|Home | About | Journals | Submit | Contact Us | Français|
Conceived and designed the experiments: DM RH SM GY. Performed the experiments: DM RH SM JM NN. Analyzed the data: DM RH FM GY. Wrote the paper: DM GY.
An omphalocele is one of the major ventral body wall malformations and is characterized by abnormally herniated viscera from the body trunk. It has been frequently found to be associated with other structural malformations, such as genitourinary malformations and digit abnormalities. In spite of its clinical importance, the etiology of omphalocele formation is still controversial. Hedgehog (Hh) signaling is one of the essential growth factor signaling pathways involved in the formation of the limbs and urogenital system. However, the relationship between Hh signaling and ventral body wall formation remains unclear.
To gain insight into the roles of Hh signaling in ventral body wall formation and its malformation, we analyzed phenotypes of mouse mutants of Sonic hedgehog (Shh), GLI-Kruppel family member 3 (Gli3) and Aristaless-like homeobox 4 (Alx4). Introduction of additional Alx4Lst mutations into the Gli3Xt/Xt background resulted in various degrees of severe omphalocele and pubic diastasis. In addition, loss of a single Shh allele restored the omphalocele and pubic symphysis of Gli3Xt/+; Alx4Lst/Lst embryos. We also observed ectopic Hh activity in the ventral body wall region of Gli3Xt/Xt embryos. Moreover, tamoxifen-inducible gain-of-function experiments to induce ectopic Hh signaling revealed Hh signal dose-dependent formation of omphaloceles.
We suggest that one of the possible causes of omphalocele and pubic diastasis is ectopically-induced Hh signaling. To our knowledge, this would be the first demonstration of the involvement of Hh signaling in ventral body wall malformation and the genetic rescue of omphalocele phenotypes.
The embryonic visceral organs transiently protrude out of the body trunk during mid-gestation, where they are covered with the peritoneal membrane. Subsequently they return to the peritoneal cavity in both mouse and human embryos. This transient embryonic hernia of the viscera is termed the physiological umbilical hernia , . According to previous reports, protrusion of the midgut loop through the umbilical ring is due to the rapid expansion in the volume of visceral organs, exceeding the space of the peritoneal cavity , . However, the molecular mechanisms underlying the ventral body wall formation, including physiological umbilical herniation, are still unclear.
An omphalocele is a major ventral body wall malformation characterized by a severe umbilical defect with herniation of visceral organs covered with peritoneum and amnion , , . The frequency is reported to be approximately 1 in 4,000 live births –. In spite of its high incidence, the cause of omphalocele is controversial; it might be due to the failure of recovery of the physiological umbilical hernia or to a midline defect at the transition zone between the ectoderm and mesoderm , –. Omphaloceles are frequently associated with other structural malformations such as cardiac, anorectal and digit malformations in more than 50% of cases , , . For instance, patients with omphalocele-exstrophy-imperforate anus-spinal defects complex (OEIS complex, OMIM: 25840) or bladder exstrophy (OMIM: %600057) exhibit defects not only in the body wall region but also in urogenital organs and its adjacent tissues, including the pelvic girdle –. Our understanding of these malformations is hampered by the complexity of these syndromes. Even the nomenclature and definitions for syndromic congenital malformations are still controversial –.
Several genetically-modified animals have been reported to display abnormalities in the body wall region. Such reports include cases of mutants of Msh-like homeobox 1 and 2 (Msx1/2), Transcription factor AP-2 alpha (Tcfap2α), Paired-like homeodomain transcription factor 2 (Pitx2), Insulin-like growth factor 2 (Igf2), Igf2 receptor (Igf2r), Transforming growth factor beta 2 and 3 (Tgfβ2/3), Bone morphogenetic protein 4 (Bmp4) and Bmp receptor type Ia (BmprIa) , -. Of note, most of these animals had accompanying limb deformities.
The Hedgehog (Hh) signaling pathway is an essential growth factor signaling pathway involved in many developmental contexts, including digit formation. One of the Hh ligands, Sonic hedgehog (Shh), is secreted from the posterior mesenchymal region of limb buds, the zone of polarizing activity. It is suggested that the anterior-posterior Shh gradient, together with a temporal gradient of exposure to Shh signaling, may specify digit number and identity –. Previous studies suggested that digit abnormalities such as polydactyly frequently accompany ectopic Hh signal induction in anterior limb buds , . Both inactivation of Patched 1 (Ptc1, a Hh signal repressor gene) and constitutive activation of Smoothened (Smo, a Hh signal transducer gene) or GLI-Kruppel family member 2 (Gli2, a Hh signal transcription factor gene) resulted in polydactylous phenotypes –. As for the mutants with body wall phenotypes, BmprIa, Msx1/2 and Tcfap2α mutants exhibited ectopic expression of Shh gene or its signaling genes (Gli1 or Ptc1) in their limb buds or other tissues –. Of note, the mutants of GLI-Kruppel family member 3 (Gli3) or Aristaless-like homeobox 4 (Alx4) also displayed body wall abnormalities, polydactylies and ectopic Hh signal activity in limb buds , , –. However, the correlation between omphalocele formation and Hh signaling has not yet been examined.
In this study, we investigated the participation of Hh signaling in ventral body wall formation and the pathogenic mechanisms leading to its malformation by utilizing a series of genetically-modified mouse systems. The phenotypic coordination of ventral body wall, digit and pelvic girdle formations is also discussed. We analyzed the lower body wall phenotypes of combinatorial mutants of Shh, Gli3 and Alx4 genes. We also analyzed conditional gain-of-function mutants of Hh signaling and revealed the Hh signal dose-dependent pathogenesis of omphalocele and pubic diastasis phenotypes. These results suggest that Hh signaling regulates omphalocele formation and shed light on the pathogenic mechanisms underlying a broad spectrum of lower body malformations.
The mutant mice used herein were Shh , Gli3Xt (XtJ) , Alx4Lst (LstJ) , , Gli1-CreERT2 , Shh-CreERT2 , Rosa26R , CAGGS-CreER™ , Rosa26-SmoM2  and del5-LacZ reporter , . The genotypes of each strain were determined as reported previously. To obtain Gli3Xt; Alx4Lst; Shh compound mutant embryos, single, double or triple heterozygous male and female mice were crossed. Noon of the day when the vaginal plug appeared was designated as embryonic day 0.5 (E0.5). Embryos for each experiment were collected from more than three independent pregnant females. All experimental procedures and protocols for animal studies were approved by the Committee on Animal Research of Kumamoto University (B22-198, B22-200, B22-201 and B22-202).
The tamoxifen (TM)-inducible Cre recombinase system removes the floxed sequence of the target genome –. TM (Sigma, St. Louis, MO, USA) was dissolved in sesame oil (Kanto chemical, Tokyo, Japan) at a final concentration of 10 mg/ml , , .
To analyze the cell contribution that responded to Hh signaling, we utilized the Gli1-CreERT2; Rosa26R system . The Gli1-CreERT2 mice were crossed with Rosa26R Cre-indicator (R26R) mice to obtain Gli1-CreERT2/+; R26R/R26R males, which were subsequently crossed with ICR females , . Time-mated ICR females were administered TM (2 mg per 40 g maternal body weight (bw)) orally with a gavage needle. Mouse embryos were processed for whole-mount X-gal staining.
For gain-of-function experiments of Hh signaling, the Rosa26-SmoM2 (R26-SmoM2) homozygous female mice were crossed with the Cre-driver mice, such as CAGGS-CreER™ transgenic male mice , , . The pregnant R26-SmoM2 females were treated with TM (1 mg, 2 mg or 4 mg per 40 g bw) orally with a gavage needle. Embryos were collected, and their morphology was investigated between mid-gestation and perinatal stages. No overt teratological effects were observed in wild-type embryos after TM administration under these conditions , , .
Mouse embryos were fixed overnight in 4% paraformaldehyde (PFA) (Sigma) with PBS, dehydrated through methanol, embedded in paraffin, and 8 µm serial sections were prepared. Hematoxylin and Eosin (HE) staining and X-gal staining were processed by standard procedures , , . For skeletal staining, dehydrated embryos were skinned, eviscerated and refixed in 95% ethanol for several days. Cartilage staining was performed for two days by incubation in 0.03% Alcian blue 8GX (Sigma), dissolved in 80% ethanol/20% acetic acid. After washing the embryos in 95% ethanol for five days, they were stained with 0.0025% Alizarin red S (Sigma) in 1% KOH for two days. Subsequently, they were treated with 1% KOH for 6 hours. Finally, the embryos were cleared with 20%, 40% and 60% glycerol and stored in 60% glycerol.
For the statistical analyses of the length of an extra digit, the length was measured with a slide gauge. Data were analyzed using a Student's t-test (two tailed). A probability of less than 0.001 was considered to indicate statistical significance. Values are given as the means±SD.
In situ hybridization was performed on PFA-fixed and dehydrated embryos. Samples were rehydrated, and incubated in 6% hydrogen peroxide solution for 1 hour. After washing in PBS containing 0.1% Tween 20, samples were incubated in 1 µg/ml ProteinaseK for 18 minutes, and refixed with fixing solution (4% PFA/0.2% glutaraldehyde) for 10 minutes. After washing with PBS containing 0.1% Tween 20, overnight incubation was performed in a buffer (50% formamide, 5x saline sodium citrate, 50 µg/ml yeast tRNA, 1% sodium dodecyl sulfate, 50 µg/ml heparin) at 65°C. Subsequent overnight hybridization was performed in a buffer with 0.5 µg/ml riboprobes at 65°C. Samples were washed in 50% formamide, 5x saline sodium citrate, 1% sodium dodecyl sulfate and 50% formamide, 2x saline sodium citrate for each 1 hour at 65°C, then 140 mM NaCl, 2.7 mM KCl, 0.1% Tween 20, 25 mM Tris-HCl (pH 7.5) for 5 minutes at room temperature before incubating with blocking solution (25% heated FBS in 140 mM NaCl, 2.7 mM KCl, 25 mM Tris-HCl (pH 7.5), 0.1% Tween 20) for 1 hour. Samples were treated with anti-digoxigenin antibody (Roche, Mannheim, Germany) in a blocking solution overnight at 4°C. After washing, samples were equilibrated in 100 mM NaCl, 50 mM MgCl2, 0.1% Tween 20, and 100 mM Tris-HCl (pH 9.5) including 2 mM levamisole (Sigma) and incubated in BM purple AP Substrate solution (Roche). Myogenin (kindly provided from Dr. Shosei Yoshida) and Gli1  probes were used. The preparation of the digoxigenin-labeled probes was performed according to the manufacturer's instructions (Roche).
Embryos were collected in PBS, rinsed in PBS and stained with 500 ng/ml Acridine Orange base (Fluka, St. Gallen, Switzerland) for 30 minutes. These procedures were performed at 37°C. Samples were then rinsed briefly in PBS, followed by fluorescence microscopy.
We analyzed the development of the embryonic body wall in a series of wild-type murine embryos. The protrusion of embryonic viscera covered with a peritoneal membrane (physiological umbilical hernia) was apparent by E12.5 (Fig. 1A,B) , . It was subsequently recovered from E16.5 onwards when the ventral body wall closed (Fig. 1C). As a result, only the umbilical cord could then be observed outside of the ventral body wall (Fig. 1C,D). We also analyzed pelvic girdle morphogenesis because patients with several congenital diseases, such as exstrophy of the cloaca, display malformations not only in the body wall region but also in the urogenital organs and the pelvic girdle –. The bilateral primordia (cartilaginous elements) of the pelvic girdle started to be perceptible from E11.5 (Fig. 1E)  and they were positioned in parallel along with the body trunk at E12.5 (Fig. 1F). Subsequently, the edges of the pubic bones started to close, but were not yet connected at the stage of the physiological umbilical hernia (at E14.5) (Fig. 1G). Consistent with the recovery of the physiological umbilical hernia, the pubic symphysis was formed at about E16.5 or later (Fig. 1H,I).
According to previous studies, several human patients and genetically-modified mouse models with body wall phenotypes often have accompanying digit abnormalities , , , , , , . Judging by the causative genes of digit abnormalities, we hypothesized that Hedgehog (Hh) signaling may also be involved in the onset of body wall malformation. To examine this hypothesis, we analyzed combinatorial mutants for Hh and putative Hh signaling related genes: Shh, Gli3 and Alx4. Hence, we analyzed the phenotypes of the hind limb, which is a well-analyzed system for examining genetic relationships among developmental genes. Wild-type and Shh+/− mice displayed normal digit morphology (Fig. 2A). Both Gli3Xt/+ and Alx4Lst/+ single heterozygotes showed preaxial polydactyly (Fig. 2B,D) , , . The size of the extra digit in Gli3Xt/+; Shh+/− mice was smaller than that of Gli3Xt/+ mice (Fig. 2C). On the other hand, this digit phenotype was completely restored in Alx4Lst/+; Shh+/− mice (Fig. 2E). Moreover, Gli3Xt/+; Alx4Lst/+ mice displayed severe polydactyly (two extra digits) (Fig. 2F). This phenotype was also partially restored by the addition of the Shh mutation (Fig. 2G). To quantify the effects of the gene mutations, we analyzed the significance of the length of the extra digit (Fig. 2H). The introduction of an additional Shh mutation significantly reduced the length of the extra digit (by a comparison between Gli3Xt/+ versus Gli3Xt/+; Shh+/−: 1.44±0.30, n=30 versus 0.90±0.24, n=26; P<0.001). On the other hand, the additional Alx4Lst mutation induced the opposite effect (Gli3Xt/+; Shh+/− versus Gli3Xt/+; Alx4Lst/+; Shh+/−: 0.90±0.24, n=26 versus 2.03±0.74, n=10; P<0.001). From these results, we suggest that both Gli3 and Alx4 genes may negatively regulate Hh signaling.
We generated graded levels of mutations for Hh signaling by introducing the Alx4Lst allele into a Gli3Xt/Xt background, and analyzed the resultant compound mutant embryos at E18.5 (Fig. 3A–D,A'–D'). The physiological umbilical hernia was recovered, and pubic symphysis was formed in wild-type embryos at E18.5 (Fig. 1D,I and Fig. 3A,A'). Decreasing wild-type Alx4 alleles accelerated the degree of omphalocele in the Gli3Xt/Xt embryos (Fig. 3B–D). In Gli3Xt/Xt; Alx4Lst/+ embryos and Gli3Xt/Xt; Alx4Lst/Lst embryos, the upper (dorsal) side of the genital tubercle was hypoplastic, in addition to the presence of an omphalocele (Fig. 3C,D). The development of the pelvic girdle also showed severe malformations in these mutants. The Gli3Xt/Xt embryos showed pubic diastasis (Fig. 3B'). The Gli3Xt/Xt; Alx4Lst/+ embryos displayed pubic diastasis and partial loss of pubic bones (Fig. 3C'). The Gli3Xt/Xt; Alx4Lst/Lst embryos showed more severe truncation and separation of the pubic bones than the Gli3Xt/Xt; Alx4Lst/+ embryos (Fig. 3D'). Thus, all of these mutants with omphalocele phenotypes displayed pubic diastasis.
We further analyzed the effects of mutations in Hh signaling related genes. The Gli3Xt/+; Alx4Lst/Lst embryos also exhibited multiple deformities, including an omphalocele, polydactyly and the loss of pubic bones and their diastasis (Fig. 3E–G). Introducing a Shh mutation could restore some of these phenotypes in Gli3Xt/+; Alx4Lst/Lst embryos (Fig. 3H–J). The omphalocele observed in Gli3Xt/+; Alx4Lst/Lst embryos (Fig. 3E) was restored completely in Gli3Xt/+; Alx4Lst/Lst; Shh+/− embryos (Fig. 3H). On the other hand, polydactyly was partially rescued, but was still observed in these mice. While Gli3Xt/+; Alx4Lst/Lst embryos displayed polydactyly (Fig. 3G), the number of extra digits was reduced in the Gli3Xt/+; Alx4Lst/Lst; Shh+/− embryos (Fig. 3J). With regard to pelvic girdle development, parts of the pubic bones were still not observed but the midline symphysis of the pelvic girdle was formed in Gli3Xt/+; Alx4Lst/Lst; Shh+/− embryos (Fig. 3I; asterisk). Taken together, these results suggest the possible involvement of Hh signaling in omphalocele and pubic diastasis phenotypes.
In order to analyze the contribution of Hh-responded cells, we utilized the Gli1-CreERT2; R26R system. In Gli1-CreERT2 mice, a TM-inducible form of Cre recombinase (CreERT2) was knocked into the Gli1 locus (Gli1-CreERT2), which is one of the direct target genes of Hh signaling , , , . The Gli1-CreERT2 hemizygotes correspond to Gli1+/− mutants, and displayed normal morphology in the ventral body wall (data not shown). By crossing Gli1-CreERT2/+; R26R/R26R males and ICR females, we could obtain Gli1-CreERT2/+; R26R/+ embryos. We treated pregnant ICR females once with 2 mg/40 g bw of TM at 8.5, 9.5, 10.5, 11.5 or 12.5 days post coitum, and embryos were collected at E14.5 (Fig. 4A–E) or at E13.5 (Fig. 4F). The recombination period in this system was estimated to occur within 6–12 hours and to continue for up to 36 hours after TM administration , . Our protocols were expected to detect Hh-responded cells during an embryonic period approximately from E8.75 to E14.0. Under these TM treatment conditions, we could not detect a significant LacZ-positive population in the ventral body wall region (Fig. 4A–F).
We also employed a reporter mouse strain (del5-LacZ) to locate active Hh signaling in vivo. The del5-LacZ model employs Gli-responsive binding sites identified in the upstream sequence of the Foxa2 gene , . In the ventral body wall region, we could not observe Hh signal activities in the del5-LacZ strain at E12.5 (Fig. 4G). This result was consistent with Hh-responded cell contribution analysis. In contrast, we observed ectopic Hh activity by del5-LacZ staining with the Gli3Xt/Xt mutation at E12.5 (Fig. 4H). These results imply that Hh signaling may not play essential roles in normal development of the embryonic ventral body wall, but may be implicated in omphalocele pathogenesis.
To assess the effects of ectopically-induced Hh signaling, we analyzed gain-of-function mutants of Hh signaling (hereafter designated as Hh-GOF) by utilizing the TM-inducible gene recombination system. Ectopic induction of Hh signaling was achieved by utilizing R26-SmoM2 and CAGGS-CreER™ mice. The CAGGS-CreER™ mice display Cre activity throughout the body upon TM treatment . The R26-SmoM2 allele possesses the constitutively activated form of Smoothened (SmoM2) and a floxed stop cassette under the ubiquitous Rosa26 promoter , . By crossing R26-SmoM2 mice and the TM-inducible form of Cre-driver mice, activation of Hh signaling was achieved. We analyzed mutant embryos that were treated once with various doses of TM (1 mg, 2 mg or 4 mg/40 g bw) at various time points on E9.5, E10.5, E11.5, E12.5 or E13.5, respectively. No noticeable toxic effects were observed for any of these TM treatment protocols , , . Upon administration of TM at E9.5, E10.5 or E11.5, mutant embryos displayed omphalocele and polydactyly phenotypes (Fig. 5B,B',D,D'; data not shown). The phenotypic differences induced by the different doses of TM were present following administration at E10.5 (Fig. 5C,C',D,D'). Omphaloceles were prominently observed in embryos from dams treated with the higher dose of TM (2 mg/40 g bw) but not with the lower dose (1 mg/40 g bw) (Fig. 5C,D). In contrast to the mutants treated with TM at E10.5, Hh-GOF mutants did not display an omphalocele even with the higher dose of TM treatment (4 mg/40 g bw) at E12.5 (Fig. 5E). On the other hand, the mutants exhibited an omphalocele induced by the lower dose of TM treatment (1 mg/40 g bw) at E9.5 (Fig. 5B). With regard to the phenotypes for digits and the pelvic girdle, the mutants with omphaloceles also showed severe polydactyly (Fig. 5B',D') compared with the non-omphalocele mutants in their hind limbs (Fig. 5C',E'). The pubic symphysis was formed in control embryos at E17.5 (Fig. 5F). The Hh-GOF mutants showed pubic diastasis when 2 mg/40 g bw of TM was administered at E10.5 (Fig. 5G). These results may indicate that the pathogenesis of omphalocele is induced by augmented Hh signaling in a time- and dose-dependent manner.
We further analyzed the Hh-GOF mutants in mid-gestation. To confirm the induction of ectopic Hh signaling, we performed gene expression analyses as one of the readouts of Hh signaling: Gli1 mRNA in Hh-GOF mutant embryos. The expression of Gli1 was observed ectopically throughout the body, including the lateral body wall in Hh-GOF mutants (Fig. 6H). We hypothesized that two potential causative factors might underlie the etiology of omphalocele formation. One could be an abnormality in the endodermal organs, such as an excess bulging out of visceral organs when the physiological umbilical hernia is observed. Another possibility could be defects in the mesodermal or ectodermal organs, such as a failure of the ventral body wall muscle formation. We expected that either or both of these factors could cause omphalocele formation. To assess these possibilities, we analyzed CAGGS-CreER™; R26-SmoM2 embryos using different TM administration protocols (TM treatment with 2 mg/40 g bw at E10.5 and harvested at E14.5, or TM treatment with 1 mg/40 g bw at E9.5 and harvested at E12.5 and E13.5). These TM administration protocols were sufficient to induce an omphalocele in later embryonic stages, and all of these mutants exhibited similar phenotypes (Fig. 5B,D). Interestingly, the volume of the herniating viscera in the peritoneal sac appeared smaller in Hh-GOF mutants than in control embryos (Fig. 6A,B; red arrow). On the other hand, excessive amount of cell death was detected by acridine orange staining in Hh-GOF mutants during the ventral body wall formation (Fig. 6D). With regard to the muscle differentiation, gene expression analyses of a muscle marker, Myogenin, suggested that the populations of muscle precursors in both the lateral body wall and limbs were decreased and distributed abnormally in Hh-GOF mutants (Fig. 6I–L). The lateral body wall of the mutants seemed to be disorganized (Fig. 6F; yellow arrowheads). Moreover, both epaxial and hypaxial muscle precursors seemed to be affected in Hh-GOF mutants (Fig. 6M; red arrowheads). These results might suggest that the pathogenesis of omphalocele in Hh-GOF mutants could be due to the failure of body wall formation and an abnormally enlarged umbilical ring associated with excessive cell death.
Recent advances in developmental biology and human embryology provide a profound understanding of the organogenesis and the pathogenesis of congenital diseases . Embryonic organogenesis is potentially influenced by genetic programs, maternal-embryonic interactions and embryonic physiological conditions , . The development of the ventral body wall displays dynamic processes, such as that observed during the formation and recovery of the physiological umbilical hernia. These processes proceed with the proper formation of adjacent structures, including body wall muscles and the pelvic girdle. Although some previous studies of genetically-modified animals have reported the processes of ventral body wall formation, our understanding of the ventral body wall formation and its related dysmorphogenesis remains incomplete. We herein reported that Hedgehog signaling is one of the causative factors for omphalocele formation, as demonstrated by utilizing a series of combinatorial mutants for Hh signaling genes and conditional gain-of-function mutants of the Hh signaling pathway. The analyses of Shh, Gli3 and Alx4 compound mutant embryos revealed that the introduction of additional Alx4Lst mutations into the Gli3Xt/Xt background resulted in the corresponding omphalocele and pubic diastasis. Moreover, the reduction of a single Shh allele restored omphalocele and pubic symphysis formation in the Gli3Xt/+; Alx4Lst/Lst embryos. The CAGGS-CreER™; R26-SmoM2 (Hh-GOF) conditional mutant analyses revealed the Hh signal-dependent omphalocele formation and pubic diastasis. This would therefore be the first demonstration of the involvement of Hh signaling in ventral body wall malformation and the genetic rescue of omphalocele formation.
In spite of recent advances in embryology and pathology, the etiology of omphalocele formation is still controversial. It has been suggested that a failure of the gut to return to the abdominal cavity after physiological herniation at appropriate developmental stages results in an omphalocele , , . According to this hypothesis, the lateral body wall closure has not been considered to be related to omphalocele formation, because the loops of the bowel are in the cord and are covered by amniotic membranes . Another possible cause of omphalocele could be a midline defect at the amnio-ectodermal transition, the transition zone between the ectoderm and mesoderm, which would result in an enlarged umbilical ring , , , . During normal development, a mature body wall covers the ventral surface surrounding the ring and the cord. With omphalocele, the mature body wall shows incomplete closure and it is localized to the periphery of the enlarged umbilical ring .
Our current studies suggested that few Hh-responded cells could contribute to the normal ventral body wall formation (Fig. 4A–F). In contrast to such results, we observed the ectopic Hh signaling in Gli3Xt/Xt mutants by utilizing the del5-LacZ Hh reporter mouse strain (Fig. 4H). In the Alx4Lst/Lst mutants, Shh expression was augmented in the cloacal epithelium, and ectopic Hh signal activity was observed in the ventral body wall region by del5-LacZ staining (data not shown). Moreover, gain-of-function mutants of Hh signaling displayed defects in body wall formation (Fig. 6F,J). These results suggested that omphalocele might be caused by ectopically-induced Hh signaling.
In this manuscript, we utilized the system for tamoxifen inducible ubiquitous activation of Hh signaling by CAGGS-CreER™; R26-SmoM2. However, the identification of such Hh-responded tissues remained unclear even when utilizing this conditional activation system. Hence, we further analyzed mutants that display specific activation of Hh signaling in the endodermal organs. To achieve endodermal activation of Hh signaling, Gli1-CreERT2; R26-SmoM2 mice and Shh-CreERT2; R26-SmoM2 mice were employed. The Gli1-CreERT2 mice and Shh-CreERT2 mice possess a tamoxifen-inducible form of Cre recombinase in the Gli1 and Shh gene loci, respectively , . Shh is specifically expressed in the endodermal epithelia of many visceral organs, and Gli1 is expressed mainly in the mesenchyme of visceral organs (Fig. S1A–C) . Regardless of the identity of Cre-driver lines, none of the mutants displayed omphalocele phenotypes (Fig. S2; data not shown). These results may suggest that Hh signal activation in endodermal organs may not be sufficient to induce such phenotypes under the current experimental conditions. In addition, the Gli1 gene is also considered to be one of the direct target genes of Hh signaling , , . Hence, the allelic combination of Gli1-CreERT2; R26-SmoM2 could result in Hh signal activation in Hh-responded tissues in such mutant embryos. Based on comparison of the phenotypes of Gli1-CreERT2; R26-SmoM2 and CAGGS-CreER™; R26-SmoM2 embryos, we would also suggest that the pathogenesis of the omphalocele was due to ectopically-induced Hh signaling (Fig. 5D and Fig. S2A').
With regard to the involvement of various mesodermal and ectodermal tissues in the pathogenesis of omphalocele formation, our previous study showed that there were abdominal wall defects with disorganized muscle layers and connective tissues in Msx1/2 double mutant mice . Likewise, abnormal Alk3-mediated BMP signaling in mesodermal tissues caused prenatal omphalocele-like defects . Moreover, the phenotypes of Pitx2 knockout mice and PITX2 mutation in humans indicate a correlation with the omphalocele formation , . The Pitx2 gene, which is expressed in abdominal muscles such as the rectus abdominis and oblique abdominis, has pivotal roles in muscle anlagen formation and maintenance , –. Altogether, these reports suggest essential roles of mesodermal or ectodermal tissues in omphalocele formation.
Another question arises regarding the onset of omphalocele formation. We showed the presence of a critical time-window for inducing omphalocele phenotypes by analyzing temporally-regulated conditional mutants of CAGGS-CreER™; R26-SmoM2 mice (Fig. 5B,D,E). Hh signal induction before the stage of physiological umbilical herniation (E9.5 and E10.5) resulted in omphalocele formation (Fig. 5B,D). In addition, activation of Hh signaling by TM injection at E12.5 did not lead to omphalocele formation (Fig. 5E). These results may imply that the onset of omphalocele formation thus begins before physiological umbilical herniation.
The pathogenic sequences of human patients with dysgenesis of the bladder and external genitalia have been reported. Such phenotypes often display exstrophy of the bladder and hypoplasia of the upper (dorsal) side of the penis , . Another characteristic symptom of bladder exstrophy is pubic diastasis (pelvic girdle separation) , . These complex congenital defects prompted the idea of coordinated urogenital organ development , . Despite the significant correlations of clinical conditions, the influence of pelvic girdle formation on such developmental coordination has previously been little analyzed.
Our current study revealed the concordant recovery of the physiological umbilical hernia and the closure of pelvic girdle at late embryonic stages (from E14.5 to E16.5) (Fig. 1B,C,G,H). According to textbook anatomy, the rectus abdominis arises from the front of the pubic symphysis and from the pubic crest , . Myocytes of the abdominal musculature are of somitic origin and the connective tissue elements within the abdominal wall, including tendons, derive from the somatopleure . Moreover, the entire pelvic elements also originate from the somatopleure , . We suggested that both body wall muscle primordia and pelvic girdle primordia arise from the lateral side of the embryonic trunk and develop toward the midline of the body (Fig. 1E–I and Fig. S3A–C). These observations imply that an unclosed pelvic girdle may be related to the phenotype with unclosed body wall muscles. In fact, Gli3Xt/Xt and Alx4Lst/Lst embryos displayed the absence of the midline muscular structures (Fig. S4C,D).
Hh signaling can affect the formation of midline structures in many developmental contexts of mammalian embryos, such as craniofacial formation. For instance, decreased Hh signal activity causes holoprosencephaly and hypotelorism, in contrast to augmented Hh signaling, which causes hypertelorism and frontonasal dysplasia . Our studies may also be supported by a broad spectrum of observations covering several processes of organogenesis. In the current study, the mutants with omphalocele phenotypes tended to show pubic diastasis. We revealed a possible association between the degree of omphalocele phenotype and pubic diastasis phenotype (Fig. 3B–D,B'–D'). In addition, mutants genetically rescued from omphalocele (Gli3Xt/+; Alx4Lst/Lst; Shh+/−) did form the pubic symphysis and normal midline muscle structures (Fig. 3I; data not shown). These results may indicate a role for Hh signaling in the formation of such midline structures, and a developmental correlation between the ventral body wall and the pelvic girdle.
While abnormal Hh signaling has been implicated as one of the major causes of polydactyly in mice, its involvement in the omphalocele formation has so far not been examined. In the current study, we showed that reduction of a single Shh allele could restore the omphalocele but could not restore the polydactyly of Gli3Xt/+; Alx4Lst/Lst embryos (Fig. 3H,J). In addition, the Hh-GOF mutants (CAGGS-CreER™; R26-SmoM2) without the omphalocele phenotype still exhibited polydactyly (Fig. 5C',E'). These results may suggest that the pathogenesis of the polydactyly phenotype seemed to be more sensitive to ectopically-induced Hh signaling than that of the omphalocele phenotype. In support of this notion, the frequency of polydactyly is higher than omphalocele, and has been reported to occur in approximately 1 in 600 human births . On the other hand, the frequency of omphalocele is relatively low, namely approximately 1 in 4,000 human births .
In the current study, we suggested that ectopically-induced Hh signaling might be one of the causes of a combination of polydactyly, omphalocele and pubic diastasis phenotypes. These results may offer a clue that can help elucidate the mechanisms underlying the formation of omphalocele and its associated syndromic malformations.
Cre recombinase activities of Shh-CreERT2 and Gli1-CreERT2in the developing gut. The Shh-CreERT2 activity was not observed in the ventral body wall at E14.5 upon tamoxifen treatment (4 mg/40 g maternal body weight) at E9.5 (A). Red arrow indicates the expression in the developing gut. The expression of Cre recombinase was detected in the endodermal epithelia of the midgut and posterior part of limb buds (B). The activity of Gli1-CreERT2 was also observed in the embryonic gut, including a part of the mesentery (C). The Gli1-CreERT2; R26R embryo was treated with 4 mg/40 g bw of tamoxifen at E10.5 and harvested at E13.5.
Augmentation of Hh signaling by utilizing Gli1-CreERT2 and Shh-CreERT2driver mouse lines. Both Gli1-CreERT2; R26-SmoM2 and Shh-CreERT2; R26-SmoM2 embryos did not display omphalocele phenotypes following administration of 4 mg/40 g bw of tamoxifen at E10.5 (A, A', B, B'). gt: genital tubercle, hl: hind limb, t: tail, uc: umbilical cord.
The expression of Myogenin in wild-type embryos at E10.5, E11.5 and E12.5. The ratio between primordia of hypaxial musculature (a, a' and a″) and epaxial musculature (b, b' and b″) was gradually increased during these stages, as hypaxial musculature (body wall muscle precursors) developed toward the midline (A-C).
Absence of midline structures in Alx4Lst/Lst and Gli3Xt/Xt embryos. Sagittal sections of a control embryo at E18.5 displayed prominent pubic symphysis (A; asterisk) and abdominal muscle structures (B). Muscles were stained with Anti-Skeletal Myosin antibody (FAST) (Sigma). Neither Alx4Lst/Lst (C) nor Gli3Xt/Xt (D) mutant embryos developed pubic symphysis or abdominal muscles, as shown by sagittal sections. b: bladder, gt: genital tubercle, om: omphalocele, r: rectum, u: urethra.
We thank members of our laboratory, Mika Kamimura, Aki Murashima, Mylah Villacorte, Yukiko Ogino, Masayo Harada and Kentaro Suzuki for comments and discussion. We would like to specially thank Drs. Alexandra L. Joyner, Shosei Yoshida, Andrew P. McMahon, Chi-chung Hui, Hiroshi Sasaki, Sanne Kuijper, Annemiek Beverdam, Toshihiko Shiroishi, Chin Chiang and Philippe Soriano for their invaluable support. We would also like to thank Drs. Pierre Chambon, Shigemi Hayashi, Sho Ohta, Shigeru Makino, Ken-ichi Yamamura, Kenji Shimamura, Shihuan Kuang, Alexander I. Agoulnik, Rolf Zeller, Richard R. Behringer and Anne M. Moon for encouragement and suggestions. We would also like to express our appreciation to Sawako Fujikawa, Keiko Horie and Yuka Endo for their assistance.
Competing Interests: The authors have declared that no competing interests exist.
Funding: This work is supported by Grant-in-Aid for Scientific Research B, for Scientific Research on Innovative Areas; Molecular mechanisms for establishment of sex differences (22132006), and the Global COE program Cell Fate Regulation Research and Education Unit from the Ministry of Education, Culture, Sports, Science, and Technology, Japan, and a grant for Child Health and Development (20-3) and Health Sciences Research Grant from the Ministry of Health, Labor, and Welfare, Japan. This work was also supported by National Institutes of Health Grant R01ES016597. The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.