-SMA is absent in the peripheral lung mesenchyme in Shh⫺/⫺ mice (28, 32), ␣-SMA immunostaining of lungs from Foxa1⫺/⫺/Foxa2⌬/⌬ mice demonstrated that ␣-SMA was decreased or absent around bronchioles lacking Foxa1 and Foxa2. Thus, the combined deletion of Foxa1 and Foxa2 inhibited mesenchymal smooth muscle cell differentiation, consistent with the decreased expression of Shh (Fig. 7, E–G). Consistent with the observed decreased Shh expression, the hedgehog-interacting protein (Hhip), myocardin (Myocd), and a number of mRNAs related to smooth muscle cell differentiation were decreased in the Foxa1⫺/⫺/Foxa2⌬/⌬ mice at E14.5, including thrombospodin 3 (Thbs3), calponin 1 (Cnn1), myosin heavy polypeptide 11 (Myh11), myosin light polypeptide 9 (Myl9), smooth muscle enteric actin ␥2 (Actg2), and the actin-associated protein transgelin (SM22␣ or Tagln) (Table I). Foxa1 and Foxa2 immunostaining was assessed in lungs from Shh⫺/⫺ mice. Nuclear staining for Foxa1 and Foxa2 was observed in the epithelial cells of respiratory tubules in Shh⫺/⫺ mice at E18.5, demonstrating that Shh is not required for the expression of Foxa proteins in the lung (data not shown). Wnt7b mRNA was significantly decreased (Fig. 7D), which is perhaps consistent with previous studies demonstrating that Foxa2 regulated Wnt7b gene transcription in vitro (33). DISCUSSION

The formation of the fetal lung occurs by a process of branching morphogenesis and depends upon complex paracrine signaling between the epithelium and the mesenchyme that in

FIG. 5. Cell autonomous effects of Foxa1 and Foxa2 deletion on gene expression in respiratory epithelial cells. Serial sections of lung from E18.5 embryos from Foxa1⫺/⫺/Foxa2⌬/⌬ mice were stained for Foxa2 (A, C, and E), mature SP-B (B), proSP-C (D), and TTF1 (F). Expression of TTF1 was not affected by the deletion of Foxa1 and Foxa2. ProSP-C and SP-B were detected in cells expressing Foxa2 (arrow). Cells lacking both Foxa1 and Foxa2 did not express either SP-B or ProSP-C (arrowhead). ProSP-C and SP-B were correlated with Foxa2 staining. Scale bar, 100 ␮m.

turn controls transcriptional programs regulating cell behavior during development. In the present study, branching and epithelial cell differentiation were perturbed after the deletion of Foxa1 and Foxa2 in lung epithelial cells, indicating that Foxa transcription factors are required for lung morphogenesis and function. Deletion of Foxa1 and Foxa2 activity inhibited the expression of Shh and Wnt7b, providing potential mechanisms by which the lack of Foxa1 and Foxa2 activity influences branching morphogenesis and differentiation of the lung. Foxa1 and Foxa2 Play Redundant Roles During Branching Morphogenesis of the Fetal Lung—Expression of Foxa1 and Foxa2 is first detected at the onset of lung bud formation. Thereafter, Foxa1 and Foxa2 are expressed in a similar temporal-spatial pattern in the respiratory epithelial cells of the conducting and peripheral airways (13). Branching morphogenesis was not perturbed in single knock-out Foxa1⫺/⫺ or Foxa2⌬/⌬ mice. In the Foxa1⫺/⫺/Foxa2⌬/⌬ mice, formation of small peripheral acinar tubules and terminal saccules was markedly disrupted in regions where deletion of both Foxa1 and Foxa2 had occurred, supporting the hypothesis that the transcription factors Foxa1 and Foxa2 act in a complementary manner during branching morphogenesis. The timing and extent of recombination with the SP-C-rtTA/ tetO7Cre system were variable and dependent upon the varying efficiency of targeted recombination of each floxed gene (21, 34). Gene targeting occurs in epithelial cells throughout the pulmonary parenchyma and occurs during early lung morphogenesis in the SP-C-rtTA mice (34). The extent of loss of Foxa2 staining in Foxa1⫺/⫺/Foxa2⌬/⌬ mice was less than in the Foxa2⌬/⌬ mice in the Foxa1⫹/⫹ background, indicating that the lack of Foxa1 may decrease the efficiency of targeting directed by the SP-C-rtTA, (tetO)7Cre system. In this mouse model,

13814


abnormalities in branching and cell differentiation correlated with the extent of deletion of both Foxa1 and Foxa2. Branching morphogenesis of the lung requires precise signaling from both epithelial and mesenchymal compartments to regulate cell proliferation, migration, and subsequent lungspecific gene expression. Many genes and associated signaling pathways influence the process of branching morphogenesis,

FIG. 6. Deletion of Foxa1 and Foxa2 alters gene expression in conducting airway epithelial cells. Serial sections were prepared from Foxa1⫺/⫺/Foxa2⌬/⌬ mice at E18.5 and stained for Foxa2 (A and C) and the airway epithelial cell differentiation markers Foxj1, a basalciliated cell marker (B), and CCSP, a non-ciliated bronchial cell marker (D). Foxj1 and CCSP were detected in cells expressing Foxa2 (arrow). Cells lacking both Foxa1 and Foxa2 did not express either Foxj1 or CCSP (arrowhead). Scale bar, 100 ␮m.

FIG. 7. Decreased Shh, ␣-SMA, and Wnt7b. Lung sections from Foxa1⫺/⫺/ Foxa2⌬/⌬ and control litter mates were prepared on E15.5 and used for in situ hybridization with an Shh riboprobe. In controls (A), Shh mRNA was detected in the epithelial cells lining respiratory tubules. In Foxa1⫺/⫺/Foxa2⌬/⌬ mice (B), Shh mRNA was absent or decreased in the epithelial cells lining the dilated respiratory tubules (arrow). Scale bar (white), 500 ␮m. Total lung RNA was prepared from fetal mouse lung at E14.5. Quantitative RT-PCR for Shh and Wnt7b mRNAs were performed. The relative concentration of each mRNA was normalized to the concentration of glyceraldehyde-3phosphate dehydrogenase mRNA in each sample. The mean of the mRNA content in control litter mates was set to 1. Relative values of Shh and Wnt7b mRNAs in Foxa1⫺/⫺/Foxa2⌬/⌬ and control mice are shown as mean ⫾ S.D. (C and D). Statistics were done by a two-tailed t test where n ⫽ 4 –5 animals per group; *, p ⬍ 0.05 versus control. Lung sections from Foxa1⫺/⫺/Foxa2⌬/⌬ and control litter mates were prepared on E15.5 and immunostained for ␣-SMA. In controls (E), ␣-SMA was detected in the smooth muscle cells around the bronchioles (arrowhead). In Foxa1⫺/⫺/Foxa2⌬/⌬ mice (F), ␣-SMA was decreased or absent around the bronchioles (arrow). Serial sections were used for Foxa2 immunostaining to demonstrate the sites of deletion of Foxa1 and Foxa2. Scale bar (black), 100 ␮m.

including ␤-catenin, Rar/Rxr, Tgf␤, Shh, Fgf10, and Bmp4 (1, 3, 35–38). Although the mechanisms by which Foxa1 and Foxa2 regulate branching morphogenesis are not understood completely, Foxa1 and Foxa2 have been implicated in several of these pathways. Foxa2 regulates Shh during early embryogenesis and central nervous system formation (39). Foxa2 positively regulated Shh transcription within the central nervous system and the axial mesoderm. Foxa1, Foxa2, and Shh are co-expressed in subsets of epithelial cell during early lung morphogenesis (39 – 43). In the present study, decreased or absent Shh mRNA was observed after the deletion of Foxa1 and Foxa2. Similarities in the disruption of branching morphogenesis seen in the lungs of Shh⫺/⫺ and Foxa1⫺/⫺/Foxa2⌬/⌬ mice support the concept that the loss of Foxa activity influences lung branching morphogenesis and mesenchymal cell differentiation, at least in part, by inhibiting Shh expression. The absence of smooth muscle differentiation as indicated by ␣-SMA staining at the site of the Foxa1 and Foxa2 deletion is consistent with decreased Shh that is required for the normal expression of myocardin and its downstream targets, including thrombospondin 3, myosin heavy polypeptide 11, myosin light polypeptide 9, calponin 1, smooth muscle enteric actin ␥2, transgelin, and ␣-SMA (28). Whereas similar branching abnormalities are observed in Shh-deficient and Foxa1⫺/⫺/Foxa2⌬/⌬ mice, changes in cell differentiation in the two models are distinct. The differentiation markers Foxj1, CCSP, SP-C, and SP-B were expressed in Shh⌬/⌬ but not in Foxa1⫺/⫺/Foxa2⌬/⌬ mice, suggesting that the phenotypic changes seen in the latter model are not consistent with lack of Shh alone. Wnt7b, a signaling molecule expressed in epithelial cells of the lung, plays a role in lung morphogenesis (44). Foxa2 trans-activates the Wnt7b promoter in NIH-3T3 cells (39). Deletion of Wnt7B


13815

TABLE I Genes related to smooth muscle cell differentiation which were decreased in Foxa1⫺/⫺/Foxa2⌬/⌬ mice at E14.5 Lung RNAs from Foxa1⫺/⫺/Foxa2⌬⌬/⌬ and age-matched control mice were prepared at E14.5. Lung cRNA was hybridized to the murine genome MOE430A chips and analyzed as described previously. Genes in the Shh pathway and associated with smooth muscle cell differentiation were significantly decreased in the Foxa1⫺/⫺/Foxa2⌬⌬/⌬ mice. The t test p value

GenBankTM

Gene name

Biological process

0.024 0.0089 0.0103 0.000236 0.0176 0.00241 0.00673 0.00988 0.012 0.0043 0.0062 0.0169 0.0271

NM_009170 NM_020259 NM_009608 NM_146386 NM_009416 NM_173756 NM_009610 NM_009922 NM_013607 XM_283793 NM_011526 NM_021467 NM_013691

Sonic hedgehog Hedgehog-interacting protein Actin, ␣, cardiac SRF co-factor protein (cardiac and smooth muscle) Tropomyosin 2, ␤ Hypothetical protein 5830457H20 Actin, ␥2, smooth muscle, enteric Calponin 1 Myosin heavy chain 11, smooth muscle Myosin, light polypeptide 9, regulatory Transgelin Troponin I, skeletal, slow 1 Thrombospondin 3

Signal transduction/pattern specification Signal transduction /organogenesis Regulation of myoblast differentiation Regulation of myoblast differentiation Regulation of myoblast differentiation Muscle development Muscle development Muscle development Muscle development Muscle development Muscle development Muscle development Cell adhesion

caused lung hypoplasia associated with defects in smooth muscle in the major pulmonary vessels (44). These findings are consistent with the present observations in Foxa1⫺/⫺/Foxa2⌬/⌬ mice, indicating that effects of their deletion may be, in part, mediated by altered Wnt7b signaling. Consistent with these in vitro studies, Wnt7b mRNA was significantly decreased in the lung of Foxa1⫺/⫺/Foxa2⌬/⌬ mice. Thus, Foxa1 and Foxa2 are required for the precise expression of Wnt7b that, in turn, may influence ␤-catenin-dependent signaling to influence lung morphogenesis or cell differentiation (33). Mechanisms by which Foxa1 and Foxa2 influence epithelial cell differentiation during lung morphogenesis are likely complex and remain to be identified. Pulmonary Epithelial Cell Differentiation Depends Upon Foxa Transcription Factors—Foxa transcription factors were shown previously to regulate cell differentiation in many organs, including the lung, central nervous system, pancreas, and liver (20, 21, 45– 48). Targeted disruption of Foxa2 in the mouse lung inhibited maturation of type II cells and caused goblet cell hyperplasia in the conducting airways, suggesting an important role of Foxa2 in the regulation of epithelial cell differentiation and function in both conducting airways and alveoli. Expression of epithelial cell markers, including SP-A, SP-B, and CCSP, was decreased, but their expression was not entirely dependent on Foxa2 (21). In contrast, immunostaining for the epithelial cell differentiation markers, SP-B, SP-C, CCSP, and Foxj1 was absent in the relatively immature columnar and cuboidal cells lining the enlarged tubules in the Foxa1⫺/⫺/Foxa2⌬/⌬ mice, indicating that their expression required Foxa activity. Overlapping and Distinct Roles of Foxa1 and Foxa2 in Regulating Pulmonary Gene Expression—Because Foxa1 and Foxa2 may bind to the same DNA motif (24 –26), Foxa1 and Foxa2 may either compete or serve complementary roles at specific transcriptional targets. Ectopic expression of Foxa1 in Foxa2⫺/⫺ embryonic stem cells partially restored the Foxa2 target gene expression (45). Foxa1 and Foxa2 were shown to transcriptionally activate expression of many of the same genes, including Titf1 (49), Cftr (50), Sftpb (51), and Scgblal (21, 52, 53). A number of genes that are expressed in the lung are regulated by both Foxa1 and Foxa2, including Sftpb (51), Scgbla1 (52, 53), lysozyme M, and lysozyme P (54). However, the distinct phenotypes in Foxa1⫺/⫺ and Foxa2⌬/⌬ mice (21–23) and the distinct expression patterns of Foxa1 and Foxa2 in many organs (13) suggest that Foxa1 and Foxa2 may also differentially regulate gene expression during embryogenesis and lung development. Foxa1 and Foxa2 may regulate gene transcription directly or

indirectly by stabilizing nucleosome positioning (55, 56). Foxa also modulates gene expression by interacting with other transcription factors, including SP1, the CCAAT/enhancer-binding protein, and the retinoic acid receptor/retinoid X receptor (57– 59). Foxa1 and Foxa2 are co-expressed with TF1 and share transcriptional targets in the respiratory epithelium (54, 60). It is therefore likely that other transcription factors interact and cooperate with Foxa1 and Foxa2 in a complex transcriptional program that regulates gene expression in the developing lung. Foxa1 mRNA was induced after the deletion of Foxa2 in respiratory epithelial cells. Likewise, Foxa2 mRNA was increased in lungs of Foxa1⫺/⫺ mice, supporting the concept that these Foxa family members play complementary roles in the regulation of branching morphogenesis and the expression of genes required in lung function after birth. Foxa proteins are utilized recurrently and at discrete times during foregut differentiation. Foxa expression is required for A-P axis formation in the early embryo, branching morphogenesis, epithelial cell differentiation prior to birth, and epithelial cell function in the postnatal lung. Acknowledgments—We thank Anne Perl, Dave Loudy, Jean Clark, and Ann Maher for valuable contributions. REFERENCES 1. Mendelsohn, C., Lohnes, D., Decimo, D., Lufkin, T., LeMeur, M., Chambon, P., and Mark, M. (1994) Development 120, 2749 –2771 2. Cardoso, W. V. (1995) Am. J. Physiol. 269, L429 –L442 3. Perl, A. K., and Whitsett, J. A. (1999) Clin. Genet. 56, 14 –27 4. Cardoso, W. V. (2001) Annu. Rev. Physiol. 63, 471– 494 5. Costa, R. H., Kalinichenko, V. V., and Lim, L. (2001) Am. J. Physiol. 280, L823–L838 6. Shu, W., Yang, H., Zhang, L., Lu, M. M., and Morrisey, E. E. (2001) J. Biol. Chem. 276, 27488 –27497 7. Pani, L., Quian, X. B., Clevidence, D., and Costa, R. H. (1992) Mol. Cell. Biol. 12, 552–562 8. Qian, X., and Costa, R. H. (1995) Nucleic Acids Res. 23, 1184 –1191 9. Hromas, R., and Costa, R. (1995) Crit. Rev. Oncol. Hematol. 20, 129 –140 10. Monaghan, A. P., Kaestner, K. H., Grau, E., and Schutz, G. (1993) Development 119, 567–578 11. Sasaki, H., and Hogan, B. L. (1993) Development 118, 47–59 12. Ang, S. L., Wierda, A., Wong, D., Stevens, K. A., Cascio, S., Rossant, J., and Zaret, K. S. (1993) Development 119, 1301–1315 13. Besnard, V., Wert, S. E., Hull, W. M., and Whitsett, J. A. (2004) Gene Expr. Patterns 5, 193–208 14. Ruiz i Altaba, A., Cox, C., Jessell, T. M., and Klar, A. (1993) Proc. Natl. Acad. Sci. U. S. A. 90, 8268 – 8272 15. Ruiz i Altaba, A., Prezioso, V. R., Darnell, J. E., and Jessell, T. M. (1993) Mech. Dev. 44, 91–108 16. Zhou, L., Lim, L., Costa, R. H., and Whitsett, J. A. (1996) J. Histochem. Cytochem. 44, 1183–1193 17. Stahlman, M. T., Gray, M. E., and Whitsett, J. A. (1998) J. Histochem. Cytochem. 46, 955–962 18. Ang, S. L., and Rossant, J. (1994) Cell 78, 561–574 19. Weinstein, D. C., Ruiz i Altaba, A., Chen, W. S., Hoodless, P., Prezioso, V. R., Jessell, T. M., and Darnell, J. E., Jr. (1994) Cell 78, 575–588 20. Zhou, L., Dey, C. R., Wert, S. E., Yan, C., Costa, R. H., and Whitsett, J. A.

13816


(1997) Dev. Dyn. 210, 305–314 21. Wan, H., Kaestner, K. H., Ang, S. L., Ikegami, M., Finkelman, F. D., Stahlman, M. T., Fulkerson, P. C., Rothenberg, M. E., and Whitsett, J. A. (2004) Development 131, 953–964 22. Kaestner, K. H., Katz, J., Liu, Y., Drucker, D. J., and Schutz, G. (1999) Genes Dev. 13, 495–504 23. Shih, D. Q., Navas, M. A., Kuwajima, S., Duncan, S. A., and Stoffel, M. (1999) Proc. Natl. Acad. Sci. U. S. A. 96, 10152–10157 24. Lai, E., Prezioso, V. R., Smith, E., Litvin, O., Costa, R. H., and Darnell, J. E., Jr. (1990) Genes Dev. 4, 1427–1436 25. Lai, E., Prezioso, V. R., Tao, W. F., Chen, W. S., and Darnell, J. E., Jr. (1991) Genes Dev. 5, 416 – 427 26. Clark, K. L., Halay, E. D., Lai, E., and Burley, S. K. (1993) Nature 364, 412– 420 27. Wert, S. E., Yoshida, M., LeVine, A. M., Ikegami, M., Jones, T., Ross, G. F., Fisher, J. H., Korfhagen, T. R., and Whitsett, J. A. (2000) Proc. Natl. Acad. Sci. U. S. A. 97, 5972–5977 28. Miller, L. A., Wert, S. E., Clark, J. C., Xu, Y., Perl, A. K., and Whitsett, J. A. (2004) Dev. Dyn. 231, 57–71 29. Irizarry, R. A., Hobbs, B., Collin, F., Beazer-Barclay, Y. D., Antonellis, K. J., Scherf, U., and Speed, T. P. (2003) Biostatistics 4, 249 –264 30. Irizarry, R. A., Bolstad, B. M., Collin, F., Cope, L. M., Hobbs, B., and Speed, T. P. (2003) Nucleic Acids Res. http://nar.oupjournals.org/cgi/ content/full/31/4/e15 31. Pepicelli, C. V., Lewis, P. M., and McMahon, A. P. (1998) Curr. Biol. 8, 1083–1086 32. Weaver, M., Batts, L., and Hogan, B. L. (2003) Dev. Biol. 258, 169 –184 33. Weidenfeld, J., Shu, W., Zhang, L., Millar, S. E., and Morrisey, E. E. (2002) J. Biol. Chem. 277, 21061–21070 34. Perl, A. K., Wert, S. E., Nagy, A., Lobe, C. G., and Whitsett, J. A. (2002) Proc. Natl. Acad. Sci. U. S. A. 99, 10482–10487 35. Hogan, B. L., Grindley, J., Bellusci, S., Dunn, N. R., Emoto, H., and Itoh, N. (1997) Cold Spring Harbor Symp. Quant. Biol. 62, 249 –256 36. Sekine, K., Ohuchi, H., Fujiwara, M., Yamasaki, M., Yoshizawa, T., Sato, T., Yagishita, N., Matsui, D., Koga, Y., Itoh, N., and Kato, S. (1999) Nat. Genet. 21, 138 –141 37. Liu, J., Tseu, I., Wang, J., Tanswell, K., and Post, M. (2000) Dev Dyn. 217, 343–360 38. Weaver, M., Dunn, N. R., and Hogan, B. L. (2000) Development 127, 2695–2704 39. Mucenski, M. L., Wert, S. E., Nation, J. M., Loudy, D. E., Huelsken, J., Birchmeier, W., Morrisey, E. E., and Whitsett, J. A. (2003) J. Biol. Chem. 278, 40231– 40238

40. Epstein, D. J., McMahon, A. P., and Joyner, A. L. (1999) Development 126, 281–292 41. Echelard, Y., Epstein, D. J., St-Jacques, B., Shen, L., Mohler, J., McMahon, J. A., and McMahon, A. P. (1993) Cell 75, 1417–1430 42. Sasaki, H., and Hogan, B. L. (1994) Cell 76, 103–115 43. Ruiz i Altaba, A., Jessell, T. M., and Roelink, H. (1995) Mol. Cell. Neurosci. 6, 106 –121 44. Shu, W., Jiang, Y. Q., Lu, M. M., and Morrisey, E. E. (2002) Development 129, 4831– 4842 45. Duncan, S. A., Navas, M. A., Dufort, D., Rossant, J., and Stoffel, M. (1998) Science 281, 692– 695 46. Sund, N. J., Ang, S. L., Sackett, S. D., Shen, W., Daigle, N., Magnuson, M. A., and Kaestner, K. H. (2000) Mol. Cell. Biol. 20, 5175–5183 47. Lin, L., Miller, C. T., Contreras, J. I., Prescott, M. S., Dagenais, S. L., Wu, R., Yee, J., Orringer, M. B., Misek, D. E., Hanash, S. M., Glover, T. W., and Beer, D. G. (2002) Cancer Res. 62, 5273–5279 48. Halmos, B., Basseres, D. S., Monti, S., D’Alo, F., Dayaram, T., Ferenczi, K., Wouters, B. J., Huettner, C. S., Golub, T. R., and Tenen, D. G. (2004) Cancer Res. 64, 4137– 4147 49. Ikeda, K., Shaw-White, J. R., Wert, S. E., and Whitsett, J. A. (1996) Mol. Cell. Biol. 16, 3626 –3636 50. Levinson-Dushnik, M., and Benvenisty, N. (1997) Mol. Cell. Biol. 17, 3817–3822 51. Bohinski, R. J., Di Lauro, R., and Whitsett, J. A. (1994) Mol. Cell. Biol. 14, 5671–5681 52. Bingle, C. D., and Gitlin, J. D. (1993) Biochem. J. 295, 227–232 53. Bingle, C. D., Hackett, B. P. Moxley, M., Longmore, W., and Gitlin, J. D. (1995) Biochem. J. 308, 197–202 54. Wan, H., Xu, Y., Ikegami, M., Stahlman, M. T., Kaestner, K. H., Ang, S. L., and Whitsett, J. A. (2004) Proc. Natl. Acad. Sci. U. S. A. 101, 14449 –14454 55. Cirillo, L. A., McPherson, C. E., Bossard, P., Stevens, K., Cherian, S., Shim, E. Y., Clark, K. L., Burley, S. K., and Zaret, K. S. (1998) EMBO J. 17, 244 –254 56. Cirillo, L. A., Lin, F. R., Cuesta, I., Friedman, D., Jarnik, M., and Zaret, K. S. (2002) Mol. Cell. 9, 279 –289 57. Braun, H., and Suske, G. (1998) J. Biol. Chem. 273, 9821–9828 58. Christoffels, V. M., Grange, T., Kaestner, K. H., Cole, T. J., Darlington, G. J., Croniger, C. M., and Lamers, W. H. (1998) Mol. Cell. Biol. 18, 6305– 6315 59. Wang, J. C., Stafford, J. M., Scott, D. K., Sutherland, C., and Granner, D. K. (2000) J. Biol. Chem. 275, 14717–14721 60. deFelice, M., Silberschmidt, D., DiLauro, R., Xu, Y., Wert, S. E., Weaver, T. E., Bachurski, C. J., Clark, J. C., and Whitsett, J. A. (2003) J. Biol. Chem. 278, 35574 –35583