Numerous studies conducted in a diversity of adult tissues have shown that certain stem cells are characterized by the expression of a protein known as the ABCG2 transporter (where ABC is ATP- binding cassette). In the adult pancreas, although various multipotent progenitors have been proposed, the ABCG2 marker has only been detected in the so-called ‘side population’ (a primitive haematopoietic cell population with a multipotential capacity). In the present study we sought to identify new ABCG2+ pancreatic cell populations and to explore whether they exhibit the properties of progenitor cells. We isolated and expanded mitoxantrone-resistant cells from pancreata of lactating rats by drug selection. These cells were characterized and maintained in different stages of differentiation using several media ‘cocktails’ plus Matrigel™ (BD Biosciences). Differentiation was assessed by RT–PCR (reverse transcription–PCR), immunocytochemistry, electron microscopy and ELISA. The expanded cell population demonstrated a phenotype of PaSCs (pancreatic stellate cells). Spontaneous cell clusters occurred during cell expansion and they showed weak expression of the transcription factor Pdx1 (pancreatic and duodenal homeobox 1). Moreover, the presence of inductive factors in the Matrigel plus exendin-4 led to an increase in Pdx1 and endocrine genes, such as insulin, islet amyloid polypeptide, glucagon, the glucose transporter GLUT2, chromogranin A and the convertases PC1/3 and PC2 were also detected. Immunocytochemical analysis showed co-localization of insulin and C-peptide, whereas ultrastructural studies revealed the presence of granules. Insulin secretion from cell clusters was detected in the cell culture medium. We identified a population of PaSCs that express the ABCG2+ transporter and have the capacity to transdifferentiate into insulin-producing cells. Although the potential therapeutic application remains to be tested, PaSCs could represent a future option for insulin replacement in diabetes research.

INTRODUCTION

The multidrug resistance transporter ABCG2 (where ABC is ATP binding cassette) has recently been identified as a molecular determinant of the SP (side population) phenotype, a primitive haematopoietic cell population with a multipotential capacity. This transporter is also expressed in a wide variety of stem cells from adult tissues. It is considered a novel stem-cell marker and it could be used, therefore, to localize and select these multipotent cells in different tissues [13].

In the case of the pancreas, there is no consensus on the nature or localization of adult stem cells [4,5]. Several authors have described transdifferentiation in different models in pancreas [610]. Moreover, experimental evidence indicates that the β-cell and other islet cells participate in intra-islet regeneration [1113]. Furthermore, a subpopulation of SP cells that co-expresses ABCG2, MDR1 (multi-drug resistance 1) and nestin (a type VI intermediate-filament protein) has been identified and isolated from the adult human pancreas, and it may be a potential source of adult multipotential stem/progenitor cells useful for the production of islet tissue for transplantation into diabetic subjects [14]. However, it is not known whether, in addition to the SP phenotype, other types of pancreatic cells, such as the mesenchymal lineage, also express the ABCG2 transporter and represent other adult stem cells that participate in endocrine differentiation.

In order to use adult pancreatic stem cells in the potential therapeutic treatment of diabetes, we require more information on the lineage of these cells, the molecular mechanisms that allow them to remain undifferentiated and the key factors that determine their differentiation [15].

PaSCs (pancreatic stellate cells) are associated with diseases such as chronic pancreatitis and fibrosis [16,17]. Their hepatic counterparts have been described as exhibiting other properties, including those of progenitor cells, but in pancreas this role awaits further investigation [18].

The aim of the present study was to identify new ABCG2+ pancreatic cell populations and to explore whether they exhibit the properties of progenitor cells.

MATERIALS AND METHODS

Selection of mitoxantrone-resistant cells and differentiation treatments

Fresh pancreata were removed from lactating rats, as approved by the Institutional Animal Care and Use Committee of the IDIBAPS Research Institute, Barcelona, Spain, and digested with collagenase to obtain a primary cell culture. The cell digestion sample was rapidly divided into two groups of homogeneous suspension cells. One group was grown in DMEM (Dulbecco's Modified Eagle Medium; Gibco–BRL) containing 25 mM glucose and supplemented with 10% (v/v) FCS (fetal-calf serum) and 100 units/ml penicillin (basal medium) and used as a control culture. The ABCG2+ cell line was generated by mitoxantrone (8 μM) selection in the second cell suspension group. This drug acted through multidrug transporter systems and was diluted in the basal medium described above. The medium was removed every 2 days and it was gently washed to remove all the dead cells (Stage 1). The cells that survived mitoxantrone exposure were then placed in fresh DMEM/F12 (containing 11.1 mM glucose) (Gibco-BRL) supplemented with 10% FCS. Mitoxantrone-resistant cells have the ability to proliferate until reaching confluence in a monolayer with a fibroblast-like appearance (Stage 2). In Stage 2, several three-dimensional structures known as ‘cell clusters’ appeared spontaneously, the so-called ‘Stage 3’. All the cultures were maintained at 37 °C in a humidified atmosphere containing 5% CO2. The cells were kept for about 2 years as an immortalized cell line. Following drug selection, the cell karyotype was analysed. The modal number of chromosomes (n=42) was normal (results not shown) and no loss of cellular phenotype was observed. Moreover, to confirm the presence of ABCG2 transporters, functional and expression studies were performed using flow cytometry, immunofluorescence and RT (reverse-transcription)–PCR techniques (see below).

To induce cell differentiation, three media were tested for 3 weeks: M 1 (Medium 1) comprised DMEM (24 mM glucose) supplemented with 10 ng/ml HGF (hepatocyte growth factor; R&D Systems), 0.5 pmol/l betacellulin (R&D Systems) and 10 mM nicotinamide (Sigma); M 2 (Medium 2) comprised DMEM/F12 (11.1 mM glucose) supplemented with 0.1 nM exendin-4 (Sigma); and M 3 (Medium 3) comprised DMEM/F12 (11.1 mM glucose) supplemented with ITS (insulin/transferrin/selenium; BD Biosciences) plus 0.1 nM exendin-4 (Sigma). Matrigel™ was supplemented in M 3 as a base membrane preparation from Engelbreth–Holm–Swarm mouse tumour cells (BD Biosciences). Matrigel™ was used, following the manufacturer's instructions, at a dilution of 1:3 and with a 1 h gelling time at 37 °C. Inverted optical microscopy was used to monitor morphological changes. In the differentiation experiments, 100 three-dimensional structures obtained from cell passages 12–15 were hand-picked and dispersed using a solution of trypsin.

Measurement of apoptotic cell death

To assess drug selection, we analysed apoptosis of cells stained with Hoechst 33342 dye. The bisbenzimidazole dye penetrates the plasma membrane and stains DNA in cells without permeabilization. In contrast with normal cells, the nuclei of apoptotic cells have highly condensed chromatin that is uniformly stained by this dye. These morphological changes may be visualized by fluorescence microscopy. After exposure to mitoxantrone in growth medium, the cells on coverslips were fixed in 4% (w/v) paraformaldehyde at room temperature (21 °C) for 20 min. They were then stained in the incubation buffer for 15 min with Hoechst 33342 dye at a concentration of 1 μg/ml. The morphological changes were examined under UV illumination using a fluorescence microscope (IX70-FL; Olympus). The dye was excited at 340 nm, and emission was filtered with a 510 nm barrier filter. To quantify the apoptotic process, cells with fragmented or condensed DNA and cells with normal DNA were counted. The results are expressed as apoptotic cells as a percentage of total cells.

Flow cytometry

Flow-cytometric analysis was performed with a FACScan flow cytometer (BD Biosciences) equipped with an argon-ion laser at 488 nm. Briefly, cells were harvested from the tissue-culture flasks following treatment with trypsin. The cell suspension, at a concentration of 1.0×106 cells/ml, was fixed for 1 h in 2% paraformaldehyde at 4 °C, washed once in PBS and stained overnight with primary antibodies, as summarized in Table 1. The labelled cells were analysed on a FACSCalibur (BD Biosciences) instrument by acquisition of 10 000 gated events. Data were stored as listmode files and analysed with CellQuest™ (BD Biosciences) data-acquisition and Summit Workstation software.

Table 1
Antibodies for the characterization of the cell culture

Anti-rabbit Cy™2 Jackson Immuno Research 711-225-152 and anti-mouse Cy™3 Jackson Immuno Research 705-225-147 were also used.

Anti-rat or human Dilution Fluorochrome Source 
Nestin 1/100 Cy™2 556309 BD Biosciences Pharmingen, Mississauga, CA, U.S.A. 
Thy-1.1 (CD90) 1/100 Fluorescein 554897 BD Biosciences Pharmingen, Mississauga, CA, U.S.A. 
N-CAM 1/100 Cy™2 L0015, USBiological, Massachusetts, U.S.A. 
GFAP 1/500 Cy™2 Chemicon International 
α-Actin 1/10 Cy™2 A0760-22 USBiological, Massachusetts, U.S.A. 
Vimentin 1/100 Cy™2 MAB3400 Chemicon International 
Chromogranin A 1/100 Cy™3 RB-9003-PO Bionova, Westinghouse, CA, U.S.A. 
Desmin 1/100 Cy™2 MCA849 Serotec, Oxford, U.K. 
C-peptide 1/100 Cy™3 4023-01 Linco Research, Missouri, U.S.A. 
Cytokeratin-19 1/1000 Cy™2 Z-0622 Dako, Carpinteria, CA, U.S.A. 
Insulin 1/2000 Cy™2 A0564, Dako, Carpinteria, CA, U.S.A. 
Anti-rat or human Dilution Fluorochrome Source 
Nestin 1/100 Cy™2 556309 BD Biosciences Pharmingen, Mississauga, CA, U.S.A. 
Thy-1.1 (CD90) 1/100 Fluorescein 554897 BD Biosciences Pharmingen, Mississauga, CA, U.S.A. 
N-CAM 1/100 Cy™2 L0015, USBiological, Massachusetts, U.S.A. 
GFAP 1/500 Cy™2 Chemicon International 
α-Actin 1/10 Cy™2 A0760-22 USBiological, Massachusetts, U.S.A. 
Vimentin 1/100 Cy™2 MAB3400 Chemicon International 
Chromogranin A 1/100 Cy™3 RB-9003-PO Bionova, Westinghouse, CA, U.S.A. 
Desmin 1/100 Cy™2 MCA849 Serotec, Oxford, U.K. 
C-peptide 1/100 Cy™3 4023-01 Linco Research, Missouri, U.S.A. 
Cytokeratin-19 1/1000 Cy™2 Z-0622 Dako, Carpinteria, CA, U.S.A. 
Insulin 1/2000 Cy™2 A0564, Dako, Carpinteria, CA, U.S.A. 

Drug uptake and retention assays

To allow cell attachment to the surface and to obtain the conditions that would allow optimal growth, the cells were seeded in 12-well tissue-culture plates (TPP, Trasadingen, Switzerland) at a final concentration of 1.0×105 cells/ml, 1 ml total volume/well, 24 h before the dye uptake/retention experiments. In the presence or absence of verapamil (5 μM), drug uptake was measured by adding the fluorescent substrate (mitoxantrone, 8 μM) to the culture basal medium for 1 h at 37 °C. After 1 h of incubation, cells were washed and resuspended in dye-free culture basal medium, and the blocker was maintained to evaluate its effect on dye retention. When dye compatibility was allowed, dead cells were excluded by simultaneous staining with 1 μg/ml propidium iodide (stock solution) [19].

Immunocytochemistry

Cells at distinct stages were plated in eight-well chamber slides (Lab-Tek™; Nalge Nunc International). After fixation in 4% paraformaldehyde, they were then blocked for 10 min at room temperature in 1% BSA and 0.2% saponin and incubated overnight at 4 °C with the primary antibody diluted in blocking solution (Table 1), standard protocols being followed. The bound antibody was visualized with a fluorescent secondary antibody (Table 1) under a Leica fluorescence microscope. BSA (1%, w/v) was used instead of primary antibodies for control slides. In differentiation experiments, total cells were isolated from cellular clusters and then spread on poly-L-lysine-coated slides, using a Cytospin 3 (Shandon) cell-preparation system. The slides were subsequently fixed in ethanol/acetone (1:1, v/v) for 10 min at −20 °C and processed by immunocytochemical techniques.

Oil Red O staining and vitamin A autofluorescence

The Oil Red O treatment for staining intracellular lipid droplets was performed in accordance with a modification of the method described by Catalano and Lillie [20]. Briefly, cells were fixed with buffered 4% (v/v) formalin for 10 min at room temperature, washed in PBS (twice, 1 min each time) and stained with a saturated solution of Oil Red O (Sigma) in acetone/ethanol (1:9, v/v) for 45 min at room temperature. Subsequently, cells were washed again under flowing cold water for 10 min, which resulted in red staining of lipid droplets. Finally, cells were examined by optical microscopy.

Vitamin A storage in lipid vesicles was detected by fluorescence microscopy using an excitation wavelength of 320–380 nm to visualize the characteristic rapidly fading blue–green fluorescence of this vitamin.

RT–PCR analysis

Total RNA was extracted from several stages of the cultures using an RNeasy extraction kit (Qiagen). To remove genomic DNA contamination, DNase I (Invitrogen) was used, the manufacturer's protocol being followed. A 1 μg portion of total RNA was reverse-transcribed in a buffer solution containing 25 nmol/l MgCl2, 100 mmol/litre Tris (pH 8.3), 500 mmol/litre KCl, RNAguard (39 units/ml; Pharmacia), M-MLV-RT (200 units/ml; BRL–Gibco), 10 mmol/l deoxynucleotide triphosphate(s) and random hexamer [d(N6)5′PO4; Pharmacia] priming. Incubations of 30 min at 42 °C, 5 min at 94 °C and 5 min at 5 °C were carried out in a total volume of 20 μl. cDNA was stored at −80 °C until used. All PCRs were performed using 38 cycles. Parallel RT–PCR reactions without reverse transcriptase were performed for each sample. PCR products were visualized with 1%-(w/v)-agarose-gel electrophoresis and ethidium bromide staining. The oligonucleotide sequences used for PCR amplification are summarized in Table 2.

Table 2
Primers for reverse transcription-polymerase chain reaction
Gene Primers Amplicon (bp) T(°C) Accession no. 
ABCG2 Forward: ACAAAAGCTGAATTAGATCAACTT 641 59 NM_181381 
 Reverse: GAGATTCACCAAGAGGCCAG    
Nestin Forward: GCGGGGCGGTGCGTGACTAC 523 62 NM_012987 
 Reverse: GCAAGGGGGAAGAGAAGGATGT    
Thy1.1 Forward: CGCTTTATCAAGGTCCTTACT 343 52 P01830 
 Reverse: GCGTTTTGAGATATTTGAAGGT    
NCAM Forward: CAGCGTTGGAGAGTCCAAAT 300 54 NM_031521 
 Reverse: TTAAACTCCTGTGGGGTTGG    
Vimentin Forward: GCCAGCAGTATGAAAGTGTG 496 60 NM_031140 
 Reverse: AGTGGGTGTCAACCAGAGGAA    
Desmin Forward: AGACTTGACTCAGGCAGCCAAT 384 60 NM_022531 
 Reverse: CGGAAGTTGAGAGCAGAGAAGG    
GFAP Forward: TGGCCACCAGTAACATGCA 538 60 NM_017009 
 Reverse: GACTCCTTAATGACCTCGCCAT    
α-Actin Forward: ATCCGATAGAACACGGCATC 500 60 RGD:1559154 
 Reverse: AGAAGAGGAAGCAGCAGTGG    
GLP 1R Forward: TCTCTTCTGCAACCGAACCT 350 55 NM_021728 
 Reverse: CTGGTGCAGTGCAAGTGTCT    
CK 19 Forward: ACAGCCAGTACTTCAAGACC 690 57 AY464140 
 Reverse: CTGTGTCAGCACGCACGTTA    
Chromogranin A Forward: CGGCTTTGGCGCTTCTGCT 401 60 NM_021655 
 Reverse: CTTGGAGGGGGCTTCTGATGCT    
Pdx-1 real-time Forward: CCGCGTTCATCTCCCTTTC   NM_022852 
 Reverse: CTCCTGCCCACTGGCTTTT    
Pdx-1 Forward: GAGCCCAGCCGCGTTCATCT 318 60 NM_022852 
 Reverse: CCCCGCTCGTTGTCCCGCTACTA    
Ngn3 Forward: TGGCGCCTCATCCCTTGGATG 160 60 MN_021700.1 
 Reverse: CAGTCACCTGCTTCTGCTTCG    
Glucagon Forward: ATCATTCCCAGCTTCCCAGA 161 57 NM_012707 
 Reverse: CGGTTCCTCTTGGTGTTCAT    
IAPP Forward: AGTCCTCCCACCAACCAATGT 220 62 NM_012586 
 Reverse: AGCACAGGCACGTTGTTGTAC    
Insulin Forward: TGGCCCTGTGGATCCG 329 52 NM_019129 
 Reverse: AGTTGCAGTAGTTCTCCAGCTGG    
GLUT2 Forward: GACACCCCACTCATAGTCACAC 270 57 P12336 
 Reverse: CAGCAATGATGAGAGCATGTG    
TBP real-time Forward: TTCGTGCCAGAAATGCTGAA   NM_001004198 
 Reverse: TTCGTGCCAGAAATGCTGAA    
TBP Forward: ACCCTTCACCAATGACTCCTATG 190 60 NM_01004198 
 Reverse: ATGATGACTGCAGCAAATCGC    
Gene Primers Amplicon (bp) T(°C) Accession no. 
ABCG2 Forward: ACAAAAGCTGAATTAGATCAACTT 641 59 NM_181381 
 Reverse: GAGATTCACCAAGAGGCCAG    
Nestin Forward: GCGGGGCGGTGCGTGACTAC 523 62 NM_012987 
 Reverse: GCAAGGGGGAAGAGAAGGATGT    
Thy1.1 Forward: CGCTTTATCAAGGTCCTTACT 343 52 P01830 
 Reverse: GCGTTTTGAGATATTTGAAGGT    
NCAM Forward: CAGCGTTGGAGAGTCCAAAT 300 54 NM_031521 
 Reverse: TTAAACTCCTGTGGGGTTGG    
Vimentin Forward: GCCAGCAGTATGAAAGTGTG 496 60 NM_031140 
 Reverse: AGTGGGTGTCAACCAGAGGAA    
Desmin Forward: AGACTTGACTCAGGCAGCCAAT 384 60 NM_022531 
 Reverse: CGGAAGTTGAGAGCAGAGAAGG    
GFAP Forward: TGGCCACCAGTAACATGCA 538 60 NM_017009 
 Reverse: GACTCCTTAATGACCTCGCCAT    
α-Actin Forward: ATCCGATAGAACACGGCATC 500 60 RGD:1559154 
 Reverse: AGAAGAGGAAGCAGCAGTGG    
GLP 1R Forward: TCTCTTCTGCAACCGAACCT 350 55 NM_021728 
 Reverse: CTGGTGCAGTGCAAGTGTCT    
CK 19 Forward: ACAGCCAGTACTTCAAGACC 690 57 AY464140 
 Reverse: CTGTGTCAGCACGCACGTTA    
Chromogranin A Forward: CGGCTTTGGCGCTTCTGCT 401 60 NM_021655 
 Reverse: CTTGGAGGGGGCTTCTGATGCT    
Pdx-1 real-time Forward: CCGCGTTCATCTCCCTTTC   NM_022852 
 Reverse: CTCCTGCCCACTGGCTTTT    
Pdx-1 Forward: GAGCCCAGCCGCGTTCATCT 318 60 NM_022852 
 Reverse: CCCCGCTCGTTGTCCCGCTACTA    
Ngn3 Forward: TGGCGCCTCATCCCTTGGATG 160 60 MN_021700.1 
 Reverse: CAGTCACCTGCTTCTGCTTCG    
Glucagon Forward: ATCATTCCCAGCTTCCCAGA 161 57 NM_012707 
 Reverse: CGGTTCCTCTTGGTGTTCAT    
IAPP Forward: AGTCCTCCCACCAACCAATGT 220 62 NM_012586 
 Reverse: AGCACAGGCACGTTGTTGTAC    
Insulin Forward: TGGCCCTGTGGATCCG 329 52 NM_019129 
 Reverse: AGTTGCAGTAGTTCTCCAGCTGG    
GLUT2 Forward: GACACCCCACTCATAGTCACAC 270 57 P12336 
 Reverse: CAGCAATGATGAGAGCATGTG    
TBP real-time Forward: TTCGTGCCAGAAATGCTGAA   NM_001004198 
 Reverse: TTCGTGCCAGAAATGCTGAA    
TBP Forward: ACCCTTCACCAATGACTCCTATG 190 60 NM_01004198 
 Reverse: ATGATGACTGCAGCAAATCGC    

Real-time PCR

Real-time PCR was performed using the double-stranded-DNA-binding dye Power SYBR Green PCR Master Mix using the ABI Prism 7900HT Sequence Detection System (Applied Biosystems). The sequences of primers used in the present study are described in Table 2. A standard curve was generated from five serial dilutions of an ARIP (rat pancreatic ductal)-cell-line-synthesized cDNA for Pdx1 (pancreatic and duodenal homeobox I). Samples were analysed in triplicate, negative controls were included, and PCR products were verified using dissociation-curve analysis immediately after RT–PCR. Expression of the target gene was determined by normalizing to the level of the housekeeping gene coding for TBP (TATA-binding protein). Results were analysed using SDS2.1 software (Applied Biosystems).

Infection of recombinant adenovirus

The adenovirus expressing mouse Ngn3 (neurogenin 3) and a control adenovirus expressing bacterial β-galactosidase were prepared with the Adeno-X system (Clontech). For viral infection, cells were incubated with adenoviruses at an MOI (multiplicity of infection) of 50 for 2 h at 37 °C. The virus-containing medium was then replaced, and cells were cultured in the basal medium for 48 h.

Assays for insulin secretion

Insulin secretion was measured by static incubation as previously described [21]. Cells were plated in 12-well plates at 1.0×106 cells/well or 100-cell clusters (mitoxantrone-resistant cells at Stage 3). Cells were pre-incubated for 1 h in KRB (Krebs–Ringer buffer) and then incubated for 60 min at 37 °C in KRB containing 120 mM NaCl, 5 mM KCl, 2.5 mM CaCl2, 1.1 mM MgCl2, 25 mM NaHCO3, 0.1% BSA and glucose at various concentrations. KRB media were collected and stored at −70 °C until assayed for insulin. Insulin was measured using an ELISA kit (Mercodia) that recognizes only mature insulin, the manufacturer's instructions being followed. This assay has <20% cross-reactivity with pro-insulin.

Electron microscopy

Cell clusters with and without Matrigel™ were pre-fixed in Karnovsky's fixative [2.5% (v/v) glutaraldehyde and 2.0% paraformaldehyde] in 0.1 mol/l phosphate buffer, pH 7.3. The tissues were fixed in 1% freshly prepared osmium tetroxide and 0.8% potassium ferrocyanide in doubly distilled water for 1 h. After three washes in cold doubly distilled water, the tissue blocks were dehydrated through ascending concentrations of acetone (30, 50, 70, 95 and 100%), and three changes of 100% acetone. They were then embedded in Spurr resin [22] and polymerized at 60 °C. The embedded blocks were sectioned using a diamond knife (Diatome) on a Ultracut UCT (Leica Microsystems) ultramicrotome. Ultrathin sections were placed on a copper grid and stained with uranyl acetate and lead citrate [23] before examination under a JEM 010 (Jeol Ltd., Tokio, Japan) electron microscope equipped with a Gatan/MegaScan model 792 digital camera (Gatan Inc.).

Statistics

Results are expressed as means±S.E.M.). Student's t test was used for paired data. A P value of <0.05 was considered significant.

RESULTS

Establishment of ABCG2+ pancreatic rat cell culture

Several pancreata of lactating rats were collected and digested with collagenase. The specific drug mitoxantrone was then used for 2 weeks to select cells. A signal of apoptosis was observed between days 2 and 4 after drug selection. This signal decreased significantly after day 6 until by day 12 of selection it totally disappeared (Figure 1). After this period, several attached cells started to grow, forming a monolayer phase with only one cellular phenotype identified as fibroblastoid-like cells (Stage 2) (Figure 3B below, panels a and b). To determine whether mitoxantrone-resistant cells were able to increase drug accumulation, the intracellular drug concentrations of mitoxantrone were determined by flow cytometry in the unselected cells or after cell selection, as described in the Materials and methods section, using verapamil as a specific ABCG2 inhibitor. As shown in Figure 2(A), there was a significant increase of the intracellular concentration of the drug in the selected cells when the mitoxantrone plus verapamil was used. This increment was 129.8%, but no significant increase was observed in the unselected cells used as a control. Moreover, gene expression of ABCG2 by RT–PCR confirmed that the primary culture of drug-resistant cells expressed this transporter (Figure 2B).

Mitoxantrone-induced apoptosis in digested fresh pancreata as measured by Hoechst 33342 staining

Figure 1
Mitoxantrone-induced apoptosis in digested fresh pancreata as measured by Hoechst 33342 staining

Results are apoptotic cells expressed as a percentage of the total number of cells (means±S.E.M. values for data obtained from four independent experiments performed in four or five wells; **P<0.001; ##P<0.01).

Figure 1
Mitoxantrone-induced apoptosis in digested fresh pancreata as measured by Hoechst 33342 staining

Results are apoptotic cells expressed as a percentage of the total number of cells (means±S.E.M. values for data obtained from four independent experiments performed in four or five wells; **P<0.001; ##P<0.01).

ABCG2 expression, and drug uptake and retention assays, in primary cell cultures (mitoxantrone-resistant cells and unselected cells)

Figure 2
ABCG2 expression, and drug uptake and retention assays, in primary cell cultures (mitoxantrone-resistant cells and unselected cells)

(A) A 1 h drug accumulation assay with and without verapamil. The cells were preincubated with 5 μM verapamil for 15 min. Subsequently, cells were treated with 8 μM mitoxantrone and assayed for drug accumulation as described in the Materials and methods section. Each result is the mean±S.D. for three experiments under each condition. Verapamil increased the intracellular concentration of mitoxantrone in the mitoxantrone-selected drug-resistant cells. A representative histogram is shown. NS, not significant. (B) ABCG2 expression in the cells from cultures was determined by RT–PCR: unselected cells (lane 1) and mitoxantrone-resistant cells at Stage 2 (lane 2). The ARIP cell line was used as a positive control of the reaction (Control), ‘-RT’ corresponds to amplification in which reverse transcriptase was excluded from the reaction (negative control).

Figure 2
ABCG2 expression, and drug uptake and retention assays, in primary cell cultures (mitoxantrone-resistant cells and unselected cells)

(A) A 1 h drug accumulation assay with and without verapamil. The cells were preincubated with 5 μM verapamil for 15 min. Subsequently, cells were treated with 8 μM mitoxantrone and assayed for drug accumulation as described in the Materials and methods section. Each result is the mean±S.D. for three experiments under each condition. Verapamil increased the intracellular concentration of mitoxantrone in the mitoxantrone-selected drug-resistant cells. A representative histogram is shown. NS, not significant. (B) ABCG2 expression in the cells from cultures was determined by RT–PCR: unselected cells (lane 1) and mitoxantrone-resistant cells at Stage 2 (lane 2). The ARIP cell line was used as a positive control of the reaction (Control), ‘-RT’ corresponds to amplification in which reverse transcriptase was excluded from the reaction (negative control).

Characterization of mitoxantrone-resistant cells

After cell drug selection, several cellular expansions without mitoxantrone were performed. During cell growth, three-dimensional structures (cell clusters) were formed in the cultures (Stage 3) with an average diameter of 29.8±11.4 μm (Figure 3B, panels a–e). There was a 2.5-fold decrease in the proliferation ratio compared with that of the monolayer (results not shown), which indicated the presence of distinct cellular stages in the same cultures. Positive staining for nestin and the cell-surface protein Thy1.1 (the two stem-cell markers proposed) was observed in the cells obtained after selection, and no evidence of positive signals was detected in the unselected cells (Figure 3A). Moreover, although CK 19 (cytokeratin 19) was negative in all passages studied from mitoxantrone-resistant cells (see Supplementary Figures S1 and S2A at http://www.BiochemJ.org/bj/421/bj4210181add.htm), nestin and Thy1.1 were positive. However, this positivity decreased in the highest passages from mitoxantrone-resistant cells, without total negativity (see Supplementary Figure S2C). To confirm these results, gene expression analysis by RT–PCR was performed for these markers in unselected cells (control cells) and in mitoxantrone-resistant cells. The results showed that, whereas no signals were detected in unselected cells, expression in the mitoxantrone-resistant cells was observed in both stages studied (Figure 3C). However, when quantification of the progenitor markers (nestin and Thy1.1) were assessed by flow cytometry, differences were detected between cells growing in mitoxantrone-resistant cells at Stage 2 and in mitoxantrone-resistant cells at Stage 3 (nestin: 85.97±10.3% in Stage 2 versus 29.5±7.9% in Stage 3; Thy1.1: 91.3±3.4% in Stage 2 versus 12.9±6.5% in Stage 3) without any significant differences in the mean fluorescence intensity (Supplementary Figure S2C). Moreover, most of the mitoxantrone-resistant cells displayed strong labelling for N-CAM (neural cell adhesion molecule) (Figure 3A) and a strong auto-fluorescence signal was also observed (Figure 3A). To examine the phenotype of the mitoxantrone-resistant cell cultures, several markers were studied: vimentin (mesenchymal lineage), desmin (intermediate filament involved in the myogenic differentiation), α-actin (smooth-muscle actin filament), GFAP (glial fibrillary acidic protein, an intermediate filament expressed in neural stem cells) and chromogranin A (present in the secretory granules of neuroendocrine cells). All these markers were detected in mitoxantrone-resistant cells at Stage 2 and Stage 3 by immunocytochemistry (Figures 4A and 4B) and confirmed by RT–PCR (Figure 4C). However, the expression of these markers was slightly lower at Stage 3 than in Stage 2, as was confirmed by flow cytometry (see Supplementary Figure S2C). As cells showed a high autofluorescence signal, Oil Red O staining was performed to identify the presence of liposoluble material in the cell cytoplasm. Positive droplets were detected in the cytoplasm, thus confirming the presence of liposoluble material in the cell cytoplasm, compatible with vitamin A, in most of the mitoxantrone-resistant cells in our primary culture. In addition, the percentage of Oil Red O staining observed in the slides was lower in Stage 2 than in Stage 3. Moreover, the characteristic fading blue–green fluorescence confirmed that these droplets corresponded to vitamin A. The overall characterization of our cultures indicated that the population selected corresponded to pancreatic stellate cells, the origin of which was mesenchymal. Moreover, the cells displayed several degrees of activation, as shown by the different amounts of cytoplasmic vitamin A recorded between Stages 2 and 3 (Figures 4A and 4B).

Establishment and characterization of a cell line from the mitoxantrone-resistant cell population

Figure 3
Establishment and characterization of a cell line from the mitoxantrone-resistant cell population

(A) Nestin, Thy1.1 and N-CAM protein expression was detected by immunostaining in culture from mitoxantrone-selected drug-resistant cells. Autofluorescence was also observed. Hovewer, no expression was observed in cultures from the unselected cells used as a control. The unselected cells were incubated in 1μg/ml DAPI (4′,6-diamidino-2-phenylindole) solution for 30 min in the dark (original magnification ×20). (B) The mitoxantrone-resistant cells were overgrown by cells with a fibroblastoid morphology (panels a and b). Spontaneously, some cells began to form three-dimensional cell clusters (panels c, d and e). (C) The immunophenotyping results were confirmed by RT–PCR of 1 μg of total RNA of the cells: unselected cells (lane 1), mitoxantrone-resistant cells at Stage 2 (lane 2) and mitoxantrone-resistant cells at Stage 3 (lane 3). The ARIP cell line was used as a control for the reaction (lane 4). The primers were designed across intron(s), when possible, and their sequences and product sizes are listed in Table 2. The thermal cycle was repeated 38 times for all the genes analysed.

Figure 3
Establishment and characterization of a cell line from the mitoxantrone-resistant cell population

(A) Nestin, Thy1.1 and N-CAM protein expression was detected by immunostaining in culture from mitoxantrone-selected drug-resistant cells. Autofluorescence was also observed. Hovewer, no expression was observed in cultures from the unselected cells used as a control. The unselected cells were incubated in 1μg/ml DAPI (4′,6-diamidino-2-phenylindole) solution for 30 min in the dark (original magnification ×20). (B) The mitoxantrone-resistant cells were overgrown by cells with a fibroblastoid morphology (panels a and b). Spontaneously, some cells began to form three-dimensional cell clusters (panels c, d and e). (C) The immunophenotyping results were confirmed by RT–PCR of 1 μg of total RNA of the cells: unselected cells (lane 1), mitoxantrone-resistant cells at Stage 2 (lane 2) and mitoxantrone-resistant cells at Stage 3 (lane 3). The ARIP cell line was used as a control for the reaction (lane 4). The primers were designed across intron(s), when possible, and their sequences and product sizes are listed in Table 2. The thermal cycle was repeated 38 times for all the genes analysed.

Mitoxantrone-resistant cells were phenotyped by immunofluorescence and RT–PCR using pancreatic stellate markers

Figure 4
Mitoxantrone-resistant cells were phenotyped by immunofluorescence and RT–PCR using pancreatic stellate markers

(A) Mitoxantrone-resistant cells at Stage 2 express the markers α-actin (Alfa-actin), GFAP, vimentin, desmin and chromogranin A. To confirm the presence of vitamin A stored in the fat droplets, Oil Red O staining was performed. (B) Disaggregated mitoxantrone-resistant cells at Stage 3 were immunophenotyped for the same markers, including the Oil Red O staining. Negative controls (Neg.) were used (original magnification ×20). (C) These results were confirmed by RT–PCR using 1 μg of total RNA from mitoxantrone-resistant cells in both stages (Stage 2 and Stage 3). Control cell lines were used as a control reaction.

Figure 4
Mitoxantrone-resistant cells were phenotyped by immunofluorescence and RT–PCR using pancreatic stellate markers

(A) Mitoxantrone-resistant cells at Stage 2 express the markers α-actin (Alfa-actin), GFAP, vimentin, desmin and chromogranin A. To confirm the presence of vitamin A stored in the fat droplets, Oil Red O staining was performed. (B) Disaggregated mitoxantrone-resistant cells at Stage 3 were immunophenotyped for the same markers, including the Oil Red O staining. Negative controls (Neg.) were used (original magnification ×20). (C) These results were confirmed by RT–PCR using 1 μg of total RNA from mitoxantrone-resistant cells in both stages (Stage 2 and Stage 3). Control cell lines were used as a control reaction.

In vitro endocrine differentiation of mitoxantrone-resistant cells (Stage 3)

Numerous studies suggest that the culture medium and the soluble factors added to the cultures play an important role in the maturation of different cell types [24]. Our aim was to demonstrate the capacity of these selected cells to differentiate towards endocrine cells by the use of media described for previous cell models (see the Materials and methods section). M 1 was used as a differentiation cocktail for 3 weeks. The mitoxantrone-resistant cells showed spontaneous differentiation when grown in basal medium DMEM/F12 (11.1 mM glucose) with 10% FCS, as indicated by the presence of a weak band of the transcription factor Pdx1. No difference in the gene coding for Pdx1 was found when the cultures were treated with M 1 (Figure 5A) [25]. Cultures obtained before cell selection (unselected cells) used as control cells and treated with the same medium (M 1) were always negative (Figure 5A). However, a weak signal corresponding to the transcript of Ngn3, the Ngn3 gene, was observed only in treated mitoxantrone-resistant cells with M 1 (Figure 5A). This gene plays a key role in β-cell differentiation [26]. To verify whether other genes implicated in the endocrine differentiation pathway could be induced in vitro, isolated cell clusters from mitoxantrone-resistant cells at Stage 3 were infected with an adenovirus–Ngn3 construct. After infection, there was an increase in Pdx-1, as well as in other transcription factors that were detected earlier, such as Neuro D and Pax-4 (paired box 4), both of which are involved in endocrine cell differentiation (Figure 5B). Quantitative real-time PCR indicated that both media (M 1 and M 2) increased Pdx-1 mRNA basal expression 3.9±2.9-fold and 4.9±3.4-fold respectively (n=3; P<0.05, Figure 5C). Moreover, M 2 supplemented with exendin-4 showed a higher increase compared with M 1. Exendin-4 is known to produce activity in islet cells in the pancreas via GLP-1 (glucagon-like peptide-1) receptors, giving rise to β-cell activity and insulin release. It also participates in control of glucose and in fat metabolism. In addition, the presence of GLP-1 receptors, confirmed by RT–PCR, indicates possible effects in these cells (Figure 5C). Finally, in order to simulate a specific cellular niche for endocrine differentiation, 100 cell clusters were cultured with Matrigel™ as a semisolid medium supplemented with exendin-4 (M 3). After 2 weeks of culture under these experimental conditions, the transcription of Pdx-1 increased. In addition, a strong insulin signal was detected for the first time. Moreover, IAPP, glucagon, GLUT2 and the convertases PC1/3 and PC2 were also detected after this treatment (Figure 5A). By contrast, expression of the transcription factor p48 and other exocrine genes, such as that coding for amylase, were not detected (results not shown). Interestingly, CK 19 expression was observed. Flow cytometry showed that 31.21±5.2 % of cells were C-peptide-positive, and 39.54±2.1 % were also positive for Pdx-1, whereas a few cells were positive for nestin (0.8±0.065%). In addition, insulin and C-peptide were co-localized in all the cells studied (Figure 6). We also observed that C-peptide co-localized with vimentin and CK 19 in some cells, but not with the other markers studied (α-actin and GFAP) (Figure 6). Moreover, ultrastructural studies showed that the monolayer cells displayed characteristics compatible with activated pancreatic stellate cells, namely extensive hypertrophy of the rough endoplasmic reticulum, abundant lysosomes in the cytoplasm, few lipid droplets and the presence of abundant fibres of collagen in the extracellular compartment (Figure 7A, panels a–d). The clusters that grew in Matrigel™ plus exendin-4 showed substantial ultrastructural changes, with a smaller and homogenous cell size, round nuclei and electron-dense homogenous chromatin, a significant increase in the number of mitochondria and lipid droplets in their cytoplasm; abundant electron-dense granules were also observed (Figure 7A, panels e–h).

Pancreatic gene expression profiles

Figure 5
Pancreatic gene expression profiles

(A) RT–PCR analysis of the gene expression profiles was performed using 1 μg of total RNA from unselected cells and mitoxantrone-resistant cells. Genes from a pancreatic lineage were identified and their expression compared using different media, namely basal medium (M 0), M 1 and M 3. Different cell lines were used as a control for the RT–PCR reactions. All primers were designed across intron(s), when possible, and their sequences and product sizes are listed in the Table 2. The thermal cycle was repeated 38 times in the gene analysis. (B) Effect of adenovirus-mediated ectopic expression of Ngn3 on key developmental transcription factors and pancreatic marker expression in mitoxantrone-resistant cells at Stage 3. Gene profiling expression of Neuro D, Pax 6, Pax 4, Pdx-1, nestin, ABCG2, insulin, glucagon, IAPP, PC1/3, PC2, CK19, Thy1.1 and N-CAM mRNA were assayed by RT–PCR in mitoxantrone-resistant cells at Stage 3 after infection with AdCMV–Ngn3 or, as a control of efficiency of infection, with AdCMV–β-gal. mRNA was harvested from mitoxantrone-resistant cells at Stage 3 24 h after infection with AdCMV–Ngn3 or AdCMV–β-gal. The numbers of PCR cycles were 25 for TBP, 30 for Ngn3 and 38 for the other transcripts. (C) Gene expression of Pdx-1 and GLP-1 receptor was detected by RT–PCR and their expression quantified by real-time RT–PCR in mitoxantrone-resistant cell cultures with M 0, M 1 and M 2. The Pdx-1 values from cell cultures with M 0 were taken as the reference and values were normalized to housekeeping TBP mRNA levels. Results in the histogram are means±S.E.M. for three independent experiments performed in duplicate; *P<0.05.

Figure 5
Pancreatic gene expression profiles

(A) RT–PCR analysis of the gene expression profiles was performed using 1 μg of total RNA from unselected cells and mitoxantrone-resistant cells. Genes from a pancreatic lineage were identified and their expression compared using different media, namely basal medium (M 0), M 1 and M 3. Different cell lines were used as a control for the RT–PCR reactions. All primers were designed across intron(s), when possible, and their sequences and product sizes are listed in the Table 2. The thermal cycle was repeated 38 times in the gene analysis. (B) Effect of adenovirus-mediated ectopic expression of Ngn3 on key developmental transcription factors and pancreatic marker expression in mitoxantrone-resistant cells at Stage 3. Gene profiling expression of Neuro D, Pax 6, Pax 4, Pdx-1, nestin, ABCG2, insulin, glucagon, IAPP, PC1/3, PC2, CK19, Thy1.1 and N-CAM mRNA were assayed by RT–PCR in mitoxantrone-resistant cells at Stage 3 after infection with AdCMV–Ngn3 or, as a control of efficiency of infection, with AdCMV–β-gal. mRNA was harvested from mitoxantrone-resistant cells at Stage 3 24 h after infection with AdCMV–Ngn3 or AdCMV–β-gal. The numbers of PCR cycles were 25 for TBP, 30 for Ngn3 and 38 for the other transcripts. (C) Gene expression of Pdx-1 and GLP-1 receptor was detected by RT–PCR and their expression quantified by real-time RT–PCR in mitoxantrone-resistant cell cultures with M 0, M 1 and M 2. The Pdx-1 values from cell cultures with M 0 were taken as the reference and values were normalized to housekeeping TBP mRNA levels. Results in the histogram are means±S.E.M. for three independent experiments performed in duplicate; *P<0.05.

Representative images of the co-immunolocalization of different markers by Cytospin 3-prepared cells obtained from disaggregated cellular clusters after treatment with M 3

Figure 6
Representative images of the co-immunolocalization of different markers by Cytospin 3-prepared cells obtained from disaggregated cellular clusters after treatment with M 3

The first, top left, panel (a) shows a representative cell cluster after treatment with M 3. The markers are visualized as follows: in red, C-peptide; in green, insulin, vimentin, CK 19, GFAP and α-actin (Alfa-actin); merges appear yellow. The insulin-secreting mouse pancreatic β-cell line MIN-6 was used as the immunohistochemical control.

Figure 6
Representative images of the co-immunolocalization of different markers by Cytospin 3-prepared cells obtained from disaggregated cellular clusters after treatment with M 3

The first, top left, panel (a) shows a representative cell cluster after treatment with M 3. The markers are visualized as follows: in red, C-peptide; in green, insulin, vimentin, CK 19, GFAP and α-actin (Alfa-actin); merges appear yellow. The insulin-secreting mouse pancreatic β-cell line MIN-6 was used as the immunohistochemical control.

Ultrastructural changes and insulin release in the mitoxantrone-resistant cells at Stage 3 after differentiation treatment with M 3

Figure 7
Ultrastructural changes and insulin release in the mitoxantrone-resistant cells at Stage 3 after differentiation treatment with M 3

(A) Transmission electron micrographs of undifferentiated cells (panels a–d) show high hypertrophy in the rough endoplasmic reticulum (rER), lipid droplets (LD), lysosomes (L) and collagenous fibres (CF). Two types of electron-dense chromatin structure were observed (Ch). However, the differentiated cells (panels e–h) presented a homogenous size with a round nucleus (N), at times indented, abundant mitochondria (M), and electron-dense granules in the cytoplasm (g). (B) Insulin secretion after 1 h of glucose stimulation at 2.8 mM versus 20 mM. The results were normalized to 100 cell clusters (n=3; *P<0.05; Student's t test).

Figure 7
Ultrastructural changes and insulin release in the mitoxantrone-resistant cells at Stage 3 after differentiation treatment with M 3

(A) Transmission electron micrographs of undifferentiated cells (panels a–d) show high hypertrophy in the rough endoplasmic reticulum (rER), lipid droplets (LD), lysosomes (L) and collagenous fibres (CF). Two types of electron-dense chromatin structure were observed (Ch). However, the differentiated cells (panels e–h) presented a homogenous size with a round nucleus (N), at times indented, abundant mitochondria (M), and electron-dense granules in the cytoplasm (g). (B) Insulin secretion after 1 h of glucose stimulation at 2.8 mM versus 20 mM. The results were normalized to 100 cell clusters (n=3; *P<0.05; Student's t test).

Finally, insulin secretions of several sets of cell clusters were measured by static incubation at low (2.8 mM) and high (20 mM) levels of glucose. The insulin-secretory response to high glucose increased by 44% (911.8±31.7 pg/cell cluster) versus low glucose (635.7±13.4 pg/cell cluster) (P<0.05) (Figure 5B). However, insulin secretion in undifferentiated cells obtained during the first expansion phase of the mitoxantrone-resistant cell (Stage 2) was below the detection limit for the two glucose treatments tested (Figure 7B).

DISCUSSION

The therapies currently used to treat diabetes are unsatisfactory, as they do not offer a cure and cannot prevent the development of the complications associated with the disease [27]. The need for new therapeutic strategies, including genetic and cellular approaches, has led researchers to explore alternative cellular sources of insulin-producing cells, such as embryonic stem cells or stem cells isolated from adult tissues [28]. There is considerable controversy regarding the progenitor of the pancreatic endocrine β-cell tissues. Ductal, acinar, SP and, more recently, multipotent fibroblast-like cells, in the adult human exocrine pancreas have all been proposed as potential progenitors [29,30]. However, the specific signals that induce transdifferentiation, as well as the proliferative signals to expand a specific type of progenitor must be identified [3133]. Moreover, to identify multipotential cells, specific cell markers are required. The mechanisms controlling islet-cell neogenesis are largely unknown. The formation of new islets, as seen from the ductal epithelium, has long been considered one of the mechanisms of normal islet growth after birth and in regeneration and suggests the presence of pancreatic stem cells. It has been documented by Bonner-Weir and Sharma [4] that there are abundant endocrine progenitor cells in the neonatal pancreas, but little is known about their relative proportions or even phenotypes. Two candidate precursor markers have been proposed: a haematopoietic stem cell marker, namely c-Kit, and a neural stem-cell marker, namely nestin. In a pre- and post-natal rat pancreas, these two markers decreased proportionally with age, coinciding with the appearance and development of endocrine cell types [34]. Multidrug transporter systems, and more specifically, the ABCG2 subtype, have been proposed as a new marker of stem cells. In the human pancreas, this marker has been identified in the primitive haematopoietic stem cells with an SP phenotype, located in the islet or around acinar cells. Other pancreatic cells that can express this marker, however, have not yet been investigated [35]. Rats and mice express the ABCG2 gene in several tissues, such as intestine, kidney and testis, but the pancreas has not been investigated to date [36].

Although the precise physiological function of these transporters in progenitor and differentiated cells is unknown, it has been postulated that they confer protection against a number of xenobiotics, thus maintaining the regenerative capacity of the tissue [37].

In the present study we isolated ABCG2+ cells with mesenchymal features from adult pancreatic tissue of lactantig rats. Because of their capacity to adhere and to divide quickly, these cells were maintained in culture, as a cell line, for more than 2 years. This approach allowed us to characterize these cells. We demonstrate that ABCG2+ cells correspond to a population of pancreatic cells known as pancreatic stellate cells. Although their origin is still being debated, they share the characteristics of mesenchymal (vimentin), neural (GFAP) and muscular (desmin and α-actin) cells. One of the main characteristics of stellate cells is their capacity to transdifferentiate. When activated, they express α-actin and produce many collagen fibres. They thus resemble myofibroblast cells and may, therefore, be involved in tissue repair [38,39]. Recently it has been described that CD133+ hepatic stellate cells exhibit the properties of progenitor cells and display the capacity to develop into endothelial-like and hepatocyte-like cells in vitro [18]. Our present study demonstrates that, although stellate cells are not currently considered to be true stem cells, they share specific markers, such as ABCG2, nestin, Thy1.1 and N-CAM, with other known adult stem cells. N-CAM participates in signal transduction and in cell-type segregation as a mediator of cellular junctions during organogenesis [40].

Although PaSCs do not usually express endocrine genes during cell expansion, spontaneous cell differentiation occurs, and these cells show weak expression of Pdx-1 (a key transcription factor in the endocrine differentiation pathway). Few studies have investigated how culture medium and additional protein components affect the viability and maturation of the cell [41]. Our results underscore the importance of defining culture-medium composition in experimental procedures in order to identify new soluble factors involved in the processes of cellular transdifferentiation. Moreover, to identify instructive signals that induce differentiation during organogenesis, it will be important to determine how such signalling networks are established and how they elicit multiple signalling responses in endodermal cells to activate appropriate genetic programmes [42]. Several signalling molecules have been implicated in induction of specific endodermal cell types. However, few of these factors have been examined in adult pancreatic tissue [43]. In the present study we show that the presence of inductive factors in the extracellular matrix, in addition to other substances participating in pancreatic differentiation, such as exendin-4 (GLP-1 analogue), are required for the transdifferentiation stages in our cellular model to proceed. GLP-1 is secreted from the L-cells of the distal ileum and colon; however, numerous studies show that GLP-1 produces an increase in the β-cell mass by inducing the neogenesis and transdifferentiation of ductal progenitors into islets cells through the expression of Pdx-1 [4446]. Soluble factors secreted by PaSCs are important in pancreatic physiology, including matrix turnover processes. However, other roles, including effects on progenitors, have not been investigated. An analysis of the PaSC secretome conducted by our group has identified three new proteins, namely PEDF (pigment epithelium-derived factor), LIF (leukaemia inhibitory factor) and Wnt5b (wingless-type mouse-mammary-tumour-virus integration site family, member 5B), that are involved in differentiation and developmental processes described in different cellular models (results not shown). Future experiments will be required to demonstrate that these proteins are able to participate in the spontaneous Pdx-1 expression detected in the culture from mitoxantrone-resistant cells and may also contribute to the high CK 19 expression detected after cell differentiation.

In summary, we found that mitoxantrone-resistant cells obtained from lactating rats pancreata, express the ABCG2 transporter, have a PaSC phenotype and are able to secrete insulin after cell differentiation.

We thank the Scientific and Technical Services of the University of Barcelona (SCT-UB) for technical support with electron microscopy, Ms Rosa Gasa for the AdCMV–Ngn3 and AdCMV–β-gal [adenovirus (cytomegalovirus) expressing Ngn3 or β-galactosidase respectively] constructs, Ms Margarita Nadal for karyotype assistance and Ms Caroline Newey for pre-submission editorial assistance.

Abbreviations

     
  • ABC

    ATP-binding cassette

  •  
  • AdCMV–β-gal

    adenovirus (cytomegalovirus) expressing β-galactosidase

  •  
  • AdCMV–Ngn3

    adenovirus (cytomegalovirus) expressing Ngn3 (neurogenin 3)

  •  
  • CIBER-BBN

    Centro de Investigación Biomédica en Red en Bioingeniería, Biomateriales y Nanomedicina

  •  
  • CIBERDEM

    CIBER de Diabetes y Enfermedades Metabólicas Asociadas

  •  
  • CK 19

    cytokeratin 19

  •  
  • DAPI

    4′,6-diamidino-2-phenylindole

  •  
  • DMEM

    Dulbecco's modified Eagle's medium

  •  
  • FCS

    fetal-calf serum

  •  
  • GFAP

    glial fibrillary acidic protein

  •  
  • GLP

    glucagon-like peptide

  •  
  • HGF

    hepatocyte growth factor

  •  
  • IAPP

    islet amyloid polypeptide

  •  
  • M 1

    M 2 and M 3, Medium 1, 2 and 3

  •  
  • N-CAM

    neural cell adhesion molecule

  •  
  • PaSC

    pancreatic stellate cell

  •  
  • Pdx1

    pancreatic and duodenal homeobox 1

  •  
  • RT–PCR

    reverse transcription–PCR

  •  
  • SP

    side population

  •  
  • TBP

    TATA-binding protein

AUTHOR CONTRIBUTION

Eugenia Mato designed the research, performed research, analysed data and wrote the paper. Maria Lucas performed research, analysed data and wrote the paper. Jordi Petriz designed the research and analysed the flow-cytometric data. Ramon Gomis and Anna Novials analysed data and wrote the manuscript Discussion.

FUNDING

Fondo de Investigación Sanitaria (FIS) [grant numbers PI020881, PI042553]; Red de Grupos en Diabetes Mellitus (RGDM) [grant number G03/212]; Instituto de Salud Carlos III (RCMN) [grant number RC033/08]; Plan Nacional de Investigación Cientifica, Desarrollo e Innovación Tecnológica, Dirección General de Investigación, Ministerio de Educación y Cienica Spain [grant numbers SAF 2003-06018, SAF 2006-07382] and by the Sardà Farriol Research Program. CIBERDEM and CIBER-BBN are ISCIII (Instituto de Salud Carlos III) projects.

References

References
1
Zhou
S.
Schuetz
J. D.
Bunting
K. D.
Colapietro
A. M.
Sampath
J.
Morris
J. J.
Lagutina
I.
Grosveld
G. C.
Osawa
M.
Nakauchi
H.
Sorrentino
B. P.
The ABC transporter Bcrp1/ABCG2 is expressed in a wide variety of stem cells and is a molecular determinant of the side-population phenotype
Nat. Med.
2001
, vol. 
7
 (pg. 
1028
-
1034
)
2
Bunting
K. D.
ABC transporters as phenotypic markers and functional regulators of stem cells
Stem Cells
2002
, vol. 
20
 (pg. 
11
-
20
)
3
Krishnamurthy
P.
Schuetz
J. D.
ABCG2/BCRP in biology and medicine
Annu. Rev. Pharmacol. Toxicol.
2005
, vol. 
46
 (pg. 
381
-
410
)
4
Bonner-Weir
S.
Sharma
A.
Pancreatic stem cells
J. Pathol.
2002
, vol. 
197
 (pg. 
519
-
526
)
5
Bonner-Weir
S.
Toschi
E.
Inada
A.
Yatoh
S.
Li
W. C.
Aye
T.
Toschi
E.
Sharma
A.
The pancreatic ductal epithelium serves as a potential pool of progenitor cells
Pediatr. Diabetes
2004
, vol. 
5
 
Suppl. 2
(pg. 
16
-
22
)
6
Von Herrath
M.
Homann
D.
Islet regeneration needed for overcoming autoimmune destruction – considerations on the pathogenesis of type 1 diabetes
Pediatr. Diabetes
2004
, vol. 
5
 
Suppl. 2
(pg. 
23
-
28
)
7
Gao
R.
Ustinov
J.
Korsgren
O.
Otonkoski
T.
In vitro neogenesis of human islets reflects the plasticity of differentiated human pancreatic cells
Diabetologia
2005
, vol. 
48
 (pg. 
2296
-
2304
)
8
Rosenberg
L.
Induction of islet cell neogenesis in the adult pancreas: the partial duct obstruction model
Microsc. Res. Tech.
1998
, vol. 
43
 (pg. 
337
-
346
)
9
Rosenberg
L.
Duguid
W. P.
Healy
M.
Clas
D.
Vinik
A. I.
Reversal of diabetes by the induction of islet cell neogenesis
Transplant. Proc.
1992
, vol. 
24
 (pg. 
1027
-
1028
)
10
Bernard
C.
Berthault
M. F.
Saulnier
C.
Ktorza
A.
Neogenesis vs. apoptosis as main components of pancreatic β-cell mass changes in glucose-infused normal and mildly diabetic adult rats
FASEB J.
1999
, vol. 
13
 (pg. 
1195
-
1205
)
11
Sharma
A
Zangen
D. H.
Reitz
P.
Taneja
M.
Lissauer
M. E.
Miller
C. P.
Weir
G. C.
Habener
J. F.
Bonner-Weir
S.
The homeodomain protein IDX-1 increases after an early burst of proliferation during pancreatic regeneration
Diabetes
1999
, vol. 
48
 (pg. 
507
-
513
)
12
Guz
Y.
Nasir
I.
Teitelman
G.
Regeneration of pancreatic β cells from intra-islet precursor cells in an experimental model of diabetes
Endocrinology
2001
, vol. 
142
 (pg. 
4956
-
4968
)
13
Dor
Y.
Brown
J.
Martinez
O. I.
Melton
D. A.
Adult pancreatic β-cells are formed by self-duplication rather than stem-cell differentiation
Nature
2004
, vol. 
429
 (pg. 
41
-
46
)
14
Lechner
A.
Leech
C. A.
Abraham
E.
Nolan
A. L.
Habener
J. F.
Nestin-positive progenitor cells derived from adult human pancreatic islets of Langerhans contain side population (SP) cells defined by expression of the ABCG2 (BCRP1) ATP-binding cassette transporter
Biochem. Biophys. Res. Commun.
2002
, vol. 
293
 (pg. 
670
-
674
)
15
Shook
D.
Keller
R.
Mechanisms, mechanics and function of epithelial-mesenchymal transitions in early development
Mech. Dev.
2003
, vol. 
120
 (pg. 
1341
-
1383
)
16
Phillips
P. A
McCarroll
J. A.
Park
S.
Wu
M. J.
Pirola
R.
Korsten
M.
Wilson
J. S.
Apte
M. V.
Rat pancreatic stellate cells secrete matrix metalloproteinases: implications for extracellular matrix turnover
Gut
2003
, vol. 
52
 (pg. 
275
-
282
)
17
Omary
M. B.
Lugea
A.
Lowe
A. W.
Pandol
S. J.
The pancreatic stellate cell: a star on the rise in pancreatic diseases
J. Clin. Invest.
2007
, vol. 
117
 (pg. 
50
-
59
)
18
Kordes
C.
Sawitza
I.
Müller-Marbach
A.
Ale-Agha
N.
Keitel
V.
Klonowski-Stumpe
H.
Häussinger
D.
CD133+ hepatic stellate cells are progenitor cells
Biochem. Biophys. Res. Commun.
2006
, vol. 
352
 (pg. 
410
-
417
)
19
Garcia-Escarp
M.
Martínez-Muñoz
V.
Sales-Pardo
I.
Barquinero
J
Domingo
J. C.
Marin
P.
Petriz
J.
Flow cytometry-based approach to ABCG2 function suggests that the transporter differentially handles the influx and efflux of drugs
Cytometry
2004
, vol. 
62
 (pg. 
129
-
138
)
20
Catalano
R. A.
Lillie
R. D.
Elimination of precipitates in Oil Red O fat stain by adding dextrin
Stain Technol.
1975
, vol. 
50
 (pg. 
297
-
299
)
21
Kawazu
S.
Kanazawa
Y.
Hayashi
M.
Ikeuchi
M.
Nakai
T.
Kosaka
K.
Monolayer culture of human fetal and adult pancreas. Static and dynamic studies of insulin release in vitro
Horm. Metab. Res.
1980
, vol. 
12
 (pg. 
354
-
360
)
22
Spurr
A. R.
A low viscosity epoxy resin embedding medium for electron microscopy
J. Ultrastruc. Res.
1969
, vol. 
26
 (pg. 
36
-
43
)
23
Reynolds
E. S.
The use of lead citrate at high pH as an electron opaque stain in electron microscopy
J. Cell Biol.
1963
, vol. 
17
 (pg. 
208
-
212
)
24
Royer
P. J.
Tanguy-Royer
S.
Ebstein
F.
Sapede
C.
Simon
T.
Barbieux
I.
Oger
R.
Gregoire
M.
Culture medium and protein supplementation in the generation and maturation of dendritic cells
Scand. J. Immunol.
2006
, vol. 
63
 (pg. 
401
-
409
)
25
Huang
H.
Tang
X.
Phenotypic determination and characterization of nestin-positive precursors derived from human fetal pancreas
Lab. Invest.
2003
, vol. 
83
 (pg. 
539
-
547
)
26
Schwitzgebel
V. M.
Scheel
D. W.
Conners
J. R.
Kalamaras
J.
Lee
J. E.
Anderson
D. J.
Sussel
L.
Johnson
J. D.
German
M. S.
Expression of neurogenin3 reveals an islet cell precursor population in the pancreas
Development
2000
, vol. 
127
 (pg. 
3533
-
3542
)
27
Shapiro
A. M.
Lakey
J. R.
Paty
B. W.
Senior
P. A.
Bigam
D. L.
Ryan
E. A.
Strategic opportunities in clinical islet transplantation
Transplantation
2005
, vol. 
79
 (pg. 
1304
-
1307
)
28
Gao
R.
Ustinov
J.
Pulkkinen
M. A.
Pulkkinen
M. A.
Lundin
K.
Korsgren
O.
Otonkoski
T.
Characterization of endocrine progenitor cells and critical factors for their differentiation in human adult pancreatic cell culture
Diabetes
2003
, vol. 
52
 (pg. 
2007
-
2015
)
29
Seeberger
K. L.
Dufour
J. M.
Shapiro
S. M.
Lakey
J. R.
Rajotte
R. V.
Korbutt
G. S.
Expansion of mesenchymal stem cells from human pancreatic ductal epithelium
Lab. Invest.
2006
, vol. 
86
 (pg. 
141
-
153
)
30
Weissman
I. L.
Anderson
D. J.
Gage
F.
Stem and progenitor cells: origins, phenotypes, lineage commitments, and trans-differentiations
Annu. Rev. Cell. Dev. Biol.
2001
, vol. 
17
 (pg. 
387
-
403
)
31
Scharfmann
R.
Control of early development of the pancreas in rodents and humans: implications of signals from the mesenchyme
Diabetologia
2000
, vol. 
43
 (pg. 
1083
-
1092
)
32
Beattie
G. M.
Levine
F.
Mally
M. I.
Otonkoski
T.
O'Brien
J. S.
Salomon
D. R.
Hayek
A.
Acid β-galactosidase: a developmentally regulated marker of endocrine cell precursors in the human fetal pancreas
J. Clin. Endocrinol. Metab.
1994
, vol. 
78
 (pg. 
1232
-
1240
)
33
Kim
S. K.
Hebrok
M.
Melton
D. A.
Notochord to endoderm signalling is required for pancreas development
Development
1997
, vol. 
124
 (pg. 
4243
-
4252
)
34
Yashpal
N. K.
Li
J.
Wang
R.
Characterization of c-Kit and nestin expression during islet cell development in the prenatal and postnatal rat pancreas
Dev. Dyn.
2004
, vol. 
229
 (pg. 
813
-
825
)
35
Fetsch
P. A.
Abati
A.
Litman
T.
Morisaki
K.
Honjo
Y.
Mittal
K.
Bates
S. E.
Localization of the ABCG2 mitoxantrone resistance-associated protein in normal tissues
Cancer Lett.
2006
, vol. 
235
 (pg. 
84
-
92
)
36
Tanaka
Y.
Slitt
A. L.
Leazeer
T. M.
Maher
J. M.
Klaassen
C. D.
Tissue distribution and hormonal regulation of the breast cancer resistence protein (Bcrp/Abcg2) in rats and mice
Biochem. Biophys. Res. Commun.
2005
, vol. 
326
 (pg. 
181
-
187
)
37
Leslie
E. M.
Deeley
R. G.
Cole
S. P.
Multidrug resistance proteins: role of P-glycoprotein, MRP1, MRP2, and BCRP (ABCG2) in tissue defense
Toxicol. Appl. Pharmacol.
2005
, vol. 
204
 (pg. 
216
-
237
)
38
Apte
M. V.
Haber
P. S.
Applegate
T. L.
Norton
I. D.
McCaughan
G. W.
Korsten
M. A.
Pirola
R. C.
Wilson
J. S.
Periacinar stellate shaped cells in rat pancreas: identification, isolation and culture
Gut
1998
, vol. 
43
 (pg. 
128
-
133
)
39
Kruse
M. L.
Hildebrand
P. B.
Timke
C.
Fölsch
U. R.
Scmidt
W. E.
Isolation, long-term culture, and characterization of rat pancreatic fibroblastoid/stellate cells
Pancreas
2000
, vol. 
23
 (pg. 
49
-
54
)
40
Esni
F.
Taljedal
I. B.
Perkl
A. K.
Cremer
H.
Christofori
G.
Semb
H.
Neural cell adhesion molecule (N-CAM) is required for cell type segregation and normal ultrastructure in pancreatic islets
J. Cell Biol.
1999
, vol. 
144
 (pg. 
325
-
337
)
41
Royer
P. J.
Tanguy-Royer
S.
Ebstein
F.
Sapede
C.
Simon
T.
Barbieux
I.
Oger
R.
Gregoire
M.
Culture medium and protein supplementation in the generation and maturation of dendritic cells
Scand. J. Immunol.
2006
, vol. 
63
 (pg. 
401
-
409
)
42
Ratineau
C.
Duluc
I.
Pourreyron
C.
Kedinger
M.
Freund
J. N.
Roche
C.
Endoderm- and mesenchyme-dependent commitment of the differentiated epithelial cell types in the developing intestine of rat
Differentiation
2003
, vol. 
71
 (pg. 
163
-
169
)
43
Sttaford
D.
White
R. J.
Kinkel
M. D.
Linville
A.
Schilling
T. F.
Prince
V. E.
Retinoids signal directly to zebrafish endoderm to specify insulin-expressing β-cells
Development
2006
, vol. 
133
 (pg. 
949
-
956
)
44
Yue
F.
Cui
L.
Johkura
K.
Ogiwara
N.
Sasaki
K.
Glucagon-like peptide-1 differentiation of primate embryonic stem cells into insulin-producing cells
Tissue Eng.
2006
, vol. 
12
 (pg. 
2105
-
2115
)
45
Abraham
E. J.
Leech
C. A.
Lin
J. C.
Zulewski
H.
Haberner
J. F.
Insulinotropic hormone glucagons-like peptide-1 differentiation of human pancreatic islet-derived progenitor cells into insulin-producing cells
Endocrinology
2002
, vol. 
143
 (pg. 
3152
-
3161
)
46
Hui
H.
Wright
C.
Perfetti
R.
Glucagon-like peptide 1 induces differentiation of islets duodenal homeobox-1-positive pancreatic ductal cells into insulin-secreting cells
Diabetes
2001
, vol. 
50
 (pg. 
785
-
796
)

Author notes

1

These two authors contributed equally to this study.

3

Present address: Institut de Recerca (Laboratori 123), Hospital Universitari Vall d'Hebron, P Vall d'Hebron 119-129, 08035 Barcelona, Spain.

Supplementary data