Plasticity in the adult rat pancreas: Transdifferentiation of exocrine to hepatocyte-like cells in primary culture

Animals.

Male Wistar rats aged 10 to 12 weeks (weight, 250–300 g) were used for the isolations (Janvier, Le Genest-St-Isle, France). Experiments were approved by the ethical committee of the Vrije Universiteit Brussel, and animals received humane care as outlined in the Guide for the Care and Use of Laboratory Animals.

Cell Isolation and Culture.

After collagenase digestion of rat pancreas, the exocrine acini were separated from other tissue components by centrifugal elutriation as described.19 In brief, acini were recovered in the fraction with an average diameter of less than 100 μm with the rotor running at 250 rpm, at a counterflow rate of 240 mL/minute (JB6 Centrifuge; Beckman, Palo Alto, CA). Acini were cultured in bacteriological Petri dishes (Nunc, Naperville, IL) in RPMI-1640 glutamax medium (Gibco BRL, Paisley, Scotland) with penicillin (75 mg/L; Continental Pharma, Brussels, Belgium), streptomycin (100 mg/L; Sigma, St Louis, MO), and 10% fetal bovine serum (Gibco BRL). The cells were kept at 37°C in a humid atmosphere of 5% CO₂ in air. For the study of cell replication, 10 μmol/L 5-bromo-2′-deoxyuridine (Sigma) was added 1 hour before harvesting the cells. Total DNA was measured as described.20 DX (Sigma) was added to the culture medium at a final concentration of 1 μmol/L. The glucocorticoid receptor antagonist RU-486 (Mifepristone; Sigma) was added at 2.5 μmol/L concentration 1 hour before addition of DX.

Freshly isolated hepatocytes were kindly provided by Dr. V. Rogiers (Free University of Brussels).

Immunocytochemistry.

Cell pellets were fixed for 1 hour in buffered 4% formaldehyde, entrapped in a 2% agarose gel (40°C; Sigma), and processed for paraffin embedding. Paraffin sections were immunostained with the streptavidin-biotin method and diaminobenzidine as chromogen.21 As primary antibodies, we used sheep polyclonal anti-albumin (Biogenesis, Poole, UK), rabbit polyclonal anti-α-amylase (Sigma), and mouse monoclonal anti-5-bromo-2′-deoxyuridine (ICN, Irvine, CA). For double immunofluorescence staining, we used an Alexa Fluor 488 conjugate (Molecular Probes, Eugene, OR) and a Sigma fluorexein isothiocyanate (FITC) conjugate as secondary antibody.

Electron Microscopy.

Cells were fixed and processed for transmission electron microscopy as described.20

Microarray Analysis.

For the micoarray analysis, freshly isolated cells and cells cultured for 5 days with DX were harvested, and total RNA (10 μg) was isolated using the Trizol RNA isolation method (Gibco BRL). Three RNA samples for each condition were analyzed. Microarray analysis were performed as described previously,22 except that we used the Enzo in vitro transcription kit containing biotin-labeled cytosine-5′-triphosphate (CTP) and uracyl-5′-triphosphate (UTP) (Enzo, Farmingdale, NY) and a fragmentation buffer containing 40 mmol/L potassium acetate and 30 mmol/L magnesium acetate. The microarray data were processed by the Affymetrix Gene Expression Analysis software and dChip (www.dchip.org).23 mRNA profiles were compared with Affymetrix (Santa Clara, CA) rat genome U34A oligonucleotide arrays. Genes were considered to be changed if they fulfilled the following criteria: (1) the average difference (AvDiff as defined by Affymetrix software on the basis of the difference between the perfect match and the mismatch probes) of 50 units or more in the condition of altered gene expression; (2) a mean fold change in expression level from three independent experiments of ≤−2 (decrease) or ≥2 (increase); (3) the expression level in each of the nine combinations having a ≤−1.5-fold or ≥1.5-fold change; or (4) P < .05.

RNA Isolation and Reverse-Transcription Polymerase Chain Reaction (RT-PCR).

Total RNA was isolated from cells and tissues using the Trizol RNA isolation method (Gibco BRL). For the semiquantitative analysis of transcripts encoding albumin, α-fetoprotein, mitochondrial cytochrome P450, α-2-macroglobulin, plasminogen activator inhibitor-1 (PAI-1), transferrin, hepatocyte nuclear factor-6 (HNF-6), cationic trypsinogen, lipase, glycosylated membrane-associated lipase, and β-actin, the total RNA was reversed transcribed and amplified as described by the manufacturer (Gibco BRL) with blanks in each assay. Primers were designed to anneal to specific sequences of rat albumin-5′: atacacccagaaagcacctc and albumin-3′: cacgaattgtgcgaatgtcac, mitochondrial cytochrome P450-5′: gcccctgctaaaagctgtgat and mitochondrial cytochrome P450-3′: ccttgacagcaggagttgcat, α-2-macroglobulin-5′: gatgcctcagctccacaaaaa and α-2-macroglobulin-3′: ttcttaaccagcaccgtgctc, PAI-1-5′: ccctctatttcaacggccaat and PAI-1-3′: cagcctggtcatgttgctctt, transferrin-5′: tgccttgacaatacccgca and transferrin-3′: ccttcccgctgatttcgaat, HNF-6-5′: agccctggagcaaactcaagt and HNF-6-3′: tggacggacgcttattttcc, cationic trypsinogen-5′: ccctcaattctcgagtgtcca and cationic trypsinogen-3′: acagccataaccccaggaaac;, lipase-5′: cgcatcattatccacggctt and lipase-3′:tcccaatagctccgaatgtcc; and glycosylated membrane-associated lipase-5′: tggaatgagccaaaaggtcg and glycosylated membrane-associated lipase-3′: gcggttttcccctcaaattg. Specific primers for control amplifications were: β-actin-5′: actatcggcaatgagcggttc and β-actin-3′: agagccaccaatccacacaga. For detection of α-fetoprotein RNA, a nested PCR was performed as described previously.24 The cycling profile for albumin was: 4 minutes at 94°C followed by 0.5 minute at 94°C, 1.5 minutes at 62°C, and 1 minute at 72°C for 30 cycles. For the other primers, the cycling profile was: 1.5 minutes at 94°C followed by 0.5 minute at 94°C, 0.5 minutes at 60°C, and 1 minute at 72°C for 10 cycles and 0.5 minute at 94°C, 0.5 minute at 58°C, and 1 minute at 72°C for 16 to 20 cycles (total of 26 cycles for transferrin; of 28 cycles for α-2-macroglobulin, PAI-1, cytochrome P450, and β-actin; and 30 cycles for HNF-6, cationic trypsinogen, trypsinogen I, lipase, and glycosylated membrane-associated lipase). Analysis of the amplified fragments was performed on ethidium bromide-stained agarose gels. All analyses were performed at least three times.

Western Blot.

Protein extraction from pancreatic cells and immunoblotting were performed as described.25 Fifty micrograms of sample were loaded on the gel. Anti-CCAAT enhancer-binding protein-β (C/EBP-β) antibody (1/100; C-19 antibody; Santa Cruz Biotechnology, Santa Cruz, CA) and horseradish peroxidase-coupled anti-mouse antibody (1/3000; Amersham, Roosendaal, The Netherlands) were used, respectively, as the first and secondary antibody. Normal rat hepatocytes were used as a positive control for C/EBP-β. All analyses were performed at least three times.

Albumin and α-Fetoprotein Expression in DX-Treated Acinar Cells.

Exocrine acini from adult rat pancreas were purified by centrifugal elutriation and were cultured in bacteriological dishes to prevent cell adhesion. In medium supplemented with 10% fetal bovine serum, the acini aggregate to form cell spheroids after a few days.16 When cells were cultured for 5 or 10 days in the presence of DX, albumin RNA became readily detectable by RT-PCR, whereas it was undetectable in cells cultured without DX (Fig. 1A). When normalized for β-actin mRNA abundance, the relative expression of albumin mRNA was approximately four times lower than that in liver. In whole pancreas and freshly isolated cells, albumin was undetectable. As shown in Fig. 1B, the fetal liver marker α-fetoprotein also becomes detectable after treatment of freshly isolated exocrine cells with DX. Albumin protein expression in DX-treated cells was demonstrated by immunocytochemistry (Fig. 2A,B), whereas there was no immunoreactivity when DX was absent from the culture medium. After 5, 10, and 15 days of culture in the presence of DX, approximately 50%, 55%, and 60%, respectively, of the cells were immunoreactive for albumin. Albumin immunoreactivity showed a granular cytoplasmic staining very similar to what was observed in freshly isolated liver hepatocytes or in liver sections (Lardon J, unpublished data, 2002). Also, the intensity of staining was comparable with that of liver hepatocytes. When cells were first cultured for 5 days in the presence of DX, followed by a culture period of another 5 days without DX, they still retained albumin expression (Lardon J, unpublished data, 2002). Thus, the albumin expression induced by DX seems to be stable even in the absence of further incubation with glucocorticoids. In the presence of the glucocorticoid receptor antagonist RU486, DX failed to induce expression of albumin (Lardon J, unpublished data, 2002). Because albumin is considered to be a hepatocyte-specific gene product, our observations suggest transdifferentiation of pancreatic cells to hepatocytes. In control cultures, amylase immunoreactivity was lost within 4 days, as we reported previously, because of dedifferentiation of the cells.16 However, the acinar cells cultured in the presence of DX retained their amylase expression. We demonstrated cells coexpressing albumin and amylase by double immunofluorescence (Fig. 3A,B). After 2 weeks of culture, amylase immunoreactivity became weak, whereas albumin staining remained intense.

Proliferation and Survival of DX-Treated Acinar Cultures.

Neither the control nor the DX-treated cells showed signs of proliferative activity, with 5-bromo-2′-deoxyuridine-labelling indices of less than 1%. The possibility of an overgrowth of a putative subset of stem cells in these cultures thus could be ruled out. The lack of proliferation and the presence of amylase-albumin double-positive cells indicate a direct transdifferentiation from acinar cells to hepatocytes. Although the total amount of DNA greatly varied between cultures, we observed consistently a better cell survival in the presence of DX. After 5, 10, and 15 days of culture, there was a survival of 38%, 29%, and 10%, respectively, of the DX-treated cells versus 25%, 8%, and 4% in the untreated cells. Microscopic observations confirmed that there was less cell death in the presence of DX and that the DX-treated cells could be kept in culture for a longer period.

mRNA Profiles of DX-Treated Acinar Cultures.

To search for other hepatocyte genes that may become activated, we compared the mRNA profile of the cells using U34A Affymetrix GeneChips. We compared the freshly isolated pancreatic exocrine cells with the cells cultured for 5 days in the presence of DX from three independent experiments. From the 8700 transcripts covered by oligo-probes on the chip, 1831 mRNAs were detectable above threshold values in the freshly isolated cells and 2825 were detectable in the DX-cultured cells. According to the criteria we specified in Methods, a total number of 263 transcripts encoding for known proteins and 217 encoding expressed sequence tags were upregulated, whereas 119 transcripts and 69 encoding expressed sequence tags were downregulated in the DX cultures when compared with the freshly isolated exocrine cells. From the upregulated genes encoding for known proteins in the DX culture, 34 (recognized by 40 probe sets) can be considered to be liver specific. These liver-specific genes can be divided over different functional classes, including plasma proteins and acute-phase proteins produced by liver, carbohydrate, amino acid and lipid metabolism, detoxification and drug metabolism, iron metabolism, urea cycle, and transcription factors (Table 1).

For some of the genes, RT-PCR was used to confirm the results of the DNA microarrays. RNA encoding α-2-macroglobulin, PAI-1, transferrin, mitochondrial cytochrome P450, and HNF-6 was expressed in the DX-treated cells, whereas it was undetectable in freshly isolated exocrine cells (Fig. 4A,B). Adult rat liver was taken as a positive control. For PAI-1, there was no RNA detectable in the normal liver; however, it is known that PAI-1 is only present in the liver during regeneration.26 We also investigated whether pancreatic-specific genes were downregulated in the DX-cultured cells. Table 2 shows that 11 of 119 downregulated transcripts are pancreas specific. We used RT-PCR to confirm the expression pattern for some of the genes. RNA encoding cationic trypsinogen, glycosylated membrane-associated lipase, and lipase was downregulated in the DX-treated cells and in the liver, where it was highly expressed in the pancreas and in the freshly isolated exocrine cells (Fig. 5). Normal rat pancreas was used as a positive control.

Expression of C/EBP-β Transcription Factor.

We analyzed the presence of C/EBP-β, a liver-selective transcription factor that was previously shown to be the “master switch” for transdifferentiation of the acinar tumour cell line AR42J into hepatocytes.14 Using Western blotting, we found that C/EBP-β protein was absent in the freshly isolated cells, but we found one protein band with the expected molecular weight of C/EBP-β, that is 36 kDa, in the cells treated with DX (Lardon, unpublished data).

Microscopic Phenotype of the Transdifferentiated Cells.

When analyzing sections from paraffin-embedded cell pellets by light microscopy, the cells in the DX-treated cultures were seen to be arranged in various ways. Sometimes a single layer of polarized cells was conspicuous at the periphery of the spheroids (Fig. 3A,B). With electron microscopy, the cells were seen to contain abundant rough endoplasmic reticulum, and in approximately 50% of the cells, zymogen granules could be found. The observed ultrastructure is not typical for acinar cells nor for hepatocytes, but may represent an intermediate between both (Lardon, unpublished data).