The gall bladder is a digestive organ whose function is to store bile. It is connected to the liver and the duodenum by bile ducts. The cystic duct fuses with the common hepatic duct to form the common bile duct (CBD). The bile produced in each liver lobe is collected in individual hepatic ducts, transported to and from the gall bladder by means of the cystic duct, and excreted into the gut by means of the CBD. Some animals such as rats and whales lack a gall bladder and cystic duct. In mice, the gall bladder arises in the foregut at almost the same time as the liver and pancreas (Lemaigre and Zaret,2004). First, the hepatic diverticulum develops as a bud of the ventral foregut endoderm at embryonic day (E) 9.0. The cranial and caudal portions of the hepatic diverticulum give rise to the hepatic and gall bladder primordia, respectively (Shiojiri,1997). The gall bladder and cystic duct also start to develop from hepatic endodermal cells and hepatoblasts just after liver primordium formation (Shiojiri,1997). Early hepatoblasts in the hepatic diverticulum migrate into the surrounding septum transversum and form the hepatic primordium and gall bladder primordium by E10.5. Bipotential hepatoblasts within the liver bud eventually give rise to hepatocytes and biliary epithelial cells (BEC), which line the bile ducts (Du Bois,1963; Shiojiri,1979).
Several mutant mouse lines have been shown to have gall bladder abnormality, such as mice deficient in the genes forkhead box F1a (Foxf1; Kalinichenko et al.,2002), one cut domain, family member 1 (Onecut1, previously called Hnf6; Clotman et al.,2002), hairy and enhancer of split 1 (Hes1; Sumazaki et al.,2004), and hematopoietically expressed homeobox (Hhex; Hunter et al.,2007). However, these mutants show not only gall bladder abnormality but also deformities in neighboring tissues. Little is known about the regulatory genes involved in mesenchymal–epithelial induction during early gall bladder development from the caudal portion of the hepatic primordium.
Leucine-rich repeat (LRR) -containing G protein-coupled receptor (LGR) family members are characterized by the presence of a seven-transmembrane domain and an extracellular domain with a long amino-terminal segment containing LRR motifs that is involved in receptor-ligand interaction (Hsu et al.,2000). The presence of a large extracellular domain is the distinctive feature that separates LGR family members from the other G protein-coupled receptors (GPCRs). Studies of LGRs from different species suggest that LGRs can be classified into three subtypes (A, B, and C) and that these three subtypes evolved during the early evolution of metazoans (Hsu,2003). Type A LGRs include the glycoprotein hormone receptors, such as follicle stimulating hormone receptor (LGR1), the luteinizing hormone receptor (LGR2) and the thyroid-stimulating hormone receptor (LGR3). Type C LGRs include the relaxin receptors LGR7 and LGR8, which are the receptors for relaxin3 and insulin-like 3 (INSL3), respectively. Type B LGRs comprise three members: LGR4 (also known as GPR48), LGR5, and LGR6, constituting a subfamily of LGRs with 17 LRRs in their putative hormone-binding domains. These are still orphan receptors and their physiological functions have not yet been determined (Hsu,2003). A putative ortholog of the Type B subfamily LGR2 in Drosophila (DLGR2) was demonstrated to be the receptor for bursicon, a hormone released at the end of the molting cycle, triggering tanning of the cuticle (Mendive et al.,2005; Luo et al.,2005). However, a bursicon ortholog has not yet been identified in vertebrates.
Recently, the developmental function of Lgr4 has been investigated using knockout mice. The generation of Lgr4 gene-interrupted mice used a secretary gene-trap system (Leighton et al.,2001), in which the expression of Lgr4 was severely attenuated by the trap vector (Mazerbourg et al.,2004). On the C57BL/6 background, a phenotype of intrauterine growth retardation associated with embryonic and perinatal lethality was described, thus preventing analysis of either the postnatal or adult phenotype. Another Lgr4 gene targeted mouse model showed embryonic/neonatal lethality with the additional phenotype of hypoplastic kidneys and an increased concentration of plasma creatinine (Kato et al.,2006). We generated an Lgr4 hypomorphic mutant (Lgr4Gt/Gt) by gene-trap insertional mutagenesis in embryonic stem cells using an exchangeable trap vector (Taniwaki et al.,2005). Our mutant mice exhibit a milder phenotype than the null mutant mice in terms of their growth and perinatal death rate. Adult males show short, dilated, and much less convoluted ducts in the epididymis (Hoshii et al.,2007), as shown by Mendive et al. (2006). These studies revealed that Lgr4 is essential for postnatal development of the male reproductive tract. Kato et al. and Mohri et al. reported that Lgr4 regulates eyelid and hair follicle development by controlling keratinocyte motility (Kato et al.,2007; Mohri et al.,2008). Weng et al. reported that Lgr4 plays an important role in the development of the ocular anterior segment (Weng et al.,2008). Taken together, it appears that Lgr4 is involved in the development of multiple tissues.
Here, we describe a new function of Lgr4. It is required for the development of the gall bladder and cystic duct, but not the development of the liver, pancreas, intrahepatic bile duct (IHBD), common hepatic duct (CBD). This is the first report of a mutant mouse that lacks only gall bladder and cystic duct.
Lgr4 Insufficiency Results in an Absence of Gall Bladder and Cystic Duct
The gene trap mouse line, B6;CB-Lgr4GtAyu21-127Imeg (Lgr4Gt) has been described previously (Hoshii et al.,2007). On a C57BL/6J genetic background, approximately 85% of homozygous mice died during the postnatal development period. However, on a CBA genetic background, 60% of the offspring survived, although they did show growth retardation (Hoshii et al.,2007). Offspring on a CBA background were used in the following studies.
Macroscopic analysis showed that 27 of 27 (100%) of adult Lgr4Gt/Gt mice exhibited an absence of the gall bladder and cystic duct, but normal common hepatic ducts and CBD (Fig. 1C). In contrast, all Lgr4+/+ (n = 37; Fig. 1A) and Lgr4+/Gt (n = 55; Fig. 1B) littermates had a normal gall bladder and cystic duct. These results suggested that Lgr4 was solely required for gall bladder and cystic duct development. Although we carried out all experiments described below using wild-type, heterozygous, and homozygous mice, the heterozygous mice showed normal development of the epithelial and mesenchymal layers in gall bladder and cystic duct when compared with wild-type mice, Thus, we only described the data on wild-type and homozygous mice in the following results.
Temporal and Spatial Expression of Lgr4 During Gall Bladder Development
Because the trapped Lgr4 allele contains the β-galactosidase/neomycin phosphotransferase fusion gene (β-geo) gene derived from the pU21 trap vector, staining for beta-galactosidase (β-gal) activity allows the identification of Lgr4-expressing cells. At Theiler stage (TS) 14 (E10.25) in Lgr4+/Gt embryos (Fig. 2A–D), weak β-gal staining was detected in the foregut, midgut, and hindgut epithelium and mesenchyme, but not in the hepatic diverticulum (Fig. 2B,D). At TS15 (E10.5) in Lgr4+/Gt embryos (Fig. 2E–H), β-gal staining was observed in the gall bladder epithelium (Fig. 2F,H). At TS16 (E10.75), a strong signal was detected in the epithelium of the gall bladder primordium (Fig. 2J,L). However, β-gal staining was not observed in the mesenchyme around the hepatic diverticulum or gall bladder primordium (Fig. 2F,H,J,L). At TS18 (E11.5) (Fig. 2M–P), the staining was more intense and spread to the ventral mesenchyme near the gall bladder (Fig. 2N,P). The staining in hepatic diverticulum and hepatic primordium was not detected at any stage. At TS19 (E12.5), the same staining pattern was observed as that at TS18 (data not shown). Similar results were obtained by endogenous gene expression analyses by in situ hybridization at E11.5, E12.5, E14.5, and E16.5 (data not shown). These results indicate that Lgr4 is involved in the development and maintenance of the gall bladder and cystic duct.
Lgr4 Is Required for Elongation of Gall Bladder Epithelium and Maintenance of the Mesenchyme Surrounding the Gall Bladder and Cystic Duct
To determine the stage when gall bladder and cystic duct development ceases, we examined gall bladder development in each genotype by histological analysis (Fig. 3). At E10.25, the liver bud enlarges and divides into two parts, the primordium of the liver and the gall bladder. At TS14 (E10.25) Lgr4Gt/Gt embryos developed a future gall bladder (Fig. 3D) that was similar to that of Lgr4+/+ embryos (Fig. 3B). At TS18 (E11.5), in Lgr4+/+ embryos (Fig. 3F) the gall bladder elongated along the liver and was surrounded by mesenchyme. In contrast, no elongation was observed in Lgr4Gt/Gt embryos (Fig. 3H), although the mesenchyme developed normally and surrounded the gall bladder and cystic duct (Fig. 3H). At E12.5, in Lgr4+/+ embryos (Fig. 3J) the gall bladder elongated ventrally and more mesenchyme appeared around the epithelium of the tubular structure. In Lgr4Gt/Gt embryos (Fig. 3L), the mesenchyme surrounding the gall bladder had mostly disappeared leaving only maintenance mesenchyme surrounding unelongated epithelium, while the gall bladder remained unelongated. To investigate whether the disappearance of the gall bladder was caused by apoptosis, we performed terminal deoxynucleotidyl transferase-mediated deoxyuridine triphosphate nick-end labeling (TUNEL) analysis of the gall bladder and CBD at E11.5. There were no differences among all genotypes (data not shown). These results suggest that Lgr4 is required for elongation of gall bladder epithelium, and maintenance of mesenchyme surrounding the gall bladder.
Reduced Lgr4 Does Not Affect the Development of the CBD, Duodenum, Hepatic Primordium, or Pancreatic Buds
We analyzed the influence of a reduction in Lgr4 on the other tissues related to the gall bladder. The presence of the CBD and duodenum can be examined by staining with Dolichos biflorus agglutinin (DBA), which is a bile duct-specific lectin. The DBA staining pattern in Lgr4Gt/Gt embryos was the same as in Lgr4+/+ embryos (Fig. 4A,B). This result suggested that BEC in the CBD and the duodenum of Lgr4Gt/Gt embryos were unaffected. Next, we examined the expression of the hepatoblast markers albumin (Alb) and alpha fetoprotein (Afp), by in situ hybridization. Alb expression was detected in the hepatic primordium and gall bladder-CBD junction in Lgr4Gt/Gt embryos as well as in Lgr4+/+ embryos (Fig. 4C,D). Afp expression was observed in the gall bladder and the regions of Alb expression in both Lgr4Gt/Gt and Lgr4+/+ embryos (Fig. 4E,F). These hepatoblast marker analyses revealed that the differentiation and distribution of hepatoblasts were normal in the hepatic primordium and biliary duct of Lgr4Gt/Gt embryos. To analyze pancreatic development, we performed immunohistochemistry for PDX1, an early pancreatic maker. PDX1-positive cells were observed in the ventral pancreatic bud (VPB) and dorsal pancreatic bud (DPB) of Lgr4Gt/Gt embryos and Lgr4+/+ embryos (Fig. 4G,H). Recently cell fate analysis has indicated that PDX1 is expressed in the proximal part of the developing CBD (Offield et al.,1996), and this factor is also present in the BEC lining the entire extrahepatic biliary tract, albeit at lower levels than in the pancreas (Sumazaki et al.,2004). As PDX1 is not expressed in the intrahepatic portion of the biliary tree, this distinguishes the two parts of the biliary tract (Lemaigre and Zaret,2004). A high magnification view of the gall bladder primordium (Fig. 4I,J) showed that weakly PDX1-positive cells were detected in the BEC of both genotypes. This result demonstrates that Lgr4Gt/Gt embryos have normal gall bladder primordium and BEC at the early stages of gall bladder development.
Taken together, these results suggest that Lgr4 is involved in gall bladder development, but not in the development of the CBD, duodenum, hepatic primordium or pancreatic buds.
Reduced Lgr4 Does Not Affect the Intrahepatic Bile Duct
To examine the development of BEC in the IHBD of Lgr4Gt/Gt mice, we analyzed liver sections with hematoxylin and eosin (H&E) staining and immunostaining using an anti-cytokeratin (CK) antibody that can specifically label BEC (van Eyken et al.,1987). At 1 week of age in both genotypes, mature bile ducts visualized as BEC staining were observed around the portal vein (Fig. 5A,B). The number and location of IHBD in Lgr4Gt/Gt mice were not significantly different from Lgr4+/+ mice. At 8 weeks of age, we obtained the same results (Fig. 5C,D) as at 1 week of age. Furthermore, we immunostained BEC in serial sections with an anti-CK antibody. The distribution of CK-positive cells in Lgr4Gt/Gt mice was not significantly different from Lgr4+/+ mice (Fig. 5E,F). Moreover, there was no β-gal staining in IHBD at E17.5. These results suggest that Lgr4 insufficiency does not affect IHBD formation.
Lgr4 insufficiency causes impaired development of the kidney (Kato et al.,2006), epididymis (Mendive et al.,2006; Hoshii et al.,2007), eyelid (Kato et al.,2007), iris (Weng et al.,2008), and hair follicle (Mohri et al.,2008). Here we demonstrate, using Lgr4 gene trapped mice, that Lgr4 regulates the development of the gall bladder and cystic duct, but not the neighboring tissues.
Several transcription factors that control biliary tract development have been identified through the production of knockout mouse lines. Foxf1 is a mesenchymal transcription factor that is expressed in mouse embryonic septum transversum. Haploinsufficient Foxf1 mice either lacked a discernible gall bladder or possessed a significantly smaller gall bladder with severe structural abnormalities, as well as a reduced number of mesenchymal cells (Kalinichenko et al.,2002). However, the fact that 2% of heterozygous mice exhibited normal gall bladder formation suggested that Foxf1 is involved in the later stages of gall bladder development. Onecut1 is expressed in hepatoblasts, the gall bladder primordium and the BEC of the developing IHBD. In Onecut1-deficient mice, the gallbladder was absent, the EHBD were abnormal and the development of the IHBD was perturbed in the prenatal period. Thus, Onecut1 is required for gall bladder formation, for normal differentiation of BEC and for proper morphogenesis of the IHBD and EHBD (Clotman et al.,2002). In the mouse, Onecut1 also controls differentiation of pancreatic precursors into endocrine cells by induction of Pdx1 expression (Jacquemin et al.,2003; Jacquemin2006). The homeobox gene hairy and enhancer of split 1 (Hes1) encodes the basic helix–loop–helix protein (Sasai et al.,1992), which represses positive helix–loop–helix proteins such as Neurogenin 3. Sumazaki et al. described that Hes1 is expressed in the EHBD epithelium throughout development, and that Hes1-deficient mice have gall bladder agenesis and severe hypoplasia of the EHBD (Sumazaki et al.,2004). The hematopoietically expressed homeobox (Hhex) gene is expressed in the ventral foregut endoderm at E8.5 (Thomas et al.,1998; Bogue et al.,2000). In Hhex-null embryos, it was shown that there was a complete failure in ventral pancreatic specification (Bort et al.,2004), while the hepatic diverticulum was specified, but reduced resulting in failure of hepatobiliary development (Keng et al.,2000; Martinez Barbera et al.,2000; Bort et al.,2006; Hunter et al.,2007). Taken together, most of the above knockout mice show a wide range of abnormalities and their phenotypes are different from those in our mice, suggesting that Lgr4 functions in a very restricted region.
In this study, we demonstrated that developmental abnormality in Lgr4Gt/Gt mice was restricted to the gall bladder and cystic duct. Although the gall bladder primordium was formed (Fig. 3D), further development of the gall bladder was not observed (Fig. 3H,L). Gall bladder development requires BEC differentiation from the foregut and proliferation by epithelial–mesenchyme interaction (Shiojiri,1997). During the elongation of the gall bladder, Lgr4 was expressed in the epithelium of the gall bladder (Fig. 2H,I). Therefore LGR4 protein in the epithelium may interact with its ligand expressed in the surrounding mesenchyme. Although the ligand for LGR4 has not yet been identified, a relationship between Lgr4 and cell motility has been reported. Gao et al. (Gao et al.,2006) reported a relationship between Lgr4 and tumorigenesis. Expression of constitutively active mutant of Lgr4 increased invasive activity and lung metastasis potency of HCT116 cells. In contrast, depletion of endogenous Lgr4 by RNA interference reduced the invasive potential of HeLa and Lewis lung carcinoma cells in vitro and in vivo. Kato et al. demonstrated that the open eyes at birth in Lgr4−/− mice was due to decreased keratinocyte motility, which was confirmed by an in vitro wound-healing scratch assay (Kato et al.,2006). Correspondingly, the kidney (Kato et al.,2006), epididymis (Mendive et al.,2006; Hoshii et al.,2007), and follicle (Mohri et al.,2008) phenotypes caused by Lgr4 deficiency are also related to epithelium formation and elongation, but the regulatory mechanisms for epithelial motility and proliferation are not known.
Weng et al. (2008) reported that the downstream target gene of Lgr4 was paired-like homeodomain transcription factor 2 (Pitx2), which is known to be a transcription factor of left–right signals (Logan et al.,1998). Pitx2−/− embryos showed a failure of ventral body wall closure, heart and lung development (Kitamura et al.,1999). We observed Pitx2 expression in the gall bladder epithelium by in situ hybridization, and this expression was unchanged in Lgr4Gt/Gt embryos (data not shown). Furthermore, Pitx2−/− embryos possessed a gall bladder-like structure (data not shown). These results suggest that Pitx2 signaling is not involved in the formation of the gall bladder, and that Lgr4 probably regulates epithelial behavior by means of another signaling pathway.
It is of interest that the mesenchyme surrounding unelongated epithelium completely disappeared by E12.5 in Lgr4Gt/Gt embryos. Kalinichenko et al. reported that Foxf1 was expressed in the embryonic septum transversum and gall bladder mesenchyme, and that it was necessary for the epithelium–mesenchyme interaction (Kalinichenko et al.,2002). Foxf1+/− gall bladders were significantly smaller and had severe structural abnormalities characterized by a deficient external smooth muscle cell layer, reduction in mesenchymal cell number, and in some cases, lack of a discernible biliary epithelial cell layer. Thus, gall bladder mesenchyme may be required for normal gall bladder development. As Lgr4 was expressed both in the epithelium of gall bladder and mesenchyme, there are two possibilities concerning the maintenance of mesenchyme, One is that the expression of Lgr4 in gall bladder epithelium is required for maintenance of the mesenchyme. The other is that Lgr4 expression in mesenchyme is required to maintain itself. Further study will be required to distinguish these possibilities.
In this report, we have described a new, morphogenetic function of Lgr4, which is that it is required for cystic duct and gall bladder development but not common hepatic duct and CBD development. This gives insight into the function of Lgr4 and the mechanisms of gall bladder and cystic duct development.
The establishment of Lgr4Gt/Gt mice was described previously (Hoshii et al.,2007). In this study, we used the F5 generation of mice. CBA/n mice were purchased from Charles River Japan Inc. (Yokohama, Japan). Mice were housed in the Center for Animal Resources and Development, Kumamoto University. All animal experiments were carried out with the approval of the Ethics Committee at the Center for Animal Resources and Development, Kumamoto University.
X-gal Staining for β-Galactosidase Activity
To detect Lgr4 expression levels, we determined the β-gal activity produced by the trapped allele. Samples were fixed for 30 min at room temperature in 1% formaldehyde, 0.2% glutaraldehyde, and 0.02% NP-40. Fixed samples were washed twice in phosphate buffered saline (PBS) and incubated for 16 hr at 37°C in a staining solution of 5 mM potassium ferricyanide, 5 mM potassium ferrocyanide, 2 mM MgCl2, and 0.5% X-gal in PBS. Samples were washed twice in PBS and then postfixed in 4% paraformaldehyde (PFA). For each experiment, rigorous controls were carried out simultaneously in Lgr4+/+ embryos, to assay background, endogenous β-gal activity. Developmental stages were classified according to the morphology, using information from the Edinburgh Mouse Atlas Project (http://genex.hgu.mrc.ac.uk/intro.html).
For the detection of apoptosis, TUNEL assays were performed using an in situ apoptosis detection kit (Wako, Osaka, Japan).
Embryos and liver tissue were fixed with 4% PFA in PBS for 16 hr at 4°C and processed for paraffin sectioning according to standard procedures. Sections were incubated with the primary antibodies in 2.5% BSA in TBS for 16 hr at 4°C. Primary antibodies were polyclonal rabbit anti-PDX1 (Chemicon International, Temecula, CA AB3243) and polyclonal anti-human keratin (Dako, Carpinteria, CA, A0575). After washing, sections were incubated with the corresponding secondary antibodies (all at 1:200 dilution) for 1 hr at room temperature. The Vectastain ABC Kit (Vector Laboratories, Burlingame, CA) was used for the avidin-biotin complex method according to the manufacturer's instructions, and detected with diaminobenzidine. We performed DBA staining using DBA-HRP (EY Laboratories, San Mateo, CA) as described previously (Shiojiri and Katayama,1988).
In Situ Hybridization
Embryos were fixed in 4% PFA in PBS for 48 h at room temperature before they were embedded into paraffin and cut into 7-μm sections. For the digoxigenin (DIG) -labeled RNA sense (control) and antisense riboprobes (Roche, Basel, Switzerland), total RNA was extracted from the embryo and then each cDNA template was polymerase chain reaction (PCR) amplified with Alb- and Afp-specific primers. The PCR products were cloned into the pGEM-T Easy Vector (Promega, Madison, WI). Slides were subsequently processed for in situ hybridization. In situ hybridization was performed on embryonic tissue sections with DIG-labeled RNA probes using a VENTANA ISH machine (Ventana Medical Systems Inc., Tucson, AZ) as described previously (Wang et al.,2008).
We thank Mrs. Michiyo Nakata and Mrs. Yumi Ohtake for their excellent work on tissue section preparation. We also thank Dr. Shoen Kume and Dr. Masaki Ohmuraya for helpful discussions.