Background
Standard therapy for solid tumors traditionally consists of different approaches, including surgical resection, hormone therapy, systemic chemotherapy and radiotherapy. However, within the last years combining the traditional approaches with molecular targeted therapies, using monoclonal antibodies and small molecule inhibitors, has become more and more important [
71]. Prime targets for this strategy for tumor control are oncogenic signaling cascades, such as the januskinase/signal transducers and activators of transcription, the tumor necrosis factor signaling or the mitogen-activated protein kinases (MAPK) pathway. Especially the MAPK signaling pathway has been shown to stimulate proliferation, cell growth, survival and resistance to chemotherapeutics and ionizing radiation IR [
7,
10,
12]. In particular the novel ATP non-competitive MEK inhibitor AZD6244 (generic names: Selumetinib, ARRY-142886) demonstrated high specificity and anti-proliferative activity in
in vitro and
in vivo models [
69]. Various research groups demonstrated, that apart from the cytostatic effects, AZD6244 also sensitized human tumor cell lines of different origins to IR, underlining the potential of the MAPK pathway as a target for radiosensitization [
9,
10,
62].
Another important oncogenic signaling cascade for a molecular targeted therapy is the phosphatidylinositol 3-kinase (PI3K)/mammalian target of rapamycin (mTOR) pathway, which is also related to proliferation and therapy resistance and which also has been validated as a target for radiosensitizing approaches in various
in vitro and
in vivo studies [
8,
19,
32,
40,
58]. Especially the dual PI3K/mTOR inhibitor NVP-BEZ235 revealed a promising radiosensitizing potential in several experiments [
20,
21,
37,
38,
49].
Although, first promising results were obtained for signaling cascade inhibitors in cancers depending on mutations of a single signaling pathway, only limited treatment success was observed, when multiple signaling cascades were deregulated [
15,
16,
27], indicating a dependency on the individual mutational background. One possible reason for this limited therapy success is the compensatory up regulation of (other) pathways by feedback loops and/or crosstalks after drug treatment. Such compensatory activation has been shown for a number of cell lines of different tumor entities pointing to its involvement in treatment resistance [
34,
35,
42]. Apart from this cell specific a priori resistance to various drugs, the perturbation of a signaling pathway can also result in an acquired drug resistance of initially responsive tumor cells, which ultimately leads to treatment failure [
31]. One approach to avoid this resistance by the induction of complementary signaling after drug treatment is to combine inhibitors of different pathways in order to achieve synergistic effects by inhibiting the complementary signaling cascades. In fact, it was proven in several
in vitro and
in vivo studies, that simultaneous perturbation of the MAPK and the PI3K/mTOR pathways resulted in enhanced effects compared to single pathway inhibition [
5,
25,
53,
66].
Especially the MEK inhibitor AZD6244 and the dual PI3K/mTOR inhibitor NVP-BEZ235 demonstrated synergistic effects in several
in vitro studies investigating various tumor entities [
24,
26,
53,
56,
59]. Furthermore, the promising
in vitro effects of the combined treatment with AZD6244 and NVP-BEZ235 were already validated in several xenografts in vivo studies with cells of different tumor entities, showing significant synergistic effects including increased tumor shrinkage and prolonged median survival after combined treatment [
17,
47,
52,
63].
Although there are several publications, validating the synergistic effects of simultaneous treatment with AZD6244 and NVP-BEZ235, to our knowledge there is no study available evaluating if these synergistic effects are enhanced, when the drugs are combined with IR. To assess the effects of simultaneous MEK and PI3K/mTOR inhibition on the MAPK and PI3K/mTOR signaling cascades and to integrate these data with the phenotypic data of the radiation response after simultaneous MEK and PI3K/mTOR inhibition, we treated glioblastoma SNB19 and lung carcinoma A549 cells with AZD6244 and NVP-BEZ235 alone and in combination. The two cell lines differ in their mutational status, as shown in Table
1, which summarizes mutations of known cancer genes in the two cell lines [
30]. As illustrated in Table
1, both cell lines have a common mutation in CDK2NA, which codes for the tumor suppressor protein p16. However, the two cell lines differ in their mutational status regarding the oncogenic MAPK and PI3K/mTOR pathways. SNB19 cells are not expressing functional phosphatase and tensin homologue (PTEN), which is a negative regulator of the PI3K/mTOR signaling cascade [
70]. A549 lung carcinoma cells in contrast do not have any known mutations in the PI3K/mTOR pathway, but a mutation in the
Kirsten rat sarcoma viral oncogene homolog (
KRAS) gene, which occurs in about 30 % of non small cell lung cancers [
6]. These mutations result in a constitutive active form of the protein [
1], ultimately leading to activation of the MAPK signaling cascade.
Table 1
Mutations of known cancer genes in the glioblastoma SNB19 and the lung carcinoma A549 cell lines [
30]
SNB19 | CDKN2A | Hom c.1_471 del 471, p.? | LOM |
| PTEN | Hom c.723_724insTT p.E242fsX15 | LOM |
| TP53 | Hom c.818G > A, p.R273H | LOM |
A549 | CDKN2A | Hom c.1_471 del 471, p.? | LOM |
| KRAS | Hom c.34G > A, p.G12S | LOM |
| STK11 | Hom c.109C > T, p.Q37X | LOM |
After determining the effects of AZD6244 and NVP-BEZ235 on the cellular proliferation rates and the expression levels of several key proteins of the MAPK (Raf-1, p-MEK1/2, MEK2, p-Erk1/2 and Erk2) and PI3K/mTOR signaling cascades (PI3K p110, PI3K p85, PTEN, p-Akt, Akt, p-mTOR, mTOR, p-S6, S6 and p-4E-BP1), we assessed the colony forming abilities, the cell cycle phase distributions, the expression levels of cell cycle related proteins (CDK1, CDK4 and p-Rb), the incidence of apoptosis markers (hypodiploid cells and poly (ADP-Ribose) polymerase (PARP) expression levels and cleavage) and the expression levels of autophagy related proteins (LC3-I and LC3-II) dependent on drug treatment and IR.
Methods
Cell culture and drug treatment
The human lung cancer cell line A549 and the human glioblastoma cell line SNB19 were obtained from the “Cell Line Services” company (Heidelberg, Germany) and routinely cultured under standard conditions (37 °C, 5 % CO2) in Dulbecco’s modified Eagle’s medium supplemented with 10 % FBS, 1 % glutamine and 1 % Penicillin-Streptomycin. For the proliferation assays cells were treated for 24 h with the indicated concentrations of AZD6244 (Selleckchem, Houston, TX, USA) and NVP-BEZ235 (Novartis Institutes for Biomedical Research, Basel, Switzerland) before measurement of the ATP content. For the other experiments of this study cells were treated 16 or 1 h prior to IR with 500 nM AZD6244 or 50 nM NVP-BEZ235, respectively. Drugs were freshly diluted from frozen aliquots stored at −20 °C. Cells treated in parallel with dimethylsulfoxide (DMSO) served as controls.
Cell viability assay
The proliferation rate was analyzed with the CellTiter-Glo Luminescent Cell Viability Assay (Promega, Madison, WI, USA) according to the manufacturer’s instructions. Serial dilutions of AZD6244 (31.25–4000 nM) or NVP-BEZ235 (3.125-400 nM) were added to exponentially growing cells and the ATP levels were determined 24 h afterwards. Furthermore, experiments with serial dilutions of AZD6244 (31.25–4000 nM) in the presence of 50 nM NVP-BEZ235 and NVP-BEZ235 (3.125–400 nM) in the presence of 500 nM AZD6244 were performed. Experiments were done in triplicates and the mean ATP content data derived from two independent experiments were normalized against DMSO-treated controls to generate dose–response curves. Further analysis of the data was basically performed as described previously [
38].
X-ray IR
IR was performed at room temperature using a 6 MV linear accelerator (Siemens, Concord, USA) at a dose rate of 2 Gy/min. After IR, cells were cultured in standard conditions for the indicated time until harvest.
Western blot
The preparation of whole cell lysates, separation according the protein size using Western blotting techniques and the detection of protein levels using protein-specific primary and species-specific peroxidase-labeled secondary antibodies were performed according standard protocols as described previously [
68]. The antibodies used in this study are specified in Additional file
1. Protein expression levels were quantified using ImageJ (NIH, Bethesda, MD, USA), normalized to β-actin levels and the relative protein expressions of the shown representative biological replicate are denoted by numbers below the corresponding blot (if changes between treatments were observed). For each Western blot experiment three independent biological replicates were performed.
Colony forming assays were performed and data were analyzed with the linear quadratic model as described elsewhere [
22]. Briefly, cells were treated with 500 nM AZD6244 and/or 50 nM NVP-BEZ235 16 h or 1 h prior to IR, respectively. Twenty-four hours after IR with graded single doses up to 8 Gy the cells were detached and seeded into 6 well plates containing drug free medium. Cells were then cultivated under standard conditions for two weeks. Colonies were stained with 0.6 % crystal violet and colonies containing more than 50 cells, were scored as survivors. Experiments were done in triplicate and each experiment was repeated at least four times.
Measurement of cell cycle phase distribution and hypodiploid cells
Cell cycle phase distributions and proportion of hypodiploid cells were assessed as described elsewhere [
55]. Briefly, samples were fixed 30 min, 24 and 48 h after IR by adding ice cold ethanol. After permeabilization and treatment with RNase A cells were stained with propidium iodide (PI) and at least 20,000 cells were assessed for their DNA content, using a flow cytometer FACSCanto II (Becton Dickinson, San Jose, CA, USA). For cell cycle analysis cell conglomerates and hypodiploid cells were excluded and deconvolution of DNA histograms was performed using the ModFit LT (Verity Software House, Topsham, ME, USA) software. Cells showing less than 80 % of the fluorescence signal of average G1-phase cells were considered to be hypodiploid.
Software and statistics
Data are represented as means ± standard deviation (SD) of at least three independent experiments. Unpaired two-sided t-tests were performed and P values <0.05 were considered to be statistically significant. For multiple comparisons the Holm-Bonferroni method of alpha error correction was applied. Statistical comparison of colony forming assays was done using the statistical software RStudio 0.96.331 (Free Software Foundation, Boston, MA, USA) along with the package CFAssay (H. Braselmann, Helmholtz Zentrum German Research Center for Environmental Health, Munich, Germany). For statistical comparisons of the cell cycle phase distribution data we tested each cell cycle phase (G1-, S- or G2/M-phase) between the different treatment groups (Control, AZD6244, NVP-BEZ235, and AZD6244 + NVP-BEZ235). Statistical significant differences in at least one of the cell cycle phases between the different treatment groups are indicated in the figure. For reasons of clarity we omitted comparisons between unirradiated and irradiated samples. Further software used in this study was Flowing Software (P. Terho, Turku Centre for Biotechnology, Turku, Finland), ImageJ (NIH, Bethesda, MD, USA) and Origin 8.5 (Microcal, Northampton, MA, USA).
Discussion
Oncogenic signaling cascades have been identified as potential molecular targets for the treatment of different tumor entities [
2,
11,
12,
18,
28,
36,
61]. However, depending on the mutational background of the cancer cell the inhibition of pathways can induce the activation of complementary signaling cascades [
31,
34,
35,
42]. A strategy to circumvent these cross activations is the simultaneous inhibition of the other complementary signaling cascades. Especially the simultaneous inhibition of the MAPK and the PI3K/mTOR pathways, which are mutated in a multiplicity of human cancers, yielded promising results in various
in vitro and
in vivo studies [
3,
14,
23,
46,
54,
64], since these signaling cascades are known to influence proliferation, cell growth, survival and resistance to chemotherapeutics and IR [
13,
50,
60].
Although, there is evidence that simultaneous treatment with the MEK inhibitor AZD6244 and the PI3K/mTOR inhibitor NVP-BEZ235 causes synergistic effects on tumor cell proliferation and induction of apoptosis [
24,
26,
53,
56,
59], little is known about the radiation response of tumor cells after simultaneous inhibition of the MAPK and the PI3K/mTOR pathways. Therefore, this study was designed to integrate the network signaling and phenotypic data of the radiation response after simultaneous MEK and PI3K/mTOR inhibition in SNB19 and A549 cells, which differ in their mutational status of the MAPK and PI3K/mTOR signaling cascade [
30].
As shown in Fig.
1, incubation with AZD6244 or NVP-BEZ235 alone resulted in a dose dependent reduction of the proliferation in SNB19 and A549, although effects of NVP-BEZ235 were of greater magnitude in A549 cells. Since there are no known mutations in the PI3K/mTOR signaling cascade in A549 cells [
30], it is likely, that the different tumor entity is the main reason for the different sensitivity towards dual PI3K/mTOR inhibition. Combining NVP-BEZ235 with AZD6244 basically resulted in the same proliferation rates as incubation with NVP-BEZ235, implicating that NVP-BEZ235 is the more effective inhibitor.
To elucidate possible reasons for the observed anti-proliferative effects, we analyzed the expression levels of certain key players of the MAPK and PI3K/mTOR signaling cascades after pathway perturbation with AZD6244 and NVP-BEZ235. As expected and in accordance with published results for different tumor entities, the treatment with AZD6244 and NVP-BEZ235 alone resulted in inhibition of the MAPK and PI3K/mTOR signaling cascade, respectively, as confirmed by reduced expression of p-Erk, p-Akt, p-S6 and p-4E-BP1 (Fig.
2). However, we also observed the induction of feedback loops by the two inhibitors, indicated by elevated levels of MEK1/2 (after MEK inhibition) and Akt phosphorylation (after PI3K/mTOR inhibition) after prolonged incubation with the inhibitors, which has already been reported for NVP-BEZ235 in other glioblastoma cell lines [
38,
41,
44]. Noteworthy, extended incubation with NVP-BEZ235 also resulted in reduced Raf-1 expression in both cell lines (Fig.
2a), indicating a crosstalk between the two signaling cascades, as depicted in our putative signaling diagram (Fig.
3) and as already published for other cell lines [
3,
23,
46].
Apart from the effects on proliferation and the signaling cascades we also assessed the clonogenic ability of SNB19 and A549 cells after IR and incubation with the inhibitors (Fig.
4). In both cell lines AZD6244 caused a radiosensitization, as reported in other studies [
9,
10,
62], although only to a moderate extent. A treatment with NVP-BEZ235 resulted in a more profound radiosensitization in both cell lines, indicating that NVP-BEZ235 is the drug which yielded the greater cytotoxic effects, when combined with IR. The NVP-BEZ235 mediated radiosensitization also is in accordance with published data [
20,
21,
51,
73]. However, the fact, that in both cell lines combining AZD6244 and NVP-BEZ235 yielded the same result as the dual PI3K/mTOR inhibitor alone, implies that no synergistic or additive effects occurred in SNB19 and A549 cells in terms of radiosensitivity. This is in contrast to previous published results of combining MAPK and PI3K/mTOR pathway inhibitors using a MEK and an Akt inhibitor in pancreatic cancer cells [
67], which again confirms that the results of pathway perturbations highly rely on the inhibitors and the genetic background of the treated cells.
To further assess the phenotypic effects in the two tested cell lines after signaling cascade inhibition, we analyzed the cell cycle phase distributions. The incubation with AZD6244 or NVP-BEZ235 caused cell cycle arrests in the G1-phase in both cell lines, as shown in Fig.
5 and Additional file
3. The combination of both inhibitors resulted in an even more profound arrest in the G1-phase in SNB19 cells, whereas no additive or synergistic effects were observed in A549 cells, which is most likely due to the extensive cytostatic effect of NVP-BEZ235 in this cell line. Our Western blot data of the cell cycle related proteins further confirm the flow cytometric data (compare Fig.
6). A treatment with both inhibitors simultaneously resulted in greatest reduction of p-Rb in SNB19 cells, which is an indicator for a blockade at the G1/S-transition checkpoint [
33,
45,
65], whereas in A549 cells treatment with NVP-BEZ235 alone already reduced p-Rb expression levels to a maximum. This enhanced cell cycle arrest in the SNB19 cell line after combined MEK and PI3K/mTOR inhibition, indicates an additive or synergistic anti-proliferative effect of AZD6244 and NVP-BEZ235, which has not been shown for glioblastoma cell lines yet, indicating a therapeutically relevant potential of combining these two inhibitors in this tumor entity.
Combining IR and AZD6244 increased the proportion of G1-phase cells in both cell lines, whereas the combination of NVP-BEZ235 and IR resulted in elevated levels of G2/M-phase cells, as shown previously for other cell lines [
37,
38]. A simultaneous inhibition of the MAPK and the PI3K/mTOR signaling cascades in irradiated cells resulted in mixed phenotypic effects as observed for the combination of IR with each inhibitor alone. Apparently the two inhibitors are somewhat counteracting each other, which might be a reason for the lack of synergy in terms of radiosensitization.
To further elucidate the effects of MEK and PI3K/mTOR inhibition, we assessed the induction of apoptosis and autophagy. As shown in Fig.
7, incubation with AZD6244 had no relevant effects on apoptosis and autophagy in SNB19 cells, whereas treatment with NVP-BEZ235 caused a slight induction of autophagy 24 h after IR. Combining both inhibitors increased cleaved PARP levels slightly, indicating apoptosis and validating a synergistic effect of the two drugs 48 h after IR in the glioblastoma cell line. However, IR of SNB19 cells increased the hypodiploid fraction and the expression level of cleaved PARP to a greater extent. Most strikingly, this radiation induced apoptosis was slightly enhanced, when irradiated cells were treated with both inhibitors simultaneously.
This is in contrast to the A549 cell line, indicating a cell line specific effect: Although the highest levels of cleaved PARP in unirradiated cells were observed, when cells were treated with both inhibitors, this was not true for irradiated A549 cells. When A549 cells were irradiated interestingly the highest rate of cleaved PARP was detected in cells treated with NVP-BEZ235 solely. Simultaneous treatment of irradiated cells with AZD6244 and NVP-BEZ235 resulted in lower levels of cleaved PARP (Fig.
7), indicating less induction of apoptosis.
Another difference between the cell lines in their response to AZD6244 and NVP-BEZ235 was observed in autophagy. When SNB19 cells were treated with NVP-BEZ235 only a slight induction of autophagy was observed 24 h after IR, as validated by a slight increase of LC3-II. However, A549 cells revealed a complete depletion of LC3-I but no increase of LC3-II. This can be due to the fact, that within the autophagic process the initially increased LC3-II levels are degraded within a few hours after induction of autophagy [
48]. This might indicate that the autophagic process is much faster in the lung carcinoma cell line. The induction of autophagy by NVP-BEZ235 was already reported for breast cancer cells lines and is of utmost interest [
37], since the role of autophagy in cancer is currently highly discussed [
29,
39,
57]. Furthermore, several research groups demonstrated, that inhibitors of the autophagic flux, such as bafilomycin A or chloroquine, can sensitize cancer cells to IR [
4,
72]. Therefore, this NVP-BEZ235-mediated induction of autophagy might be exploited to further enhance the radiosensitization with autophagy inhibitors.
Conclusions
In this study we clearly demonstrated, that our novel approach of combining AZD6244 and NVP-BEZ235 with IR resulted in pathway perturbations and cell line specific effects. In SNB19 cells mainly synergistic cytostatic effects were observed after simultaneous treatment with AZD6244 and NVP-BEZ235, whereas the combination of the same drugs induced apoptosis to a much greater extent in A549 cells. However, apart from the synergistic effects in unirradiated cells, our study also clearly shows, that there are no additive effects in terms of radiosensitivity, when the two radiosensitizers are used in combination.
One major question, which arises from the data presented in this study, is, if the different effects of the combined treatment with AZD6244 and NVP-BEZ235 in the glioblastoma SNB19 and lung carcinoma A549 cells rely on the different mutational background (SNB19 cells are TP53 and PTEN mutated, whereas A549 cells are KRAS mutated) or on the different tumor entity. A systematic approach addressing this question will be the subject of forthcoming in vitro and in vivo experiments, improving our knowledge about how specific mutations and the cancer cell origin affect the impact of inhibiting the MAPK and the PI3K/mTOR pathways. The data generated in these forthcoming experiments will help to improve our putative signaling network, which we presented in this study, ultimately facilitating valid predictions of how the effects of an inhibitor depend on certain mutations and/or the tumor cell origin. This will pave the way for a personalized therapy regimen, which is based on the genetic background of an individual cancer, in order to enhance therapy outcome.
Competing interests
The authors declare that they have no competing interests.
Authors’ contributions
SK, MF and CSD conceived the study. SK and CSD designed and coordinated the experiments and analyzed the data. SK wrote the first draft of the manuscript. SK, MF and CSD contributed to the writing of the manuscript. SK, MF and CSD developed the structures and arguments for the paper, made critical revisions and approved the final versions. All authors read and approved the final manuscript.