Inhibitors of the PI3-kinase/Akt pathway induce mitotic catastrophe in non-small cell lung cancer cells

Authors


Abstract

Non-small cell lung cancer cells (NSCLC) are more resistant to anticancer treatment as compared with other types of cancer cells. Recently (Hemström et al., Exp Cell Res 2005;305:200–13) we showed that apoptosis of U1810 NSCLC cells induced by the staurosporine analog PKC 412 correlated with inhibition of Akt and ERK1/2, suggesting the involvement of these kinases in cell survival. Here we investigated the contribution of the PI3-kinase/Akt and MEK/ERK pathways to survival of NSCLC cells. The two signaling pathways were studied by using different combinations of the PI3-kinase inhibitors LY-294002 and wortmannin, the Akt activator Ro 31-8220, the MEK inhibitor PD 98059 and PKC 412. PI3-kinase inhibitors induced apoptosis-like death in U1810 cells. H157 cells in general were relatively resistant to PI3 kinase/Akt inhibitors yet these compounds sensitized cells to the DNA-damaging drug VP-16, while Ro 31-8220 could not. PD 98059 only had a sensitizing effect on H157 cells when combined with PI3-kinase inhibition and VP-16. Morphological data indicated that LY-294002 and PKC 412 induced cell death at anaphase and metaphase, respectively, suggesting death by mitotic catastrophe. Analyzes of cells blocked in G2/M-phase by nocodazol revealed that LY-294002 increased, while PKC 412 decreased histone H3 phosphorylation, suggesting that LY-294002 allowed, while PKC 412 inhibited cells to leave M-phase. Flow cytometric analysis of cell cycle distribution demonstrated that LY-294002 allowed cells to leave G2/M phase, while PKC 412 inhibited cytokinesis, resulting in formation of multinucleated cells. These results indicate that sensitization of NSCLC cells by PI3-kinase inhibition involves interplay between cell cycle regulation, mitotic catastrophe and apoptosis. © 2006 Wiley-Liss, Inc.

Cell lines derived from non-small cell lung carcinoma (NSCLC) are most often characterized by cellular resistance towards anticancer drugs and radiation.1 Efficiency of action of DNA-damaging drugs used in anticancer therapy depends on the successful ability to induce growth arrest and to activate the cell death machinery.2 Apoptosis induced by DNA damage is typically associated with activation of a family of proteases, the caspases, as a result of a sequence of mitochondria-mediated events. The caspases cleave many proteins, eventually leading to biochemical and morphological apoptosis-specific changes. In addition to activation of the caspases, mediated by release of cytochrome c, several proteins, such as apoptosis inducing factor (AIF) and endonuclease G are also released from mitochondria, translocate to the nuclei and induce chromatin condensation and cell death independently of caspases.

Cell cycle regulation is also an important determinant for efficiency of cell death in response to DNA damage. For example, topoisomerase inhibitors are suggested to be particularly effective in S-phase when DNA replication occurs,3 while irradiated cells are most sensitive at G2/M phase.4 DNA-damaged cells are removed by cell death following activation of the DNA damage checkpoints in G1 and G2-phases of the cell cycle.5 There are many mechanisms for cells being resistant to DNA damage.6 For example, resistant lung cancer cells have an increased activity of DNA-dependent protein kinase (DNA-PK) and DNA repair, correlating with a decreased incidence of cell death, suggesting involvement of DNA-PK in tumor survival.7, 8 Aberrant signaling of other kinases, such as the Ras/MEK/ERK and PI3-kinase/Akt pathways is also contributes to cell death resistance. In NSCLC-patients phospho-Akt overexpression confers a stage-independent survival disadvantage9 and in NSCLC cell lines a high activity of Akt promotes cellular survival and resistance to radiation and chemotherapy.10 Extracellular regulated kinase ½ (ERK1/2) activation in NSCLC cells is associated with advanced tumors.11

To improve treatment efficiency of lung tumors it is necessary to find new pathways to activate cell death. One possibility is to drive the cells prematurely out of G2 phase, before the DNA-damaged cells are ready to re-enter the cell cycle. Depending on different factors the fate of a cell prematurely entering mitosis can significantly alter. In cells with an intact spindle checkpoint upon DNA damage and escape from G2 there is a delay at metaphase. Rieder and Maiato12 (and references therein) define the outcome of such a delay at metaphase of DNA-damaged cells prematurely entering metaphase as mitotic catastrophe. The fate of such a delayed mitosis may be death by apoptosis, death by necrosis, production of two or more aneuploid cells or exit from mitosis as a 4N cell ultimately leading to aneuploidy, to death from senescence or apoptosis. Lately another definition of mitotic catastrophe as an apoptotic cell death occurring either as a result of DNA damage or incomplete DNA replication upon failure to activate the G2/M checkpoints was suggested.13 Suppression of the apoptotic machinery of these cells may lead to asymmetric cell division resulting in the generation of tetraploid cells with aneuploid offspring.

Recently we showed that the protein kinase C (PKC) inhibitor PKC 412, which is an analog of the kinase inhibitor staurosporine, induces apoptosis in U1810 NSCLC cells, characterized by a drop in mitochondrial membrane potential, caspase activation and an increase in the number of cells with an apoptosis-like nuclear morphology.14 PKC 412 also sensitizes U1810 cells to treatment with etoposide and radiation. Interestingly, single treatment of cells with PKC 412 leads to the formation of tetraploid cells, indicating that PKC 412 disturbs mitosis-related signaling.15 PKC 412 treatment is also associated with a concentration-dependent decrease in the phosphorylation of Akt and of the MEK target, ERK1/2.14 The staurosporine analog Ro 31-8220 had an opposite effect on cells; exposure to this compound leads to an increase in Akt and ERK phosphorylation, and this compound cannot sensitize cells to DNA damage. Since these two compounds equally well inhibit the activity of classical PKC isoforms, we hypothesized that the cell death induced by PKC 412 might relate to inhibition of the PI3-kinase pathway and/or MEK-related signaling. To our knowledge there is no investigations performed regarding the involvement of PI3-kinase and MEK in mitotic catastrophe of lung cancer cells. In this report, using one inhibitor of MEK and three different inhibitors and one activator of PI3-kinase signaling, the involvement of these pathways in mitotic catastrophe and apoptosis in two NSCLC cell lines was investigated.

Abbreviations:

AIF, apoptosis inducing factor; Akt, serine/threonine kinase, or protein kinase B; DNA-PK, DNA-dependent protein kinase; ERK1/2, extracellular regulated kinase ½; MEK, MAPK (mitogen-activated protein kinase)/ERK kinase; PI3-kinase, phosphatidylinositol 3-kinase; PKC, protein kinase c; NSCLC, nonsmall cell lung cancer.

Material and methods

Cell lines, culture conditions and treatment

U1810 and H157 cells were maintained at 37°C, 5% CO2 in RPMI 1640 medium supplemented with 10% heat-inactivated fetal calf serum, 2 mM L-glutamine, penicillin (100 U/ml) and streptomycin (100 U/ml). Cells were seeded at least 24 hr prior to treatment with PKC 412 (1–10 μM), Ro 31-8220 (1–10 μM), LY-294002 (10–30 μM), wortmannin (30 μM), PD 98059 (10–30 μM), etoposide (2.5 μM) and nocodazol (0.3 μM). All experiments were performed at ambient room temperature and at a cell density allowing exponential growth.

Antibodies and reagents

PKC 412 (CGP 41251, Novartis), Ro-31 8220 (Calbiochem), LY-294002, wortmannin and PD 98059 (all from Sigma) were dissolved in DMSO to a 10 mM concentration. Etoposide (VePeside, VP-16) was purchased in solution (20 mg/ml) from Bristol Myers Squibb. Nocodazol (Sigma) was dissolved in DMSO (0.5 mg/ml). All further dilutions were made in fresh RPMI 1640 medium. Antibodies used for western blot were anti-PARP (1:1,000, mouse monoclonal, Biomol Research Laboratories), anti-G3PDH (1:3,000, rabbit polyclonal, Trevigen), antibodies targeting phosphorylated ERK1/ERK2 (Thr202/Tyr204), Akt (ser-473), Cdc2 p34 (Tyr 15) and anti-nonphosphorylated ERK1/ERK2 and Akt (1:1,000, rabbit polyclonal, New England Biolabs), antiphospho-Histone H3 (Ser10) (1:2,000, anti-mouse monoclonal, New England Biolabs), Cdc2 p34 (1:500, mouse monoclonal, Santa Cruz); secondary antibodies (1:10,000, anti-mouse, anti-rabbit, or anti-goat peroxidase-conjugated, Pierce).

Western blot analysis

Untreated or treated cells were harvested by trypsinization and washed in PBS before addition of lysis buffer (20 mM MOPS, 50 mM β-glycerolphosphate, 50 mM sodium fluoride, 1 mM sodium vanadate, 5 mM EGTA, 2 mM EDTA, 1% NP40, 1 mM DTT, 1 mM benzamidine, 1 mM PMSF, 10 μg/ml leupeptin, 10 μg/ml aprotinin). Samples were kept on ice and pulse sonicated for 1 min before addition of loading buffer. Samples were boiled for 4 min, loaded on 15% SDS polyacrylamide gels, which were run at 130 V and proteins were transblotted onto nitrocellulose membranes for 2 hr at 100 V. Membranes were blocked for at least 1 hr in 5% nonfat milk, 0.1% Tween-20 and 0.1% NaN3 in PBS, pH 7.4 and probed with primary antibody (diluted in 1% bovine serum albumin, 0.1% Tween-20 and 0.1% NaN3 in PBS, pH 7.4) overnight. Membranes were washed (two times in PBS, once in 0.15% Tween-20 in PBS) and exposed to secondary antibody (diluted 1:10,000 in 2.5% milk in PBS) for 1 hr followed by washing (PBS 2×, 0.15% Tween-20 in PBS 1×). Protein bands were visualized by ECL or ECL plus (Amersham Biosciences) according to manufacturer's instructions. Membranes were stripped in a buffer (60 mM Tris, 2% SDS, 0.7% mercaptoethanol) at 50°C for 30 min, reblocked and restained with antibody.

DNA morphology assessment

The amount of apoptotic cells was measured quantitatively by assessing the percentage of cells with fragmented or condensed nuclei. After trypsination and washing in PBS, cells were loaded on a slide, allowed to dry and subsequently fixed in ethanol:acetone (1:1, 5 min). Smears were washed in PBS and DNA was stained with Hoechst 33342 (Molecular Probes; 1 μg/ml in PBS, 20 min). Following washing (5 min PBS 2×) slides were mounted in PBS:glycerol (1:1) and sealed. At least 350 nuclei were counted per sample.

Cell cycle analysis

For estimation of the distribution of cells in different phases of the cell cycle, FACS-analysis of cells stained with propidium iodide (PI) was performed. Upon harvesting cells were fixed in 75% of ethanol, left at 4°C for at least 2 hr and thereafter stored at −20°C until analysis. Fixed cells were washed once in PBS, stained in a PI-containing solution (50 μg propidium iodide/ml, 50 μg RNAse A/ml, 0.1 M EDTA in PBS, pH 7.4) and analyzed on a Becton Dickinson flow cytometer.

Results

Differences in apoptosis induction and ERK-phosphorylation upon treatment of U1810 and H157 cells with staurosporine analogs

Apoptosis-inducing abilities of PKC 412 and Ro 31-8220 were investigated in U1810 and H157 cells (Figs. 1a and 1b). As was previously shown,14 a 24 hr exposure of U1810 cells to PKC 412 induced a concentration-dependent increase in cell death, whereas Ro 31-8220 was a much weaker apoptosis inducer compared to PKC 412 (Fig. 1a). Upon treatment with 10 μM PKC 412 and Ro 31-8220, the amount of U1810 cells that exhibited a condensed and/or fragmented nuclei was 24% and 5%, respectively. H157 cells were resistant to both PKC 412 and Ro 31-8220. The number of cells with apoptotic nuclear morphology after 24 hr incubation with these compounds was similar to the control level of 2% (Fig. 1a). Caspases cleave many different target proteins, including the nuclear protein PARP, which is often used as a biochemical marker of apoptosis. Therefore the cleavage of PARP upon PKC 412 and Ro 31-8220 treatment was analyzed (Fig. 1b). In the U1810 cells PKC 412 induced formation of the 85 kD PARP cleavage product, whereas upon treatment with Ro 31-8220 the cleavage was not as pronounced. In the H157 cells neither PKC 412, nor Ro 31-8220 induced PARP cleavage. At later time points (48 and 72 hr) upon treatment with 10 μM PKC 412 there was a minor increase in PARP cleavage whereas Ro 31-8220 had no effect (data not shown).

Figure 1.

Apoptosis induction and kinase phosphorylation mediated by PKC 412 and Ro 31-8220. The percentage of Hoechst-stained U1810 and H157 cells exhibiting condensed and/or fragmented nuclei upon 24 hr treatment with PKC 412 or Ro 31-8220 (a). The error bars represent standard deviation. PARP cleavage following 24 hr exposure to 10 μM PKC 412 and 10 μM Ro 31-8220 was visualized by binding of an antibody to the full length PARP of 116 kD and the 85 kD cleavage product upon subjecting protein samples to western blotting (b). Effect of PKC 412 and Ro 31-8220 on the phosphorylation of Akt and ERK1/2 upon 1 hr treatment of H157 cells with PKC 412 and Ro 31-8220 (c). Following treatment protein extracts were analyzed by western blotting. G3PDH was used for control of equal loading.

Earlier,14 we showed that PKC 412 and Ro 31-8220 had opposite effects on AKT phosphorylation: PKC 412 decreased while Ro 31-8220 increased the phosphorylation level in a concentration-dependent manner. Both agents were also shown to decrease the phosphorylation of ERK, PKC 412 being the most potent inhibitor. These two agents inhibited activity of conventional PKC isoforms equally well, and therefore we hypothesized that inhibition of PI3-kinase and MEK pathways might contribute to apoptosis induced by PKC 412 in U1810 cells. Since the H157 cells were resistant to both PKC 412 and Ro 31-8220 we investigated whether differences in resistance correlate with the level of protein phosphorylation. Akt and ERK phosphorylation following a 1 hr exposure to PKC 412 and Ro 31-8220 was analyzed in H157 cells (Fig. 1c). In a concentration-dependent manner PKC 412 and Ro 31-8220 decreased and increased Akt phosphorylation, respectively. Both agents induced an increase in ERK phosphorylation, compared to control. Taken together, these results show that both U1810 and H157 cells were resistant to apoptosis induction by Ro 31-8220, whereas only U1810 cells responded to PKC 412. In both cell lines Ro 31-8220 increased, and PKC 412 decreased Akt phosphorylation, while both compounds only in H157 cells increased ERK phosphorylation.

U1810 cells were more sensitive to inhibition of the PI3-kinase pathway than H157 cells, while both cell lines exhibited resistance to MEK inhibition

As shown in Figure 1, there are differences between the two cell lines in phosphorylation and in the level of apoptosis in response to treatment with the kinase inhibitors PKC 412 and Ro 31-8220. Taking into consideration the reports showing the involvement of the PI3-kinase and the MEK pathways in treatment resistance, we decided to further clarify the role of these pathways in cell death of U1810 and H157 cells by experimentally exploring three different possibilities: (1) The higher sensitivity of U1810 cells to PKC 412 treatment might be explained by survival dependency of the PI3-kinase/Akt pathway, whereas the survival of H157 cells might depend upon other mechanism(s). (2) Both cell lines are equally sensitive to treatment in the case of simultaneous inhibition of PI3-kinase and the MEK/ERK pathways. (3) Cell death is dependent upon inhibition of the MEK/ERK pathway only. To explore these three possibilities apoptosis upon treatment with the PI3-kinase/Akt inhibitors LY-294002 and wortmannin and the MEK inhibitor PD 98059 alone, or PI3-kinase inhibition combined with PD 98059 were compared. For evaluation of the involvement of the MEK/ERK pathway in survival of H157 cells, experiments using the Akt inhibitor/ERK activator PKC 412 in combination with PD 98059 were also included.

Apoptosis and phosphorylation-related effects upon treatment with PI3-kinase inhibitors LY-294002 and wortmannin and the MEK inhibitor PD 98059 were analyzed in both cell lines. The phosphorylation level of Akt and ERK1/2, downstream targets of PI3-kinase and MEK, respectively, was visualized by western blot technique upon one-hour treatment of H157 (Fig. 2a) and U1810 (Fig. 2b) cells. Akt phosphorylation in H157 cells was decreased in response to wortmannin and LY-294002 in a concentration-dependent manner, although the effect of wortmannin was stronger. Upon exposure of H157 cells to PD 98059, ERK phosphorylation was also significantly decreased. In U1810 cells both LY-294002 and wortmannin significantly inhibited Akt phosphorylation and treatment of U1810 cells with PD 98059 and LY-294002 decreased ERK phosphorylation in a concentration-dependent manner, PD 98059 being the most potent inhibitor.

Figure 2.

Effects of MEK inhibitor PD 98059 (PD) and/or the PI3-kinase inhibitors LY-294002 (Ly) and wortmannin (Wm) on Akt and ERK1/2 phosphorylation and induction of apoptotic nuclear morphology. Akt phosphorylation upon a 1 hr treatment of H157 (a) and U1810 (b) cells with Ly, PD or Wm, and of ERK 1/2 phosphorylation following exposure to Ly or PD. Protein extracts were subjected to Western blot analysis. G3PDH was used for control of equal loading. The percentage of H157 (c) and U1810 (d) cells exhibiting condensed and/or fragmented nuclei upon treatment with PD and/or Ly and Wm. Following a 24 hr treatment cells were fixed and stained with Hoechst. The error bars show standard deviation.

Analysis of abilities of LY-294002, wortmannin and PD 98059, individually or in combination, to induce apoptosis in H157 cells revealed that LY-294002 and wortmannin induced appearance of condensed and/or fragmented nuclei in 4% and 6% of cells, respectively, compared to a control level of below 1% (Fig. 2c). Upon single treatment with PD 98059, 2% of cells were classified as apoptotic, and PD 98059 could not increase the apoptotic response to wortmannin (4%) and PKC 412 (3%). Cotreatment with PD 98059 and LY-294002 increased the percentage of cells with fragmented and/or condensed nuclei to 8%, compared to the 4% of apoptotic cells obtained upon treatment with LY-294002 alone. Similar experiments performed in U1810 cells demonstrated that the PI3-kinase inhibitors could significantly increase the level of cells with fragmented and/or condensed nuclei (Fig. 2d). Thus, LY-294002 and wortmannin increased the percentage of cells with apoptotic nuclear morphology up to 22% and 28%, respectively, compared to a control value of 2%. Upon single treatment with PD 98059, only 3% of cells were classified as apoptotic. Combinatory treatment of PD 98059 with LY-294002 or wortmannin was not able to further increase the percentage of apoptotic cells as compared to action of a single agent.

Since LY-294002 inhibited both Akt and ERK phosphorylation, examination of the influence on apoptosis of simultaneous inhibition of PI3-kinase and MEK pathway was also performed. Low concentration of LY-294002 (10 μM) did not decrease ERK phosphorylation and combination of this drug with PD 98059 was unable to increase the percentage of apoptotic cells, compared to the effect of LY-294002 alone. Thus, the U1810 cells were more sensitive to inhibition of the PI3-kinase pathway, compared to H157 cells. MEK inhibition did not significantly influence apoptosis induction in either of the cell lines analyzed, except for a slight increase in apoptosis upon co-treatment of H157 cells with PD 98059 and LY-294002.

LY-294002 induced aberrant mitosis with apoptotic features

Upon treatment with LY-294002 in addition to a small increase in the number of cells with condensed apoptotic-like nuclear morphology (Fig. 2) appearance of a fraction of H157 cells with another type of nuclear changes was observed. These cells resembled anaphase cells, yet exhibiting condensed or partially condensed DNA (Fig. 3a, upper and middle panel). Upon wortmannin treatment some of H157 cells also exhibited a similar morphology, whereas PKC 412 did not induce accumulation of these cells (data not shown). The typical apoptotic morphology of U1810 cells dying due to treatment with LY-294002 is presented in Figure 3a (lower panel). Treatment of H157 cells with LY-294002 induced aberrant mitosis with partial nuclear condensation in 15% of cells, compared to a control mitosis level of 4% (Fig. 3b). MEK-inhibition did not increase the level of mitosis-like cells either when used alone (3%), or in combination with LY-294002 (13%). A 24 hr treatment with LY-294002 resulted in a small increase in PARP cleavage, which was further enhanced upon addition of PD 98059 (Fig 3c). At 48 hr incubation with LY-294002 the level of PARP cleavage was even higher, irrespective of presence or absence of PD 98059 in the incubation medium. At this time point the amount of mitosis-like cells was decreased to control level and the percentage of cells with condensed and/or fragmented nuclei was increased (data not shown). Taken together these data showed that in H157 cells LY-294002 induced an aberrant mitosis with apoptotic features, leading with time to clear apoptosis-related changes, while in the U1810 cells LY-294002 induced a faster cell death characterized by nuclear condensation and fragmentation with absence of mitosis-like features.

Figure 3.

LY-294002-induced nuclear changes. Typical nuclear morphology of Hoechst-stained untreated and treated for 24 hr with LY-294002 (Ly) H157 and U1810 cells (a). The percentage of cells with mitotic catastrophe nuclear morphology (b) following a 24 hr exposure to Ly and/or PD 98059 (PD) was estimated by counting of Hoechst-stained cells. The error bars represent standard deviation. PARP cleavage (c) upon 24 hr and 48 hr treatment with 30 μM Ly alone, or in combination with 30 μM PD. G3PDH was used as a control for equal loading.

LY-294002 and wortmannin but not PD 98059 sensitized NSCLC cells to DNA-damage

Treatment with LY-294002 resulted in mitotic catastrophe of H157 cells. As mentioned above, DNA damage due to incomplete replication or due to the action of some chemical compounds is a prerequisite for mitotic catastrophe to occur. To understand if the combination of PI3-kinase inhibition and DNA damage could further enhance cell death, H157 cells were exposed to combinatory treatment with PI3-kinase inhibitors and the topoisomerase II inhibitor VP-16 (Fig 4). Cotreatment of H157 cells with LY-294002 and VP-16 increased the amount of cells with apoptosis-like condensed and/or fragmented nuclei, compared to separate use of these compounds (Fig 4a). LY-294002 in combination with VP-16 induced appearance of 18% of apoptotic cells, compared to 1% as a result of single VP-16 treatment. The highest number of dying cells (45%) was seen when VP-16 was combined with LY-294002 and PD 98059. The latter alone could not alter the effect of VP-16. Wortmannin also sensitized cells to DNA damage, although the percentage of dying cells was not as high as that after LY-294002 exposure. Wortmannin in combination with VP-16 increased the amount of apoptotic cells up to 12%, and cotreatment of cells with wortmannin, PD 98059 and VP-16 further increased death to 21%.

Figure 4.

Cell death-related changes upon exposure of cells to different combinations of the kinase inhibitors LY-294002 (Ly, 30 μM), wortmannin (Wm, 30 μM), PKC 412 (PKC, 1 μM), PD 98059 (PD, 30 μM) and the DNA-damaging drug VP-16 (2.5 μM). The percentage of cells exhibiting an apoptosis-like nuclear morphology with condensed and/or fragmented nuclei was assessed in H157 (a) and U1810 (b) cells. Following a 24 hr exposure cells were stained with Hoechst and at least 350 cells per sample were counted. The error bars show standard deviation. PARP processing in H157 (c) and U1810 (d) cells upon 24 hr treatments was analyzed by Western blot. G3PDH is included for control of equal loading.

In U1810 cells no increase in the percentage of cells with condensed and/or fragmented nuclei as a result of combined treatment of either LY-294002 with VP-16 or PD 98059 with VP-16 was observed (Fig. 4b). The MEK inhibitor alone could not sensitize U1810 cells to VP-16, whereas the combination of VP-16 with wortmannin, or with wortmannin and PD 98059, resulted in 39% and 47% of dying cells, respectively (Fig. 4b).

Analysis of PARP cleavage demonstrated that in H157 cells upon treatment with VP-16 alone or in combination with LY-294002, wortmannin and PD 98059, proteolysis of this protein correlated with the respective changes in the number of cells with condensed nuclei (Figs. 4a and 4c). In the U1810 cells the level of PARP cleavage also correlated with the changes in the percentage of condensed and/or fragmented nuclei, except upon co-treatment with VP-16 and LY-294002 (Fig. 4d). PARP cleavage upon treatment with PKC 412 in combination with VP-16 was used as a positive control (Fig. 4d). Thus, a combination of wortmannin or LY-294002 sensitized H157 and U1810 cells to DNA damage; the highest sensitizing effect was observed in the H157 cell line. Based on nuclear morphology and PARP cleavage the cell death in these cases was apoptosis-like. PD 98059 alone could not sensitize cells to DNA damage, but MEK inhibition in the H157 cells further increased the apoptosis induced by DNA damage in combination with PI3-kinase inhibition.

Combination of PKC 412 with VP-16 induced mitotic catastrophe with apoptotic features

Since PKC 412 sensitizes U1810 cells to DNA damage14 and PI3-kinase inhibitors could sensitize H157 cells to VP-16 as well (Figs. 4a and 4c), sensitizing effect of PKC 412 to DNA damage was also investigated in H157 cells. Typical nuclear morphology upon Hoechst staining of untreated or exposed to PKC 412 in combination with VP-16 cells as well as cells treated with LY-294002 in combination with PD 98059 and VP-16 is presented in Figure 5a. PKC 412 in combination with VP-16 induced nuclear condensation (Fig. 5a, middle panel), but the morphology was not typical for cells dying by apoptosis (a morphology that is exemplified in the lower panel by H157 cells treated with a combination of LY-294002, VP-16 and PD 98059). Instead, some of the dying cells upon combined treatment with PKC 412 and VP-16 exhibited a morphology resembling cells at metaphase (middle panel, lower right corner), while others were characterized by a larger ring formation of condensed chromatin. A fraction of cells with a similar nuclear morphology also appeared upon treatment of U1810 cells with PKC 412 or PKC 412 in combination with VP-16 (data not shown).

Figure 5.

Nuclear changes in H157 cells upon treatment with 10 μM PKC 412 (PKC) alone or in combination with 30 μM PD 98059 (PD) and/or 2.5 μM VP-16. Hoechst-stained cells show the typical morphological changes upon treatment with either a combination of PKC 412 and VP-16 or 30 μM LY 294002 (Ly), PD and VP-16 (a). The percentage of Hoechst stained cells exhibiting condensed chromatin (b). Data from experiments when cells were treated with 10 μM Ro 31-8220 (Ro) in combination with VP-16 and PD are also included. The error bars show standard deviation. PARP cleavage in cells treated with PKC or Ro in different combinations with PD and VP-16 (c). G3PDH is included as a control for equal loading.

The percentage of H157 cells with condensed chromatin upon treatment with PKC 412 in combination with VP-16 increased to 19% as compared to single treatment with these compounds (6% and 2%, respectively)(Fig. 5b). PD 98059 tended to, but could not significantly increase the percentage of dying cells when combined with PKC 412 and VP-16 (24%). Since Ro 31-8220 increased Akt phosphorylation this agent was used as a negative control for PI3-kinase inhibition. Ro 31-8220 in combination with VP-16 and PD 98059 induced nuclear condensation in less than 4% cells (Fig. 5b). To confirm morphological data PARP cleavage was analyzed in these cells. The level of PARP cleavage (Fig. 5c) at 24 hr incubation with VP-16 in combinations with PKC 412, PD 98059 and Ro 31-8220 correlated with the differences in the number of cells with condensed chromatin upon identical treatments (Fig. 4b). Taken together these data show a sensitization effect of PKC 412 to DNA damage in H157 cells, resulting in cell death with apoptotic features, yet with a mitotic catastrophe-like morphology. MEK inhibition could not significantly increase apoptosis.

PKC 412 treatment inhibited cytokinesis and increased the percentage of metaphase cells

Induction of apoptosis with nontypical nuclear morphology in response to combination of PKC 412 and VP-16 (Fig. 5) tempted us to more precisely investigate morphological mitosis-related changes in both cell lines upon treatment with PKC 412. FACS analysis of cells with PI-stained DNA revealed in these cell lines an increase in the fraction of G2 cells and formation of tetraploid cells in response to PKC 412 (Fig. 6a). However, as a result of combination of PKC 412 with VP-16 no multinucleate cells were observed. Instead there was an increase in G2 and S-phase cells resembling the data obtained after treatment with VP-16 alone. PKC 412 treatment of U1810 cells mediated a temporary increase in the percentage of metaphase cells (Figs. 6b and 6c). A typical metaphase cell upon treatment of U1810 cells with PKC 412 is presented in Figure 6b. Following 6 hr incubation the amount of metaphase cells in untreated samples was below 1%, while in response to PKC 412 a four-fold increase was documented (Fig. 6c). After 12 hr the level of metaphase cells was reduced to control level. This decrease in the number of metaphase cells coincided with an increase in apoptosis to 9% at 12 hr as compared to 3% at 6 hr (Fig. 6d). Thus, these data suggested that PKC 412 when used alone inhibited cytokinesis leading to formation of polyploid cells. However, as a result of combination of PKC 412 with VP-16 polyploid cells disappeared, and instead cells were detected at S and G2 phases, as was also the case for VP-16 only-treated cells. At early time points upon addition of PKC 412 there was an increase in the number of U1810 cells at metaphase. With time these metaphase cells disappeared, coinciding with an increase in apoptosis, suggesting a link between mitosis-related changes and apoptosis.

Figure 6.

Cell cycle-related changes mediated by PKC 412. The different phases of the cell cycle (a) visualized by FACS analysis upon PI-staining of cells treated for 24 hr with PKC 412, 2.5 μM VP-16 or PKC 412 in combination with VP-16. H157 and U1810 cells were treated with 10 μM PKC 412 and 1 μM of PKC 412, respectively. Typical metaphase cell (b) upon treatment of U1810 cells with PKC 412. Samples were stained with Hoechst. Changes in the level of metaphase-cells upon 6 hr and 12 hr treatment of U1810 cells with 10 μM PKC 412 (c). Time-dependent increase in the percentage of Hoechst-stained U1810 cells exhibiting a condensed and/or fragmented nucleus (d). The error bars in c and d show standard deviation.

LY-294002 increased G2/M-phase-related histone H3 phosphorylation whereas PKC 412 decreased this event

As shown above, PKC 412 inhibited cytokinesis and induced metaphase-related apoptosis, which is different from LY-294002-induced mitotic catastrophe, when, according to morphological data (Fig. 3), cell death occurred in anaphase. To further investigate the role of mitosis in death induced by LY-294002 FACS analysis of cell cycle was performed and revealed predominant accumulation of both H157 and U1810 cells in G1 phase (Fig. 7a). Upon treatment with wortmannin cells responded in a similar manner as upon LY-294002 treatment, while Ro 31-8220 slightly increased the amount of cells in G2 phase (data not shown). Since histone H3 phosphorylation is considered to be one of the events essential for allowing cells to leave the spindle check point,16 western blot technique was used to analyze phosphorylation of this protein in both cell lines. To perform these experiments cells were treated with nocodazol that interferes with microtubule formation and does not permit cells to leave metaphase. Upon treatment with nocodazol the amount of cells in G2/M phase was increased in both cell lines (Fig. 7b). Nocodazol treatment increased the level of histone H3 phosphorylation in both cell lines as compared to control (Fig. 7c). Treatment with PKC 412 resulted in decrease of this phosphorylation while incubation with LY-294002 increased histone H3 phosphorylation in nocodazol-treated and untreated H157 cells. In U1810 cells, upon a 1 hr treatment with LY-294002 there was no significant increase in histone H3 phosphorylation in samples from untreated or nocodazol-treated cells. However, upon a five-hour treatment of U1810 cells with LY-294002 there was a clear increase in the level of histone H3 phosphorylation (data not shown). Thus, combined these data showed that LY-294002 allowed cells to leave G2/M-phase, which is accompanied with increased G2/M-phase-related phosphorylation of histone H3 (a marker for M-phase cells and an essential signal for cells to leave the spindle checkpoint), whereas PKC 412 decreased this phosphorylation.

Figure 7.

Cell cycle-related changes mediated by kinase inhibition. Representative FACS data of PI-stained H157 and U1810 cells subjected to 24 hr treatment with 30 μM LY-294002 (Ly) (a). Increase in the number of G2/M cells of nocodazol-treated U1810 and H157 cells (b). Cells were treated for 22 hr with 0.3 μM nocodazol (noc) and thereafter stained with PI and analyzed by FACS. Western blot data show histone H3 phosphorylation in cells treated with 30 μM Ly or PKC 412 (PKC) (c). H157 and U1810 cells were treated with 10 μM PKC 412 and 1 μM of PKC 412, respectively. When indicated, cells were pretreated with 0.3 μM nocodasol (noc), washed 3× 5 min with PBS, and thereafter new medium was added. In some experiments kinase inhibitor was also included in treatment. G3PDH is shown for control of equal loading.

Incubation of VP-16-treated NSCLC cells with LY-294002, wortmannin or PKC 412 resulted in a fast PARP cleavage

Earlier14 we showed that the order of exposure of cells to chemotherapy agents increased the treatment efficiency. Post-treatment of VP-16- or radiation-exposed cells with PKC 412 being inferior to co-/or pre-treatment of cells. Here we demonstrate by using western blot technique that LY-294002 and wortmannin could induce apoptosis-related cleavage of PARP by being added after DNA-damage. Upon a four-hour exposure of VP-16-treated U1810 and H157 cells to LY-294002, wortmannin or PKC 412 there was a clear increase in PARP cleavage when compared to controls (Fig. 8). Thus, inhibition of the PI3-kinase/Akt pathway rapidly induced apoptosis-related PARP cleavage in DNA-damaged cells.

Figure 8.

PARP-cleavage upon post-treatment of VP-16-exposed cells with PI3-kinase inhibitors. Cells were treated for 28 hr with either 2.5 μM VP-16 alone, or for 24 hr with VP-16 and four additional hours in combination with 30 μM LY-294002 (Ly), 30 μM wortmannin (Wm), or 1 μM (U1810)/10 μM (H157) PKC 412 (PKC). G3PDH is shown for control of equal loading.

Discussion

Resistance to available treatments characterizes NSCLC and new strategies have to be developed to combat this type of malignant cells. In recent years there has been a growing interest in the attempts to use kinase inhibitors for treatment of NSCLC. Kinases are often involved in complex systems, including regulation of life and death. In order to enable effective use of kinase inhibitors in treatment of NSCLC the understanding of in which combinations and why these drugs influence tumor growth is required. It has been proposed that mechanisms associated with regulation of the cell division machinery also have to be activated to enable cell death by apoptosis.17 Even though this statement has not been proven, there are numerous evidences supporting the close interplay between the cell cycle regulatory mechanisms and the apoptotic machinery.18, 19, 20 PI3-kinase activity is known to be involved in maintaining survival of NSCLC cells. Here, to our knowledge for the first time, we demonstrated that PI3-kinase inhibition in NCSLC results in cell death with characteristics of mitotic catastrophe. As mentioned above, there is no consensus in the definition of mitotic catastrophe as a type of cell death; the only characteristic that unifies these cells is a failure in mitosis due to premature entering into M-phase, which in some situations might be preceded by DNA damage.

Investigation of nuclear morphology of H157 cells clearly indicated appearance of mitotic catastrophe following PI3-kinase inhibition. Upon LY-294002 treatment and to a small extent also upon exposure to wortmannin (data not shown) nuclei of H157 cells were characterized by morphology resembling that of cells in anaphase in combination with the characteristic condensation and fragmentation of cells undergoing apoptosis. Eventually, at 48 hr the mitosis/apoptosis-like cells disappeared, coinciding with an increase in the number of cells with apoptosis-like condensed nuclei and PARP cleavage suggesting that cell death was initiated by some disruption at anaphase, which led to activation of the apoptotic machinery. Co-treatment with PKC 412 and etoposide induced mitotic catastrophe, which was supported by observation of correlation between an increase in PARP cleavage and an increase in a number of cells with a nuclear metaphase-like morphology. Most U1810 cells exhibited a more classical apoptosis-like nuclear morphology compared to the formation of mitosis-like nuclei with partial apoptotic nuclear condensation of H157 cells. However, at early time points, PKC 412 induced an increase in the number of metaphase U1810 cells. The percentage of metaphase cells decreased at 12 hr, a time point when the amount of cells with apoptotic morphology increased, suggesting a mitosis-related apoptosis.

Biochemical data also suggested appearance of mitosis-related changes upon inhibition of the PI3-kinase/Akt pathway. Histone H3 phosphorylation has been proposed as one of the signals ensuring that the cell is ready to leave metaphase.16 The nuclear morphology indicated that PKC 412 induced cell death in metaphase, while upon LY-294002 treatment mitosis was interrupted at anaphase. These morphological changes of the H157 and U1810 cells correlated well with changes in phosphorylation of histone H3; PKC 412 decreased while LY-294002 increased histone H3 Ser-10 phosphorylation. In Drosophila reduced histone H3 phosphorylation is associated with a failure in cytokinesis.21 The increase in the number of multinucleate cells upon PKC 412 treatment suggests that PKC 412-induced reduction in histone H3 phosphorylation was also associated with an aberrant cytokinesis. LY-294002 decreased the amount of cells in G2, and thus, in this case the increase in histone H3 phosphorylation indicated that cells were allowed, or even stimulated, to leave metaphase.

Even though the nuclear morphology of U1810 cells treated with LY-294002 did not show characteristics of mitosis, the reason for onset of death might be the same in both cell lines; it could be that U1810 cells, once they are triggered to die, are more effective in activating mechanisms leading to characteristics of apoptosis, such as nuclear condensation. Upon LY-294002 treatment also a higher number of U1810 cells were triggered to die compared to H157 cells. It has been suggested that a rather small increase in the missegregation of chromosomes might be incompatible with viability of cells with chromosomal instability, and therefore agents that trigger the spindle checkpoint are believed to have some degree of selectivity for aneuploid cells.22 When analyzing the cell cycle of U1810 cells a sub-G1 population appeared to not correlate with apoptosis, indicating aneuploidity. Thus, since PI3-kinase inhibition leads to aberrant mitosis, one possible reason for the higher sensitivity of U1810 cells is the aneuploidity of this cell line.

MEK1 inhibition could not induce apoptosis, either when used alone, or when combined with etoposide, suggesting that in both cell lines activity of the MEK1 pathway was not a primary cause of resistance. However, in the H157 cells PD 98059 sensitized cells to apoptosis induced by LY-294002 alone or in combination with etoposide, whereas sensitization was not as clear when MEK1 inhibition was combined with PKC 412 and etoposide. MEK1 has been shown to interact with spindle checkpoint proteins,23 and the activity of MEK1 is reduced upon anaphase entry.24 If LY-294002-induced cell death was dependent on the ability of cells to leave metaphase, MEK1 inhibition might have conferred an additional signal to LY-294002-treated H157 cells to leave metaphase, while upon PKC 412 treatment escape from metaphase was potently inhibited and therefore the decrease in the MEK1 signal had no effect. Interestingly, MEK inhibition also sensitizes H157 cells to death induced by the microtubule stabilizing drug paclitaxel that blocks cell cycle progression in mitosis.25 The reason why PD 98059 could not sensitize U1810 cells is not clear but might be explained by cellular differences in the level of checkpoint aberration. The two cell lines also seem to have differences in ERK regulation since PKC 412 increases the phosphorylation of ERK in H157 cells and decreases it in U1810 cells. Since in H157 cells the lower concentration of PKC 412 induces the highest increase in ERK phosphorylation it is likely that PKC 412 with different potency affects various signalling pathways that interfere with ERK. The understanding of the complexity of ERK regulation in NSCLC cells requires further investigations.

It is not always clear if stimulation of the spindle checkpoint is favorable or inhibitory for cell death induction. Silencing of the spindle checkpoint protein BubR1 sensitized breast and ovarian cells to death induced by paclitaxel and nocodazol,26 whereas in another study intact spindle function of BubR1 and Mad2 was favorable for cell death induction by DNA-damaging agents.27 The multinucleated cells induced by PKC 412 disappeared when PKC 412 was combined with etoposide, and the disappearance of these multinucleated cells correlated with an increase in the number of apoptotic metaphase cells, indicating that the DNA damage activated apoptosis at the spindle checkpoint. These data show that a block at the spindle checkpoint might be beneficial for cell death induction if the cell is capable to induce apoptosis due to this block. However, if the chances of survival upon a premature anaphase are low, a bypass of the spindle checkpoint might also increase cell death, as was shown with LY-294002.

All three PI3-kinase/Akt inhibitors, PKC 412, LY-294002 and wortmannin, sensitized NSCLC cells to DNA damage, whereas the PI3-kinase activator Ro 31-822028 did not, clearly suggesting a role for PI3-kinase in NSCLC survival. The question remains whether onset of apoptosis upon combined treatment with VP-16 and PI3-kinase inhibition in all cases occurred according with the same mechanism at the same stage of the cell cycle as cell death caused by single use of compounds inhibiting the PI3-kinase pathway? When used alone, all three inhibitors of the PI3-kinase pathway allowed at least a fraction of cells to leave G2 and, as discussed above, morphological and biochemical data suggested a mitosis-related death. FACS analysis showed that in response to post-treatment with PI3-kinase inhibitors of DNA-damaged cells in all cases PARP cleavage occurred when most cells had doubled their DNA. The question is whether the DNA-damaged cells died by mitotic catastrophe due to premature exit from G2 induced by PI3-kinase/Akt inhibition, or if cell death was induced when the cells were still at G2-phase. There are many publications in which appearance of apoptotic cells was described simultaneously with activation of mechanisms associated with G2/M transition without showing any morphological signs of mitotic catastrophe.18, 19 Thus, it is possible that cell death due to DNA damage in combination with inhibition of the PI3-kinase pathway occurred at the same stages of mitosis as those observed upon single treatment with the kinase inhibitors. Morphological data obtained as a result of combined treatment with PKC 412 and VP-16 supports this idea. The more “clear” apoptotic nuclear morphology most often observed upon combination of DNA damage with PI3-kinase inhibition could be due to a more potent triggering of apoptosis in response to DNA damage-related mechanisms, leading to a faster cell death not displaying signs of mitosis. It is also possible that upon DNA damage, initiation of early events in G2/M transition by PI3-kinase inhibition is sufficient for triggering of apoptosis, without requirement to reach/leave metaphase for cell death to occur. According to another scenario inhibitors of the PI3-kinase pathway may sensitize DNA-damaged cells to apoptosis by mechanisms unrelated to mitosis regulation, an issue that is still under investigation.

Upon treatment of cells with PI3-kinase inhibitors, in addition to cell line-dependent differences in apoptosis induction, the level of apoptosis and cell cycle related changes varied dependent on the type of inhibitor used. Wortmannin is a more potent and specific PI3-kinase inhibitor, compared to LY-294002,29 which was also documented when the phosphorylation of the PI3-kinase downstream target Akt was analyzed in the H157 cells.

As mentioned above,10 the downstream PI3-kinase target Akt is implicated in treatment resistance of NSCLC cells, suggesting Akt inhibition as a mechanism important for cell death induced by the inhibitors of the PI3-kinase pathway. To our surprise we found no correlation between the level of Akt inhibition and apoptosis induction. In H157 cells LY-294002 was a better apoptosis inducer compared to the more potent Akt inhibitor wortmannin. In U1810 cells, even though treatment with 30 μM LY-294002 induced a higher percent of dead cells compared to a 10 μM concentration, there was no difference in Akt phosphorylation, suggesting involvement of other kinases in apoptosis induction. On the other hand, phosphorylation status upon treatment with PI3-kinase inhibitors is difficult to compare, since there might be changes over time. Moreover, it has been shown that the balance in the level of Akt activity is important for cell cycle regulation: too little as well as too much Akt activity trigger G2 arrest.30 Thus, Akt might play a role in apoptosis induction of the lung cancer cells, yet other kinases are most likely involved as well. For example, other members of the PI3-kinase family, such as DNA-PK, might be involved, since activity of this kinase was implicated in resistance of NSCLC.8 It is known that PI3-kinase activity is able to increase the level of cyclin E and D and to inhibit FOXO transcription factors that regulate transcription of proteins essential for the G1/S transition, such as p27.31 This observation might explain the accumulation of G1 cells upon treatment with wortmannin and LY-294002.

Since PI3-kinase inhibition induces mitotic catastrophe possible effector proteins are likely to be involved in checkpoint mechanisms. One of these proteins is Chk-1 that upon inhibition allows cells to prematurely leave G2-phase.32 A DN Chk-1 stimulates histone H3 Ser 10 phosphorylation in HeLa cells,33 suggesting involvement of Chk-1 inhibition in LY-294002-induced apoptosis. However, LY-294002 treatment of MDCK cells increases Chk-1 kinase activity, yet in this case the increase in apoptosis correlates with a G2 block,34 suggesting cell type-dependent differences in Chk-1 regulation. Of interest is also the observation that Chk-1 is one of the downstream targets of Akt. However, there are discrepancies concerning the outcome of Akt-mediated phosphorylation of Chk-1, moreover different phosphorylation sites in Chk-1 might give different outcomes. In MDCK cells LY-294002 triggers apoptosis and G2 arrest, which is inhibited by AKT-mediated phosphorylation of Chk-1 on site Ser 280 leading to inhibition of Chk-1.34, 35 In hematopoietic cells inhibition of GSK3 kinase that is a kinase negatively regulated by Akt, leads to enhancement of Chk-1 (Ser 345) phosphorylation.36 Etoposide treatment of these hematopoietic cells induces G2/M arrest in an Akt- and Chk-1-dependent manner indicating that in these cells increased Akt activity and the correlating increase in phosphorylation of Chk-1 stimulate G2 arrest. In A549, Calu1 and H596 NSCLC cell lines the staurosporine analog and Chk-1 inhibitor UCN-01 potentiate cisplatin-induced apoptosis, which is correlated with abrogation of the S and G2 checkpoints.37 UCN-01 can also sensitize A549 NSCLC cells to perifosine, which is correlated with a decrease in Akt phosphorylation and increase in PARP cleavage,38 supporting Akt involvement in cell death. However, UCN-01 inhibits purified Chk-1 from phosphorylation of a Cdc25C motif,39 making it difficult to understand if the death of NSCLC cells is dependent on Akt or direct Chk-1 inhibition. Our data show that the PI3-kinase/Akt pathway protects NSCLC cells from mitotic catastrophe. In very few studies the involvement of the PI3-kinase/Akt pathway in mitotic catastrophe was investigated. Recently was shown that Akt activity protects U87MG human glioma cells from mitotic catastrophe induced by temozolomide,40 and geldanamycin-induced mitotic catastrophe is associated with a decrease in Akt phosphorylation.41 In lung cancer cells, to our knowledge, the involvement of the PI3-kinase/Akt pathway in mitotic catastrophe has not been investigated. In fact, very few studies were performed to analyze the involvement of mitotic catastrophe in death of lung cancer cells. One reason is that cells dying due to mitotic catastrophe show characteristics of necrosis or apoptosis. Therefore, at the time points at which cell death is investigated, signs of mitotic catastrophe might be difficult to trace. One can argue that the most important result of tumor treatment is cell death rather than the understanding of the way of cell execution. However, to be able to develop new strategy to kill lung cancer cells, we believe that it is important to take into consideration cell cycle-related changes, pathways of apoptosis regulation and cellular signalling. All these three components interact with each other and are involved in regulation of cell survival and death. Therefore, understanding of the precise mechanisms related to crosstalk between these components will help to efficiently eliminate lung cancer cells.

Acknowledgements

The authors thank Professor Sten Orrenius for permanent support. This work was supported by grants from the Swedish and Stockholm Cancer Societies, and EC-RTD grant (to B.Z.).

Ancillary