FGF-Regulated ETV Transcription Factors Control FGF-SHH Feedback Loop in Lung Branching

Abstract

The mammalian lung forms its elaborate tree-like structure following a largely stereotypical branching sequence. While a number of genes have been identified to play essential roles in lung branching, what coordinates the choice between branch growth and new branch formation has not been elucidated. Here we show that loss of FGF-activated transcription factor genes, Etv4 and Etv5 (collectively Etv), led to prolonged branch tip growth and delayed new branch formation. Unexpectedly, this phenotype is more similar to mutants with increased rather than decreased FGF activity. Indeed, an increased Fgf10 expression is observed, and reducing Fgf10 dosage can attenuate the Etv mutant phenotype. Further evidence indicates that ETV inhibits Fgf10 via directly promoting Shh expression. SHH in turn inhibits local Fgf10 expression and redirects growth, thereby initiating new branches. Together, our findings establish ETV as a key node in the FGF-ETV-SHH inhibitory feedback loop that dictates branching periodicity.

Figure 1. Inactivation of Etv in the lung epithelium led to epithelial branching defects

(A) Etv5 inactivation was efficient as evidenced by the clear reduction of full-length transcripts in the Etv mutant lung (E11.5: 1.0 for controls, 0.14 for Etv mutants, p=0.005; E12.5: 1.0 for controls, 0.15 for Etv mutants, p=0.02, n=3 each group). (B-I) Representative control (B,D,F,H) and Etv mutant (C,E,G,I) whole lungs with epithelium outlined by anti-E-Cadherin immunohistochemical staining. In the mutant, the tip dilation phenotype was already apparent at E10.5 (arrowhead in C). The reduced tip number phenotype was apparent shortly after the initiation of secondary branching at E11.5; the position of the bud for the future accessory lobe was shifted more posteriorly compared to control (arrowheads in D and E). (J) Tip area is increased in the left lung at the indicated stages (E12.5: 1.0 for controls and 3.08 for Etv mutants, p=0.001; E13.5: 1.0 for controls and 2.05 for Etv mutants, p=0.018). (K) Tip number is decreased in the Etv mutant Left lung at the indicated stages (E11.5: 1.0 for controls, 0.5 for Etv mutants, p=.004; E12.5: 1.0 for controls, 0.55 for Etv mutants, p=0.011; E13.5: 1.0 for controls, 0.67 for Etv mutants, p=0.024. Actual tip numbers at the three stages shown are: E11.5: control 3.50 ± 0.57 versus mutant 1.75 ± 0.50; E12.5: control 9.67± 1.15 versus mutant 5.33 ± 0.58; E13.5: control 16.00 ± 2.00 versus mutant 10.67 ± 1.53). Examples of how the tip areas were defined were indicated by arrowheads and outlined in the insets in F-G. Quantification was carried out in n≥4 samples for each genotype and stage. Data are presented as standard error of the means (+SEM), as in graphs in all figures. See also Figure S1.

Figure 2. Cell differentiation is largely normal in Etv mutant lungs

(A-H) Immunofluorescent labeling of airway cell types using indicated markers for club cells (A,B), pulmonary neuroendocrine cells (C,D), ciliated cells (E,F) and basal cells at E18.5 (A-F) or adult (G,H). (I,J) Wholemount RNA in situ hybridization of Scgb1a1 to outline the airway of E18.5 left lobe. (K) Quantification of E18.5 airway phenotype. The terminal bronchiole size was calculated by measuring the tips of Scgb1a1 outlined bronchiole alveolar junction (for ratio: 1.0 for controls and 1.35 for Etv mutants, p=0.011, n=3). The secondary bronchi number was calculated by counting all branches off the main left lung bronchus (for ratio: 1.0 for controls and 0.67 for mutants, p=0,016, n=3; actual average tip number 6±0 for controls and 4±0.8 for mutants). The terminal bronchiole number was calculated by counting the tips of Scgb1a1 outlined bronchiole alveolar junction (for ratio: 1.0 for controls and 0.47 for mutants, p=2.8×10⁻⁶, actual tip number 29±0.6 for controls and 14±0.5 for mutants). See also Figure S2.

Figure 3. Etv mutant lungs exhibited an increase in FGF signaling activity

(A-F) Representative E11.5 whole lung immunofluorescently stained for E-Cadherin (red) and pERK (green). pERK staining in the mutant lung was increased in level and expanded in domain compared to control. (G-I) Fgf10 expression was increased as shown by qRT-PCR at stages indicated (E11.5: 1.0 for controls, 2.72 for Etv mutants, p=0.03; E12.5: 1.0 for controls, 2.29 for Etv mutants, p=0.01, n=3 each group) (G) and RNA in situ hybridization at E12.5 (H,I). See also Figure S3.

Figure 4. Reducing Fgf10 gene dosage in the Etv mutant background led to attenuation of the tip dilation phenotype

(A-C) Representative E13.5 lung of indicated genotype with the epithelium outlined by anti-E-cadherin immunohistochemical staining. (D) Introducing the Fgf10 mutant allele attenuated the branch tip area phenotype (area: 1.0 for controls, 2.5 for Etv mutants, 1.22 for Etv;Fgf10 mutants; p=0.009 for Etv mutants versus Etv;Fgf10 mutants, and p=0.075 for controls versus Etv;Fgf10 mutants). (E) Branch tip number was not attenuated by introducing the Fgf10 mutant allele (tip number: 1.0 for controls, 0.562 for Etv mutants, 0.654 for Etv;Fgf10 mutants, p=0.32). Quantification was carried out in n=3 samples for each genotype. See also Figure S4.

Figure 5. SHH signaling was decreased in the Etv mutant lung

(A-F) Representative whole mount RNA in situ hybridization of left lobes of E12.0-E12.5 lungs with each set of control and mutant as littermates. (G) Quantification of expression by qRT-PCR (Shh: 1.0 for controls, 0.56 for Etv mutants, p=0.027; Gli1: 1.0 for controls, 0.77 for Etv Mutants, p=0.013; Ptch1: 1.0 for controls, 0.61 for Etv mutants, p=0.022). Quantification was carried out in n≥3 samples for each.

Figure 6. SAG treatment of Etv mutant lungs led to attenuation of tip dilation phenotype

(A-H) Representative images of E11.5 lungs cultured for 24 hours in either DMSO (A,B,E,F) or SAG (C,D,G,H). In the control lungs, SAG treatment did not affect tip size (areas outlined by dashes, 1.0 DMSO versus 1.01 SAG, p=0.89) (B,D). In the Etv mutant lungs, SAG treatment led to a reduction of tip size towards the size of the control (2.10 DMSO versus 1.22 SAG, p=0.015) (F,H). (I) Quantification of the right lung epithelial tip size (n=3 for each). See also Figure S5.

Figure 7. ETV controls Shh expression through putative binding sites in a long-range enhancer

(A) The MACS1 lung epithelium enhancer lies approximately 800kb upstream of the Shh transcriptional start site, and contains three highly conserved ETV binding sites (red) and one highly conserved NKX2-1 binding site (blue). (B) The three ETV binding sites were each mutated to nucleotides previously shown to abolish binding (de Launoit et al., 1998). (C) Relative luciferase activity from MLE12 cells transfected with either wild-type (wt) or mutant (mut) MACS1 enhancer; together with either Etv5 empty vector, wild-type (wt) Etv5 vector, or mutant (mut) Etv5 with disrupted DNA binding domain vector (1.0 for WT MACS1+no Etv5, 2.14 for WT MACS1+wt Etv5, 1.06 for wt MACS1+mut Etv5, 0.88 for mut MACS1+no Etv5, and 1.19 mut MACS1+wt Etv5, n=3 for each group). (D,E) Representative β-gal staining of transgenic lungs carrying lacZ reporter driven by either wt or mut Shh enhancer. (F) Percent recovery compared to inputs of either MACS1, a control fragment approximately 1.2 kb upstream of MASC1 (control region 1), or a control fragment approximately 800kb downstream of MACS1 near the Shh gene (control region 2), by anti-Flag antibody against ETV5-Flag, or no antibody control. The extract was prepared from lung epithelial MLE12 cells with overexpression of Etv5-Flag and Nkx2-1 plasmids. (G) A model of Etv regulation of lung epithelial branching. The outside circle represents the growth and bifurcation that constitute each reiterated cycle of branching. In the Etv mutant, decrease in Shh and increase in Fgf10 leads to prolonged growth and delayed branching. See also Figure S6.