Radiotherapy and Oncology xxx (2014) xxx–xxx
Contents lists available at ScienceDirect
Radiotherapy and Oncology journal homepage: www.thegreenjournal.com
Original article
Contextual regulation of pancreatic cancer stem cell phenotype and radioresistance by pancreatic stellate cells Osama Al-Assar 1, Fevzi Demiciorglu, Serena Lunardi, Maria Manuela Gaspar-Carvalho, William Gillies McKenna, Ruth M. Muschel, Thomas B. Brunner ⇑ The Radiobiology Research Institute, MRC/CR-UK Gray Institute for Radiation Oncology and Biology, Department of Oncology, University of Oxford, Churchill Hospital, UK
a r t i c l e
i n f o
Article history: Received 16 August 2013 Received in revised form 12 February 2014 Accepted 18 March 2014 Available online xxxx Keywords: Pancreatic cancer Cancer stem-like cells Radioresistance Epithelial mesenchymal transition Pancreatic stellate cells
a b s t r a c t Background and purpose: Progression of pancreatic ductal adenocarcinoma (PDAC) is promoted by desmoplasia induced by pancreatic stellate cells (PSC). Contributory to this progression is epithelial mesenchymal transition (EMT), which shares many characteristics with the cancer stem cell (CSC) hypothesis. We investigated the role of these processes on the radioresponse and tumorigenicity of pancreatic cancer cells. Materials and methods: We used an in vitro sphere model and in vivo xenograft model to examine the role of PSC in EMT and CSC processes. Results: We demonstrated that PSC enhanced the CSC phenotype and radioresistance of pancreatic cancer cells. Furthermore, the expression of several EMT and CSC markers supported enhanced processes in our models and that translated into remarkable in vivo tumorigenicity. Multi-dose TGFb neutralizing antibody inhibited the EMT and CSC processes, sensitized cells to radiation and reduced in vivo tumorigenicity. A proteomic screen identified multiple novel factors that were regulated by PSC in pancreatic cells. Conclusion: These results are critical in highlighting the role of PSC in tumor progression and radioresistance by manipulating the EMT and CSC processes. TGFb and the novel factors identified are important targets for better therapeutic outcome in response to PSC mediated mechanisms. Ó 2014 Elsevier Ireland Ltd. All rights reserved. Radiotherapy and Oncology xxx (2014) xxx–xxx
Pancreatic ductal adenocarcinoma is an aggressive and highly lethal disease, with less than 1% 5-year survival. The poor prognosis of PDAC is attributed in part to the presence of a strong desmoplastic reaction driven by pancreatic stellate cells (PSC), which are the main source of excessive extracellular matrix production in chronic pancreatitis and pancreatic adenocarcinoma [3]. The presence of PSC is advantageous for tumor growth in vivo [14]. A subpopulation of cells called cancer stem cells (CSC) has been isolated from PDAC [15]. CSCs are thought to be a determining factor in chemo and radioresistance [7]. However, the model used is important in determining the outcome [1]. A link between CSC and epithelial mesenchymal transition (EMT) has been demonstrated [19]. Inducers of EMT include transforming growth factor b (TGFb) [9,26], which is also involved in PSC activation [18]. Activated PSC drive the desmoplastic reaction characteristic of ⇑ Corresponding author. Current address: University Hospitals of Freiburg, Department of Radiation Oncology, Robert-Koch-Str. 3, D-79106 Freiburg im Breisgau, Germany. E-mail address: [email protected] (T.B. Brunner). 1 Current address: MRC Unit of Molecular Haematology, The Weatherall Institute of Molecular Medicine, University of Oxford, John Radcliffe Hospital, Oxford OX3 9DS, UK.
pancreatic cancer resulting in extensive collagen types I, III and fibronectin production [24]. Furthermore, in vivo models of pancreatic cancer show an EMT defect that is attributed to deletion in TGFb signaling including DPC4/smad4 and transcriptional intermediary factor (TIF)-1c [31]. In this study, we investigated the role of PSC in the induction of CSC and EMT and the effect on the radioresponse of pancreatic cancer cells in vitro and in vivo. An improved understanding of the role of stellate cells in the contextual induction of stem cell phenotype is critical in the development of more effective treatments. Materials and methods Complete description of Materials and methods is provided in the Supplementary Materials due to space limitation. Cell culture The human pancreatic cancer cell lines PSN-1, Panc-1 and MiaPaCa-2 were originally obtained from the American Type Tissue Collection. The human pancreatic stellate cell line (hPSC) was kindly provided by Dr. Atsushi Masamune [20]. The LTC-14 cells
http://dx.doi.org/10.1016/j.radonc.2014.03.014 0167-8140/Ó 2014 Elsevier Ireland Ltd. All rights reserved.
Please cite this article in press as: Al-Assar O et al. Contextual regulation of pancreatic cancer stem cell phenotype and radioresistance by pancreatic stellate cells. Radiother Oncol (2014), http://dx.doi.org/10.1016/j.radonc.2014.03.014
2
Pancreatic stellate cells induce cancer stem cells
used in the supplementary figures were rat stellate cells [30]. The different cell lines were cultured in the recommended medium as suggested by the supplier. Indirect co-culture of pancreatic cancer cells-pancreatic stellate cells and sphere formation Human PSC were seeded in chemically-defined medium and single pancreatic cancer cell suspensions were resuspended in a mixture of Matrigel (BD Biosciences) and serum free medium, and plated into culture inserts of 1.0 lm pore size (BD Biosciences). After 10 days of incubation, spheres were counted and measured for size. Spheres were then retrieved using Matrigel Recovery Solution (BD Biosciences) according to the manufacturer’s instructions. Sodium dodecyl sulfate–polyacrylamide gel electrophoresis and Western blotting Spheres protein extracts were resolved on a NuPAGE 4–12% Bis–Tris mini gel (Invitrogen) and transferred onto a Hybond-C Extra membrane (Amersham BioSciences). Membranes were cut in several sections and incubated with several antibodies for quantitative analysis of the different samples on the same membrane using chemiluminescence (ECL, Pierce). Supplementary Table 1 summarizes all antibody concentrations used for the corresponding technique. ImageJ was used to evaluate the densitometric differences employing integrated intensity.
tumors were fixed in a 4% neutral solution of formaldehyde and cut into 10 lm sections. The antibodies used and their working concentrations are described in Supplementary Table 1. Sections used for immunohistochemistry were counterstained with hematoxylin and viewed using a standard microscope. Sections used for immunofluorescence were viewed under either an epifluorescent or confocal microscope. ImageJ was used to evaluate the average pixel density differences in whole images. Flow cytometry Panc-1 spheres grown with or without human PSC were mechanically dissociated and washed twice with ice-cold phosphate-buffered saline (PBS). Then they were incubated with the non-enzymatic Cell Recovery solution (BD Biosciences) for up to 1 h at 4 °C on a rotary shaker. Cells obtained from mechanically dissociated spheres were stained with the corresponding antibodies, as outlined in the Supplementary Table 1, and analyzed using BD FACSCalibur machine. Enzyme-linked immunosorbent assay (ELISA) and neutralizing TGFb antibody treatment Quantikine Human TGFb1 Immunoassay (R&D Systems, Inc.) was performed as described by the manufacturer. E-Cadherin ELISA was assayed as described by the manufacturer (MesoScale). In vivo tumorigenicity and TCD50 assay
Immunohistochemical and immunofluorescence staining of formalinfixed frozen tissues All animal procedures were carried out in accordance with current U.K. legislation under an approved project license. Excised
All animal procedures were carried out in accordance with the current UK legislation under an approved project license. For the in vivo tumorigenicity assay, spheres were disaggregated, injected and followed up to 6 months. For the TCD50 experiments, single
A
B
Fig. 1. The effect of human pancreatic stellate cells (PSC) on the sphere forming ability of three pancreatic cancer cell lines. Panel A shows representative pictures of spheres established from pancreatic cancer cell lines co-cultured with or without human PSC, 7–12 days after plating into Matrigel. Panel B is a numerical representation for the pictures in panel A as assessed for both size and number. The asterisks represent statistical significance for differences using the Student’s t-test (a = 0.05). The error bars represent SE of different wells from three different plates.
Please cite this article in press as: Al-Assar O et al. Contextual regulation of pancreatic cancer stem cell phenotype and radioresistance by pancreatic stellate cells. Radiother Oncol (2014), http://dx.doi.org/10.1016/j.radonc.2014.03.014
O. Al-Assar et al. / Radiotherapy and Oncology xxx (2014) xxx–xxx
dose c-irradiation (14, 20, 30 or 40 Gy) was delivered when tumor volumes reached 60–80 mm3 and volumes were calculated using the formula p/6 length width height. The tumor control probability (TCP) was modeled using the logistic model,
TCP ¼
3
ability in two pancreatic cancer cell lines strongly supporting a role for a secreted factor mediating this effect (Supplementary Fig. 1C). Co-culture conditions resulted in an increase in the expression of different stem cell markers (CD24, CD44 and CD326) and down regulation of E-cadherin indicative of EMT induction in Panc-1 spheres (Supplementary Fig. 2A and B).
1 ð1 þ ðD50=DÞrÞ
Results
Pancreatic stellate cells enhance radioresistance in vitro and in vivo Clonogenic survival after radiotherapy is a direct test of stem cell eradication [7]. Fig. 2A shows how hPSC enhance the clonogenic survival of disaggregated PSN-1 and Panc-1 spheres grown in serum free medium. In addition, whole PSN-1 and Panc-1 spheres co-cultured with hPSC were more radioresistant than spheres grown without PSC without disaggregation (Fig. 2B). In line with these observations, we also observed enhanced radioresistance of SW1222 colorectal cancer cells when co-cultured with tumor associated myofibroblasts isolated from patients (Supplementary Fig. 3). As a single surviving CSC can induce recurrent growth in vivo, we used the tumor control probability 50% assay (TCD50). Xenograft tumors grown from cancer cells co-injected with hPSC required a significantly higher dose to achieve the same level of local control compared with tumors that formed after injection of PSN-1 cells only (Fig. 2C).
Co-culture with human pancreatic stellate cells enhances the cancer stem cell phenotype in 3D serum free conditions Sphere formation, as a functional approach, is particularly useful to enrich CSC subpopulations [7]. In order to evaluate the possible effects of hPSC on stemness, three pancreatic cancer cell lines (PSN-1, MiaPaCa-2, and Panc-1), were indirectly co-cultured with PSC in three-dimensional serum free conditions and the sphereforming ability was compared. The presence of PSC enabled all cell lines to form more and larger spheres (Fig. 1). The effect was not limited to the hPSC cell line but confirmed with the rat LTC-14 pancreatic stellate cell line (Supplementary Fig. 1A). Importantly, this effect was specific to stellate cells and not observed with normal fibroblasts (Supplementary Fig. 1B). The addition of conditioned medium from hPSC had a strong effect on the sphere forming
A
B
C
Fig. 2. The effect of pancreatic stellate cells (PSC) on the radioresponse of pancreatic cancer cells in vitro and in vivo. Panel A shows the clonogenic survival of disaggregated pancreatic cancer spheres from two cell lines after co-culturing with or without stellate cells. Spheres were pooled from different wells and different plates and were cultured in a chemically defined medium. Panel B represents the radioresponse of Panc-1 spheres after irradiation in situ with or without stellate cells. The spheres were assessed for both size and number changes. The error bars represent SE of the mean of several wells from different plates. The presence of stellate cells gave an ANOVA p-value of 0.04 and 0.073 after the different doses of radiation for Panc-1 and PSN-1 sphere size, respectively. The ANOVA p-values were 0.05 and 0.10 for the average sphere number/filed after the different doses of radiation for the Panc-1 and PSN-1 cells, respectively. Panel C is a tumor control probability 50% (TCD50) assay of PSN-1 xenograft tumors established from injection with cancer cells or cancer and stellate cells. Logistic and probability analyses give almost identical tumor control probability.
Please cite this article in press as: Al-Assar O et al. Contextual regulation of pancreatic cancer stem cell phenotype and radioresistance by pancreatic stellate cells. Radiother Oncol (2014), http://dx.doi.org/10.1016/j.radonc.2014.03.014
4
Pancreatic stellate cells induce cancer stem cells
Pancreatic stellate cells promote epithelial mesenchymal transition in vitro and in vivo Western blotting showed increased expression of vimentin, a mesenchymal marker, and b-catenin, in PSN-1 and Panc-1 spheres co-cultured with PSC (Fig. 3A and B). Additionally, E-cadherin, an epithelial marker, was down-regulated in the presence of PSC, which was also verified using an ELISA for E-cadherin (Fig. 3C). We next performed co-immunofluorescence staining of PSN-1 (Fig. 4A) and Panc-1 (Fig. 4B) xenograft sections from tumors formed after co-injection with and without PSC to verify our findings in vivo. In tumor cells, vimentin expression was usually distributed diffusely in the cytoplasm, although perinuclear staining was seen in some cells as well. Positive staining for vimentin was also detected in stromal cells as expected (Figs. 4A and B). Immunohistochemistry re-confirmed the expression pattern of E-cadherin and vimentin in immunofluorescence in addition to b-catenin in PSN-1 xenograft sections (Supplementary Fig. 4A). Nuclear localization and expression of b-catenin was dramatically enhanced in the sections that had both PSN-1 and stellate cells co-injections (Supplementary Fig. 4A). This was also seen in Panc-1 cells co-injected with PSC (data not shown). Immunofluorescence of other EMT markers including N-cadherin and slug also supported the original findings in the PSN-1 xenograft sections (Supplementary Fig. 4B). Pancreatic stellate cells induce stemness as defined by cell surface markers The amount of CSC has recently been suggested to be influenced by the stromal compartment [29]. We aimed to assess the influence of PSC on the representation of CSC within pancreatic tumors
A
by performing immunofluorescence staining for cell surface markers, namely CD24 and CD326 (EpCAM). Immunofluorescence analysis of PSN-1 and Panc-1 ± PSC xenograft sections demonstrated that the co-expression of two cell surface markers in cancer cells was significantly induced by PSC (Fig. 4B). Confocal immunofluorescence was performed to assess the overlap between EMTinduced cells (increased vimentin expression/loss of E-cadherin) and CSC, defined by co-expression of CD24 and CD326. A strong but incomplete overlap between the EMT and CSC markers was observed (Fig. 4C).
The presence of stellate cells enhances the in vivo tumorigenicity of disaggregated pancreatic cancer spheres To assess the effect of stellate cells on the induction of stemness in pancreatic cancer cells, we used an in vivo tumorigenicity assay of limiting dilutions. We found that spheres co-cultured with stellate cells had a strong enhancing effect on tumorigenicity (Fig. 5). Cell numbers as low as 10 cells disaggregated from spheres co-cultured with hPSC induced tumors in all of the animals compared with none of the animals in the control group for the same cell number. The growth kinetics of both groups was similar after injection of 10,000 and 100 cells, however, a prolonged delay in growth onset was observed in the disaggregated spheres grown without PSC for the lower cell numbers. The growth kinetics of the two groups was different only at 100 cells (Student’s t-test p-value 0.025). To confirm the sphere culture system as a stem cell model we additionally found that culturing cells as spheres enhances in vivo tumorigenicity compared with the injection of cells from monolayers as shown in Fig. 5C.
B
Fig. 3. Expression of different EMT markers in spheres co-cultured with or without human pancreatic stellate cells (PSC). Panel A shows the expression of vimentin, b-catenin, and Slug in both PSN-1 and Panc-1 spheres co-cultured with PSC compared with mono-cultured cells. The molecular weight markers are shown to the left of the panel. The minus and plus signs represent co-cultures without or with PSC, respectively. Panel B represents the ratio of densitometric values normalized to GAPDH and divided over the value without PSC (dotted line). Panel C shows an ELISA of E-cadherin expression in PSN-1 and Panc-1 spheres co-cultured with or without PSC. The error bars are SE of multiple experiments.
Please cite this article in press as: Al-Assar O et al. Contextual regulation of pancreatic cancer stem cell phenotype and radioresistance by pancreatic stellate cells. Radiother Oncol (2014), http://dx.doi.org/10.1016/j.radonc.2014.03.014
O. Al-Assar et al. / Radiotherapy and Oncology xxx (2014) xxx–xxx
A
5
B
C
Fig. 4. Representative immunofluorescence staining of xenograft sections of tumors grown from PSN-1 or Panc-1 pancreatic cancer co-injected with or without pancreatic stellate cells (PSC). The sections were co-stained for EMT markers (E-Cad and vimentin) and cancer stem cell markers (CD24 and CD326). The IgG negative control for both panels is shown for each panel. Panels A and B are immunofluorescence images of staining for EMT and cancer stem cell markers, consecutively. Panel C represents confocal microscopy images of both EMT and cancer stem cell markers in PSN-1 xenograft tumors. The graphs in panel C represent the overlap of fluorescence signals for the different markers along the diagonal line drawn in the DAPI panel. The same pattern was observed in xenograft sections from different animals for both cell lines. The scale bars represent 200 lm.
Attenuation of TGFb modulates pancreatic cancer cells stemness as defined by the sphere-forming ability and reverses the profound effect of pancreatic stellate cells The release of TGFb in the co-culture medium was increased in the presence of stellate cells in the three pancreatic cancer cell lines used (Supplementary Fig. 5A). The expression of TGFb was also enhanced in xenograft tumors in the presence of stellate cells in vivo and was considerably weaker in cancer cells in sections without the co-injected stellate cells (Supplementary Fig. 5B). Attenuating TGFb by adding multiple doses of a neutralizing antibody had a strong effect on the sphere forming ability of different pancreatic cancer cell lines in monoculture and co-culture conditions (Fig. 6A). Concurrent with a reduction in CSC sphere forming ability, there was a down regulation of EMT markers and the TGFb non-canonical p-MAPK signaling since these cells harbor SMAD4 mutation (Fig. 6B). In addition, multi-dose neutralizing TGFb radio-sensitized disaggregated spheres in a classical clonogenic assay (Fig. 6C). Under the same conditions, there was a marked reduction of tumorigenicity in vivo (Fig. 6D); nonetheless, the growth kinetics was similar for all groups when tumors occurred. A single dose of neutralizing TGFb antibody had a restricted effect on the sphere forming ability (Supplementary Fig. 5C).
Discussion We here report that PSC enhance stemness in pancreatic cancer spheres and xenograft tumors, how this observation is linked with a rise in EMT features and how this impacts therapeutically on cancer stem cell survival after radiotherapy. Radiotherapy plays an important role in the multimodal treatment of pancreatic cancer but long-term local control with radiotherapy only is not achievable and this is a typical example for persistent CSC due to insufficient dose [6]. These findings are underscored by enhanced tumorigenicity in SCID mice with PSC and by the enhanced radiation dose required to locally control xenograft tumors containing PSC. We hypothesized a role for PSC in the induction of EMT and CSC in PDAC with a direct effect on treatment. To test this hypothesis we used three dimensional (3D) sphere formation in vitro and xenograft in vivo models. The prominent feature of sphere formation assays is that each sphere is accepted to derive from a single cell and is therefore clonal. Since self-renewal and differentiation can be theoretically investigated at the single-cell level, sphereforming assays are commonly used in stem cell biology [25]. We found that co-culturing of spheres with PSC results in an increase of both size and number of spheres in three different pancreatic cancer cell lines, and the differences were statistically significant (Fig. 1A and B). We used the recommended cell density
Please cite this article in press as: Al-Assar O et al. Contextual regulation of pancreatic cancer stem cell phenotype and radioresistance by pancreatic stellate cells. Radiother Oncol (2014), http://dx.doi.org/10.1016/j.radonc.2014.03.014
6
Pancreatic stellate cells induce cancer stem cells
A
B
C
Fig. 5. In vivo tumorigenicity assay of PSN-1 spheres co-cultured with or without human pancreatic stellate cells (PSC) in SCID mice. Panel A is a Kaplan–Meier graph showing tumor formation-1 of disaggregated spheres that were co-cultured with or without human PSC. The inset table and values represent incidence summary and the statistical significance of the difference between the tumor forming ability of the two groups after six-months monitoring. Panel B is the growth kinetics of the different groups in panel A stratified according to the number of cells injected. Panel C is a table representation of the in vivo tumorigenicity of disaggregated spheres after co-culture with or without PSC versus monolayer cells in SCID mice up to 9 months after injection.
from other studies to maximize clonality and the cells were embedded into Matrigel to eliminate sphere aggregation [8]. We used time lapse imaging (data not shown) to eliminate the possibility of aggregation in our studies and Supplementary Fig. 6 shows fluorescent dye retention in spheres supporting clonal expansion of spheres. Our 3D culture conditions resulted in enhanced combined expression of the three cancer stem cell markers (CD24, CD44 and CD326) when co-cultured with stellate cells (Supplementary Fig. 2A and B). These markers were shown to be cancer stem cell markers that resulted in increased tumorigenicity in vivo [15]. Other groups showed that culturing cells under 3D conditions enriches stem cell markers [12,17]. Furthermore, the same spheres in our 3D-model showed regulation of EMT markers, including E-cadherin, using fluorescence imaging, flow cytometry and western blotting (Figs. 2 and 4). The expression of E-cadherin in pancreatic cancer was cell line dependent in a previous study [16]. Both MiaPaCa-2 and Panc-1 cell lines showed down regulation of E-cadherin and strong expression of transcriptional activators of mesenchymal markers [16]. We confirmed our in vitro results using an in vivo xenograft model of two pancreatic cancer cell lines (Fig. 4, and Supplementary Fig. 4). The argument for an orthotopic model to fully assess tumor behavior [32] would make it difficult if not impossible to assess the role of PSC since they are already present in the pancreas and can become activated. The presence of PSC in our in vivo experiments resulted in b-catenin expression being predominantly nuclear in tumor cells adjacent
to the stromal compartment unlike cells in the tumor center, characteristic of EMT induction. Also characteristic of EMT induction is cadherin switching with a profound effect on cell type and behavior [33], which we observed for E and N-cadherins in our xenograft sections (Fig. 4 and Supplementary Fig. 4). Interestingly, the congruence between the EMT and CSC markers was incomplete in our experiments (Fig. 4C), which has been reported and requires further examination [19]. The up-regulation of the CSC phenotype and EMT in the context of PSC presence had remarkable translational in vitro (Fig. 2A and B) and in vivo (Fig. 2C) radiation effects. A number of other groups showed that transcriptional regulators of EMT are advantageous for the survival of cells undergoing EMT [10,13,27]. In addition, EMT induction in both MiaPaCa-2 and Panc-1 resulted in a more resistant phenotype to chemotherapeutic drugs [4]. To prove that the in vitro assay results and CSC marker profiles up-regulation in response to PSC had real functional significance on stemness in vivo, we used an in vivo tumorigenicity assay. There was a great advantage for PSC co-culture in vitro on the development of tumors in vivo (Fig. 5) and was in total agreement with the TCD50 results (Fig. 2C), strongly supporting the enrichment of CSC fraction in our model. The evidence for the correlation between tumorigenicity as a result of CSC serial dilution and TCD50 was previously shown in 13 different tumor experimental models (reviewed in [6]). The differences in tumorigenicity were not matched by changes in the growth kinetics of tumors arising from spheres co-cultured with or without PSC, except for tumors
Please cite this article in press as: Al-Assar O et al. Contextual regulation of pancreatic cancer stem cell phenotype and radioresistance by pancreatic stellate cells. Radiother Oncol (2014), http://dx.doi.org/10.1016/j.radonc.2014.03.014
O. Al-Assar et al. / Radiotherapy and Oncology xxx (2014) xxx–xxx
A
C
7
B
D
Fig. 6. The effect of TGFb attenuation on the stem cell phenotype, EMT marker expression and in vivo tumorigenicity. Panel A shows the effect of multiple doses of neutralizing TGF-b antibody on the sphere forming ability of different pancreatic cancer cell lines co-cultured indirectly with human pancreatic stellate cells (PSC). All results were statistically significant using the Student’s t-test (a = 0.05) for the effect of TGFb on sphere size and number. Fresh antibody was added for five days and the spheres were analyzed on day 8. The error bars represent SE of different wells from three different plates. Panel B shows the expression of different EMT markers using western blotting in the co-culture conditions and after the addition of TGFb neutralizing antibody. Panel C represents a clonogenic assay of TGFb pre-treated spheres that were disaggregated and subsequently treated with different radiation doses. The error bars represent SE of the mean of several wells from different plates. Panel D is a Kaplan– Meier graph showing in vivo tumorigenicity assay of PSN-1 disaggregated spheres co-cultured with or without human PSC in female SCID mice after a 5-day treatment with neutralizing TGFb antibody in vitro. The results are statistically significant as shown in the inset values. Also shown in panel D is the growth kinetics of the different tumor groups. A total cell number of 1000 was injected for each group of 5 mice and the experiment was allowed to run for 60 days and the animals were terminated when they reached 100 mm3.
originating from injections with 30 cells. This implies the contextual regulation of CSC percentage and the sphere conditions enriched for stemness compared with the 2D monolayer culture (Fig. 5C). TGFb is a factor with paradox roles in tumor suppression and promotion (Fig. 7) [5]. Inhibition of TGFb signaling in PSN-1, Panc-1 and MiaPaCa-2 cell lines, using multiple doses of neutralizing TGFb antibody, resulted in the suppression of the CSC 3D sphere formation phenotype and down regulation of EMT markers in our sphere forming assay (Fig. 6A and B). In addition, multipledose neutralizing TGFb antibody had a radio-sensitizing effect on disaggregated spheres co-cultured with or without human PSC in Panc-1 and PSN-1 pancreatic cancer cell lines (Fig. 6C) and reduced in vivo tumorigenicity in PSC-1 cell (Fig. 6D). Unlike Panc-1 and MiaPaCa-2, the inhibition of the CSC and EMT markers in the PSN-1 cell line is DPC4/Smad4 independent due to the homologous deletion of this gene [22]. In this context, we showed inhibition of non-canonical TGFb signaling through pMAPK in the spheres, by western blotting (Fig. 6B). Our results are in agreement with published data showing a role for TGFb in the self-renewal and maintenance of stem cells in breast cancer [19,23] and glioma samples [2]. However, other
groups showed that other members of the TGFb superfamily were key players in the sphere forming ability of primary pancreatic cancer cells, but not TGFb [17]. These discrepancies might be due to the presence of a cross-talk between TGFb, WNT, Notch and Sonic hedgehog in EMT and stem cell induction [28]. We identified several targets that have the potential of being very valuable in understanding the mechanism of CSC phenotype and EMT induction, consequently leading to better tumor control (Supplementary Fig. 7). One of the factors identified in our screen is CXCL8 (Supplementary Fig. 7B), which has been shown to be involved in stromal signaling in pancreatic cancer [21] and in breast cancer stem cell maintenance [11]. In summary, our results show that PSC enhance both CSC phenotype and EMT process in vitro and in vivo. Diagram 1 highlights some of the targets we investigated in this study, including the interplay of these targets with TGFb. Our results also emphasize the importance of contextual stromal factors in affecting the CSC representation in tumors, leading to direct translational effects on tumorigenesis and therapeutic outcome. In addition, continued investigation is required to verify our potential therapeutic targets in vivo, and ultimately their clinical use to improve therapy.
Please cite this article in press as: Al-Assar O et al. Contextual regulation of pancreatic cancer stem cell phenotype and radioresistance by pancreatic stellate cells. Radiother Oncol (2014), http://dx.doi.org/10.1016/j.radonc.2014.03.014
8
Pancreatic stellate cells induce cancer stem cells
Tumor supressing
Tumor promong TGFβ Levels Wnt
?
EMT
KRAS
TP53 ink4A
MET
MAPK SMAD4
?
Vimenn
TP53 ink4A c-Myc
? Proteome Array
FGF
MAPK SMAD4
Β-Catenin FAK
?
Cancer stem cell phenotype
N-Cadherin CXCR4
Connecve Tissue Growth Factor
Integrins
Immune cells
? Acvaon
?
KRAS
GSK3
E-Cadherin
Frizzled
Stromal modificaon
Endothelial cells
PSC PSC: Pancreac stellate cells EMT: Epithelial mesenchymal transion MET: Mesenchymal epithelial transion Fig. 7. The diagram depicts the role of pancreatic stellate cells (PSCs) in transforming growth factor b (TGFb) signaling pathways in the context of pancreatic cancer cells (PCCs). PCCs are shown in the top half of the diagram and star-shaped PSCs in the bottom half and endothelial as well as immune cells are other cell types of the tumor stroma. Tumor suppressing effects of TGFb are lost with rising levels of the growth factor and converted to tumor promoting effects on PCCs and with epithelial– mesenchymal transition (EMT). EMT is reversible and then called MET. TGFb is produced by both PCCs and PSCs and TGFb downstream signaling in PCCs is often defective (e.g. DPC-4 mutation) and thereby contributes to paracrine signaling. PCCs that have undergone EMT will express a number of factors such as N-cadherin, vimentin and bcatenin. At the same time the expression of cancer stem cell markers is enhanced.
Financial support Medical Research Council (UK) and Cancer Research, UK. Conflict of interest None declared. Acknowledgements We acknowledge the Medical Research Council (UK) and Cancer Research UK for their core funding. We would like to thank Emmanouil Fokas, Karla Watson and Magdalena Flieger for their help with the animal experiments. We would also like to thank John Fenwick for performing the probability analysis for the TCD50 assay, and Prof. Sir Walter Bodmer’s laboratory for the SW1222 cell line and Neil Ashley for supplying us with the primary colorectal myofibroblasts. Last, we thank Helio Pais for bioinformatics assistance with the proteome analysis. Appendix A. Supplementary data Supplementary data associated with this article can be found, in the online version, at http://dx.doi.org/10.1016/j.radonc.2014. 03.014. References [1] Al-Assar O, Muschel RJ, Mantoni TS, McKenna WG, Brunner TB. Radiation response of cancer stem-like cells from established human cell lines after sorting for surface markers. Int J Radiat Oncol Biol Phys 2009;75:1216–25. [2] Anido J, Saez-Borderias A, Gonzalez-Junca A, et al. TGF-beta receptor inhibitors target the CD44(high)/Id1(high) glioma-initiating cell population in human glioblastoma. Cancer Cell 2010;18:655–68.
[3] Apte MV, Park S, Phillips PA, et al. Desmoplastic reaction in pancreatic cancer: role of pancreatic stellate cells. Pancreas 2004;29:179–87. [4] Arumugam T, Ramachandran V, Fournier KF, et al. Epithelial to mesenchymal transition contributes to drug resistance in pancreatic cancer. Cancer Res 2009;69:5820–8. [5] Bachman KE, Park BH. Duel nature of TGF-beta signaling: tumor suppressor vs. tumor promoter. Curr Opin Oncol 2005;17:49–54. [6] Baumann M, Krause M, Hill R. Exploring the role of cancer stem cells in radioresistance. Nat Rev Cancer 2008;8:545–54. [7] Brunner TB, Kunz-Schughart LA, Grosse-Gehling P, Baumann M. Cancer stem cells as a predictive factor in radiotherapy. Semin Radiat Oncol 2012;22:151–74. [8] Chojnacki A, Weiss S. Production of neurons, astrocytes and oligodendrocytes from mammalian CNS stem cells. Nat Protoc 2008;3:935–40. [9] Ellenrieder V, Hendler SF, Ruhland C, Boeck W, Adler G, Gress TM. TGF-betainduced invasiveness of pancreatic cancer cells is mediated by matrix metalloproteinase-2 and the urokinase plasminogen activator system. Int J Cancer 2001;93:204–11. [10] Escriva M, Peiro S, Herranz N, et al. Repression of PTEN phosphatase by Snail1 transcriptional factor during gamma radiation-induced apoptosis. Mol Cell Biol 2008;28:1528–40. [11] Ginestier C, Liu S, Diebel ME, et al. CXCR1 blockade selectively targets human breast cancer stem cells in vitro and in xenografts. J Clin Invest 2010;120:485–97. [12] Gou S, Liu T, Wang C, et al. Establishment of clonal colony-forming assay for propagation of pancreatic cancer cells with stem cell properties. Pancreas 2007;34:429–35. [13] Haslehurst AM, Koti M, Dharsee M, et al. EMT transcription factors snail and slug directly contribute to cisplatin resistance in ovarian cancer. BMC Cancer 2012;12:91. [14] Hwang RF, Moore T, Arumugam T, et al. Cancer-associated stromal fibroblasts promote pancreatic tumor progression. Cancer Res 2008;68:918–26. [15] Li C, Heidt DG, Dalerba P, et al. Identification of pancreatic cancer stem cells. Cancer Res 2007;67:1030–7. [16] Li C, Lee CJ, Simeone DM. Identification of human pancreatic cancer stem cells. Methods Mol Biol 2009;568:161–73. [17] Lonardo E, Hermann PC, Mueller MT, et al. Nodal/Activin signaling drives selfrenewal and tumorigenicity of pancreatic cancer stem cells and provides a target for combined drug therapy. Cell Stem Cell 2011;9:433–46. [18] Von Mahadevan D, Hoff DD. Tumor-stroma interactions in pancreatic ductal adenocarcinoma. Mol Cancer Ther 2007;6:1186–97. [19] Mani SA, Guo W, Liao MJ, et al. The epithelial–mesenchymal transition generates cells with properties of stem cells. Cell 2008;133:704–15.
Please cite this article in press as: Al-Assar O et al. Contextual regulation of pancreatic cancer stem cell phenotype and radioresistance by pancreatic stellate cells. Radiother Oncol (2014), http://dx.doi.org/10.1016/j.radonc.2014.03.014
O. Al-Assar et al. / Radiotherapy and Oncology xxx (2014) xxx–xxx [20] Masamune A, Watanabe T, Kikuta K, Satoh K, Shimosegawa T. NADPH oxidase plays a crucial role in the activation of pancreatic stellate cells. Am J Physiol 2008;294:G99–G108. [21] Matsuo Y, Raimondo M, Woodward TA, et al. CXC-chemokine/CXCR2 biological axis promotes angiogenesis in vitro and in vivo in pancreatic cancer. Int J Cancer 2009;125:1027–37. [22] Moore PS, Sipos B, Orlandini S, et al. Genetic profile of 22 pancreatic carcinoma cell lines. Analysis of K-ras, p53, p16 and DPC4/Smad4. Virchows Arch 2001;439:798–802. [23] Morel AP, Lievre M, Thomas C, Hinkal G, Ansieau S, Puisieux A. Generation of breast cancer stem cells through epithelial–mesenchymal transition. PLoS One 2008;3:e2888. [24] Omary MB, Lugea A, Lowe AW, Pandol SJ. The pancreatic stellate cell: a star on the rise in pancreatic diseases. J Clin Invest 2007;117:50–9. [25] Pastrana E, Silva-Vargas V, Doetsch F. Eyes wide open: a critical review of sphere-formation as an assay for stem cells. Cell Stem Cell 2011;8:486–98. [26] Peinado H, Quintanilla M, Cano A. Transforming growth factor beta-1 induces snail transcription factor in epithelial cell lines: mechanisms for epithelial mesenchymal transitions. J Biol Chem 2003;278:21113–23.
9
[27] Robson EJ, Khaled WT, Abell K, Watson CJ. Epithelial-to-mesenchymal transition confers resistance to apoptosis in three murine mammary epithelial cell lines. Differentiation 2006;74:254–64. [28] Scheel C, Eaton EN, Li SH, et al. Paracrine and autocrine signals induce and maintain mesenchymal and stem cell states in the breast. Cell 2011;145: 926–40. [29] Singh A, Settleman J. EMT, cancer stem cells and drug resistance: an emerging axis of evil in the war on cancer. Oncogene 2010;29:4741–51. [30] Sparmann G, Hohenadl C, Tornoe J, et al. Generation and characterization of immortalized rat pancreatic stellate cells. Am J Physiol 2004;287: G211–219. [31] Vincent DF, Yan KP, Treilleux I, et al. Inactivation of TIF1gamma cooperates with Kras to induce cystic tumors of the pancreas. PLoS Genet 2009;5: e1000575. [32] Vonlaufen A, Phillips PA, Xu Z, et al. Pancreatic stellate cells and pancreatic cancer cells: an unholy alliance. Cancer Res 2008;68:7707–10. [33] Wheelock MJ, Shintani Y, Maeda M, Fukumoto Y, Johnson KR. Cadherin switching. J Cell Sci 2008;121:727–35.
Please cite this article in press as: Al-Assar O et al. Contextual regulation of pancreatic cancer stem cell phenotype and radioresistance by pancreatic stellate cells. Radiother Oncol (2014), http://dx.doi.org/10.1016/j.radonc.2014.03.014
Contents lists available at ScienceDirect
Radiotherapy and Oncology journal homepage: www.thegreenjournal.com
Original article
Contextual regulation of pancreatic cancer stem cell phenotype and radioresistance by pancreatic stellate cells Osama Al-Assar 1, Fevzi Demiciorglu, Serena Lunardi, Maria Manuela Gaspar-Carvalho, William Gillies McKenna, Ruth M. Muschel, Thomas B. Brunner ⇑ The Radiobiology Research Institute, MRC/CR-UK Gray Institute for Radiation Oncology and Biology, Department of Oncology, University of Oxford, Churchill Hospital, UK
a r t i c l e
i n f o
Article history: Received 16 August 2013 Received in revised form 12 February 2014 Accepted 18 March 2014 Available online xxxx Keywords: Pancreatic cancer Cancer stem-like cells Radioresistance Epithelial mesenchymal transition Pancreatic stellate cells
a b s t r a c t Background and purpose: Progression of pancreatic ductal adenocarcinoma (PDAC) is promoted by desmoplasia induced by pancreatic stellate cells (PSC). Contributory to this progression is epithelial mesenchymal transition (EMT), which shares many characteristics with the cancer stem cell (CSC) hypothesis. We investigated the role of these processes on the radioresponse and tumorigenicity of pancreatic cancer cells. Materials and methods: We used an in vitro sphere model and in vivo xenograft model to examine the role of PSC in EMT and CSC processes. Results: We demonstrated that PSC enhanced the CSC phenotype and radioresistance of pancreatic cancer cells. Furthermore, the expression of several EMT and CSC markers supported enhanced processes in our models and that translated into remarkable in vivo tumorigenicity. Multi-dose TGFb neutralizing antibody inhibited the EMT and CSC processes, sensitized cells to radiation and reduced in vivo tumorigenicity. A proteomic screen identified multiple novel factors that were regulated by PSC in pancreatic cells. Conclusion: These results are critical in highlighting the role of PSC in tumor progression and radioresistance by manipulating the EMT and CSC processes. TGFb and the novel factors identified are important targets for better therapeutic outcome in response to PSC mediated mechanisms. Ó 2014 Elsevier Ireland Ltd. All rights reserved. Radiotherapy and Oncology xxx (2014) xxx–xxx
Pancreatic ductal adenocarcinoma is an aggressive and highly lethal disease, with less than 1% 5-year survival. The poor prognosis of PDAC is attributed in part to the presence of a strong desmoplastic reaction driven by pancreatic stellate cells (PSC), which are the main source of excessive extracellular matrix production in chronic pancreatitis and pancreatic adenocarcinoma [3]. The presence of PSC is advantageous for tumor growth in vivo [14]. A subpopulation of cells called cancer stem cells (CSC) has been isolated from PDAC [15]. CSCs are thought to be a determining factor in chemo and radioresistance [7]. However, the model used is important in determining the outcome [1]. A link between CSC and epithelial mesenchymal transition (EMT) has been demonstrated [19]. Inducers of EMT include transforming growth factor b (TGFb) [9,26], which is also involved in PSC activation [18]. Activated PSC drive the desmoplastic reaction characteristic of ⇑ Corresponding author. Current address: University Hospitals of Freiburg, Department of Radiation Oncology, Robert-Koch-Str. 3, D-79106 Freiburg im Breisgau, Germany. E-mail address: [email protected] (T.B. Brunner). 1 Current address: MRC Unit of Molecular Haematology, The Weatherall Institute of Molecular Medicine, University of Oxford, John Radcliffe Hospital, Oxford OX3 9DS, UK.
pancreatic cancer resulting in extensive collagen types I, III and fibronectin production [24]. Furthermore, in vivo models of pancreatic cancer show an EMT defect that is attributed to deletion in TGFb signaling including DPC4/smad4 and transcriptional intermediary factor (TIF)-1c [31]. In this study, we investigated the role of PSC in the induction of CSC and EMT and the effect on the radioresponse of pancreatic cancer cells in vitro and in vivo. An improved understanding of the role of stellate cells in the contextual induction of stem cell phenotype is critical in the development of more effective treatments. Materials and methods Complete description of Materials and methods is provided in the Supplementary Materials due to space limitation. Cell culture The human pancreatic cancer cell lines PSN-1, Panc-1 and MiaPaCa-2 were originally obtained from the American Type Tissue Collection. The human pancreatic stellate cell line (hPSC) was kindly provided by Dr. Atsushi Masamune [20]. The LTC-14 cells
http://dx.doi.org/10.1016/j.radonc.2014.03.014 0167-8140/Ó 2014 Elsevier Ireland Ltd. All rights reserved.
Please cite this article in press as: Al-Assar O et al. Contextual regulation of pancreatic cancer stem cell phenotype and radioresistance by pancreatic stellate cells. Radiother Oncol (2014), http://dx.doi.org/10.1016/j.radonc.2014.03.014
2
Pancreatic stellate cells induce cancer stem cells
used in the supplementary figures were rat stellate cells [30]. The different cell lines were cultured in the recommended medium as suggested by the supplier. Indirect co-culture of pancreatic cancer cells-pancreatic stellate cells and sphere formation Human PSC were seeded in chemically-defined medium and single pancreatic cancer cell suspensions were resuspended in a mixture of Matrigel (BD Biosciences) and serum free medium, and plated into culture inserts of 1.0 lm pore size (BD Biosciences). After 10 days of incubation, spheres were counted and measured for size. Spheres were then retrieved using Matrigel Recovery Solution (BD Biosciences) according to the manufacturer’s instructions. Sodium dodecyl sulfate–polyacrylamide gel electrophoresis and Western blotting Spheres protein extracts were resolved on a NuPAGE 4–12% Bis–Tris mini gel (Invitrogen) and transferred onto a Hybond-C Extra membrane (Amersham BioSciences). Membranes were cut in several sections and incubated with several antibodies for quantitative analysis of the different samples on the same membrane using chemiluminescence (ECL, Pierce). Supplementary Table 1 summarizes all antibody concentrations used for the corresponding technique. ImageJ was used to evaluate the densitometric differences employing integrated intensity.
tumors were fixed in a 4% neutral solution of formaldehyde and cut into 10 lm sections. The antibodies used and their working concentrations are described in Supplementary Table 1. Sections used for immunohistochemistry were counterstained with hematoxylin and viewed using a standard microscope. Sections used for immunofluorescence were viewed under either an epifluorescent or confocal microscope. ImageJ was used to evaluate the average pixel density differences in whole images. Flow cytometry Panc-1 spheres grown with or without human PSC were mechanically dissociated and washed twice with ice-cold phosphate-buffered saline (PBS). Then they were incubated with the non-enzymatic Cell Recovery solution (BD Biosciences) for up to 1 h at 4 °C on a rotary shaker. Cells obtained from mechanically dissociated spheres were stained with the corresponding antibodies, as outlined in the Supplementary Table 1, and analyzed using BD FACSCalibur machine. Enzyme-linked immunosorbent assay (ELISA) and neutralizing TGFb antibody treatment Quantikine Human TGFb1 Immunoassay (R&D Systems, Inc.) was performed as described by the manufacturer. E-Cadherin ELISA was assayed as described by the manufacturer (MesoScale). In vivo tumorigenicity and TCD50 assay
Immunohistochemical and immunofluorescence staining of formalinfixed frozen tissues All animal procedures were carried out in accordance with current U.K. legislation under an approved project license. Excised
All animal procedures were carried out in accordance with the current UK legislation under an approved project license. For the in vivo tumorigenicity assay, spheres were disaggregated, injected and followed up to 6 months. For the TCD50 experiments, single
A
B
Fig. 1. The effect of human pancreatic stellate cells (PSC) on the sphere forming ability of three pancreatic cancer cell lines. Panel A shows representative pictures of spheres established from pancreatic cancer cell lines co-cultured with or without human PSC, 7–12 days after plating into Matrigel. Panel B is a numerical representation for the pictures in panel A as assessed for both size and number. The asterisks represent statistical significance for differences using the Student’s t-test (a = 0.05). The error bars represent SE of different wells from three different plates.
Please cite this article in press as: Al-Assar O et al. Contextual regulation of pancreatic cancer stem cell phenotype and radioresistance by pancreatic stellate cells. Radiother Oncol (2014), http://dx.doi.org/10.1016/j.radonc.2014.03.014
O. Al-Assar et al. / Radiotherapy and Oncology xxx (2014) xxx–xxx
dose c-irradiation (14, 20, 30 or 40 Gy) was delivered when tumor volumes reached 60–80 mm3 and volumes were calculated using the formula p/6 length width height. The tumor control probability (TCP) was modeled using the logistic model,
TCP ¼
3
ability in two pancreatic cancer cell lines strongly supporting a role for a secreted factor mediating this effect (Supplementary Fig. 1C). Co-culture conditions resulted in an increase in the expression of different stem cell markers (CD24, CD44 and CD326) and down regulation of E-cadherin indicative of EMT induction in Panc-1 spheres (Supplementary Fig. 2A and B).
1 ð1 þ ðD50=DÞrÞ
Results
Pancreatic stellate cells enhance radioresistance in vitro and in vivo Clonogenic survival after radiotherapy is a direct test of stem cell eradication [7]. Fig. 2A shows how hPSC enhance the clonogenic survival of disaggregated PSN-1 and Panc-1 spheres grown in serum free medium. In addition, whole PSN-1 and Panc-1 spheres co-cultured with hPSC were more radioresistant than spheres grown without PSC without disaggregation (Fig. 2B). In line with these observations, we also observed enhanced radioresistance of SW1222 colorectal cancer cells when co-cultured with tumor associated myofibroblasts isolated from patients (Supplementary Fig. 3). As a single surviving CSC can induce recurrent growth in vivo, we used the tumor control probability 50% assay (TCD50). Xenograft tumors grown from cancer cells co-injected with hPSC required a significantly higher dose to achieve the same level of local control compared with tumors that formed after injection of PSN-1 cells only (Fig. 2C).
Co-culture with human pancreatic stellate cells enhances the cancer stem cell phenotype in 3D serum free conditions Sphere formation, as a functional approach, is particularly useful to enrich CSC subpopulations [7]. In order to evaluate the possible effects of hPSC on stemness, three pancreatic cancer cell lines (PSN-1, MiaPaCa-2, and Panc-1), were indirectly co-cultured with PSC in three-dimensional serum free conditions and the sphereforming ability was compared. The presence of PSC enabled all cell lines to form more and larger spheres (Fig. 1). The effect was not limited to the hPSC cell line but confirmed with the rat LTC-14 pancreatic stellate cell line (Supplementary Fig. 1A). Importantly, this effect was specific to stellate cells and not observed with normal fibroblasts (Supplementary Fig. 1B). The addition of conditioned medium from hPSC had a strong effect on the sphere forming
A
B
C
Fig. 2. The effect of pancreatic stellate cells (PSC) on the radioresponse of pancreatic cancer cells in vitro and in vivo. Panel A shows the clonogenic survival of disaggregated pancreatic cancer spheres from two cell lines after co-culturing with or without stellate cells. Spheres were pooled from different wells and different plates and were cultured in a chemically defined medium. Panel B represents the radioresponse of Panc-1 spheres after irradiation in situ with or without stellate cells. The spheres were assessed for both size and number changes. The error bars represent SE of the mean of several wells from different plates. The presence of stellate cells gave an ANOVA p-value of 0.04 and 0.073 after the different doses of radiation for Panc-1 and PSN-1 sphere size, respectively. The ANOVA p-values were 0.05 and 0.10 for the average sphere number/filed after the different doses of radiation for the Panc-1 and PSN-1 cells, respectively. Panel C is a tumor control probability 50% (TCD50) assay of PSN-1 xenograft tumors established from injection with cancer cells or cancer and stellate cells. Logistic and probability analyses give almost identical tumor control probability.
Please cite this article in press as: Al-Assar O et al. Contextual regulation of pancreatic cancer stem cell phenotype and radioresistance by pancreatic stellate cells. Radiother Oncol (2014), http://dx.doi.org/10.1016/j.radonc.2014.03.014
4
Pancreatic stellate cells induce cancer stem cells
Pancreatic stellate cells promote epithelial mesenchymal transition in vitro and in vivo Western blotting showed increased expression of vimentin, a mesenchymal marker, and b-catenin, in PSN-1 and Panc-1 spheres co-cultured with PSC (Fig. 3A and B). Additionally, E-cadherin, an epithelial marker, was down-regulated in the presence of PSC, which was also verified using an ELISA for E-cadherin (Fig. 3C). We next performed co-immunofluorescence staining of PSN-1 (Fig. 4A) and Panc-1 (Fig. 4B) xenograft sections from tumors formed after co-injection with and without PSC to verify our findings in vivo. In tumor cells, vimentin expression was usually distributed diffusely in the cytoplasm, although perinuclear staining was seen in some cells as well. Positive staining for vimentin was also detected in stromal cells as expected (Figs. 4A and B). Immunohistochemistry re-confirmed the expression pattern of E-cadherin and vimentin in immunofluorescence in addition to b-catenin in PSN-1 xenograft sections (Supplementary Fig. 4A). Nuclear localization and expression of b-catenin was dramatically enhanced in the sections that had both PSN-1 and stellate cells co-injections (Supplementary Fig. 4A). This was also seen in Panc-1 cells co-injected with PSC (data not shown). Immunofluorescence of other EMT markers including N-cadherin and slug also supported the original findings in the PSN-1 xenograft sections (Supplementary Fig. 4B). Pancreatic stellate cells induce stemness as defined by cell surface markers The amount of CSC has recently been suggested to be influenced by the stromal compartment [29]. We aimed to assess the influence of PSC on the representation of CSC within pancreatic tumors
A
by performing immunofluorescence staining for cell surface markers, namely CD24 and CD326 (EpCAM). Immunofluorescence analysis of PSN-1 and Panc-1 ± PSC xenograft sections demonstrated that the co-expression of two cell surface markers in cancer cells was significantly induced by PSC (Fig. 4B). Confocal immunofluorescence was performed to assess the overlap between EMTinduced cells (increased vimentin expression/loss of E-cadherin) and CSC, defined by co-expression of CD24 and CD326. A strong but incomplete overlap between the EMT and CSC markers was observed (Fig. 4C).
The presence of stellate cells enhances the in vivo tumorigenicity of disaggregated pancreatic cancer spheres To assess the effect of stellate cells on the induction of stemness in pancreatic cancer cells, we used an in vivo tumorigenicity assay of limiting dilutions. We found that spheres co-cultured with stellate cells had a strong enhancing effect on tumorigenicity (Fig. 5). Cell numbers as low as 10 cells disaggregated from spheres co-cultured with hPSC induced tumors in all of the animals compared with none of the animals in the control group for the same cell number. The growth kinetics of both groups was similar after injection of 10,000 and 100 cells, however, a prolonged delay in growth onset was observed in the disaggregated spheres grown without PSC for the lower cell numbers. The growth kinetics of the two groups was different only at 100 cells (Student’s t-test p-value 0.025). To confirm the sphere culture system as a stem cell model we additionally found that culturing cells as spheres enhances in vivo tumorigenicity compared with the injection of cells from monolayers as shown in Fig. 5C.
B
Fig. 3. Expression of different EMT markers in spheres co-cultured with or without human pancreatic stellate cells (PSC). Panel A shows the expression of vimentin, b-catenin, and Slug in both PSN-1 and Panc-1 spheres co-cultured with PSC compared with mono-cultured cells. The molecular weight markers are shown to the left of the panel. The minus and plus signs represent co-cultures without or with PSC, respectively. Panel B represents the ratio of densitometric values normalized to GAPDH and divided over the value without PSC (dotted line). Panel C shows an ELISA of E-cadherin expression in PSN-1 and Panc-1 spheres co-cultured with or without PSC. The error bars are SE of multiple experiments.
Please cite this article in press as: Al-Assar O et al. Contextual regulation of pancreatic cancer stem cell phenotype and radioresistance by pancreatic stellate cells. Radiother Oncol (2014), http://dx.doi.org/10.1016/j.radonc.2014.03.014
O. Al-Assar et al. / Radiotherapy and Oncology xxx (2014) xxx–xxx
A
5
B
C
Fig. 4. Representative immunofluorescence staining of xenograft sections of tumors grown from PSN-1 or Panc-1 pancreatic cancer co-injected with or without pancreatic stellate cells (PSC). The sections were co-stained for EMT markers (E-Cad and vimentin) and cancer stem cell markers (CD24 and CD326). The IgG negative control for both panels is shown for each panel. Panels A and B are immunofluorescence images of staining for EMT and cancer stem cell markers, consecutively. Panel C represents confocal microscopy images of both EMT and cancer stem cell markers in PSN-1 xenograft tumors. The graphs in panel C represent the overlap of fluorescence signals for the different markers along the diagonal line drawn in the DAPI panel. The same pattern was observed in xenograft sections from different animals for both cell lines. The scale bars represent 200 lm.
Attenuation of TGFb modulates pancreatic cancer cells stemness as defined by the sphere-forming ability and reverses the profound effect of pancreatic stellate cells The release of TGFb in the co-culture medium was increased in the presence of stellate cells in the three pancreatic cancer cell lines used (Supplementary Fig. 5A). The expression of TGFb was also enhanced in xenograft tumors in the presence of stellate cells in vivo and was considerably weaker in cancer cells in sections without the co-injected stellate cells (Supplementary Fig. 5B). Attenuating TGFb by adding multiple doses of a neutralizing antibody had a strong effect on the sphere forming ability of different pancreatic cancer cell lines in monoculture and co-culture conditions (Fig. 6A). Concurrent with a reduction in CSC sphere forming ability, there was a down regulation of EMT markers and the TGFb non-canonical p-MAPK signaling since these cells harbor SMAD4 mutation (Fig. 6B). In addition, multi-dose neutralizing TGFb radio-sensitized disaggregated spheres in a classical clonogenic assay (Fig. 6C). Under the same conditions, there was a marked reduction of tumorigenicity in vivo (Fig. 6D); nonetheless, the growth kinetics was similar for all groups when tumors occurred. A single dose of neutralizing TGFb antibody had a restricted effect on the sphere forming ability (Supplementary Fig. 5C).
Discussion We here report that PSC enhance stemness in pancreatic cancer spheres and xenograft tumors, how this observation is linked with a rise in EMT features and how this impacts therapeutically on cancer stem cell survival after radiotherapy. Radiotherapy plays an important role in the multimodal treatment of pancreatic cancer but long-term local control with radiotherapy only is not achievable and this is a typical example for persistent CSC due to insufficient dose [6]. These findings are underscored by enhanced tumorigenicity in SCID mice with PSC and by the enhanced radiation dose required to locally control xenograft tumors containing PSC. We hypothesized a role for PSC in the induction of EMT and CSC in PDAC with a direct effect on treatment. To test this hypothesis we used three dimensional (3D) sphere formation in vitro and xenograft in vivo models. The prominent feature of sphere formation assays is that each sphere is accepted to derive from a single cell and is therefore clonal. Since self-renewal and differentiation can be theoretically investigated at the single-cell level, sphereforming assays are commonly used in stem cell biology [25]. We found that co-culturing of spheres with PSC results in an increase of both size and number of spheres in three different pancreatic cancer cell lines, and the differences were statistically significant (Fig. 1A and B). We used the recommended cell density
Please cite this article in press as: Al-Assar O et al. Contextual regulation of pancreatic cancer stem cell phenotype and radioresistance by pancreatic stellate cells. Radiother Oncol (2014), http://dx.doi.org/10.1016/j.radonc.2014.03.014
6
Pancreatic stellate cells induce cancer stem cells
A
B
C
Fig. 5. In vivo tumorigenicity assay of PSN-1 spheres co-cultured with or without human pancreatic stellate cells (PSC) in SCID mice. Panel A is a Kaplan–Meier graph showing tumor formation-1 of disaggregated spheres that were co-cultured with or without human PSC. The inset table and values represent incidence summary and the statistical significance of the difference between the tumor forming ability of the two groups after six-months monitoring. Panel B is the growth kinetics of the different groups in panel A stratified according to the number of cells injected. Panel C is a table representation of the in vivo tumorigenicity of disaggregated spheres after co-culture with or without PSC versus monolayer cells in SCID mice up to 9 months after injection.
from other studies to maximize clonality and the cells were embedded into Matrigel to eliminate sphere aggregation [8]. We used time lapse imaging (data not shown) to eliminate the possibility of aggregation in our studies and Supplementary Fig. 6 shows fluorescent dye retention in spheres supporting clonal expansion of spheres. Our 3D culture conditions resulted in enhanced combined expression of the three cancer stem cell markers (CD24, CD44 and CD326) when co-cultured with stellate cells (Supplementary Fig. 2A and B). These markers were shown to be cancer stem cell markers that resulted in increased tumorigenicity in vivo [15]. Other groups showed that culturing cells under 3D conditions enriches stem cell markers [12,17]. Furthermore, the same spheres in our 3D-model showed regulation of EMT markers, including E-cadherin, using fluorescence imaging, flow cytometry and western blotting (Figs. 2 and 4). The expression of E-cadherin in pancreatic cancer was cell line dependent in a previous study [16]. Both MiaPaCa-2 and Panc-1 cell lines showed down regulation of E-cadherin and strong expression of transcriptional activators of mesenchymal markers [16]. We confirmed our in vitro results using an in vivo xenograft model of two pancreatic cancer cell lines (Fig. 4, and Supplementary Fig. 4). The argument for an orthotopic model to fully assess tumor behavior [32] would make it difficult if not impossible to assess the role of PSC since they are already present in the pancreas and can become activated. The presence of PSC in our in vivo experiments resulted in b-catenin expression being predominantly nuclear in tumor cells adjacent
to the stromal compartment unlike cells in the tumor center, characteristic of EMT induction. Also characteristic of EMT induction is cadherin switching with a profound effect on cell type and behavior [33], which we observed for E and N-cadherins in our xenograft sections (Fig. 4 and Supplementary Fig. 4). Interestingly, the congruence between the EMT and CSC markers was incomplete in our experiments (Fig. 4C), which has been reported and requires further examination [19]. The up-regulation of the CSC phenotype and EMT in the context of PSC presence had remarkable translational in vitro (Fig. 2A and B) and in vivo (Fig. 2C) radiation effects. A number of other groups showed that transcriptional regulators of EMT are advantageous for the survival of cells undergoing EMT [10,13,27]. In addition, EMT induction in both MiaPaCa-2 and Panc-1 resulted in a more resistant phenotype to chemotherapeutic drugs [4]. To prove that the in vitro assay results and CSC marker profiles up-regulation in response to PSC had real functional significance on stemness in vivo, we used an in vivo tumorigenicity assay. There was a great advantage for PSC co-culture in vitro on the development of tumors in vivo (Fig. 5) and was in total agreement with the TCD50 results (Fig. 2C), strongly supporting the enrichment of CSC fraction in our model. The evidence for the correlation between tumorigenicity as a result of CSC serial dilution and TCD50 was previously shown in 13 different tumor experimental models (reviewed in [6]). The differences in tumorigenicity were not matched by changes in the growth kinetics of tumors arising from spheres co-cultured with or without PSC, except for tumors
Please cite this article in press as: Al-Assar O et al. Contextual regulation of pancreatic cancer stem cell phenotype and radioresistance by pancreatic stellate cells. Radiother Oncol (2014), http://dx.doi.org/10.1016/j.radonc.2014.03.014
O. Al-Assar et al. / Radiotherapy and Oncology xxx (2014) xxx–xxx
A
C
7
B
D
Fig. 6. The effect of TGFb attenuation on the stem cell phenotype, EMT marker expression and in vivo tumorigenicity. Panel A shows the effect of multiple doses of neutralizing TGF-b antibody on the sphere forming ability of different pancreatic cancer cell lines co-cultured indirectly with human pancreatic stellate cells (PSC). All results were statistically significant using the Student’s t-test (a = 0.05) for the effect of TGFb on sphere size and number. Fresh antibody was added for five days and the spheres were analyzed on day 8. The error bars represent SE of different wells from three different plates. Panel B shows the expression of different EMT markers using western blotting in the co-culture conditions and after the addition of TGFb neutralizing antibody. Panel C represents a clonogenic assay of TGFb pre-treated spheres that were disaggregated and subsequently treated with different radiation doses. The error bars represent SE of the mean of several wells from different plates. Panel D is a Kaplan– Meier graph showing in vivo tumorigenicity assay of PSN-1 disaggregated spheres co-cultured with or without human PSC in female SCID mice after a 5-day treatment with neutralizing TGFb antibody in vitro. The results are statistically significant as shown in the inset values. Also shown in panel D is the growth kinetics of the different tumor groups. A total cell number of 1000 was injected for each group of 5 mice and the experiment was allowed to run for 60 days and the animals were terminated when they reached 100 mm3.
originating from injections with 30 cells. This implies the contextual regulation of CSC percentage and the sphere conditions enriched for stemness compared with the 2D monolayer culture (Fig. 5C). TGFb is a factor with paradox roles in tumor suppression and promotion (Fig. 7) [5]. Inhibition of TGFb signaling in PSN-1, Panc-1 and MiaPaCa-2 cell lines, using multiple doses of neutralizing TGFb antibody, resulted in the suppression of the CSC 3D sphere formation phenotype and down regulation of EMT markers in our sphere forming assay (Fig. 6A and B). In addition, multipledose neutralizing TGFb antibody had a radio-sensitizing effect on disaggregated spheres co-cultured with or without human PSC in Panc-1 and PSN-1 pancreatic cancer cell lines (Fig. 6C) and reduced in vivo tumorigenicity in PSC-1 cell (Fig. 6D). Unlike Panc-1 and MiaPaCa-2, the inhibition of the CSC and EMT markers in the PSN-1 cell line is DPC4/Smad4 independent due to the homologous deletion of this gene [22]. In this context, we showed inhibition of non-canonical TGFb signaling through pMAPK in the spheres, by western blotting (Fig. 6B). Our results are in agreement with published data showing a role for TGFb in the self-renewal and maintenance of stem cells in breast cancer [19,23] and glioma samples [2]. However, other
groups showed that other members of the TGFb superfamily were key players in the sphere forming ability of primary pancreatic cancer cells, but not TGFb [17]. These discrepancies might be due to the presence of a cross-talk between TGFb, WNT, Notch and Sonic hedgehog in EMT and stem cell induction [28]. We identified several targets that have the potential of being very valuable in understanding the mechanism of CSC phenotype and EMT induction, consequently leading to better tumor control (Supplementary Fig. 7). One of the factors identified in our screen is CXCL8 (Supplementary Fig. 7B), which has been shown to be involved in stromal signaling in pancreatic cancer [21] and in breast cancer stem cell maintenance [11]. In summary, our results show that PSC enhance both CSC phenotype and EMT process in vitro and in vivo. Diagram 1 highlights some of the targets we investigated in this study, including the interplay of these targets with TGFb. Our results also emphasize the importance of contextual stromal factors in affecting the CSC representation in tumors, leading to direct translational effects on tumorigenesis and therapeutic outcome. In addition, continued investigation is required to verify our potential therapeutic targets in vivo, and ultimately their clinical use to improve therapy.
Please cite this article in press as: Al-Assar O et al. Contextual regulation of pancreatic cancer stem cell phenotype and radioresistance by pancreatic stellate cells. Radiother Oncol (2014), http://dx.doi.org/10.1016/j.radonc.2014.03.014
8
Pancreatic stellate cells induce cancer stem cells
Tumor supressing
Tumor promong TGFβ Levels Wnt
?
EMT
KRAS
TP53 ink4A
MET
MAPK SMAD4
?
Vimenn
TP53 ink4A c-Myc
? Proteome Array
FGF
MAPK SMAD4
Β-Catenin FAK
?
Cancer stem cell phenotype
N-Cadherin CXCR4
Connecve Tissue Growth Factor
Integrins
Immune cells
? Acvaon
?
KRAS
GSK3
E-Cadherin
Frizzled
Stromal modificaon
Endothelial cells
PSC PSC: Pancreac stellate cells EMT: Epithelial mesenchymal transion MET: Mesenchymal epithelial transion Fig. 7. The diagram depicts the role of pancreatic stellate cells (PSCs) in transforming growth factor b (TGFb) signaling pathways in the context of pancreatic cancer cells (PCCs). PCCs are shown in the top half of the diagram and star-shaped PSCs in the bottom half and endothelial as well as immune cells are other cell types of the tumor stroma. Tumor suppressing effects of TGFb are lost with rising levels of the growth factor and converted to tumor promoting effects on PCCs and with epithelial– mesenchymal transition (EMT). EMT is reversible and then called MET. TGFb is produced by both PCCs and PSCs and TGFb downstream signaling in PCCs is often defective (e.g. DPC-4 mutation) and thereby contributes to paracrine signaling. PCCs that have undergone EMT will express a number of factors such as N-cadherin, vimentin and bcatenin. At the same time the expression of cancer stem cell markers is enhanced.
Financial support Medical Research Council (UK) and Cancer Research, UK. Conflict of interest None declared. Acknowledgements We acknowledge the Medical Research Council (UK) and Cancer Research UK for their core funding. We would like to thank Emmanouil Fokas, Karla Watson and Magdalena Flieger for their help with the animal experiments. We would also like to thank John Fenwick for performing the probability analysis for the TCD50 assay, and Prof. Sir Walter Bodmer’s laboratory for the SW1222 cell line and Neil Ashley for supplying us with the primary colorectal myofibroblasts. Last, we thank Helio Pais for bioinformatics assistance with the proteome analysis. Appendix A. Supplementary data Supplementary data associated with this article can be found, in the online version, at http://dx.doi.org/10.1016/j.radonc.2014. 03.014. References [1] Al-Assar O, Muschel RJ, Mantoni TS, McKenna WG, Brunner TB. Radiation response of cancer stem-like cells from established human cell lines after sorting for surface markers. Int J Radiat Oncol Biol Phys 2009;75:1216–25. [2] Anido J, Saez-Borderias A, Gonzalez-Junca A, et al. TGF-beta receptor inhibitors target the CD44(high)/Id1(high) glioma-initiating cell population in human glioblastoma. Cancer Cell 2010;18:655–68.
[3] Apte MV, Park S, Phillips PA, et al. Desmoplastic reaction in pancreatic cancer: role of pancreatic stellate cells. Pancreas 2004;29:179–87. [4] Arumugam T, Ramachandran V, Fournier KF, et al. Epithelial to mesenchymal transition contributes to drug resistance in pancreatic cancer. Cancer Res 2009;69:5820–8. [5] Bachman KE, Park BH. Duel nature of TGF-beta signaling: tumor suppressor vs. tumor promoter. Curr Opin Oncol 2005;17:49–54. [6] Baumann M, Krause M, Hill R. Exploring the role of cancer stem cells in radioresistance. Nat Rev Cancer 2008;8:545–54. [7] Brunner TB, Kunz-Schughart LA, Grosse-Gehling P, Baumann M. Cancer stem cells as a predictive factor in radiotherapy. Semin Radiat Oncol 2012;22:151–74. [8] Chojnacki A, Weiss S. Production of neurons, astrocytes and oligodendrocytes from mammalian CNS stem cells. Nat Protoc 2008;3:935–40. [9] Ellenrieder V, Hendler SF, Ruhland C, Boeck W, Adler G, Gress TM. TGF-betainduced invasiveness of pancreatic cancer cells is mediated by matrix metalloproteinase-2 and the urokinase plasminogen activator system. Int J Cancer 2001;93:204–11. [10] Escriva M, Peiro S, Herranz N, et al. Repression of PTEN phosphatase by Snail1 transcriptional factor during gamma radiation-induced apoptosis. Mol Cell Biol 2008;28:1528–40. [11] Ginestier C, Liu S, Diebel ME, et al. CXCR1 blockade selectively targets human breast cancer stem cells in vitro and in xenografts. J Clin Invest 2010;120:485–97. [12] Gou S, Liu T, Wang C, et al. Establishment of clonal colony-forming assay for propagation of pancreatic cancer cells with stem cell properties. Pancreas 2007;34:429–35. [13] Haslehurst AM, Koti M, Dharsee M, et al. EMT transcription factors snail and slug directly contribute to cisplatin resistance in ovarian cancer. BMC Cancer 2012;12:91. [14] Hwang RF, Moore T, Arumugam T, et al. Cancer-associated stromal fibroblasts promote pancreatic tumor progression. Cancer Res 2008;68:918–26. [15] Li C, Heidt DG, Dalerba P, et al. Identification of pancreatic cancer stem cells. Cancer Res 2007;67:1030–7. [16] Li C, Lee CJ, Simeone DM. Identification of human pancreatic cancer stem cells. Methods Mol Biol 2009;568:161–73. [17] Lonardo E, Hermann PC, Mueller MT, et al. Nodal/Activin signaling drives selfrenewal and tumorigenicity of pancreatic cancer stem cells and provides a target for combined drug therapy. Cell Stem Cell 2011;9:433–46. [18] Von Mahadevan D, Hoff DD. Tumor-stroma interactions in pancreatic ductal adenocarcinoma. Mol Cancer Ther 2007;6:1186–97. [19] Mani SA, Guo W, Liao MJ, et al. The epithelial–mesenchymal transition generates cells with properties of stem cells. Cell 2008;133:704–15.
Please cite this article in press as: Al-Assar O et al. Contextual regulation of pancreatic cancer stem cell phenotype and radioresistance by pancreatic stellate cells. Radiother Oncol (2014), http://dx.doi.org/10.1016/j.radonc.2014.03.014
O. Al-Assar et al. / Radiotherapy and Oncology xxx (2014) xxx–xxx [20] Masamune A, Watanabe T, Kikuta K, Satoh K, Shimosegawa T. NADPH oxidase plays a crucial role in the activation of pancreatic stellate cells. Am J Physiol 2008;294:G99–G108. [21] Matsuo Y, Raimondo M, Woodward TA, et al. CXC-chemokine/CXCR2 biological axis promotes angiogenesis in vitro and in vivo in pancreatic cancer. Int J Cancer 2009;125:1027–37. [22] Moore PS, Sipos B, Orlandini S, et al. Genetic profile of 22 pancreatic carcinoma cell lines. Analysis of K-ras, p53, p16 and DPC4/Smad4. Virchows Arch 2001;439:798–802. [23] Morel AP, Lievre M, Thomas C, Hinkal G, Ansieau S, Puisieux A. Generation of breast cancer stem cells through epithelial–mesenchymal transition. PLoS One 2008;3:e2888. [24] Omary MB, Lugea A, Lowe AW, Pandol SJ. The pancreatic stellate cell: a star on the rise in pancreatic diseases. J Clin Invest 2007;117:50–9. [25] Pastrana E, Silva-Vargas V, Doetsch F. Eyes wide open: a critical review of sphere-formation as an assay for stem cells. Cell Stem Cell 2011;8:486–98. [26] Peinado H, Quintanilla M, Cano A. Transforming growth factor beta-1 induces snail transcription factor in epithelial cell lines: mechanisms for epithelial mesenchymal transitions. J Biol Chem 2003;278:21113–23.
9
[27] Robson EJ, Khaled WT, Abell K, Watson CJ. Epithelial-to-mesenchymal transition confers resistance to apoptosis in three murine mammary epithelial cell lines. Differentiation 2006;74:254–64. [28] Scheel C, Eaton EN, Li SH, et al. Paracrine and autocrine signals induce and maintain mesenchymal and stem cell states in the breast. Cell 2011;145: 926–40. [29] Singh A, Settleman J. EMT, cancer stem cells and drug resistance: an emerging axis of evil in the war on cancer. Oncogene 2010;29:4741–51. [30] Sparmann G, Hohenadl C, Tornoe J, et al. Generation and characterization of immortalized rat pancreatic stellate cells. Am J Physiol 2004;287: G211–219. [31] Vincent DF, Yan KP, Treilleux I, et al. Inactivation of TIF1gamma cooperates with Kras to induce cystic tumors of the pancreas. PLoS Genet 2009;5: e1000575. [32] Vonlaufen A, Phillips PA, Xu Z, et al. Pancreatic stellate cells and pancreatic cancer cells: an unholy alliance. Cancer Res 2008;68:7707–10. [33] Wheelock MJ, Shintani Y, Maeda M, Fukumoto Y, Johnson KR. Cadherin switching. J Cell Sci 2008;121:727–35.
Please cite this article in press as: Al-Assar O et al. Contextual regulation of pancreatic cancer stem cell phenotype and radioresistance by pancreatic stellate cells. Radiother Oncol (2014), http://dx.doi.org/10.1016/j.radonc.2014.03.014
