Molecular profiling of docetaxel cytotoxicity in breast cancer cells: uncoupling of aberrant mitosis and apoptosis

Abstract

Among microtubule-targeting agents, docetaxel has received recent interest owing to its good therapeutic index. Clinical trials have underlined its potential for the treatment of advanced breast cancer, although little is known about its molecular mode of action in this context. We characterized the molecular changes induced by docetaxel in two well-known human breast carcinoma cell lines. Two mechanisms of action according to drug concentration were suggested by a biphasic sensitivity curve, and were further validated by cell morphology, cell cycle and cell death changes. Two to four nanomolar docetaxel induced aberrant mitosis followed by late necrosis, and 100 nM docetaxel induced mitotic arrest followed by apoptosis. Passing through mitosis phase was a requirement for hypodiploidy to occur, as shown by functional studies in synchronized cells and by combining docetaxel with the proteasome inhibitor MG132. Transcriptional profiling showed differences according to cell line and docetaxel concentration, with cell cycle, cell death and structural genes commonly regulated in both cell lines. Although p53 targets were mainly induced with low concentration of drug in MCF7 cells, its relevance in the dual mechanism of docetaxel cytotoxicity was ruled out by using an isogenic shp53 cell line. Many of the genes shown in this study may contribute to the dual mechanism by which docetaxel inhibits the growth of breast cancer cells at different concentrations. These findings provide a basis for rationally enhancing docetaxel therapy, considering lower concentrations, and better drug combinations.

Introduction

In spite of the availability of novel antineoplastic agents, breast cancer remains as one of the most important causes of death by cancer in the developed world (Veronesi et al., 2005). Among novel drugs for the treatment of advanced breast cancer are those that target microtubules, and constitute one of the most effective classes of chemotherapeutics for survival prolongation in advanced disease. Currently, taxanes are the most important antimicrotubule drugs used in chemotherapy, and include two members, paclitaxel and docetaxel. Docetaxel is semisynthesized from an inactive taxoid precursor, 10-deacetyl baccatin III, extracted from the needles of the European yew, Taxus baccata. It has shown high-cytotoxic activity in several solid tumours (Aapro, 1996; Consolini et al., 1998), including advanced breast cancer (Dieras et al., 1996; Fumoleau et al., 1996). Although docetaxel has been approved as adjuvant therapy in patients with early, high-risk breast cancer (Montero et al., 2005), cumulative systemic toxicity after prolonged and high-dose therapy has fuelled efforts to enhance the anticancer effects of lower doses.

At the molecular level, taxanes act by stabilizing tubulin heterodimers, impairing mitosis and cell proliferation in tumours (Jordan et al., 1996). Molecular effectors of docetaxel have been related with activation of signalling pathways, such as the jun N-terminal kinase pathway (Wang and Wieder, 2004), induction of p27 protein (Brown et al., 2004; Nawrocki et al., 2004), Raf-1 kinase (Blagosklonny et al., 1996; Caraglia et al., 2005) and phosphorylation of Bcl-2 (Blagosklonny et al., 1996; Berchem et al., 1999; Wang and Wieder, 2004). It has been reported that docetaxel increases Bcl-2 phosphorylation and downregulates bcl-xl, inhibiting the antiapoptotic function of Bcl-2 family (Haldar et al., 1997; Boudny and Nakano, 2002). In addition, increased p21 and p53 levels have been related to apoptosis following docetaxel treatment in human leukaemia cells (Avramis et al., 1998). In spite of previous efforts, a better understanding of the mechanisms of action and alterations in key molecular pathways induced by docetaxel is still needed for the designing of rational chemotherapy schedules in breast cancer.

In the present study, we conducted in vitro evaluations of the cytotoxic effects of different concentrations of docetaxel against two well-known human breast cancer cell lines. We evaluated cell death and cell cycle parameters together with high-throughput gene transcription analysis after drug exposure. In all these levels, two clearly different mechanisms of action were found according to drug concentration. Mitotic arrest was followed by apoptosis at high concentrations, whereas aberrant mitosis was induced at low concentrations and it was followed by necrosis. This dual mechanism of action of docetaxel was found to be independent of p53 activity.

The present data will help to understand the molecular mechanisms of action of docetaxel, and it is likely to contribute to a more accurate design of therapeutic strategies using this drug. The identification of altered gene expression induced by docetaxel demonstrates additional biological activity in breast cancer cells, and provides information regarding potential efficacy of using lower doses, and combining other agents with docetaxel.

Results

Biphasic growth of breast cancer cells in response to increasing concentrations of docetaxel

Sensitivity to docetaxel was assessed in MCF7 and MDA-MB-231 breast cancer cells. As shown in Figure 1a, cells are sensitive to nanomolar concentrations of docetaxel, with a biphasic curve of growth consistently found in both cell lines. At concentrations below 10 nM docetaxel, there was a first slope of growth inhibition, followed by a recovery, and a second decay after 50 nM docetaxel. At lower concentrations, MDA-MB-231 cells showed a higher degree of growth inhibition (IC75∼2 nM) when compared to MCF7 cells (IC75∼4 nM) (Figure 1a). Low concentrations of docetaxel were able to induce significant growth inhibition in both cell lines after 24 h of exposure, as shown in time course experiments (Figure 1b and c). By light microscopy examination, a high proportion of detached cells was observed after 24 h of exposure to 100 nM docetaxel in both cell lines (Figure 1d, and data not shown). However, a concentration of docetaxel corresponding to the first decay in cell growth (4 nM in MCF7 cells) was not able to induce the same pattern of cell detachment.

Figure 1

Effects of docetaxel on cell growth. (a) Cell viability after treatment with docetaxel. MCF7 and MDA-MB-231 cells were seeded in 24-well microtiter plate. After 24 h, different concentrations of docetaxel were added to triplicate wells and percent of surviving cells was evaluated after 48 h using a crystal violet assay. (b, c) Time course growth was evaluated at different time points after exposure to low concentrations of docetaxel in MCF7 (b) and MDA-MB-231 (c) cells. (d) Exponentially growing MCF7 cells were incubated without drug, with 4 nM docetaxel, or with 100 nM docetaxel for 24 h. Cell morphology was evaluated by light microscopy.

In summary, both cancer cell lines were highly sensitive to nanomolar concentrations of docetaxel and responded in a similar way with a biphasic growth curve, with MDA-MB-231 cells showing higher sensitivity at lower concentrations of drug.

Hypodiploidy and cell cycle arrest are alternatively induced by different concentrations of docetaxel

Biphasic growth after docetaxel exposure is consistent with the existence of at least two different mechanisms of action for this agent. In fact, a remarkable difference between low and high concentrations of docetaxel was found when studying cell cycle profiles with propidium iodide (PI) staining of DNA (Figure 2a and b). At low concentrations, a subG1 (hypodiploid=less than 2n DNA content) population was evident both in MCF7 and MDA-MB-231 cells after 24 h of exposure (31 and 8% subG1, respectively), suggesting DNA fragmentation or aneuploidy. SubG1 population was evident after as low as 2 nM docetaxel in MDA-MB-231, and 4 nM in MCF7 cells. In contrast, few subG1 cells were observed with 100 nM docetaxel in both cell lines, and more than 80% of cells were instead arrested in G2/M phase of the cell cycle, with MDA-MB-231 cells showing a higher subG1 fraction at this concentration of docetaxel. Staining of cells with phospho-MPM-2 antibodies confirmed that cells were arrested in mitosis after 24 h of 100 nM docetaxel (mitotic index of around 38% compared to 3% in non-treated control cells) (Figure 2c). Phospho-MPM-2 increase was evident as early as 8 h with 100 nM docetaxel, but started to normalize at 48 h (data not shown), indicating that cells were re-entering cell cycle with 4n content of DNA (mitotic slippage). Remarkably, hypodiploid cells were mainly observed at concentrations of docetaxel below 10 nM, and this finding was not observed in the related agent, paclitaxel (Figure 3a). As shown in Figure 3a, subG1 accumulation at 1 nM concentration was 1.8 and 19% for paclitaxel and docetaxel, respectively. In addition, to rule out the effect of docetaxel in other phases of the cell cycle, we studied the synthesis of DNA by bromodeoxyuridine (BrdU) uptake. In both MCF7 and MDA-MB-231 cells, BrdU uptake was not affected after 24 h of exposure to low concentrations of docetaxel (Figure 3b). In contrast, the G2/M arrest induced by 100 nM docetaxel was complete, and no DNA synthesis was observed.

Figure 2

Effects of docetaxel on cell cycle (a). DNA content was measured on asynchronously growing cells using PI-staining and fluorescence-activated cell sorter analysis at 8, 24 and 48 h following the start of treatment with low and high concentrations of docetaxel. Cell cycle profiles are shown for MCF7 (a), and MDA-MB-231 cells (b). (c) Mitotic index was assessed by flow cytometry using the MPM-2 antibody in MCF7 cells, untreated or treated with 4 or 100 nM docetaxel. Percentage of mitotic cells is shown, as well as the cell cycle profile after PI staining. MPM-2 expression is shown in y axis, and DNA content in x axis.

Figure 3

Effects of docetaxel on cell cycle (b). (a) DNA content was assessed with PI as in Figure 2. Increasing concentrations of paclitaxel and docetaxel were added to exponentially growing MCF7 cells. (b) BrdU uptake (y axis) vs DNA content (x axis) was assessed as described in Materials and methods after 24 h of treatment with low and high concentrations of docetaxel. Percentage of S-phase cells (considered as cells positive for BrdU staining) is shown for each dot plot.

In conclusion, two different mechanisms of action of docetaxel were suggested by sensitivity assays, microscopy examination, and cell cycle analysis. The effect induced at lower concentration seems to be specific for docetaxel, whereas mitotic arrest induced at higher concentrations is essentially the same as that observed after paclitaxel treatment. The data with BrdU uptake confirmed the G2/M arrest induced by 100 nM docetaxel, and ruled out arrest in other phases by lower concentrations of this agent.

Apoptosis induction after mitotic arrest with 100 nM docetaxel

Detached cells in culture plates could correspond either to cells arrested in mitosis or to apoptotic cells. In addition, the DNA fragmentation observed with low concentrations of docetaxel could also correspond to cell death, either late apoptosis or necrosis induced by therapy. In MCF7 cells, apoptosis was mainly induced after exposure to 100 nM docetaxel, as assessed by Annexin-V staining (11 vs 20%, for 4 vs 100 nM docetaxel, respectively) (Figure 4a). Separate Annexin-V staining of adherent and non-adherent cells confirmed that most of the apoptotic cells were confined to the detached fraction, probably entering apoptosis after mitotic arrest (47% of apoptotic cells after 100 nM docetaxel, Figure 4b). However, for the lower concentration of docetaxel (4 nM), no clear population of early apoptosis cells was observed, and instead, dead cells appeared as double positive (Annexin-V positive, and PI positive) cells, indicating necrosis (Figure 4b). Similar results were obtained with MDA-MB-231 cells (data not shown).

Figure 4

Cell death after docetaxel exposure. (a) Apoptosis was evaluated after treating MCF7 cells with 4 and 100 nM docetaxel, and staining with Annexin-V at 24 h. Flow cytometry profile represents Annexin-V-FITC staining in x axis and PI in y axis. The number represents the percentage of early apoptotic cells in each condition (lower right quadrant). (b) In addition, apoptosis was assessed by Annexin-V staining of MCF7 cells as in (a), after separating adherent and detached cells in the culture plates.

In this sense, apoptosis was a relatively late finding, occurring after arrest in mitosis in cells treated with 100 nM docetaxel. In contrast, hypodiploid cells induced by 2–4 nM docetaxel did not stain for Annexin-V, in agreement with the induction of aneuploidy instead of DNA fragmentation owing to cell death. However, at longer time points double-positive cells were seen, indicating non-apoptotic cell death.

Docetaxel targets different phases of the cell cycle according to drug concentration

The induction of hypodiploid cells by 2–4 nM docetaxel can be dependent on DNA fragmentation of cells in a cell cycle-independent way. Alternatively, hypodiploid cells can be the result of asymmetrical cell division after impaired mitosis. This issue was functionally assessed by synchronizing MDA-MB-231 cells with double-thymidine block and exposing to docetaxel in different points of the cell cycle. In this manner, we can rule out the cell cycle dependency of this process. Preliminary studies showed that cells were almost totally arrested in G1 just before release from double-thymidine block (Figure 5a). In addition, 6 h after release, more than 80% of cells showed a 4n content of DNA with a mitotic index of 2%, indicating synchronization in G2. Therefore, cells were treated with 2 and 100 nM docetaxel just after release from double-thymidine block (sensitivity of G1 cells), and 6 h after release (sensitivity of G2 cells). Non-synchronized cells were also treated as a control. As can be seen in Figure 5b, control cells were synchronized in G2 after 6 h of culture, and moved to G1 after 6 additional hours. When G1-arrested cells were treated with 2 nM docetaxel for 6 h, no hypodiploid cells were observed (0.9%). In contrast, when cells were treated 6 h after release from block (G2 cells) with 2 nM docetaxel, subG1 accumulation was significantly higher (11%) (Figure 5b). Thus, subG1 accumulation after 2 nM docetaxel is probably dependent on cells passing through mitosis, as these cells show higher sensitivity. Similar to control cells, G1 cells treated with 100 nM docetaxel were arrested in G2, 6 h after release. In addition, cells treated with 100 nM docetaxel after reaching G2 were kept with 4n DNA content for additional 6 h. However, the mitotic index increased from 5 to 62%, indicating that they progressed from G2 to M and stayed arrested in this phase. No subG1 accumulation was observed in any of the samples treated with this high concentration.

Figure 5

Cell cycle-dependent effect of docetaxel. (a) Double-thymidine synchronization of MDA-MB-231 cells. (b) Cells were left untreated or treated immediately after release from synchronization (0 h) or 6 h after release, with 2 or 100 nM docetaxel. Cell cycle profiles and mitotic index were determined by PI and MPM-2 staining, respectively. (c) Antagonist effect of docetaxel and the proteasome inhibitor MG132, in MDA-MB-231 cells. Cell cycle profile after 24 h of 2 nM docetaxel, alone or in combination with 10 μ M of the proteasome inhibitor MG132.

Previous studies in our laboratory have shown a schedule-dependent enhancement of the cytotoxicity induced by a different taxoid drug, paclitaxel, when combined with the proteasome inhibitor MG132 (Hernandez-Vargas et al., unpublished observations). This effect was related with the ability of both drugs to selectively target different phases of the cell cycle. To further study the dependence of low concentration docetaxel on cell cycle, we combined docetaxel with MG132 (which induces cell cycle arrest before mitosis). As predicted, the induction of hypodiploid cells with 2 nM docetaxel was drastically reduced when combined with MG132 (from 18 to 2.5%, Figure 5c). According to our premise, this result favours the possibility that the hypodiploid population induced by low concentrations of docetaxel is dependent on cell cycle.

The findings with cell synchronization and combination with MG132, demonstrate that passing through mitosis is a requirement for 2 nM docetaxel to induce a hypodiploid population in MDA-MB-231 cells. These data also indicate that MG132 may have an antagonist effect on the antitumour activity of docetaxel in vitro.

Aberrant mitosis and multinucleation are alternatively induced with two different concentrations of docetaxel

To confirm our previous findings, the morphology of MCF7 cells was analysed by immunofluorescence after 4 and 100 nM docetaxel. As shown in Figure 6, a main finding after both concentrations of docetaxel was the presence of multipolar spindles. However, although multinucleated cells were seen after both concentrations of docetaxel, 100 nM docetaxel mainly produced multinucleated cells. Instead, cells exposed to 4 nM docetaxel were small, and mainly with single nucleus (Figure 6).

Figure 6

Immunofluorescence assay. Microtubules and nuclear staining are shown separately (α-tubulin and DAPI, respectively) and combined (merge), for MCF7 cells untreated or exposed to 4 or 100 nM docetaxel for 24 h.

This confirms that hypodiploidy observed with 4 nM docetaxel is the consequence of asymmetrical divisions originated by aberrant mitosis. In contrast, cells treated with 100 nM docetaxel do not undergo cytokinesis after mitotic arrest, and remain with a higher content of DNA.

Transcriptional profile of breast cancer cells exposed to docetaxel

As a strategy to understand the mode of action of docetaxel in breast cancer cells, differential gene regulation was studied by cDNA microarrays. As two potential mechanisms of docetaxel cytotoxicity were described in our preliminary data, two different concentrations of drug were also used in MCF7 and MDA-MB-231 cells. Shown in Table 1 are selected genes with high regulation after different conditions of docetaxel therapy, relative to non-treated cells. Ontogeny analysis revealed genes related with multiple cell functions, although the most important groups included cell cycle, structural genes, antioxidative stress and apoptosis (Supplementary Information 1, and data not shown).

Table 1 Induced (↑) and suppressed (↓) genes in MCF7 and MDA-MB-231 cells after exposure to docetaxel

In agreement with preliminary data, cDNA microarrays analysis revealed dose-dependent patterns of gene expression, identifying genes and pathways potentially related with the mechanisms of docetaxel cytotoxicity. Known targets of p53 were induced after exposure of MCF7 cells to docetaxel (CDKN1A/p21, TNFRSF6/Fas, Tp53INP1 and SOD2). However, this was more evident when using low concentrations of docetaxel (4 nM). In contrast, 100 nM docetaxel was associated with the regulation of G2/M-related transcripts (e.g. AURKA/STK15, PLK1 and VEGFC), as well as genes related with checkpoint activation (BUB1B and MAD2L1). Some of these data were validated by quantitative reverse transcriptase–polymerase chain reaction (qRT–PCR) and immunoblotting (Figure 7). Interestingly, although TNFRSF6/FAS was only induced at 4 nM docetaxel in MCF7 cells, qRT–PCR analysis showed a high upregulation of this transcript also at high concentrations (Figure 7b). Moreover, it was induced in MDA-MB-231 cells at similar levels, after docetaxel exposure (Figure 7c). Because of its relevance in the induction of cell death, TNFRSF6/FAS could be a common mechanism of action of docetaxel, independent of cell type and drug concentration. Validations of microarrays data included cells treated simultaneously with the related taxoid, paclitaxel, to check for specificity of these findings. For these studies, 100 nM paclitaxel was used, which induces cell cycle changes similar to those induced by 100 nM docetaxel. Some differences were seen, including a lower accumulation of p53 and H2AX, and no changes in Bax and Bcl-2 proteins (Figure 7a). Interestingly, p21 protein levels were higher, in spite of lower p53 accumulation. In a similar way, 4 nM docetaxel induced less accumulation of p53 but higher levels of p21, when compared with 100 nM docetaxel. These findings are in agreement with a recent analytical report considering p53 stabilization as a marker of the duration of mitotic arrest (Blagosklonny, 2006). In contrast, prolonged mitotic arrest induced by 100 nM docetaxel was associated with a higher accumulation of Bax and phosphorylation of Bcl-2 proteins (Figure 7a).

Figure 7

Validation of gene expression profiles obtained by microarrays analysis in breast cancer cell lines treated with docetaxel. Microarrays results in MCF7 and MDA-MB-231 cells were validated by immunoblotting (a) or TaqMan quantitative real-time PCR (b, c). Western blot analysis of protein expression is shown for MCF7 and MDA-MB-231 cells (a) after treatment with docetaxel. α-Tubulin is included as a control. qRT–PCR expression of selected transcripts is shown for MCF7 (b) and MDA-MB-231 (c) cells, after 24 h of docetaxel exposure. Relative induction is shown in a fold scale.

In MDA-MB-231 cells, p53 transactivation activity was no evident. However, some interesting genes were also observed (Table 1). With 2 nM docetaxel, induction of EPAS1/HIF1A was observed, in agreement with an antiangiogenic effect of this drug (Escuin et al., 2005). In addition, downregulation of HSPA8 is an interesting finding, as an antiapoptotic role for this protein has been recently described (Stankiewicz et al., 2005). Some common genes for both concentrations of docetaxel included DCT, TYR, TYRP1 and MAF. Consistent with a stronger arrest in mitosis with 100 nM docetaxel, also MDA-MB-231 cells showed a higher expression of mitosis-related genes, such as PLK1 and PLAU. In immunoblots analysis, Bax and Bcl-2 phosphorylation was observed with both concentrations of docetaxel, although Bcl-2 phosphorylation seemed higher after 100 nM concentration. Finally, in MCF7 cells DNA damage was mainly evident at higher concentrations of docetaxel, as assessed by phospho-H2AX immunoblotting (Figure 7a). However, this finding was especially evident in MDA-MB-231 cells, after both concentrations of docetaxel. This is in agreement with the higher sensitivity of MDA-MB-231 cells to docetaxel.

Docetaxel response in p53-deficient MCF7 cells

Although the global effect of docetaxel was similar in the two breast cancer cell lines studied here, some differences between both cell lines were also evident: a higher sensitivity of MDA-MB-231 cells, with a higher trend to show hypodiploid cells and DNA damage. In addition, transcriptional profile was essentially different, showing induction of p53 targets after 4 nM docetaxel in MCF7 cells. One main difference between both cell lines is the status of the p53 protein, as a point mutation in MDA-MB-231 cells has been shown to inhibit p53 function (Olivier et al., 2002). Because of the relevance of p53 in the sensitivity of tumour cells to chemotherapy agents, we studied the response to docetaxel in MCF7 cells made deficient for p53 by RNA interference (shp53). Cell cycle changes were similar after both concentrations of docetaxel when comparing shp53 and control shGFP MCF7 cells (Figure 8). SubG1 population was evident at 4 nM docetaxel, and G2/M arrest was the main finding after 100 nM docetaxel, in both shp53 and shGFP cells.

Figure 8

Cell cycle effect of docetaxel in p53-deficient cells. (a) DNA content was assessed by PI staining of shp53 MCF7 cells vs control shGFP MCF7 cells (b), after exposure to 4 and 100 nM docetaxel. Percentage of cells in each cell cycle phase is shown in the bar graph.

Considering these results, none of the mechanisms of action described for low and high concentrations of docetaxel seems to be strictly dependent on a functional p53 protein. In this sense, the differences observed between the p53-wt MCF7 and the p53-mt MDA-MB-231 cell lines in this study, are probably not dependent on p53 status.

Discussion

We have shown that differences in the concentration of docetaxel will influence its mechanism of action at the molecular level, including cell cycle, cell death and transcriptional profile. Docetaxel (100 nM) is able to induce mitotic arrest starting as early as 8 h after exposure. In contrast, lower concentrations (2–4 nM) induce a high accumulation of cells in the subG1 region. In addition, by studying dependency of this effect on cell cycle phase, we concluded that docetaxel at lower concentrations is able to induce aberrant mitosis and aneuploidy. This finding was further validated by immunofluorescence analysis. Currently, there are no experimental data to explain these findings, although several lines of evidence can shed light onto this issue. Docetaxel is easily accumulated in cells. By using very low doses, this accumulation is avoided, and a transient impairment of microtubules dynamic is achieved. In MCF7 cells, this impairment produces a moderate stabilization of p53 protein, followed by an induction of p53 transactivation activity once mitotic arrest is overcome. In contrast, docetaxel accumulation after higher doses of the drug induces a sustained arrest in mitosis with the consequent inhibition of transcription and higher p53 accumulation. Owing to an unknown mechanism, this difference in mitotic arrest duration also affects the postmitotic consequences, that is, aberrant mitosis with production of hypodiploid cells (at 2–4 nM docetaxel) vs mitosis slippage without cytokinesis producing large multinucleated cells (at 100 nM docetaxel). Moreover, as a consequence of this, the type of cell death will also be different, that is, mainly non-apoptotic after asymmetrical division, and mainly apoptotic after multinucleation. In this sense, our results with p53-wild-type MCF7 cells, are in agreement with the model proposed by Blagosklonny (2006). According to this model, mitotic arrest will induce p53 accumulation owing to the lack of transcriptional regulation of p53 during mitosis. After slippage, cells recover transcriptional ability and p53 protein will induce transcriptional targets, including p21. Therefore, our findings of p53 target activation after low concentrations of docetaxel can be the result of transient arrest in mitosis and accumulation of p53 followed by activation of transcriptional targets of p53 after aberrant mitosis. In contrast, prolonged arrest after 100 nM docetaxel will induce a higher accumulation of p53, but transcription will be stopped for longer time, so targets like p21 will not be activated. In agreement with Blagosklonny hypothesis, our data point to a mechanism regulating the different consequences of docetaxel exposure according to the duration of mitotic arrest, depending on the accumulation of proteins (including p53 family of transcription factors) in a moment of the cell cycle when there is no transcriptional activity. Timing of arrest would determine the amount, and probably also the quality or type, of protein/s that will accumulate. Therefore, more studies are required at the post-transcriptional level, to complement our findings and support this hypothesis.

Interestingly, the finding that low-dose (but not high dose) docetaxel induces p21, suggests a possible mechanism for mitotic catastrophe induced by the low dose. The situation here may parallel the work of Chang et al. (2000), where transient induction of p21 was shown to induce mitotic catastrophe. In that case, this outcome was associated with asynchronous resynthesis of mitosis-executing and mitosis-regulating proteins after release from p21 inhibition. In agreement with this interpretation, gene expression data show several of the transcripts for these proteins (BUB1B, MAD2L1, CENPA, PLK1) as induced by high-dose docetaxel, but not by the low dose (which induces p21, potentially inhibiting these genes).

This mechanism of cytotoxicity is not valid for MDA-MB-231 cells with non-functional p53 protein. In this sense, the cell response to paclitaxel has been described to be dependent on p53 status, although this last point is more controversial (Giannakakou et al., 2002). In addition to transcriptional profiles, some differences between MCF7 and MDA-MB-231 cells were found in this study. MDA-MB-231 cells showed a higher sensitivity at lower concentrations of docetaxel, and an increased trend to produce subG1 cells after treatment. However, when using an isogenic cell line, we found no difference in response to docetaxel according to p53 functionality. In fact, the same pattern of mitotic arrest at high concentration and aberrant mitosis at low concentrations was found after docetaxel exposure. This also gives support to the consistency of our findings and its relevance, considering the high frequency of p53 mutations in breast tumour cells.

The consequences of a better understanding of the mechanisms of action of docetaxel are important. Mitotic perturbations will determine the main mechanism of cell death after docetaxel exposure. Similar to our results, Fabbri et al. (2006) found that lower concentrations of docetaxel were not able to induce early apoptosis (Annexin+, PI− cells) in a bladder cancer cell line, and instead it induced a necrotic type of cell death. In a similar way, the ability of paclitaxel to induce cell death by independent mechanisms has been studied in more detail than docetaxel. It has been described that paclitaxel-dependent cytotoxicity requires cells to be first arrested at mitosis and upon mitotic slippage to enter apoptotic cell death (Blajeski et al., 2001). Moreover, differences in mechanism of action according to drug concentration have also been described with paclitaxel. In the early work of Torres and Horwitz (1998), aberrant mitosis was induced by low concentrations of paclitaxel followed by non-apoptotic cell death, and high concentrations induced mitotic arrest with apoptosis. In addition, according to Giannakakou et al. (2001), low concentrations of paclitaxel (between 6 and 12 nM) induce p53, p21 and G1/G2 arrest in MCF7 cells, instead of mitotic arrest. We observed a similar activation of p53 pathway after docetaxel exposure in MCF7 cells, with upregulation of p21 and Tp53INP1. However, we could not find a clear difference in cell cycle and cell death profile in paclitaxel, as we found for docetaxel, in a concentration-dependent manner.

Induction of cell death and cell cycle arrest has been considered as the main mechanisms for drug-dependent inhibition of cell growth. However, alternative modes of cytotoxicity, as aberrant mitosis, are of increasing interest as potentially induced by chemotherapy, especially in cells that may be apoptosis deficient (Morse et al., 2005). In accordance to our results, Chen et al. (2003) described that low concentrations of mitotic inhibitors (10 nM EpoB or paclitaxel) induce aberrant mitosis, and higher concentrations induce mitotic arrest followed by mitotic slippage after 36 h. In addition, our results in MCF7 and MDA-MB-231 cells indicating aberrant mitosis, complement those obtained by other authors by microscopy examination (Paoletti et al., 1997; Morse et al., 2005) of breast cancer cells treated with docetaxel. According to the study of Morse et al., the primary mechanism of death is mitotic catastrophe as assessed by scoring of micronucleated cells and cells undergoing aberrant mitosis. At concentrations of docetaxel similar to those used in our study (10 and 100 nM), non-apoptotic cell death was the main mechanism of cytotoxicity, and it was mainly dependent on the induction of aberrant mitosis. However, no subG1 accumulation was described in that study, and authors conclude that the damage causes the improper segregation of chromosomes resulting in aberrant mitosis, multiple micronuclei and an eventual necrosis-like death. In a similar way, aberrant mitosis with low concentration docetaxel was followed by non-apoptotic cell death in our study. Interestingly, although longer time points would be necessary to evaluate the final effect of different concentrations of docetaxel in breast cancer cells, long-term assays show that cells undergoing mitotic catastrophe eventually result in cell death, reducing overall clonogenic survival (Roninson et al., 2001; Blagosklonny et al., 2002). Moreover, in agreement with several studies, mitotic death normally ends in necrosis (Mansilla et al., 2006).

In a similar way to MG132, other studies have shown schedule-dependent inhibition of taxanes cytotoxicity. It was previously shown that doxorubicin could interfere with the cytotoxic effect of docetaxel on both mitotic arrest and apoptotic cell death, when administered before or simultaneously (Zeng et al., 2000). However, no antagonism was observed when docetaxel was added before doxorubicin. Nawrocki et al. (2004) studied the effect of combining docetaxel with the proteasome inhibitor bortezomib, for the treatment of pancreatic cancer cells. Although the combination was shown to be effective in vivo (probably related with an enhancement of the antiangiogenic effect of docetaxel), in vitro studies showed that the drug combination inhibited docetaxel-induced DNA fragmentation in one of the cell lines tested. Although this mechanism is not fully understood, it is possible that MG132 counteracts docetaxel-induced aberrant mitosis when administered in first place by arresting cells before entering mitosis.

Several previous reports have studied patterns of gene expression in response to docetaxel. Brown et al. (2004) generated MCF7 and MDA-MB-231 cell lines resistant to docetaxel, and by cDNA microarrays identified several candidates for mediating such resistance. Reduced expression of p27 was considered as a potential mechanism of resistance in their study. Li et al. (2004) studied gene regulation in prostate cancer cells in response to 2 nM docetaxel for 6, 36 and 72 h. In a similar way to our microarrays data, the expression of tubulin was decreased, whereas the expression of microtubule-associated proteins was increased, confirming the microtubule-targeting effect of docetaxel. Clustering analysis showed downregulation of some genes for cell proliferation and cell cycle and upregulation of genes involved in apoptosis and cell cycle arrest. Yoo et al. (2002) have reported docetaxel-induced gene expression patterns in head and neck cancer cells by using cDNA microarrays and PowerBlot immunoblotting technique. Interestingly, as we found for low concentration docetaxel in MCF7 cells, they showed that docetaxel regulated some genes that are related to cell cycle, apoptosis, angiogenesis and tyrosine kinase signal transduction. They found that docetaxel induced apoptosis is, in part, because of the Fas death receptor upregulation in HNSSC cell lines and not FasL.

To our knowledge, this is the first study correlating two molecular mechanisms of docetaxel activity with a transcriptional profile analysis. These data have been validated in two non-related breast cancer cell lines, making it relevant for future studies assessing taxanes action and the identification of novel targets for chemotherapy. The differences described here may have clinical implications, providing a rational approach to lower the therapeutic concentrations of docetaxel-based chemotherapy. In this sense, at the clinical setting, it has been shown that weekly regimen of docetaxel (with lower concentrations of drug) can achieve a clinical response comparable to traditional 3-weekly regimens (with higher concentrations of drug), although the side effects profile is improved (Baker et al., 2004). In addition, the cytotoxicity of docetaxel mediated by mechanisms alternative to apoptosis may be partly responsible for their efficacy in treating breast cancers, which may be apoptosis deficient. Finally, the dependency on cell cycle phases should also be considered when designing novel strategies for combination chemotherapy, as cells synchronized in specific phases of the cell cycle can be more sensitive to a certain concentration of docetaxel.

Materials and methods

Cell culture and drug sensitivity assays

Human breast carcinoma cell lines MCF7 and MDA-MB-231 were cultured in Dulbecco's modified Eagle medium with Glutamax-1 (Gibco-BRL, Grand Island, NY, USA) supplemented with 10% foetal bovine serum (Gibco-BRL). Docetaxel (Taxotere) was obtained from Aventis Pharma SA (Paris, France), and stored at a concentration of 10 mg/ml (12.6 mM) in 13% w/w ethanol at 4°C. Paclitaxel (Sigma-Aldrich, St Louis, MO, USA) was dissolved in dimethyl sulfoxide to make a stock concentration of 20 mM and stored at −70°C. MG132 (Calbiochem, La Jolla, CA, USA) was used at 10 μ M.

Growth inhibition was evaluated using a crystal violet assay as described previously (Carmichael et al., 1987). Docetaxel was tested at scalar concentrations ranging from 0 to 1000 nM for 24–96 h. The optical density of treated cells was determined at a wavelength of 595 nm.

Flow cytometry analysis of cell cycle and cell death

Cell cycle distribution was measured with and without exposure to docetaxel. After culture at different time points, samples were harvested (including detached cells), suspended in phosphate-buffered saline (PBS), fixed in 70% ethanol, and their DNA content was evaluated after PI staining, as described previously (Hernández-Vargas et al., 2006). Fluorescence-activated cell sorting analysis was carried out using a FACScan flow cytometer (Becton Dickinson, San Diego, CA, USA) and CellQuest software. In some experiments, cells were synchronized using the double-thymidine block protocol before cell cycle analysis. In brief, 2 mM thymidine was added to culture plates with less than 50% confluence for 17 h. Cells were released from first block by washing and replacing with fresh medium. After 8 h, cells were exposed again to 2 mM thymidine, and released 15 h later by washing and replacing with fresh medium. The BD Pharmingen BrdU Flow Kit (BD Biosciences, San Jose, CA, USA) was used to detect cells entering and progressing through the S phase (DNA synthesis) of the cell cycle. BrdU was added 45 min before harvesting of cells, and incorporation to the cells was assessed according to the manufacturer's recommendations. Simultaneous staining of DNA with 7-amino-actinomycin D was used in combination with BrdU, followed by two-colour flow cytometric analysis.

The Annexin-V-FITC Apoptosis Detection Kit (BD Biosciences) was used to detect apoptosis by flow cytometry. Cells were exposed to several conditions of docetaxel treatment, and after 0–48 h, they were harvested (including detached cells), and processed according to the manufacturer's instructions.

Immunofluorescence and confocal microscopy analysis

For immunofluorescence experiments, cells were plated on sterile 10-mm glass coverslips, and allowed to adhere for 24 h, after this time, cells were left untreated or treated with 2 and 100 nM docetaxel. After 24 h, they were fixed in methanol (−20°C, 5 min). Cells were blocked in 5% bovine serum albumin-PBS and then incubated with anti-α-tubulin (Sigma-Aldrich) as primary antibody at 1:1000 dilution. The secondary antibody, Alexa-488-coupled anti-mouse immunoglobulin (Ig) (Molecular Probes, Eugene, OR, USA), was applied at 1:100 dilution. Cell nuclei were stained using 4,6-diaminidino-2-phenylindole (DAPI) (Molecular Probes). Staining was examined using a confocal ultra-spectral-microscope (TCS-SP-2-AOBS-UV, Leica, Wetzlar, Germany). Post-capture image analysis and processing of confocal image stacks were performed using the Leica Confocal software.

Microarrays hybridization and data analysis

Total RNA from cell lines untreated and treated with several concentrations of docetaxel during 24 and 48 h was extracted from duplicate plates. Three micrograms of total RNA from samples and Universal Human Reference RNA (Stratagene, La Jolla, CA, USA) were used for amplification using the Superscript Choice System (Life Technologies Inc., Gaithersburg, MD, USA) and in vitro transcription with Megascript T7 (Ambion, Austin, TX, USA), as described previously (Moreno-Bueno et al., 2003; Hernández-Vargas et al., 2006). The cDNA array chip is the CNIO Oncochip manufactured by the CNIO Genomic Unit (http://bioinfo.cnio.es/data/oncochip). It contains ∼11 500 clones corresponding to 9300 different genes (UniGene clusters) selected by its molecular function, with emphasis on cancer related processes. Microarray data were quantified using the GenePix Pro 6.0 program (Axon Instruments Inc., Union City, CA, USA). The microarrays raw data discussed in this publication have been deposited in NCBIs Gene Expression Omnibus (GEO, http://www.ncbi.nlm.nih.gov/geo/) and are accessible through GEO Series accession number GSE5149 (submitter H Hernández-Vargas). For statistical analysis, we selected genes whose expression differed by a factor of at least threefold with respect to the reference pool. The SOTA and TreeView programs (http://gepas.bioinfo.cnio.es/) were used for clustering analysis, assuming Euclidean distances between genes. Each condition of treatment was performed in triplicate arrays, with RNA obtained from independent experiments.

Quantitative real time RT–PCR

Quantitative real-time PCR (TaqMan) was performed with ABI PRISM 7700 Sequence Detection System Instrument and software (Applied Biosystems, Foster City, CA, USA), using the manufacturer's recommended conditions. The comparative threshold cycle (C_t) method was used to calculate the amplification factor, and the amount of target and endogenous reference was determined from a standard curve for each experimental sample. The sequence of oligonucleotides and TaqMan probes used for the analysis of AURKA/STK15, CDKN1A/p21, TNFSRF6/Fas and Tp53INP1 were obtained using the Assays-by-Design (SM) File Builder program (Applied Biosystems). Expression of β2-microglobulin was used as internal standard.

Immunoblotting

Protein lysates obtained from cells treated with docetaxel for up to 48 h and untreated control cells, were measured using a colorimetric detection assay (BCA Protein Assay, Pierce, Rockford, IL, USA). Equal amounts of protein lysates (20–30 μg) were separated by sodium dodecyl sulfate–polyacrylamide gel electrophoresis on 12% gels, and electrotransferred to Immobilon-P membranes (Millipore Corporation, Bedford, MA, USA). Primary antibodies included anti-p53 DO7 (Novocastra, Newcastle, UK), anti-p21 (Oncogene, Boston, MA, USA), anti-BCL-2 (Dako, Glostrup, Denmark), anti-Bax (Santa Cruz Biotechnology, Heidelberg, Germany), anti-phospho-H2AX (Upstate, Lake Placid, NY, USA) and anti-α-tubulin (Sigma). After incubation with peroxidase-conjugated secondary antibody (anti-mouse IgG-horseradish peroxidase; Dako) protein expression was detected using ECL Western blotting reagents (Amersham Biosciences, Uppsala, Sweden).

RNA interference to shut-down p53

The vectors expressing shRNA's against p53 (shp53) and GFP (shGFP, control) have been described previously (Grandori et al., 2003; Tarsounas et al., 2004). Phoenix cells (from ATCC) carrying the amphotropic receptor were used to generate retroviral supernatants. The retrovirus infection protocol is as described previously (von Kobbe et al., 2004). Briefly, MCF7 cells were infected twice, in two rounds of 24 h infection. Stably infected cell clones were selected in puromycin (2 μg/ml) (Sigma).