Background
Pancreatic ductal adenocarcinoma (PDAC) is one of the most lethal cancers with less than 5% of overall patient survival after 5 years. Local and distant invasion, resistance to chemotherapy and radiotherapy and lack of early detection are responsible for this poor prognosis. Gemcitabine (2',2'-difluorodeoxycytidine, a pyrimidine nucleoside analogue) chemotherapy, is the standard treatment of the patients. The combination of gemcitabine with other chemo- or biotherapies has resulted in a very limited prognostic improvement. Recently, a high throughput RNAi screen identified the checkpoint kinase 1 (CHK1) as a gene conferring resistance to gemcitabine in pancreatic cancer cells [
1]. CHK1 is a key component of the cell cycle checkpoints that are activated by genomic and replicative stress (review in [
2]). This checkpoint activation is known to facilitate DNA repair. Consequently, CHK1 may play an important role in the resistance of tumor cells to genotoxic therapy, raising the possibility that inhibitors of checkpoint kinases may be useful adjuvant agents in chemotherapy of cancer. In the case of pancreatic cancer, in vitro and in vivo studies have shown that CHK inhibitors enhance the antitumor activity of gemcitabine [
3‐
5].
The MultiCellular Tumor Spheroid (MCTS) model is generally considered as a better model than two dimensional culture to predict the in vivo response to drug treatments [
6‐
8] and it is now widely accepted that MCTS reproduce more accurately the tumor microenvironment than monolayer cell cultures. While growing, spheroids display a gradient of proliferating cells from the outer cell layers with quiescent cells located more centrally. When deprived of oxygen and glucose, central cells die and a necrotic zone is formed. This cell heterogeneity is similar to that found in avascular micro-regions of tumors [
9]. It is well established that solid tumor environment induces the level of drug resistance to many chemotherapeutic agents. This phenomenon, called multicellular resistance [
10], emerges as soon as cancer cells have established contacts with surrounding cells or extracellular matrix, i.e. its microenvironment. In MCTS, cancer cells can acquire this multicellular resistance by interacting efficiently in 3 dimensions with their environment [
10‐
12].
In order to contribute to the discovery of new anti pancreatic cancer agents or new potent combinations with gemcitabine, we describe here the development and the validation of a new spheroid model mimicking the structure and chemo resistance of pancreatic solid tumors compared to conventional 2D cell culture models. We also present the spatio-temporal parameters of the biological response of gemcitabine alone or combined with a CHK1 inhibitor, CHIR-124.
Materials and methods
Reagents
Gemcitabine was purchased from Sigma. CHIR-124 was a generous gift of Dr Alain Pierré (Institute de Recherche Servier).
Cell culture
Capan-2 pancreatic cancer cells were cultured in DMEM/F12 (Invitrogen, France) containing 10% FCS with 2 mmol/l glutamine and penicillin/streptomycin in a humidified atmosphere of 5% CO2 at 37°C. Capan-2 cells were transduced with a lentiviral vectors coding for fused green -emitting fluorescent proteins to Geminin [
13].
Spheroid generation
Spheroids were prepared according to [
14]. A Capan-2 cell suspension containing 10
4 cells/ml of DMEM/F12 supplemented with EGF (20 ng/ml) (Invitrogen) and B27 (Invitrogen) was prepared. 100 μl of this cell suspension were plated on each well of poly-HEMA-coated 96-well plates. The plates were centrifugated at 200 g during 6 min and then incubated in a humidified atmosphere of 5% CO2 at 37°C. By using this technique we obtained single spheroids in each well, the variation of size between spheroids is less than 10%. In order to generate quiescent spheroids, after a first 4 days growth phase in defined medium (DMEM/F12 supplemented with EGF and B27), spheroids were washed twice with media containing 10% FCS, and then incubated with this media during 1-6 days.
Spheroid viability quantification
Spheroid viability was quantified by ATP monitoring with the Perkin Elmer ATPlite™ assay system. This system is based on the production of light caused by the reaction of ATP, a cell viability marker present in cell lysate, with added luciferase and D-luciferin. We adapted ATPlite assay procedure for spheroid application, especially concerning spheroid dissociation and cell lysis. Then 100 μl of mammalian cell lysis solution (ATPlite kit) were added to each well containing one spheroid in 100 μl of culture medium. The plate was shaken for 20 min. In order to read luminescent signal, 75 μl of the cell lysate was transferred to a black 96-well plate. Then 37 μl of DMEM/F12 medium containing 10% FCS and 37 μl of ATPlite kit substrate solution were added. After 15 min of shaking, the luminescence signal was read on an Envision® plate reader (Perkin Elmer).
Immunofluorescence on frozen sections
Capan-2 spheroids were rinsed with PBS and fixed in 4% neutral-buffered formalin (Sigma) for 2 h. After fixation, spheroids were processed for 5 μm frozen sections. Sections were incubated overnight at 4°C with antibodies directed against cleaved form of PARP (rabbit monoclonal, Epitomics, 1/1000), or γH2AX phosphorylated (mouse monoclonal, Upstate, 1/200) and Ki67 (rabbit polyclonal, Santa Cruz, 1/200). After washing in PBS/Triton 0.1% v/v, the secondary antibody was applied (Alexa 488-anti-mouse or Alexa 594-anti-rabbit, Molecular Probes, 1/800, for 1 h at room temperature). To determine cell cycle repartition, sections of Capan-2 spheroids expressing the green FUCCI probe were directly analyzed by fluorescence imaging. The observations were based on the examination of 3 sections from at least 5 spheroids. Each experiment has been repeated a minimum of 3 times.
Cytotoxicity assays
Spheroids were generated using 1000 cells in 100 μl per well as indicated in spheroid generation section. After 4 days of culture, chemotherapeutic agents or combinations were added (10 μl/well). Spheroid viability was evaluated by ATP quantification after 72 h compound treatment. Tests were performed in triplicate and the data presented are from at least three separate experiments. ATP content percentage was calculated with regard to non-treated spheroid and showed cell growth inhibition and/or toxicity. The 50% effective concentration (EC50) of a compound is the concentration which provokes 50% of the maximal effect (ATP production decrease) of this drug. Curve fittings were performed with GraphPad (San Diego, CA) Prism version 4.0 software using the sigmoidal dose-response to determine EC50 values.
Discussion
Standard chemotherapeutic drugs have limited effect in large-scale clinical trials for pancreatic cancer. Because of the very poor prognosis of this type of cancer, novel approaches are therefore urgently needed. Most in vitro screening approaches are based on monolayer culture of pancreatic cancer cells but it is well established that tumor microenvironment plays an important role in response to chemotherapy. It is therefore of major importance that more predictive pharmacological models be developed for the assessment of new therapeutic strategies.
Multicellular Tumor Spheroids are of particular interest as they offer a level of intermediate complexity that recapitulate the three-dimensional organization of a tumor and integrate the notion of microenvironment. The production of 500-600 μm large spheroids from various epithelial cancer cell lines has already been shown for colon, breast, prostate and kidney but not pancreas with the liquid overlay technology [
18]. Spheroids from several pancreatic ductal adenocarcinoma (PDAC) cell lines were obtained on micro-patterned culture plates but no pharmacological analysis were presented with these models [
19]. Recently, PDAC cell lines grown in 3D collagen microenvironment were shown to proliferate in the presence of gemcitabine whereas they stopped growing when cultivated on tissue culture plastic indicating that 3D cell organisations could have an impact on pancreatic cancer cell drug sensitivity [
20]. Then, the development of new MCTS models represents an interesting way to improve the discovery of new treatment. By using the in vivo validated gemcitabine and CHIR124 molecules [
3], we show here that our Capan-2 MCTS model for pancreatic cancer could detect effective drug combinations.
In this study we developed an "automation friendly" spheroid model of Capan-2 pancreatic cancer cell spheroids in 96 well-plates. We chose ATP quantification to measure the effect of chemical compounds on cell viability and proliferation. We showed that epidermal growth factor (EGF) was necessary to maintain Capan-2 cell proliferation in a 3-D context, whereas it was not the case in monolayer. It is well known that EGF plays an important role in pancreatic cancer progression and EGF and its ligand over-expression have been frequently observed in pancreatic cancer [
21,
22]. A recent study reporting the effects of EGF ligands in different culture conditions of ovarian cancer cells clearly showed that in contrast to monolayer culture, spheroids facilitated growth stimulatory activity of EGF ligands [
23]. This EGF dependent-proliferation of spheroids emphasized the relevance of this model by comparison with cell monolayer and with tumor context. Moreover, the EGFR systems and associated signaling pathway could be promising targets for pancreatic cancer treatment [
24]. Consequently Capan-2 cell spheroid appears to be a relevant model to screen for EGF signaling targeting compounds.
A proliferation gradient was observed for spheroids around 600 μm diameter: proliferative cells were located in the outer layer whereas quiescent cells were located more centrally. It has been previously shown that when the central cells become deprived of oxygen and glucose, cell death and necrosis occur [
9]. According to this, we found that apoptotic cells were detected in the spheroid center after 7 days when the spheroid size reached 600 μm. This proportion greatly increased until day 12. The characterization of the proliferation gradient in the spheroid of different sizes clearly showed that there was a window to test antitumoral compounds. This window started when proliferation gradient was established (after 4 days) but before central necrosis appeared at onset of treatment (before 7 days).
Most in vitro studies on the response of pancreatic cancer cell to gemcitabine were based on monolayer cell culture. A study reports that gemcitabine was less potent when cancer cells were grown as multilayer compared to monolayer cultures [
25]. It is well established that for many chemotherapeutic drugs a solid tumor environment results in an increased level of drug resistance, a phenomenon called the multicellular resistance. Multicellular resistance emerges as soon as cancer cells have established contacts with their microenvironment, homologous cells, heterologous cells or extracellular matrix [
10,
26]. This contact dependent resistance can be observed when cell are cultured as spheroid. Spheroid culture of glioblastoma cells are less sensitive to gemcitabine than monolayer cells [
27]. Our results show that pancreatic Capan-2 cells cultured as spheroids are also less sensitive to gemcitabine than Capan-2 monolayer. This result agrees with a recent study showing that a 3-D collagen microenvironment protects pancreatic cancer cells from gemcitabine-induced proliferation arrest [
20]. Spheroid permeability, presence of quiescent and hypoxic cells could explain this resistance [
10]. Our observation that gemcitabine potency was reduced in quiescent Capan-2 spheroid suggests that pancreatic cancer cell proliferation status plays a role in gemcitabine response.
DNA damage induced by gemcitabine results in activation of S cell cycle checkpoint and apoptosis [
15]. In addition to assess the global cytotoxicity of anticancer agents, the spheroid model allows to image cell response in function of their position within the spheroid [
12,
28]. H2AX phosphorylation, which has been demonstrated as a pharmacodynamic indicator of gemcitabine-induced stalled replication forks [
15], was first used to image gemcitabine response in Capan-2 spheroid. The establishment of gemcitabine-induced S phase checkpoint was characterized by using Capan-2 cells expressing the Fucci reporters corresponding to the fluorescent protein geminin-mAG (green) that is expressed in S/G2/M phases of the cell cycle. Our results show that 16 h after gemcitabine addition only the cells located in the outer cell layer are targeted by gemcitabine. Indeed, cells of the outer cell layer are those with damaged DNA and accumulated in the S/G2/M phases of the cell cycle. This spatially confined DNA damage may result from limited drug penetration or a low sensitivity of non-proliferating cells in deeper spheroid layers. Our results do not discriminate between these two hypotheses. One limitation to gemcitabine efficacy is its poor penetration in human tumors [
22,
29,
30]. Using a multicellular layer method to study drug penetration it has been shown that the penetration of gemcitabine in multicellular cell layer is independent of cell concentration but decrease with the thickness of the layer [
31]. 48 h after gemcitabine addition, cells arrested in the S phase remained located in the outer cell layer whereas DNA damages and apoptosis were detected throughout the spheroid suggesting that DNA damage rather than cell cycle arrest are correlated with apoptosis. This result agrees with a previous study showing that in spheroids the persistence of DNA damage determined by γH2AX staining predicted clonogenic cell survival [
32].
One field of spheroid interest is the study of drug combination. Inhibition of CHK1 represents a targeted approach to selectively enhance the cytotoxicity of DNA-damaging agents in tumor cells. Whereas, p53-deficient cells have been preferentially killed by the combination of a DNA damaging agent, which arrest the cell cycle in G2, followed by CHK1 inhibitor, p53-proficient tumors could potentially be targeted by concurrent administration of an antimetabolite and a CHK1 inhibitor [
33]. CHK1 inhibitors are able to potentiate gemcitabine cytotoxicity in vitro and in vivo [
3,
4,
16,
34,
35]. In agreement with these results our data show that gemcitabine and a CHK1 inhibitor (CHIR-124) exert a synergistic cytotoxic effect on Capan-2 spheroid. This synergic cytotoxic effect was associated with an increase in the ability of gemcitabine to trigger DNA damage and apoptosis. Taken together these data indicate that the spheroid model provide new information concerning the role of cancer cell microenvironment on the gemcitabine and CHK1 inhibitor pancreatic cancer cell response.
Competing interests
The authors declare that they have no competing interests.
Authors' contributions
ID, CF, FS, LD, PC and AV performed experiments. FA, BD and AV directed the study. ID, FA, BD and AV wrote the paper. All authors read and approved the final manuscript.