The bHLH transcription factor Mist1 is required to maintain exocrine pancreas cell organization and acinar cell identity

Abstract

The pancreas is a complex organ that consists of separate endocrine and exocrine cell compartments. Although great strides have been made in identifying regulatory factors responsible for endocrine pancreas formation, the molecular regulatory circuits that control exocrine pancreas properties are just beginning to be elucidated. In an effort to identify genes involved in exocrine pancreas function, we have examined Mist1, a basic helix-loop-helix transcription factor expressed in pancreatic acinar cells. Mist1-null (Mist1(KO)) mice exhibit extensive disorganization of exocrine tissue and intracellular enzyme activation. The exocrine disorganization is accompanied by increases in p8, RegI/PSP, and PAP1/RegIII gene expression, mimicking the molecular changes observed in pancreatic injury. By 12 m, Mist1(KO) mice develop lesions that contain cells coexpressing acinar and duct cell markers. Analysis of the factors involved in cholecystokinin (CCK) signaling reveal inappropriate levels of the CCK receptor A and the inositol-1,4,5-trisphosphate receptor 3, suggesting that a functional defect exists in the regulated exocytosis pathway of Mist1(KO) mice. Based on these observations, we propose that Mist1(KO) mice represent a new genetic model for chronic pancreas injury and that the Mist1 protein serves as a key regulator of acinar cell function, stability, and identity.

Figure 1.

Mist1 is expressed specifically in acinar cells during pancreatic development. (A) Immunohistochemistry using a Mist1-specific antibody reveals nuclear expression in putative acinar cells surrounding duct tissue (arrows) at E14.5. (B) By postnatal day 1 (PN1), Mist1 expression is wide spread in the pancreas, remaining exclusively in acinar cells. (C–H) Colocalization of CK-20 (C) or CK-7 (F) with Mist1 (D and G, red) and DAPI (blue) in adult wild-type mice indicates that Mist1 is not expressed in intra- or interlobular ducts, respectively. Combining the staining patterns (E and H) reveals an exclusion of Mist1 from duct cell nuclei (arrows).

Figure 3.

Mist1 ^KO mice have defective acinar cell organization and polarity. (A and B) The absence of Mist1 expression is confirmed in Mist1^KO mice (B) using immunohistochemistry. Wild-type mice (A) show specific expression of Mist1 in the nuclei of acinar cells. Islet and duct cells remain Mist1 negative. (C–F) Histochemical analysis reveals acinar cell disorganization in 3-m-old Mist1^KO mice. (C) Stained cryostat sections from wild-type animals show exocrine acinus formation with well-defined cellular polarity and clearly demarcated ER borders (delineated by bracket). This organization is not detected in sections from Mist1^KO mice (D). At this magnification, nuclear dysplasia (arrows) is observed also. Epoxy sections stained with toluidine blue confirm the disorganization of the ZGs. In wild-type mice (E), ZGs accumulate in clusters (arrows) at the apical border of the acinar cells, whereas nuclei remain at the basal border of the cell. In Mist1^KO mice (F), ZGs extend throughout the cell with no specific localization observed. Immunolocalization of CPA (costained with DAPI) confirms the loss of ZG organization in Mist1^KO mice. The wild-type pancreas exhibits a well-defined cellular organization (G), whereas in null pancreatic tissue CPA localization completely surrounds individual nuclei (H).

Figure 2.

Insertion of LacZ into the Mist1 locus reveals an expression pattern linked to exocrine pancreatic development. (A) Recombination of the Mist1 gene using a targeting vector in which the LacZ gene replaces the coding region of the Mist1 gene (Mist1^LacZ). The wild-type allele of Mist1 contains two exons (I and II; white boxes) with the entire coding region (red box) contained within the second exon. Probe fragments (A and B) used for Southern hybridization are indicated. B, BamHI; R, EcoRI; C, ClaI. (B) Southern blot analysis confirms the genotype of offspring from Mist1^LacZ mice. Digestion of tail DNA with BamHI produces a differential hybridization pattern with probe A. Using this external probe, the wild-type allele (10.5 kb) is distinguished from the null allele (6.7 kb). Similar blots were performed using the LacZ gene as a probe (Probe B). (C–G) β-Gal in Mist1^LacZ mice first appears at E10.5. Whole-mount staining reveals an absence of β-gal at E9.0 in the distal region of the foregut (C). At E10.5 (D), a circular rim of β-gal expression just caudal and medial to the forelimb bud appears. (E) Higher magnification suggests that this staining is specific to the distal region of the foregut. (F and G) Sections from whole-mount β-gal staining at E10.5 counterstained with eosin. β-Gal–positive cells are observed in the wall of the foregut and in a small bud of cells extending from the foregut to the neural tube (NT). (F) Higher magnification of a region from (G) shows that both the pancreatic bud and adjacent foregut contain β-gal–expressing cells, indicating Mist1 expression at the beginning of exocrine pancreas morphogenesis. At E12.5 (H), β-gal staining highlights the two pancreatic buds (inset). Staining at E14.5 indicates that LacZ expression is limited to the budding acinar cells, whereas duct cells are negative (I and J). At E15.5 (K), sagittal sections show that β-gal is limited to only the pancreas (arrow). B, brain; H, heart; L, liver; P, pancreas.

Figure 4.

Mist1-null mice develop exocrine pancreatic lesions by 12 m of age. (A–H) To reveal the overall morphology of the tissue, cryostat sections were stained with methylene blue (A–D) or β-gal histochemistry and counterstained with nuclear fast red (E–H). Wild-type littermates exhibit typical acinar cell organization (A). In contrast, pancreatic tissue of Mist1^KO mice shows progressive deterioration, resulting in lesions throughout the exocrine tissue (B–D). Initially, lesions contain presumptive acinar tissue (arrow) that maintains expression of the Mist1 gene locus (β-gal positive) (F). Eventually, circular duct-like structures are observed within the lesions (C). The appearance of presumptive ducts containing β-gal–positive cells (G, open arrows) suggests that acinar cells may have developed a duct cell phenotype. In ∼10% of the Mist1^KO mice, the majority of pancreatic tissue is lost (D) with few Mist1-LacZ–expressing cells remaining (H). (I–K) Exocrine tissue in Mist1-null animals is selectively targeted for trauma. Pancreatic tissue from a severely affected animal (I) reveals extreme disruption of the exocrine tissue (J) while retaining islets and insulin expression (K). Arrows mark identical positions on the individual serial sections. The asterisks indicate a large area that is CPA and insulin negative.

Figure 5.

Lesions within Mist1 ^KO pancreatic tissue contain cells that coexpress the Mist1 locus and duct cell markers. (A and B) Immunohistochemistry with CK-20–specific antibodies reveals the normal accumulation of duct tissue within wild-type animals (A). In contrast, the Mist1^KO pancreas contains large accumulations of duct cells throughout the exocrine tissue (B, open arrow). Identical staining patterns are obtained with antibodies against CK-7 (unpublished data). Insets reveal normal insulin-positive islets in these animals. (C–E) Staining for CK-20 and β-gal reveals apparent coexpression of acinar- and duct-specific markers in the same cells. The filled and open arrows point to β-gal–positive and –negative nuclei, respectively, within CK-20–positive cells. (F–H) Immunohistochemistry for SMA on wild-type (F) and null (G) tissue sections. SMA expression is limited to blood vessels in the wild-type mouse. However, several SMA-positive cells are observed throughout the null tissue with one particular lesion shown (H). Higher magnification (inset, costained with DAPI) indicates that these cells likely represent small blood vessels (filled arrow) and stellate cells (open arrow). Staining of null sections with Gomori's trichrome (H) also reveals the vast accumulating connective tissue (green, arrows) found in many fibrotic lesions.

Figure 6.

Lesions in Mist1^KO mice contain cells expressing both duct and acinar cell markers. (A–C) Staining for amylase (A) and CK-20 (B) reveals coexpression of acinar- and duct- specific markers in the same cells (combined in C). The insets show a higher magnification of one particular acinus that coexpresses CK-20 and amylase. (D–I) Confocal microscopy on a similar lesion at 0.7-μm optical section resolution confirms the coexpression of amylase (D and G) and CK-20 (E and H). Cells can be observed that express either amylase (open arrow), CK-20 (blue arrow), or both cell markers (white arrow) (F and I). G–I show a higher magnification of the areas highlighted in D–F. The arrows in F demarcate the edge of the particular lesion. The cells outside the lesion are primarily amylase-positive acinar cells.

Figure 7.

Progressive loss of β-catenin is associated with increased acinar cell disorganization in Mist1 ^KO mice. (A) Western blot analysis of protein samples from wild-type (+/+) and Mist1^KO (−/−) mice reveals a gradual loss of β-catenin expression in Mist1^KO acinar tissue. (B) The loss of β-catenin expression is observed in all Mist1^KO mice at 12 m of age, whereas Mist1^LacZ mice (+/−) exhibit normal levels of β-catenin. E-cadherin expression remains relatively normal in Mist1^KOmice. (C–F) Loss of β-catenin is specific to the acinar tissue in Mist1^KO mice. The exocrine tissue of Mist1-null mice exhibits a complete absence of β-catenin protein (compare C with D), whereas the islets (I, insets) and duct cells (arrows) continue to express normal β-catenin levels. Similar analysis with a ZO-1–specific antibody (E and F) reveals that tight junctions are maintained in Mist1^KO mice, albeit surrounding severely distended lumens (arrows).

Figure 8.

Mist1 ^KO mice present biochemical and molecular markers of pancreatic injury. (A) Western blot analysis for pro-CPA (47 kD) reveals the accumulation of the activated CPA enzyme (35 kD), specifically in Mist1^KO animals (arrowhead). (B and C) Intracellular digestion is observed in Mist1^KO mice. In B, a presumptive lysosome is observed fusing with a mitochondrion (arrow). Many digested mitochondria can be seen within each acinar cell, and autophagocytic bodies are observed commonly in the Mist1^KO samples (C, arrow). (D) Northern hybridizations were used to examine RegI/PSP and PAP1/RegIII gene expression at 4 and 8 m of age. In both cases, a significant increase in expression is observed in Mist1^KO mice. A similar increase is detected in p8 transcript levels in Mist1^KOmice compared with wild-type litter mates. Liver RNA samples are included as negative controls.

Figure 9.

Mist1 ^KO mice exhibit alterations in the CCK signaling pathway. (A) Western blot analysis for IP₃R1, -2, and -3, and VAMP2 expression. A specific loss of IP₃R3 expression is observed in 6-m-old Mist1^KO mice, whereas VAMP2 protein levels are elevated slightly in null animals. (B) Northern hybridization of pancreatic RNA at 3 wk and 8 and 14 m of age reveals increased expression of CCK AR in Mist1^KO mice (−/−) compared with wild-type (+/+) animals.