|Home | About | Journals | Submit | Contact Us | Français|
Fibroblast growth factor (FGF) signaling to the epithelium and mesenchyme mediated by FGF10 and FGF9, respectively, controls cecal formation during embryonic development. In particular, mesenchymal FGF10 signals to the epithelium via FGFR2b to induce epithelial cecal progenitor cell proliferation. Yet the precise upstream mechanisms controlling mesenchymal FGF10 signaling are unknown. Complete deletion of Fgf9 as well as of Pitx2, a gene encoding a homeobox transcription factor, both lead to cecal agenesis. Herein, we used mouse genetic approaches to determine the precise contribution of the epithelium and/or mesenchyme tissue compartments in this process. Using tissue compartment specific Fgf9 versus Pitx2 loss of function approaches in the gut epithelium and/or mesenchyme, we determined that FGF9 signals to the mesenchyme via Pitx2 to induce mesenchymal Fgf10 expression, which in turn leads to epithelial cecal bud formation.
The cecum forms a pouch contiguous with the gastrointestinal (GI) tract, lying between the ileum and the large intestine or colon. Hosting a large reservoir of microbes, the cecum plays an important role in the digestion of small food particles and complex carbohydrates from plant matter. Thus, this part of the gut tends to be more prominent in herbivores and omnivores than obligate carnivores (Backhed et al., 2005; Eckburg et al., 2005). In mouse embryogenesis, the cecum starts to form at E10.5 as a mesenchymal expansion, followed by an epithelial evagination (Burns et al., 2004). Epithelial evagination initiation is then followed by elongation (growth), differentiation and arrest. To date, however, the signals defining cecal development remain incompletely understood.
The cecum, like other parts of the intestine, is composed of two layers: an endoderm-derived epithelium and the surrounding mesoderm-derived mesenchyme. Epithelial-mesenchymal interactions are required for proper budding morphogenesis and differentiation in many organs including the gut (Cardoso, 2001; Koike and Yasugi, 1999; Shannon and Hyatt, 2004). Fibroblast growth factors, key elements of epithelial-mesenchymal interactions in many tissues, have been described as major players in controlling cecum formation. We have shown that loss of FGFR2b signaling, in Fgf10 or Fgfr2IIIb knock-out (K.O.) embryos, results in the formation of a mesenchymal expansion, but the epithelium fails to proliferate and bud (Burns et al., 2004; Fairbanks et al., 2004). Moreover, the guts of Fgf9 null embryos display complete cecal agenesis, with the absence of both mesenchymal expansion and epithelial budding. This is accompanied with decreased mesenchymal proliferation as well as complete lack of Bmp4 and Fgf10 expression. In turn, absence of Fgf10 resulted in decreased epithelial proliferation (Zhang et al., 2006). During embryonic development, Fgf9 is mostly found in the epithelium of the cecum but is also detected at lower levels in the mesenchyme. However, compartment specific (epithelial vs. mesenchymal) deletion of Fgf9 in the cecum has not yet been studied.
Pitx2 is a member of the homeobox gene family that encodes a transcription factor initially identified as a gene mutated in Axenfeld-Rieger Syndrome type I, a rare autosomal dominant disorder that affects the development of the teeth, eyes and umbilicus (Semina et al., 1996a). In the GI tract, Pitx2 is mainly expressed in the mesenchyme of the developing cecum (Burns et al., 2004) and it was recently reported that classical Pitx2 inactivation in mouse leads to cecal agenesis (Nichol and Saijoh, 2011). In addition, it has been shown that over-expression of Hoxd12, another homeobox gene, phenocopies the loss of Fgf9 and leads to cecal agenesis and loss of Fgf10 and Pitx1 expression (Zacchetti et al., 2007).
Developmental studies of branching processes in different organs, e.g. cecum and lung, suggest that the mechanisms controlling branching are substantially conserved between organs. We have reported previously that deletion of mesenchymal β-catenin in the embryonic lung results in a loss of Pitx2 and Fgf10 expression. This was associated with impaired epithelial and mesothelial FGF9 signaling to the mesenchyme due to decreased expression of Fgfr2-IIIc (De Langhe et al., 2008). These results led us to propose that an FGF9/Pitx2/FGF10 signaling pathway controls lung bud formation. In the current study, we have used tissue specific Fgf9 and Pitx2 loss of function approaches in the gut epithelium and mesenchyme to show that this signaling axis is active in the developing gut, and demonstrate its importance for cecal formation.
Dermo1-Cre (C57Bl/6 background), and Fgf9flox/flox mice were obtained from Dr. David Ornitz (Washington University, Saint Louis, MO (Yu et al., 2003)) and Dr. Fen Wang (Institute of Biosciences and Technology, Houston, TX (Lin et al., 2006)) respectively. Pitx2flox/flox mice were previously described (Gage et al., 1999). Dermo1-Cre mice were crossed with Pitx2flox/flox mice to generate [Dermo1-Cre; Pitx2flox/+] that were then crossed with Pitx2flox/flox mice to generate [Dermo1-Cre; Pitx2flox/flox] mutant embryos (called hereafter Pitx2Dermo1-Cre). Shh-Cre mice were purchased from The Jackson Laboratory and were used to inactivate Pitx2 specifically in the epithelium. Shh-Cre mice were crossed with Pitx2flox/flox mice to generate [Shh-Cre, Pitx2f/+] that were then crossed with Pitx2flox/flox mice to generate [Shh-Cre; Pitx2flox/flox] mutant embryos (called hereafter Pitx2Shh-Cre). To inactivate Fgf9 in the mesenchyme, we used Dermo1-Cre as described above for Pitx2 inactivation. CMV-Cre mice were also used to completely inactivate Fgf9 throughout the embryo including both epithelium and mesenchyme of the gut. Animal experiments were performed under the research protocol (31-08) approved by the Animal Research Committee at Children's Hospital Los Angeles and in strict accordance with the recommendations in the Guide for the Care and Use of Laboratory Animals of the National Institutes of Health. The approval identification for Children's Hospital Los Angeles is AAALAC A3276-01.
Microdissected guts were fixed in 4% paraformaldehyde (PFA) for 20 min and dehydrated in ethanol. The samples were washed twice in PBS for 10 min, transferred and stored in 70% ethanol until use. Whole-mount in situ hybridization protocol was performed as described (Winnier et al., 1995). The following mouse cDNAs were used as templates for the synthesis of digoxigenin-labeled riboprobes: a 528 bp fragment of Fgf9 (provided by Dr. Ornitz), a 642 bp fragment of Shh (a kind gift from Dr Andrew McMahon, Harvard University, Boston, MA), a 584 bp fragment of Fgf10 (Bellusci et al., 1997), a 1.5 kb full-length mouse Bmp4 (Winnier et al., 1995), and a 559 bp fragment of Pitx2 present in all 3 Pitx2 isoforms (De Langhe et al., 2008). Sense probes were used for negative controls on E12.5 wild type ceca.
Intraperitoneal injection of 0.2 mL bromodeoxyuridine (BrdU, Amersham Biosciences UK) was given to pregnant females (4 pregnant females) carrying mutant and littermate control embryos at E12.5. The females were sacrificed 15 min later and the embryos were immediately placed in ice-cold Hank's solution. The ceca were dissected from the embryos, fixed in 4% PFA, gradually dehydrated in ethanol and processed for paraffin sectioning. The ceca (n=6) were uniformly dissected and oriented away from the label of the embedding cassette or slide, whith a short segment of the Ileum and a segment of the colon toward label. The embedded specimens were sectioned at 5 μm. The sections were re-hydrated and the antigen was retrieved by boiling the slides for 15 min in a microwave in 10 mM sodium citrate (pH 6.0). The slides were incubated for one hour with monoclonal anti-BrdU antibody (Clone BU-1) RPN 202 as recommended by the manufacturer (Amersham Biosciences, UK). Cy3-labeled anti-mouse secondary antibodies were used. The slides were then mounted using Vectashield containing DAPI and photographed. The epithelial and mesenchymal cells of the cecum were counted separately for the number of total cells and BrdU-labeled cells. The boundaries of the mutant ceca were defined by the mesenchymal thickness and the curvature angles on each side, as illustrated in figure 4, panel K. The results are reported as the percentage of BrdU-positive cells. Tissues from females not injected with BrdU, and sections stained with secondary antibody alone were used as negative controls. No staining was observed in these specimens.
RNA was extracted from individually microdissected ceca from Pitx2Dermo1-Cre mutant and littermate control embryos at E12.5 (n=8). One microgram of RNA was reverse-transcribed into cDNA using Transcriptor High Fidelity cDNA Synthesis Kit (Roche Applied Science, Indianapolis, USA) according to the manufacturer's instructions. cDNA (2 microliters) was used for dual color Hydrolysis Probe – Universal probe library based real time PCR, using the LightCycler 480 from Roche Applied Science. Mouse GAPDH gene assay (Roche applied Science) was used as the reference gene. The sets of primers and probe used for each gene examined are Bmp4 (Forward: GAGGAGTTTCCATCACGAAGA, Reverse: GCTCTGCCGAGGAGATCA, Probe 89), Fgf9 (Forward: GGGGAGCTGTATGGATCAGA, Reverse: TCCCGTCCTTATTTAATGCAA, Probe 12), Fgf10 (Forward: CGGGACCAAGAATGAAGACT, Reverse: AACAACTCCAGATTTCCACTGA, Probe 80), Pitx2 (Forward: CCTTACGGAAGCCCGAGT, Reverse: CCAAGCCATTCTTGCACA, Probe 40), Fgfr2b (Forward: CCCTACCTCAAGGTCCTGAA, Reverse: CATCCATCTCCGTCACATTG, Probe 21), Fgfr2c (Forward: TGCATGGTTGACAGTTCTGC, Reverse: TGCAGGCGATTAAGAAGACC, Probe 60). mRNA and water were used as negative controls.
Ceca were microdissected from wild type C57Bl/6 mice at E12.5, and placed atop polycarbonate filters (13mm diameter, 8 μm pore size, Whatman) in 1 mL DMEM/F12 supplemented with 5% Fetal Bovine Serum and 1% Penicillin/Streptomycin. Ceca were then incubated with or without 250 ng/mL of human recombinant FGF9 (n=4 in duplicates) at 37°C for 12 hours in a moist atmosphere (5% CO2).
NIH 3T3 murine fibroblasts (ATCC, #CRL-1658) were grown in DMEM with 10% FBS to 80% confluence. Cultures were transfected with plasmid encoding human Pitx2 (pCI-HAPitx2a) (Kozlowski et al., 2000) or empty vector (pCI) in Opti-MEM (Gibco Life Technologies) using Lipofectamine 2000 (Invitrogen), following the manufacturer's recommended protocol. 48h after transfection, cellular RNA was isolated for qRT-PCR analysis of Pitx2, Fgf9, and Fgf10 expression.
Statistical analyses were performed using Graphpad Prism software. All data are expressed as mean ± SEM. Comparisons of the changes between controls and DTG were performed using paired wilcoxon test. A p≤0.05 was considered significant. The proliferation and quantitative PCR analyses were performed on mutants and control littermates from at least 4 pregnant females. Each litter carried 1-3 mutants and several controls. The statistical analyses were performed on paired samples; in each experiment, n designates the number of embryos.
Previous studies have shown that Pitx2 is expressed in several organs including the left lateral plate mesoderm, eye, brain, pituitary glands, mandible, heart, lung, limbs and teeth (Arakawa et al., 1998); (De Langhe et al., 2008; Gage and Camper, 1997; Green et al., 2001; Kitamura et al., 1997; Muccielli et al., 1996; Semina et al., 1996b). It is also expressed in the developing mouse gut at E12.5, the midgut at E16.5, and in the cecum at E10.5 and E11.5 (Burns et al., 2004; Campione et al., 1999; Hjalt et al., 2000). However, the specific cellular compartments expressing Pitx2 have not been clearly defined. To examine the spatial expression of Pitx2 in the developing GI tract, we carried out whole mount in situ hybridization (WISH) for Pitx2 on whole isolated GI tracts between E12.5 and E14.5. At these stages, Pitx2 was strongly expressed in the small intestine and in the cecum, but absent in the colon as shown in Fig.1A-C. Vibratome sections of the WISH samples showed that Pitx2 is expressed in both epithelium and mesenchyme of the cecum at all the developmental stages studied (Fig.1D-F). Interestingly, mesenchymal Pitx2 expression is stronger at the apex of the bud compared to the stalk suggesting that it is directly involved in the proximal-distal growth of the cecal tube. However, in the small intestine Pitx2 expression is restricted to the epithelium early on at E12.5 (Fig. 1G), while at E14.5, it is expressed in both epithelium and mesenchyme (Fig. 1H).
It was recently reported that complete inactivation of Pitx2 results in cecal agenesis (Nichol and Saijoh, 2011). Since Pitx2 is expressed in both epithelium and mesenchyme of the cecum as shown in Fig. 1, it is unclear whether it is mesenchymal and/or epithelial Pitx2 which is/are required for cecal budding and elongation. We therefore examined the development of the cecum in mice with Pitx2 specifically deleted from the mesenchyme. Using the Dermo1-Cre driver line, Pitx2 deletion in the mesenchyme resulted in embryonic lethality around E17.5 (data not shown). Therefore control and Pitx2Dermo1-Cre embryos at E11.5, E12.5, E14.5 and E16.5 were used for this study. We first confirmed specific Pitx2 deletion in the mesenchyme of Pitx2Dermo1-Cre but not control (littermate Dermo1-Cre; Pitx2f/+) ceca at E11.5 (Fig. 2A,B). Note that Pitx2 expression in the epithelium of the mutant gut was still maintained. Dermo1-Cre; Pitx2f/+ animals were morphologically similar to their control littermates at all stages (>70 embryos were analyzed for each genotype at each time point). At E12.5, the control cecum normally developed into a mesenchymal protrusion surrounding a layer of epithelial bud (Fig. 2C). In E12.5 Pitx2Dermo1-Cre mutants, both the mesenchymal protrusion and epithelial buds were absent (n=70/70 embryos examined) even though the characteristic bending of the gut at that location was still visible (Fig. 2D). At E14.5 the wild type cecum continued to develop (Fig. 2E) while the mutant cecum did not show any evidence of budding or elongation (Fig. 2F). A similar observation was made at E16.5 (Fig. 2G,H). In contrast to mesenchymal deletion, removing Pitx2 from the epithelium using the Shh-Cre driver line did not affect cecal formation (Fig. 2I,J). Thus, we conclude that mesenchymal Pitx2 but not epithelial Pitx2 is required for cecal bud formation. Interestingly, we also observed that the small intestines of Pitx2Dermo1-Cre embryos are shorter than those of control littermates (Fig. 2L; 78.6% of the length of the controls at E14.5, p=0.0125, n=7, white arrows indicate the location of the cecum), whereas the colon lengths were similar (Fig. 2M). Additionally, around 65% of the mutants developed Meckel's diverticulum (black arrow, Fig. 2L) on the small intestine.
At E12.5, the epithelial and mesenchymal buds were normal in control embryos (Fig. 3A), whereas no epithelial or mesenchymal buds were observed in the mutant as shown by H&E staining (Fig. 3B). Since both mesenchymal and epithelial budding was affected in the mutant cecum, we assessed cell proliferation in these compartments using BrdU incorporation. Both epithelial and mesenchymal proliferation were reduced in the Pitx2Dermo1-Cre cecum (Fig. 3D,F) as compared to controls (Fig. 3C,E). Proliferation was reduced from 36.9% to 23.1% in the mesenchyme (Fig. 3G) (p=0.05, n=6) and from 45.2% to 27.4% in the epithelial cells (p=0.05, n=6) (Fig. 3G).
Loss of either Fgf9 or Fgf10 reduces proliferation in the epithelium and mesenchyme of the cecum (Burns et al., 2004; Zhang et al., 2006). Furthermore, the abnormalities we observed in the Pitx2Dermo1-Cre mutant cecum phenocopy those reported in Fgf9 mutants. Therefore, we examined the expression of Fgf9, Fgf10 and Bmp4 in Pitx2Dermo1-Cre and control littermates using WISH and RT-qPCR. Fgf10 was expressed in the mesenchyme of the cecum at E12.5 (Fig. 4A). However, in Pitx2Dermo1-Cre embryos, it was absent from the cecal mesenchyme as shown by WISH (Fig. 4B). Bmp4 was expressed exclusively in the mesenchyme of the wild type cecum at E12.5 (Fig. 4C) and this expression was maintained in the Pitx2Dermo1-Cre mutant embryos (Fig. 4D). In contrast to Bmp4, Fgf9 was expressed in both epithelium and mesenchyme of control cecum at E12.5 (Fig. 4E), whereas in Pitx2Dermo1-Cre animals it was only expressed in the epithelium (Fig. 4F, compare insets in E and F). WISH using sense probes for Fgf10, Bmp4 and Fgf9 on E12.5 wild-type ceca did not show any staining (Fig. 4G-I respectively).
The WISH data results were confirmed by qRT-PCR. The results showed a significant decrease in Pitx2, Fgf10 and Fgf9 levels in the mutant ceca as compared to controls (p=0.024, p=0.025 and p=0.05 respectively; Fig. 4J). No significant changes were detected in expression of Bmp4 or Fgfr2c. Moreover, we found a statistically significant reduction in Fgfr2b (p=0.023) in the Pitx2 conditional knockouts. These results suggest that maintenance of Fgf10 expression requires mesenchymal Pitx2. Conversely, we previously reported that deletion of Fgf10 in the cecum does not affect the expression of mesenchymal Pitx2 (Burns et al., 2004). Taken together, these observations indicate that Fgf10 is downstream of mesenchymal Pitx2. In addition, unlike what was observed in the lung (De Langhe et al., 2008), it does not appear that mesenchymal Pitx2 regulates Fgfr2c expression.
Specific mesenchymal Pitx2 inactivation mice show a cecum phenotype similar to the Fgf9 knockout. However, Fig. 4 shows that epithelial Fgf9 expression does not depend on mesenchymal Pitx2 expression. To test whether FGF9 can control Pitx2 expression, we exposed in vitro cultures of E12.5 cecum for 12 hours to 250 ng/mL of recombinant FGF9 in vitro. FGF9-treated ceca showed significant increase in Pitx2 and Fgf10 expression, assessed by qRT-PCR, as compared to controls (Fig. 5A), while no changes were observed in the expression of Fgfr2b or Fgfr2c. To investigate whether mesenchymal Fgf9 controls the expression of Pitx2 in the mesenchyme, we generated mutant embryos with mesenchymal (Fgf9Dermo1-Cre) or global (Fgf9CMV-Cre) deletion of Fgf9, using Fgf9f/f mice crossed with Dermo1-Cre and CMV-Cre respectively. At E14.5, the wild type embryo has a well-developed epithelial cecal bud surrounded by a thick layer of mesenchyme (Fig. 5B). Fgf9Dermo1-Cre embryos with specific inactivation in the mesenchyme displayed a hypoplastic cecum with a narrowed epithelial lumen surrounded by a thinner mesenchyme (Fig. 5C) as compared to the control. By comparison, Fgf9CMV-Cre embryos showed a more severe phenotype with a substantially attenuated epithelial and mesenchymal cecal bud (Fig. 5D). After deletion of mesenchymal Fgf9, Pitx2 was still strongly expressed albeit at a lower level in the Fgf9Dermo1-Cre cecum as compared to control (Fig. 5E, F). Global deletion of Fgf9 completely abolished the expression of Pitx2 in the mesenchyme of the Fgf9CMV-Cre gut (Fig. 5G). However, Pitx2 expression is still detectable in the epithelium at a significant level. In contrast, Shh expression was not affected by either mesenchymal or global Fgf9 deletion (Fig. 5H-J). We conclude that both mesenchymal and epithelial Fgf9 are important for the formation of the cecum.
To test whether Pitx2 can directly induce the expression of Fgf10 in the absence of Fgf9, We used an in vitro model of 3T3 cells to overexpress Pitx2. We first confirmed the absence of Fgf9 in these cells by PCR. The transfection of these cells with Pitx2a vector resulted in a 7-fold increase of Pitx2 as assessed by qRT-PCR (data not shown). Moreover, the overexpression of Pitx2 resulted in a significant increase (40%) in the expression of Fgf10 as compared to the control or the empty vector (Fig. 6). In contrast, no change in Fgf10 was induced by the empty vector.
In humans, the cecum is an important functional part of the GI tract. It is populated by a biodiverse microbiota that play an important role in the digestion of complex carbohydrates delivered from the small intestine (Backhed et al., 2005; Eckburg et al., 2005), and is also a significant site of intestinal immune activity. In the mouse, the cecum develops in 3 phases: bud initiation, bud elongation and bud arrest. To date, the molecular mechanisms regulating these phases are poorly understood. It was recently demonstrated that Pitx2 null embryos display cecal agenesis (Nichol and Saijoh). In this report, we show that mesenchymal Pitx2 regulates both the formation of the cecal mesenchymal bud and the subsequent epithelial bud induction. Absence of mesenchymal Pitx2 translates into cecal agenesis. However, Pitx2Dermo1-Cre mutants display a less severe phenotype than the Pitx2 null, which was expected. This is due in part to the fact that Pitx2 expression in the epithelium is still maintained. Pitx2 is mainly implicated in controlling left-right asymmetry during embryonic development (Liu et al., 2001) and is also known to regulate cell proliferation through cyclinD1 expression in several organs such as the gonads (Rodriguez-Leon et al., 2008). Consistent with previous reports, our data show that cecal agenesis is associated with decreased proliferation in both the epithelium and mesenchyme of the cecum (Burns et al., 2004; Zhang et al., 2006). However, deletion of epithelial Pitx2 did not adversely affect cecal formation.
Pitx2 has two homologues, Pitx1 and Pitx3. Pitx1 is expressed in the epithelium and mesenchyme of the developing gut and the cecum (Lanctot et al., 1997; Zacchetti et al., 2007). It is absent in the cecum of mice lacking HoxD cluster genes (homeobox domain) that also present with cecal agenesis. Fgf10 expression is also drastically reduced in these mice (Zacchetti et al., 2007), suggesting a role for Pitx1 in cecal formation as well as Fgf10 expression. However, Pitx1 null embryos display normal cecum formation (Supplementary Fig. S1). Pitx3 is mainly expressed in skeletal muscle, eye, midbrain and forebrain as demonstrated by WISH and Pitx3-GFP or Pitx3-LacZ reporter mice (Coulon et al., 2007; Grealish et al., 2010; L'Honore et al., 2007; Zhao et al., 2004). Pitx3 expression in developing gut or cecum has not been reported. This indicates a unique role for mesenchymal Pitx2 in coordinating cecal formation. However, we cannot exclude a possible functional redundancy between Pitx1 and Pitx2 in the epithelium.
We have previously shown that, in lung development, Pitx2 positively regulates Fgfr2c transcription. Through this mechanism, Pitx2 directly impacts the ability of the mesenchyme to respond to FGF9 and thus to maintain Fgf10 expression (De Langhe et al., 2008). Moreover, Fgf10 is strongly expressed in the mesenchyme of the cecum (El Agha et al., 2012). These results led us to propose that an FGF9/Pitx2/FGF10 signaling pathway could control bud formation. Our current results demonstrate that this model applies, at least partially, to the cecum as we found that epithelial FGF9 controls mesenchymal Pitx2 expression, which in turn controls mesenchymal Fgf10 expression. However, Fgfr2c expression was not changed upon mesenchymal Pitx2 inactivation, suggesting that Pitx2 does not control Fgfr2c or that other compensatory pathways could be involved. Further studies will be needed to characterize the regulators of Fgfr2c expression in the cecal mesenchyme.
Fibroblast Growth factors 9 and 10 along with their cognate receptors FGFR2c and FGFR2b, respectively, are known regulators of cecal development. Fgf9 is essential for mesenchymal cecal bud induction and controls Fgf10 expression. In turn, the FGF10/FGFR2b pathway is critical for the subsequent elongation phase of the epithelial cecal bud into the newly formed cecal mesenchyme (Burns et al., 2004; Fairbanks et al., 2004; Zhang et al., 2006). FGF10, from the mesenchyme, signals to the epithelium via its receptor FGFR2b. The FGF10/FGFR2b pathway is known to induce epithelial branching in several organ systems including lung, salivary glands and mammary glands (Min et al., 1998; Parsa et al., 2008; Sekine et al., 1999; Steinberg et al., 2005). Both Fgf10 and Fgfr2b null mice fail to develop a cecal epithelial bud (Burns et al., 2004) but still exhibit a significant mesenchymal bud while Fgf9 null embryos exhibited a complete absence of the epithelial and mesenchymal cecal bud. In this report we show that, similar to total Fgf9 knockout, deletion of Pitx2 in the mesenchyme results in the complete absence of mesenchymal cecal bud formation. Moreover, we showed that Pitx2 induces the expression of Fgf10 in the absence of Fgf9, therefore positioning Pitx2 downstream of Fgf9. These observations, along with the in vitro cecal cultures in the presence of recombinant FGF9, confirm our hypothesis that FGF9 acts upstream of Pitx2 during cecal development. In addition, the decrease of Fgf10 and Fgfr2b expression in Pitx2Dermo1-Cre mutants is consistent with what was observed in the Fgf9 null mutants.
It has been proposed that epithelial but not mesenchymal Fgf9 is required for cecal budding and elongation (Zhang et al., 2006). Our results demonstrate that mesenchymal Fgf9 is also an important player in the elongation of the cecal bud as mesenchymal Fgf9 deletion resulted in the formation of a shorter cecum and a thinner mesenchyme, accompanied with a decrease in Pitx2 expression. However, global Fgf9 deletion resulted in a much more severe phenotype, displaying only an attempt to form a rudimentary epithelial bud and a complete absence of mesenchymal Pitx2 expression. This rudimentary bud formation upon complete Fgf9 deletion could be due to variable or partial penetrance of the previously described Fgf9 null phenotype. However, as implied by global Fgf9 inactivation, epithelial Fgf9 is likely the key player in the initial steps of cecal formation.
Fgf10 expression is restricted to the caudal cecal mesenchyme while Fgf9 is expressed in both the epithelium and the mesenchyme of the developing small intestine, cecum and colon (Zhang et al., 2006). Interestingly, in the small intestine Pitx2 expression is restricted to the epithelium during early stages of development (E11.5-E12.5) and then starts to be expressed in the mesenchyme around E13.5. Our data indicate that mesenchymal Pitx2 expression is critical for the proper extension of the small intestine during development. Fgf9 null mutants showed a decrease in the length of the small intestine (about 80% of the control length at E14.5) (Zhang et al., 2006) similar to that observed for Pitx2Dermo1-Cre mutants. In both mutants (Fgf9 and Pitx2), this difference is noticeable mostly from E14.5 onwards, a time-point when Pitx2 is expressed in the mesenchyme. Taken together, these findings suggest that Pitx2 function downstream of Fgf9 is important for the elongation of the small intestine. This possibility is supported by the fact that the development of the colon, where Fgf9 is expressed but not Pitx2, is not affected by the deletion of Fgf9 (Geske et al., 2008). We therefore speculate that an important role of Pitx2 to control the elongation of the small intestine. Further analyses to test that hypothesis could be carried out. For example, expressing Pitx2 in the colon mesenchyme (where it is normally absent) could help determine whether or not this is sufficient to trigger excessive elongation.
In conclusion, Fgf9 compartment-specific deletions suggest that both epithelial and mesenchymal FGF9 are important for the proper formation of the cecum and that both sources control mesenchymal Pitx2 expression. In addition, the phenotype of Pitx2Dermo1-Cre mice suggests that FGF9 controls mesenchymal proliferation and acts mainly through mesenchymal Pitx2 to induce Fgf10 expression. In turn, FGF10 controls cecal epithelial budding and elongation (Fig. 7).
D.A.A. acknowledges previous support of the American Lung Association and current support from the American Heart Association. S.B. acknowledges support from NIH (HL086322, HD052609 and HL074832) and the Excellence Cluster in Cardio-Pulmonary system, Giessen, Germany. M.R.F. is supported by NIH awards DK077956 and DK090295. Tracy Grikscheit acknowledges support from CIRM (California Institute for Regenerative Medecine, RN2009461). We would like to thank Jonathan Branch for the maintenance of the different mouse colonies needed for this project and Clarence Wigfall for critical reading of the manuscript.
Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.