SEARCH

SEARCH BY CITATION

Keywords:

  • beta cells;
  • insulin;
  • snail;
  • endoderm;
  • islet;
  • Langerhans;
  • delamination

Abstract

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. RESULTS
  5. DISCUSSION
  6. EXPERIMENTAL PROCEDURES
  7. Acknowledgements
  8. REFERENCES
  9. Supporting Information

During development, pancreatic endocrine cells are specified within the pancreatic epithelium. They subsequently delaminate out of the epithelium and cluster in the mesenchyme to form the islets of Langerhans. Neurogenin3 (Ngn3) is a transcription factor required for the differentiation of all endocrine cells and we investigated its role in their delamination. We observed in the mouse pancreas that most Ngn3-positive cells have lost contact with the lumen of the epithelium, showing that the delamination from the progenitor layer is initiated in endocrine progenitors. Subsequently, in both mouse and chick newly born endocrine cells at the periphery of the epithelium strongly decrease E-cadherin, break-down the basal lamina and cluster into islets of Langerhans. Repression of E-cadherin is sufficient to promote delamination from the epithelium. We further demonstrate that Ngn3 indirectly controls Snail2 protein expression post-transcriptionally to repress E-cadherin. In the chick embryo, Ngn3 independently controls epithelium delamination and differentiation programs. Developmental Dynamics 240:589–604, 2011. © 2011 Wiley-Liss, Inc.


INTRODUCTION

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. RESULTS
  5. DISCUSSION
  6. EXPERIMENTAL PROCEDURES
  7. Acknowledgements
  8. REFERENCES
  9. Supporting Information

The pancreas is an organ playing an essential role as a dual exocrine and endocrine gland (Johansson and Grapin-Botton,2002; Habener et al.,2005). The endocrine function relies on the islets of Langerhans, which are clusters of cells separated from the branched exocrine tree and dispersed in the pancreas. Islets are composed of five endocrine cell types, each secreting a specific hormone (Johansson and Grapin-Botton,2002; Habener et al.,2005; Lyttle et al.,2008). The clustered organization is essential for the metabolic function of the islets, notably synchronized and synergistic secretion of insulin by beta cells coupled by gap-junctions (Bosco et al.,1989; Bavamian et al.,2007). Pancreatic islets are polyclonal (Deltour et al.,1991; Desgraz and Herrera,2009) and form through the aggregation of endocrine cells originating from different locations. These cells can then proliferate, especially in the perinatal and early postnatal period to expand islet mass. Because all pancreatic endocrine cells arise from epithelial progenitors during embryonic development, they must delaminate from the epithelium, migrate in the mesenchyme, and aggregate into clusters. There are some evidences suggesting that newly generated endocrine cells delaminate by achieving an epithelial-to-mesenchymal transition (EMT; Chiang and Melton,2003; Rukstalis and Habener,2007; Cole et al.,2009). EMT is a process by which an epithelial cell becomes mesenchymal and which involves profound phenotypic changes that include the loss of cell–cell adhesion, the loss of cell polarity, and the acquisition of migratory and invasive properties (Thiery et al.,2009; Acloque et al.,2009). Achieving a successful EMT induces the cell to exit the epithelium, a process commonly referred to as delamination. After embryonic day (E) 14.5 in mouse embryos, it was reported that endocrine cells exhibit low E-cadherin expression and up-regulate N-cadherin, although it is unclear when this switch occurs (Hutton et al.,1993; Johansson et al.,2010). Moreover, Snail2 (previously known as Slug), a zinc finger transcription factor that promotes EMT, was shown to be expressed in the pancreas (Rukstalis and Habener,2007). After delamination, endocrine cells migrate in the mesenchyme. Several genes have been shown to play a role in this process such as integrins, Rac1 and matrix metalloproteases (Miralles et al.,1998; Cirulli et al.,2000; Miettinen et al.,2000; Yebra et al.,2003; Perez et al.,2005; Greiner et al.,2009). Finally, endocrine cells aggregate into clusters. Cadherin function is required for β-cell aggregation (Dahl et al.,1996). Indeed, expression of a dominant negative E-cadherin construct, inhibiting the function of all cadherins, resulted in perturbed islets aggregation. N-cadherin was recently shown to be dispensable for endocrine cells aggregation, suggesting that this function depends on E-cadherin or redundant activities of N- and E-cadherin (Johansson et al.,2010). However, the role of E-cadherin in delamination has not been investigated (Dahl et al.,1996).

During development, pancreatic endocrine cells derive from pancreas progenitors expressing Pdx1 (Gu et al.,2002) and Sox9 (Piper et al.,2002; Seymour et al.,2007), which then transiently express Neurogenin 3, a protein that is necessary for endocrine cell differentiation (Apelqvist et al.,1999; Gradwohl et al.,2000; Schwitzgebel et al.,2000; Grapin-Botton et al.,2001; Gu et al.,2002; Johansson et al.,2007; Desgraz and Herrera,2009). Similarly, neurons differentiate upon expression of Ngn1, 2, or 3 (Sommer et al.,1996; Ma et al.,1999; Hand et al.,2005; Ge et al.,2006). In the cortex, Ngn2 also promotes the migration of neurons through long distances to reach their final position and form the typical layered organization (Hand et al.,2005; Ge et al.,2006). In this tissue, although the binding of Ngn2 to DNA is required for inducing differentiation it is not necessary for migration (Hand et al.,2005; Ge et al.,2006). The migration program is activated as the RhoA GTPase, a regulator of actin dynamics, is repressed. The repression of RhoA by Ngn2 appears to rely on the displacement of the coactivator CREBS binding protein (CBP) from the RhoA promoter to that of NeuroD or other direct differentiation targets of Ngn2. We investigated whether Ngn3 similarly controls endocrine cell movement in the pancreas.

To better understand whether and how Ngn3 controls delamination of pancreatic endocrine cells, we analyzed in detail the different steps of pancreatic endocrine cell delamination. We found that Ngn3-positive endocrine progenitor cells lose apical contact, a feature previously reported as an early event in EMT. Subsequently, they exhibit all the features of classical EMT. Using the chicken as a model system, we showed that Ngn3 is sufficient to induce this EMT and that similar mechanisms occur in the mouse pancreas, under the control of Ngn3. Chick pancreas development is very similar to mice and humans although the ratios between insulin-positive and glucagon-positive cells are different in the three species (Dieterlen-Lievre and Beaupain,1974; Rawdon and Larsson,2000). Finally, we show that Ngn3 indirectly controls Snail2 expression, which likely leads to a subsequent EMT in endocrine cells.

RESULTS

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. RESULTS
  5. DISCUSSION
  6. EXPERIMENTAL PROCEDURES
  7. Acknowledgements
  8. REFERENCES
  9. Supporting Information

Ngn3-Positive Endocrine Progenitors Initiate Delamination From the Progenitor Layer

Although it is well known that endocrine cells delaminate from the epithelium to form islets, it is unclear at which step they do so. We hypothesized that this process may start in Ngn3-positive endocrine progenitors. We characterized the shape, location, and polarization of Sox9-positive pancreas progenitors, Ngn3-positive endocrine progenitors and endocrine cells in the embryonic day (E) 14.5 mouse epithelium (Fig. 1). We immunostained for the apical marker atypical PKC (aPKC), an essential component of the apical complex. The aPKC colocalizes with mucin at the apical membrane (data not shown, see also Kesavan et al.,2009, for active aPKC). Down-regulation of the aPKC/Cdc42/Par6 apical complex is known to lead to cell delamination (Georgiou et al.,2008). At E14.5, all of 801 Sox9-high multipotent progenitors sampled using thick confocal stacks in the centre of four independent pancreata were polarized, lining the lumen and expressing aPKC on their apical side (Fig. 1A,B). In contrast, we observed that the majority of Ngn3-positive cells were located at the edge of the epithelium (Fig. 1C). We found that, in E14.5 wild-type pancreas, 77.9 ± 2.9% of a total of 1,005 Ngn3-positive cells from 4 wild-type pancreata did not have apical aPKC localization demonstrating that these cells have started delamination from the progenitor layer (Fig. 1C–F). Moreover, of the 22% of Ngn3-positive cells that were still polarized, most exhibited very small aPKC area and were pear-shaped (Fig. 1D), suggesting that these cells are losing apicobasal polarity. At this stage, most glucagon- and insulin-positive cells resided in large clusters excluded from the epithelium (not shown). However many were observed in association with or in the epithelium either isolated or in small clusters (Fig. 1A,B). None of 927 glucagon- or insulin-positive cells scored in 4 pancreata in or abutting the epithelium showed aPKC at the membrane. These results show that a delamination program from the Sox9-positive progenitor layer is initiated in Ngn3-positive proendocrine cells and that delamination from the progenitor layer is completed by the time endocrine cells have differentiated. Then, we analyzed whether Ngn3-positive endocrine progenitors are still epithelial cells based on E-cadherin expression. Quantification showed that all Ngn3-positive cells were E-cadherin positive (total of 1,005 cells on n = 4 pancreas; Fig. 1C–E). Finally, we observed that Ngn3-positive cells do not express the mesenchymal marker Vimentin whereas endocrine cells are Vimentin-positive (Fig. 1G,H). Although Ngn3-positive endocrine progenitors have delaminated from the progenitor layer, they are still epithelial and have not delaminated from the epithelium and thus completed an EMT.

thumbnail image

Figure 1. Neurogenin3-positive endocrine progenitors lose apical polarity. A–E: Confocal optical sections from 50- to 60-μm-thick sections of embryonic day (E) 14.5 wild-type mouse pancreas. A: Sox9-positive cells (nuclear green) lining a duct exhibit aPKC staining (membrane green). Cell membranes are shown in red (E-cadherin). Although some Sox9-positive cells appear aPKC-negative on the optical sections displayed, they are aPKC-positive on the optical sections above or below the ductal lumen running toward the top right and bottom right corners. Two endocrine cells (blue) in the epithelium are devoid of aPKC. B: Cell membranes lined by E-cadherin (red) allows for cell boundary definition. C: Most Ngn3-positive cells are at the edge of the epithelium and negative for aPKC staining. D: Of the 22% of Ngn3-positive cells that are still polarized, most cells exhibit a very small aPKC area and are pear-shaped. E: A few cells were still fully polarized. Note that Ngn3-positive cells are E-cadherin–positive. Scale bar = 5 μm. F: Graph showing the results of quantification of aPKC expression and polarization of Ngn3-positive cells. Histograms present the average percentage of 1,005 Ngn3-positive cells over n = 4 embryos *P = 0.028. G,H: Costaining of Ngn3 and Vimentin on sections of E14.5 wild-type mouse pancreas shows that Ngn3-positive cells are not Vimentin-positive. Only endocrine cells are Vimentin-positive. Scale bar = 10 μm.

Download figure to PowerPoint

Ngn3 Independently Induces Differentiation and Delamination From the Epithelium

It has been shown already that overexpression of Ngn3 in the chicken endoderm induces pancreatic endocrine differentiation (Grapin-Botton et al.,2001). In this system, Ngn3 also triggered delamination out of the epithelium, migration in the mesenchyme and aggregation. To better mimic the transient expression of Ngn3 seen in pancreas progenitors, we designed a construct in which Ngn3 was flanked by two LoxP sites (Fig. 2A). Coelectroporating this construct with a plasmid expressing Cre recombinase downstream of a CMV promoter led to transient Ngn3 expression in the gut epithelium and the pancreas until Cre recombinase had accumulated enough to flox out Ngn3. This recombination event resulted in the expression of TauGFP, which was cloned downstream of the floxed Ngn3 (Fig. 2A). We analyzed the differentiation and delamination of electroporated cells 48 and 72 hr after electroporation. Most cells were found in the small intestine where the quantification was performed 72 hr postelectroporation and a smaller number in the pancreas. As shown before, sustained Ngn3 expression from this vector in the absence of Cre-induced delamination out of the epithelium and differentiation into endocrine cells (Fig. 2A; Supp. Fig. S1, which is available online; Table 1). Delamination out of epithelium was also observed in the pancreas upon sustained Ngn3 expression (Fig. 2A and S1), although the total number of cells targeting this organ was too low to perform quantifications of statistical significance. Transient Ngn3 overexpression was also able to induce differentiation into glucagon-expressing endocrine cells (Figs. 2A; Supp. Fig. S1; Table 1). However, they rarely delaminated out of the epithelium. No delamination from the epithelium was observed in electroporated cells in the pancreas (Fig. 2A, S1). This shows that sustained expression of Ngn3 is necessary to induce delamination from the epithelium. However, both sustained and transient overexpression lead to similar delamination from the progenitor layer, resulting in basally located, E-cadherin–positive cells that had lost apical contact (Table 1). These experiments also show that Ngn3 can induce a differentiation program without inducing the epithelium delamination program. Conversely, delamination without endocrine differentiation was also observed in chicken ectoderm (Fig. 2B). When pCIG5 was electroporated in the surface ectoderm, green fluorescent protein (GFP) -positive cells remained in the monolayered surface ectoderm. Conversely, when the vector expressing sustained Ngn3 was electroporated in the ectoderm, most electroporated cells delaminated from the surface ectoderm and invaded the underlying mesenchyme (Fig. 2B), without differentiating into any of the endocrine cell types (data not shown). However, they expressed Tuj1 (Fig. 2B) and NeuN (not shown), two neuronal markers. These results show that Ngn3 initiates a delamination program in multiple germ layers and this function is independent of its ability to trigger endocrine differentiation.

thumbnail image

Figure 2. Delamination and differentiation are two distinct programs that can be uncoupled. A: Schematic representation of mNgn3 expression vector electroporated in the chicken endoderm in presence or absence of CMV::Cre. The pictures show frozen transversal sections of the small intestine and pancreas, stained for E-cadherin (blue), green fluorescent protein (GFP; marking electroporated cells in the absence of cre due to coelectroporation of a GFP-expression vector or cells having floxed-out Ngn3 in the presence of Cre) and glucagon (red). Sustained expression of Ngn3 induces basal accumulation of targeted cells, their delamination out of the epithelium and differentiation of a subset of cells into glucagon-positive cells both in the small intestine and pancreas. Note that the high variation in GFP levels results in some GFP masking glucagon expression and some glucagon-positive red cells in which GFP cannot be distinguished. When a Cre-expressing vector, under the control of a CMV promoter is coelectroporated, Ngn3 expression is transient, and cells do not delaminate from the epithelium. However, they do differentiate and accumulate basally. B: Electroporation of mNgn3 in the surface ectoderm of chicken embryos. Pictures show transversal frozen sections of the gut stained for Tuj1 (red), DAPI (4′,6′-diamidino-2-phénylindole, nuclei, blue), and GFP (expressed from pCIG vector). Unlike pCIG empty vector, expression of Ngn3 induces delamination of electroporated cells from the surface ectoderm. These cells express the neuronal marker Tuj1. C: Electroporation of the Ngn3ER[RIGHTWARDS ARROW]AQ mutant construct in the chicken endoderm. Pictures show transversal frozen sections of the gut stained for glucagon or E-cadherin (red), DAPI (nuclei, blue) and GFP (expressed from pCIG vector). Ngn3ER[RIGHTWARDS ARROW]AQ induces neither endocrine differentiation nor delamination from the epithelium. Scale bars = 20 μm in A–C. l, lumen.

Download figure to PowerPoint

Table 1. Quantification of Delamination From Progenitor Layer and Epithelium, Glucagon Differentiation Upon Sustained and Transient Ngn3 Overexpression in the Chick Intestine
 Sustained Ngn3Transient Ngn3
  • Electroporation of the floxed-Ngn3 expression vector described in Figure 2A in the presence of Cre recombinase leads to transient overexpression, whereas in the absence of Cre recombinase, expression is sustained. The cells located in or outside the small intestine epithelium where quantified in the presence (n = 5) and the absence (n = 3) of Cre recombinase. On the same embryos, the cells only in contact with the basal lamina (no apical contact) were quantified. Their differentiation into glucagon cells was also measured. ns: the two values on the same row are not significantly different;

  • *

    : the two values on the same row are significantly different with P < 0.05.

% cells out of epithelium30.7 ± 10.15.6 ± 4.0*
% basal among epithelial40.3 ± 11.250.4 ± 5.9 ns
% glucagon-positive8.5 ± 3.957.4 ± 29.1*

Finally, we also analyzed if the transcriptional activity of Ngn3 was required to induce the endocrine differentiation or delamination programs. It was described that mutations in the basic domain of Ngn1 and 2 transcription factors by the substitution of two amino acids at positions 123–124 (NR to AQ) resulted in the translation of a protein that cannot bind the E-boxes and, therefore, is transcriptionally inactive (Sun et al.,2001; Lee and Pfaff,2003; Hand et al.,2005). Yet it is still localized in the nucleus (Hand et al.,2005). We similarly cloned a mutant form of Ngn3 in which amino acids 92 and 93 of the basic domain were mutated from ER to AQ. When this mutant construct was overexpressed in the developing chicken endoderm, cells did not differentiate into endocrine cells although the protein was expressed. Moreover, they did not delaminate out of the epithelium (Fig. 2C). These results show that transcriptional activity of Ngn3 is required to control both differentiation and delamination programs.

Ngn3 Induces an EMT in Endocrine Cells

As we observed that Ngn3-positive cells were delaminating, we hypothesized that Ngn3 induced an EMT in proendocrine cells. To test this hypothesis, we overexpressed Ngn3 in the developing chicken endoderm (Fig. 3). As it was shown before, overexpression of Ngn3 in the endoderm was sufficient to trigger pancreatic endocrine differentiation (Fig. 3A). We observed that 48 hr after Ngn3 overexpression electroporated cells down-regulated E-cadherin before delaminating from the epithelium (Fig. 3B). They delaminated out of the epithelium by breaking down the basal lamina (Fig. 3C). These events were observed as early as 28 hr after electroporation (data not shown). Electroporated cells also up-regulated the intermediate filament Vimentin (Fig. 3D) and lost the typical epithelial apical and basal distribution of actin cytoskeleton, another hallmark of EMT (Fig. 3E). Taken together, these results show that overexpression of Ngn3 is sufficient to induce an EMT in the chicken endoderm, and, therefore, delamination out of the epithelium.

thumbnail image

Figure 3. Electroporation of Ngn3 alone in the chicken endoderm induces an epithelial-to-mesenchymal transition. Frozen transversal sections of the chicken gut 48 hr after electroporation of a Ngn3-expressing vector stained for DAPI (4′,6′-diamidino-2-phénylindole nuclei, blue), green fluorescent protein (GFP; expressed from pCIG::Ngn3 vector) and EMT markers (red). A: Electroporation of Ngn3 induces endocrine differentiation (glucagon; red). B: Before delaminating from the epithelium, Ngn3-electroporated cells strongly decrease E-cadherin levels. C: When electroporated cells delaminate, they break down basal lamina. D: Electroporated cells up-regulate the mesenchymal marker Vimentin. E: While delaminating from the epithelium, cells exhibit a loss of the apical and basal distribution of cortical actin. Scale bar = 50 μm in A–D, 25 μm in E.

Download figure to PowerPoint

We subsequently investigated whether in the mouse pancreas Ngn3 could also induce an EMT. To be able to observe groups of Ngn3-positive cells, we initially used a transgenic mouse in which the Pdx1 promoter drives a Ngn3-ER fusion protein that can be activated by tamoxifen injections (Fig. 4A). This mouse line was shown to expand Ngn3 expression in Pdx1-positive progenitors within 24 hr of tamoxifen injection while retaining endogenous expression levels (Johansson et al.,2007). In the absence of tamoxifen few endocrine cells differentiated, as in wild-type mice (Johansson et al.,2007). We injected tamoxifen at E9.5 and dissected embryos 48 hr after injection (E11.5). At this stage, we observed that endocrine cells had differentiated and had down-regulated E-cadherin (Fig. 4B). Endocrine cells broke down basal lamina while delaminating from the epithelium (Fig. 4C). Moreover, they also up-regulated the mesenchymal marker Vimentin (Fig. 4D) and lost the typical epithelial apical and basal distribution of actin cytoskeleton (Fig. 4D). These events were also observed in wild-type pancreas. Indeed at E14.5, endocrine clusters exhibited low E-cadherin staining, breakdown of basal lamina (Supp. Fig. S2) and expression of mesenchymal markers such as Vimentin (Fig. 1G,H) and N-cadherin (not shown). As shown in Figure 1C–E, Ngn3-positive cells were still E-cadherin–positive. Moreover, Ngn3-positive endocrine progenitor cells were never Vimentin positive (Fig. 1G,H). Taken together, these results show that endocrine cells have lost multiple epithelial characters, although they still remain in close contact with the pancreatic epithelium. They achieve an EMT to delaminate out of the epithelium.

thumbnail image

Figure 4. Ngn3 induces epithelial-to-mesenchymal transition (EMT) of endocrine cells in the mouse pancreas. B–D: Immunostaining on embryonic day (E) 11.5 of the pancreas of the transgenic mouse line Pdx1::Ngn3ERTM represented in (A) 48 hr after tamoxifen injection. Transversal sections were immunostained for the different EMT markers. Newly differentiated endocrine cells exhibit all the signs of an EMT: loss of E-cadherin, breakdown of the basal lamina (Laminin) and up-regulation of the mesenchymal marker Vimentin. Scale bar = 50 μm.

Download figure to PowerPoint

Loss of E-cadherin Is Sufficient to Trigger Delamination Out of the Endodermal Epithelial Layer

Our observations showed that the down-regulation of E-cadherin was an early event during Ngn3-induced EMT. We investigated whether E-cadherin down-regulation was sufficient to induce delamination of endocrine cells. To achieve this aim, we electroporated a dominant-negative form of E-cadherin in chick endoderm (Fig. 5C). Cells electroporated with an empty pCIG5 alone could not exit the epithelium, whereas 35% of the cells electroporated with dominant-negative E-cadherin could delaminate from the epithelium as early as 28 hr after electroporation (average value on n = 4 embryos; Fig. 5A,B). Therefore, E-cadherin down-regulation was sufficient to drive delamination out of the epithelium.

thumbnail image

Figure 5. Loss of E-cadherin function is sufficient to trigger delamination out of the epithelium. A: Immunostainings on frozen transversal sections of chicken endoderm 28 hr after electroporation with a plasmid coding for a dominant negative form of E-cadherin. E-cadherin staining (red) and green fluorescent protein (GFP) fluorescence allow visualization of delamination of electroporated cells out of the epithelium. Note that the E-cadherin antibody targeted to the cytoplasmic part of the protein also marks electroporated cells. The white dotted lines show the boundary between epithelium and mesenchyme. B: Quantification of cell delamination out of the epithelium. Cells were scored for being in or out of the epithelium (n = 4 embryos, ***P = 0.0001). C: Schematic representation of dominant negative E-cadherin construct (adapted from Dahl et al.,1996). This construct was cloned in the expression vector pCIG, allowing tracing of GFP-positive electroporated cells. SP, signal peptide; PP, prepeptide; EC, extracellular domain; TM, transmembrane domain; CP, cytoplasmic domain. Scale bar = 50 μm in A.

Download figure to PowerPoint

E-Cadherin Is Transcriptionally Regulated in Endocrine Cells

As we observed a decrease of immunofluorescent staining of E-cadherin in endocrine cells, we investigated at which level E-cadherin was down-regulated. For that purpose, we monitored mRNA levels of E-cadherin in the wild-type mouse pancreas at E14.5 by in situ hybridization (Fig. 6A). We observed a marked reduction of E-cadherin mRNA in glucagon-positive cells, showing that E-cadherin was strongly down-regulated at the transcriptional level in these cells.

thumbnail image

Figure 6. E-cadherin is transcriptionally regulated in endocrine cells. A: Transverse paraffin sections of embryonic day (E) 14.5 wild-type pancreas stained by in situ hybridization to detect E-cadherin mRNA expression pattern (blue) and by immunofluorescence for glucagon-positive endocrine cells. The E-cadherin mRNA levels are strongly reduced in glucagon-positive cells. B: Graph showing representative results from n = 3 luciferase assay experiments each containing triplicates of each condition in NMuMG cells. To address the role of Ngn3 on E-cadherin repression, the 364-bp wild-type promoter element of E-cadherin driving the firefly luciferase was cotransfected with an empty vector (pMES), Ngn3 (pMES::Ngn3), or Snail1 and Snail2 (pCX::Snail1 and pCX::Snail2), two known repressors of E-cadherin. The levels of E-cadherin transcription was addressed either 24 hr (white bars) or 48 hr (black bars) after transfection. Snail1 and Snail2 can both decrease E-cadherin transcription whereas Ngn3 was not able to directly repress E-cadherin transcription when compared with empty vector control. ns: not significant, ***P < 0.0001. Scale bar = 50 μm in A.

Download figure to PowerPoint

Because Ngn3 was expressed since E-cadherin down-regulation and because another basic helix–loop–helix transcription factor, Acheate-Scute homologue 1a (Ash1a), had been shown to directly repress E-cadherin in vitro (Osada et al.,2008), we tested whether Ngn3 could directly bind to the E-cadherin promoter and repress its transcription. Indeed, the E-cadherin promoter contains three E-boxes (Batlle et al.,2000; Osada et al.,2008): one in the first intron and two upstream of the transcription start site (−90 and −70 base pairs). The two latter have been described to mediate E-cadherin repression by Snail1 and Snail2 (Batlle et al.,2000). We cotransfected plasmids in which the E-cadherin promoter drives firefly luciferase expression together with Ngn3 in epithelial NMuMG cells (Fig. 6B). In the absence of known pancreas progenitor cell lines, this epithelial cell line was chosen based on its robust endogenous E-cadherin expression and its common use as an EMT model (Miettinen et al.,2000; Hirano et al.,2008). Known repressors of E-cadherin such as Snail1 and Snail2 could efficiently repress transcription of E-cadherin when compared with control transfection (Fig. 6B). However, cotransfection of Ngn3 with the wild-type E-cadherin promoter did not modify the levels of transcription of E-cadherin (Fig. 6B). With a possibility that Ngn3 may control more distant E-cadherin regulatory elements, these results suggest that Ngn3 does not directly block E-cadherin expression or may require convergent activity of factors that are absent in NMuMG cells.

Ngn3 Controls Snail2 Expression in the Pancreas

It has been shown that Snail2 is expressed in the pancreas in a subset of Ngn3-positive cells and in a subset of hormone-positive cells (Rukstalis and Habener,2007). To investigate whether Snail2 expression requires Ngn3, we analyzed its expression in the Ngn3 knock-out pancreas by immunofluorescence. As published before, in wild-type control embryos Snail2 was expressed in a salt-and-pepper pattern resembling that of Ngn3 (Rukstalis and Habener,2007). As predicted for a target of Ngn3, Snail2-positive cells were located at the edge of the epithelium (Fig. 7A). In Ngn3 knock-out pancreas, the nuclear staining of Snail2 was completely lost, showing that Ngn3 controls Snail2 protein levels.

thumbnail image

Figure 7. The epithelial-to-mesenchymal transition (EMT) inducer Snail2 is expressed in the pancreas and requires Ngn3-expression. A: Transverse frozen section of embryonic day (E) 14.5 wild-type versus Ngn3 knock-out mouse pancreas immunostained for E-cadherin (green) and Snail2 (red). The pictures show that the specific nuclear staining of Snail2 is completely lost in Ngn3 knock-out. B: Adjacent sections were stained with Snail1, Snail2 (antisense and sense) and Ngn3 RNA probes. Immunostaining for E-cadherin (red) and glucagon (green) was performed after the in situ hybridization, and was used to localize pancreatic epithelium and identify endocrine cells. In wild-type embryos, Snail1 and Snail2 are expressed in the epithelium, but much lower than in the mesenchyme. In Pdx1::Ngn3ER™ ; Ngn3−/− transgenic pancreas, Snail2 is not up-regulated. Dotted lines show the boundary between epithelium and mesenchyme. Scale bars = 50μm. dp, dorsal pancreas; g. ep., gut epithelium. C: Quantitative real-time polymerase chain reaction to address the relative levels of Snail1, 2, and 3 mRNAs in known expression tissues and compared with the expression levels in the pancreas. Snail1 and 2 are known to be expressed in numerous tissues such as neural crest cells (NCC), frontonasal region (FRN), developing limbs and heart during development (Nieto et al.,1992; Oram et al.,2003). Snail3 was described to be expressed at high levels in embryonic skeletal muscles and postnatal thymus (Zhuge et al.,2005). As negative control tissues, we used the liver for Snail1 and 2, and the lungs for Snail3 (Nieto et al.,1992; Oram et al.,2003; Zhuge et al.,2005). mRNAs were collected from tissues dissected from n = 3 embryos at least and pooled for cDNA synthesis. The mRNA level in the E11.5 wild-type pancreas is arbitrary set as one. D: Analysis of the mRNA levels of Snail1 and 2 in Ngn3 wild-type, knock-out and rescued mouse embryonic pancreas at E11.5 and E15.5. The analysis of mRNA levels shows that there are no significant changes in the expression of Snail genes depending on Ngn3 expression levels. Errors bars represent standard deviation between triplicates; ns: not significant.

Download figure to PowerPoint

To investigate whether Snail2 mRNA expression was controlled by Ngn3 and whether other snail family genes may also play a redundant function, we used quantitative reverse transcriptase-polymerase chain reaction (RT-PCR) to compare the levels of Snail1, 2, and 3 mRNAs in the pancreas to known expression sites and negative control tissues (Nieto et al.,1992; Oram et al.,2003; Zhuge et al.,2005). We found that Snail1 and 2 were both expressed in the E11.5 and E15.5 pancreas at similar levels as compared to neural crest cells, frontonasal region, developing limbs and hearts, whereas Snail3 was expressed at very low levels when compared with known expression sites (Fig. 7C). We observed no significant changes in Snail1 and 2 mRNA levels in Ngn3 knock-out pancreas, either at E11.5 or E15.5 (Fig. 7D). These results show that the regulation of Snail2 by Ngn3 is post-transcriptional. We also monitored the mRNA levels of Snail1 and Snail2 by in situ hybridization (Fig. 7B). Both Snail1 and 2 mRNAs were expressed at very low level in the wild-type pancreatic epithelium and mesenchyme at E11.5 (Fig. 7B, top panels) and E14.5 (data not shown). Levels in control tissues are shown for comparison in Supp. Fig. S3. No punctae of expression were seen, suggesting that these mRNAs were not enriched in endocrine cells or endocrine progenitors. Furthermore, no up-regulation of either Snail1 or Snail2 was detected by in situ hybridization in endocrine cells or when Ngn3 was misexpressed throughout the pancreatic epithelium (Fig. 7B, bottom panels). These results suggest that Snail2 is expressed at low levels throughout the pancreas and that an indirect post-transcriptional regulation by Ngn3 increases Snail2 protein levels in endocrine progenitors and drives EMT of endocrine cells.

Snail2 Is Sufficient to Promote EMT in the Chicken Endoderm

Finally, we tested if Snail2 is sufficient to induce EMT in the chicken endoderm. We coelectroporated the Snail2 gene together with pCIG5 to trace GFP-positive electroporated cells and analyzed embryos 28 (data not shown) and 48 hr (Fig. 8) after electroporation. We observed that electroporated cells decreased E-cadherin as early as 28 hr after electroporation (data not shown). However, the number of cells out of the epithelium was low, suggesting that the delamination may still be in progress. Forty-eight hours after electroporation, the decrease of E-cadherin was more obvious (Fig. 8A), and we observed cells delaminating out of the epithelium after breaking down the basal lamina (Fig. 8B). This was observed in gut endoderm (n = 4) as well as in the pancreas (n = 4). None of the Snail2-electroporated cells in the duodenum expressed glucagon, showing that delamination is not sufficient to turn on endocrine differentiation. These results show that Snail2 is sufficient to induce EMT in the chicken endoderm and, therefore, delamination out of the epithelium. Snail2 likely plays a similar role in the endogenous, developing pancreatic endocrine cells.

thumbnail image

Figure 8. Overexpression of Snail2 in the endoderm is sufficient to induce epithelial-to-mesenchymal transition (EMT). Transversal frozen sections of chicken gut coelectroporated with pCX::Snail2 and pCIG immunostained for E-cadherin (A) or Laminin (B) (red). Green fluorescent protein (GFP) expression allows tracing of electroporated cells and DAPI stains nuclei (blue). Forty-eight hours after electroporation, Snail2 repressed E-cadherin in the chicken gut (A) and induces delamination of electroporated cells (A and B). These cells brake down the basal lamina (B). The white dotted lines show the boundary between epithelium and mesenchyme. Scale bar = 50 μm.

Download figure to PowerPoint

DISCUSSION

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. RESULTS
  5. DISCUSSION
  6. EXPERIMENTAL PROCEDURES
  7. Acknowledgements
  8. REFERENCES
  9. Supporting Information

Ngn3 is a key transcription factor that initiates pancreatic endocrine differentiation during development. Here, we show that Ngn3 triggers the first step of EMT, namely apical constriction and loss of apical contacts. The first detectable morphogenetic events are seen in cells that express Ngn3 but not yet hormones. We demonstrate that the transient Ngn3-expression wave occurring in endocrine progenitors is sufficient to trigger delamination out of the progenitor layer but not out of the epithelium. Epithelial delamination is only seen when Ngn3 expression is artificially maintained. It was shown by a single-cell transcript analysis that most Ngn3-positive cells express Vimentin and E-cadherin mRNAs (Chiang and Melton,2003). Although the Vimentin transcript is already present in these cells, our experiments reveal that the protein is not detectable in Ngn3-positive cells but is evident in hormone-expressing cells. Ngn3-positive cells are still epithelial cells expressing E-cadherin. E-cadherin is known to be a very stable protein and has to be regulated both at the protein and the transcriptional levels to efficiently promote EMT (Arnold et al.,2008; Hirano et al.,2008; Zahn et al.,2008). Therefore, subtle changes in E-cadherin may have already started but are not detectable yet. All endocrine cells, both isolated in the epithelium and clustered in emerging islets, express Vimentin and low levels of E-cadherin, two hallmarks of EMT. This demonstrates that the E-cadherin decrease and the Vimentin up-regulation occur before the aggregation of endocrine cells. We further show that the decrease in E-cadherin is sufficient to promote epithelial escape. In the pancreas, islets never migrate long distances away from the epithelial ducts. This may be due to the fact that E-cadherin is never completely lost. This lowering in E-cadherin levels may also facilitate endocrine cell clustering and segregation from other epithelial progenitors. Indeed it was previously shown that homotypic adhesion occurs preferentially in cells with similar level and profile of cadherin expression (Nose et al.,1988; Steinberg and Takeichi,1994). The maintenance of low level of E-cadherin at all stages, even in endocrine cells, suggests a partial EMT. This is in agreement with previous studies on the role of cadherins during islet formation (Dahl et al.,1996; Johansson et al.,2010). Indeed, it was shown that E-cadherin function is required for proper aggregation of endocrine cells after delamination (Dahl et al.,1996).

In other systems, EMT can be triggered by various signaling pathways (Jamora et al.,2003; Nawshad and Hay,2003; Medici et al.,2006), but these pathways usually converge on zinc finger transcription factors of the Snail family which directly control downstream effectors. In the pancreas, Snail2 may be the transcription factor that induces EMT downstream of Ngn3. Indeed, when Ngn3 is absent, Snail2 nuclear protein is lost. However, we did not observe changes at the mRNA level by quantitative PCR, a result further supported by the different microarrays performed in our laboratory comparing Ngn3−/− and Ngn3−/− pancreata rescued with the Pdx1::Ngn3ERTM transgene (unpublished data). These results suggest that a target of Ngn3 that should control protein expression, stability or nuclear localization of Snail2. Snail2 is a highly unstable protein with a half-life of 8 to 10 hr. It is believed that during neural crest cell migration, degradation of Snail2 is controlled by polyubiquitination of the N-terminal part (Vernon and LaBonne,2006). In addition, phosphorylation sites have been identified in Snail1 that control the nuclear localization and activity in both positive (Yang et al.,2005) and negative (Dominguez et al.,2003; Zhou et al.,2004; Yook et al.,2005) manners. However, similar sites do not appear to regulate Snail2 stability in vivo in Xenopus (Vernon and LaBonne,2006). Whether Ngn3 induces a protein that similarly affects Snail2 post-transcriptional modifications remains to be elucidated.

In this study, we also show that Ngn3 independently governs the endocrine differentiation and delamination programs. Indeed, transient overexpression of Ngn3 in the chick small intestine and pancreas does not induce delamination of endocrine cells from the epithelial layer whereas cells still differentiate. Furthermore, while analyzing the differentiation and delamination of Ngn3-electroporated cells, we observed that the cells that differentiate are usually those exhibiting the lowest GFP intensity or the most transient expression. Either the duration of Ngn3 expression, or the amount of protein produced in a cell is important to control the delamination program. This hypothesis is in agreement with a recent article demonstrating different transcriptional activities by different deletion constructs of Ngn3 (Rosenberg et al.,2010). Although we show that endocrine differentiation and delamination are two concurrent processes during normal pancreas development, they are independent. In the developing cortex, Ngn2 has also been shown to control migration independently of differentiation, as the two programs can be uncoupled (Hand et al.,2005; Ge et al.,2006). Some targets that are activated are transcriptionally controlled by Ngn2 binding to E-boxes in their promoter (Ge et al.,2006). This is the case for the differentiation program exemplified by the activation of NeuroD as well as for migration target genes such as Dcx (Ge et al.,2006). However, other targets such as RhoA must be down-regulated to allow migration of newly generated neurons. Repression of RhoA is achieved independently of E-boxes, the natural Neurogenin binding elements, through recruitment of CBP by Ngn2 on other direct target gene promoters, leading to a displacement of the transcriptional activator from the RhoA promoter (Ge et al.,2006). It was shown that phosphorylation of Ngn2 at tyrosine 241 controls the migration program in vivo (Hand et al.,2005). However, such a phosphorylation site is absent in Ngn3. In the pancreas, we show that E-box–mediated transcriptional activity is required to induce both differentiation and delamination of endocrine cells. This is in agreement with Rosenberg et al. who recently used another Ngn3 mutant abolishing binding to E-boxes (Rosenberg et al.,2010). The similarities with the cortical neuron migration program may be limited as these cells never delaminate from the epithelium and, therefore, may not engage in EMT before migrating longer distances.

This study reveals that delamination of cells on the endocrine differentiation pathway progresses in a stepwise manner (Fig. S4). At the stage of Ngn3 expression, cells become bottle-shaped and lose apical contact before they express hormones. At this stage, it was previously shown that Snail2 expression is turned on. Snail2 protein is indeed found in a subset of Ngn3-positive and a subset of hormone-expressing cells (Rukstalis and Habener,2007). Because transient expression of Ngn3 is not sufficient to induce E-cadherin repression and epithelial delamination, it is likely that in the developing pancreas other proteins synergize with Ngn3 to induce Snail2 or repress E-cadherin. NeuroD, which can trigger delamination in the chick assay, may be one of them. Snail2 is a potent E-cadherin repressor. E-cadherin is indeed transcriptionally repressed by the time hormones are expressed. We have shown that the decrease of E-cadherin is sufficient to trigger delamination out of the epithelium. This delamination is accompanied by other hallmarks of EMT: breakdown of the basal lamina next to endocrine cells and induction of Vimentin and N-cadherin. Committed endocrine cells are, therefore, released in the mesenchyme. We identified transcription factors that coordinate delamination: Ngn3 initiates it and Snail2 is a trigger of EMT effectors. We further revealed that Ngn3 indirectly controls Snail2 nuclear protein but not RNA. Finally, we clarified the role of some of these effectors: E-cadherin down-regulation is sufficient to trigger delamination from the epithelium. This study establishes a framework for the identification of other molecules that control each of the delamination steps and the future analysis of delamination by time-lapse microscopy.

EXPERIMENTAL PROCEDURES

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. RESULTS
  5. DISCUSSION
  6. EXPERIMENTAL PROCEDURES
  7. Acknowledgements
  8. REFERENCES
  9. Supporting Information

Mice

For tamoxifen treatment in vivo, pregnant females were injected intraperitoneally with 200 μl of a 10 mg/ml TM (Sigma) solution in corn oil (Sigma). For genotyping of Pdx1: :Ngn3ERTM transgenic animals, tail biopsies were lysed in Direct PCR Lysis Reagent (Viagen Biotech) and analyzed by PCR, with the (5′-caccatggctcctcatcccttgg-3′) and (5′-cctcatg tctcctgaagc-3′) primers. For the Ngn3 knock-out locus, the primers were Ngn3 5′ (5′-tctcgcctcttctggctttc-3′); Ngn3 3′ (5′-cggcagatttgaatgagggc-3′) to detect the wild-type allele and neo 5′ (5′-gcagcgcatcgccttctatc-3′) and Ngn3 3′ to detect the mutant allele. All three primers were included in a single PCR reaction.

In Ovo Electroporation

Fertilized White Leghorn chicken eggs (E. Pavillard, Orny, Switzerland) were incubated at 38°C in a humidified atmosphere. Electroporation was performed on HH13 embryos, corresponding to approximately 60 hr of incubation. A hole was made in the shell on the side of the air chamber and a window was then cut on top of the egg. A solution of 2 μg/μl DNA in 1× phosphate buffered saline (PBS), 1 mM MgCl2, 3 mg/ml carboxymethylcellulose (Sigma), and 50 μg/ml Nile Blue Sulfate (Applichem), was injected in the yolk, in close contact with the endoderm. 1× PBS was added to the top of the embryo to prevent the embryo from drying and to establish an even electrical current. Tungsten electrodes were placed under (negative) and above (positive) the embryo, along the anteroposterior axis. Three square pulses of 15 volts and 50 ms each were applied to the embryo (BTX ECM830 Electro Square Porator). After electroporation, windows were closed with tape, and reincubated at 38°C from 28 hr to 72 hr before fixation.

Constructs

The lox-Ngn3 construct was engineered by inserting the PCR amplified mNgn3 in the lox-MCS containing pBS242 vector from Ivan Rodriguez and subsequently subcloned upstream of the coding sequence of tauGFP obtained in p313 (gift from I. Rodriguez). The insert was eventually cloned into pCAGGS. Ngn3ER92AQ was mutated by recombinant PCR and the open reading frame inserted in pCIG using the Gateway system. The PCIG-Gateway-Ngn3 was functionally checked by electroporating chicken endoderm, which resulted in glucagon-cell differentiation and delamination. In similar experiments pCIG-Gateway-Ngn3ER92AQ expressed Ngn3 protein but did not trigger differentiation nor delamination. The E-cadherin dominant negative construct is a gift from J. Johansson and H. Semb (Dahl et al.,1996). The original construct was made by deleting most of the extracellular domain of mouse E-cadherin cDNA by digesting with EagI and BstEII restriction enzymes. A Myc-Tag was added at the 5′ of the construct, and cloned in pBS. This vector was digested with EcoRI and blunted before ligation into pCIG5. The ligation products were transformed in DH5a competent bacteria. Clones were verified by digestion and sequenced (Microsynth AG).

Snail1 and Snail2 constructs were gifts from Angela Nieto (Sefton et al., 1998). The following constructs were used for electroporation or probe synthesis: pCX::cSnail2; pCX::cSnail1; pcDNA3::mSnail2; pcDNA3::mSnail1. PCIG::mNgn3 construct was made from a PCR amplification of the mouse coding sequence and cloned into pCIG5. PMES and pMES::mNgn3 were gifts from Gérard Gradwohl. In electroporation tests, we ascertained that pMES::mNgn3 induced endocrine differentiation and delamination as pCIG::mNgn3.

Immunofluorescence

Mouse or chicken embryos were fixed in 4% paraformaldehyde (PFA) during 30 min at room temperature. After several washes in 1× PBS, embryos were equilibrated in 15% sucrose (Applichem) in Phosphate Buffer 0.12 M before embedding in 7.5% gelatin (Sigma), 15% sucrose in Phosphate Buffer 0.12 M. Gelatin blocks were subsequently frozen in −65°C 2-Methylbutane (isopentane, Acros Organics;) and kept at −80°C till sectioning. Frozen sections were then dried at room temperature, and washed and pretreated before blocking. Primary antibodies were incubated overnight at 4°C. The following primary antibodies were used : Glucagon (1:100; Zymed, 18-0064 and 1:400; Linco, 4031-01F), Insulin (1:100; Dako, A0564), E-cadherin (1:100; BD Transduction Lab, C20820), Laminin (1:150; Sigma, L9393 and 1:150; DSHB, 3H11), Vimentin (1:100; NeoMarkers, MS-129), Phalloïdin-Alexa 488 (1:100; Molecular Probes, A12379), Phalloïdin-Texas Red (1:100; Molecular Probes, T7471), aPKC (1:600; Santa Cruz, c-20 SC-216), Ngn3 (1:500; Beta Cell Biology Consortium, 2011), Snail2 (1:500; gift from M. Rukstalis), Sox9 (1:500, Chemicon AB5809), Tuj1 (1:100, R&D MAB1195). Secondary antibodies were incubated 1 hr at room temperature. Alexa Fluor secondary antibodies (all from Molecular Probes - Invitrogen) were used for multicolor detection. To stain nuclei the following chemicals were used: 4′,6′-diamidino-2-phénylindole (DAPI, 50 ng/μl; Sigma) and Draq5 (1:500; Biostatus). For whole-mount immunofluorescence, dissected E14.5 wild-type pancreata were fixed in 4% PFA for 5 min at 4°C and washed several times in 1× PBS. Tissues were dehydrated in methanol series and kept at −20°C until staining was performed. After rehydration in methanol series, tissues were permeabilized in 1× PBS + 0.2% Triton X-100 for 10 min and blocked in 1% bovine serum albumin (BSA) in 1× PBS, 0.2%–0.5% Tx100 at 4°C overnight. Primary antibodies were diluted in blocking solution and tissues were incubated for 24 hr at 4°C. Tissues were washed extensively in 1× PBS, 0.2% Tx100 at 4°C all day or overnight. Secondary antibodies were incubated in blocking solution at 4°C overnight. Stained tissues were washed extensively in 1× PBS, 0.2% Tx100 at 4°C. Nuclei were stained with DAPI or Draq5 for 2–3 hr. After washing in 1× PBS, tissues were post-fixed then washed in PBS and mounted in Dako fluorescent mounting medium. All fluorescence microscopy images for sections were taken with a Zeiss Axioplan 2 or a Leica DM5500 imaging motorized upright microscope, or a Leica SP2 inverted confocal microscope. Images were then treated with Adobe Photoshop CS4: light and contrast were adjusted (levels), pictures were reframed and resolution changed to match figure size. All observations were realized on at least n = 4 embryos.

In Situ Hybridization

To synthesize RNA probes, plasmids were linearized using appropriated enzymes, and RNA polymerases T3, T7 (Promega), and SP6 (New England Biolabs) were used, in presence of UTP-digoxigenin (Roche). The following clones were used to synthesize RNA probes: mNgn3 (Anne Grapin-Botton), mSnail2, and mSnail1 (gifts from A. Nieto; Sefton et al., 1998), mE-cadherin (gift from L. Larue). Mouse embryos were fixed in 4% PFA overnight at 4°C. After several washes, they were dehydrated in ethanol and xylene series before embedding in paraffin. Paraffin blocks were sectioned at 4 to 8 μm thickness. Slides were dewaxed in xylene, rehydrated in ethanol series and 1× PBS, and pretreated with 0.2 N HCl (15 min, room temperature), proteinase K in 1× PBS (time and temperature depending on the embryonic stage analyzed), and 0.6% anhydride acid in triethanolamine solution pH 8.0 (two times 5 min at room temperature). After blocking in hybridization mix (50% formamide, 5× standard saline citrate pH 4.5, 2% Blocking Solution [Roche], 5 mM ethylenediaminetetraacetic acid, 0.05% Chaps, 50 μg/ml heparin, 1 μg/ml yeast total RNA), the slides were incubated overnight with a 1 μg/ml RNA probe in hybridization mix. Slides were washed and incubated with anti-digoxigenin-alcaline phosphatase antibody (1:2,000, Roche; #11-093-274-910). Staining was developed with a 4.5 μl/ml nitro blue tetrazolium solution (NBT stock: 75 mg/ml in nanopure water, Eurobio) and 7 μl/ml of BCIP (5-bromo-4-chloro-3-indolyl phosphate) solution (BCIP stock: 25 mg/ml in nanopure water, Biotium) and slides mounted in 50% glycerol in 1× PBS. Images were acquired with a Leica DM5500 wide field microscope. All observations were realized on at least n = 4 embryos.

Quantification of Cell Polarization in Three Dimensions

E14.5 pancreatic tissue from ICR mice was stained using the whole-mount protocol described above and then sectioned (50–60 μm) with a Vibratome (Leica VTS 1200). Confocal imaging was done using Zeiss LSM710 or Leica SP5 White Laser microscopes. The ×63 objectives were used with ×1.3 zoom, and z-stacks were taken with 0.25-μm interval. Cell counting was done in a reconstructed 3D using IMARIS software (Bitplane AG, Zurich, Switzerland). Cells of interest (Ngn3-positive, Sox9-positive, insulin-positive, and glucagon-positive) were marked with a measurement tool in a plane using oblique slicer. Cell borders were identified by an epithelial membrane marker, E-cadherin and the apical membrane was detected using aPKC. We ascertained that each cell was scored only once by checking in three dimensions (3D) and that all cells were counted. Because the number of Sox9-positive cells is high, we could not quantify all of them but proceeded to avoid sampling biases. We chose a confocal stack at the center of the organ and a second one 20 μm below to avoid counting a cell twice. In these stacks all Sox9-positive cells with high levels were scored for aPKC membrane localization. Sox9-low cells which can be either Ngn3-positive or cells differentiating into acinar cells at terminal end buds were not counted. To quantify the polarity of endocrine cells, all glucagon- or insulin-positive in or abutting the epithelium were quantified in four pancreata. The majority of cells were already in large clusters outside of epithelium and were not counted.

Quantitative Real-Time PCR

Tissues were dissected in 1× PBS and frozen in RLT lysis buffer from Qiagen RNeasy Protect mini kit (Qiagen; #74124) and kept at −80°C. Samples were then thawed and mechanically disrupted on ice. Total RNA were extracted using the RNeasy Protect Miniki; Qiagen. If needed, RNAs were concentrated using the following protocol: 5 μg of linear acrylamide (Ambion; #9520), 1/10 of 3 M Sodium Acetate pH 5.2 (Acros Organics; #21711-0010) and 2.5 volumes of 100% ethanol were added to the samples and incubated at −20°C overnight. Samples were centrifuged at 12,000 × g, 4°C during 30 min and washed in cold 70% ethanol. The pellets were resuspended in nanopure water on ice for 3 hr. Concentration and quality of RNA was evaluated with a NanoDrop Spectrophotometer ND-100 and Agilent 2100 Bioanalyzer using Pico Chips (Agilent RNA 6000 Pico Chips). Reverse transcription of cDNA was carried out using random primers (Promega) and MMTV Reverse transcriptase enzyme (Invitrogen). For each tissue analyzed, 1 μg of total RNA was used. The qPCR was carried out using the Power Sybr Green mix (Applied Biosystems). To quantify specific gene expression levels, the following primer pairs were used: mSnail1: For – gctgcttcgagccatagaac ; Rev – ctcaaagaaggtggcctgaa; mSnail2: For – tctgcagacccactctgatg; Rev – agcagccagactcctcatgt; mSnail3: For – actggagacactgagagaagc; Rev – gtagggggtcactgggattg; E-cadherin: For – gtctaccaaagtgacgctgaa; Rev: gggtacacgctgggaaacat. The qPCR was made with the Step one plus real time PCR systems from Applied Biosystems, samples were analyzed as triplicates and expression levels were calculated with the manufacturer's software using the ΔΔCt method and Microsoft Excel. When testing each probe for the first time, qPCR products were run on gel by electrophoresis to address the specificity of amplification and the size of amplicons.

Cell Culture, Cell Transfection, and Luciferase Assay

NMuMG cells (gift from Curzio Rüegg) were grown in Dulbecco's Modified Eagle Medium (DMEM) supplemented with 10% fetal calf serum (FCS), 1% Fungizone, 1% penicillin/streptomycin (all from Gibco). Cells were split every 3 to 4 days depending on the growth rate. For transfections, 50,000 cells were plated in 24-well plates (Costar; #3524), and grown for 24 hr before transfection. Transient transfection was carried out with the TransFast Transfection Reagent (Promega; #E2431). A total amount of 1 μg of DNA was transfected, using 3 μl of TransFast reagent. Efficiency of transfection was analyzed 24 or 48 hr after transfection using a Leica DC200 inverted microscope. The E-cadherin wild-type promoter constructs were a gift from Takashi Takahashi (Osada et al.,2008). Luciferase assay was performed using the Dual-Luciferase Reporter Assay System according to the manufacturer's instructions (Promega). The E-cadherin promoter construct was transfected in NMuMG cells together with pMES empty vector, pMES::mNgn3, pCX::Snail1 or pCX::Snail2. Cells were also transfected with pGL3::RenillaTK to normalize for transfection efficiency. Both Firefly and Renilla luciferase activity were measured with a Lumat LB Tube Luminometer (Berthold). For all conditions tested, Firefly luciferase activity was normalized with that of Renilla luciferase. Each transfection was made in triplicates and the experiments performed three times.

Statistics

Error bars represent standard deviations and P values were calculated using the nonparametric Mann-Whitney U-test.

Acknowledgements

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. RESULTS
  5. DISCUSSION
  6. EXPERIMENTAL PROCEDURES
  7. Acknowledgements
  8. REFERENCES
  9. Supporting Information

M.G. designed, carried out most experiments and drafted the manuscript. Y.H.K. carried out the analysis of polarity of Ngn3-positive cells. K.J. and K.K. carried out the experiments on transient overexpression of Ngn3 in the chicken endoderm. A.G.B. designed the experiments, carried out the cloning and electroporation of Ngn3ER92AQ mutant construct. All authors read, edited, and approved the final manuscript. The authors thank G. Gradwohl, I. Rodriguez, P. Herrera, E. Semb, A. Nieto, H. Acloque, H. Osada, M. Rukstalis, C. Rüegg, and L. Larue for sharing reagents and useful comments. We thank A. Griffa and J.C. Floyd Sarria for their help with image analysis. We also thank Y. Pfister for his technical assistance and to R. MacDonald and M. Rukstalis for discussions.

REFERENCES

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. RESULTS
  5. DISCUSSION
  6. EXPERIMENTAL PROCEDURES
  7. Acknowledgements
  8. REFERENCES
  9. Supporting Information
  • Acloque H, Adams MS, Fishwick K, Bronner-Fraser M, Nieto MA. 2009. Epithelial-mesenchymal transitions: the importance of changing cell state in development and disease. J Clin Invest 119: 14381449.
  • Apelqvist A, Li H, Sommer L, Beatus P, Anderson DJ, Honjo T, Hrabe de Angelis M, Lendahl U, Edlund H. 1999. Notch signalling controls pancreatic cell differentiation. Nature 400: 877881.
  • Arnold SJ, Hofmann UK, Bikoff EK, Robertson EJ. 2008. Pivotal roles for eomesodermin during axis formation, epithelium-to-mesenchyme transition and endoderm specification in the mouse. Development 135: 501511.
  • Batlle E, Sancho E, Franci C, Dominguez D, Monfar M, Baulida J, Garcia De Herreros A. 2000. The transcription factor snail is a repressor of E-cadherin gene expression in epithelial tumour cells. Nat Cell Biol 2: 8489.
  • Bavamian S, Klee P, Britan A, Populaire C, Caille D, Cancela J, Charollais A, Meda P. 2007. Islet-cell-to-cell communication as basis for normal insulin secretion. Diabetes Obes Metab 9( suppl 2): 118132.
  • Bosco D, Orci L, Meda P. 1989. Homologous but not heterologous contact increases the insulin secretion of individual pancreatic B-cells. Exp Cell Res 184: 7280.
  • Chiang MK, Melton DA. 2003. Single-cell transcript analysis of pancreas development. Dev Cell 4: 383393.
  • Cirulli V, Beattie GM, Klier G, Ellisman M, Ricordi C, Quaranta V, Frasier F, Ishii JK, Hayek A, Salomon DR. 2000. Expression and function of alpha(v)beta(3) and alpha(v)beta(5) integrins in the developing pancreas: roles in the adhesion and migration of putative endocrine progenitor cells. J Cell Biol 150: 14451460.
  • Cole L, Anderson M, Antin PB, Limesand SW. 2009. One process for pancreatic beta-cell coalescence into islets involves an epithelial-mesenchymal transition. J Endocrinol 203: 1931.
  • Dahl U, Sjodin A, Semb H. 1996. Cadherins regulate aggregation of pancreatic beta-cells in vivo. Development 122: 28952902.
  • Deltour L, Leduque P, Paldi A, Ripoche MA, Dubois P, Jami J. 1991. Polyclonal origin of pancreatic islets in aggregation mouse chimaeras. Development 112: 11151121.
  • Desgraz R, Herrera PL. 2009. Pancreatic neurogenin 3-expressing cells are unipotent islet precursors. Development 136: 35673574.
  • Dieterlen-Lievre F, Beaupain D. 1974. [Immunocytological study of the differentiation of the endocrine pancreas in the chick embryo. I. Islands of Langerhans]. Gen Comp Endocrinol 22: 6269.
  • Dominguez D, Montserrat-Sentis B, Virgos-Soler A, Guaita S, Grueso J, Porta M, Puig I, Baulida J, Franci C, Garcia de Herreros A. 2003. Phosphorylation regulates the subcellular location and activity of the snail transcriptional repressor. Mol Cell Biol 23: 50785089.
  • Ge W, He F, Kim KJ, Blanchi B, Coskun V, Nguyen L, Wu X, Zhao J, Heng JI, Martinowich K, Tao J, Wu H, Castro D, Sobeih MM, Corfas G, Gleeson JG, Greenberg ME, Guillemot F, Sun YE. 2006. Coupling of cell migration with neurogenesis by proneural bHLH factors. Proc Natl Acad Sci U S A 103: 13191324.
  • Georgiou M, Marinari E, Burden J, Baum B. 2008. Cdc42, Par6, and aPKC regulate Arp2/3-mediated endocytosis to control local adherens junction stability. Curr Biol 18: 16311638.
  • Gradwohl G, Dierich A, LeMeur M, Guillemot F. 2000. neurogenin3 is required for the development of the four endocrine cell lineages of the pancreas. Proc Natl Acad Sci U S A 97: 16071611.
  • Grapin-Botton A, Majithia AR, Melton DA. 2001. Key events of pancreas formation are triggered in gut endoderm by ectopic expression of pancreatic regulatory genes. Genes Dev 15: 444454.
  • Greiner TU, Kesavan G, Stahlberg A, Semb H. 2009. Rac1 regulates pancreatic islet morphogenesis. BMC Dev Biol 9: 2.
  • Gu G, Dubauskaite J, Melton DA. 2002. Direct evidence for the pancreatic lineage: NGN3+ cells are islet progenitors and are distinct from duct progenitors. Development 129: 24472457.
  • Habener JF, Kemp DM, Thomas MK. 2005. Minireview: transcriptional regulation in pancreatic development. Endocrinology 146: 10251034.
  • Hand R, Bortone D, Mattar P, Nguyen L, Heng JI, Guerrier S, Boutt E, Peters E, Barnes AP, Parras C, Schuurmans C, Guillemot F, Polleux F. 2005. Phosphorylation of Neurogenin2 specifies the migration properties and the dendritic morphology of pyramidal neurons in the neocortex. Neuron 48: 4562.
  • Hirano M, Hashimoto S, Yonemura S, Sabe H, Aizawa S. 2008. EPB41L5 functions to post-transcriptionally regulate cadherin and integrin during epithelial-mesenchymal transition. J Cell Biol 182: 12171230.
  • Hutton JC, Christofori G, Chi WY, Edman U, Guest PC, Hanahan D, Kelly RB. 1993. Molecular cloning of mouse pancreatic islet R-cadherin: differential expression in endocrine and exocrine tissue. Mol Endocrinol 7: 11511160.
  • Jamora C, DasGupta R, Kocieniewski P, Fuchs E. 2003. Links between signal transduction, transcription and adhesion in epithelial bud development. Nature 422: 317322.
  • Johansson JK, Voss U, Kesavan G, Kostetskii I, Wierup N, Radice GL, Semb H. 2010. N-cadherin is dispensable for pancreas development but required for beta-cell granule turnover. Genesis 48: 374381.
  • Johansson KA, Grapin-Botton A. 2002. Development and diseases of the pancreas. Clin Genet 62: 1423.
  • Johansson KA, Dursun U, Jordan N, Gu G, Beermann F, Gradwohl G, Grapin-Botton A. 2007. Temporal control of neurogenin3 activity in pancreas progenitors reveals competence windows for the generation of different endocrine cell types. Dev Cell 12: 457465.
  • Kesavan G, Sand FW, Greiner TU, Johansson JK, Kobberup S, Wu X, Brakebusch C, Semb H. 2009. Cdc42-mediated tubulogenesis controls cell specification. Cell 139: 791801.
  • Lee SK, Pfaff SL. 2003. Synchronization of neurogenesis and motor neuron specification by direct coupling of bHLH and homeodomain transcription factors. Neuron 38: 731745.
  • Lyttle BM, Li J, Krishnamurthy M, Fellows F, Wheeler MB, Goodyer CG, Wang R. 2008. Transcription factor expression in the developing human fetal endocrine pancreas. Diabetologia 51: 11691180.
  • Ma Q, Fode C, Guillemot F, Anderson DJ. 1999. Neurogenin1 and neurogenin2 control two distinct waves of neurogenesis in developing dorsal root ganglia. Genes Dev 13: 17171728.
  • Medici D, Hay ED, Goodenough DA. 2006. Cooperation between snail and LEF-1 transcription factors is essential for TGF-beta1-induced epithelial-mesenchymal transition. Mol Biol Cell 17: 18711879.
  • Miettinen PJ, Huotari M, Koivisto T, Ustinov J, Palgi J, Rasilainen S, Lehtonen E, Keski-Oja J, Otonkoski T. 2000. Impaired migration and delayed differentiation of pancreatic islet cells in mice lacking EGF-receptors. Development 127: 26172627.
  • Miralles F, Battelino T, Czernichow P, Scharfmann R. 1998. TGF-beta plays a key role in morphogenesis of the pancreatic islets of Langerhans by controlling the activity of the matrix metalloproteinase MMP-2. J Cell Biol 143: 827836.
  • Nawshad A, Hay ED. 2003. TGFbeta3 signaling activates transcription of the LEF1 gene to induce epithelial mesenchymal transformation during mouse palate development. J Cell Biol 163: 12911301.
  • Nieto MA, Bennett MF, Sargent MG, Wilkinson DG. 1992. Cloning and developmental expression of Sna, a murine homologue of the Drosophila snail gene. Development 116: 227237.
  • Nose A, Nagafuchi A, Takeichi M. 1988. Expressed recombinant cadherins mediate cell sorting in model systems. Cell 54: 9931001.
  • Oram KF, Carver EA, Gridley T. 2003. Slug expression during organogenesis in mice. Anat Rec A Discov Mol Cell Evol Biol 271: 189191.
  • Osada H, Tomida S, Yatabe Y, Tatematsu Y, Takeuchi T, Murakami H, Kondo Y, Sekido Y, Takahashi T. 2008. Roles of achaete-scute homologue 1 in DKK1 and E-cadherin repression and neuroendocrine differentiation in lung cancer. Cancer Res 68: 16471655.
  • Perez SE, Cano DA, Dao-Pick T, Rougier JP, Werb Z, Hebrok M. 2005. Matrix metalloproteinases 2 and 9 are dispensable for pancreatic islet formation and function in vivo. Diabetes 54: 694701.
  • Piper K, Ball SG, Keeling JW, Mansoor S, Wilson DI, Hanley NA. 2002. Novel SOX9 expression during human pancreas development correlates to abnormalities in Campomelic dysplasia. Mech Dev 116: 223226.
  • Rawdon BB, Larsson LI. 2000. Development of hormonal peptides and processing enzymes in the embryonic avian pancreas with special reference to co-localisation. Histochem Cell Biol 114: 105112.
  • Rosenberg LC, Lafon ML, Pedersen JK, Yassin H, Jensen JN, Serup P, Hecksher-Sorensen J. 2010. The transcriptional activity of Neurog3 affects migration and differentiation of ectopic endocrine cells in chicken endoderm. Dev Dyn 239: 19501966.
  • Rukstalis JM, Habener JF. 2007. Snail2, a mediator of epithelial-mesenchymal transitions, expressed in progenitor cells of the developing endocrine pancreas. Gene Expr Patterns 7: 471479.
  • Schwitzgebel VM, Scheel DW, Conners JR, Kalamaras J, Lee JE, Anderson DJ, Sussel L, Johnson JD, German MS. 2000. Expression of neurogenin3 reveals an islet cell precursor population in the pancreas. Development 127: 35333542.
  • Sefton M, Sánchez S, Nieto MA. 1998. Conserved and divergent roles for members of the snail family of transcription factors in the chick and mouse embryo. Development 125: 31113121.
  • Seymour PA, Freude KK, Tran MN, Mayes EE, Jensen J, Kist R, Scherer G, Sander M. 2007. SOX9 is required for maintenance of the pancreatic progenitor cell pool. Proc Natl Acad Sci U S A 104: 18651870.
  • Sommer L, Ma Q, Anderson DJ. 1996. neurogenins, a novel family of atonal-related bHLH transcription factors, are putative mammalian neuronal determination genes that reveal progenitor cell heterogeneity in the developing CNS and PNS. Mol Cell Neurosci 8: 221241.
  • Steinberg MS, Takeichi M. 1994. Experimental specification of cell sorting, tissue spreading, and specific spatial patterning by quantitative differences in cadherin expression. Proc Natl Acad Sci U S A 91: 206209.
  • Sun Y, Nadal-Vicens M, Misono S, Lin MZ, Zubiaga A, Hua X, Fan G, Greenberg ME. 2001. Neurogenin promotes neurogenesis and inhibits glial differentiation by independent mechanisms. Cell 104: 365376.
  • Thiery JP, Acloque H, Huang RY, Nieto MA. 2009. Epithelial-mesenchymal transitions in development and disease. Cell 139: 871890.
  • Vernon AE, LaBonne C. 2006. Slug stability is dynamically regulated during neural crest development by the F-box protein Ppa. Development 133: 33593370.
  • Yang Z, Rayala S, Nguyen D, Vadlamudi RK, Chen S, Kumar R. 2005. Pak1 phosphorylation of snail, a master regulator of epithelial-to-mesenchyme transition, modulates snail's subcellular localization and functions. Cancer Res 65: 31793184.
  • Yebra M, Montgomery AM, Diaferia GR, Kaido T, Silletti S, Perez B, Just ML, Hildbrand S, Hurford R, Florkiewicz E, Tessier-Lavigne M, Cirulli V. 2003. Recognition of the neural chemoattractant Netrin-1 by integrins alpha6beta4 and alpha3beta1 regulates epithelial cell adhesion and migration. Dev Cell 5: 695707.
  • Yook JI, Li XY, Ota I, Fearon ER, Weiss SJ. 2005. Wnt-dependent regulation of the E-cadherin repressor snail. J Biol Chem 280: 1174011748.
  • Zahn C, Jaschke A, Weiske J, Hommel A, Hesse D, Augustin R, Lu L, Hong W, Florian S, Scheepers A, Joost HG, Huber O, Schurmann A. 2008. ADP-ribosylation factor-like GTPase ARFRP1 is required for trans-Golgi to plasma membrane trafficking of E-cadherin. J Biol Chem 283: 2717927188.
  • Zhou BP, Deng J, Xia W, Xu J, Li YM, Gunduz M, Hung MC. 2004. Dual regulation of Snail by GSK-3beta-mediated phosphorylation in control of epithelial-mesenchymal transition. Nat Cell Biol 6: 931940.
  • Zhuge X, Kataoka H, Tanaka M, Murayama T, Kawamoto T, Sano H, Togi K, Yamauchi R, Ueda Y, Xu Y, Nishikawa S, Kita T, Yokode M. 2005. Expression of the novel Snai-related zinc-finger transcription factor gene Smuc during mouse development. Int J Mol Med 15: 945948.

Supporting Information

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. RESULTS
  5. DISCUSSION
  6. EXPERIMENTAL PROCEDURES
  7. Acknowledgements
  8. REFERENCES
  9. Supporting Information

Additional Supporting Information may be found in the online version of this article.

FilenameFormatSizeDescription
DVDY_22544_sm_suppinfofig1.tif5411KSupp. Fig. S1. Delamination from progenitor layer and epithelium, glucagon differentiation upon sustained and transient Ngn3 overexpression in the chick intestine and pancreas. A–D: Electroporation of the floxed-Ngn3 expression vector described in Figure 2A in the absence of Cre recombinase, expression is sustained (A,B), whereas presence of Cre recombinase leads to transient overexpression (C,D). The sections are those presented as overlays in Figure 2A. The pictures show frozen transversal sections of the small intestine (A,C) and pancreas (B,D), stained for E-cadherin (purple), green fluorescent protein (GFP, marking electroporated cells in the absence of cre due to coelectroporation of a GFP-expression vector or cells having floxed-out Ngn3 in the presence of Cre) and glucagon (red). Sustained expression of Ngn3 induces basal accumulation of targeted cells, their delamination out of the epithelium and differentiation of a subset of cells into glucagon-positive cells both in the small intestine and pancreas. When a Cre-expressing vector, under the control of a CMV promoter is coelectroporated, Ngn3 expression is transient, and cells do not delaminate from the epithelium. However, they do differentiate and accumulate basally. Scale bar: 20 μm.
DVDY_22544_sm_suppinfofig2.tif2905KSupp. Fig. S2. Low E-cadherin and basal lamina breakdown in embryonic day (E) 14.5 wild type pancreas. The E14.5 wild-type mouse pancreas immunostained for the different EMT markers. A: Confocal images show that endocrine cells (glu+ins) exhibit reduced E-cadherin at the membrane both after and before clustering (observe the basal membrane of the isolated endocrine cell on the right). B: Breakdown of the basal lamina can be seen in isolated cell and cluster (Laminin). Scale bar = 20 μm.
DVDY_22544_sm_suppinfofig3.tif4429KSupp. Fig. S3. Expression pattern of Snail1 and Snail2 in the embryonic day (E) 9.5 mouse embryo. In situ hybridization performed on E9.5 mouse embryo transversal paraffin sections. Control of specificity and quality of Snail2 and Snail1 mRNA probes by analysis of the expression pattern in the mouse embryo at E9.5. Probes stain neural crest cell derivative and branchial arches as it was described previously (Nieto et al., 1992; Oram et al., 2003). fb, forebrain vesicle; ov, otic vesicle; ba, branchial arches.
DVDY_22544_sm_suppinfofig4.tif118KSupp. Fig. S4. Model of epithelial-to-mesenchymal transition (EMT) sequence of events. At the stage of Ngn3 expression, cells become pear-shaped and lose apical contact before they express hormones. At this stage, it was previously shown that Snail expression is turned on. Snail expression is indeed found in a subset of Ngn3-positive and a subset of hormone-expressing cells and we show that its expression requires Ngn3. Snail2 is a potent E-cadherin repressor. E-cadherin is indeed transcriptionally repressed by the time hormones are expressed. We have shown that the decrease of E-cadherin is sufficient to trigger delamination out of the epithelium. This delamination is accompanied by other hallmarks of EMT: breakdown of the basal lamina next to endocrine cells and induction of vimentin and N-cadherin.

Please note: Wiley Blackwell is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.