Conditional deletion of WT1 in the septum transversum mesenchyme causes congenital diaphragmatic hernia in mice

*For correspondence: chapuli@ uma.es Competing interests: The authors declare that no competing interests exist. Funding: See page 14 Received: 15 March 2016 Accepted: 08 September 2016 Published: 19 September 2016 Reviewing editor: Margaret Buckingham, Institut Pasteur, France Copyright Carmona et al. This article is distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use and redistribution provided that the original author and source are credited. Conditional deletion of WT1 in the septum transversum mesenchyme causes congenital diaphragmatic hernia in mice Rita Carmona 1,2 , Ana Can ˜ete 1,2 , Elena Cano 3 , Laura Ariza 1,2 , Anabel Rojas 4,5 , Ramon Mun ˜oz-Cha ´puli 1,2 * 1 Department of Animal Biology, University of Ma´laga, Ma´laga, Spain; 2 Andalusian Center for Nanomedicine and Biotechnology (BIONAND), Ma´laga, Spain; 3 Max Delbru€ck Center for Molecular Medicine, Berlin, Germany; 4 Andalusian Center of Molecular Biology and Regenerative Medicine (CABIMER), Sevilla, Spain; 5 Centro de Investigacio´ n Biome´ dica en Red de Diabetes y Enfermedades Metabo´ licas Asociadas (CIBERDEM), Sevilla, Spain Abstract Congenital diaphragmatic hernia (CDH) is a severe birth defect. Wt1-null mouse embryos develop CDH but the mechanisms regulated by WT1 are unknown. We have generated a murine model with conditional deletion of WT1 in the lateral plate mesoderm, using the G2 enhancer of the Gata4 gene as a driver. 80% of G2-Gata4 Cre ;Wt1 fl/fl embryos developed typical Bochdalek-type CDH. We show that the posthepatic mesenchymal plate coelomic epithelium gives rise to a mesenchyme that populates the pleuroperitoneal folds isolating the pleural cavities before the migration of the somitic myoblasts. This process fails when Wt1 is deleted from this area. Mutant embryos show Raldh2 downregulation in the lateral mesoderm, but not in the intermediate mesoderm. The mutant phenotype was partially rescued by retinoic acid treatment of the pregnant females. Replacement of intermediate by lateral mesoderm recapitulates the evolutionary origin of the diaphragm in mammals. CDH might thus be viewed as an evolutionary atavism. DOI: 10.7554/eLife.16009.001 Introduction Congenital diaphragmatic hernia (CDH) is a severe birth defect, characterized by incomplete formation or muscularization of the diaphragm and, as a consequence, herniation of the stomach, intestines, liver or spleen into the pulmonary cavities, leading to pulmonary hypoplasia. CDH occurs approximately in 1 out of 3000 births, accounting for about 8% of all severe congenital anomalies. 80–90% of all the cases are posterolateral hernias, also known as Bochdalek-type CDH, characterized by a defect in the postero (dorsal in mice) lateral area of the diaphragm. In most cases (>85%), this defect is located at the left side (Pober et al., 2010). The classical hypothesis is that Bochdalek CDH is due to a failure of the fusion of pleuroperitoneal folds (PPFs) with the septum transversum (ST). PPFs are lateral rims of tissue connecting caudally with the nephric ridges and anteriorly with the ST (Mayer et al., 2011). Some authors describe that this fusion rather occurs with the posthepatic mesenchymal plate (PHMP), an accumulation of mesenchymal cells derived from the ST and located in the posterodorsal margin of the liver lobes (Iritani et al., 1984). Although all these tissues form an anatomical continuum, we will refer to the posterior, dorsolateral areas of the liver as PHMP, and PPFs to the tissue folds located in the dorsal part of the coelomic cavity, marking the limit between the peritoneal and the pleural cavities. The clinical aspects of CDH are well studied but its etiology is still poorly known (Klaassens et al., 2006). Animal models are therefore very valuable in order to investigate the Carmona et al. eLife 2016;5:e16009. DOI: 10.7554/eLife.16009 1 of 17 RESEARCH ARTICLE cellular and molecular processes leading to CDH in humans. Recently, it has been described in mice that deletion of GATA4 in the PPFs mesenchyme using a Prrx1 Cre driver leads to the development of amuscular, weak areas in the diaphragm and CDH (Merrell et al., 2015). This paper assumes that Prrx1 is expressed in the mesenchyme of the pleuroperitoneal folds (PPFs) but not in the ST, and in fact, it suggests that the ST contributes minimally to the definitive diaphragm. However, this report does not distinguish between PPFs and PHMP and the phenotype does not display diaphragmatic discontinuities, neither does it show a prevalence on the left side, which are common features in human CDH. The retinoic acid (RA) signaling pathway has also been involved in diaphragmatic development. A classical animal model for CDH consists in the treatment of pregnant rats with nitrofen, a substance that inhibits the synthesis of RA. It has been shown that nitrofen treatment leads to reduced size of the PPFs/PHMP, especially in the left side, decreasing cell proliferation in this area (Clugston and Green, 2007;Clugston et al., 2010) and also decreasing expression of WT1 and GATA4 (Dingeman et al., 2011,2013). RA treatment in this model partially rescues the pulmonary hypoplasia defect (Montedonico et al., 2008;Sugimoto et al., 2008). The phenotype of this animal model is thus, quite similar to the human one. To date, only a few examples of CDH in human have been associated to mutations in WT1 locus, such as the Denys-Drash, Meacham and WAGR syndromes (Scott et al., 2005;Suri et al., 2007; Antonius et al., 2008). WT1 is expressed in the coelomic epithelium of the septum transversum of mouse embryos by the stage E9.0. Loss of function of the Wilms’ tumor suppressor gene Wt1 in mice leads to defective diaphragms (Kreidberg et al., 1993;Clugston et al., 2006). However, the mechanism by which WT1 contributes to diaphragm development is still unknown. To study the role of WT1 in CDH we performed loss-of-function experiments by conditionally inactivating the Wt1 gene in the ST/PHMP/PPFs mesenchyme. The use of WT1 conditional knockout overcomes the early embryonic death caused by systemic deficiency of WT1. We used a driver based on the G2 enhancer of the Gata4 gene (Rojas et al., 2005). This enhancer drives expression of Gata4 in the lateral plate mesoderm from the stage E7.5, and by the stage E9.5 is active in the septum transversum and proepicardium, ceasing its activity by E12.5 The activity of this enhancer is completely absent in the intermediate mesoderm. Our findings indicate that WT1 is involved in the generation of the mesenchyme of the ST/PHMP/PPFs continuum through epithelial-mesenchymal transition and they provide a novel perspective on the genesis of the Bochdalek hernia and the evolutionary origin of the diaphragm. Results The septum transversum, posthepatic mesenchymal plate and pleuroperitoneal folds contain heterogeneous mesenchymal populations In normal E10.5 embryos, the posterior and dorsal margin of the liver shows an accumulation of mesenchymal tissue, which extends from the dorsal mesenterium of the liver to the lateral tips of the lobes (Figure 1). This mesenchymal layer is the PHMP described by Iritani (1984). The PPFs, by this developmental stage, appear as a pair of outgrowths of the body wall located at both sides of the lung buds. (Figure 1A,B). They are also constituted of mesenchymal cells lined by the coelomic epithelium. The G2-Gata4 enhancer directed LacZ reporter expression in both PPF and PHMP at E10.5 and E11.5 in the mouse embryo (Figure 1A). However, no beta galactosidase activity was observed in the posterior and medial part of the septum transversum, where the liver is connected with the digestive tract (asterisk in Figure 1A), indicating that the G2 enhancer is not active in this specific domain. PHMP and PPFs are not anatomically independent entities. The lateral tips of the PHMP appear continuous with the PPFs at the cephalic level while they are separated more caudally (Figure 1B). The continuous part of the PHMP/PPFs ensemble apparently forms itself by the anterior growth of the peritoneal recess, which separates the liver sides from the body wall. Thus, the mesenchymal population of the ST originates the PHMP and also the most cephalic portion of the PPFs. The lateral closure of the pleural cavities occurs between E11.5 and E12.5 by posterior growth of the crescent- Carmona et al. eLife 2016;5:e16009. DOI: 10.7554/eLife.16009 2 of 17 Research article Developmental Biology and Stem Cells Human Biology and Medicine Figure 1. Cell lineages in the septum transversum (ST), posthepatic mesenchymal plate (PHMP) and pleuroperitoneal folds (PPF). (A) G2-LacZ embryo at the stage E11.5. The G2 enhancer is active in the liver (LI) mesothelium and posthepatic mesenchymal plate (PHMP). However, the central area of the septum transversum shows no activity of this enhancer (asterisk). (B)Wt1 Cre ;R26R LacZ embryo, E10.5. Cells from the WT1 lineage appear in blue. Series of transverse sections from cephalic (left) to caudal (right) levels. The anterior peritoneal recess (PR) is splitting the liver from the body wall. The PHMP is continuous with the PPF at cephalic levels. Note the abundance of mesenchymal cells in the anterior PHMP and how this mesenchyme is less abundant in posterior levels. The PPF is continuous with the renal ridge (RR). The mesenchymal cells of all these structures belong to a WT1-expressing cell lineage. (C)Wt1 LacZ embryo, E11.5. Activation of the Wt1 reporter can be seen in the liver mesothelium. PHMP, PPF and RR. The RRs appear at the level of the stomach (STO). OE: oesophagus. (D)Wt1 Cre ;R26R YFP embryo, E11.5. Cells from the WT1-expressing cell lineage (green) are more abundant at Figure 1 continued on next page Carmona et al. eLife 2016;5:e16009. DOI: 10.7554/eLife.16009 3 of 17 Research article Developmental Biology and Stem Cells Human Biology and Medicine like continuum formed by the PHMP and the PPFs, a process which is parallel to the fast growth of the liver lobes during these stages (Figure 1B,C). As seen in the Wt1 Cre labeled cells, the PPFs continue posteriorly as the nephric ridges, the progenitor tissue of the mesonephros (Figure 1B and C). A loose mesenchymal tissue appears in the PPF between the more compact arrangement of the cells in the PHMP and in the renal ridges, but the commissural area of the PPF is always constituted by compact mesenchyme similar to that from the PHMP (arrow in Figure 2C). All these mesenchymal populations are derived from a WT1- expressing cell lineage, as shown by two models, the Wt1 Cre ;R26R LacZ (Figure 1B) and the Wt1 Cre ; R26R YFP (Figure 1D). Expression of the Wt1 gene is also revealed in the PPF and PHMP mesenchyme by the WT1-LacZ reporter and by immunohistochemistry (Figures 1C and 3A, respectively). Epithelial and mesenchymal cells of the PHMP and most anterior PPFs derive from a cell lineage in which Gata4 expression is driven by the G2 enhancer. This is demonstrated by the expression of YFP in G2-Gata4 Cre ;R26R YFP embryos, and supports the existence of a process of cell migration from PHMP to PPFs (Figure 1E,F). However, the nephric ridges and the mesonephros derived from them show no YFP+ cells in this model, indicating that the progenitors of these tissues derived from a different cell lineage. Thus, G2-Gata4 lineage tracing allows to establishing a well-defined limit between two lateral mesodermal territories that we will call herein the G2 + and the G2 - domains (see figure 3B,C). Interestingly, the gonad and the adrenal stroma were YFP+ (Figure 1G), and this can be related with the early expression of GATA4 in the coelomic epithelium, which will cover the genital but not the renal ridges (Figure 1H). Incidentally, the GATA4 positive cells from the central part of the ST do not express YFP in G2-Gata4 Cre ;R26R YFP embryos (Figure 1F), an observation consistent with the lack of reporter activity in the G2- Gata4 LacZ embryos (Figure 1A). These data reveal a significant heterogeneity between central and lateral mesenchymal populations in the ST. A clear asymmetry was observed in the distribution of PHMP mesenchyme in embryos of E10.5- E12.5 (Figure 1D,E). Mesenchymal cells were always more abundant in the right PHMP than in the left, which might favor the herniation in the left side. We estimated, by image analysis, the volume taken up by the YFP+ cells in the right and left PHMP of four G2-Gata4 Cre ; R26R YFP embryos. The R/ L ratio of these volumes was 1,21 and 1,45 in two E11.5 embryos, and 1,53 and 1,61 in two E12,5 embryos (mean = 1,45; S.E.M = 0,086). The t-test for one sample compared with an expected value = 1 (symmetrical distribution) gave a result of 5,21 (p-value=0,0069 for three d.o.f). Thus, the right PHMP contains about 50% more cells from the G2-Gata4 lineage than the left one. WT1 conditional knockout mice display Bochdalek’s hernia This is a basically descriptive study of the phenotype of mouse embryos with conditional deletion of Wt1 in lateral mesoderm. We have included in the study 36 out 104 mutant embryos detected by genotyping. The rest of them had been previously used for a study of the cardiac phenotype (Cano et al., 2016). To understand the role of WT1 positive cells in the development of diaphragm, we conditionally inactivated Wt1 gene using the described G2-Gata4 Cre line. Deletion of WT1 in the PHMP/PPFs continuum, i.e. in the G2 + domain, causes a phenotype characterized by defects in the inflow tract of the heart, coronary vessels (Cano et al., 2016) and, importantly, Bochdalek’s hernia, most frequently Figure 1 continued the right PHMP (R) as compared with the left one (L). Note the lack of YFP+ cells in the central area of the ST. AO: aorta; LU: lung. (E) G2-Gata4 Cre ; R26R YFP embryo, E11.5. Cells from the lineage expressing GATA4 under the control of the G2 enhancer are stained in green, and RALDH2 is stained in red. Most cells in the PHMP and PPFs belong to the G2-Gata4 lineage, but they are very scarce in the central area of the ST. Abundant YFP+ cells are present dorsally to the aorta, and around the notochord (NC). (F) G2-Gata4 Cre ;R26R YFP embryo, E11.5. GATA4 immunostaining in red. Colocalization of GATA4 and YFP is observed in the PHMP and PPF, but the central area of the ST shows GATA4 expression not driven by the G2 enhancer. (G) G2- Gata4 Cre ; R26R YFP embryo, E11.5. Smooth muscle alpha-actin is stained in red. Gonads (GON) and adrenals (AD) contain a large number of G2-Gata4 lineage cells, but they are scarcer into the mesonephros (MN). Data reused, with permission, from Figure 4A, Mun ˜oz-Cha ´puli et al. Developmental Dynamics, Special Issue: Mechanisms of Morphogenesis, 245:307–322 (2016). Ó2015 Wiley Periodicals, Inc. (H) Immunolocalization of GATA4 (red) in an E10.5 G2-Gata4 Cre ; R26R YFP embryo. The G2 lineage is shown in green. There is a sharp boundary (arrow) between the mesothelial cells expressing GATA4 of the genital ridge (GR), close to the dorsal mesentery (DM), and the mesothelial cells of the renal ridges (RR), which are GATA4-. AO: aorta; OE: oesophagus. DOI: 10.7554/eLife.16009.002 Carmona et al. eLife 2016;5:e16009. DOI: 10.7554/eLife.16009 4 of 17 Research article Developmental Biology and Stem Cells Human Biology and Medicine located in the left side (Figure 2). The genotypes of the embryos obtained at different ages show the expected percentage of mutants at all embryonic stages (about 25%) except for the stage E15.5, when we only obtained four mutants among 60 embryos (Table 1). The number of embryos studied in older stages (42 embryos, nine mutants, 21,4%), was not enough to confirm if the mutation causes some lethality in late gestation. We analyzed WT1 mutant embryos at different stages of development. Defect in the PHMP in G2-Gata4 Cre ;Wt1 fl/fl mutant embryos could be observed as early as E10.5, since less mesenchymal cells are present between the coelomic epithelium and the hepatoblasts when compared with control littermates (Figure 2A–D). At later stages, the closure of the pleural cavity by growth of the ST/ PPFs crescent is delayed with respect to the controls, and the mesenchymal cells of the PHMP are Figure 2. Phenotype of the G2-Gata4 Cre ; Wt1 fl/fl embryos. (A–D) Wildtype (A,C) and mutant (B,D), E10.5 littermates. The larger amount of mesenchymal cells in the posthepatic mesenchymal plate (PHMP) is evident in the wildtype, especially in the right side. Serial sections corresponding to the connection between the PHMP and the pleuroperitoneal folds (PPF) are shown in C. Note the presence of compact mesenchyme in the PHMP and also in the closest part of the PPF (arrow in C). A corresponding section of the mutant in shown in D. Note the lack of PHMP, the coelomic epithelium lying directly on the hepatic tissue (arrows in D) and the limit of the septum transversum (ST) mesenchymal cells (arrowhead in D), which do not extend laterally. (E) Serial sections of a G2-Gata4 Cre ; Wt1 fl/fl E11.5 embryo at levels equivalent to those shown in Figure 1C. Despite the presence of normal PHMP in the anterior part, the posterior areas of the liver lack of lateral mesenchymal cells (arrow). The mesenchyme is restricted to the central ST. Renal ridges (RR) appear at a level corresponding to the entrance of the oesophagus (OE) into the ST. MG: mesogastrium. (F) Wildtype E13.5 embryo showing complete isolation of the pleural cavities by the pleuroperitoneal membranes that constitute the main part of the diaphragm (D). (G) G2- Gata4 Cre ; Wt1 fl/fl E14.5 embryo at eight different levels showing left diaphragmatic defect with herniation of the left liver lobe (LI) into the pleural cavity and severe hypoplasia of the left lung (LL). Ectopic muscle appear in the mediastinum (MM). Adrenals (AD), mesonephros (MN), gonads (GON) and spleen (SP) appear normal. LU: lungs; H: heart; RL: right lung; STO: stomach. DOI: 10.7554/eLife.16009.003 Carmona et al. eLife 2016;5:e16009. DOI: 10.7554/eLife.16009 5 of 17 Research article Developmental Biology and Stem Cells Human Biology and Medicine located more medially, not in the lateral tips of the liver lobes. This is more evident at the left side (arrow in Figure 2E). Thus, these cells cannot migrate towards the PPFs precluding caudal growth of the lateral commissures of the pleuroperitoneal septa and closure of the pleural cavities. Five E12.5 mutant embryos studied already showed strong reduction of the PHMP size as compared with control littermates, but the diaphragmatic defect of the WT1 conditional knockout embryos became clearly apparent by E13.5, when the pleural cavities are almost completely isolated in wildtype embryos (Figure 2F). Of 23 mutant embryos analyzed by the stages E13.5 or older, only four of them showed normal development of the diaphragm, while the rest of embryos showed different degrees of diaphragmatic defects (Table 1). Three of the embryos showed thin, membranous diaphragm in the left side and 16 showed a wide defect in the diaphragm, frequently with herniation of the liver into the left pleural cavity and hypoplasia of the left lung (Figure 2G). One of them (E13.5) showed the defect only at the right side and the CDH was present in the left side in the rest of them. Thus, about 80% of the E13.5 or older mutant embryos showed abnormal diaphragmatic development. Renal ridges appear well developed in mutant E11.5 embryos at a level corresponding to the entrance of the oesophagus into the ST (Figure 2E). The presence of intermediate mesoderm in these thoracic renal ridges was confirmed by ISH of Pax2 (Figure 4). In wildtype embryos of the same age thoracic renal ridges can also be seen, but located more caudally, at the level of the stomach (Figure 1C). Figure 3. WT1, GATA4 and RALDH2 expression in G2-Gata4 Cre ; Wt1 fl/fl embryos. (A,B) Immunolocalization of WT1 (A) and GATA4 (B) in wildtype E11.5 embryos. The posthepatic mesenchymal plate (PHMP) and pleuroperitoneal folds (PPF) show abundant positive cells. However, the WT1 immunoreactive cells are present in the base of the PPF and the adjacent body wall, where GATA4+ cells are absent. This difference defines the G2+ and the G2- domains. Data in A reused, with permission, from Figure 3E, Mun ˜oz-Cha ´puli et al. Developmental Dynamics, Special Issue: Mechanisms of Morphogenesis, 245:307–322 (2016). Ó2015 Wiley Periodicals, Inc. (C–D) Immunolocalization of WT1 (C) and GATA4 (D) in G2-Gata4 Cre ; Wt1 fl/fl E11.5 embryos. The PHMP and the G2+ domain of the PPF show GATA4+ cells but no WT1+ cells due to the conditional deletion of this gene in the G2+ domain. WT1 expression remains in the G2- domain of the base of the PPF and body wall. GATA4 expression is normal in most cephalic PPF, but immunoreactivity disappears in more caudal areas of PPF, as shown in D’ and D’’. (E–H) Immunolocalization of RALDH2 in E10.5 (E,F) and E11.5 (G,H) wildtype and G2-Gata4 Cre ; Wt1 fl/fl mutant embryos. RALDH2 immunoreactivity is high in PPF and the adjacent body walls. Note strong immunoreactivity in the renal ridges (RR) of the E11.5 mutant embryo. However, RALDH2 immunoreactivity is reduced or absent in the PHMP and liver mesothelium of the mutant embryos, as compared with the controls. CV: cardinal veins; LU: lungs. DOI: 10.7554/eLife.16009.004 Carmona et al. eLife 2016;5:e16009. DOI: 10.7554/eLife.16009 6 of 17 Research article Developmental Biology and Stem Cells Human Biology and Medicine The sharp boundary between the above explained G2 + and G2  domains was clearly observed by the immunolocalization of WT1 and GATA4 protein in G2-Gata4 Cre ;Wt1 fl/fl mutant embryos (Figure 3). WT1 immunoreactivity is strong in the PHMP and PPF of the wildtype embryos and localizes in the coelomic epithelium and, with a lower intensity, in mesenchymal cells (Figure 3A). However, WT1 is clearly downregulated in the G2 + domain of the mutant embryos indicating an efficient Table 1. Number of embryos studied and % of G2-Gata4 Cre ;Wt1 fl/fl found. The five E12.5 embryos showed abnormal posthepatic mesenchymal plates at the left side, but they were not computed for the diagnosis of the diaphragmatic hernia, that was performed only on the 23 embryos at the stages E13.5 or older. Stage Number of embryos studied Number and% G2Cre+; Wt1fl/fl Embryos analyzed for phenotype No defect Mild defect Diaphragmatic hernia E10–10.5 42 11 (26,2%) 3 E11.5 89 21 (23,6%) 5 E12.5 76 21 (27,6%) 5 (5) E13,5 65 19 (27,1%) 13 1 1 11* E14,5 53 17 (32,1%) 4 2 2 E15,5 60 4 (6,7%) 4 2 2 E16,5 21 4 (19,0%) 1 1 E17,5– 18,5 21 5 (23,4%) 1 1 Total 427 102 (23,6%) 36 4 (17,4%) 3 (13,0%) 16 (69,6%) *We have included seven embryos from females fed with control diet in the RA-rescue experiment. One embryo showed CDH only at the right side. DOI: 10.7554/eLife.16009.005 Figure 4. Localization of Pax2 expression by in situ hybridization. (A) Control E11.5 embryo. Pax2 is expressed in the neural tube (NT). The pleuroperitoneal folds (PPF) lack of Pax2 expression. (B) G2-Gata4 Cre ; Wt1 fl/fl E11.5 embryo. Pax2 is expressed in the PPF at the level of the lung buds (LU). The expression is stronger in the tubular structures that are developing in the left PPF (insert). DOI: 10.7554/eLife.16009.006 Carmona et al. eLife 2016;5:e16009. DOI: 10.7554/eLife.16009 7 of 17 Research article Developmental Biology and Stem Cells Human Biology and Medicine excision of the WT1 floxed allele (Figure 3C). Immunolocalization of GATA4 protein marked the border between both G2 domains, confirming the absence of endogenous GATA4 protein and the inactivity of the G2 enhancer in the G2 - domain. (Figure 3B,D). The presence of GATA4+ cells in the more cephalic part of the PPF supports a migration of PHMP cells towards the PPFs. Since WT1 is involved in epithelial-mesenchymal transition (EMT) of the epicardium by repression of E-cadherin and activation of Snail1 (Martı´nez-Estrada et al., 2010), we checked the expression of E-cadherin in the developing PHMP of G2-Gata4 Cre ;Wt1 fl/fl mutant embryos. Control E11.5 embryos show an E-cadherin negative coelomic epithelium over a layer of mesenchymal cells in the developing PHMP (Figure 5A–D). However, mutant embryos show expression of E-cadherin in the epithelium of the corresponding area and lack of E-cadherin negative mesenchymal cells (Figure 5E–H). Systemic WT1 loss of function leads to complete lack of PHMP We have compared the G2-Gata4 Cre ;Wt1 fl/fl CDH phenotype with that displayed by the systemic WT1 knockout mouse embryos, which do not survive beyond E13.5. Despite the early embryonic death of these models, we have observed a more severe defect, with a complete lack of PHMP and displaying, in some cases, persistent thoracic nephric ridges in E12.5 and E13.5 embryos (Figure 6). In these embryos only the central part of the ST is conserved, i.e. the area where the G2 enhancer is not active, as shown in Figure 1E. Retinoic acid partially rescues the defects in the diaphragm of WT1 conditional knockout embryos The retinoic acid (RA) signaling pathway has been involved in the formation of diaphragm. To test whether RA signaling might be disturbed in our CDH model, we checked the expression of RALDH2 by immunofluorescence in G2-Gata4 Cre ; Wt1 fl/fl mutant and control embryos. RALDH2 expression is found in the PHMP and PPF of wildtype E10.5-E11.5 embryos (Figure 3E,G). The PPF of the G2- Figure 5. Immunolocalization of E-cadherin in the left posthepatic mesenchymal plate (PHMP) of four E11.5 control embryos (A–D) and four E11.5 G2- Gata4 Cre ; Wt1 fl/fl embryos (E–H). E-cadherin appears upregulated in the coelomic epithelium of the PHMP of the mutant embryos (arrows in E–H). In wildtype embryos the coelomic epithelium of the PHMP lacks of E-cadherin expression (arrows in A–D) and covers a layer of E-cadherin-negative mesenchymal cells (asterisks). DOI: 10.7554/eLife.16009.007 Carmona et al. eLife 2016;5:e16009. DOI: 10.7554/eLife.16009 8 of 17 Research article Developmental Biology and Stem Cells Human Biology and Medicine Gata4 Cre ; Wt1 fl/fl mutant embryos show an increased immunoreactivity for RALDH2 but we found a reduced or absent expression of this enzyme in the PHMP (Figure 3F,H). This probably reflects, as discussed below, the persistence of the intermediate mesoderm of the nephric ridges into the PPF. In fact, RALDH2 is normally expressed at high levels in the intermediate mesoderm even in systemic WT1 knockout embryos, which also show decreased RALDH2 immunoreactivity in the liver mesothelium (Figure 6C,D). RA treatment partially rescues the pulmonary phenotype and promotes alveologenesis in the nitrofen model of induced CDH in rats (Montedonico et al., 2008;Sugimoto et al., 2008). In order to check the effects of a RA treatment on our mutant phenotype, in a first experiment we fed pregnant females with a RA-supplemented diet. For ethical reasons we decided to limit the study to four litters in which we detected eight G2-Gata4 Cre ; Wt1 fl/fl mutant embryos. Five of them (62,5%) displayed a complete septation of the pleural cavity by E15.5, i.e. about three times the proportion of normal phenotypes found in the non-treated mutants (Z test = 2,09, two-tailed p-value= 0,037). Control littermates were not affected by the RA-treatment. Figure 6. Diaphragmatic phenotype in systemic WT1 loss of function. (A,B) WT1-/- E13.5 embryos. The complete lack of posthepatic mesenchymal plate is evident, mesenchymal cells are absent from the lateral ends of the liver lobes and remain restricted to the central area of the septum transversum (ST), close to the oesophagus (OE). The pleural cavities are continuous with the peritoneal cavity. Note the renal ridges (RR) persisting at thoracic level, and the abnormal adhesion of the right liver lobe with the body wall (arrow in B). AO: aorta; LI: liver; LU: lungs. (C,D) Immunolocalization of RALDH2 in WT1-/- embryos, E11.5 (A) and E12.5 (B). RALDH2 immunoreactivity is strong in the renal ridges (RR), but weak or absent in the liver (LI) mesothelium (MES). AO: aorta; LU: lung. DOI: 10.7554/eLife.16009.008 Carmona et al. eLife 2016;5:e16009. DOI: 10.7554/eLife.16009 9 of 17 Research article Developmental Biology and Stem Cells Human Biology and Medicine Delgado I, Carrasco M, Cano E, Carmona R, Garcı ´a-Carbonero R, Marı ´n-Go ´ mez LM, Soria B, Martı ´n F, Cano DA, Mun ˜oz-Cha´puli R, Rojas A. 2014. GATA4 loss in the septum transversum mesenchyme promotes liver fibrosis in mice. Hepatology 59:2358–2370. doi: 10.1002/hep.27005 Dingemann J, Doi T, Gosemann J-H, Ruttenstock EM, Nakazawa N, Puri P. 2013. Decreased expression of GATA4 in the diaphragm of nitrofen-induced congenital diaphragmatic hernia. Birth Defects Research Part B: Developmental and Reproductive Toxicology 98:139–143. doi: 10.1002/bdrb.21037 Dingemann J, Doi T, Ruttenstock E, Puri P. 2011. Expression of the Wilm’s tumor gene WT1 during diaphragmatic development in the nitrofen model for congenital diaphragmatic hernia. Pediatric Surgery International 27:159–163. doi: 10.1007/s00383-010-2795-y Dressler GR, Deutsch U, Chowdhury K, Nornes HO, Gruss P. 1990. Pax2, a new murine paired-box-containing gene and its expression in the developing excretory system. Development 109:787–795. Essafi A, Webb A, Berry RL, Slight J, Burn SF, Spraggon L, Velecela V, Martinez-Estrada OM, Wiltshire JH, Roberts SG, Brownstein D, Davies JA, Hastie ND, Hohenstein P. 2011. A wt1-controlled chromatin switching mechanism underpins tissue-specific wnt4 activation and repression. Developmental Cell 21:559–574. doi: 10. 1016/j.devcel.2011.07.014 Goodrich ES. 1930. Studies on the Structure and Development of Vertebrates. London: Macmillan. Guadix JA, Carmona R, Mun ˜oz-Cha´puli R, Pe´ rez-Pomares JM. 2006. In vivo and in vitro analysis of the vasculogenic potential of avian proepicardial and epicardial cells. Developmental Dynamics 235:1014–1026. doi: 10.1002/dvdy.20685 Hayashi S, Fukuzawa Y, Rodrı ´guez-Va ´zquez JF, Cho BH, Verdugo-Lo´ pez S, Murakami G, Nakano T. 2011. Pleuroperitoneal canal closure and the fetal adrenal gland. The Anatomical Record: Advances in Integrative Anatomy and Evolutionary Biology 294:633–644. doi: 10.1002/ar.21351 Ijpenberg A, Pe´ rez-Pomares JM, Guadix JA, Carmona R, Portillo-Sa´nchez V, Macı ´as D, Hohenstein P, Miles CM, Hastie ND, Mun ˜oz-Cha´puli R. 2007. Wt1 and retinoic acid signaling are essential for stellate cell development and liver morphogenesis. Developmental Biology 312:157–170. doi: 10.1016/j.ydbio.2007.09.014 Iritani I. 1984. Experimental study on embryogenesis of congenital diaphragmatic hernia. Anatomy and Embryology 169:133–139. doi: 10.1007/BF00303142 Klaassens M, Scott DA, van Dooren M, Hochstenbach R, Eussen HJ, Cai WW, Galjaard RJ, Wouters C, Poot M, Laudy J, Lee B, Tibboel D, de Klein A. 2006. Congenital diaphragmatic hernia associated with duplication of 11q23-qter. American Journal of Medical Genetics Part A 140:1580–1586. doi: 10.1002/ajmg.a.31321 Kreidberg JA, Sariola H, Loring JM, Maeda M, Pelletier J, Housman D, Jaenisch R. 1993. WT-1 is required for early kidney development. Cell 74:679–691. doi: 10.1016/0092-8674(93)90515-R Kurpios NA, Iban ˜es M, Davis NM, Lui W, Katz T, Martin JF, Izpisu ´a Belmonte JC, Tabin CJ. 2008. The direction of gut looping is established by changes in the extracellular matrix and in cell:cell adhesion. PNAS 105:8499– 8506. doi: 10.1073/pnas.0803578105 Martı´nez-Estrada OM, Lettice LA, Essafi A, Guadix JA, Slight J, Velecela V, Hall E, Reichmann J, Devenney PS, Hohenstein P, Hosen N, Hill RE, Mun ˜oz-Chapuli R, Hastie ND. 2010. Wt1 is required for cardiovascular progenitor cell formation through transcriptional control of Snail and E-cadherin. Nature Genetics 42:89–93. doi: 10.1038/ng.494 Mayer S, Metzger R, Kluth D. 2011. The embryology of the diaphragm. Seminars in Pediatric Surgery 20:161– 169. doi: 10.1053/j.sempedsurg.2011.03.006 Merrell AJ, Ellis BJ, Fox ZD, Lawson JA, Weiss JA, Kardon G. 2015. Muscle connective tissue controls development of the diaphragm and is a source of congenital diaphragmatic hernias. Nature Genetics 47:496– 504. doi: 10.1038/ng.3250 Montedonico S, Nakazawa N, Puri P. 2008. Congenital diaphragmatic hernia and retinoids: searching for an etiology. Pediatric Surgery International 24:755–761. doi: 10.1007/s00383-008-2140-x Moore AW, Schedl A, McInnes L, Doyle M, Hecksher-Sorensen J, Hastie ND. 1998. YAC transgenic analysis reveals Wilms’ tumour 1 gene activity in the proliferating coelomic epithelium, developing diaphragm and limb. Mechanisms of Development 79:169–184. doi: 10.1016/S0925-4773(98)00188-9 Norden J, Grieskamp T, Christoffels VM, Moorman AF, Kispert A. 2012. Partial absence of pleuropericardial membranes in Tbx18- and Wt1-deficient mice. PLoS One 7:e45100. doi: 10.1371/journal.pone.0045100 Norden J, Grieskamp T, Lausch E, van Wijk B, van den Hoff MJ, Englert C, Petry M, Mommersteeg MT, Christoffels VM, Niederreither K, Kispert A. 2010. Wt1 and retinoic acid signaling in the subcoelomic mesenchyme control the development of the pleuropericardial membranes and the sinus horns. Circulation Research 106:1212–1220. doi: 10.1161/CIRCRESAHA.110.217455 Paris ND, Coles GL, Ackerman KG. 2015. Wt1 and b-catenin cooperatively regulate diaphragm development in the mouse. Developmental Biology 407:40–56. doi: 10.1016/j.ydbio.2015.08.009 Pober BR, Russell MK, Ackerman KG. 2010. Congenital diaphragmatic hernia overview. In: Pagon R. A, Adam M. P, Ardinger H. H, Wallace S. E, Amemiya A, Bean L. J. H, Bird T. D, Fong C. T, Mefford H. C, Smith R. J. H, Stephens K (Eds). GeneReviewsÒ[Internet]. University of Washington: Seattle. p 1993–2015. Preger-Ben Noon E, Barak H, Guttmann-Raviv N, Reshef R. 2009. Interplay between activin and Hox genes determines the formation of the kidney morphogenetic field. Development 136:1995–2004. doi: 10.1242/dev. 035592 Rojas A, De Val S, Heidt AB, Xu SM, Bristow J, Black BL. 2005. Gata4 expression in lateral mesoderm is downstream of BMP4 and is activated directly by Forkhead and GATA transcription factors through a distal enhancer element. Development 132:3405–3417. doi: 10.1242/dev.01913 Carmona et al. eLife 2016;5:e16009. DOI: 10.7554/eLife.16009 16 of 17 Research article Developmental Biology and Stem Cells Human Biology and Medicine Scott DA, Cooper ML, Stankiewicz P, Patel A, Potocki L, Cheung SW. 2005. Congenital diaphragmatic hernia in WAGR syndrome. American Journal of Medical Genetics Part A 134:430–433. doi: 10.1002/ajmg.a.30654 Sugimoto K, Takayasu H, Nakazawa N, Montedonico S, Puri P. 2008. Prenatal treatment with retinoic acid accelerates type 1 alveolar cell proliferation of the hypoplastic lung in the nitrofen model of congenital diaphragmatic hernia. Journal of Pediatric Surgery 43:367–372. doi: 10.1016/j.jpedsurg.2007.10.050 Suri M, Kelehan P, O’neill D, Vadeyar S, Grant J, Ahmed SF, Tolmie J, McCann E, Lam W, Smith S, Fitzpatrick D, Hastie ND, Reardon W. 2007. WT1 mutations in Meacham syndrome suggest a coelomic mesothelial origin of the cardiac and diaphragmatic malformations. American Journal of Medical Genetics Part A 143A:2312–2332. doi: 10.1002/ajmg.a.31924 Volpe MV, Wang KTW, Nielsen HC, Chinoy MR. 2008. Unique spatial and cellular expression patterns of Hoxa5, Hoxb4, and Hoxb6 proteins in normal developing murine lung are modified in pulmonary hypoplasia. Birth Defects Research Part A 82:571–584. doi: 10.1002/bdra.20481 von Gise A, Zhou B, Honor LB, Ma Q, Petryk A, Pu WT. 2011. WT1 regulates epicardial epithelial to mesenchymal transition through b-catenin and retinoic acid signaling pathways. Developmental Biology 356: 421–431. doi: 10.1016/j.ydbio.2011.05.668 Wessels A, van den Hoff MJ, Adamo RF, Phelps AL, Lockhart MM, Sauls K, Briggs LE, Norris RA, van Wijk B, Perez-Pomares JM, Dettman RW, Burch JB. 2012. Epicardially derived fibroblasts preferentially contribute to the parietal leaflets of the atrioventricular valves in the murine heart. Developmental Biology 366:111–124. doi: 10.1016/j.ydbio.2012.04.020 Carmona et al. eLife 2016;5:e16009. DOI: 10.7554/eLife.16009 17 of 17 Research article Developmental Biology and Stem Cells Human Biology and Medicine