Loss of Pancreas upon Activated Wnt Signaling Is Concomitant with Emergence of Gastrointestinal Identity

Abstract

Organ formation is achieved through the complex interplay between signaling pathways and transcriptional cascades. The canonical Wnt signaling pathway plays multiple roles during embryonic development including patterning, proliferation and differentiation in distinct tissues. Previous studies have established the importance of this pathway at multiple stages of pancreas formation as well as in postnatal organ function and homeostasis. In mice, gain-of-function experiments have demonstrated that activation of the canonical Wnt pathway results in pancreatic hypoplasia, a phenomenon whose underlying mechanisms remains to be elucidated. Here, we show that ectopic activation of epithelial canonical Wnt signaling causes aberrant induction of gastric and intestinal markers both in the pancreatic epithelium and mesenchyme, leading to the development of gut-like features. Furthermore, we provide evidence that β -catenin-induced impairment of pancreas formation depends on Hedgehog signaling. Together, our data emphasize the developmental plasticity of pancreatic progenitors and further underscore the key role of precise regulation of signaling pathways to maintain appropriate organ boundaries.

Fig 1. Loss of pancreatic cell identity in Pdx1-Cre; Ctnnb1^tm1Mmt/+ pancreas.

Hematoxylin/eosin staining of E13.5 pancreatic sections showing reduced branching and multiple dilations in the pancreatic epithelium of Pdx1-Cre; Ctnnb1^tm1Mmt/+ pancreata (A) compared to control pancreata (B). Boxed areas in A and B are shown in higher magnification in A' and B', respectively. The enzyme Carboxipeptidase A1 (Cpa1, red) is located in the acinar progenitor cell population found at the tips of the E13.5 branching epithelium stained for the epithelial maker E-cadherin (green) in control mice (C). Cpa1-positive cells (arrowheads) are reduced in E13.5 Pdx1-Cre; Ctnnb1^tm1Mmt/+ pancreatic epithelium (D). The pancreatic ductal epithelium in E13.5 Pdx1-Cre; Ctnnb1^tm1Mmt/+ embryos display loss of apical marker Muc-1 (green) (arrowheads) but normal basal (laminin, blue) and basolateral marker (E-cadherin, green) in dilated areas (F) compared to control pancreata (E). Insets show higher magnification pictures. (G) Quantification of Cpa1-positive area. Marked reduction in Pdx-1 expression in Pdx1-Cre Ctnnb1^tm1Mmt/+ pancreatic epithelium (I) at E13.5 compared to control pancreata (H). Note that in Pdx1-Cre; Ctnnb1^tm1Mmt/+ pancreatic epithelium, Pdx-1 expression is dramatically reduced in cells displaying nuclear β-catenin accumulation while cells with only membranous β-catenin localization express significant Pdx-1 levels (I' and I''). Decreased expression of the embryonic pancreatic transcription factors NKX2.2 (J, K) and NKX6.1 (L, M) in E13.5 Pdx1-Cre; Ctnnb1^tm1Mmt/+ pancreata. (N) Quantitative RT-PCR analysis of the expression of pancreatic transcription factors in E13.5 and E15.5 embryonic pancreas. (O) Quantitative RT-PCR analysis of the expression of canonical Wnt signaling target genes at E13.5 and E15.5 embryonic pancreas. Results in N and O are expressed as fold relative to levels in control pancreata (n = 3–6 embryos from each genotype). Data are presented as mean ± SEM; * p < 0.05 (Student's t-test). Scale bars 100 μm, (in A for A-B; in E for E-H; in I for I-J).

Fig 2. β-catenin stabilization induces activation of gastrointestinal genes.

Staining for PAS (A, B) and Alcian Blue (C, D) in Pdx1-Cre; Ctnnb1^tm1Mmt/+ and control newborn pancreata. Immunohistochemical staining for the gastric-specific mucin Muc5ac (E, F), intestine-specific mucin Muc2 (G, H) in control and mutant Pdx1-Cre; Ctnnb1^tm1Mmt/+ newborn pancreata. Arrowheads mark cells lining the epithelium positive for the gastrointestinal markers. Cdx2 accumulation in pancreatic cysts of newborn Pdx1-Cre; Ctnnb1^tm1Mmt/+ mice in low (I) and high (J) magnification pictures. Cdx2 is not detected in newborn control pancreata (M). Sox2 accumulation in pancreatic cysts of newborn Pdx1-Cre; Ctnnb1^tm1Mmt/+ mice in low (K) and high (L) magnification pictures. No expression of Sox2 is detected in control pancreata (N). The boxed areas in I and K are shown at higher magnification in J and L, respectively. Arrowheads in J and L indicate the areas shown in the insets at higher magnification. (O) Coexpression of Sox2 and Cdx2 in epithelial cells of pancreatic cysts in newborn Pdx1-Cre; Ctnnb1^tm1Mmt/+ mice. The boxed area in O is shown at higher magnification in the inset. (P) Coexpression of β-galactosidase and Sox2 gastric marker in epithelial cells (arrowheads) of pancreatic cysts in newborn Pdx1-Cre; R26R; Ctnnb1^tm1Mmt/+ mice. (Q) Analysis of the expression level of transcription factors specific of intestine (Cdx2) and stomach (Sox2 and Pitx1) showed a significant increase in expression in the pancreas of E15.5 Pdx1-Cre; Ctnnb1^tm1Mmt/+ embryos. Results are expressed as fold relative to levels in control pancreata (n = 6 embryos from each genotype). Data are presented as mean ± SEM; ** p < 0.01 *** p < 0.001 (Student's t-test). st, stomach; d, duodenum. Scale bars, 100 μm.

Fig 3. Increased mesenchyme in pancreas with activated β-catenin signaling.

Gomori staining shows increased mesenchyme surrounding the epithelium of the pancreatic cystic structures in newborn Pdx1-Cre; Ctnnb1^tm1Mmt/+ mice (B) compared to control mice (A). Arrowheads in B indicate the region enlarged in the inset. Immunofluorescence for antibodies against laminin and smooth muscle actin (SMA) suggest formation of smooth muscle in newborn Pdx1-Cre; Ctnnb1^tm1Mmt/+ pancreas (D). SMA-positive layers are only found associated with big blood vessels in control pancreata (C). Tuj1-positive cells are intercalated in the SMA-positive layer surrounding the cysts in newborn Pdx1-Cre Ctnnb1^tm1Mmt/+ pancreata (F) compared to control pancreata (E). Increased pancreatic mesenchyme in Pdx1-Cre Ctnnb1^tm1Mmt/+ embryos (H) compared to control embryos (G) is observed during early stages of pancreas formation. Increased proliferation (marked by Ki-67 immunostaining) of epithelial and mesenchymal cells in E13.5 Pdx1-Cre; Ctnnb1^tm1Mmt/+ embryonic pancreata (J) compared to control pancreata (I). Note that proliferative epithelial cells in control embryonic pancreata are preferentially located at the tips of the branching epithelium (yellow asterisks). Pancreatic epithelium is outlined in white dashed line. Quantification of proliferating epithelial (K) and mesenchymal (L) cells in E13.5 embryonic pancreas. Results are shown as percentage of KI67⁺ cells. (n = 3 embryos from each genotype). (M) Analysis of the expression level of mesenchymal gastrointestinal markers shows a significant increase in expression in the pancreas of E13.5 and E15.5 Pdx1-Cre; Ctnnb1^tm1Mmt/+ embryos. Results are expressed as fold relative to levels in control pancreata (n = 3–6 embryos from each genotype). Data in K, L and M are presented as mean ± SEM; ** p< 0.01, *** p< 0.001 (Student's t-test). Scale bars 100 μm.

Fig 4. Early β-catenin activation in foregut endoderm severely impacts pancreas formation.

Low magnification pictures of hematoxylin/eosin staining of paraffin sections of E18.5 Rfx6-Cre; Ctnnb1^tm1Mmt/+ (A) and control mice (F). Boxed areas are regions that are shown in higher magnification in right panels for each antibody marker used in consecutive paraffin sections. The pancreatic remnant of E18.5 Rfx6-Cre; Ctnnb1^tm1Mmt/+ mice displayed membranous β-catenin localization (B) and no accumulation of Sox2 (B') and Cdx2 (B''). The stomach of E18.5 Rfx6-Cre; Ctnnb1^tm1Mmt/+ mice displayed nuclear β-catenin localization (C), expression of Sox2 (C') but not Cdx2 (C'). The dilated area of the duodenum located proximal to the stomach (D) displayed nuclear β-catenin accumulation concomitant with increased expression of Sox2 (D') and Cdx2 (D''). However, the duodenal area distal to the stomach did not display nuclear β-catenin localization (E). In this area, Cdx2 (E'') but not Sox2 (E'), was expressed. In E18.5 control mice, pancreas (G), stomach (H) and duodenum (I) displayed β-catenin membranous localization. No Sox2 (G') or Cdx2 (G'') was expressed in control pancreas. In control stomach, Sox2 was expressed (H') but not Cdx2 (H''). In control duodenum, Cdx2 was expressed (I'') but not Sox2 (I'). d, duodenum; dp, dorsal pancreas; li, liver; p, pancreas, spleen; st, stomach; vp, ventral pancreas. Scale bars 100 μm (in B for B-I).

Fig 5. Inhibition of the Hedgehog pathway partially rescues pancreas formation in pancreas with activated β-catenin signaling.

Increased Shh expression in the embryonic (E13.5) epithelium of Pdx1-Cre; Ctnnb1^tm1Mmt/+ pancreas (B) compared to control pancreata (A). The boxed areas are shown at higher magnification in the adjacent panels (A' and B'). (C) Fold change and p-values from upregulated components of the hedgehog pathway in E13.5 Pdx1-Cre; Ctnnb1^tm1Mmt/+ pancreas observed in microarray analysis. (D) Increase in expression of Hh pathway genes in embryonic E15.5 Pdx1-Cre; Ctnnb1^tm1Mmt/+ pancreata as assessed by quantitative PCR. Results are expressed as fold relative to levels in control pancreata (n = 6 embryos from each genotype). Data are presented as mean ± SEM; *p < 0.05 **p < 0.01 (Student's t-test). (E-J) Hh inhibition partially rescues pancreas formation in cultured pancreatic rudiments. Images show confocal z-stacks of pancreatic explants in whole-mounts stained for Pdx1. Foreguts from E10.5 Pdx1-Cre; Ctnnb1^tm1Mmt/+ and control mice were cultured in the presence or absence of 10 μM cyclopamine for 3 consecutive days. Vehicle-treated Pdx1-Cre Ctnnb1^tm1Mmt/+ pancreatic buds displayed a reduced and disorganized epithelium positive for Pdx-1 (F) compared to control pancreatic buds (E). (G) Cyclopamine-treated Pdx1-Cre; Ctnnb1^tm1Mmt/+ pancreatic buds displayed an increase in Pdx-1-positive epithelium and a recovery of the normal epithelium morphology. (H-J) Whole-mount immunostaining for Muc-1 (green) and Pdx-1 (red) showing reduced branching and decreased Muc-1 expression in vehicle-treated Pdx1-Cre; Ctnnb1^tm1Mmt/+ pancreatic buds (I) compared to control cyclopamine-treated control pancreatic buds (H). Addition of cyclopamine partially rescued these phenotypes (J). Images are z-stacks of serial confocal sections. (K) Quantification of Pdx-1-positive pancreatic epithelium in cultured explants. Results are shown as normalized ratios compared to DMSO-treated mutant Pdx1-Cre; Ctnnb1^tm1Mmt/+ pancreatic explants (n = 6 explants per group, 3 independent experiments). Data are presented as mean ± SEM. ** p< 0.01 (one-way ANOVA). Scale bars 100 μm, (in A for A-B; in E for E-G; in H for H-J).