Hedgehog signaling is required for effective regeneration of exocrine pancreas

Abstract

Although both endocrine and the exocrine pancreas display a significant capacity for tissue regeneration and renewal, the existence of progenitor cells in the adult pancreas remains uncertain. Using a model of cerulein-mediated injury and repair, we demonstrate that mature exocrine cells, defined by expression of an Elastase1 promoter, actively contribute to regenerating pancreatic epithelium through formation of metaplastic ductal intermediates. Acinar cell regeneration is associated with activation of Hedgehog (Hh) signaling, as assessed by up-regulated expression of multiple pathway components, as well as activation of a Ptch-lacZ reporter allele. Using both pharmacologic and genetic techniques, we also show that the ability of mature exocrine cells to accomplish pancreatic regeneration is impaired by blockade of Hh signaling. Specifically, attenuated regeneration in the absence of an intact Hh pathway is characterized by persistence of metaplastic epithelium expressing markers of pancreatic progenitor cells, suggesting an inhibition of redifferentiation into mature exocrine cells. Given the known role of Hh signaling in exocrine pancreatic cancer, these findings may provide a mechanistic link between injury-induced activation of pancreatic progenitors and subsequent pancreatic neoplasia.

Figure 1

Cerulein-induced exocrine pancreatic injury is characterized by proliferative regeneration and expansion of nestin-expressing cells. Panels A–D: Histologic progression of cerulein-induced pancreatic epithelial injury and regeneration. Images represent representative H&E-stained sections of mouse pancreatic tissue harvested on indicated day following final injection of cerulein. The peak of exocrine injury is observed at around day 2 (panel B), with gradual abatement and onset of regeneration by day 5 (panel C), and near complete restitution by day 7 (panel D). Panels E–H: Exocrine regeneration is accompanied by robust proliferation assessed by BrDU incorporation. Compared with control pancreata, over 50-fold higher BrDU incorporation is observed at day 2 in the exocrine compartment (panel H). Panels I–L: Expansion of nestin-expressing, E-cadherin-positive epithelial cells following cerulein-mediated injury. In untreated pancreas (I), nestin expression is confined to occasional E-cadherin-negative mesenchymal cells. By 16 hours (J), multiple E-cadherin-positive cells (red) stain for nestin (green), consistent with nestin up-regulation in the residual epithelial compartment. Nestin expression is observed in the metaplastic ductal intermediates seen at 40 hours (L).

Figure 2

Cerulein-induced exocrine pancreatic injury is characterized by transient activation of the Hh-signaling pathway. Panels A–D: Immunohistochemical staining demonstrates injury-associated up-regulation of Shh expression with subsequent down-regulation following regeneration. Panels E–H: Temporal evolution of Hh pathway activation during cerulein-mediated injury and regeneration, as assessed by X-gal staining of pancreatic tissue harvested from Ptc-lacZ reporter mice. No exocrine expression of Shh ligand or evidence of Hh pathway activation is seen in PBS-treated pancreata (panels A and E). Shh ligand expression and Hh pathway activation are observed at days 1 and 2 (panels B and C and F and G, respectively), with abatement by day 5 (panels D and H, respectively). Panels I and J: Quantitative real-time PCR on RNA obtained from cerulein-treated mice demonstrates profound up-regulation of the Hh transcriptional target Gli1 (panel I), as well as a second Hh ligand, Indian Hh (panel J), in the injured pancreata. When normalized to the corresponding transcript levels in PBS-treated mouse pancreata, up-regulation of Hh pathway components begins as early as 2 hours, reaching peak levels between 48 and 64 hours following final cerulein injection. Error bars indicate standard error of the mean for each measurement. X-axis: hours after last cerulein injection; y-axis: relative fold expression normalized to PBS-treated mice. Panel K: Semiquantitative RT-PCR assay confirms similar up-regulation expression of Smo and Ptch1 within 2 hours following final dose of cerulein (time of harvest after last cerulein injection is indicated above each lane; β-actin is used as loading control).

Figure 3

Cyclopamine abrogates cerulein-induced up-regulation of the Hh-signaling pathway. Ptc1-LacZ reporter mice were treated with cerulein alone (Cae) or with cerulein plus cyclopamine (Cae + Cyc), and histochemical staining for β-galactosidase was performed on whole mounts of harvested pancreata. Widespread but transient up-regulation of β-galactosidase is seen in cerulein-treated mice (panels A–D), consistent with activation of the Hh-signaling pathway. β-galactosidase over expression begins on day 1 (panel A), peaks at day 2 (panel B), and is progressively down-regulated on days 3 and 7 (panels C and D), coinciding with exocrine regeneration. In contrast, pancreatic tissue from mice receiving cerulein plus cyclopamine (panels E–H) shows minimal activation of β-galactosidase activity compared with baseline expression observed in PBS-treated mice (panel I). Further documentation of the diminished Hh response associated with cyclopamine treatment was obtained using quantitative real-time PCR for Gli1 transcripts (panel J), using pancreata harvested after cerulein treatment alone (Cae, blue bars) or cerulein plus cyclopamine treatment (Cae + Cyc, brown bars). X-axis indicates day following final cerulein injection, and y-axis indicates relative fold expression of Gli1 compared with PBS-treated control mice. Values represent mean ± SEM.

Figure 4

Persistence of extraislet Pdx1 expressing metaplastic intermediates in cerulein pancreatitis in the setting of pharmacologic blockade of Hh signaling at day 7 postcerulein injury. Pdx1 expression in the pancreata of mice with retained Hh activity is restricted to the islet cells. In contrast, in the setting of Hh blockade (Cae + Cyc), persistent metaplastic intermediates (E-cadherin positive, amylase negative) are seen that label with nuclear Pdx1 (green channel). Acinar derivation of the metaplastic intermediates is suggested by the occasional presence of amylase, E-cadherin, and Pdx1 coexpressing acinar structures within the field of injury. Specificity of Pdx1 labeling in the epithelial compartment is confirmed by the presence of both Pdx1-expressing (arrow) and Pdx1-negative (arrowhead) structures within the same field.

Figure 5

Genetic interruption of Hh signaling results in impaired tissue regeneration following cerulein injury. Panels A–L: Representative H&E-stained sections of mouse pancreas on days 2, 3, and 7 following injection of either PBS or cerulein. Pancreatic tissue was harvested from either wild-type (Wt) (panels A–C), Pdx1-Cre; Smoothened^fl/fl (panels D–F), Elastase1-CreER^T2;Smoothened^fl/fl (panels G–I), or Elastase1-CreER^T2;Smoothened^fl/wt (panels J–L) mice. In panels A–C, PBS-treated wild-type mice demonstrate no tissue response. Analogously, tamoxifen-induced Ela-CreERT2;Smoothened^lox/wt heterozygous mice (panels J–L), which retain Smoothened function, demonstrate a tissue regeneration response that is identical to wild-type mice treated with cerulein (see Figure 1), with a peak of injury on day 2, and near complete regeneration by day 7. In contrast, interruption of Hh signaling either throughout the epithelial compartment (Pdx1-Cre;Smoothened^fl/fl mice), or exclusively in acinar cells (Elastase1-CreER^T2;Smoothened^fl/fl mice), resulted in an impaired regenerative response identical to that observed following cyclopamine treatment. Panels M and N: Quantification of initial injury and subsequent regeneration, using 2D fluorescent intensity profiles as demonstrated in Supplementary Figure 3 (see Supplementary Figure 3 online at www.gastrojournal.org. (M) Similar reductions in amylase staining are observed in each group on day 2, indicating similar severity of initial cerulein-mediated injury in each group. (N) Failure to regenerate amylase-expressing cells on day 7 in the setting of either pharmacologic or genetic blockade of Hh signaling. Either systemic blockade of Hh signaling by cyclopamine or deletion of Smoothened either throughout pancreatic epithelium or specifically in acinar cells results in failure to regenerate exocrine cell mass.