- Email: [email protected]
Cdkn1a is a key mediator of rat pancreatic stellate cell senescence
Pancreatology 13 (2013) 254e262
Contents lists available at SciVerse ScienceDirect
Pancreatology journal homepage: www.elsevier.com/locate/pan
Original article
Cdkn1a is a key mediator of rat pancreatic stellate cell senescence Brit Fitzner, Andreas Lange, Sarah Müller, Robert Jaster* Department of Medicine II, Division of Gastroenterology, University Medicine Rostock, E.-Heydemann-Str. 6, D-18057 Rostock, Germany
a r t i c l e i n f o
a b s t r a c t
Article history: Received 16 January 2013 Received in revised form 27 February 2013 Accepted 8 March 2013
Background/objectives: Completion of pancreatic wound healing requires termination of pancreatic stellate cell (PSC) activation to prevent fibrosis. Besides induction of apoptosis and return to a quiescent phenotype, senescence of PSC followed by immune cell-mediated cytolysis represents a potential mechanism. Here, we have studied if the cell cycle inhibitor cyclin-dependent kinase inhibitor 1A (Cdkn1a, p21/Waf1), expression of which is increased in senescent rat PSC, plays a causative role in the senescence process. Methods: Senescence was induced by doxorubicin treatment. The functions of Cdkn1a were analyzed using two approaches, treatment of primary rat PSC with siRNA and tetracycline-regulated overexpression of Cdkn1a in immortalized rat cells. Expression of senescence-associated b-galactosidase (SA b-Gal) was used as a surrogate marker of senescence. Results: The knockdown of Cdkn1a significantly attenuated the growth-inhibitory effect of doxorubicin and strongly diminished the portion of SA b-Gal-positive cells. Overexpression of Cdkn1a enhanced both the antiproliferative effect of doxorubicin and induction of senescence. In primary PSC, doxorubicin treatment was associated with increased expression of interleukin-6 (IL-6) and matrix metalloproteinase (MMP)-9, while expression of the activation marker a-smooth muscle actin (a-SMA), p53, Cdk1 and Rad54 was diminished. The application of Cdkn1a siRNA specifically antagonized the effects of doxorubicin on the expression of p53, Cdk1 and Rad54 but not IL-6 and a-SMA, while MMP-9 expression and also activity were even enhanced. Conclusions: Cdkn1a plays a direct role in the process of rat PSC senescence. Additional Cdkn1aindependent pathways may contribute to the partial maintenance of a gene expression profile typical of senescent PSC. Copyright Ó 2013, IAP and EPC. Published by Elsevier India, a division of Reed Elsevier India Pvt. Ltd. All rights reserved.
Keywords: Chronic pancreatitis Fibrosis Cellular ageing
1. Introduction The wound healing reaction in the pancreas involves proliferation and activation of pancreatic stellate cells (PSC), the main source of extracellular matrix (ECM) proteins in the diseased organ. In chronic pancreatitis and in pancreatic cancer, persistent activation of PSC results in pancreatic fibrosis, which is considered a Abbreviations: a-SMA, a-smooth muscle actin; BrdU, 5-bromo-20 -deoxyuridine; Cdkn1a, cyclin-dependent kinase inhibitor 1A; CP, chronic pancreatitis; DAPI, 40 ,6Diamidino-2-phenylindole; doxo, doxorubicin; ECM, extracellular matrix; FCS, fetal calf serum; HA, hemagglutinin; IL-6, interleukin-6; IMDM, Iscove’s Modified Dulbecco’s Medium; MMP-9, matrix metalloproteinase 9; PI, propidium iodide; PSC, pancreatic stellate cell; PVDF, polyvinylidene difluoride; SA b-Gal, senescenceassociated b-galactosidase; SASP, senescence-associated secretory phenotype; SEM, standard error of the mean; tet, tetracycline; X-Gal, 5-bromo-4-chloro-indolyl-b-Dgalactopyranoside. * Corresponding author. Tel.: þ49 381 494 7349; fax: þ49 381 494 7482. E-mail address: [email protected] (R. Jaster).
progression factor of the tumor disease. Although the mechanisms leading to perpetuation of PSC activation are incompletely understood, it is likely that cytokines and growth factors derived from injured acinar cells, inflammatory cells, tumor cells and activated PSC themselves play a key role in this process [1,2]. The pathogenetic relevance of pancreatic fibrosis has also raised the question how PSC activation may be terminated. Previous studies of several laboratories, including our own one, have identified factors such as vitamin A derivates and ligands of the peroxisome proliferator activated receptor g, that favor exhibition of a quiescent PSC phenotype [3e7]. However, since activation of PSC is associated with cell proliferation, additional mechanisms appear necessary to eliminate a surplus of ECM-producing cells. One obvious possibility is that PSC, after completing their repair function, may undergo apoptosis, and indeed experimental evidence in support of this hypothesis has been provided [7e10]. In addition, studies in the field of liver fibrosis have suggested another possibility: The process of cellular senescence may render stellate cells susceptible to
1424-3903/$ e see front matter Copyright Ó 2013, IAP and EPC. Published by Elsevier India, a division of Reed Elsevier India Pvt. Ltd. All rights reserved. http://dx.doi.org/10.1016/j.pan.2013.03.009
B. Fitzner et al. / Pancreatology 13 (2013) 254e262
cytolysis mediated by immune cells; specifically natural killer cells [11]. Using a rat model of chronic pancreatitis, we have recently shown that senescent cells, most likely PSC, are also detectable in the fibrotic pancreas [12]. Cellular senescence represents a state of irreversible cell cycle arrest and is considered a key barrier mechanism against tumorigenesis [13,14]. In addition to telomere shortage in the course of multiple cell divisions, diverse forms of cellular stress have been implicated in the induction of senescence [11,13,14]. In our previous study we found that senescence of isolated rat PSC can be triggered by long-term culture, the DNA-damaging agent doxorubicin (doxo), hydrogen peroxide and a low concentration of the pro-apoptotic substance staurosporine [12]. Senescent rat PSC mimicked nonproliferating quiescent PSC also in that they expressed low levels of the activation marker a-smooth muscle actin (a-SMA). In contrast to quiescent PSC however, senescent stellate cells expressed high levels of senescence-associated b-galactosidase (SA b-Gal) and displayed a gene expression profile compatible with a senescence-associated secretory phenotype (SASP); e.g. synthesis of large amounts of interleukin-6 (IL-6). Exhibition of SASP and expression of SA b-Gal are two out of many established surrogate markers of cellular senescence [11,13,14]. Another typical phenomenon represents the overexpression of cell cycle inhibitors. Indeed, we observed that rat PSC senescence is accompanied by increased mRNA levels of Cdkn1a (p21/Waf1) [12]. In this study, we have addressed the question if expression of Cdkn1a is an epiphenomenon of the senescence process, or plays an active role in its progression. Using two independent experimental approaches, application of siRNA against Cdkn1a and overexpression of the protein, we found that doxo-induced senescence of rat PSC is mediated both through Cdkn1a-dependent and eindependent pathways. 2. Methods 2.1. Cell isolation and culture Rat PSC were isolated by collagenase digestion of the pancreas followed by Nycodenz density gradient centrifugation as previously described [15]. They were cultured in Iscove’s Modified Dulbecco’s Medium (IMDM) supplemented with 17% fetal calf serum, 1% nonessential amino acids (dilution of a 100 stock solution), 100 U/ml penicillin and 100 mg/ml streptomycin at 37 C in a 5% CO2 humidified atmosphere. The cells were grown on culture plates to subconfluency, harvested by trypsination and recultured at equal seeding densities. All experiments were performed with cells recultured no more than two times. Stellate cell lines were cultured in IMDM supplemented with 10% fetal calf serum (FCS), non-essential amino acids, 100 U/ml penicillin and 100 mg/ml streptomycin at 37 C in a 5% CO2 humidified atmosphere. Tetracycline (tet) was applied as described below. 2.2. Inhibition of Cdkn1a expression using siRNA siRNAs were purchased from Riboxx (Radebeul, Germany). Out of four siRNAs tested, the following two guide strand sequences displayed the strongest inhibitory effect on Cdkn1a expression and were selected for functional studies: 50 -UAAAGAUAAACAGUCCAGCCCCC-30 , and 50 -AUAGAAAUCUGUUAGGCUGCCCC-30 (subsequently termed siRNA1 and siRNA2, respectively). The guide strand sequence 50 -UAAGCACGAAGCUCAGAGUCCCCC-30 served as non-silencing control. For transfection with siRNA, PSC were trypsinized, centrifuged and resuspended in RPMI 1640 culture medium supplemented with TrueFect-Lipo transfection reagent
255
according to the suggestions of the manufacturer (United Biosystems, Herndon, VA, USA) and siRNA at a final concentration of 100 nM. Afterwards, the cells were seeded into culture dishes of variable size according to the experimental requirements and allowed to attach for 4 h before the transfection medium was substituted by regular culture medium supplemented with doxo at 25 ng/ml as indicated. To ensure maintenance of silencing in longterm studies (>48 h) with siRNA-transfected cells, the transfection procedure was repeated every 2e3 days, except of that PSC were not trypsinized prior to siRNA application. Efficiency of siRNA treatment was monitored by analyzing expression of Cdkn1a mRNA using real-time PCR. 2.3. Vector construction and expression of Cdkn1a The cDNA of the murine Cdkn1a gene (AF457188; complete coding sequence) was cloned into the Nhe I site of pTRE2purHA (confers puromycin resistance; Invitrogen, Karlsruhe, Germany), and the correct sequence and orientation of the insert was verified by dideoxy sequencing. The resulting constructs encode for a Cdkn1a protein with an N-terminal hemagglutinin (HA)-tag, expression of which is controlled by a tet response element. For Cdkn1a overexpression studies, we used LTC-Tet cells [7], an immortalized rat PSC line that stably expresses the transcriptional activator protein tTA, which is inactivated upon tet binding. LTC-Tet cells were transfected with the pTRE2purHA-Cdkn1a construct by lipofection. Therefore, cells growing at a low density (5% confluency) in 6-well plates were incubated with plasmid DNA (2 mg per well), using FuGENE HD Transfection Reagent (Roche, Mannheim, Germany) according to the vendor’s specifications. After transfection, the cells were cultured in medium supplemented with tet (1 mg/ml) for 48 h, before puromycin was added at 2 mg/ml. Upgrowing resistant cells were tested for a tet-dependent Cdkn1a expression profile, and clones with a strong induction of Cdkn1a transgene expression after tet withdrawal were selected. Unless otherwise stated, puromycin-resistant cells (referred to as LTCCdkn1a cells) were permanently cultured with tet (1 mg/ml). The data shown were verified in three independent cell clones of LTCCdkn1a. 2.4. Quantification of DNA synthesis Cell proliferation was assessed using a 5-bromo-20 -deoxyuridine (BrdU) labeling and detection enzyme-linked immunosorbent assay kit (Roche). In experiments with siRNA, PSC resuspended in transfection medium supplemented as described above were seeded at equal densities into 96-well plates and treated with doxo (at 25 ng/ml) as indicated. For overexpression studies, LTC-Cdkn1a cells were washed free of tet, trypsinized and also seeded into microplates. Tet and doxo (10 ng/ml, since a dose of 25 ng/ml displayed toxic effects) were added as specified below. In all studies, BrdU labeling was initiated 24 h after drug application or withdrawal by adding labeling solution at a final concentration of 10 mM to the culture medium. After another incubation period of 24 h, the labeling was stopped, and BrdU uptake was measured according to the manufacturer’s instructions. 2.5. Detection of SA b-Gal Transfected PSC and LTC-Cdkn1a cells (exposed to tet as indicated) were cultured on coverslips for 7 days in the presence and absence of doxo, respectively. Afterwards, SA b-Gal-positive cells were stained using a specific kit (New England BioLabs, Frankfurt, Germany), whereas the nuclei were counterstained with DAPI according to standard protocols. The assay used in this study
256
B. Fitzner et al. / Pancreatology 13 (2013) 254e262
monitors enzymatic activity, which is required to cleave the substrate X-Gal. Since the reaction is performed at pH 6 where other cellular b-galactosidases are inactive, the assay is specific for SA b-gal. Subsequently, specimens were analyzed by light and fluorescence microscopy. Image evaluation was performed as previously described [12], using an ImageMagick-based software to distinguish between SA b-Gal-positive and -negative cells. 2.6. Immunoblotting Protein extracts of cultured cells were prepared, adjusted to identical protein concentrations and subjected to immunoblot analysis as published before [15], using PVDF membrane for protein transfer. The following primary antibodies were employed: Anti-bactin (# 4970, New England BioLabs, Frankfurt, Germany), Anti-aSMA (# A2547, SigmaeAldrich, Deisenhofen, Germany) and Anti-Cdkn1a (# SX118, BD Biosciences, Heidelberg, Germany). The blots were developed using LI-COR reagents for an OdysseyÒ Infrared Imaging System as previously described [16]. Signal intensities of the investigated proteins were quantified by means of the OdysseyÒ software version 3.0, and normalized to b-actin prior to the comparison of expression levels. 2.7. Quantitative reverse transcriptase PCR using real-time TaqManÔ technology Transfected PSC were grown in 12-well plates for 7 days, before total RNA was isolated with peqGOLD TriFast reagent (Peqlab Biotechnologie, Erlangen, Germany) following the manufacturer’s instructions. After DNAse digestion of genomic DNA, RNA was reverse transcribed into cDNA using Maxima Probe Irox qPCR Master Mix (Fisher Scientific, Schwerte, Germany) and random hexamer priming. Relative quantification of target cDNA levels by real-time PCR was performed in an ABI Prism 7000 sequence detection system employing TaqManÔ Universal PCR Master Mix and the following Assay-on-DemandÔ rat gene-specific fluorescently labeled TaqManÔ MGB probes (instrument and reagents: Life Technologies, Darmstadt, Germany): Rn00589996_m1 (Cdkn1a/p21), Rn00570728_m1 (Cdk1), Rn00561420_m1 (IL-6), Rn00579162_m1 (MMP-9), Rn00755717_m1 (p53), Rn01416981_m1 (Rad54) and Rn01527840_m1 (Hprt). PCR was performed under the following conditions: 95 C for 10 min, 50 cycles of 15 s at 95 C, 1 min at 60 C. Relative expression of each mRNA compared with the housekeeping gene Hprt was calculated according to the equation DCt ¼ Cttarget CtHPRT. The relative amount of target mRNA in control cells and cells treated as indicated was expressed as 2(DDCt), where DDCttreatment ¼ DCtsample DCt control. The reactions were performed in triplicate, and repeated at least 6 times with independent samples.
resuspended in 2 ml ice-cold 70% ethanol for at least 12 h at 4 C. Following further washing steps, the cells were incubated for 20 min in 400 ml PBS containing 0.1 mg/ml RNase A (Roche) at 37 C. After the addition of 50 mg propidium iodide (PI; SigmaeAldrich) per ml, the samples were subjected to cytofluorometric analysis. 10.000 events were measured for each sample and the data stored in list-mode for further analysis. The cell cycle distribution was calculated using the Cellquest program. Cells of the Sub-G1 peak were considered apoptotic. 2.9. Quantification of procollagen I carboxy terminal propeptide (PICP) To monitor collagen type I synthesis and secretion, supernatants from PSC cultures (pretreated as indicated) were collected, cleared by centrifugation and stored at 80 C until assayed. PICP levels were measured using a rat PICP-specific ELISA according to the instructions of the manufacturer (Uscn Life Science, Houston, TX, USA). Measured PICP concentrations were normalized with respect to unequal cell densities based on the quantification of b-actin levels in total cellular lysates by Western blotting. 2.10. Zymography Supernatants of PSC cultures were obtained and processed as described above, mixed 1:1 with Zymogram Sample Buffer (BioRad Laboratories, München, Germany) and loaded onto 8% SDSe polyacrylamide gels supplemented with 1% gelatine. After electrophoresis at 4 C, the gels were soaked in 1 Zymogram Renaturation Buffer (BioRad) with gentle shaking (2 h at room temperature), before they were incubated overnight at 37 C in 1 Zymogram Development Buffer (BioRad). Finally, lysis bands were visualized by Coomassie staining (30 min at room temperature followed by three washing steps with distilled water) and, taking advantage of the fluorescence of the stain [17], quantitated using the OdysseyÒ Infrared Imaging System. The raw data were normalized as described for PICP. 2.11. Statistical analysis Results are expressed as mean standard error of the mean (SEM) for the indicated number of separate cultures per
2.8. Detection of dead cells and analysis of cellular DNA content by flow cytometry To quantify dead cells by flow cytometry, PSC of passage 1 were exposed to doxo (25 ng/ml), or left untreated. After 7 days, the cells were harvested by trypsination, resuspended in FACS-buffer (PBS pH 7,4; 0.5% bovine serum albumin; 0.1% sodium azide) and kept on ice until measurement. Subsequently, the samples were labeled with propidium iodide (PI; 10 mg/ml). PI-positive (dead) cells were quantified using a FACSCalibur cytometer (BD Biosciences) and the Cellquest program. For the detection of the cellular DNA content, the following protocol was employed: PSC transfected with siRNA were grown for 48 h in 6-well plates in the presence and absence of doxo as indicated. To obtain a single cell suspension, they were trypsinized, pelleted by centrifugation, washed twice with PBS and
Fig. 1. Inhibition of Cdkn1a expression with siRNA. Cultured PSC were transfected with Cdkn1a-specific siRNAs or a non-silencing oligonucleotide (controls; all siRNAs at 100 nM) and exposed to doxo (at 25 ng/ml) as indicated. After an incubation period of 16 h, mRNA expression of Cdkn1a and Hprt (housekeeping control) was analyzed by real-time PCR, and relative amounts of Cdkn1a mRNA were calculated. One hundred percent Cdkn1a expression corresponds to controls cultured without doxo. Data of n 5 independent experiments (with duplicate samples) were used to calculate mean values and SEM. *P < 0.05 versus controls cultured without doxo, #P < 0.05 versus control cells cultured with doxo (columns 5 and 6 vs. column 4).
B. Fitzner et al. / Pancreatology 13 (2013) 254e262
experimental protocol. Statistical significance was analyzed using the indicated statistical test. P < 0.05 was considered to be statistically significant. 3. Results To study the role of Cdkn1a in PSC senescence, expression of Cdkn1a in rat PSC was inhibited by means of siRNA. As shown in Fig. 1, two different siRNAs against Cdkn1a strongly diminished both basal and doxo-induced Cdkn1a expression. An efficient knock-down was observed 16 h after transfection (Fig. 1), and maintained over at least 7 days (data not shown). Analysis of cell proliferation revealed a small but significant growth advantage of PSC transfected with Cdkn1a siRNA1 under
257
basal conditions (no doxo). More importantly, both Cdkn1a siRNAs significantly attenuated the growth-inhibitory effect of doxo (Fig. 2A), implicating Cdkn1a in the action of the drug. Consistent with these findings, more doxo-treated PSC were found in the S-phase of the cell cycle when the cells were transfected with siRNA against Cdkn1a (Fig. 2B). Under all conditions (doxo and/or siRNA), the portion of cells in the sub-G1 peak remained below 4%, suggesting that only very few cells underwent apoptosis. As indicated by the results of PI staining, the total portion of dead (apoptotic and necrotic) cells after doxo treatment was approximately 10%, suggesting that cytotoxic effects only modestly contributed to doxo-mediated growth inhibition. Fig. 2C shows the results for cells treated doxo in the absence of siRNA. Application of the Cdkn1a siRNAs did at least
Fig. 2. Downregulation of Cdkn1a reduces the antiproliferative effect of doxorubicin. (A and B) Cultured PSC were transfected with Cdkn1a-specific or control siRNA (at 100 nM) and exposed to doxo (at 25 ng/ml) for 48 h as indicated. (A) Cell proliferation was assessed with the BrdU DNA-incorporation (inc) assay as described in the “Materials and methods” section. One hundred percent cell proliferation corresponds to controls cultured without doxo. (B) Cells in the S-phase of the cell cycle were detected by FACS analysis based on their DNA content and related to the total number of cells. (A and B) Data from n 10 separate cultures were used to calculate mean values and SEM. *P < 0.05 versus controls cultured without doxo, #P < 0.05 versus control cells cultured with doxo (columns 5 and 6 vs. column 4). (C) PSC were cultured with or without doxo (at 25 ng/ml) for 7 days; the time corresponding to the conditions of the senescence assays (Fig. 3). Afterwards, they were subjected to a cytofluorometric quantification of dead (PI-positive) cells and apoptotic (SubG1 peak) cells (expressed as percent of total cells). Data from n 6 separate cultures were used to calculate mean values and SEM. *P < 0.05 versus controls cultured without doxo.
258
B. Fitzner et al. / Pancreatology 13 (2013) 254e262
not increase the small portion of dead cells in doxo-treated cultures (data not shown). In agreement with our previous studies [12], we found that doxo treatment for 7 days induced expression of the senescence marker SA b-Gal in a significant portion of cultured rat PSC while inhibiting expression of the activation marker a-SMA (Figs. 3 and 4). In the presence of non-silencing control siRNA, senescence of almost one third of the cells was observed. Treatment with either of the two Cdkn1a siRNAs reduced this portion by roughly 50% (Fig. 3). In contrast, doxo-dependent inhibition of a-SMA expression was not diminished by the Cdkn1a siRNAs but even modestly enhanced (Fig. 4). We have previously reported that doxo-induced senescence of rat PSC is associated with characteristic changes of gene expression, including upregulation of IL-6 and downregulation of p53 mRNA levels [12]. As shown in Fig. 5, treatment of PSC with Cdkn1a siRNAs did not interfere with doxo-mediated IL-6 expression (A), but completely antagonized the inhibitory effect of the drug on the expression of p53 (B). In a similar manner, the Cdkn1a-specific siRNAs also blocked downregulation of cyclin-dependent kinase 1 (Cdk1) (C) and Rad54 (D) mRNA levels by the DNA-damaging agent. In contrast, doxo-triggered expression of MMP-9 (E) was strongly enhanced in cells treated with Cdkn1a siRNA. To validate the latter findings, we also quantified MMP-9 activities in PSC supernatants by zymography (Fig. 6A). Again, a doxo-dependent increase was observed that was strongly enhanced by treatment with Cdkn1aspecific siRNA. In contrast, activities of MMP-2 remained unchanged (data not shown). As indicated by the measurements of PICP levels in the same supernatants, doxo treatment of control cultures had no significant effect on the secretion of collagen type 1 (Fig. 6B). Here, a knockdown of Cdkn1a lead to significantly
Fig. 4. Inhibition of a-SMA expression by doxorubicin is independent of Cdkn1a. PSC were treated as described in Fig. 3. (A) Expression of a-SMA and the housekeeping protein b-actin was analyzed by immunoblotting. (B) For quantitative evaluation, signal intensities were determined, and the ratio a-SMA/b-actin was calculated. A ratio of one hundred percent corresponds to controls transfected with non-silencing siRNA and cultured without doxo (column 1). Data from n ¼ 9 independent samples per data point were used to calculate mean values and SEM. *P < 0.05 versus controls cultured without doxo, #P < 0.05 versus control cells cultured with doxo (columns 4 vs. 3 and 6 vs. 5, respectively).
Fig. 3. Cdkn1a-specific siRNA inhibits doxorubicin-dependent senescence of PSC. Cultured PSC were transfected with Cdkn1a-specific or control siRNA (at 100 nM) and exposed to doxo (at 25 ng/ml) for 7 days as indicated. Afterwards, the cells were stained for expression of SA b-Gal and the nuclei where counterstained with DAPI. Panel (A) shows typical microphotographs of PSC treated with non-silencing control siRNA and Cdkn1a-specific siRNA2, respectively, in the presence and absence of doxo. (B) SA b-Gal-positive cells were determined based on a quantitative image evaluation and related to the total cell number. The results are averaged values (SEM) of 12 independent experiments. *P < 0.05 versus controls cultured without doxo, #P < 0.05 versus control cells cultured with doxo (columns 5 and 6 vs. column 4).
B. Fitzner et al. / Pancreatology 13 (2013) 254e262
259
Fig. 5. Effects of Cdkn1a-specific siRNA on PSC gene expression. PSC were transfected with Cdkn1a-specific or non-silencing control siRNA (at 100 nM) and treated with doxo (at 25 ng/ml) for 7 days as indicated. The mRNA expression of IL-6 (A), p53 (B), Cdk1 (C), RAD54 (D), MMP-9 (E) and Hprt (housekeeping control) was analyzed by real-time PCR, and relative amounts of target mRNA were determined. The results are averaged values (SEM) of at least 6 independent experiments. One hundred percent mRNA expression corresponds to controls cultured without doxo. *P < 0.05 versus controls cultured without doxo, #P < 0.05 versus control cells cultured with doxo (columns 5 and 6 vs. column 4).
decreased levels of PICP (comparison of doxo-treated cultures), suggesting different roles of Cdkn1a in the regulation of MMP-9 and collagen type 1 expression. To verify the biological effects of Cdkn1a in stellate cells of the pancreas, Tet-regulated overexpression of a Cdkn1a transgene was employed as an independent approach. Therefore, cells of the immortalized rat pancreatic stellate cell line LTC-Tet cells (stably expressing the tTA protein [7]), were transfected with the pTRE2purHA-Cdkn1a vector as described above. LTC-Cdkn1a cells cultured with tet at 1 mg/ml did not express detectable amounts of Cdkn1a protein (Fig. 7A), suggesting low levels of the endogenous Cdkn1a protein and suppression of transgene transcription. Withdrawal of tet induced a strong expression of the transgenic protein, which was associated with an inhibition of DNA synthesis and a significant enhancement of the antiproliferative effect of doxo (Fig. 7B). Analysis of SA b-Gal expression revealed that expression of the Cdkn1a transgene alone was not sufficient to induce senescence of a significant portion of LTC-Cdkn1a cells (Fig. 7C). In this immortalized cell line, the effect of doxo was also weaker than in primary cells (Fig. 3) but nevertheless clear (18% vs. 3% SA b-Galpositive cells in the presence and absence of doxo, respectively).
When tet withdrawal and doxo treatment were combined, a nearly 2-fold increase of SA b-Gal-positive cells was observed, suggesting that expression of the Cdkn1a transgene significantly enhanced induction of cellular senescence by doxo. 4. Discussion Increased levels of the cell cycle inhibitor Cdkn1a are a typical characteristic of senescent cells in various tissues, including stellate cells of liver and pancreas [11,12]. Expression of Cdkn1a is regulated both through p53-dependent and p53-independent pathways [18e 20]. Originally implicated in cell cycle checkpoints in G1 and S phases by inhibiting activities of cyclin E-CDK2 and cyclin A-CDK2 complexes, Cdkn1a has later on also been shown to be capable of causing cell cycle arrest in the G2 and M phases (reviewed in [21]). Although a stop of cell cycle progression is a prerequisite of cellular senescence, it is not necessarily also sufficient for its induction. In this study, we have therefore addressed the question if expression of Cdkn1a is an essential component of the senescence process in pancreatic stellate cells, or an epiphenomenon only. Using a siRNA approach, we found that a knockdown of Cdkn1a not
260
B. Fitzner et al. / Pancreatology 13 (2013) 254e262
Fig. 6. Effects of a Cdkn1a knockdown on MMP-9 activity and collagen type I secretion. PSC were treated as described in Fig. 3 before supernatants were collected and subjected to the quantification of MMP activities by zymography (A) and PICP levels by ELISA (B). (A) The upper panel shows a typical example of a zymography gel. MMP-9 was identified based on the molecular weight of 92 kDa and by running a MMP-9 standard (data not shown). Lower panel: Quantified band intensities were used to calculate mean values and SEM (n ¼ 3). (B) Data obtained from 7 independent samples were combined to calculate mean values and SEM. #P < 0.05 versus control cells cultured with doxo (columns 4 vs. column 3). There were no statistically significant differences between controls cultured without doxo (column 1) and all other groups. Only siRNA2 was tested.
only stimulated cell proliferation, but also diminished doxoinduced senescence of rat PSC. Furthermore, overexpression of Cdkn1a in a rat stellate cell line strongly enhanced the induction of senescence by the DNA-damaging agent. Together, these data provide evidence that the Cdkn1a protein is directly involved in the process of rat PSC senescence. However, we observed that the knockdown only reduced the number of senescent cells whereas even under the conditions of overexpression the majority of cells remained in a non-senescent stage. These data need to be interpreted cautiously, since the expression of Cdkn1a was only knocked down but not abolished, whereas overexpression was performed in immortalized cells. Nevertheless, our results support the hypotheses that expression of high levels of Cdkn1a alone is not sufficient to trigger senescence, and Cdkn1a-independent pathways contribute to PSC senescence as well. With the following exceptions, the gene expression data obtained in this study are in agreement with the concept of Cdkn1a as a mediator of doxo effects in PSC: The knock-down of Cdkn1a had no effect on IL-6 mRNA expression and even strongly increased the expression of MMP-9; an effect that was also verified at the level of MMP-9 enzyme activity. Furthermore, cells treated with siRNA against Cdkn1a secreted less collagen type I. Since doxo treatment alone did not significantly alter PICP levels, the latter finding could
not be directly linked to the senescence process in this study and should be investigated further before drawing final conclusions. Here and in our previous experiments [12] we found that senescent PSC, like quiescent PSC, expressed lower levels of a-SMA protein than fully activated stellate cells. Another common feature of quiescent and senescent PSC is their low or absent proliferative activity. In contrast to quiescent PSC, senescent cells are, however, characterized by an increased size and a typical “flat” morphology [12]. Furthermore, they do also not contain significant amounts of Vitamin A-storing fat droplets (data not shown). Interestingly, the knock-down of Cdkn1a inhibited the doxo-dependent increase of SA b-Gal-positive cells but not the decrease of a-SMA protein expression. Together, these data suggest that some of the doxo effects in PSC are at least less sensitive to the downregulation of the Cdkn1a protein level than others. More specifically, they support the hypothesis of two primarily independent effects of doxo, induction of PSC senescence and inhibition of PSC activity (indicated by reduced levels of a-SMA), which may be separated from each other through the regulation of Cdkn1a protein levels. PSC have recently been shown to display characteristics of stem cells [22,23], and proposed to play a role not only in (pathological) fibrosis but also in normal tissue homeostasis and regeneration [23e25]. In this regard it is interesting to note that PSC cultures
B. Fitzner et al. / Pancreatology 13 (2013) 254e262
261
Fig. 7. Tet-regulated overexpression of Cdkn1a. LTC-Cdkn1a cells were cultured in the presence and absence of tet (1 mg/ml) and doxo (10 ng/ml) as indicated. (A) 48 h after tet withdrawal, expression of Cdkn1a (insert, upper panel) and b-actin (insert, lower panel) was analyzed by immunoblotting using specific antibodies. For quantitative evaluation, signal intensities were determined, and the ratio Cdkn1a/b-actin was calculated. A ratio of one hundred percent corresponds to cells cultured with tet. Data from n ¼ 12 independent samples were used to calculate mean values and SEM. *P < 0.05 versus controls cultured with tet. (B) Cell proliferation was assessed with the BrdU DNA-incorporation (inc) assay as described in the “Materials and methods” section. (C) On day 7 after tet withdrawal and doxo application, SA b-Gal-positive cells were determined based on a quantitative image evaluation and related to the total cell number. (B and C) Data from n 11 separate cultures were used to calculate mean values and SEM. *P < 0.05 versus controls cultured with tet/without doxo, #P < 0.05 versus control cells cultured with doxo (column 4 vs. 3).
never underwent complete senescence; neither in response to doxo (this study) nor to other triggers such as long-term culture and oxidant stress [12], suggesting a persistent self-renewal capacity. This interesting aspect of PSC biology should be further addressed in follow-up studies. Disclosure statement All authors declare to have no conflict of interest. Acknowledgments We gratefully acknowledge the excellent technical assistance of Mrs. Katja Bergmann and Mrs. Katrin Sievert-Küchenmeister. This work was supported by a grant from the Deutsche Forschungsgemeinschaft (to RJ). References [1] Erkan M, Adler G, Apte MV, Bachem MG, Buchholz M, Detlefsen S, et al. StellaTUM: current consensus and discussion on pancreatic stellate cell research. Gut 2012;61:172e8. [2] Jaster R, Emmrich J. Crucial role of fibrogenesis in pancreatic diseases. Best Pract Res Clin Gastroenterol 2008;22:17e29. [3] McCarroll JA, Phillips PA, Santucci N, Pirola RC, Wilson JS, Apte MV. Vitamin A inhibits pancreatic stellate cell activation: implications for treatment of pancreatic fibrosis. Gut 2006;55:79e89. [4] Jaster R, Hilgendorf I, Fitzner B, Brock P, Sparmann G, Emmrich J, et al. Regulation of pancreatic stellate cell function in vitro: biological and molecular effects of all-trans retinoic acid. Biochem Pharmacol 2003;66:633e41.
[5] Froeling FE, Feig C, Chelala C, Dobson R, Mein CE, Tuveson DA, et al. Retinoic acid-induced pancreatic stellate cell quiescence reduces paracrine Wnt-bcatenin signaling to slow tumor progression. Gastroenterology 2011;141: 1486e97. 1497.e1-14. [6] Masamune A, Kikuta K, Satoh M, Sakai Y, Satoh A, Shimosegawa T. Ligands of peroxisome proliferator-activated receptor-g block activation of pancreatic stellate cells. Biol Chem 2002;277:141e7. [7] Jaster R, Lichte P, Fitzner B, Brock P, Glass Ä, Karopka T, et al. Peroxisome proliferator-activated receptor g overexpression inhibits pro-fibrogenic activities of immortalized rat pancreatic stellate cells. J Cell Mol Med 2005;3: 670e82. [8] Klonowski-Stumpe H, Fischer R, Reinehr R, Lüthen R, Häussinger D. Apoptosis in activated rat pancreatic stellate cells. Am J Physiol Gastrointest Liver Physiol 2002;283:G819e26. [9] Rickmann M, Vaquero EC, Malagelada JR, Molero X. Tocotrienols induce apoptosis and autophagy in rat pancreatic stellate cells through the mitochondrial death pathway. Gastroenterology 2007;132:2518e32. [10] Manapov F, Muller P, Rychly J. Translocation of p21(Cip1/WAF1) from the nucleus to the cytoplasm correlates with pancreatic myofibroblast to fibroblast cell conversion. Gut 2005;54:814e22. [11] Krizhanovsky V, Yon M, Dickins RA, Hearn S, Simon J, Miething C, et al. Senescence of activated stellate cells limits liver fibrosis. Cell 2008;134:657e67. [12] Fitzner B, Müller S, Walther M, Fischer M, Engelmann R, Müller-Hilke B, et al. Senescence determines the fate of activated rat pancreatic stellate cells. J Cell Mol Med 2012;16:2620e30. [13] Krizhanovsky V, Xue W, Zender L, Yon M, Hernando E, Lowe SW, et al. Implications of cellular senescence in tissue damage response, tumor suppression, and stem cell biology. Cold Spring Harb Symp Quant Biol 2008;73: 513e22. [14] Campisi J, d’Adda di Fagagna F. Cellular senescence: when bad things happen to good cells. Nat Rev Mol Cell Biol 2007;8:729e40. [15] Jaster R, Sparmann G, Emmrich J, Liebe S. Extracellular signal-regulated kinases are key mediators of mitogenic signals in rat pancreatic stellate cells. Gut 2002;51:579e84. [16] Rateitschak K, Karger A, Fitzner B, Lange F, Wolkenhauer O, Jaster R. Mathematical modelling of interferon-g signalling in pancreatic stellate cells
262
[17]
[18]
[19] [20]
B. Fitzner et al. / Pancreatology 13 (2013) 254e262 reflects and predicts dynamics of STAT1 pathway activity. Cell Signal 2010;22: 97e105. Luo S, Wehr NB, Levine RL. Quantitation of protein on gels and blots by infrared fluorescence of Coomassie blue and Fast Green. Anal Biochem 2006;350:233e8. Phalke S, Mzoughi S, Bezzi M, Jennifer N, Mok WC, Low DH, et al. p53Independent regulation of p21Waf1/Cip1 expression and senescence by PRMT6. Nucleic Acids Res 2012;40:9534e42. el-Deiry WS, Tokino T, Velculescu VE, Levy DB, Parsons R, Trent JM, et al. WAF1, a potential mediator of p53 tumor suppression. Cell 1993;75:817e25. Dulic V, Kaufmann WK, Wilson SJ, Tlsty TD, Lees E, Harper JW, et al. p53dependent inhibition of cyclin-dependent kinase activities in human fibroblasts during radiation-induced G1 arrest. Cell 1994;76:1013e23.
[21] Romanov VS, Pospelov VA, Pospelova TV. Cyclin-dependent kinase inhibitor p21(Waf1): contemporary view on its role in senescence and oncogenesis. Biochemistry (Mosc) 2012;77:575e84. [22] Mato E, Lucas M, Petriz J, Gomis R, Novials A. Identification of a pancreatic stellate cell population with properties of progenitor cells: new role for stellate cells in the pancreas. Biochem J 2009;421:181e91. [23] Kordes C, Sawitza I, Götze S, Häussinger D. Stellate cells from rat pancreas are stem cells and can contribute to liver regeneration. PLoS One 2012;7:e51878. [24] Means AL. Pancreatic stellate cells: small cells with a big role in tissue homeostasis. Lab Invest 2013;93:4e7. [25] Riopel MM, Li J, Liu S, Leask A, Wang R. b1 integrin-extracellular matrix interactions are essential for maintaining exocrine pancreas architecture and function. Lab Invest 2013;93:31e40.
Contents lists available at SciVerse ScienceDirect
Pancreatology journal homepage: www.elsevier.com/locate/pan
Original article
Cdkn1a is a key mediator of rat pancreatic stellate cell senescence Brit Fitzner, Andreas Lange, Sarah Müller, Robert Jaster* Department of Medicine II, Division of Gastroenterology, University Medicine Rostock, E.-Heydemann-Str. 6, D-18057 Rostock, Germany
a r t i c l e i n f o
a b s t r a c t
Article history: Received 16 January 2013 Received in revised form 27 February 2013 Accepted 8 March 2013
Background/objectives: Completion of pancreatic wound healing requires termination of pancreatic stellate cell (PSC) activation to prevent fibrosis. Besides induction of apoptosis and return to a quiescent phenotype, senescence of PSC followed by immune cell-mediated cytolysis represents a potential mechanism. Here, we have studied if the cell cycle inhibitor cyclin-dependent kinase inhibitor 1A (Cdkn1a, p21/Waf1), expression of which is increased in senescent rat PSC, plays a causative role in the senescence process. Methods: Senescence was induced by doxorubicin treatment. The functions of Cdkn1a were analyzed using two approaches, treatment of primary rat PSC with siRNA and tetracycline-regulated overexpression of Cdkn1a in immortalized rat cells. Expression of senescence-associated b-galactosidase (SA b-Gal) was used as a surrogate marker of senescence. Results: The knockdown of Cdkn1a significantly attenuated the growth-inhibitory effect of doxorubicin and strongly diminished the portion of SA b-Gal-positive cells. Overexpression of Cdkn1a enhanced both the antiproliferative effect of doxorubicin and induction of senescence. In primary PSC, doxorubicin treatment was associated with increased expression of interleukin-6 (IL-6) and matrix metalloproteinase (MMP)-9, while expression of the activation marker a-smooth muscle actin (a-SMA), p53, Cdk1 and Rad54 was diminished. The application of Cdkn1a siRNA specifically antagonized the effects of doxorubicin on the expression of p53, Cdk1 and Rad54 but not IL-6 and a-SMA, while MMP-9 expression and also activity were even enhanced. Conclusions: Cdkn1a plays a direct role in the process of rat PSC senescence. Additional Cdkn1aindependent pathways may contribute to the partial maintenance of a gene expression profile typical of senescent PSC. Copyright Ó 2013, IAP and EPC. Published by Elsevier India, a division of Reed Elsevier India Pvt. Ltd. All rights reserved.
Keywords: Chronic pancreatitis Fibrosis Cellular ageing
1. Introduction The wound healing reaction in the pancreas involves proliferation and activation of pancreatic stellate cells (PSC), the main source of extracellular matrix (ECM) proteins in the diseased organ. In chronic pancreatitis and in pancreatic cancer, persistent activation of PSC results in pancreatic fibrosis, which is considered a Abbreviations: a-SMA, a-smooth muscle actin; BrdU, 5-bromo-20 -deoxyuridine; Cdkn1a, cyclin-dependent kinase inhibitor 1A; CP, chronic pancreatitis; DAPI, 40 ,6Diamidino-2-phenylindole; doxo, doxorubicin; ECM, extracellular matrix; FCS, fetal calf serum; HA, hemagglutinin; IL-6, interleukin-6; IMDM, Iscove’s Modified Dulbecco’s Medium; MMP-9, matrix metalloproteinase 9; PI, propidium iodide; PSC, pancreatic stellate cell; PVDF, polyvinylidene difluoride; SA b-Gal, senescenceassociated b-galactosidase; SASP, senescence-associated secretory phenotype; SEM, standard error of the mean; tet, tetracycline; X-Gal, 5-bromo-4-chloro-indolyl-b-Dgalactopyranoside. * Corresponding author. Tel.: þ49 381 494 7349; fax: þ49 381 494 7482. E-mail address: [email protected] (R. Jaster).
progression factor of the tumor disease. Although the mechanisms leading to perpetuation of PSC activation are incompletely understood, it is likely that cytokines and growth factors derived from injured acinar cells, inflammatory cells, tumor cells and activated PSC themselves play a key role in this process [1,2]. The pathogenetic relevance of pancreatic fibrosis has also raised the question how PSC activation may be terminated. Previous studies of several laboratories, including our own one, have identified factors such as vitamin A derivates and ligands of the peroxisome proliferator activated receptor g, that favor exhibition of a quiescent PSC phenotype [3e7]. However, since activation of PSC is associated with cell proliferation, additional mechanisms appear necessary to eliminate a surplus of ECM-producing cells. One obvious possibility is that PSC, after completing their repair function, may undergo apoptosis, and indeed experimental evidence in support of this hypothesis has been provided [7e10]. In addition, studies in the field of liver fibrosis have suggested another possibility: The process of cellular senescence may render stellate cells susceptible to
1424-3903/$ e see front matter Copyright Ó 2013, IAP and EPC. Published by Elsevier India, a division of Reed Elsevier India Pvt. Ltd. All rights reserved. http://dx.doi.org/10.1016/j.pan.2013.03.009
B. Fitzner et al. / Pancreatology 13 (2013) 254e262
cytolysis mediated by immune cells; specifically natural killer cells [11]. Using a rat model of chronic pancreatitis, we have recently shown that senescent cells, most likely PSC, are also detectable in the fibrotic pancreas [12]. Cellular senescence represents a state of irreversible cell cycle arrest and is considered a key barrier mechanism against tumorigenesis [13,14]. In addition to telomere shortage in the course of multiple cell divisions, diverse forms of cellular stress have been implicated in the induction of senescence [11,13,14]. In our previous study we found that senescence of isolated rat PSC can be triggered by long-term culture, the DNA-damaging agent doxorubicin (doxo), hydrogen peroxide and a low concentration of the pro-apoptotic substance staurosporine [12]. Senescent rat PSC mimicked nonproliferating quiescent PSC also in that they expressed low levels of the activation marker a-smooth muscle actin (a-SMA). In contrast to quiescent PSC however, senescent stellate cells expressed high levels of senescence-associated b-galactosidase (SA b-Gal) and displayed a gene expression profile compatible with a senescence-associated secretory phenotype (SASP); e.g. synthesis of large amounts of interleukin-6 (IL-6). Exhibition of SASP and expression of SA b-Gal are two out of many established surrogate markers of cellular senescence [11,13,14]. Another typical phenomenon represents the overexpression of cell cycle inhibitors. Indeed, we observed that rat PSC senescence is accompanied by increased mRNA levels of Cdkn1a (p21/Waf1) [12]. In this study, we have addressed the question if expression of Cdkn1a is an epiphenomenon of the senescence process, or plays an active role in its progression. Using two independent experimental approaches, application of siRNA against Cdkn1a and overexpression of the protein, we found that doxo-induced senescence of rat PSC is mediated both through Cdkn1a-dependent and eindependent pathways. 2. Methods 2.1. Cell isolation and culture Rat PSC were isolated by collagenase digestion of the pancreas followed by Nycodenz density gradient centrifugation as previously described [15]. They were cultured in Iscove’s Modified Dulbecco’s Medium (IMDM) supplemented with 17% fetal calf serum, 1% nonessential amino acids (dilution of a 100 stock solution), 100 U/ml penicillin and 100 mg/ml streptomycin at 37 C in a 5% CO2 humidified atmosphere. The cells were grown on culture plates to subconfluency, harvested by trypsination and recultured at equal seeding densities. All experiments were performed with cells recultured no more than two times. Stellate cell lines were cultured in IMDM supplemented with 10% fetal calf serum (FCS), non-essential amino acids, 100 U/ml penicillin and 100 mg/ml streptomycin at 37 C in a 5% CO2 humidified atmosphere. Tetracycline (tet) was applied as described below. 2.2. Inhibition of Cdkn1a expression using siRNA siRNAs were purchased from Riboxx (Radebeul, Germany). Out of four siRNAs tested, the following two guide strand sequences displayed the strongest inhibitory effect on Cdkn1a expression and were selected for functional studies: 50 -UAAAGAUAAACAGUCCAGCCCCC-30 , and 50 -AUAGAAAUCUGUUAGGCUGCCCC-30 (subsequently termed siRNA1 and siRNA2, respectively). The guide strand sequence 50 -UAAGCACGAAGCUCAGAGUCCCCC-30 served as non-silencing control. For transfection with siRNA, PSC were trypsinized, centrifuged and resuspended in RPMI 1640 culture medium supplemented with TrueFect-Lipo transfection reagent
255
according to the suggestions of the manufacturer (United Biosystems, Herndon, VA, USA) and siRNA at a final concentration of 100 nM. Afterwards, the cells were seeded into culture dishes of variable size according to the experimental requirements and allowed to attach for 4 h before the transfection medium was substituted by regular culture medium supplemented with doxo at 25 ng/ml as indicated. To ensure maintenance of silencing in longterm studies (>48 h) with siRNA-transfected cells, the transfection procedure was repeated every 2e3 days, except of that PSC were not trypsinized prior to siRNA application. Efficiency of siRNA treatment was monitored by analyzing expression of Cdkn1a mRNA using real-time PCR. 2.3. Vector construction and expression of Cdkn1a The cDNA of the murine Cdkn1a gene (AF457188; complete coding sequence) was cloned into the Nhe I site of pTRE2purHA (confers puromycin resistance; Invitrogen, Karlsruhe, Germany), and the correct sequence and orientation of the insert was verified by dideoxy sequencing. The resulting constructs encode for a Cdkn1a protein with an N-terminal hemagglutinin (HA)-tag, expression of which is controlled by a tet response element. For Cdkn1a overexpression studies, we used LTC-Tet cells [7], an immortalized rat PSC line that stably expresses the transcriptional activator protein tTA, which is inactivated upon tet binding. LTC-Tet cells were transfected with the pTRE2purHA-Cdkn1a construct by lipofection. Therefore, cells growing at a low density (5% confluency) in 6-well plates were incubated with plasmid DNA (2 mg per well), using FuGENE HD Transfection Reagent (Roche, Mannheim, Germany) according to the vendor’s specifications. After transfection, the cells were cultured in medium supplemented with tet (1 mg/ml) for 48 h, before puromycin was added at 2 mg/ml. Upgrowing resistant cells were tested for a tet-dependent Cdkn1a expression profile, and clones with a strong induction of Cdkn1a transgene expression after tet withdrawal were selected. Unless otherwise stated, puromycin-resistant cells (referred to as LTCCdkn1a cells) were permanently cultured with tet (1 mg/ml). The data shown were verified in three independent cell clones of LTCCdkn1a. 2.4. Quantification of DNA synthesis Cell proliferation was assessed using a 5-bromo-20 -deoxyuridine (BrdU) labeling and detection enzyme-linked immunosorbent assay kit (Roche). In experiments with siRNA, PSC resuspended in transfection medium supplemented as described above were seeded at equal densities into 96-well plates and treated with doxo (at 25 ng/ml) as indicated. For overexpression studies, LTC-Cdkn1a cells were washed free of tet, trypsinized and also seeded into microplates. Tet and doxo (10 ng/ml, since a dose of 25 ng/ml displayed toxic effects) were added as specified below. In all studies, BrdU labeling was initiated 24 h after drug application or withdrawal by adding labeling solution at a final concentration of 10 mM to the culture medium. After another incubation period of 24 h, the labeling was stopped, and BrdU uptake was measured according to the manufacturer’s instructions. 2.5. Detection of SA b-Gal Transfected PSC and LTC-Cdkn1a cells (exposed to tet as indicated) were cultured on coverslips for 7 days in the presence and absence of doxo, respectively. Afterwards, SA b-Gal-positive cells were stained using a specific kit (New England BioLabs, Frankfurt, Germany), whereas the nuclei were counterstained with DAPI according to standard protocols. The assay used in this study
256
B. Fitzner et al. / Pancreatology 13 (2013) 254e262
monitors enzymatic activity, which is required to cleave the substrate X-Gal. Since the reaction is performed at pH 6 where other cellular b-galactosidases are inactive, the assay is specific for SA b-gal. Subsequently, specimens were analyzed by light and fluorescence microscopy. Image evaluation was performed as previously described [12], using an ImageMagick-based software to distinguish between SA b-Gal-positive and -negative cells. 2.6. Immunoblotting Protein extracts of cultured cells were prepared, adjusted to identical protein concentrations and subjected to immunoblot analysis as published before [15], using PVDF membrane for protein transfer. The following primary antibodies were employed: Anti-bactin (# 4970, New England BioLabs, Frankfurt, Germany), Anti-aSMA (# A2547, SigmaeAldrich, Deisenhofen, Germany) and Anti-Cdkn1a (# SX118, BD Biosciences, Heidelberg, Germany). The blots were developed using LI-COR reagents for an OdysseyÒ Infrared Imaging System as previously described [16]. Signal intensities of the investigated proteins were quantified by means of the OdysseyÒ software version 3.0, and normalized to b-actin prior to the comparison of expression levels. 2.7. Quantitative reverse transcriptase PCR using real-time TaqManÔ technology Transfected PSC were grown in 12-well plates for 7 days, before total RNA was isolated with peqGOLD TriFast reagent (Peqlab Biotechnologie, Erlangen, Germany) following the manufacturer’s instructions. After DNAse digestion of genomic DNA, RNA was reverse transcribed into cDNA using Maxima Probe Irox qPCR Master Mix (Fisher Scientific, Schwerte, Germany) and random hexamer priming. Relative quantification of target cDNA levels by real-time PCR was performed in an ABI Prism 7000 sequence detection system employing TaqManÔ Universal PCR Master Mix and the following Assay-on-DemandÔ rat gene-specific fluorescently labeled TaqManÔ MGB probes (instrument and reagents: Life Technologies, Darmstadt, Germany): Rn00589996_m1 (Cdkn1a/p21), Rn00570728_m1 (Cdk1), Rn00561420_m1 (IL-6), Rn00579162_m1 (MMP-9), Rn00755717_m1 (p53), Rn01416981_m1 (Rad54) and Rn01527840_m1 (Hprt). PCR was performed under the following conditions: 95 C for 10 min, 50 cycles of 15 s at 95 C, 1 min at 60 C. Relative expression of each mRNA compared with the housekeeping gene Hprt was calculated according to the equation DCt ¼ Cttarget CtHPRT. The relative amount of target mRNA in control cells and cells treated as indicated was expressed as 2(DDCt), where DDCttreatment ¼ DCtsample DCt control. The reactions were performed in triplicate, and repeated at least 6 times with independent samples.
resuspended in 2 ml ice-cold 70% ethanol for at least 12 h at 4 C. Following further washing steps, the cells were incubated for 20 min in 400 ml PBS containing 0.1 mg/ml RNase A (Roche) at 37 C. After the addition of 50 mg propidium iodide (PI; SigmaeAldrich) per ml, the samples were subjected to cytofluorometric analysis. 10.000 events were measured for each sample and the data stored in list-mode for further analysis. The cell cycle distribution was calculated using the Cellquest program. Cells of the Sub-G1 peak were considered apoptotic. 2.9. Quantification of procollagen I carboxy terminal propeptide (PICP) To monitor collagen type I synthesis and secretion, supernatants from PSC cultures (pretreated as indicated) were collected, cleared by centrifugation and stored at 80 C until assayed. PICP levels were measured using a rat PICP-specific ELISA according to the instructions of the manufacturer (Uscn Life Science, Houston, TX, USA). Measured PICP concentrations were normalized with respect to unequal cell densities based on the quantification of b-actin levels in total cellular lysates by Western blotting. 2.10. Zymography Supernatants of PSC cultures were obtained and processed as described above, mixed 1:1 with Zymogram Sample Buffer (BioRad Laboratories, München, Germany) and loaded onto 8% SDSe polyacrylamide gels supplemented with 1% gelatine. After electrophoresis at 4 C, the gels were soaked in 1 Zymogram Renaturation Buffer (BioRad) with gentle shaking (2 h at room temperature), before they were incubated overnight at 37 C in 1 Zymogram Development Buffer (BioRad). Finally, lysis bands were visualized by Coomassie staining (30 min at room temperature followed by three washing steps with distilled water) and, taking advantage of the fluorescence of the stain [17], quantitated using the OdysseyÒ Infrared Imaging System. The raw data were normalized as described for PICP. 2.11. Statistical analysis Results are expressed as mean standard error of the mean (SEM) for the indicated number of separate cultures per
2.8. Detection of dead cells and analysis of cellular DNA content by flow cytometry To quantify dead cells by flow cytometry, PSC of passage 1 were exposed to doxo (25 ng/ml), or left untreated. After 7 days, the cells were harvested by trypsination, resuspended in FACS-buffer (PBS pH 7,4; 0.5% bovine serum albumin; 0.1% sodium azide) and kept on ice until measurement. Subsequently, the samples were labeled with propidium iodide (PI; 10 mg/ml). PI-positive (dead) cells were quantified using a FACSCalibur cytometer (BD Biosciences) and the Cellquest program. For the detection of the cellular DNA content, the following protocol was employed: PSC transfected with siRNA were grown for 48 h in 6-well plates in the presence and absence of doxo as indicated. To obtain a single cell suspension, they were trypsinized, pelleted by centrifugation, washed twice with PBS and
Fig. 1. Inhibition of Cdkn1a expression with siRNA. Cultured PSC were transfected with Cdkn1a-specific siRNAs or a non-silencing oligonucleotide (controls; all siRNAs at 100 nM) and exposed to doxo (at 25 ng/ml) as indicated. After an incubation period of 16 h, mRNA expression of Cdkn1a and Hprt (housekeeping control) was analyzed by real-time PCR, and relative amounts of Cdkn1a mRNA were calculated. One hundred percent Cdkn1a expression corresponds to controls cultured without doxo. Data of n 5 independent experiments (with duplicate samples) were used to calculate mean values and SEM. *P < 0.05 versus controls cultured without doxo, #P < 0.05 versus control cells cultured with doxo (columns 5 and 6 vs. column 4).
B. Fitzner et al. / Pancreatology 13 (2013) 254e262
experimental protocol. Statistical significance was analyzed using the indicated statistical test. P < 0.05 was considered to be statistically significant. 3. Results To study the role of Cdkn1a in PSC senescence, expression of Cdkn1a in rat PSC was inhibited by means of siRNA. As shown in Fig. 1, two different siRNAs against Cdkn1a strongly diminished both basal and doxo-induced Cdkn1a expression. An efficient knock-down was observed 16 h after transfection (Fig. 1), and maintained over at least 7 days (data not shown). Analysis of cell proliferation revealed a small but significant growth advantage of PSC transfected with Cdkn1a siRNA1 under
257
basal conditions (no doxo). More importantly, both Cdkn1a siRNAs significantly attenuated the growth-inhibitory effect of doxo (Fig. 2A), implicating Cdkn1a in the action of the drug. Consistent with these findings, more doxo-treated PSC were found in the S-phase of the cell cycle when the cells were transfected with siRNA against Cdkn1a (Fig. 2B). Under all conditions (doxo and/or siRNA), the portion of cells in the sub-G1 peak remained below 4%, suggesting that only very few cells underwent apoptosis. As indicated by the results of PI staining, the total portion of dead (apoptotic and necrotic) cells after doxo treatment was approximately 10%, suggesting that cytotoxic effects only modestly contributed to doxo-mediated growth inhibition. Fig. 2C shows the results for cells treated doxo in the absence of siRNA. Application of the Cdkn1a siRNAs did at least
Fig. 2. Downregulation of Cdkn1a reduces the antiproliferative effect of doxorubicin. (A and B) Cultured PSC were transfected with Cdkn1a-specific or control siRNA (at 100 nM) and exposed to doxo (at 25 ng/ml) for 48 h as indicated. (A) Cell proliferation was assessed with the BrdU DNA-incorporation (inc) assay as described in the “Materials and methods” section. One hundred percent cell proliferation corresponds to controls cultured without doxo. (B) Cells in the S-phase of the cell cycle were detected by FACS analysis based on their DNA content and related to the total number of cells. (A and B) Data from n 10 separate cultures were used to calculate mean values and SEM. *P < 0.05 versus controls cultured without doxo, #P < 0.05 versus control cells cultured with doxo (columns 5 and 6 vs. column 4). (C) PSC were cultured with or without doxo (at 25 ng/ml) for 7 days; the time corresponding to the conditions of the senescence assays (Fig. 3). Afterwards, they were subjected to a cytofluorometric quantification of dead (PI-positive) cells and apoptotic (SubG1 peak) cells (expressed as percent of total cells). Data from n 6 separate cultures were used to calculate mean values and SEM. *P < 0.05 versus controls cultured without doxo.
258
B. Fitzner et al. / Pancreatology 13 (2013) 254e262
not increase the small portion of dead cells in doxo-treated cultures (data not shown). In agreement with our previous studies [12], we found that doxo treatment for 7 days induced expression of the senescence marker SA b-Gal in a significant portion of cultured rat PSC while inhibiting expression of the activation marker a-SMA (Figs. 3 and 4). In the presence of non-silencing control siRNA, senescence of almost one third of the cells was observed. Treatment with either of the two Cdkn1a siRNAs reduced this portion by roughly 50% (Fig. 3). In contrast, doxo-dependent inhibition of a-SMA expression was not diminished by the Cdkn1a siRNAs but even modestly enhanced (Fig. 4). We have previously reported that doxo-induced senescence of rat PSC is associated with characteristic changes of gene expression, including upregulation of IL-6 and downregulation of p53 mRNA levels [12]. As shown in Fig. 5, treatment of PSC with Cdkn1a siRNAs did not interfere with doxo-mediated IL-6 expression (A), but completely antagonized the inhibitory effect of the drug on the expression of p53 (B). In a similar manner, the Cdkn1a-specific siRNAs also blocked downregulation of cyclin-dependent kinase 1 (Cdk1) (C) and Rad54 (D) mRNA levels by the DNA-damaging agent. In contrast, doxo-triggered expression of MMP-9 (E) was strongly enhanced in cells treated with Cdkn1a siRNA. To validate the latter findings, we also quantified MMP-9 activities in PSC supernatants by zymography (Fig. 6A). Again, a doxo-dependent increase was observed that was strongly enhanced by treatment with Cdkn1aspecific siRNA. In contrast, activities of MMP-2 remained unchanged (data not shown). As indicated by the measurements of PICP levels in the same supernatants, doxo treatment of control cultures had no significant effect on the secretion of collagen type 1 (Fig. 6B). Here, a knockdown of Cdkn1a lead to significantly
Fig. 4. Inhibition of a-SMA expression by doxorubicin is independent of Cdkn1a. PSC were treated as described in Fig. 3. (A) Expression of a-SMA and the housekeeping protein b-actin was analyzed by immunoblotting. (B) For quantitative evaluation, signal intensities were determined, and the ratio a-SMA/b-actin was calculated. A ratio of one hundred percent corresponds to controls transfected with non-silencing siRNA and cultured without doxo (column 1). Data from n ¼ 9 independent samples per data point were used to calculate mean values and SEM. *P < 0.05 versus controls cultured without doxo, #P < 0.05 versus control cells cultured with doxo (columns 4 vs. 3 and 6 vs. 5, respectively).
Fig. 3. Cdkn1a-specific siRNA inhibits doxorubicin-dependent senescence of PSC. Cultured PSC were transfected with Cdkn1a-specific or control siRNA (at 100 nM) and exposed to doxo (at 25 ng/ml) for 7 days as indicated. Afterwards, the cells were stained for expression of SA b-Gal and the nuclei where counterstained with DAPI. Panel (A) shows typical microphotographs of PSC treated with non-silencing control siRNA and Cdkn1a-specific siRNA2, respectively, in the presence and absence of doxo. (B) SA b-Gal-positive cells were determined based on a quantitative image evaluation and related to the total cell number. The results are averaged values (SEM) of 12 independent experiments. *P < 0.05 versus controls cultured without doxo, #P < 0.05 versus control cells cultured with doxo (columns 5 and 6 vs. column 4).
B. Fitzner et al. / Pancreatology 13 (2013) 254e262
259
Fig. 5. Effects of Cdkn1a-specific siRNA on PSC gene expression. PSC were transfected with Cdkn1a-specific or non-silencing control siRNA (at 100 nM) and treated with doxo (at 25 ng/ml) for 7 days as indicated. The mRNA expression of IL-6 (A), p53 (B), Cdk1 (C), RAD54 (D), MMP-9 (E) and Hprt (housekeeping control) was analyzed by real-time PCR, and relative amounts of target mRNA were determined. The results are averaged values (SEM) of at least 6 independent experiments. One hundred percent mRNA expression corresponds to controls cultured without doxo. *P < 0.05 versus controls cultured without doxo, #P < 0.05 versus control cells cultured with doxo (columns 5 and 6 vs. column 4).
decreased levels of PICP (comparison of doxo-treated cultures), suggesting different roles of Cdkn1a in the regulation of MMP-9 and collagen type 1 expression. To verify the biological effects of Cdkn1a in stellate cells of the pancreas, Tet-regulated overexpression of a Cdkn1a transgene was employed as an independent approach. Therefore, cells of the immortalized rat pancreatic stellate cell line LTC-Tet cells (stably expressing the tTA protein [7]), were transfected with the pTRE2purHA-Cdkn1a vector as described above. LTC-Cdkn1a cells cultured with tet at 1 mg/ml did not express detectable amounts of Cdkn1a protein (Fig. 7A), suggesting low levels of the endogenous Cdkn1a protein and suppression of transgene transcription. Withdrawal of tet induced a strong expression of the transgenic protein, which was associated with an inhibition of DNA synthesis and a significant enhancement of the antiproliferative effect of doxo (Fig. 7B). Analysis of SA b-Gal expression revealed that expression of the Cdkn1a transgene alone was not sufficient to induce senescence of a significant portion of LTC-Cdkn1a cells (Fig. 7C). In this immortalized cell line, the effect of doxo was also weaker than in primary cells (Fig. 3) but nevertheless clear (18% vs. 3% SA b-Galpositive cells in the presence and absence of doxo, respectively).
When tet withdrawal and doxo treatment were combined, a nearly 2-fold increase of SA b-Gal-positive cells was observed, suggesting that expression of the Cdkn1a transgene significantly enhanced induction of cellular senescence by doxo. 4. Discussion Increased levels of the cell cycle inhibitor Cdkn1a are a typical characteristic of senescent cells in various tissues, including stellate cells of liver and pancreas [11,12]. Expression of Cdkn1a is regulated both through p53-dependent and p53-independent pathways [18e 20]. Originally implicated in cell cycle checkpoints in G1 and S phases by inhibiting activities of cyclin E-CDK2 and cyclin A-CDK2 complexes, Cdkn1a has later on also been shown to be capable of causing cell cycle arrest in the G2 and M phases (reviewed in [21]). Although a stop of cell cycle progression is a prerequisite of cellular senescence, it is not necessarily also sufficient for its induction. In this study, we have therefore addressed the question if expression of Cdkn1a is an essential component of the senescence process in pancreatic stellate cells, or an epiphenomenon only. Using a siRNA approach, we found that a knockdown of Cdkn1a not
260
B. Fitzner et al. / Pancreatology 13 (2013) 254e262
Fig. 6. Effects of a Cdkn1a knockdown on MMP-9 activity and collagen type I secretion. PSC were treated as described in Fig. 3 before supernatants were collected and subjected to the quantification of MMP activities by zymography (A) and PICP levels by ELISA (B). (A) The upper panel shows a typical example of a zymography gel. MMP-9 was identified based on the molecular weight of 92 kDa and by running a MMP-9 standard (data not shown). Lower panel: Quantified band intensities were used to calculate mean values and SEM (n ¼ 3). (B) Data obtained from 7 independent samples were combined to calculate mean values and SEM. #P < 0.05 versus control cells cultured with doxo (columns 4 vs. column 3). There were no statistically significant differences between controls cultured without doxo (column 1) and all other groups. Only siRNA2 was tested.
only stimulated cell proliferation, but also diminished doxoinduced senescence of rat PSC. Furthermore, overexpression of Cdkn1a in a rat stellate cell line strongly enhanced the induction of senescence by the DNA-damaging agent. Together, these data provide evidence that the Cdkn1a protein is directly involved in the process of rat PSC senescence. However, we observed that the knockdown only reduced the number of senescent cells whereas even under the conditions of overexpression the majority of cells remained in a non-senescent stage. These data need to be interpreted cautiously, since the expression of Cdkn1a was only knocked down but not abolished, whereas overexpression was performed in immortalized cells. Nevertheless, our results support the hypotheses that expression of high levels of Cdkn1a alone is not sufficient to trigger senescence, and Cdkn1a-independent pathways contribute to PSC senescence as well. With the following exceptions, the gene expression data obtained in this study are in agreement with the concept of Cdkn1a as a mediator of doxo effects in PSC: The knock-down of Cdkn1a had no effect on IL-6 mRNA expression and even strongly increased the expression of MMP-9; an effect that was also verified at the level of MMP-9 enzyme activity. Furthermore, cells treated with siRNA against Cdkn1a secreted less collagen type I. Since doxo treatment alone did not significantly alter PICP levels, the latter finding could
not be directly linked to the senescence process in this study and should be investigated further before drawing final conclusions. Here and in our previous experiments [12] we found that senescent PSC, like quiescent PSC, expressed lower levels of a-SMA protein than fully activated stellate cells. Another common feature of quiescent and senescent PSC is their low or absent proliferative activity. In contrast to quiescent PSC, senescent cells are, however, characterized by an increased size and a typical “flat” morphology [12]. Furthermore, they do also not contain significant amounts of Vitamin A-storing fat droplets (data not shown). Interestingly, the knock-down of Cdkn1a inhibited the doxo-dependent increase of SA b-Gal-positive cells but not the decrease of a-SMA protein expression. Together, these data suggest that some of the doxo effects in PSC are at least less sensitive to the downregulation of the Cdkn1a protein level than others. More specifically, they support the hypothesis of two primarily independent effects of doxo, induction of PSC senescence and inhibition of PSC activity (indicated by reduced levels of a-SMA), which may be separated from each other through the regulation of Cdkn1a protein levels. PSC have recently been shown to display characteristics of stem cells [22,23], and proposed to play a role not only in (pathological) fibrosis but also in normal tissue homeostasis and regeneration [23e25]. In this regard it is interesting to note that PSC cultures
B. Fitzner et al. / Pancreatology 13 (2013) 254e262
261
Fig. 7. Tet-regulated overexpression of Cdkn1a. LTC-Cdkn1a cells were cultured in the presence and absence of tet (1 mg/ml) and doxo (10 ng/ml) as indicated. (A) 48 h after tet withdrawal, expression of Cdkn1a (insert, upper panel) and b-actin (insert, lower panel) was analyzed by immunoblotting using specific antibodies. For quantitative evaluation, signal intensities were determined, and the ratio Cdkn1a/b-actin was calculated. A ratio of one hundred percent corresponds to cells cultured with tet. Data from n ¼ 12 independent samples were used to calculate mean values and SEM. *P < 0.05 versus controls cultured with tet. (B) Cell proliferation was assessed with the BrdU DNA-incorporation (inc) assay as described in the “Materials and methods” section. (C) On day 7 after tet withdrawal and doxo application, SA b-Gal-positive cells were determined based on a quantitative image evaluation and related to the total cell number. (B and C) Data from n 11 separate cultures were used to calculate mean values and SEM. *P < 0.05 versus controls cultured with tet/without doxo, #P < 0.05 versus control cells cultured with doxo (column 4 vs. 3).
never underwent complete senescence; neither in response to doxo (this study) nor to other triggers such as long-term culture and oxidant stress [12], suggesting a persistent self-renewal capacity. This interesting aspect of PSC biology should be further addressed in follow-up studies. Disclosure statement All authors declare to have no conflict of interest. Acknowledgments We gratefully acknowledge the excellent technical assistance of Mrs. Katja Bergmann and Mrs. Katrin Sievert-Küchenmeister. This work was supported by a grant from the Deutsche Forschungsgemeinschaft (to RJ). References [1] Erkan M, Adler G, Apte MV, Bachem MG, Buchholz M, Detlefsen S, et al. StellaTUM: current consensus and discussion on pancreatic stellate cell research. Gut 2012;61:172e8. [2] Jaster R, Emmrich J. Crucial role of fibrogenesis in pancreatic diseases. Best Pract Res Clin Gastroenterol 2008;22:17e29. [3] McCarroll JA, Phillips PA, Santucci N, Pirola RC, Wilson JS, Apte MV. Vitamin A inhibits pancreatic stellate cell activation: implications for treatment of pancreatic fibrosis. Gut 2006;55:79e89. [4] Jaster R, Hilgendorf I, Fitzner B, Brock P, Sparmann G, Emmrich J, et al. Regulation of pancreatic stellate cell function in vitro: biological and molecular effects of all-trans retinoic acid. Biochem Pharmacol 2003;66:633e41.
[5] Froeling FE, Feig C, Chelala C, Dobson R, Mein CE, Tuveson DA, et al. Retinoic acid-induced pancreatic stellate cell quiescence reduces paracrine Wnt-bcatenin signaling to slow tumor progression. Gastroenterology 2011;141: 1486e97. 1497.e1-14. [6] Masamune A, Kikuta K, Satoh M, Sakai Y, Satoh A, Shimosegawa T. Ligands of peroxisome proliferator-activated receptor-g block activation of pancreatic stellate cells. Biol Chem 2002;277:141e7. [7] Jaster R, Lichte P, Fitzner B, Brock P, Glass Ä, Karopka T, et al. Peroxisome proliferator-activated receptor g overexpression inhibits pro-fibrogenic activities of immortalized rat pancreatic stellate cells. J Cell Mol Med 2005;3: 670e82. [8] Klonowski-Stumpe H, Fischer R, Reinehr R, Lüthen R, Häussinger D. Apoptosis in activated rat pancreatic stellate cells. Am J Physiol Gastrointest Liver Physiol 2002;283:G819e26. [9] Rickmann M, Vaquero EC, Malagelada JR, Molero X. Tocotrienols induce apoptosis and autophagy in rat pancreatic stellate cells through the mitochondrial death pathway. Gastroenterology 2007;132:2518e32. [10] Manapov F, Muller P, Rychly J. Translocation of p21(Cip1/WAF1) from the nucleus to the cytoplasm correlates with pancreatic myofibroblast to fibroblast cell conversion. Gut 2005;54:814e22. [11] Krizhanovsky V, Yon M, Dickins RA, Hearn S, Simon J, Miething C, et al. Senescence of activated stellate cells limits liver fibrosis. Cell 2008;134:657e67. [12] Fitzner B, Müller S, Walther M, Fischer M, Engelmann R, Müller-Hilke B, et al. Senescence determines the fate of activated rat pancreatic stellate cells. J Cell Mol Med 2012;16:2620e30. [13] Krizhanovsky V, Xue W, Zender L, Yon M, Hernando E, Lowe SW, et al. Implications of cellular senescence in tissue damage response, tumor suppression, and stem cell biology. Cold Spring Harb Symp Quant Biol 2008;73: 513e22. [14] Campisi J, d’Adda di Fagagna F. Cellular senescence: when bad things happen to good cells. Nat Rev Mol Cell Biol 2007;8:729e40. [15] Jaster R, Sparmann G, Emmrich J, Liebe S. Extracellular signal-regulated kinases are key mediators of mitogenic signals in rat pancreatic stellate cells. Gut 2002;51:579e84. [16] Rateitschak K, Karger A, Fitzner B, Lange F, Wolkenhauer O, Jaster R. Mathematical modelling of interferon-g signalling in pancreatic stellate cells
262
[17]
[18]
[19] [20]
B. Fitzner et al. / Pancreatology 13 (2013) 254e262 reflects and predicts dynamics of STAT1 pathway activity. Cell Signal 2010;22: 97e105. Luo S, Wehr NB, Levine RL. Quantitation of protein on gels and blots by infrared fluorescence of Coomassie blue and Fast Green. Anal Biochem 2006;350:233e8. Phalke S, Mzoughi S, Bezzi M, Jennifer N, Mok WC, Low DH, et al. p53Independent regulation of p21Waf1/Cip1 expression and senescence by PRMT6. Nucleic Acids Res 2012;40:9534e42. el-Deiry WS, Tokino T, Velculescu VE, Levy DB, Parsons R, Trent JM, et al. WAF1, a potential mediator of p53 tumor suppression. Cell 1993;75:817e25. Dulic V, Kaufmann WK, Wilson SJ, Tlsty TD, Lees E, Harper JW, et al. p53dependent inhibition of cyclin-dependent kinase activities in human fibroblasts during radiation-induced G1 arrest. Cell 1994;76:1013e23.
[21] Romanov VS, Pospelov VA, Pospelova TV. Cyclin-dependent kinase inhibitor p21(Waf1): contemporary view on its role in senescence and oncogenesis. Biochemistry (Mosc) 2012;77:575e84. [22] Mato E, Lucas M, Petriz J, Gomis R, Novials A. Identification of a pancreatic stellate cell population with properties of progenitor cells: new role for stellate cells in the pancreas. Biochem J 2009;421:181e91. [23] Kordes C, Sawitza I, Götze S, Häussinger D. Stellate cells from rat pancreas are stem cells and can contribute to liver regeneration. PLoS One 2012;7:e51878. [24] Means AL. Pancreatic stellate cells: small cells with a big role in tissue homeostasis. Lab Invest 2013;93:4e7. [25] Riopel MM, Li J, Liu S, Leask A, Wang R. b1 integrin-extracellular matrix interactions are essential for maintaining exocrine pancreas architecture and function. Lab Invest 2013;93:31e40.