If you don't remember your password, you can reset it by entering your email address and clicking the Reset Password button. You will then receive an email that contains a secure link for resetting your password
If the address matches a valid account an email will be sent to __email__ with instructions for resetting your password
Correspondence: Address correspondence to: Beatriz Sosa-Pineda, PhD, Northwestern University Feinberg School of Medicine, 303 E Superior Street, SQI 8-520, Chicago, Illinois 60611.
An interactive regulatory network assembled through the induction and downregulation of distinct transcription factors governs acinar cell maturation. Understanding how this network is built is relevant for protocols of directed pancreatic acinar differentiation. The murine transcription factor Prox1 is highly expressed in multipotent pancreatic progenitors and in various mature pancreatic cell types except for acinar cells. In this study, we investigated when is Prox1 expression terminated in developing acinar cells and the potential involvement of its activity in acinar cell specification/differentiation. We also investigated the effects of sustained Prox1 expression in acinar maturation and maintenance.
Methods
Prox1 acinar expression was analyzed by immunofluorescence and confocal microscopy. Prox1-null embryos (Prox1GFPCre/Δ), Prox1AcOE transgenic mice, histologic and immunostaining methods, transmission electron microscopy, functional assays, and quantitative RNA and RNA-sequencing methods were used to investigate the effects of Prox1 functional deficiency and sustained Prox1 expression in acinar maturation and homeostasis.
Results
Immunostaining results reveal transient Prox1 expression in newly committed embryonic acinar cells. RNA-sequencing demonstrate precocious expression of multiple “late” acinar genes in the pancreas of Prox1GFPCre/Δ embryos. Prox1AcOE transgenic mice carrying sustained Prox1 acinar expression have relatively normal pancreas development. In contrast, Prox1AcOE adult mice have severe pancreatic alterations involving reduced acinar gene expression, abnormal acinar secretory granules, acinar atrophy, increased endoplasmic reticulum stress, and mild chronic inflammation.
Conclusion
Prox1 transient expression in early acinar cells is necessary for correct sequential gene expression. Prox1 expression is terminated in developing acinar cells to complete maturation and to preserve homeostasis.
The exocrine pancreas comprises more than 95% of the organ’s mass and consists of acinar cells that produce the zymogens of digestive enzymes and ducts of different calibers that solubilize the secreted zymogens in a bicarbonate fluid and transport the secretions to the duodenum. Pancreatic acinar cells have the greatest rate of protein synthesis of any mammalian organ, and they are uniquely bestowed with abundant endoplasmic reticulum (ER) to support that role.
In the mature pancreas, acinar cell functions are exclusively dedicated to the synthesis, quality control, transport, packaging, storage, and regulated exocytosis of proteins.
In mice, pancreatic acinar cells originate from a multipotent progenitor cell (MPC) population that emerges at the periphery of the epithelial plexus at around embryonic day [E] 11.5 and subsequently segregates to the tips of the branches at around E12.5 to E14.5.
MPCs are distinguished by the expression of carboxypeptidase A (Cpa) and the transcription factors (TFs) Ptf1a and c-Myc. As MPCs undergo acinar cell specification, they continue to express high levels of Cpa and Ptf1a and lose the expression of c-Myc.
Early differentiating acinar cells begin to express amylase and other digestive enzymes at around E13.5, and in the following days, this cell population continues to proliferate and expand. Acinar cell expansion is accompanied by an orderly program of gene expression that prolongs beyond postnatal stages and is governed by an interactive transcriptional regulatory network.
Acinar cell specification, differentiation, and maintenance are entirely dependent on the activity of a basic helix-loop-helix TF named Ptf1a. In the early embryonic pancreas, Ptf1a associates with another TF (Rbpj) to form a new regulatory complex (PTF1-J) that promotes growth and regulates morphogenesis.
PTF1-J activity induces an “early” acinar gene program that includes Rbpjl and its encoded TF, Rbpjl. The accumulation of Rbpjl gradually replaces Rbpj in a new complex with Ptf1a named PTF1-L, whose activity is necessary to complete acinar cell maturation.
In immature acinar cells, PTF1-L promotes a “late” acinar gene program that includes the genes of most secreted digestive enzymes (including all trypsinogens) and genes encoding proteins required for packaging, intracellular transport, exocytosis, and other functions.
Acinar cell development and maintenance also require the activities of Mist1/Bhlha15 and Lrh1/Nr5a2. In the pancreas, the TF Mist1 controls the expression of proteins involved in regulated exocytosis and various secretory proteins,
The nuclear hormone receptor family member NR5A2 controls aspects of multipotent progenitor cell formation and acinar differentiation during pancreatic organogenesis.
As an example, we published that the homeodomain TF Prox1 is highly expressed in MPC in the mouse embryonic pancreas and is undetected in acinar cells in the postnatal murine pancreas.
We also demonstrated that Prox1 loss-of-function induces premature expression of a few acinar genes (eg, Amy) and simultaneously reduces the expression of several endocrine genes (eg, Neurog3, Ins, Gcg, Nkx6.1) in the midgestation (E12.5–E13.5) murine pancreas.
Those initial results argued that Prox1 activity opposes acinar specification while favoring endocrine specification in MPCs. Alternatively, Prox1 function could delay acinar cell commitment in MPCs to enable the expansion of this progenitor population.
To better clarify the functional role of Prox1 in pancreatic acinar development, we examined its expression in murine pancreatic tissues dissected at preacinar and postacinar specification stages. We also used bulk RNA sequencing (RNAseq) to compare gene expression profiles between control and Prox1-nullizygous pancreatic specimens dissected at the acinar specification/commitment stage. Additionally, we produced a novel transgenic mouse model to investigate if sustained Prox1 expression is permissive or incompatible with acinar cell maturation and/or maintenance. We report that Prox1 is transiently expressed in early acinar cells in the murine embryonic pancreas and reveal a new role of this TF during pancreas development involving the establishment of temporal gene expression in newly specified acinar cells. In addition, we demonstrate that Prox1 activity is incompatible with acinar cell maturation since its persistent expression reduces digestive protein expression, impairs secretory granule formation, increases ER stress, and promotes a pathology reminiscent of mild chronic pancreatitis.
Prox1GFPCre/Δ mice were crossed with Prox1loxP/+ to generate Prox1-null mice. Ptf1Cre/+ mice (expressing Cre recombinase in pancreatic progenitors and pancreatic acinar cells) are described elsewhere.
Ptf1Cre/+ mice and CAG-loxP-eGFP-STOP-loxP-Prox1-IRES-β-gal mice were crossed to produce Ptf1aCre/+;CAG-loxP-eGFP-STOP-loxP-Prox1-IRES-β-gal transgenic progeny (hereafter named Prox1AcOE). Prox1GFPCre/Δ mice and Prox1loxP/+ mice were on an Naval Medical Research Institute genetic background. The other mice in the study were on a mixed genetic background. All animal experiments were performed in accordance with protocols reviewed and approved by the Institutional Animal Care and Use Committee at Northwestern University. Mice were treated in accordance with the criteria outlined in the Guide for the Care and Use of Laboratory Animals of the National Institutes of Health.
RNA Isolation and Quantitative PCR
E13.5 and E17.5 pancreatic RNAs were isolated using TRIzol Reagent (Invitrogen, Whaltham, MA, USA) and the RNeasy Plus Micro Kit (50) (Cat.74034; Qiagen, Hilden, Germany). P4, P7, and P30 pancreata were collected in RNAlater (Cat AM7021; Invitrogen, Whaltham, MA, USA), incubated overnight at 4 °C, and frozen at −80 °C before isolating the RNA using TRIzol Reagent and the PureLink RNA Mini Kit (Cat. 12183018A; Invitrogen, Whaltham, MA, USA) following the manufacturer’s instructions. The cDNA was synthesized using the iScript Reverse Transcription Supermix for RT-qPCR (Cat 1,708,840; Bio-Rad, Hercules, CA, USA) following the manufacturer’s instructions. iTaq Universal SYBR R Green Supermix (Cat. 1725121; Bio-Rad, Hercules, CA, USA) was used for quantitative PCR. qRT-PCR was performed on a Mastercycler Realplex machine (Eppendorf, Hamburg, Germany). Expression levels were determined with gene-specific primers. Data were analyzed with the relative standard curve method and were normalized to beta-actin (ActB) or Gapdh expression. Table A6 lists the primers used for quantitative PCR.
Tissue Processing
Mouse embryos and newborn and postnatal pancreata (up to P21) were dissected and prepared as described in the following text. P30 and older mice were anesthetized with 2, 2, 2-Tribromoethanol (Avertin) (Cat. T48402; Sigma-Aldrich, St Luis, MO, USA) and perfused intracardially with 4% paraformaldehyde before the pancreas was harvested. Tissues or embryos used for cryosectioning were fixed overnight in 4% paraformaldehyde at 4 °C, incubated one more night in 30% sucrose at 4 °C, and embedded in Tissue-Tek OCT Compound, (Cat. 4583; Sakura, Torrance, CA, USA). Pancreata used for paraffin sectioning were fixed overnight in 4% paraformaldehyde at 4 °C, dehydrated in a graded ethanol series, and embedded in paraffin.
Histology
The paraffin sections (8 μm) were stained with H&E for morphology analysis. The paraffin sections were stained with Sirius red (Direct Red 80, Cat. 365548; Sigma-Aldrich, St. Luis, MO, USA) or Masson’s trichrome (Connective Tissue Stain Kit, Cat. ab150686; Abcam, Cambridge, UK,) to visualize collagen fibers, following the manufacturer’s instructions. Bright-field pictures were captured using a Zeiss Axioscop 2 Plus microscope and processed with Adobe Photoshop 7.0 (Adobe Systems).
Immunofluorescence and Imaging
Immunostaining of cryosections was performed as previously reported.
Briefly, The 10-μm frozen sections were incubated in blocking solution (Cat. 11096176001; Roche, Mannheim, Germany) for 30 minutes. The sections were then incubated overnight with primary antibodies, fluorescent-conjugated secondary antibodies for 2 hours, and 4',6-diamidino-2-phenylindole (Cat. D1306; Applied Biosystems, Waltham, MA, USA) to stain the nuclei. All incubations were at room temperature. Images were acquired with a Zeiss Axioscop 2 fluorescence microscope or with a confocal Leica DM 2500 microscope and processed with Adobe Photoshop CC (Adobe Systems). Table A7 lists the antibodies used for immunofluorescence.
Immunohistochemistry and Imaging
Immunostaining of paraffin sections was performed as previously reported.
Briefly, the 8-μm sections were deparaffinized in xylene, rehydrated in ethanol, and incubated with citrate antigen retrieval buffer in a 2100-Retriever (BioVendor Laboratory, Kassel, Germany). After incubation in blocking solution (Cat. 11096176001; Roche, Mannheim, Germany) for 30 minutes, the sections were incubated with primary antibodies overnight and with biotinylated secondary antibodies for 2 hours. The sections were then incubated with ABC reagent (Cat. PK-4000; Vector Laboratories, Burlingame, CA) for 30 min and with DAB solution (Cat. SK-4100; Vector Laboratories, Burlingame, CA) to visualize the immunocomplexes. Finally, the slides were counterstained with 40% Mayer’s hematoxylin (Cat. MHS16-500ML; Sigma-Aldrich, St. Luis, MO, USA) before mounting. All incubations were at room temperature. Images were acquired with a Zeiss Axioscop microscope and processed with Adobe Photoshop CC (Adobe Systems). Table A7 lists the antibodies used for immunohistochemistry.
Phalloidin Staining
Rhodamine Phalloidin (Cat. R415; Thermo Fisher Scientific, Whaltham, MA, USA) was used to stain F-actin fibers following the manufacturer’s instructions. Briefly, the Rhodamine Phalloidin reagent (5 μL of a 200 U/mL stock solution dissolved in 100 μL 1X phosphate buffered saline) was added to the slides after immunofluorescence staining and incubated for 20 minutes at room temperature. Afterward, the slides were stained with 4',6-diamidino-2-phenylindole (Cat. D1306; Applied biosystems, Foster City, CA, USA).
Transmission Electron Microscopy
P30 pancreatic specimens were fixed in 0.1 M sodium cacodylate buffer pH7.3 containing 2% paraformaldehyde and 2.5% glutaraldehyde and postfixed with 2% osmium tetroxide in unbuffered aqueous solution, rinsed with distilled water, stained with 3% uranyl acetate, rinsed with distilled water, dehydrated in ascending grades of ethanol, transitioned with propylene oxide, and embedded in resin mixture of the EMbed 812 kit (Cat. RT14120; Electron Microscopy Sciences, Hatfield, PA), cured in a 60 °C oven. Samples were sectioned on a Leica Ultracut UC6 Ultramicrotome. The 1-μm-thick sections were collected and stained with Toluidine Blue O, and 70-nm sections were collected on 200-mesh copper grids. Thin sections were stained with uranyl acetate and Reynolds lead citrate. A FEI Tecnai Spirit G2 120kV transmission electron microscope located at the Northwestern Center for Advanced Microscopy was used for imaging.
Morphometric Analyses
To quantify the acinar, beta cell, and ductal cell areas, the entire pancreas was sectioned at 10 μm with a cryostat, and each 10th consecutive section was incubated with antibodies for E-cadherin, amylase, insulin, and Sox9. The E-cadherin+ area was calculated per field using ImageJ 1.37v software (NIH, Bethesda, USA) and used as reference of total parenchyma area. The amylase+, insulin+, and Sox9+ areas were measured in the same corresponding field using ImageJ 1.37v software to calculate the percentage of acinar, beta cell, and ductal areas with respect to the total parenchyma area. To measure the percentage of acinar tissue expressing the Prox1/β-gal transgene, the entire pancreas was sectioned at 10 μm with a cryostat, and each 10th consecutive section was incubated with antibodies for β-galactosidase and amylase. The amylase+ and β-gal+ areas were calculated using ImageJ 1.37v software. To quantify the number of apoptotic cells per field, the entire pancreas was sectioned at 10 μm with a cryostat, and each 10th consecutive section was incubated with antibodies for cleaved Caspase-3 (Asp175). The Casp3+ cells were counted in at least 6 different fields per section using ImageJ 1.37v software. To measure acinar cell size, the entire pancreas of 3 Prox1AcOE mice was sectioned at 10 μm, and each 10th consecutive section was incubated with antibodies for E-cadherin, β-galactosidase, and amylase (at least 10 sections per pancreas were collected). ImageJ 1.37v software was used to determine the size of individual Ecad+/β-gal+/amy+ and Ecad+/β-gal−/amy+ acinar cells (the E-cadherin signal was used to delimit the cell border). At least 20 cells per field were counted. To measure the size and number of acinar secretory granules, the pancreas of 2 control and 2 Prox1AcOE mice was processed using a transmission electron microscope as described previously. At least 3 pictures per specimen with 5 to 10 acinar cells per field were selected, and the secretory granules were measured and counted with ImageJ 1.37v software.
Edema and Fibrosis Scoring
H&E staining and Sirius red staining were performed separately on 8-μm pancreatic sections, and images were acquired and processed as previously described (6 images from 3 slides covering distinct pancreatic areas, per genotype, per age [1 month and 5 months of age], and per condition [H&E or Sirius red staining]). The severity of edema and fibrosis was blindly graded by a semiquantitative assessment as described in the study by Moreno et al.
The edema scores are as follows: 0 (absent), 1 (focally increased between lobules), 2 (diffusely increased between lobules), and 3 (acini disrupted and separated). The fibrosis scores are as follows: 0 (absent), 1 (rare or around ductal margins), 2 (in the parenchyma, <50% of the lobules), and 3 (in the parenchyma, >50% of the lobules).
Immune Infiltrate Quantification
Immunostaining with anti-CD68 antibodies was performed as described previously on 8-μm pancreatic sections (6 images from 3 slides covering distinct pancreatic areas, per genotype, per age [1 month and 5 months of age]). The images were acquired and processed using ImageJ 1.37v software (NIH, Bethesda, USA) to measure the CD68+ area and the total pancreas area per field.
Serum Amylase
Mice were fasted overnight and then administered a bolus of cerulein (70μg/Kg, Cat. C9026; Sigma-Aldrich, St Louis, MO, USA). After 1 hour, the mice were anesthetized by isofluorane inhalation, and blood was extracted from the retro-orbital plexus. The blood was allowed to coagulate at room temperature for 30 minutes and centrifuged at 2000 rpm for 15 minutes to separate the serum. Amylase activity was measured using the Magle Life Sciences Phadebas Amylase Test (Cat. MSPP-1301; VWR Avantor, USA), following the manufacturer’s instructions. The diluted serum samples were incubated with the substrate for 15 minutes at 37 °C, filtered, and the absorbance at 620 nm was measured using a Gen5 Synergy 2 plate reader (BioTek, Vermont, USA). Amylase activity was calculated interpolating the absorbance values in the standard curve provided by the manufacturer and multiplied by the dilution factor. Amylase activity was expressed as micromoles per second per liter (μmol/s∗L).
Trypsin Activity
Mice were euthanized, and the whole pancreas was dissected. Intrapancreatic trypsin activity was measured using the Trypsin Activity Colorimetric Assay kit (Cat. MAK290; Sigma-Aldrich, St Louis, MO, USA) following the manufacturer’s instructions. Briefly, each pancreas was homogenized in 4 volumes of trypsin assay buffer, centrifuged, and 50μL of clear supernatant was transferred to a 96-well plate. The samples were incubated with 1μL of chymotrypsin inhibitor (10mM) for 10 minutes at room temperature. Fifty microliters of trypsin substrate diluted 1:25 in assay buffer was added, and the plate was incubated at 25 °C. The change in absorbance (405 nm) was recorded every 15 minutes in a Gen5 Synergy 2 plate reader for 2 hours. Protein concentration was measured by Bicinchoninic acid assay using the Pierce BCA Protein Assay Kit (Cat. 23225;Thermo Fisher Scientific, Waltman, MA, USA) and reading the absorbance at 562 nm in a Gen5 Synergy2 plate reader. Trypsin activity was calculated as indicated by the manufacturer, normalized to the total protein in the assay mix, and expressed as mU/mL/mg.
RNA Sequencing
E13.5 pancreatic RNA was isolated using TRIzol Reagent (Invitrogen, Whaltham, MA) and the RNeasy Plus Micro Kit (50) (Cat. 74034; Qiagen, Hilden, Germany). P7 pancreata were collected in RNAlater (Cat. AM7021; Invitrogen, Whaltham, MA) and incubated overnight at 4 °C before isolating RNA using TRIzol Reagent and the PureLink RNA Mini Kit (Cat. 12183018A; Invitrogen, Whaltham, MA. USA). RNA quality and quantity were measured using the Agilent 2100 Bioanalyzer System (Agilent Technologies). RNA libraries were prepared using the Tru-Seq mRNA-Seq Library Prep and sequenced using an Illumina HiSeq 4000 instrument at the Northwestern University NUSeq Core facility. The cutoff for determining significantly differentially expressed genes was a false discovery rate (FDR)–adjusted P value < .05 using the Benjamini-Hochberg method. Gene set expression analysis bioinformatics tools were used to further compare gene expression profiles between P7 control and Prox1AcOE pancreata. Personnel at the NUSeq Bioinformatics Core performed the analyses of the data sets and assisted the interpretation of the results. RNAseq data have been deposited in Gene Expression Omnibus under accession numbers GSE178173 and GSE178174.
Western Blot Analysis
Pancreatic tissues were isolated, flash-frozen, and powdered in liquid nitrogen. The tissue samples were sonicated for 30 seconds in ice-cold RIPA buffer (Cat No. 89900; Thermo-Fisher, Waltman, MA) with complete Protease Inhibitor Cocktail tablets (Cat No. 11697498001; Roche, Indianapolis, IN). The lysates were centrifuged at 4 °C for 15 minutes, and protein was quantified by Bicinchoninic acid assay using the Pierce BCA Protein Assay Kit (Cat No. 23225; Thermo Fisher Scientific, Waltman, MA, USA). A total of 30 μg of protein was loaded per lane. Sodium dodecyl sulfate/polyacrylamide gel electrophoresis, protein transfer, and Western blotting were performed using standard techniques. Table A7 lists the antibodies used for Western blot. SuperSignal West Pico Chemiluminescent Substrate Pierce (Cat No. 34580; Thermo Fisher Scientific, Waltman, MA, USA) was used, and images were obtained using the iBright Imaging Systems (Thermo Fisher Scientific, Waltman, MA, USA). Band intensity was quantified using ImageJ 1.37v software (NIH, Bethesda, USA).
Statistics
The data were analyzed using GraphPad Prism software (version 8; GraphPad Software, San Diego, CA). Data are shown as the mean ± standard error of mean. Statistical significance was determined using 2-tailed unpaired Student t test to compare control and Prox1-null or control and Prox1AcOE groups. Levels of significance are ∗P < .05, ∗∗P < .01, and ∗∗∗P < .001, and P > .05 is not significant. Each quantitative experiment was repeated at least 3 times. We consider biological replicates as those mice or tissues subjected to the same experimental test and technical replicates as individual samples or tissues subjected to the same analysis.
Results
Prox1 is Expressed During a Short Time Window in Early Developing Acinar Cells
Prox1 expression fluctuates in all epithelial cell lineages in the mouse embryonic pancreas and becomes restricted to endocrine, ductal, and centroacinar cells (and excluded from acinar cells) in the postnatal murine pancreas.
we hypothesized that its expression must be terminated prior or shortly after acinar cell specification. To investigate this notion, we performed immunofluorescence experiments in the pancreas of wild-type mouse embryos dissected at various developmental stages: acinar cell specification/commitment (∼E13.5), early acinar cell differentiation (∼E15.5), perinatal stage (E18.5), and postnatal stage (P5). This analysis shows moderate-to-high levels of Prox1 expression in newly specified acinar precursors identified with the specific marker Cpa1
and located at the periphery of the expanding epithelial plexus (Figure 1A) in E13.5 pancreata. Also, immunofluorescence results show low but persistent expression of Prox1 in early differentiating acinar cells (amylase+) located at the tip of the branching epithelium (Figure 1B) in E15.5 pancreata. In contrast, Prox1 expression is absent in acinar cells and is visible in ductal cells, centroacinar cells, and islet cells, in the pancreas of E18.5 mouse embryos (Figure 1C) and P5 mouse pups (Figure 1D). These results demonstrate that Prox1 expression persists in newly specified/committed acinar cells during a short time window and is terminated in that cell lineage toward the end of gestation.
Figure 1Prox1 is transiently expressed in developing acinar cells. Immunofluorescence images of (A) E13.5, (B) E15.5, (C) E18.5, and (D) P5 wild-type pancreata stained for Prox1 (green in A, red in B-D) and co-stained with the acinar markers: carboxypeptidase A1 (red in A), amylase (green in B and blue in C, D), and the epithelial marker E-cadherin (blue in A and green in C, D). Prox1 is expressed in early developing acinar cells (E13.5, E15.5; white arrowheads in A, B) and is absent in perinatal (E18.5; white arrowheads in C) and postnatal (P5; arrowheads in D) acinar cells. White arrows indicate Prox1+ “trunk” cells (A, B) and Prox1+ ductal cells (C, D). Yellow arrows indicate Prox1+ lymphatic cells (A) and Prox1+ centroacinar cells (C, D). Yellow arrowhead indicates Prox1+ islet cells (D). Images are representative from at least 3 independent experiments. Scale bars = 25 μm.
To investigate if Prox1 transient expression is relevant for acinar cell development, we compared gene expression profiles between wild-type (control) and Prox1-nullizygous (Prox1GFPCre/Δ) pancreata dissected at E13.5 using bulk RNAseq. In agreement to our previous findings at E12.5,
the new RNAseq data reveal reciprocal changes in acinar and endocrine gene expression in the E13.5 Prox1-null pancreas. Specifically, lack of Prox1 function upregulates multiple acinar genes (Table A1), downregulates numerous endocrine genes (Table A2), and reduces the expression of a few hepatic genes (eg, Afp, Alb) (Table A2), in the E13.5 murine pancreas. To further examine the effects of Prox1 functional deficiency in acinar gene expression, we compared the levels of “early” and “late” acinar transcripts identified in other studies
between E13.5 control and Prox1-null pancreata using quantitative real-time polymerase chain reaction (qRT-PCR). The results of this analysis show slightly increased expression of a few “early” acinar transcripts
(Amy2b, Cpa1, and Pnliprp1) in the E13.5 Prox1-null pancreas, although among those transcripts, only Cpa1 is significantly changed (Figure 2). In contrast, numerous “late” acinar transcripts
appear significantly upregulated in E13.5 Prox1-null pancreata, including several trypsinogen transcripts (eg, Prss1, Prss2, Prss3, Try10, Try4, Try12, and Ctrl) and transcripts encoding granule secretory proteins (eg, Sync and Gp2) (Figure 2). These new results support and expand our previous report of premature expression of 2 “late” acinar genes (Prss2 and Try4) in the pancreas of E12.5 Prox1-null embryos.
Finally, the combined results from RNAseq and qRT-PCR analyses show that lack of Prox1 function does not affect the expression of transcripts encoding the key regulators of acinar cell specification and differentiation Ptf1a, Rpbjl, and Rbpj
in the embryonic pancreas (Figure 2 and Table A1). Therefore, it is probable that Prox1 expression restricts or delays the activity of PTF1-L in newly committed acinar cells, since the “late” acinar transcripts showing upregulation in E13.5 Prox1-null pancreata are known targets of the transcriptional complex formed by Ptf1a and Rbpjl.
Figure 2PTF1-L target genes are significantly upregulated in the E13.5 Prox1-null pancreas. Expression levels of acinar and liver transcripts in the pancreas of E13.5 Prox1-null embryos relative to the expression in E13.5 control pancreas quantified by qRT-PCR. Gene expression was normalized to Gapdh. Results are expressed as mean ± SEM (n = 4 samples with 3-4 pancreata per genotype). ∗P < .05; ∗∗P < .01; ∗∗∗P < .001 by 2-tailed Student t test.
Prox1 Continuous Expression Does not Affect Morphogenesis or Suppresses Differentiation in Embryonic Acinar Cells
The rapid extinction of Prox1 expression in embryonic acinar cells raises the possibility that its activity is nonpermissive for acinar cell maturation. To test this hypothesis, we crossed Ptf1aCre/+ mice
to produce Ptf1aCre/+;CAG-loxP-eGFP-STOP-loxP-Prox1-IRES-β-gal (or Prox1AcOE) transgenic mice carrying sustained Prox1 expression in embryonic, postnatal, and adult pancreatic acinar cells (Figure 3A). Since Prox1AcOE mice are heterozygous for Ptf1a,
all comparisons in our study used pancreatic specimens from Ptf1aCre/+ mice as controls (which are not bona fide surrogates of wild-type pancreata). We found that Prox1AcOE transgenic mice are viable and look normal at birth, and they survive to adulthood (at least until 5 months of age, which is the end point of the study). The analysis of Prox1AcOE embryos dissected at E15.5 using double-immunostaining for β-gal (Figure 3A) and Prox1 demonstrates broad expression of these proteins throughout the pancreas, including the epithelial branching tips where the early developing acinar cells are located (Figure 3B). In contrast, Prox1 levels are low in the epithelial branching tips, and the β-gal reporter is not expressed (Figure 3B) in the pancreas of E15.5 control littermates. Double-immunostaining for amylase and Prox1 in E18.5 and P8 Prox1AcOE pancreatic specimens shows broad, moderate Prox1 expression in acinar cells at E18.5 (Figure 3C) and more extensive areas of acinar tissue expressing high levels of Prox1 at P8 (Figure 3D). Conversely, double-immunostaining experiments demonstrate that Prox1 is not expressed in acinar cells in E18.5 and P8 control pancreata (Figure 3C, D). Additional results of hematoxylin and eosin (H&E) staining reveal comparable acinar morphology and overall pancreas architecture between control and Prox1AcOE pancreata at E18.5 (Figure 3C) and P8 (Figure 3D). To further examine the Prox1AcOE newborn pancreas, we stained this organ from 2 P4 Prox1AcOE pups and 2 P4 control littermates with antibodies for the pancreas epithelium marker E-cadherin (Figure 3E), the acinar marker amylase (Figure 3E), the islet marker insulin (Figure 3E), and the ductal marker Sox9 (Figure 3E) and performed morphometric analyses. We found that the overall pancreas size (ie, total E-cadherin+ area) is ∼25% smaller in P4 Prox1AcOE mice than in control littermates (Figure 3F). In contrast, the proportion of insulin+ cells, ductal cells, and acinar cells is no different between Prox1AcOE pancreata and control pancreata at P4 (Figure 3F). These results demonstrate that while Prox1 continuous expression has no major consequences for fetal acinar morphogenesis or differentiation, it reduces postnatal pancreas growth.
Figure 3Persistent expression of Prox1 does not affect fetal acinar cell morphogenesis. (A) Schematic of the CAG-Prox1-IRES-β-gal transgene and the production of Prox1AcOE (Ptf1aCre/+;CAG-Prox1-IRES-β-gal) transgenic mice. (B) Immunofluorescence images of E15.5 control (Ptf1Cre/+) and Prox1AcOE pancreata co-stained for Prox1 (green) and β-gal (red). Prox1 expression (arrows) is lower at the branching tips than at the trunk region (arrowheads) in the control pancreas. Prox1 and β-gal (arrows) are uniformly expressed in the Prox1AcOE pancreas. (C, D, left panels) H&E staining images show no obvious histologic differences between control and Prox1AcOE pancreata at E18.5 (C) and P8 (D). (C, D, middle and right panels) Immunofluorescence images of E18.5 (C) and P8 (D) pancreata co-stained for Prox1 (red), amylase (green), and E-cadherin (blue). Prox1 is not expressed in acinar cells (arrowheads) in the control pancreas at E18.5 (C) and P8 (D). Prox1 acinar expression (arrowheads) increases from E18.5 (C) to P8 (D) in the Prox1AcOE pancreas. (E, top) Immunofluorescence images of P4 pancreata co-stained for Prox1 (red) and E-cadherin (blue). Prox1 is expressed in ductal (d), islet (isl), and centroacinar cells (arrowheads) in control and Prox1AcOE pancreata. Acinar cells (arrows, Ac) lack Prox1 expression in control pancreata and show high Prox1 expression in Prox1AcOE pancreata. (E) Immunofluorescence images of P4 pancreata co-stained for amylase (red) and E-cadherin (blue), insulin (red) and E-cadherin (blue), and Sox9 (red) and E-cadherin (blue) show similar abundance of acinar tissue (amylase+, arrows), beta cells (insulin+, arrows), and ductal cells (Sox9+, arrows) between control and Prox1AcOE pancreata. (F) Quantification of total pancreatic area (Ecad+) and the percentage of acinar (amylase+), beta cell (insulin+), and ductal (Sox9+) areas in P4 control and Prox1AcOE pancreata (average ± std. dev. n = 2 specimens per genotype). (C, D) The green channel is not shown in the images on the right for better comparison of Prox1 expression. Control tissues are Ptf1Cre/+. Images are representative from at least 3 independent experiments. Scale bars = 50 μm.
Acinar Cell Morphology is Gradually Disrupted Upon Persistent Prox1 Ectopic Expression
Different to our findings at P8, H&E staining results show focal areas where acinar cells have scarce cytoplasm and enlarged lumens in the pancreas of Prox1AcOE mice at preweaning (P15, Figure 4A) and postweaning (P23, Figure 4A) stages. H&E staining also reveals large areas with abnormal acinar tissue, immune cell infiltrates, and conspicuous edema (Figure 4B) in the pancreas of 1-month-old Prox1AcOE mice. In contrast, none of the former alterations are noticed in the pancreas of control littermates (Figure 4B). In addition, many acini show abnormal globular shapes and mislocalized nuclei in the pancreas of P30 Prox1AcOE mice (Figure 4C), whereas all acini have polygonal-shaped and basolateral-located nuclei in the pancreas of P30 control littermates (Figure 4C). The immunohistochemistry results reveal that the areas of abnormal acinar tissue are infiltrated with macrophages in the pancreas of 1-month-old Prox1AcOE mice (Figure 4D). In contrast, only a few resident macrophages are visible in the pancreas of control littermates (Figure 4D). Finally, immunostaining for the apoptosis marker active Caspase-3 shows numerous acinar cells expressing this protein in the pancreas of P30 Prox1AcOE mice (Figure 4E), and almost no acinar cells were positive for active Caspase 3 in the pancreas of control littermates (Figure 4E). These results demonstrate that the acinar tissue is gradually lost due to apoptosis and an inflammatory process ensues in the pancreas of Prox1AcOE young adult mice.
Figure 4Histologic defects, inflammation, and acinar apoptosis are prevalent in P30 Prox1AcOE pancreata. (A) H&E staining images of control and Prox1AcOE pancreata dissected at P15 and P23 show focal areas with no acini and aberrant ductal structures (arrows) in the Prox1AcOE specimens. (B) H&E staining images of P30 control and Prox1AcOE pancreata show edema (asterisks, right panel), immune cells infiltrate (arrowhead, right panel), and abnormal acini (arrows, right panel) in the Prox1AcOE specimen. (C) H&E staining images of P30 control and Prox1AcOE pancreata show acini with basolateral nuclei and polygonal shape (arrows, left panel) in the control specimen, and globular nonpolarized acini (arrows, right panel) in the Prox1AcOE specimen. (D) Immunohistochemistry images of P30 control and Prox1AcOE pancreata stained for F4/80 show abundant macrophages (arrows) in the Prox1AcOE specimen. (E) Immunohistochemistry images of P30 control and Prox1AcOE pancreata stained for active Caspase-3 show a single Caspase-3+ cell in the control specimen (arrow, left panel) and numerous Caspase-3+ cells in the Prox1AcOE specimen (arrows, right panel). Right, Quantification of apoptosis (ie, active Caspase-3+ cells per field) in P30 control and Prox1AcOE pancreata. Results are expressed as mean ± SEM (n = 6 samples, 3 pancreata per genotype). Symbols correspond to different specimens. ∗∗∗P < .001 by 2-tailed Student t test. Control tissues are Ptf1Cre/+. All images are representative from at least 3 independent experiments. Scale bars = 100 μm (A, top panels, B), 50 μm (A, bottom panels, D, E), 25 μm (C).
We investigated if the morphologic alterations in Prox1AcOE adult pancreata correlate with the extent of Prox1 ectopic acinar expression using immunostaining approaches. This analysis reveals extensive (albeit mosaic) Prox1 expression in the acinar lobules of P30 Prox1AcOE pancreata (Figure 5A). In contrast, Prox1 is expressed in islets, ducts, and centroacinar cells but not in acinar cells (Figure 5A) in the pancreas of control littermates. The comparison of Prox1+ and Prox1− acini within the same Prox1AcOE pancreatic specimen shows that the Prox1+ acini are smaller, have less cytoplasm, and their nucleus is mislocalized relative to the adjacent Prox1− acini that have abundant cytoplasm and basolateral nuclei (Figure 5A). Also, double-immunofluorescence experiments to compare β-gal+ and β-gal− acini within individual P15 and P30 Prox1AcOE pancreatic specimens show normal apical distribution of the protein mucin in acini that do not express β-gal (Figure 5B) and very low or no mucin expression in adjacent β-gal+ acini (Figure 5B). Moreover, the results of double-immunofluorescence and morphometric analyses reveal that the average size of β-gal+ acinar cells is significantly smaller than the average size of adjacent β-gal− acinar cells (Figure 5C). To quantify the extent of Prox1/β-gal acinar expression in the pancreas of Prox1AcOE mice of various ages, we performed double-immunostaining experiments using antibodies for β-gal and amylase. The results show extensive β-gal expression in acinar cells at E18.5 and P7 (Figure 5D) and mosaic expression of β-gal in acinar cells at 1 month and 3 months of age (Figure 5D). Quantification of those results demonstrates that >80% of the Prox1AcOE acinar tissue expresses Prox1/β-gal at the perinatal and neonatal stages (Figure 5E), and less than 50% of the Prox1AcOE acinar tissue expresses Prox1/β-gal in adult stages (Figure 5E). In conclusion, we discovered that Prox1 persistent expression causes severe morphologic alterations and promotes apoptosis in adult acinar cells. In turn, the lost acinar tissue is gradually replaced by acinar cells that do not express the Prox1/β-gal transgene.
Figure 5Prox1 ectopic acinar expression is increasingly mosaic in Prox1AcOE pancreata. (A) Immunohistochemistry images of P30 control and Prox1AcOE pancreata stained for Prox1 show restricted expression in centroacinar, ductal, and islet (isl) cells (arrows, left panel) in the control specimen, and broad mosaic acinar expression in the Prox1AcOE specimen (arrows, right panel; arrowheads indicate Prox1− acini). (B) Immunofluorescence images of P15 and P30 Prox1AcOE pancreata co-stained for mucin (red), E-cadherin (green), and β-galactosidase (blue) show mucin (arrows) expression on the apical side of β-gal− acini and very low or absent mucin expression in β-gal+ acini (arrowheads; yellow arrowheads indicate a few β-gal+/mucin+ acini in the P15 specimen). (C, left) Confocal immunofluorescence image of P30 Prox1AcOE pancreata stained for E-cadherin (red) and β-galactosidase (blue) shows mosaic expression of β-gal in the acini. The β-gal+ acinar cells (arrows) look smaller than the β-gal− acinar cells (arrowheads). (C, right) Results of morphometric analysis show that the median size of the β-gal+ acinar cells is smaller than that of the β-gal− acinar cells in P30 Prox1AcOE pancreata (n = 3). ∗∗∗P < .001 by 2-tailed Student t test. (D) Immunofluorescence images of E18.5, P7, 1- and 3-month-old pancreata co-stained for β-galactosidase (red) and amylase (green) show increasingly mosaic expression of Prox1/β-gal in the pancreas of Prox1AcOE mice from E17.5 to 3 months of age. (E) Quantification of the percentage of β-gal+ area in Prox1AcOE pancreata of various ages. Results are expressed as mean ± SEM (n = 12 samples, 3 pancreata per genotype). ∗∗∗P < .001 by 2-tailed Student t test. Control tissues are Ptf1Cre/+. All images are representative from at least 2 independent experiments. Scale bars = 50 μm (A, B [P30], C, D), 25 μm (B [P15]).
Prox1AcOE Adult Mice Display Considerable Acinar Atrophy and Features of Mild Chronic Pancreatitis
To investigate the long-term consequence of sustained Prox1 acinar expression, we analyzed the pancreas of 5-month-old Prox1AcOE mice by histologic and immunostaining approaches. H&E staining shows portions of relatively intact acinar tissue (Figure 6A) and variable areas with extensive acinar atrophy, edema, lipomatosis, and few normal acini (Figure 6A) in 5-month-old Prox1AcOE pancreata. Also, Sirius red and Masson’s trichrome staining reveal scattered regions with abundant-to-moderate collagen deposition in those tissues, especially around areas where acinar atrophy is prominent (Figure 6B). The results of immunostaining show that the Prox1+ acini are smaller, have unusual globular shape, and lack basolateral nuclear polarity (Figure 6C) in comparison to the surrounding Prox1− acini (Figure 6C). Moreover, the areas showing persistent Prox1 acinar expression appear infiltrated with macrophages (Figure 6C), and the Prox1+ acinar cells display very deficient expression of the secretory enzymes amylase and elastase (Figure 6C) in comparison to the adjacent Prox1− acinar cells (Figure 6C). As inflammation, fibrosis and acinar atrophy are features of chronic pancreatitis, we measured several typical pathologic indicators
in control and Prox1AcOE adult mice. The results of cerulein-induced amylase secretion did not reveal significant differences between control and Prox1AcOE mice at 2 months and 5 months of age (Figure 6D). On the other hand, a nonsignificant trend of increased intrapancreatic trypsin activity (Figure 6E) and decreased pancreas weight/body weight (Figure 6F) was uncovered in 3-month-old Prox1AcOE mice. Furthermore, significant increases in the abundance of CD68+ immune cells (Figure 6G), edema score (Figure 6H), and fibrotic area (Figure 6I) were revealed in the pancreas of Prox1AcOE mice at 1 month and 5 months of age. These results demonstrate that sustained Prox1 acinar expression gradually induces a pathologic process recapitulating features of mild chronic pancreatitis.
Figure 6Tissue damage and fibrosis are prevalent in the Prox1AcOE adult pancreas. (A) H&E staining images of the 5-month-old Prox1AcOE pancreas showing areas with intact acinar tissue (left), scarce immune infiltrates (arrow, left), and lipid deposits (arrowhead, left) and areas with extensive acinar atrophy (right), lipomatosis (arrowhead, right), and abundant immune infiltrates (arrows, right). (B) Images of Sirius red (left) and mason-trichrome (right) staining of the 5-month-old Prox1AcOE pancreas showing moderate collagen deposits and fibrosis (arrows) in areas with conspicuous acinar atrophy. (C, D) Immunostaining of adjacent sections of a 5-month-old Prox1AcOE pancreas shows that the Prox1+ acini (C, “Prox1” arrows) are smaller and have different morphology compared to the surrounding Prox1– acini; they are infiltrated with macrophages (C, “F4/80” arrows) and show very deficient expression of amylase (D, “Amylase” arrows) and elastase (D, “Elast” arrows). The arrowheads indicate normal-looking, Prox1– acini. The immunofluorescence image is stained with elastase (red), Prox1 (green), and E-cadherin (blue). (E) Quantification of serum amylase in 2-month-old (2 MO) and 5-month-old (5 MO) control (ctrl) and Prox1AcOE (OE) mice (n = 14 [2 MO, ctrl], 18 [2 MO, OE], 6 [5 MO, ctrl], and 7 [5 MO, OE]). (F) Quantification of intrapancreatic trypsin activity in 2 MO control and Prox1AcOE mice (n = 5 ctrl and 6 OE). (G) Quantification of pancreas weight/body weight in 2 MO control and Prox1AcOE mice (n = 4 ctrl and 5 OE). (H–J) Quantification of macrophage area (H, CD68+ cells), edema score (I, H&E staining), and fibrotic area (J, Sirius red staining), in 1-month-old (1 MO) and 5-month-old (5 MO) control and Prox1AcOE mice. Results are expressed as mean ± SEM (n = 6 samples and 3 pancreata per genotype). ∗∗∗P < .001 by 2-tailed Student t test. Images are representative from at least 3 independent experiments. Control mice are Ptf1Cre/+. Scale bars = 100 μm (A, B, C [Prox1]), 50 mm (C [F4/80] and D).
Bulk RNAseq Reveals Enhanced ER Stress and Activation of the Unfolded Protein Response in Prox1AcOE Adult Pancreata
To uncover the underlying mechanisms driving the severe Prox1AcOE acinar pathology, we compared gene expression profiles between Prox1AcOE and control (Ptf1cre/+) pancreatic specimens dissected at P7 by bulk RNAseq. This experiment was performed at P7 since no histologic signs of acinar atrophy or conspicuous inflammation were noticed in Prox1AcOE pancreata prior to P15 (Figures 3 and 4). Evaluation of the RNAseq results using gene set expression analysis identifies inflammation and fibroblast growth factor signaling amongst the most significantly upregulated pathways in P7 Prox1AcOE pancreata (Figure 7A). Also, the gene set expression analysis results show that pathways involving protein translation are significantly downregulated in the transgenic pancreas (Figure 7A and B). The analysis of individual changes in gene expression reveals that ∼80 transcripts show significant upregulation (Table A3) in P7 Prox1AcOE pancreata. Some transcripts in that list encode proteins induced by ER stress (eg, Fgf21)
and proteins that mediate unfolded protein response (UPR) activation downstream of ER stress (eg, Eif2ak3/PERK, Atf5, Atf3, Hspa9/mortalin, Reg3b, and Ddit3/CHOP).
qRT-PCR approaches were used to validate some of the former results, and the results demonstrate persistent upregulation of transcripts encoding markers of ER stress and UPR activation (Fgf21, sXbp1, Atf4, Grp78/Bip, Ddit3/CHOP, and Edem)
in P30 Prox1AcOE pancreata (Figure 7C). These findings indicate that sustained Prox1 expression induces apoptosis in acinar cells via increasing ER stress and promoting UPR activation.
Figure 7Indicators of ER stress and UPR activation are increased in the pancreas of Prox1AcOE mice at P7. (A) Results of gene set enrichment analysis (GSEA) show upregulation of interferon signaling and immune response pathways, and downregulation of pathways involving protein translation, in P7 Prox1AcOE pancreata (FDR<0.05). Three independent biological replicates per genotype were used for the RNAseq experiment. (B) GSEA plots show that the most downregulated pathways in P7 Prox1AcOE pancreata relate to protein translation. (C) QRT-PCR data showing that markers of ER stress (Fgf21) and unfolded protein response (UPR) activation (tXbp1, usXbp1, sXbp1, Grp78, Ddit, Atf4, and Edem) are increased in P30 Prox1AcOE (OE) pancreata. Gene expression was normalized to Gapdh. Control tissues are Ptf1Cre/+. Results are expressed as mean ± SEM. ∗P < .05; ∗∗P < .01; ∗∗∗P < .001 by 2-tailed Student t test (n = 6 [control], 5 [Prox1AcOE] pancreata). Control tissues are Ptf1Cre/+.
Secretory Granule Defects are Prevalent in the Prox1AcOE Adult Pancreas
Analysis of the RNAseq data sets identified ∼230 transcripts showing significant downregulation in the pancreas of P7 Prox1AcOE mice (Table A4). Many of those genes encode acinar digestive enzymes (eg, Pga5, Pnliprp2, Ctrl, Cel, Pla2g1b, Prss53, Prss8, Try10, Cpn1, Pnlip, Try4, Cela3b, and Pnliprp1), acinar secretory granule proteins (eg, Gp2, Zg16, and Sync), regulators of pancreatic fluid secretion (eg, Aqp12), and acinar TFs (eg, Rbpjl, and Xbp1) (Table A4). We were intrigued by the discovery that some of the most downregulated genes in P7 Prox1AcOE pancreata encode major acinar secretory granule proteins (Gp2, Zg16, Aqp12, and Sync).
To follow up and expand those results, the expression of Gp2, Zg16, and Aqp12 was analyzed in Prox1AcOE pancreatic specimens dissected at E17.5, P4, and P30 by qRT-PCR. The results show relatively normal expression of Gp2, Zg16, and Aqp12 at E17.5 (Figure 8A) and significant downregulation of those transcripts at P4 and P30 (Figure 8A) in Prox1AcOE pancreata. This analysis also demonstrates significantly reduced expression of Ctrl from E17.5 onward (Figure 8A) and significant upregulation of the ER stress marker Fgf21 as early as P4 (Figure 8A) in Prox1AcOE pancreata. We also compared the P7 Prox1AcOE RNAseq data sets with the E13.5 Prox1-nullizygous RNAseq data sets and uncovered opposite changes in the expression of Ctrl, Sync, Prss2, and Gp2 in the Prox1 gain-of-function and loss-of-function conditions (Figure A1). These findings suggest that Prox1 activity represses Ctrl, Sync, Prss2, and Gp2 in acinar cells.
Figure 8Secretory granule defects increase with age in Prox1AcOEpancreata. (A) Top: Expression levels of Prox1 and the acinar transcripts Ptf1a, Mist1, Gp2, Zg16, Aqp12, Ctrl1, and Fgf21 in control (C) and Prox1AcOE (OE) pancreata at E17.5 and P4, quantified by qRT-PCR (n = 6 per genotype, per stage). Bottom: Expression levels of Prox1, Ctrl, Gp2, Zg16, and Aqp12 in control (C) and Prox1AcOE (OE) pancreata at P30, quantified by qRT-PCR (n = 6 [control] and 5 [Prox1AcOE]). Gene expression was normalized to ActB. Control tissues are Ptf1Cre/+. Results are expressed as mean ± SEM. ∗P < .05; ∗∗P < .01; ∗∗∗P < .001 by 2-tailed Student t test. (B) Transmission electron microscopy images of P30 control and Prox1AcOE pancreata show differences in the size, number, and localization of secretory granules. Blue arrows indicate nuclei, yellow arrows indicate centroacinar cells, and red asterisks indicate lumens. (C) Quantification of the number of electron-dense granules per cell (n = 13 cells [control] and 16 cells [Prox1AcOE]) and the granule cell area (in pixels2 x 103) (n = 744 granules [control] and 987 granules [Prox1AcOE]). ∗∗∗P < .001 by 2-tailed Student t test. Three to five separate fields were used for quantification. Control tissues are Ptf1Cre/+. (D) Immunofluorescence images of the P30 Prox1AcOE pancreas stained for the secretory granule proteins SNAP29 (red), VAMP8 (red), and Syncollin/SYCN (red), or for F-actin fibers (with rhodamine phalloidin [red]), show very different expression of those proteins in β-gal+/Prox1+ (arrows) and β-gal–/Prox1– (arrowheads) acini. Sections were co-stained for E-cadherin and β-gal or Prox1. Nuclei were stained with DAPI (gray). (E) The immunofluorescence image of P15 Prox1AcOE pancreata co-stained for Mist-1 and β-galactosidase shows comparable Mist1 expression in β-gal+ (arrow) and β-gal– (arrowhead) acini. (F) Left and middle: Immunofluorescence images of P30 Prox1AcOE pancreata co-stained for Mist-1 (red) and β-galactosidase (blue) show high Mist1 expression in β-gal– acini (arrow) and very low or no Mist1 expression in β-gal+ acini (arrowheads). Right: The immunofluorescence image of P30 Prox1AcOE pancreata co-stained for connexin 32 (red) and β-galactosidase (blue) shows connexin 32 proteins in β-gal– acini (arrow) but not in β-gal+ acini (arrowhead). Images are representative from at least 2 independent experiments. Scale bars = 5 mm (B), 25 mm (D, F), 50 mm (E).
) (Table A3), we analyzed the ultrastructure of those organelles in P30 control and Prox1AcOE pancreatic specimens by transmission electron microscopy. This analysis shows large zymogen granules located on the apical side of acinar cells in control pancreata (Figure 8B). In contrast, the zymogen granules look smaller and appear distributed throughout the cytoplasm in acinar cells of Prox1AcOE pancreata (Figure 8B). The results of morphometric analysis corroborate that the acinar secretory granules are significantly more numerous and smaller in the pancreas of Prox1AcOE mice in comparison to those granules in control pancreata (Figure 8C). We also stained the pancreas of P30 Prox1AcOE mice with antibodies for β-gal and various zymogen granule proteins and uncovered marked differences between β-gal+ acini and β-gal− acini. Specifically, this analysis reveals normal apical localization of SNAP29 proteins
Pancreatic acinar cells express vesicle-associated membrane protein 2- and 8-specific populations of zymogen granules with distinct and overlapping roles in secretion.
Pancreatic acinar cells express vesicle-associated membrane protein 2- and 8-specific populations of zymogen granules with distinct and overlapping roles in secretion.
proteins appear concentrated toward the apical side in β-gal− acini (Figure 8D), and they show diffuse and very low expression in β-gal+ acini (Figure 8D). We also compared the distribution of F-actin filaments between β-gal+ acini and β-gal− acini in the pancreas of P30 Prox1AcOE mice. In acinar cells, polymerized F-actin fibers contribute to regulated exocytosis by coating the zymogen granules and stabilizing the fusion of membranes.
The results of immunofluorescence and phalloidin staining uncovered normal apical distribution of actin filaments (F-actin) in β-gal− acini (Figure 8D) and diffuse mislocalized distribution of F-actin in β-gal+ acini (Figure 8D). These collective results demonstrate that Prox1 ectopic expression alters the morphology and localization of secretory granules and disrupts apicobasal polarity in pancreatic acinar cells.
Published studies showed that Mist1 activity establishes and maintains the secretory acinar phenotype in multiple exocrine glands, including the pancreas
. Since we uncovered a severe secretory granule phenotype in adult pancreatic acinar cells expressing ectopic Prox1, we investigated the expression of Mist1 in Prox1AcOE pancreatic specimens dissected at different ages. The results of qRT-PCR analysis and RNA sequencing show comparable Mist1 transcript expression between control and Prox1AcOE pancreata at E17.5, P4, and P7 (Figure 8A and Table A4). Also, the results of double-immunofluorescence analysis using anti-β-gal and anti-Mist1 antibodies show very similar Mist1 expression in β-gal− (“normal”) acini and β-gal+ (“Prox1-positive”) acini (Figure 8E) in the pancreas of P15 Prox1AcOE mice. In contrast, the double-immunofluorescence results reveal high Mist1 expression in β-gal−acini (Figure 8F) and very low or no Mist1 expression in β-gal+ acini (Figure 8F) in the pancreas of P30 Prox1AcOE mice. Western blot results corroborate very reduced expression of Mist1 proteins in the pancreas of P30 Prox1AcOE mice (Figure A2). Additional data from double-immunofluorescence analysis demonstrate expression of the gap junction protein connexin 32 (a known Mist1 pancreatic target
) in β-gal− (“normal”) acini (Figure 8G) and lack of connexin 32 expression in β-gal+ (“Prox1-positive”) acini (Figure 8G) in the pancreas of P30 Prox1AcOE mice. Taken together, our study uncovered a progressive decay in Mist1 protein expression in adult acinar cells expressing ectopic Prox1. Most probably, the lack of Mist1 activity triggered many of the severe secretory granule alterations described earlier and contributed to aggravate the overall Prox1AcOE pancreatic pathology.
Discussion
Pancreatic acinar cell maturation is governed by an interactive regulatory network that is gradually assembled through the induction and termination of specific TF functions.
In this study, we show that the homeodomain TF Prox1 is transiently expressed in newly committed acinar cells to temporarily arrest the expression of genes encoding secretory digestive zymogens and other “late” acinar proteins.
We also demonstrate that Prox1 activity is detrimental for acinar cell maturation since its persistent expression downregulates multiple genes encoding secretory enzymes and major zymogen granule proteins, it increases ER stress and induces apoptosis, and it disrupts secretory granule morphology. Furthermore, Prox1 ectopic acinar expression elicits a gradual pathologic process involving inflammation, fibrosis, and other features typical of mild chronic pancreatitis. These new findings demonstrate that Prox1 expression must be tightly regulated during acinar cell development to enable proper maturation and to preserve homeostasis.
The reassessment of E13.5 Prox1-nullizygous pancreata by RNAseq analysis uncovered numerous alterations in acinar gene expression that almost certainly involve the activity of PTF1-L. In newly committed acinar cells, PTF1-J and PTF1-L complexes sequentially induce an array of genes encoding digestive enzymes, metabolic proteins, and secretory machinery components.
At the onset of acinar differentiation (E12.5–E13.5), PTF1-J activates Cpa1, Amy2b, Cela, Rbpjl, and a few other “early” acinar genes. Once PTF1-L proteins begin to accumulate (∼E15.5), a “late” acinar differentiation program is induced that includes genes encoding digestive proteases (eg, Prss2), packaging proteins, secretory proteins, and metabolic regulators.
reported >3-fold Prox1 upregulation in the pancreas of E17.5 Rbpjl-nullizygous mouse embryos. Using microarray analysis, we previously uncovered ∼2-fold increase in Amy2b expression and ∼48-fold increase in Prss2 expression in the pancreas of E12.5 Prox1-nullizygous embryos.
Our new results of RNAseq not only expand the list of acinar genes whose expression is affected by the loss of Prox1 activity in the embryonic pancreas. In addition, those results reveal that the most significantly upregulated acinar genes in the E13.5 Prox1- Prox1-null pancreas are targets of PTF1-L, whereas in those mutant tissues, the expression of Ptf1a, Rbpj, and Rbpjl is not affected. These novel findings and the results in the study by Masui et al.
advocate a model whereby Prox1 function restrains PTF1-L activity or prevents PTF1-L formation in early acinar cells. Once the PTF1-L complexes start to accumulate, the expression of Prox1 is terminated, and the “late” acinar gene program is fully activated.
A second prediction of the former model is that sustained Prox1 expression could impair differentiation and/or maturation in acinar cells via antagonizing the activity of PTF1-L. To explore this possibility, we produced Prox1ACOE transgenic mice in which Prox1 expression is induced and maintained in the acinar cell lineage. Contrary to our initial hypothesis, the analysis of E17.5 Prox1ACOE embryos uncovered minimal effects in acinar cell development under Prox1 continuous expression. However, it is possible that a specific threshold of Prox1 activity is required to affect acinar gene expression since multiple PTF1A target genes showed increased downregulation in the Prox1ACOE pancreas between P4 and P7, and we noticed increasing abundance of Prox1 proteins in the Prox1ACOE acinar tissue after birth. Many of the PTF1A target genes downregulated in the Prox1ACOE postnatal pancreas encode secretory enzymes (Try10, Try4, Try5, Ctrl, Prss1, Cel, Trypsinogen 7, Pnliprp1/2, and Rnase1), nonsecretory proteins (Gatm, Gls2, Gcat, Tmed11, Derl3, Lfgn, Dmbt1, Gal, and Tff2), secretory granule proteins (Gp2, Zg16, Aqp12, and Sync), regulators of ER protein folding (Cabp2 and Erp27), and acinar TFs (Rbpjl and Spdef).
showed downregulation in P7 Prox1ACOE pancreata, whereas in those tissues, the expression of Ptf1a, Mist1, and Nr5a2 was not affected. Furthermore, although less severe, the Prox1ACOE adult acinar phenotype recapitulates many features reported in mice with Ptf1a deletion in adult acinar cells, including loss of apicobasal polarity, altered zymogen granule distribution, increased ER stress, apoptosis, and reduced cell size.
We uncovered several indicators of enhanced ER stress in the pancreas of Prox1AcOE transgenic mice, such as upregulation of Fgf21 (encoding a protective acinar cell secretagogue induced by ER stress)
at P7, and increased expression of canonical indicators of ER stress and UPR activation (eg, Atf5, Eif2ak3/PERK, Ddit3/CHOP, Hspa9, sXbp1, Atf4, Grp78/Bip, and Edem)
from P7 onward. Under conditions in which the ER capacity is exceeded, the UPR is activated to inhibit protein synthesis and to promote the degradation of misfolded proteins, and if the damage is excessive or unresolved, the UPR induces cell death.
Accordingly, the increased acinar cell death that we reported in the pancreas of Prox1ACOE mice could be the consequence of activation of the UPR downstream of mounting ER stress. How sustained Prox1 expression increases ER stress in acinar cells is unclear, although this alteration could involve impaired or reduced PTF1A activity
Furthermore, we do not rule out that excessive production of Prox1/β-gal proteins contributed to increase ER stress and UPR activation in acinar cells of our Prox1ACOE mouse model.
Prox1ACOE adult mice develop a progressive pancreatic pathology involving inflammation, fibrosis, tissue atrophy, and other features reminiscent of chronic pancreatitis. We discovered multiple secretory morphologic alterations involving mislocalized and abnormally small zymogen granules, deficient expression of secretory proteins, aberrant expression of secretory granule proteins implicated in Ca2+ regulated exocytosis, a disorganized mesh of F-actin fibers, and loss of apicobasal polarity, in the pancreas of Prox1ACOE adult mice. Furthermore, we identified 3 genes encoding essential zymogen granule components (Gp2, Zg16, and Aqp12) amongst the earliest and most downregulated genes in Prox1ACOE pancreata. While the former observations suggest that Prox1 activity directly interferes with zymogen granule formation, several results in our study indicate that the loss of Mist1 expression is a major driver of the Prox1ACOE adult secretory phenotype. For instance, we found that Mist1 transcripts and Mist1 proteins have normal expression in Prox1ACOE pancreata at P7 to P15, and at those stages, the majority of Prox1/βgal+ acinar cells show no obvious morphologic alterations. In contrast, Prox1/βgal+ acinar cells show very deficient expression of Mist1 and its target connexin 32 in addition to multiple morphologic secretory alterations in Prox1ACOE adult pancreata. The former results argue that as Mist1 expression declines in Prox1+ acinar cells, the secretory machinery dismantles, and the pathologic process aggravates. Moreover, it is possible that early-on in the Prox1ACOE pathologic process, the expression of Mist1 helps to mitigate a rise in ER stress via inducing the expression of FGF21.
Why Mist1 expression lost relatively late in the acinar cells of Prox1ACOE mice is unclear, although this alteration could be the result of a specific threshold of Prox1 activity and loss of other acinar TF functions. Interestingly, despite all the previous acinar secretory granule defects, the response to cerulein-induced amylase secretion was relatively normal in Prox1ACOE mice. However, we caution that the interpretation of this result is complicated by the fact that all adult transgenic mice retained a sizeable intact acinar mass in their pancreas.
In summary, we uncovered a novel role of Prox1 in newly committed acinar cells involving the temporal regulation of “late” acinar genes (particularly PTF1-L target genes) and demonstrate that sustained Prox1 activity is detrimental for acinar cell maturation and homeostasis possibly through interference with PTF1A functions. The results in our study concur with other publications showing the pathologic effects of sustained or increased expression of TFs that should be downregulated or terminated in acinar cells, such as c-Myc,
Moreover, our findings recommend monitoring the expression of Prox1 in protocols of directed pancreatic acinar differentiation of embryonic or induced-pluripotent stem cells to ensure that proper maturation is achieved.
Acknowledgement:
The authors thank Ming-Yi Chiang and Wanshu Ma (Northwestern University) for technical support, Guillermo Oliver (Northwestern University) for providing the CAG-Prox1-IRES-LacZ mice, and the services of the NUSeq Core Facility and the Center for Advanced Microscopy (Northwestern University).
Authors’ Contributions:
Beatriz Sosa-Pineda and Angelica S. Martinez-Ramirez conceived, designed, and directed the study and wrote the manuscript. Angelica S. Martinez-Ramirez performed most of the experiments with the assistance of Thomas L. Borders. Leena Paul initiated the study. Xinkun Wang conducted RNA sequencing experiments. Matthew Schipma performed RNA sequencing bioinformatics analysis. Farida Korobova assisted the electron microscopy analysis. Christopher V. Wright provided the Ptf1cre/+ mice.
The nuclear hormone receptor family member NR5A2 controls aspects of multipotent progenitor cell formation and acinar differentiation during pancreatic organogenesis.
Pancreatic acinar cells express vesicle-associated membrane protein 2- and 8-specific populations of zymogen granules with distinct and overlapping roles in secretion.
Conflicts of Interest: The authors disclose no conflicts.
Funding: This study was funded by Northwestern University Feinberg Medical School, National Institutes of Health/National Institute of Diabetes and Digestive and Kidney Diseases grant RO1 DK106266 (BS-P), and Northwestern University's Center for Advanced Microscopy/Cancer Center Support Grant (NCI CA060553).
Ethical Statement: The corresponding author, on behalf of all authors, jointly and severally, certifies that their institution has approved the protocol for any investigation involving humans or animals and that all experimentation was conducted in conformity with ethical and humane principles of research.
Data Transparency Statement: Data and analytic methods will be made available to other researchers upon acceptance of the manuscript. Study materials will be available upon request.
00091-7&id=gr1.jpg){kind=link}
00091-7&id=gr2.jpg){kind=link}
00091-7&id=gr3.jpg){kind=link}
00091-7&id=gr4.jpg){kind=link}
00091-7&id=gr5.jpg){kind=link}
00091-7&id=gr6.jpg){kind=link}
00091-7&id=gr7.jpg){kind=link}
00091-7&id=gr8.jpg){kind=link}