• lung;
  • fibroblast growth factor;
  • gene transcription;
  • Ets;
  • Elf5


  1. Top of page
  2. Abstract
  7. Acknowledgements

Fibroblast growth factor (FGF) signaling has been shown to be essential for many aspects of normal lung development. To determine epithelial targets of FGF signaling, we cultured embryonic day (E) 11.5 mouse lungs for 24 hr in the presence or absence of the FGF receptor antagonist SU5402, which inhibited branching morphogenesis. Affymetrix gene chip analysis of treated and control epithelia identified several genes regulated by FGF signaling, including Elf5, a member of the Epithelial-specific Ets family of transcription factors. SU5402 reduced Elf5 expression in mesenchyme-free cultures of E12.5 epithelium, demonstrating that the inhibition was direct. In situ hybridization revealed that Elf5 had a dynamic pattern of expression during lung development. We found that expression of Elf5 was induced by FGF7 and FGF10, ligands that primarily bind FGFR2b. To further define the pathways by which FGFs activate Elf5 expression, we cultured E11.5 lung tips in the presence of compounds to inhibit FGF receptors (SU5402), PI3-Kinase/Akt–mediated signaling (LY294002), and MAP Kinase/Erk-mediated signaling (U0126). We found that SU5402 and LY294002 significantly reduced Elf5 expression, whereas U0126 had no effect. LY294002 also reduced Elf5 expression in cultures of purified epithelium. Finally, pAkt was coexpressed with Elf5 in the proximal epithelial airways of E17.5 lungs. These results demonstrate that Elf5 is an FGF-sensitive transcription factor in the lung with a dynamic pattern of expression and that FGF regulation of Elf5 by means of FGFR2b occurs through the PI3-Kinase/Akt pathway. Developmental Dynamics 236:1175–1192, 2007. © 2007 Wiley-Liss, Inc.


  1. Top of page
  2. Abstract
  7. Acknowledgements

In the mouse, the lung buds are first evident at embryonic day 9.5 (E9.5) as two endodermal outgrowths from the foregut penetrating the adjacent splanchnic mesoderm. After this initial event, the pulmonary tree forms by a process referred to as branching morphogenesis, which is characterized by repeated outgrowth and branching of the epithelium. This process is highly governed by epithelial–mesenchymal interactions by means of intercellular secreted paracrine factors and propagated through their receptors (Shannon and Hyatt, 2004; Warburton and Bellusci, 2004; Cardoso and Lu, 2006). One group of secreted factors that has been intensely studied in embryonic lung development is the fibroblast growth factor (FGF) family. FGF1 (Cardoso et al., 1997), FGF2 (Lebeche et al., 1999), FGF7 (Tichelaar et al., 2000), FGF9 (Colvin et al., 2001), FGF10 (Sekine et al., 1999), and FGF18 (Usui et al., 2004) have all been implicated in lung development. The FGF family of secreted ligands induces an array of cellular responses through alternately spliced isoforms of FGF receptors (FGFR1, FGFR2, FGFR3, and FGFR4) on cell surfaces, which activate signaling pathways depending on cell type and on interactions with other signaling cascades. FGFR1, FGFR2, and FGFR3 bind FGFs as either IIIb or IIIc splice variants in the third Ig-like domain, giving each variant its own specificity for FGF ligand binding and subsequent activation (Dailey et al., 2005). FGFR4, however, does not have splice variants of this third Ig-like domain (Eswarakumar et al., 2005). Formation of the FGF–FGFR complex stimulates tyrosine autophosphorylation, which promotes signaling through ERK, PI3-Kinase, and PLCγ pathways (Dailey et al., 2005; Eswarakumar et al., 2005).

All four FGFRs are detectable in the lung (Powell et al., 1998). Although mice with a deletion of either FGFR3 or FGFR4 have normal lung development, mice with null mutations for both FGFR3 and FGFR4 have postnatal lung alveoligenesis defects (Weinstein et al., 1998). In the lung, the epithelium-specific isoform IIIb of FGFR2 (hereafter FGFR2b) has been the most intensely studied in branching morphogenesis. The importance of FGFR2b to lung morphogenesis is demonstrated by the fact that the deletion of FGFR2b causes an agenesis of the lung (De Moerlooze et al., 2000). FGF10 signals through FGFR2b during the morphogenesis of many organs, including the lung (Ohuchi et al., 2000). FGF10 expression in E9.5 splanchnic mesoderm is required for the outgrowth of the first epithelial lung buds from the foregut endoderm, because FGF10 null mice also exhibit complete agenesis of the lung (Min et al., 1998; Sekine et al., 1999; Ohuchi et al., 2000). Distinct domains of high FGF10 expression govern lung branching morphogenesis by directing epithelial migration through the mesenchyme by chemotaxis (Bellusci et al., 1997; Park et al., 1998). FGF7, which also signals through FGFR2b but elicits different morphological responses than FGF10, has been shown to be important for lung epithelial proliferation and differentiation in vitro and in transgenic models (Lebeche et al., 1999; Matsui et al., 1999; Tichelaar et al., 2000; Hyatt et al., 2002). Because FGF7 and FGF10 are secreted by the mesenchyme, but activate epithelial FGFR2b to regulate gene expression, they are thought to be important mediators of epithelial–mesenchymal interactions in the developing lung.

We were interested in uncovering genes regulated by FGF signaling in the epithelium that might be important for embryonic lung development, particularly those genes affecting proliferation and/or differentiation. To do this, we used Affymetrix microarray analysis to compare the transcriptomes of distal epithelium from E11.5 lung explants treated in vitro with either vehicle or the FGF receptor antagonist SU5402. We found that the Ets family transcription factor Elf5 (E74-like Factor 5, also known as Ese2 [Epithelial-Specific Ets 2]), was one of the genes decreased when FGF signaling was inhibited. In situ hybridization (ISH) analysis showed that Elf5 is only expressed in the epithelium throughout embryonic lung development and persists postnatally. Elf5 expression was directly up-regulated in E11.5 lung explant cultures treated with exogenous FGF7 or FGF10 and in transgenic mouse lungs misexpressing FGF7 or FGF10. Elf5 was coexpressed with phospho-Akt (pAkt) in proximal airways in late gestation. Finally, Elf5 expression was reduced in explant cultures treated with the pAkt inhibitor LY294002, suggesting that FGFs regulate Elf5 by means of the PI3-Kinase/Akt pathway.


  1. Top of page
  2. Abstract
  7. Acknowledgements

SU5402 Blocks Embryonic Lung Branching in Culture

To block FGF receptor signaling in the embryonic lung, we cultured whole E11.5 lungs in the presence of FGF receptor antagonist SU5402 for 2 days (Fig. 1). The 10 μM concentration of SU5402 used in these experiments has been shown to effectively block FGF signaling while having no effect on platelet-derived growth factor–induced tyrosine kinase activity (Mohammadi et al., 1997). SU5402 was originally shown to inhibit signaling through FGFR1, but subsequent studies have shown that it also inhibits signaling through other FGF receptors (Maeda et al., 2005; Bernard-Pierrot et al., 2006; Masih-Khan et al., 2006; Urban et al., 2006). Dimethyl sulfoxide (DMSO) controls displayed typical lung explant branching morphogenesis (Fig. 1A,D,G). SU5402-treated lung explant cultures, however, showed a complete cessation of branching (Fig. 1B,E,H). To control for the possibility that the effects of SU5402 were due to nonselective cellular toxicity leading to cell death, we treated cultures with SU5402 for 1 day, then washed the explants and placed them in DMSO-containing medium for an additional 2 days (Fig. 1C,F,I,J). We found that branching resumed after FGF signaling resumed after the removal of SU5402 (Fig. 1I,J), indicating that SU5402 was not nonspecifically toxic to the explants. Furthermore, sections of lungs treated with SU5402 for 2 days showed no apparent difference from controls in the number of dead cells with pyknotic nuclei (data not shown).

thumbnail image

Figure 1. Treatment of embryonic day (E) 11.5 lung explants with fibroblast growth factor (FGF) receptor antagonist SU5402 halts lung branching morphogenesis. A,D,G: Compared with day 0 (A), DMSO-treated E11.5 lung explants branched when cultured for 1 (D) or 2 (G) days. Arrows indicate the formation of new branches. B, E, H: Compared with Day 0 (B), SU5402-treated lung explants showed no increase in branching after 1 (E) or 2 (H) days. The potential toxicity of SU5402 was tested in washout experiments. C, F: Lungs were cultured with SU5402 for 1 day, whereupon branching ceased. I,J: Branching resumed when these cultures were washed free of SU5402 and cultured with DMSO-containing medium for an additional 2 days. All photomicrographs were taken at the same original magnification.

Download figure to PowerPoint

SU5402 Treatment Decreases Expression of FGF-Responsive Genes in the Epithelium of E11.5 Lung Explants

We were most interested in identifying epithelial genes whose expression decreased in the absence of FGF signaling. To achieve this, we cultured E11.5 lungs in the presence of SU5402 or DMSO for 1 day, separated distal tip epithelium from the mesenchyme, isolated RNA from the treated and untreated epithelia, and performed Affymetrix gene chip analysis. A comparison was also made between treated and untreated mesenchyme. Given the known importance of FGFs in the development of the lung, it is not surprising that the total number of genes we found to be affected by SU5402 was large. To make our analysis more manageable, we chose to focus on those genes that decreased in response to SU5402 that were expressed exclusively in the lung epithelium. We found that SU5402 treatment down-regulated 136 epithelial genes (Table 1). These included genes that have been previously described to be FGF-sensitive, such as Sftpa, Sftpb, and Sftpc (Cardoso et al., 1997; Matsui et al., 1999; Tichelaar et al., 2000; Yano et al., 2000; Clark et al., 2001; Hyatt et al., 2004). Of the 136 genes decreased in SU5402-treated epithelium, there were five transcription factors (Hnf4a, Atoh1, Nkx3-1, Trp63, and Elf5). Because of the known importance of the Ets family of transcription factors in lung development (Liu et al., 2003; Lin et al., 2006), we chose Elf5 for further investigation.

Table 1. Epithelial Genes Down-regulated by SU5402
Fold changeGenBankCommonDescription
26.247NM_008261Hnf4aHepatic nuclear factor 4, alpha
15.337NM_010921Nkx3-1NK-3 transcription factor, locus 1
5.714NM_007500Atoh1Atonal homolog 1
4.545NM_010125Elf5E74-like factor 5
3.003NM_011641Trp63Transformation related protein 63
Regulation of surface tension
14.728NM_147779SftpbSurfactant associated protein B
14.124NM_023134Sftpa1Surfactant associated protein A1
3.413NM_011359SftpcSurfactant associated protein C
31.348NM_011402Slc34a2Solute carrier family 34, member 2
16.207NM_026228Slc39a8Solute carrier family 39, member 8
10.627NM_173388Slc43a2Solute carrier family 43, member 2
8.850NM_019810Slc5a1Solute carrier family 5, member 1
8.403NM_172658Slco4c1Solute carrier organic anion transporter family, member 4C1
7.194NM_023908Slco3a1Solute carrier organic anion transporter family, member 3a1
5.814NM_009700Aqp4Aquaporin 4
4.525NM_080440Slc8a3Solute carrier family 8, member 3
3.717NM_177341Trpm3Transient receptor potential melastatin 3
3.497NM_007469Apoc1Apolipoprotein C-I
3.257NM_033314Slco2a1Solute carrier organic anion transporter family, member 2a1
3.135NM_172469Clic6Chloride intracellular channel 6
2.994NM_008423Kcnd1Potassium voltage-gated channel, Shal-related family, member 1
2.825NM_031197Slc2a2Solute carrier family 2, member 2
2.410NM_177909Slc9a9Solute carrier family 9, member 9
ATP binding/kinase activity
33.898NM_027756Mfap3lMicrofibrillar-associated protein 3-like
24.752NM_133653Mat1aMethionine adenosyltransferase I, alpha
15.528NM_011084Pik3c2gPhosphatidylinositol 3-kinase, C2 domain containing, gamma polypeptide
8.130XM_355911Pcsk6Proprotein convertase subtilisin/kexin type 6
6.757NM_145508Dyrk3Dual-specificity tyrosine-(Y)-phosphorylation regulated kinase 3
5.650NM_021330Prkcb1Acid phosphatase 1, soluble
4.444NM_080728Myh7Myosin, heavy polypeptide 7, cardiac muscle, beta
3.571NM_007595Camk2bCalcium/calmodulin-dependent protein kinase II, beta
2.532NM_178907Mapkapk3Mitogen-activated protein kinase-activated protein kinase 3
Ion binding   
23.148NM_172613Atp13a4ATPase type 13A4
4.566NM_011246Rasgrp1RAS guanyl releasing protein 1
4.292NM_021543Pcdh8Protocadherin 8
3.534XM_181420Cgref1Cell growth regulator with EF hand domain 1
3.497NM_026993Ddah1Dimethylarginine dimethylaminohydrolase 1
2.381NM_021509Moxd1Monooxygenase, DBH-like 1
Protein binding   
5.155XM_193582Lin7aLin 7 homolog a
4.348NM_009936Col9a3Procollagen, type IX, alpha 3
3.788XM_140451Lama3Laminin, alpha 3
3.521NM_029633Clasp2CLIP associating protein 2
2.857NM_144557MyripMyosin VIIA and Rab interacting protein
2.681NM_008733NrapNebulin-related anchoring protein
2.571NM_026495Btbd14aBTB domain containing 14A
Enzymatic activity
109.290NM_007801CtshCathepsin H
15.244NM_172466Adamts18A disintegrin-like and metalloprotease with thrombospondin type 1 motif, 18
8.850NM_001001446Cyp2c44Cytochrome P450, family 2, subfamily c, polypeptide 44
6.098NM_138595GldcGlycine decarboxylase
5.882NM_013777Akr1c12Aldo-keto reductase family 1, member C13
5.814NM_027988Noxo1NADPH oxidase organizer 1
5.025NM_001004143Usp22Ubiquitin-specific protease 22
4.950NM_001003719Garnl1GTPase activating RANGAP domain-like 1
4.854NM_009658Akr1b3Aldo-keto reductase family 1, member B3
4.762NM_145451Gpx6Glutathione peroxidase 6
4.651NM_175177Bdh3-hydroxybutyrate dehydrogenase
4.566NM_013778Akr1c13Aldo-keto reductase family 1, member C13
2.817NM_011523Synj2Synaptojanin 2
2.415NM_183284Spink2Serine protease inhibitor, Kazal type 2
13.245NM_133192Gpr74G protein-coupled receptor 74
8.065NM_016982Vpreb1Pre-B lymphocyte gene 1
3.788NM_011350Sema4fSema domain, immunoglobulin domain, TM domain, and short cytoplasmic domain
3.472NM_172572Rhbdl6Rhomboid, veinlet-like 6
78.125NM_008970PthlhParathyroid hormone-like peptide
7.519NM_145435PyyPeptide YY
Secreted factor   
20.040NM_010702Lect2Leukocyte cell-derived chemotaxin 2
8.772NM_008371Il7Interleukin 7
3.846NM_011340Serpinf1Serine proteinase inhibitor, clade F, member 1
2.915NM_009528Wnt7bWingless-related MMTV integration site 7B
256.410NM_008230HdcHistidine decarboxylase
38.314NM_146101Habp2Hyaluronic acid binding protein 2
13.966NM_010014Dab1Disabled 1
13.123XM_132070Cyt1Cytokine like 1
11.025NM_178723Zfp533Zinc finger protein 533
8.696NM_133735Ptcd1Pentatricopeptide repeat domain 1
8.475NM_019910DcppDemilune cell and parotid protein
7.576NM_146010Tm4sf3Tetraspanin 8
6.849NM_030143Ddit4lDNA-damage-inducible transcript 4-like
6.410NM_153801Srd5a2l2Steroid 5 alpha-reductase 2-like 2
5.650NM_011815FybFYN binding protein
5.181NM_182930Plekha6Pleckstrin homology domain containing, family A member 6
3.817NM_001029978Tceal3Transcription elongation factor A (SII)-like 3
3.460NM_138685Wfdc15WAP four-disulfide core domain 15
3.215NM_025458Tmed6Transmembrane emp24 protein transport domain containing 6
2.941NM_145470Depdc6DEP domain containing 6

Whole-mount in situ hybridization (WM-ISH) using an antisense Elf5 probe on E11.5 lungs showed that Elf5 is expressed in the distal tip epithelium (Fig. 2A). Reverse transcriptase-polymerase chain reaction (RT-PCR) of separated E11.5 distal lung mesenchyme and epithelium confirmed that Elf5 was expressed only in the epithelium (Fig. 2B). We then cultured individual distal lung tips in medium containing DMSO or SU5402. We found that tips cultured with DMSO began branching by 24 hr (Fig. 2C,D), whereas tips cultured with SU5402 had denser mesenchyme and did not branch (Fig. 2E,F). When we measured Elf5 expression in these cultures by real-time PCR, we found that SU5402 significantly decreased Elf5 mRNA levels by 66% (Fig. 2G,H).

thumbnail image

Figure 2. Elf5 is expressed in lung epithelium and requires fibroblast growth factor (FGF) stimulation. A: Embryonic day (E) 11.5 lungs were examined for Elf5 expression by whole-mount in situ hybridization (WM-ISH). Arrows indicate epithelial Elf5 expression in the distal lung tips. B: Reverse transcriptase-polymerase chain reaction (RT-PCR) for Elf5 in purified distal epithelium (LgE) and mesenchyme (LgM) from E11.5 lungs confirmed that expression was restricted to the epithelium; Sftpc was used as a marker for lung epithelium. C–F: E11.5 lung tips (C,E) were cultured for 1 day in medium containing DMSO (D) or SU5402 (F) for 24 hr. The epithelium of SU5402-treated tips did not branch, and the mesenchyme appeared denser. G: Qualitative gel analysis by RT-PCR suggested that Elf5 expression was decreased by SU5402 treatment. H: This suggestion was confirmed by quantitative real-time PCR, which showed that Elf5 expression decreased by 66% (*P < 0.02; n = 3). All photomicrographs were taken at the same original magnification (×6) except panel A, which was taken at ×5.

Download figure to PowerPoint

Because distal tip cultures contain both epithelium and mesenchyme, and because both tissue compartments contain FGF receptors, we could not ascertain whether the effects of blocking FGF signaling on Elf5 expression were direct on the epithelium or mediated through the mesenchyme. To address these possibilities, we cultured mesenchyme-free distal tip epithelium for 3 days in a complex medium (BFGM) that we have previously shown will support type II cell differentiation of embryonic respiratory epithelium (Shannon et al., 1999), then treated the rudiments with SU5402 for 2 days. We found that, whereas DMSO-treated rudiments continued to expand in culture (Fig. 3A,C,E), blocking FGF signaling with SU5402 inhibited further expansion, and in some cases caused the rudiments to decrease in size (Fig. 3B,D,F). The majority of dilation in control cultures was likely due to FGF-induced fluid secretion as has been previously reported (Zhou et al., 1996; Graeff et al., 1999). The apparent collapse of SU5402-treated epithelium was probably due to the inhibition of fluid secretion, because epithelial expansion resumed after SU5402 was washed out and the cultures maintained for an additional 2 days in medium containing DMSO (Fig. 3G). This observation is also consistent with our microarray data showing a decrease in several transport genes. Importantly, SU5402 significantly reduced Elf5 mRNA levels by 52% in lung epithelial cultures (Fig. 3H,I), indicating that FGFs directly affect Elf5 expression. The decrease in Elf5 expression was not due to cell death, because SU5402-treated epithelia showed no increase in immunostaining for cleaved caspase 3 compared with DMSO-treated controls (data not shown).

thumbnail image

Figure 3. Elf5 expression decreased in purified epithelium cultures treated with SU5402. Embryonic day (E) 12.5 distal lung epithelia were separated from mesenchyme and cultured for 3 days in BFGM (see the Experimental Procedures section). A–F: After 3 days of culture, the medium was replaced with BFGM containing dimethyl sulfoxide (DMSO, A,C,E) or SU5402 (B,D,F) and cultures were maintained for 2 additional days. Rudiments cultured with DMSO continued to expand, whereas those treated with SU5402 stopped expanding and in some cases collapsed. G: These rudiments resumed expansion after SU5402 was washed out and replaced with DMSO-containing medium for 2 days. H: Qualitative gel analysis of reverse transcriptase-polymerase chain reaction (RT-PCR) products indicated that blocking fibroblast growth factor (FGF) signaling with SU5402 directly inhibited Elf5 expression. I: This finding was confirmed by real-time PCR, which showed that Elf5 expression was reduced by 52% (*P < 0.02; n = 3). All photomicrographs were taken at the same original magnification.

Download figure to PowerPoint

Elf5 Is Expressed Throughout Lung Development and Continues Postnatally

Using WM-ISH, we detected Elf5 in distal tip lung epithelium but not in the proximal lung or tracheal epithelium on day E11.5 (Fig. 2A). Tissue section ISH revealed that Elf5 expression was expressed in the more distal branching epithelium on day E14.5 (Fig. 4A,B), but by E15.5 (Fig. 4C,D) and E16.5 (Fig. 4E,F), it was also expressed in the proximal epithelium. On day E17.5 and E18.5, Elf5 expression appeared more intense in the proximal bronchiolar and tracheal epithelium than in the distal epithelium (Fig. 4G–J). Lungs from postnatal day 1 pups and adult mice still expressed Elf5 in the proximal epithelial airways, whereas Elf5 was very low or absent in the distal epithelium (Fig. 4K–P).

thumbnail image

Figure 4. Elf5 expression in the developing lung. Sections from mouse lungs on different days of embryonic (E) and postnatal (PN) development were hybridized with a radiolabeled cRNA probe for Elf5. A–D: During the pseudoglandular stage of lung development, Elf5 is expressed predominantly in the distal epithelium, although some proximal expression can be seen. E–J: Expression in the proximal epithelium becomes more pronounced in the canalicular stage (E–H), and this finding continues until just before birth on day E18.5 (I,J). K–N: Expression in the distal epithelium is almost completely extinguished on PN1 (K,L), a pattern that persists through adulthood (M,N). O,P: High magnification of adult lung showing high Elf5 expression in the airway epithelium. Arrows indicate proximal epithelium, while asterisks specify distal parenchyma. All photomicrographs were taken at the same original magnification (×4) except panels O and P (×20).

Download figure to PowerPoint

FGFs Up-regulate Elf5 Expression in E11.5 Explant Cultures and Transgenic Lungs

To determine which FGFs were able to induce Elf5 expression, E11.5 lung explants were cultured for 2 days with engrafted heparin beads that had been soaked in Hanks' balanced salt solution (HBSS, control) or in one of the following FGFs: FGF1, FGF2, FGF7, FGF9, FGF10, or FGF18 (Fig. 5), an approach that has been used by others to elucidate the role of FGFs in the developing lung (Park et al., 1998; Lebeche et al., 1999; Izvolsky et al., 2003; Liu et al., 2003; Lu et al., 2005; White et al., 2006). We then performed WM-ISH to detect Elf5 expression (Fig. 5C,F,I,L,O,R,U). Engrafting explant cultures with control beads did not cause a local increase Elf5 expression in the lung or induce it in tracheal epithelium; however, Elf5 expression in the distal epithelium appeared to increase with time in culture (Fig. 5A–C) when compared with uncultured time 0 E11.5 lungs (Fig. 2A). Beads soaked in FGF2, FGF7, FGF9, or FGF10 (Fig. 5G–I, J–L, M–O, P–R, respectively) all caused morphological changes. FGF7 (Fig. 5J–L) and FGF10 (Fig. 5P–R) beads caused a localized epithelial dilation and expansion, which a previous study has shown is due to increased proliferation and chemotaxis of the epithelium adjacent to the bead (Park et al., 1998). These cultures showed increased Elf5 expression in the lung epithelium closest to the bead (Fig. 5L,R). Notably, we observed Elf5 expression in tracheal epithelium adjacent to FGF7 and FGF10 beads. Because tracheal epithelium does not normally express Elf5 at this stage, these results indicate that FGF7 and FGF10 can induce Elf5 expression. FGF2 (Fig. 5G–I) and FGF9 (Fig. 5M–O) beads caused an expansion of the mesenchyme and dilation of the distal lung epithelium, but did not appear to affect the tracheal epithelium. FGF2 and FGF9 also appeared to increase Elf5 expression in the lung epithelium, but, in contrast to FGF7 and FGF10 beads, had no effect on Elf5 expression in the tracheal epithelium (Fig. 5I,O). These data suggest that the influence of FGF2 and FGF9 on the distal lung epithelium may be indirectly mediated by means of the mesenchyme.

thumbnail image

Figure 5. Fibroblast growth factor 7 (FGF7) and FGF10 directly induce Elf5. A–U: Embryonic day (E) 11.5 lung explants were cultured for 2 days with engrafted heparin beads soaked in Hanks' balanced salt solution (A–C), FGF1 (D–F), FGF2 (G–I), FGF7 (J–L), FGF9 (M–O), FGF10 (P–R), or FGF18 (S–U). After culture, WM-ISH was used to detect Elf5 expression (C,F,I,L,O,R,U). Asterisks indicate bead placements. FGF2 (G–I), FGF7 (J–L), FGF9 (M–O), and FGF10 (P–R) beads all caused epithelial dilation and expansion. FGF2 (H) and FGF9 (N) beads also induced an increase in mesenchymal mass. FGF2 (I), FGF7 (L), FGF9 (O), and FGF10 (R) beads all increased Elf5 expression in the lung epithelium. However, only FGF7 (L) and FGF10 (R) increased Elf5 in lung epithelium immediately adjacent to the bead and induced Elf5 in the tracheal epithelium (arrows). The induction of Elf5 in the tracheal epithelium indicates a direct effect on the epithelium. All photomicrographs were taken at the same original magnification.

Download figure to PowerPoint

Because FGF7 and FGF10 increased Elf5 expression in E11.5 lung explants, we investigated if this response persisted in E18.5 and adult lungs, when expression of Elf5 in the distal lung epithelium has diminished (Fig. 4I–P). A previous study demonstrated that targeted misexpression of FGF10 in the lung epithelium induced adenomatous growths within the lungs that exhibited markers of peripheral lung epithelial differentiation (Clark et al., 2001). We examined both E18.5 and adult lungs from these transgenic mice for expression of Elf5 by ISH and found that it was highly expressed within the hyperplastic epithelium, but was negative or at low levels in surrounding tissue that appeared histologically normal (Fig. 6E–H). Lungs from E18.5 mice misexpressing FGF7 in the lung epithelium exhibited widespread epithelial hyperplasia consistent with cystadenomatoid malformation, and had no apparent normal tissue (Tichelaar et al., 2000). Section ISH revealed that FGF7 induced robust Elf5 expression in virtually all of the epithelial cells in these lungs (Fig. 6C,D), whereas strong Elf5 expression was only seen in the proximal epithelium of control lungs (Fig. 6A,B).

thumbnail image

Figure 6. Elf5 is induced in the lungs of mice conditionally misexpressing fibroblast growth factor-7 (FGF7) or FGF10. ISH was performed to detect Elf5 in the lungs of embryonic day (E) 18.5 and adult mice that conditionally express FGF7 or FGF10 under control of the rat CCSP promoter, which targets both proximal and distal epithelium in transgenic mice. A,B: Relative to the peripheral lung epithelium, high Elf5 expression was seen in the proximal lung epithelium of control, single transgenic E18.5 CCSP-rtTA mice. C,D: Misexpression of FGF7 caused widespread epithelial hyperplasia, with no areas of normal lung morphology. Elf5 was uniformly induced throughout the hyperplastic epithelium. E–H: Misexpression of FGF10 caused focal adenomatous tumors that existed next to apparently normal lung parenchyma in both E18.5 (E,F) and adult (G,H) lungs. Elf5 was strongly induced in the adenomatous lesions (arrow), but remained absent in the normal peripheral epithelium (asterisks). All photomicrographs were taken at the same original magnification.

Download figure to PowerPoint

Elf5 Expression Is Up-regulated Through the PI3-Kinase Pathway

FGF receptors are known to propagate their signals three ways: the MAP Kinase/Erk pathway, PI3-Kinase/Akt pathway, and the PLCγ pathway. We determined the effects of inhibiting the MAP Kinase/Erk and PI3-Kinase/Akt pathways on Elf5 expression; the PLCγ pathway was not investigated in this study. To evaluate if these pathways regulate Elf5 expression, E11.5 lung tips were cultured in the presence of DMSO, or SU5402, or LY294002 (a PI3-Kinase inhibitor), or U0126 (a MAP ERK kinase inhibitor). Lung tips were used in these experiments because they are the site of highest Elf5 expression on day E11.5. DMSO-treated lung tips began to branch after 1 day in culture (Fig. 7B), whereas tips treated with either SU5402, or LY294002, or U0126 did not branch (Fig. 7C–E). Real-time PCR was used to assess Elf5 expression after 1 day in culture. As described above, SU5402 significantly decreased Elf5 expression to 40% of control levels. LY294002 significantly decreased Elf5 expression by 76%. In contrast, U0126 had no effect on Elf5 expression (Fig. 7F,G).

thumbnail image

Figure 7. Elf5 is downstream of the PI3-Kinase/Akt signaling pathway in cultured distal lung tips. A–E: Embryonic day (E) 11.5 lung tips (A) were cultured for 1 day in medium containing either dimethyl sulfoxide (DMSO, control, B), or the fibroblast growth factor (FGF) receptor inhibitor SU5402 (C), or the PI3-Kinase inhibitor LY294002 (D), or the MAP Kinase inhibitor U0126 (E). Whereas control tips began to branch in culture (B), branching in all other treatment groups was inhibited. F: Qualitative gel analysis of reverse transcriptase-polymerase chain reaction (RT-PCR) products indicated that Elf5 mRNA content was inhibited by SU5402 and LY294002. G: These data were confirmed by quantitative real-time PCR, which showed that SU5402 and LY294002 treatment significantly (*P < 0.001; n = 4) decreased Elf5 expression by 58% and 76%, respectively). All photomicrographs were taken at the same original magnification.

Download figure to PowerPoint

To determine whether LY294002 directly affected Elf5 expression in the epithelium and not through an indirect mechanism by means of the mesenchyme, we cultured E12.5 lung epithelium in BFGM minus FGF1 for 3 days; FGF1 was omitted from the medium based on the results in Figure 5F that FGF1 had no effect on Elf5 expression. This initial culture period allowed the lung epithelial rudiments to adapt and expand in culture. The rudiments were then treated for an additional day with either DMSO, or LY294002, or U0126 (Fig. 8A–F). No obvious change in morphology was distinguishable after 1 day of LY294002 or U0126 treatment. However, LY294002 significantly reduced Elf5 expression by 61%, whereas U0126 caused no significant change (Fig. 8G). These data suggest that Elf5 expression is maintained through FGF Receptor signaling by means of the PI3-Kinase/Akt pathway.

thumbnail image

Figure 8. PI3-Kinase/Akt signaling in the epithelium regulates Elf5 expression. A–F: Purified epithelial rudiments from embryonic day (E) 12.5 distal lung tips enrobed in Matrigel were cultured in BFGM minus fibroblast growth factor-1 (FGF1) for 3 days, then for an additional day in the presence of either DMSO (A,B), or LY294002 (C,D) or U0126 (E,F). The additions of LY294002 or U0126 had no obvious effect on rudiment morphology. G: Quantitative analysis by real-time PCR, demonstrated that LY294002 significantly (*P < 0.03; n = 3) decreased Elf5 expression by 61%, whereas U0126 treatment caused no significant change. All photomicrographs were taken at the same original magnification.

Download figure to PowerPoint

To further explore the relationship of activated pAkt and Elf5 in the lungs of later gestation embryos in vivo, we examined the expression of pAkt (Fig. 9A,C, page 10) and Elf5 (Fig. 9B,D) by immunohistochemistry and ISH, respectively, on adjacent sections of E17.5 lungs. We observed that Elf5 and pAkt colocalized in the proximal airway epithelium, further supporting a role for the PI3-Kinase/Akt pathway in maintaining Elf5 expression.

thumbnail image

Figure 9. Elf5 is coexpressed with phosphorylated Akt (pAkt) in proximal airways. A–D: Immunohistochemistry for pAkt (A,C) and ISH for Elf5 (B,D) were performed on consecutive sections of embryonic day (E) 17.5 lungs. The results show that pAkt was strongly present in the proximal airway epithelium (arrows) and that this finding correlated with the pattern of Elf5 expression. All photomicrographs were taken at the same original magnification.

Download figure to PowerPoint


  1. Top of page
  2. Abstract
  7. Acknowledgements

SU5402 Down-regulates FGF-Sensitive Genes in the Lung Epithelium

FGFs are known mediators of the epithelial–mesenchymal interactions that drive lung morphogenesis and differentiation. For example, FGF9 produced by the endodermal and mesothelial epithelium is required for normal proliferation in the lung mesenchyme; deletion of FGF9 results in reduced mesenchymal mass and lung hypoplasia (Colvin et al., 2001). FGF10 is dynamically expressed in distinct mesenchymal domains during the pseudoglandular stage (Bellusci et al., 1997) and patterns epithelial branching of the pulmonary tree by acting as a chemoattractant for the embryonic lung epithelium (Park et al., 1998). FGF10 null mice completely lack lungs because they are unable to generate primary epithelial buds (Min et al., 1998; Sekine et al., 1999). Furthermore, mice harboring a hypomorphic FGF10 allele display a decrease in embryonic lung branching, indicating that FGF10 is required for secondary branching (Mailleux et al., 2005). FGF7, which is produced by lung mesenchymal cells, is a potent mitogen for lung epithelial cells, both in vivo (Ulich et al., 1994; Simonet et al., 1995; Tichelaar et al., 2000) and in vitro (Shiratori et al., 1996; Bellusci et al., 1997; Cardoso et al., 1997; Shannon et al., 1999), as well as a strong inducer of alveolar type II cell differentiation (Chelly et al., 1999; Shannon et al., 1999). The differences in response of the lung epithelium to FGF10 and FGF7, even though both ligands signal primarily through FGFR2b, may be due to differential interactions with low-affinity proteoglycan coreceptors (Izvolsky et al., 2003; Shannon et al., 2003; Sedita et al., 2004). The differential response of epithelium to FGF10 and FGF7 is not limited to the lung. Similar results have been reported for embryonic submandibular salivary gland (Steinberg et al., 2005) and ureteric bud (Qiao et al., 2001) epithelia.

We treated E11.5 lung explants in vitro with the FGF receptor antagonist SU5402 to block FGF signaling during the pseudoglandular stage of lung development, a period when lung patterning predominates. The effectiveness of SU5402 in blocking FGF signaling was evident from our observation that branching ceased in SU5402-treated explants but continued in DMSO controls. This lack of branching was identical to that seen when a dominant-negative form of FGFR2 was expressed in the embryonic lung epithelium in vivo (Peters et al., 1994). The effects of SU5402 were not due to nonselective cellular toxicity, because branching resumed after the explants were washed free of SU5402 and switched to control medium. We, therefore, believed that this was a good model system for identifying FGF-dependent genes in the developing lung. Because we were most interested in FGF-regulated genes in the lung epithelium, microarray analysis was performed on purified distal epithelium isolated from SU5402-treated vs. control lungs. We identified a list of FGF-sensitive, epithelium-specific genes shown in Table 1. This list contained some genes previously reported to be FGF-regulated in the lung, such as Sftpa (Cardoso et al., 1997; Matsui et al., 1999), Sftpb (Cardoso et al., 1997; Matsui et al., 1999; Clark et al., 2001), and Sftpc (Hyatt et al., 2004). Tm4sf3, Myh7, and Ctsh were also identified in a previous study as epithelial genes up-regulated in response to the treatment of FGF10 (Lu et al., 2005). The presence of these known FGF-regulated genes provided additional validation that the genes on our list were truly FGF-sensitive. We were interested in finding FGF-sensitive transcription factors, because they likely serve as upstream regulators of other epithelial genes. Our analysis revealed five transcription factors (Hnf4a, Atoh1, Nkx3-1, Trp63, and Elf5) whose expression decreased in response to SU5402. Because of the demonstrated roles for Ets family transcription factors in lung development (Liu et al., 2003; Lin et al., 2006), we chose to further investigate Elf5.

The Ets family of transcription factors includes approximately 30 members that bind GGAA/T cis elements in the promoters of target genes. Ets transcription factors have been reported to be both transcriptional activators and repressors (Tootle and Rebay, 2005) and are important in many cellular processes involving development, proliferation, apoptosis, and differentiation (Oikawa and Yamada, 2003). The FGF-dependent Ets proteins Pea3 and Erm have previously been shown to be expressed in the embryonic lung and to play a role in lung development (Liu et al., 2003). Expressing a dominant-negative form of Erm in the lung epithelium, Liu et al. showed that Erm and Pea3 are needed for distal differentiation of alveolar type I and type II cells. Furthermore, studies in our laboratory have shown that Erm directly regulates Sftpc transcription by interacting with Ttf1 through its Ets domain (Lin et al., 2006).

Elf5 is a member of the subfamily of Ets transcription factors known as Epithelial Specific Ets, which are expressed only in the epithelial cells (Feldman et al., 2003). Elf5 is expressed primarily in tissues containing glandular epithelium, such as lung, trachea, kidney, prostate, salivary gland, stomach, and mammary gland (Oettgen et al., 1999; Lapinskas et al., 2004; Zhou et al., 2005). Elf5 is essential for embryogenesis, because Elf5 null embryos die at E7.5 from a lack of extraembryonic ectoderm (Donnison et al., 2005; Zhou et al., 2005). Elf5 has previously been reported to be involved in epithelial differentiation in various organ systems. Elf5 is expressed in differentiating keratinocytes (Oettgen et al., 1999; Tummala and Sinha, 2006), transactivates epithelial-specific genes in mammary tissue (Thomas et al., 2000), and is also essential in mammary alveolar differentiation during lactation (Zhou et al., 2005; Harris et al., 2006; Oakes et al., 2006). Elf5 can interact with Ets binding sites in keratinocytes-, prostate-, and salivary gland–specific genes (Oettgen et al., 1999). Elf5 has been previously shown to play a role in regulating gene expression in the A549 and IB3-1 lung epithelial cell lines, where it activates cytokeratin 18 gene expression (Yaniw and Hu, 2005). Using RT-PCR and whole-mount and tissue section ISH, we confirmed that Elf5 is expressed in the epithelium and absent from the surrounding mesenchyme in both developing and postnatal lungs. Furthermore, Elf5 expression in lung epithelium is dynamic: It is initially expressed exclusively in the distal epithelium, but becomes more highly expressed in the proximal epithelium during the later stages of development when distal expression is decreasing. Our data support a model in which FGF10, which is expressed in distal lung mesenchyme during branching morphogenesis (Bellusci et al., 1997), most likely induces Elf5 expression in the distal epithelium. In late gestation, FGF10 expression may decrease in the distal lung, causing a parallel decrement of Elf5 expression. This expression pattern may be significant in the regulation of Sftpc, which our lab has shown is up-regulated by Erm (Lin et al., 2006). Sftpc expression increases in late-gestation distal epithelium, when there is a concomitant decrease in Elf5 expression, suggesting that an absence of Elf5 in the distal epithelium may be needed for robust Sftpc expression. The high expression of Elf5 in the proximal epithelium at this stage is consistent with the possibility that Elf5 may play a role in suppressing Sftpc.

Elf5 Is an FGF-Sensitive, Epithelium-Specific Gene in the Lung

To confirm the Affymetrix gene chip results, we cultured E11.5 distal lung tips in the presence of SU5402 or DMSO and demonstrated by real-time PCR that SU5402 significantly decreased Elf5 expression. Because FGFs also affect the mesenchyme (White et al., 2006), we used mesenchyme-free cultures of lung epithelium to evaluate if SU5402 directly inhibited Elf5 expression. That SU5402 significantly decreased Elf5 expression in cultures of purified lung epithelium showed that FGFs directly regulate Elf5 expression in lung epithelium.

Because several FGFs are expressed in the developing lung, we added FGF1, FGF2, FGF7, FGF9, FGF10, and FGF18 individually to E11.5 lung explant cultures by means of heparin beads and compared their effects on Elf5 expression. We observed that lungs engrafted with FGF2-, FGF7-, FGF9-, and FGF10-soaked beads all appeared morphologically distinct from cultures engrafted with HBSS-soaked beads. Our results are consistent with those observed by other investigators using FGF-soaked beads in lung cultures. FGF7 beads have been shown to induce dilated cystic lumina in lung explants (Park et al., 1998). We also observed dilation in the FGF7 bead cultures after 1 day of culture, although this finding was less pronounced after 2 days. Similar to our observations, FGF10 beads have been shown to induce chemotaxis of lung epithelium in culture (Park et al., 1998). We observed both epithelial and mesenchymal expansion in response to FGF9 beads, which is identical to previous results (White et al., 2006). Similar to cultures with FGF9 beads, we also observed mesenchymal and epithelial expansion in the cultures engrafted with FGF2 beads. A recent publication has also shown that FGF2 added to the medium caused expansion of both the epithelium and mesenchyme of cultured lungs from embryonic Shh null mice (van Tuyl et al., 2006). A notable finding in the study of Liu et al. (2003) was the observation that FGF2 caused a localized increase in expression of the Ets factors Pea3 and Erm in both the epithelium and mesenchyme. We did not observe an increase in Elf5 expression in the vicinity of the FGF2 bead that appeared stronger than that seen in other regions of the explant. FGF1 beads did not induce any appreciable mesenchymal or epithelial expansion, nor did they have any effect on Elf5 expression. Park et al. observed a modest epithelial chemoattraction toward an FGF1 bead, but this finding occurred at a concentration of ligand 5× higher than that used in our studies (Park et al., 1998).

Although FGF2-, FGF7-, FGF9-, and FGF10-soaked beads all induced morphological changes and increased Elf5 expression, only FGF7 and FGF10 beads strongly induced Elf5 expression in the epithelium immediately adjacent to the bead. Notably, this induction was also seen in tracheal epithelium, which does not normally express Elf5. These results are similar to those of Liu et al. (2003), who showed that FGF7 and FGF10 beads induced Pea3 and Erm expression in tracheal epithelium, whereas FGF2 and FGF9 beads did not. Tracheal epithelium expresses FGFR2b (Cardoso et al., 1997), the cognate receptor for FGF7 and FGF10 (Igarashi et al., 1998; Zhang et al., 2006). Previous studies from our laboratory have shown that activation of FGFR2b in tracheal epithelium by FGF7 can reprogram it to express distal lung epithelial differentiation, as evidenced by expression of Sftpc and production of osmiophilic lamellar bodies (Shannon et al., 1999). We propose that FGF7 and FGF10 directly induce Elf5 in tracheal epithelium by means of activation of FGFR2b and that this induction may be another indication of the reprogramming of tracheal epithelium to a distal lung phenotype.

In contrast to FGF7 and FGF10, the effects of FGF2 and FGF9 on Elf5 expression appeared to be more generalized, and we did not observe higher Elf5 expression immediately adjacent to the beads. Furthermore, while Elf5 expression was increased in the lung epithelium, we observed no effects of FGF2 and FGF9 on the tracheal epithelium. It has been previously reported that FGF9 null embryos have decreased FGF10 expression in the lung mesenchyme (Colvin et al., 2001), and FGF7 and FGF10 expression increased in cultured embryonic lungs with the addition of FGF9 (del Moral et al., 2006; White et al., 2006). Exogenous FGF9 has also been shown to increase expression of genes that are known to be FGF7 and FGF10 responsive (White et al., 2006). FGF9 is expressed in the epithelium and mesothelium of the embryonic lung (Colvin et al., 1999), but signals using FGFR splice forms that are found in lung mesenchyme (White et al., 2006; Zhang et al., 2006). The site of FGF2 production in the lung is less clear, with some investigators immunolocalizing FGF2 to the epithelium (Gonzalez et al., 1990) while others do not (Han et al., 1992). Like FGF9, however, FGF2 activates signaling using FGFR splice variants that are present in lung mesenchyme (Zhang et al., 2006). Therefore, the effects of FGF9 on Elf5 expression are likely indirect, resulting from increased levels of FGF7 and/or FGF10 in the lung mesenchyme, which may in turn be due to increased mesenchymal mass. Tracheal mesenchyme may not have the competence to increase expression of FGF7 or FGF10 upon stimulation with FGF2 or FGF9, which could explain the lack of effect of FGF2 and FGF9 on Elf5 expression in the tracheal epithelium. It is also interesting that we observed no effect of FGF1, which activates all FGF receptors, including FGFR2b (Zhang et al., 2006), on Elf5 expression. This finding suggests that there is another level of complexity involved in the binding of FGF1 and its activation of signaling in the lung epithelium, such as heparan sulfate proteoglycans, which can modulate binding of FGF1 to FGF receptors (Ornitz, 2000; Pellegrini, 2001).

We also examined if Elf5 could be induced in vivo by FGF7 and FGF10 during the saccular stage of lung development on E18.5, when Elf5 is normally expressed at low levels in distal epithelial cells. We found that strong Elf5 expression was induced in the lungs of transgenic E18.5 fetuses misexpressing either FGF7 or FGF10. This finding was also seen in the lungs of adults when FGF10 was misexpressed in the lung epithelium, but only in areas of adenomatous hyperplasia, which may suggest a role for Elf5 in lung epithelial tumorigenesis. Supporting this concept, a previous study has shown that Elf5 is increased in mouse mammary tumor models (Galang et al., 2004). In contrast, however, another study demonstrated that many human carcinoma cell lines derived from Elf5-expressing tissue cease to express Elf5, suggesting that Elf5 has tumor suppressing properties (Zhou et al., 1998). Further studies will be needed to elucidate the importance of Elf5 in tumorigenesis.

Elf5 Induction Is Downstream of PI3-Kinase Signaling

We wished to determine the signaling pathway(s) downstream of FGF receptors that control Elf5 expression. The two signaling pathways most studied as downstream components of FGF signaling are the PI3-Kinase/Akt and MAP Kinase/Erk pathways. The PI3-Kinase pathway has been traditionally described as being important for proliferation, cell survival, and anti-apoptosis (Song et al., 2005), whereas the MAP Kinase/Erk pathway is active in proliferation, differentiation, and cell motility (Huang et al., 2004). Given the caveat that they have not been tested on all known kinases, LY294002 and U0126 are generally accepted as inhibitors of PI3-Kinase/Akt and MAP Kinase/Erk signaling, respectively. Wang et al. showed that 10 μM LY294002 specifically decreased pAKT and lung branching in E11.5 mouse lung explants (Wang et al., 2005). Kling et al. have previously shown that lung explants treated with 20 μM U0126 had decreased pERK as well as decreased branching (Kling et al., 2002). We found that Elf5 expression decreased in lung tip explants that were treated with LY294002, but not with U0126, indicating that PI3-Kinase signaling by means of activated Akt maintains Elf5 expression. The importance of PI3-Kinase in Elf5 expression was further demonstrated by our observation that Elf5 expression decreased when mesenchyme-free lung epithelial rudiments were treated with LY294002, but did not change with U0126 treatment. That U0126 inhibited epithelial branching in tips but had no effect on Elf5 expression suggests that Elf5 is not sufficient to support branching morphogenesis. U0126 likely inhibits one (or more) of the other genes involved in the highly complex process of lung branching. In addition, we found that E17.5 airway epithelium coexpressed pAkt and Elf5, which correlates with our in vitro data. Our observation that an FGF can use the PI3-Kinase but not the Map/Kinase/Erk pathway to regulate a specific function is not novel. Chandrasekher et al. have shown that FGF7 stimulated the PI3-Kinase pathway during wound healing of corneal epithelial cells, but inhibition of the MAP Kinase/ERK pathway had no effect (Chandrasekher et al., 2001).

In this study, we have identified genes in embryonic lung epithelium that are FGF-responsive. We have shown that one of these, Elf5, is dynamically expressed during lung development. FGF7 and FGF10, which signal through the FGFR2b splice variant in lung epithelium, were most effective at stimulating Elf5, and this required activation of the PI3-Kinase/Akt pathway. Our data do not, however, exclude the possibility that other signaling molecules that activate PI3-Kinase/Akt may also regulate Elf5. The function of Elf5 in the lung is not clear. Previous studies have reported that Elf5 can activate tissue-specific genes in the epithelia of mammary gland, keratinocytes, prostate gland, and salivary gland (Oettgen et al., 1999; Thomas et al., 2000; Zhou et al., 2005; Harris et al., 2006; Oakes et al., 2006; Tummala and Sinha, 2006). Future studies will be needed to identify Elf5 target genes in the lung epithelium and to assess the function of Elf5 in lung morphogenesis.


  1. Top of page
  2. Abstract
  7. Acknowledgements

Embryo and Tissue Isolation

All protocols involving the use of mice were reviewed and approved by the Institutional Animal Care and Use Committee at Cincinnati Children's Medical Center. Timed-pregnant FVB/N female mice (Taconic) were killed to obtain tissues at specific developmental time points. Lungs from E11.5 and E12.5 embryos were dissected and used for explant culture experiments. Lungs from E11.5, E12.5, E14.5, 15.5, 16.5, 17.5, and 18.5 embryos, as well as from postnatal day 1 (PN1) and adult mice were fixed for section or WM-ISH.

Transgenic Mice

Transgenic mice conditionally expressing FGF7 or FGF10 in the lung epithelium were generated from two transgenic mouse lines: (1) “activator” mice (CCSP-rtTA) expressing reverse tetracycline transactivator (rtTA) under control of a 2.3-kb element from the rat Clara cell secretory protein (CCSP) promoter, which confers expression in both proximal and distal lung epithelial cells; and (2) “responder” transgenic mice ((tetO)7-FGF7 or (tetO)7-FGF10), in which FGF7 or FGF10 are expressed by a minimal CMV promoter under control of tetracycline response elements (Tichelaar et al., 2000; Clark et al., 2001). (tetO)7-FGF7 or (tetO)7-FGF10 mice were crossed with CCSP-rtTA transgenic mice to generate offspring that conditionally expressed FGF7 or FGF10 in the lung epithelium upon treatment with doxycycline.

Lung Explant Culture

Lungs or lung tips dissected from E11.5 mouse embryos were cultured on 8.0 μM Whatman Nuclepore membranes (Whatman International, Clifton, NJ) floating on 1 ml of DMEM/F12 medium (Invitrogen) plus 5% fetal bovine serum (FBS; Sigma) in a three-well dissecting dish. Medium additions included 0.1% DMSO or SU5402 (10 μM) or LY294002 (20 μM) or U0126 (20 μM; all from EMD Biosciences, La Jolla, CA). Whole lung and lung tip explants were cultured for 1 day and then harvested for RNA isolation in 4 M guanidinium isothiocyanate (GITC).

To test the effects of individual FGFs, E11.5 lung/trachea complexes were engrafted with heparin beads (Sigma) soaked in different individual FGFs and then cultured for 2 days. Beads were soaked for 3–4 hr while rocking in 10 μl of FGF1 (R&D Systems, Minneapolis MN), FGF2 (Promega, Madison, WI), FGF7 (Peprotech Inc., Rocky Hill, NJ), FGF9 (R&D Systems), FGF10 (Peprototech Inc.), or FGF18 (Peprotech Inc.). All FGFs were used at 100 ng/μl. The beads were then rinsed once in HBSS and then grafted into lung explants. After 2 days of culture, lungs were fixed for WM-ISH.

Microarray Analysis

After E11.5 lung explants were cultured for 1 day with either 0.1% DMSO or 10 μM SU5402, the distal tips were removed and separated into purified epithelial and mesenchymal populations as previously described (Shannon et al., 1998). RNA from three independent experiments was purified from the isolated epithelium and mesenchyme using an RNeasy Micro kit (Qiagen, Valencia, CA), amplified using an Ovation Biotin RNA amplification and labeling system (NuGen Technologies, San Carlos, CA), and hybridized to murine genome MOE430 chips (contains approximately 45,000 gene entries) according to the manufacturer's protocol (Affymentrix, Santa Clara, CA). Affymetrix Microarray Suite 5.0 was used to scan and quantify gene expression under default scan settings. Differentially expressed genes between SU5402- vs. DMSO-treated epithelium samples were identified by Student t-test at a P value ≤ 0.05 (Genespring GX 7.3, Silicon Genetics, Redwood City, CA). Data were further analyzed using additional filters for candidates that are only expressed in epithelium and down-regulated by SU5402. The filters include a minimal of twofold decrease in SU5402 treated samples, a minimal of two present calls by the Affymetrix algorithm in three epithelium samples, and not present in mesenchymal samples. Gene ontology analysis was preformed using the publicly available Web-based tool David (database for annotation, visualization, and integrated discovery; Dennis et al., 2003). Functional classification is based on both Gene ontology analysis and literature information (Pubmed

Mesenchyme-Free Lung Epithelium Cultures

Distal tips were removed from E12.5 lungs, and the epithelium was separated from mesenchyme as previously published (Shannon et al., 1998); this procedure yields epithelial rudiments that are not contaminated by residual mesenchymal cells (Deterding and Shannon, 1995). Epithelial rudiments were enrobed in growth-factor reduced Matrigel (BD Biochemical) and covered with medium formulated previously in our laboratory (Shannon et al., 1999) that supports both proliferation and differentiation of embryonic lung epithelium. This medium (hereafter BFGM) consists of DMEM/F12 containing 5% charcoal-stripped FBS, insulin 10 μg/ml, cholera toxin 1 μg/ml, human recombinant epidermal growth factor (EGF) 25 ng/ml, human recombinant hepatocyte growth factor (HGF) 10 ng/ml, FGF7 (Peprotech) 25 ng/ml, and FGF1 100 ng/ml. For the SU5402 treatment experiments, epithelial rudiments were cultured in BFGM for the first 3 days, then cultured for another 2 days in BFGM additionally containing SU5402 (10 μM) or 0.1% DMSO. For the LY294002 and U0126 treatment experiments, epithelial rudiments were cultured for 3 days in BFGM minus FGF1 then cultured for an additional 24 hr with BFGM containing either LY294002 (20 μM), or U0126 (20 μM), or 0.1% DMSO. Tissue was excised from the Matrigel at the end of culture, and RNA was isolated using an RNeasy Micro Kit.


RNA was converted to cDNA using ImProm-II reverse transcriptase (Promega). Real-Time PCR was performed using SYBR green I mix (Roche Molecular Biochemicals), 2.0 mM (for Elf5 primer set I) or 2.5 mM (for β-actin and Elf5 SuperArray primer sets) MgCl2, and 0.25 mM of each primer in a Smart Cycler (Cepheid, Sunnyvale, CA). The sequences of Elf5 primer set I were 5′-GGACTCCGTAACCCATAGCA-3′ and 5′-TACTGGTCGCAGCAGAATTG-3′, which gives a product of 207 base pairs. Reaction conditions for Elf5 primer set I were 95°C for 300 sec and then cycled through 95°C for 10 sec, 58°C for 15 sec, 72°C for 20 sec until fluorescence reached threshold (approximately 30–35 cycles); the efficiency of this amplification was 86%. Real-Time PCR analysis of mesenchyme-free epithelial cultures treated with LY294002 and U0126 was done using RT2 PCR Primers for Elf5 (SuperArray Bioscience Corporation, Frederick, MD) following manufacturer's instructions to give a product of 154 base pairs; the efficiency of this amplification was 93%. The sequences of the β-actin primers were 5′-TGGAATCCTGTGGCATCCATGAAAC-3′ and 5′-TAAAACGCAGCTCAGTAACAGTCCG-3′, which gives a product of 350 base pairs. Reaction conditions for β-actin were 95°C for 150 sec and then cycled through 95°C for 10 sec, 60°C for 15 sec, 72°C for 20 sec until fluorescence reached threshold. Elf5 and β-actin relative expression values were obtained from standard curves (r2 ≥ 0.99) created for Elf5 and β-actin primers using a dilution series of control cDNA. Normalized data were calculated by dividing the Elf5 value by the β-actin value for each sample and are presented as percentages of DMSO controls. RT-PCR for SP-C expression was used to demonstrate that preparations of purified lung mesenchyme were not contaminated with epithelial cells. The sequences of the primers were 5′-CATACTGAGATGGTCCTTGAG-3′ and 5′-TCTGGAGCCATCTTCATGATG-3′, which gave a product of 207 bp.

In Situ Hybridization

A full-length Elf5 clone was generated using RT-PCR of Ell.5 lung cDNA. The sequences of the primers used were 5′-TGAAAGCCTTCTGTCTGGACC-3′ and 5′-CCATCAAATGCGCCTGGTGT-3′. The 836-bp product was then cloned into pcDNA3.1 (Invitrogen). This construct was subsequently used as template to generate Elf5 ISH probes by PCR. The sequences of the primers used were 5′-GGGGTACCCCTTGAAAGCCTTCTGTCTGG-3′ and 5′-CGGGATCCCGAGGGCTTCTGACTTAACCA-3′, which produced a 642-base pair product lacking the sequence for the Ets domain, and additionally contained KpnI and BamHI restriction ends to facilitate cloning into pGEM-4Z. This template was linearized, and a Riboprobe kit (Promega) was used to make either digoxigenin or 33P-UTP labeled probes.

Samples for WM-ISH were fixed in 4% paraformaldehyde (PFA) and then dehydrated to 100% methanol for storage at 4°C. WM-ISH was performed as previously described (Wilkinson, 1992). Tissue samples for tissue section ISH were fixed in 4% PFA and then dehydrated to 70% ethanol and then embedded in paraffin. The hybridization procedure was performed as previously described (Deterding and Shannon, 1995), with the exception that 33P-UTP was used to radiolabel probes.


pAkt was detected in paraffin sections with a 1:50 dilution of a rabbit monoclonal antibody against pAkt (Ser473; Cell Signaling Technology, Danvers, MA) using the manufacturer's protocol. Biotinylated anti-rabbit secondary antibody was used along with Vectastain ABC kit (Vector Laboratories) to visualize pAkt staining.


Real-Time PCR data were statistically analyzed using GraphPad Prizm version 4.0 (GraphPad Software, San Diego, CA). Data from SU5402-treated lung tip cultures (Fig. 2H) and SU5402-treated epithelium cultures (Fig. 3I) were analyzed using paired t-tests. Data from lung tip cultures (Fig. 7G) and lung epithelium cultures treated with DMSO, LY294002, and U0126 (Fig. 8G) were analyzed by one-way analysis of variance.


  1. Top of page
  2. Abstract
  7. Acknowledgements

We thank Kalpana Srivastava, Xiaofei Shangguan, Michael Burhans, Nambirajan Sundaram, Kathleen McCormick-Shannon, and Yanhua Wang for their expert technical assistance.


  1. Top of page
  2. Abstract
  7. Acknowledgements