|Home | About | Journals | Submit | Contact Us | Français|
The key role played by of Fgf10 during early lung development is clearly illustrated in Fgf10 knockout mice, which exhibit complete lung agenesis. However, Fgf10 is continuously expressed throughout lung development suggesting extended as well as additional roles for FGF10 at later stages of lung organogenesis. We previously reported that the enhancer trap Mlcv1v-nLacZ-24 transgenic mouse strain functions as a reporter for Fgf10 expression and displays decreased endogenous Fgf10 expression (Mailleux et al., 2005). In this paper, we have generated an allelic series to determine the impact of Fgf10 dosage on lung development. We report that 80% of the newborn Fgf10 hypomorphic mice die within 24 hours of respiratory failure. These mutant lungs display severe hypoplasia, dilation of the distal airways and large hemorrhagic areas. Epithelial differentiation and proliferation studies indicate a specific decrease in the percentile of TTF1 and SP-B expressing cells correlating with reduced epithelial cell proliferation and associated with a decrease in activation of the canonical Wnt signaling in the epithelium. Analysis of vascular development shows a reduction in PECAM expression at E14.5, which is associated with a simplification of the vascular tree at E18.5. We also show a decrease in α-SMA expression in the respiratory airway suggesting defective formation of the alveolar smooth muscle cells. At the molecular level, these defects are associated with a decrease in Vegfa and Pdgfa expression likely resulting from the decrease of the epithelium/mesenchymal ratio in the Fgf10 hypomorphic lungs. Thus, our results indicate that FGF10 plays a pivotal role in maintaining epithelial progenitor cell proliferation as well as coordinating alveolar smooth muscle cell formation and vascular development.
Fibroblast Growth Factor 10 (FGF10) is responsible for directed outgrowth of the lung endoderm (Bellusci et al., 1997). In the developing lung, Fgf10 is expressed in the distal mesenchyme at sites where prospective epithelial buds will appear. Moreover, its dynamic pattern of expression and its ability to induce epithelial expansion and budding in organ cultures have led to the hypothesis that FGF10 governs the directional outgrowth of lung buds during branching morphogenesis (Bellusci et al., 1997). Furthermore, FGF10 was shown to induce chemotaxis of the distal lung epithelium (Park et al., 1998; Weaver et al., 2000). Consistent with these observations, Fgf10 null mutants show multiple organ defects including complete lung agenesis (Min et al., 1998; Sekine et al., 1999), therefore excluding the study of subsequent FGF10 roles during lung development. FGF10 is the main ligand for Fibroblast Growth Factor Receptor 2b (FGFR2b) during embryonic development as demonstrated by the remarkable similarity of phenotypes exhibited by embryos where these genes have been inactivated (Min et al. 1998; Sekine et al. 1999, Burns et al., 2004; De Moerlooze et al., 2000; Mailleux et al., 2002; Burns et al., 2004; del Moral et al., 2006). Moreover, inhibition of FGFR2b signaling at embryonic day 14.5 (E14.5) using a transgenic mouse line expressing a soluble FGFR2b (FGFR2b-HFc) under the control of an inducible lung-specific, Surfactant protein C promoter (SpC-rtTA), resulted in decreased epithelial morphogenesis before birth and caused severe emphysema at maturity (Hokuto et al., 2003). Interestingly, Fgf10 expression levels increase as the lung develops between E11.5 and E18.5, (Bellusci et al., 1997), suggesting that Fgf10 most likely plays an extended and vital role during late lung organogenesis.
We have previously shown that a transgenic mouse line with the β-galactosidase gene under the control of Fgf10 regulatory sequences can be used to monitor Fgf10 expression in the heart, lung and somites (Kelly et al., 2001; Mailleux et al., 2005; Veltmaat et al., 2006). Originally, Kelly et al. (2001) sought to express LacZ under the control of the myocardial ventricular-slow skeletal muscle Myosin Alkali-light chain (Mlc1v) promoter. In one of four founders, the expression pattern in the developing heart suggested that LacZ was under the control of Fgf10 regulatory sequences. In addition, analysis of the integration site showed that the Mlc1v-LacZ cassette had integrated 120 kb upstream of the Fgf10 gene. Based on the expression pattern as well as on the site of insertion, the authors proposed that LacZ expression was under the control of Fgf10 regulatory sequences. We reported that the insertion of the Mlc1v-LacZ cassette disrupted the endogeneous expression of Fgf10 (Mailleux et al., 2005). The Mlc1v-LacZ mice were crossed with Fgf10+/− mice to generate an allelic series to determine the effect of decreasing Fgf10 expression on parabronchial smooth muscle cell formation. We also demonstrated that Fgf10 identifies a new population of parabronchial smooth muscle cell progenitors located in the sub-mesothelial mesenchyme (Mailleux et al., 2005).
In this paper, we report our analysis of Fgf10 hypomorphic (Fgf10+/−; Mlc1v-LacZ+/−) lungs at various developmental stages. Newborn Fgf10 hypomorphic mice exhibit respiratory failure and consequently die within 48 hours. Analysis shows severe lung hypoplasia, dilation of the distal airways and large hemorrhagic areas. Therefore, this mouse represents a unique opportunity to study the role of FGF10 in coordinating epithelial morphogenesis with the differentiation of the mesenchyme, in particular with the formation of alveolar smooth muscle cells and the vasculature.
TOPGAL expression was monitored by β-galactosidase activity using wholemount and histological revelation as described by Kelly et al. (1995). The day of vaginal plug was considered as embryonic day 0.5 (E0.5). Embryos were dissected in 1x PBS at desired stages and fixed in 4% PFA. Following the fixation, the embryos were washed in 1x PBS and stained in X-gal solution. The stained lungs were sectioned (25μm) in a vibrotome.
The Mlc1v-nLacZ-24 line (called for simplification Mlc1v-LacZ or Mlc1v in this paper) has been previously described (Kelly et al., 2001). The transgene containing an nLacZ reporter gene (containing a nuclear localization signal) is integrated upstream of the Fgf10 gene. Fgf10+/−; Mlc1v-LacZ+/− embryos were generated by crossing Fgf10+/− and Mlc1v-LacZ+/− mice (Kelly et al., 2001; Sekine et al., 1999) on a C57BL/6 background. Fgf10+/− littermates were used as control embryos at different developmental stages. The Fgf10- and Mlc1v-LacZ+ alleles were genotyped as described previously (Kelly et al., 2001; Mailleux et al., 2002). The number of Fgf10+/−; Mlc1v-LacZ+/− embryos used in this study (47 in total) at the different stages was as follows: E12.5 (n=7), E13.5 (n=3), E14.5 (n=6), E16.5 (n=2), E17.5 (n=10); E18.5 (n=7); PN (n=12).
Total RNA was extracted from individual Fgf10+/− and Fgf10+/−; Mlc1v+/− E14.5 embryonic lungs (n= 3 for each genotype) using the RNeasy kit (Gibco BRL) according to the manufacturer’s instructions. DNA contaminations were removed in the total RNA using Turbo DNAse (Ambion). Total RNA was reverse-transcribed using the Superscript-III first strand super mix (Invitrogen) following the manufacturer’s recommendations. 5 μg of the total RNA was used to prepare cDNA from the isolated total RNA using oligodT primers. 25 pg cDNA was used for each of the real time PCR reaction using the primers and probes designed by the online Roche software: Probe finder version 2.20, https://www.roche-applied-science.com/sis/rtpcr/upl/adc.jsp. All real time PCR reactions were performed with Roche: FastStart TaqMan® Probe Master kits, according to the manufacturer’s instructions in Roche Light Cycler 1.5 Real Time PCR machine. 18s ribosomal RNA was used as an internal control for all analysis.
Newborn lungs (3 Fgf10+/−; Mlc1v-LacZ+/− and 3 control) were fixed overnight in paraformaldehyde, rinsed in PBS twice for five minutes, transferred to 70% ethanol overnight and stored in 100% ethanol. The samples were then embedded in paraffin and sections (5 μm) were cut. The number of cells expressing SP-B and TTF1 was quantified using immunochemistry protocol with Envision + HRP system (Dakocytomation). Alternatively, expression of SP-B was also determined using in situ hybridization on sections using a specific mouse SP-B probe (gift from Dr. Jeffrey Whitsett). Positive Cells were scored in random portions of a section in eight photomicrographs (50X magnification). A total number of 3000 cells were counted per sample. The results are presented as a ratio of total number of positive cells/total number of cells. Lungs from three mutant and three control mice were taken at birth.
The following antibodies were used for immunochemistry on 5 μm paraffin sections: a mouse monoclonal antibody against α-SMA (Sigma) at a dilution of 1/5000, a mouse monoclonal antibody against Phospho-p44/42 MAPK (cell Signaling: 20G11) at a dilution of 1/50, a mouse monoclonal antibody against Thyroid Transcription factor-1 (NeoMarkers), a rabbit polyclonal antibody against Surfactant protein B (Chemicon), anti-PECAM/CD31 antibody (BD Pharmage), anti-Laminin antibodies (Sigma), anti-Elastin polyclonal antibodies (EPC) at a dilution of 1/200. Slides were mounted with Vectashield (Vector Labs) containing DAPI. Immunohistochemistry for β-catenin was performed with the Envision kit from Dako cytomation, β-catenin antibodies (BD Biosciences) was used at 1:100. Photomicrographs were taken using a Leica DMRA fluorescence microscope with a Hamamatsu Digital CCD Camera.
To visualize the vessels of the vascular system of E18.5 embryonic lung, we used Batson’s #17 Anatomical Corrosion Kit (Polysciences). The pregnant female mouse was euthanized according to institute regulations. The abdominal area was opened and 5 U of heparin (1000 U/ml) were injected into the inferior vena cava. Embryos were then isolated, the sternum exposed and approximately 4 ml of a methyl mathacrylate based resin was injected into the right ventricle of the heart. The embryos were cured for 3 hours in ice-cold water and were subsequently soaked in distilled water overnight at 4°C. Embryos were then digested in maceration solution using approximately three times volume of tissue being digested and were incubated at 50°C for 72 hours or until all tissue was cleared. Maceration solution was changed every 3 hours. Images of the cast were documented using a Leica MZ125 with SPOT v3.2.0 digital camera.
Dehydrated E14.5 and E17.5 lungs (3 Fgf10+/−; Mlc1v-LacZ+/− and 3 control for each stage) were fixed overnight in paraformaldehyde, rinsed in PBS twice for five minutes, transferred to 70% EtOH overnight and stored in 100% EtOH. The samples were then embedded in paraffin and sections (5 μm) were cut. The PCNA staining kit (Zymed Labs Inc. 93-1143) was used for immunostaining. At E14.5, the total numbers of cells in the epithelium as well as the number of PCNA positive cells in the epithelium were scored in three photomicrographs (64x magnification), in random portions of a section of three different mutant and control lungs each. A similar count was done for mesenchymal cells. At E17.5, as it was no longer possible to distinguish between epithelium and mesenchyme, the total number of cells versus the total number of PCNA positive cells were counted for each genotype. A total number of 3000 cells were counted per sample. The significance in proliferation between control and hypomorphic lungs was evaluated by one-tailed paired t-test. P values less than 0.05 were considered to be statistically significant.
In situ zymogram assays were carried out using fluorescein isothyocyanate labeled gelatin (Molecular Probes, Eugene, Oregon) to detect gelatinase activity (corresponding to MMP2 and MMP9 expressed at high levels in the smooth muscle cells). The reaction product was visualized using fluorescence microscopy.
We recently published that the insertion of the Mlvc1v-LacZ cassette 120 kb upstream of the transcriptional start site of the Fgf10 gene, resulted in a general decrease in Fgf10 expression in the embryo (Mailleux et al., 2005; Veltmaat et al., 2006). We took advantage of this characteristic of the Mlc1v-LacZ line to generate an allelic series to determine the impact of Fgf10 dosage on lung development. Mlc1v-LacZ+/− line was crossed with Fgf10neo null allele heterozygote line (denoted as Fgf10-) to generate Fgf10+/+; Mlc1v-LacZ−/− (wild type), Fgf10+/+; Mlc1v-LacZ+/−, Fgf10+/−; Mlc1v-LacZ−/− (called thereafter Fgf10+/−) and Fgf10+/−; Mlc1v-LacZ+/− (called also thereafter Fgf10 hypomorph) embryos at different developmental stages. As the lungs of embryos corresponding to the first three genotypes (Fgf10+/+; Mlc1v-LacZ−/− (wild type), Fgf10+/+; Mlc1v-LacZ+/− and Fgf10+/−; Mlc1v-LacZ−/− (or Fgf10+/−)) were phenotypically identical throughout all the developmental stages examined (data not shown), we considered as control for this study, among these 3 genotypes, the one corresponding to the lowest expression of Fgf10 (Fgf10+/−). 47 Fgf10+/−; Mlc1v-LacZ+/− compound heterozygous embryos were generated at expected Mendelian ratios and examined at different stages. As expected, real time PCR experiments show that Fgf10 expression is reduced by 27% in E14.5 Fgf10+/−; Mlc1v-LacZ+/− lungs compared to corresponding Fgf10+/− lungs (Fig. 1A,B; n=3, P=0.005).
The Fgf10+/−; Mlc1v-LacZ+/− embryos displayed defects in lung development. At E12 and E12.5, the lung phenotype was characterized by a decrease in epithelial branching (Fig. 1C-F). Consistent with our hypothesis that this phenotype results from reduced Fgf10 levels, a similar phenotype was observed in transgenic lungs over-expressing Sprouty2, a negative regulator of FGF10 signaling (Mailleux et al., 2001). Primary lobe formation in the mutant lungs was generally not affected, except for the accessory lobe, which was absent in 9 of the 52 Fgf10+/−; Mlc1v-LacZ+/− lungs generated (see arrow in Fig. 1G,H). In addition, the remaining lobes in the mutant lungs were substantially smaller in volume than their control counterparts. In order to rule out general growth retardation as the cause for the hypoplastic lungs, the body weight (Fig. 1I) and lung weight (data not show) of control (Fgf10+/−) and Fgf10 hypomorphic embryos at different developmental stages were measured (the number of embryo examined were n=3 at E14.5, n=5 at E17.5 and n=7 at birth, for control and mutant). While the body weight between control and mutant embryos were not significantly different at a given developmental stage (P=0.28 at E14.5; P=0.05 at E17.5 and P=0.26 at birth), there is a clear reduction in the lung weight of the mutant embryos compared to the littermate controls at the three stages examined (7.5 mg ± 2.8, 16.8 mg ± 2.1; 20,2 mg ± 0.6 vs. 13.7 mg ± 2.1, 34.8 mg ± 10.6, 45.7 mg ± 15 at E14.5, E17.5 and birth, respectively, P<0.01 for all three stage examined). This observation is confirmed by the calculation of normalized wet weight ratios (wet weight of the lung divided by the wet weight of the embryo; 1.7% ± 0.1, 1.9% ± 0.4 and 1.7% ± 0.4 for the Fgf10+/−; Mlc1v-LacZ+/− mutants at E14.5, E17.5 and birth, respectively; versus 2.9% ± 0.7, 3.5% ± 0.7 and 3.4% ± 1.0 for the corresponding control embryos or pups, respectively, Fig. 1J,). The differences observed at each stage are statistically significant (P<0.01 for all three stages examined). Thus, the lung growth defects observed in Fgf10 hypomorphic embryos are not linked to general growth retardation.
Analysis of PCNA positive cells as a read out for proliferation was carried out in control and mutant lungs at E14.5 and E17.5 (Fig. 2A-D). At E14.5, epithelial proliferation is drastically reduced in the mutant vs. control lung (11.0% ± 2 vs. 26.4% ± 4.1; n=3 for each genotype, P<0.05). Interestingly, mesenchymal proliferation is not affected in mutant (Fig. 2C; 5.6% ± 1.4 vs. 5.4% ± 1.2, respectively). The decreased proliferation in the epithelium is associated with a lower number of P-ERK positive cells in the Fgf10 hypomorphic lung epithelium (47% and 38% decrease in the distal epithelium of Fgf10 hypomorph vs. control lung at E13.5 and E14.5 respectively, n=3, data not shown, P<0.05). At E17.5, we could no longer distinguish between epithelial and mesenchymal cells. A general reduction in cell proliferation in the mutant is observed (32.8% ± 8. vs. 21.6% ± 4.5; n=3 for each genotype; P<0.05). Wnt signaling is crucial during lung development to control the maintenance and proliferation of epithelial progenitors at least up to E14.5 (De Langhe et al., 2005; Mucenski et al., 2003). We therefore examined the status of canonical Wnt signaling in Fgf10+/−; Mlc1v-LacZ+/− lungs by genetically introducing the TOPGAL allele into these mice. Previous studies indicate that canonical Wnt signaling can be detected in the epithelium of the lung using the TOPGAL reporter mice (De Langhe et al., 2005; Okubo and Hogan, 2004; Shu et al., 2005), where β-galactosidase expression is a reporter for canonical Wnt signaling activation (DasGupta and Fuchs, 1999). Fig. 2E-J show that reduction in FGF10 signaling results in drastic down regulation of canonical Wnt signaling in the distal lung epithelium at E13.5 (n=3 for each genotype). β-catenin staining at E12.5 shows decreased nuclear β-catenin expression in the Fgf10 hypomorph vs. control lung (Fig. 2K-N). Altogether, these observations suggest that FGF10 in the mesenchyme is upstream of Wnt signaling in the epithelium.
Immunohistochemistry was used to determine potential epithelial differentiation defects in Fgf10+/−; Mlc1v-LacZ+/− vs. Fgf10+/− (control) lungs at birth. Antibodies against Surfactant proteins SP-A, SP-B and SP-C as well as the general lung epithelial marker TTF-1, were used. A general reduction in the number of epithelial cells positive for these markers was detected in mutant lungs (data not shown for SP-A and SP-C). In particular, the percentile of cells positive for Surfactant Protein B was reduced by 40% in mutant vs. control (Fgf10+/−) lungs (Fig. 3A-C; 12.0% ± 0.2 vs. 18.0% ± 0.2, respectively, P<0.001). Interestingly, it has been reported that a mutation in the SP-B gene in human is neonatal lethal (Nogee et al., 1994). Therefore, the reduction in Surfactant Protein B expression could explain some of the respiratory failure observed at birth. However, we cannot conclude that epithelial differentiation per se is affected as we also observe a comparable reduction in the percentile of cells positive for the general epithelial marker TTF-1 in the mutant vs. control lung (Fig. 3D-F; 25.0% ± 1.0 vs. 42.0% ± 5.0, respectively; n=3 for each genotype, P<0.01). Interestingly the ratio SP-B:TTF1 in control and mutant lungs is also similar (43% vs. 48%, respectively) suggesting that the reduction in the percentile of cells positive for SP-B and TTF-1 in the mutant lung reflects a general decrease in epithelial proliferation as observed at earlier stages (Fig. 2). Similar observation has been made with the ratio SP-C: TTF1 (60% vs. 55% in control vs. mutant lungs).
Histological analysis of the control and Fgf10+/−; Mlc1v-LacZ+/− lungs at E13.5 and E14.5 showed no abnormalities in the proximal-distal patterning of the lung epithelium. In both control and transgenic lungs, the bronchiolar proximal epithelium was columnar, whereas the distal respiratory epithelium was low cubical (data not shown). At E17.5, seven out of ten lungs from Fgf10+/−; Mlc1v-LacZ+/− embryos displayed overly expanded distal airways with readily apparent defects in the septation process (compare Fig. 4A and B). Similar defects were found at postnatal stages (Fig. 4D-F). These changes in peripheral pulmonary alveolar size were quantified at E17.5 and P2 by morphometric measurement of mean linear intercepts (MLI) and radio-alveolar counts (RAC) (Chen et al., 2005; Fig. 4H, I). Statistically significant increased MLI and associated decreased RAC in mutant vs. control (P<0.001 for the two stages examined) indicate dilation of the distal airways and failure to undergo secondary septation. These structural defects in the respiratory airways were accompanied by early lethality. Out of 12 pups analyzed after birth, 9 (80%) died within 24 hours, 3 died within 48 hours (Fig. 4G). All of them exhibited severe gasping indicating respiratory distress. Interestingly, the 3 pups that survived more than 1 day after birth did not grow (data not shown) and died 48 hours after birth probably as a consequence of respiratory failure as well as limb abnormalities (Veltmaat et al., 2006 and data not shown) and gut defects (Sala et al., 2006).
Immunofluorescence for α-smooth muscle actin (α-SMA) was used to detect alveolar smooth muscle cells at E18.5 in the respiratory airway. Fig. 5A shows the expression of α-SMA in E18.5 control lungs. SMA expression was severely reduced in corresponding Fgf10+/−; Mlc1v-LacZ+/− lungs indicating defective alveolar smooth muscle cell formation (n=3 for each genotype). In addition, in situ zymogram assays using fluorescein isothyocyanate labeled gelatin to detect gelatinase activity (corresponding to MMP2 and MMP9) was also performed. In this assay, the gelatin with a fluorescent tag does not fluoresce unless it is cleaved and the reaction product is visualized using fluorescence microscopy. Smooth muscles in control lung exhibit a high level of MMP activity (Fig. 5C) while a clear reduction in gelatinase activity is observed in the Fgf10+/−; Mlc1v-LacZ+/− lung (Fig. 5D). In agreement with reduced formation of smooth muscle cells in Fgf10+/−; Mlc1v-LacZ+/− lung, Elastin deposition was also decreased in the mutant lung compared to the control (data not shown). Thus the results indicate that there is a defect in the formation of alveolar smooth muscle cells in Fgf10+/−; Mlc1v-LacZ+/− lungs.
We have investigated potential defects in the development of the vascular system in Fgf10 hypomorphic lungs using light microscopy, immunofluorescence and vascular casts. Fig. 6A and B show marked reduction in the number of red blood cells present in the E13.5 Fgf10+/−; Mlc1v-LacZ+/− vs. Fgf10+/− (control) lungs. A general decrease in the expression of PECAM, a marker for endothelial cells, was observed in E14.5 Fgf10+/−; Mlc1v-LacZ+/− vs. control lungs (n=3 for each genotype). Antibodies against the extracellular matrix protein Laminin were also used to visualize the status of the microvasculature in the lung mesenchyme during the pseudoglandular stage (Fig. 6E). Formation of the microcapillaries in Fgf10+/−; Mlc1v-LacZ+/− lung was impaired (Fig. 6F). Finally, vascular corrosion casts of the control and mutant lungs at E18.5 indicate a clear simplification of the vascular tree (Fig. 6G-g’; n=2 for control and Fgf10+/−; Mlc1v-LacZ+/−). At P2, all of the lungs examined exhibited large hemorrhagic areas (compare Fig. 6H and I). Such defects were already observed at E17.5 (Fig. 1H).
We performed real time PCR quantification of mRNA expressed in the Fgf10+/−; Mlc1v-LacZ+/− vs. Fgf10+/− lungs. Since the Fgf10+/−; Mlc1v-LacZ+/− lungs exhibited defects in vascular development and alveolar smooth muscle cell formation, expression levels of related genes were determined. Fig. 7 shows representative real-time PCR results for the relative expression of Fgf10, Vegfa, Pdgfa, and Pdgfb in Fgf10+/−; Mlc1v-LacZ+/− vs. Fgf10+/− lungs. Consistent with previously reported results, we observe a 27% decrease in Fgf10 expression (Mailleux et al., 2005). Vegfa is expressed by the distal lung epithelium and acts on the adjacent endothelial cells, expressing VEGF-R to stimulate their proliferation (Del Moral et al., 2006).Vegfa expression was reduced by 28% in the Fgf10 hypomorphic lungs compared to control lungs. Pdgfa and Pdgfb are also expressed by the distal lung epithelium. These growth factors act on the alveolar smooth muscle cell progenitors present in the mesenchyme to induce their proliferation (Bostrom et al., 1996; 2002). Pdgfa and Pdgfb expression were reduced about 60% and 16%, respectively. The expression of Hif1a and Glut1, as indicators of potential hypoxic conditions in lung tissues (Mobasheri et al., 2005), was not significantly changed between Fgf10+/− and Fgf10+/−; Mlc1v-LacZ+/− at E14.5. These results support the conclusion that changes in gene expression in the Fgf10+/−; Mlc1v-LacZ+/− lungs are solely the consequence of decreasing Fgf10 expression and not due to hypoxia.
Fgf10−/− embryos display lung agenesis precluding functional analysis of Fgf10 during subsequent lung development. Here, we show that, in addition to its use as a reporter for Fgf10 expression, the Mlc1v-LacZ transgenic line also represents a hypomorphic allele of Fgf10. This allowed us to determine the in vivo role of Fgf10 during the later phases of lung development. Newborn Fgf10+/−; Mlc1v-LacZ+/− compound heterozygote mice exhibit severe respiratory defects and most die within 24 hours of birth. During embryogenesis, the Fgf10+/−; Mlc1v-LacZ+/− lungs display significant hypoplasia, associated with dilated distal airways. Analysis of vascular development shows a reduction in PECAM expression at E14.5. This reduction is associated with simplification of the vascular tree at E18.5. E18.5 mutant lungs also display reduced expression of α-SMA in the respiratory airways indicating abnormal alveolar smooth muscle cell formation. These defects are associated with a decrease in Vegfa and Pdgfa expression, both expressed in the epithelium.
The notion that appropriate levels of Fgf10 are needed for normal lung development to occur has been unclear so far. Our results indicate that the lungs of Fgf10+/− (control) mice are phenotypically normal. However, if Fgf10 levels decrease by 30% compared to heterozygous levels, distal lung hypoplasia occurs. This is a likely consequence of significant reduction in epithelial proliferation in the Fgf10 hypomorph vs. control lung. Evidence is accumulating that Fgf10 dosage is critical for proper development of certain organs while others remain unaffected. The importance of Fgf10 dosage for the development of salivary gland and lacrimal gland in humans has been shown (Entesarian et al., 2005). Individuals with autosomal dominant aplasia of lacrimal and salivary glands (ALSG) exhibit hypoplastic or missing parotid and submandibular glands. ALSG was mapped to 5p13.2-5q13.1, which includes the FGF10 gene. Subsequently, all family members with ALSG were found to be heterozygous for Fgf10. Complementary studies in adult Fgf10+/− mice revealed that Fgf10 heterozygotes have no parotid and smaller submandibular glands (Jaskoll et al., 2005). Our recent results further demonstrate that the morphogenesis of the mammary glands and distal region of the colon are also Fgf10 dose-dependent (Veltmaat et al., 2006; Sala et al., 2006).
Our results show that Fgf10+/−; Mlc1v-LacZ+/− pups exhibit severe lung hypoplasia as a result of reduced Fgf10 expression. The Fgf10+/−; Mlc1v-LacZ+/− lungs displayed decreased numbers of SP-A, SP-B and SP-C expressing cells at birth, suggesting that the development of the type-II alveolar cell lineage may be affected. However, as the ratio SPB: TTF1 (TTF1 being a general marker of lung epithelial cells) between Fgf10+/− and Fgf10+/−; Mlc1v-LacZ+/− lung is similar, it is unlikely that the decrease in surfactant protein (SP)-positive cells reflects a defect in epithelial differentiation. Our results suggest that the decrease in epithelial cell proliferation during the pseudoglandular stage is the underlying cause of the reduction in SP-positive cells.
Many in vitro studies imply FGFs as signals of proliferation and differentiation (for reviews see Cardoso, 2001; Cardoso and Lu 2006; Warburton et al., 2000). In vivo, the loss of function studies of Fgfr2b or Fgf10 shows clearly that this signaling pathway is vital in the development of the bronchial tree structure (Arman et al., 1999; De Moerlooze et al., 2000; Min et al., 1998; Sekine et al., 1999). Nevertheless, the phenotypes of the null mutants make it impossible to establish the functional role of FGF10 and FGFR2b in late stages of pulmonary development. In the case of Fgfr2b, the inducible over-expression of the soluble dominant negative FGFR2b receptor from the end of the pseudoglandular phase (E14.5) disturbs alveogenesis after birth. By contrast, a similar inhibition of FGFR2b signals after birth does not lead to a defect in alveolar formation suggesting that FGFR2b signaling is not needed for lung homeostasis (Hokuto et al., 2003). The authors reported that FGFR2b signaling is indeed critical during the early pseudoglandular stage. Decrease in FGFR2b signaling during this early developmental phase impacts negatively alveogenesis. Interestingly, other FGFR2b ligands (FGF1, 3, 7) are expressed from E13.5 onwards in lung, either in the epithelium or the mesenchyme (Bellusci et al., 1997; Cardoso, 2001; Lebeche et al., 1999). There are therefore many FGFR2b ligands that could potentially compensate for the loss of Fgf10. However, Fgf10 hypomorphic mice have alveolar defects, as the likely consequence of events taking place during early embryonic development, when Fgf10 is the only functional FGFR2b ligand expressed in the lung. Our results therefore demonstrate that FGF10 plays a unique function in the E9.5-E13.5 pseudoglandular period specifically in the amplification of pulmonary epithelial progenitors.
A similar emphysematous-like, non-inflammatory lung phenotype and postnatal death has been described for the Elastin and Mmp-2 null mice (Kheradmand et al., 2002; Wendel et al., 2000). In 15% of Mmp-2−/− newborns, a syndrome of acute respiratory distress is associated with improper deposition of elastin fibers (Kheradmand et al., 2002). Interestingly, a specific decrease in gelatinase activity at birth in Fgf10+/−; Mlc1v-LacZ+/− lung mesenchyme was revealed by in situ zymography and is associated with diminished SMA expression and Elastin deposition. These data suggest that reduced Fgf10 expression disturbs the formation of the smooth muscle cells in pre-alveolar structures, possibly reflecting hypoplasia of the alveolar SMC progenitors. This conclusion is supported by our observation that Pdgfa expression is strongly decreased in Fgf10+/−; Mlc1v-LacZ+/− lungs. Interestingly, Pdgfa has also been shown to be downstream of canonical Wnt signaling in the lung epithelium (De Langhe et al., 2005). In addition, β-catenin has been shown to be a downstream target of FGF10 in the lung epithelium (Lu et al., 2005). This is in harmony with the results of our recent Affymetrix arrays showing a 13% decrease in β-catenin expression in Fgf10 hypomorph vs. control lung at E14.5 (data not shown). However, it is unlikely that decreased transcription of β-catenin per se is enough to explain the reduction in Wnt signaling observed in the Fgf10 hypomorphic lungs (Fig. 2). Our results show indeed a decrease in nuclear β-catenin localization in the epithelium of Fgf10 hypomorph vs. control lungs (Fig. 2). It has recently been shown that the well-known PI3K/AKT pathway is capable, via the phosphorylation of β-catenin on serine 522, to increase the amount of nuclear β-catenin (He et al., 2007). It is therefore likely that in the lung epithelium, FGF10 controls β-catenin signaling in such a way rather than at the level of the transcription of β-catenin gene. More work will have to be done to determine the exact mechanism of action of FGF10 on the Wnt/β-catenin signaling. Our result that TOPGAL expression is decreased in Fgf10+/−; Mlc1v-LacZ+/− lungs, therefore strongly suggest that FGF10, upstream of β-catenin signaling, controls Pdgfa expression in the lung epithelium.
Decreased PECAM and Laminin expression in E14.5 Fgf10 hypomorphic lung suggest that the simplified vascular tree observed later on, at E18.5, is likely the consequence of abnormal vascular development starting early on during the pseudoglandular stage. Interestingly, we have demonstrated that lung development during this pseudoglandular stage is under the unique control of Fgf10. Our results therefore suggest a link between mesenchymal FGF10, which acts directly on the epithelium and the development of the vascular system. This observation has important implications as epithelial morphogenesis, mostly mediated by FGF10, has to be tightly controlled with vascular development to obtain a functional lung at birth. Our results indicate that there is a 28% decrease in Vegfa expression in the Fgf10 hypomorph vs. control lung. Interestingly, a similar decrease in TTF1 expression is observed in the mutant lungs indicating that the decrease in Vegfa is likely due to the reduction in the number of epithelial cells and not because FGF10 controls directly, at the transcriptional level, Vegf expression in the epithelium.
In conclusion, our results indicate that FGF10 plays a crucial role throughout the first half of the pseudoglandular stage in maintaining the activation of the Wnt canonical pathway in epithelial progenitor cells and allow their proliferation. In addition, we have demonstrated that FGF10 is essential in coordinating alveolar smooth muscle cell formation and vascular development.
This work was funded by ALA and an NIH RO1 HL074832 (to SB), HL60231, 75773, 44060, 44977 (to DW), HL056590 and 073471 (to PM), CBCRP and CHLA institutional award (to JV); ARC and FRM (to AA), CHLA institutional award (to PdM), ALA and CHLA institutional award (to SDL). We would like to thank Dennis Mock for his technical help with the analysis of the Affymetrix arrays.
Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.