Background
Breast cancer is the most common type of cancer in women and accounts for one third of all cancers in women. The incidence of breast cancer is continuously increasing, with more than one million reported new cases diagnosed per year worldwide [
1‐
3]. Among these cases, 20-30% present with metastatic or locally advanced disease, and another 30% will develop recurrent or metastatic disease [
3].
Chemotherapy is often used to treat breast cancer in both the adjuvant and neoadjuvant settings and often involves the administration of anthracyclines together with (or followed by) taxanes. Typical taxanes used in current treatment regimens for breast cancer include paclitaxel (Taxol) and docetaxel (Taxotere), while typical anthracyclines include doxorubicin (Adriamycin) and epirubicin (Pharmorubicin) [
4,
5]. Docetaxel belongs to the group of taxanes that were first introduced into clinical use during the 1990’s. Both paclitaxel and docetaxel bind to β-tubulin in assembled tubulin, thereby reducing depolymerisation [
6]. Taxanes stabilise microtubules and dampens microtubule dynamics to prevent the normal formation of mitotic spindles [
6]. This leads to chronic activation of the spindle assembly checkpoint (SAC), which in turn leads to mitotic arrest [
7]. Extended mitotic arrest eventually leads to cell death [
8]. We have recently shown that the taxanes also strongly induce TNF alpha production, which may also promote apoptosis through binding to its receptor TNFR1 [
9]. On the other hand, doxorubicin is an anthracycline antibiotic, which intercalates with DNA. While the mechanisms of action are not yet completely understood, one very important component of the activity of doxorubicin is its interaction with chromatin and its constitutive components: DNA and histones. These interactions lead to chromatin unfolding and aggregation. This chromatin structural disruption is likely to interfere with DNA replication and transcription, which eventually leads to cell apoptosis. It was also suggested that doxorubicin interacts with key cellular enzymes such as topoisomerases I and II. Topoisomerase II mediated DNA damage by doxorubicin is followed by G1 and G2 growth arrest and induction of apoptosis, which correlates with tumor response and patient outcomes [
1,
10‐
12].
Resistance to chemotherapy can occur prior to drug treatment (primary or innate resistance) or develop over time following exposure to a given chemotherapeutic agent (acquired resistance) [
13]. Chemoresistance presents a major obstacle to therapy and leaves few effective treatment options [
5]. Both innate and acquired resistance to taxanes and doxorubicin are common as more breast cancer patients receive these drugs [
1,
14]. The most established
in vitro mechanism for resistance to more than one chemically unrelated class of agents (multidrug resistance) is the overexpression of drug efflux proteins. The best known drug efflux proteins are members of the ATP-binding cassette (ABC) superfamily, including P-glycoprotein [Pgp; also called multidrug resistance protein (MDR) or ABCB1], the multidrug resistance-associated protein 1 [MRP-1, also called ABCC1], and the breast cancer resistance protein [BCRP, also called ABCG2]. ABC transporter substrates include a diverse array of compounds, many of them structurally unrelated. These proteins protect cells and tissues by exporting potential toxins, including anticancer agents from cells in normal tissues and cancer cells [
4]. In general, ABCB1 transports large hydrophobic compounds, whereas ABCC1 and ABCG2 transport both hydrophobic drugs and large anionic compounds [
15]. ABC proteins have been implicated in both taxane and doxorubicin resistance in breast cancers [
1,
3,
4,
14]. When 60 cell lines were tested, it was found that the lower the ABCB1 expression level, the greater the sensitivity to paclitaxel in the cell lines [
16]. However, in clinic studies the results are controversial. One study shows that increased ABCB1 expression level is correlated with shortened disease-free survival [
17]. Some other studies show that no correlation between ABCB expression level and response to either paclitaxel or docetaxel treatment in breast cancer patients [
18]. On the other hand, both ABCC1 and ABCG2 mediate resistance to doxorubicin, but not paclitaxel [
5,
19].
Resistance may also arise from the expression of proteins underlying a specific drug’s mechanism of action. For example, taxanes operate by binding to β-tubulin. Taxane-resistant cancer cells may have altered expression and function of certain β-tubulin isotypes, caused by mutations in β-tubulin, and increased microtubule dynamics associated with altered microtubule-associated protein (MAP) expression [
3,
4,
14,
20‐
23]. Altered expression of the topoisomerase IIa gene (TOP2A), which encodes the enzyme target of the anthracyclines, may confer anthracycline resistance [
24].
Chemoresistance is a major factor involved in poor response and reduced overall survival in patients with locally advanced and metastatic breast cancer. Chemoresistance is a very challenging and complex phenomenon involving a number of complex mechanisms. Elucidating these mechanisms is crucial to understanding how to improve the use of taxane and doxorubicin in cancer treatment. Although extensive studies have been carried out to understand chemoresistance in breast cancer both in vitro and clinically, many questions remain unanswered. In previous research, we established several drug-resistant MCF-7 cell lines by exposing MCF-7 cells to increasing concentrations of specific chemotherapy drugs [
25]. Our study showed that while drug transporters were induced during selection for drug resistance (which reduced drug accumulation into tumour cells), additional drug-transporter-independent mechanisms must play important roles [
25]. In the current study, we used two of our preciously created resistant cell lines, doxorubicin-resistant MCF-7 cells (MCF-7
DOX) and docetaxel-resistant MCF-7 cells (MCF-7
TXT), to study the mechanisms underlying the acquired drug resistance, with emphasis on the resistance to taxanes in MCF-7
TXT cells. We show that MCF-7
TXT cells are ten times more resistant to both docetaxel and paclitaxel than the sensitive wild type parental cell line (MCF-7
CC). MCF-7
DOX cells are eight times more resistant to doxorubicin than MCF-7
CC cells. However, MCF-7
TXT cells are not cross-resistant to doxorubicin and MCF-7
DOX cells are not cross-resistant to taxanes. We also showed that multiple mechanisms are involved in the resistance to taxanes in MCF-7
TXT cells. Firstly, the selected chemo-resistant cell lines express higher levels of certain ABC proteins. The expression level of ABCB1 is very high only in the MCF-7
TXT cells and the expression level of ABCC1 is very high only in the MCF-7
DOX cells. The expression level of ABCG2 is similar in both the selected chemo-resistant and the parental MCF-7 cell lines. Moreover, MCF-7
TXT cells are also more resistant to taxane-induced mitotic spindle disruption and M phase arrest, which leads to apoptosis. The microtubule dynamics of MCF-7
TXT cells are insensitive to the docetaxel treatment, which may partially explain why docetaxel is less effective in inducing M-phase arrest and apoptosis in MCF-7
TXT cells in comparison with MCF-7
CC cells. Finally, MCF-7
TXT cells express relatively higher levels of β-2 and β-4 tubulin and relatively lower levels of β-3 tubulin than both MCF-7
CC and MCF-7
DOX cells. The subcellular localization of various β-tubulin isoforms in MCF-7
TXT cells is also different from that in MCF-7
CC and MCF-7
DOX cells.
Discussion and conclusion
To better understand the mechanisms underlying acquired resistance to taxanes in breast cancer, we utilized previously established cell lines in which MCF-7 breast cancer cells were selected for survival in increasing concentrations of doxorubicin (MCF-7
DOX cells) or docetaxel (MCF-7
TXT cells) [
25]. A cell line selected under identical conditions in the absence of drug (MCF-7
CC) was used as a control.
We showed that MCF-7
TXT cells that are resistant to docetaxel are cross resistant to other drugs in the same group such as paclitaxel, but are not resistant to doxorubicin, a different type of cancer drug (Figure
1). Similarly, we showed that MCF-7
DOX cells are resistant to doxorubicin, but not resistant to both docetaxel and paclitaxel. These results demonstrate that the acquired chemoresistance in this instance is specific to the selection agent and it is not a consequence of the establishment of mechanisms of multidrug resistance. Our finding is different from a previous report showing that drug resistant MCF-7 cells lines also develops cross-resistance to structurally unrelated cancer drugs [
27]. However, in this previous report, it is also shown that selected paclitaxel-resistant MCF-7 cell is not cross-resistant to doxorubicin, which is consistent to our data. Nevertheless, our data suggest that the acquired resistance can be specific and chemotherapy using combined drugs or alternative drugs may overcome the resistance. Indeed, sequential single-agent therapy or combination therapy have been used in breast cancer treatment to overcome drug resistance [
28].
We further showed that the selected chemoresistant cell lines do have higher expression level of certain ABC transporter proteins (Figure
2). The expression level of ABCB1 is very high only in MCF-7
TXT cells and the expression level of ABCC1 is very high only in MCF-7
DOX cells. The expression level of ABCG2 is similar in both the selected chemoresistant and the parental MCF-7 cell lines and likely did not play a role in the drug-resistant phenotypes of these cell lines. These observations are consistent with our previous findings regarding the transcription of these ABC transporter genes in various selected MCF-7 cells [
25]. We also showed previously that MCF-7
TXT cells have lower accumulation of paclitaxel and MCF-7
DOX cells have lower accumulation of doxorubicin [
25]. While all of these three ABC proteins have been implicated in multiple drug resistance including taxanes and doxorubicin [
1,
3,
4,
14,
29], our results suggest that the specific member of ABC transporter proteins that are induced during the selection process may be different depending upon the selection agent. Our results indicate that the resistance to taxanes in MCF-7
TXT cells is associated with high expression level of the ABCB1 protein, but not ABCC1 and ABCG2, which is consistent with previous findings in other cell lines [
19]. Although ABCB1, ABCC1 and ABCG2 are all implicated in the resistance to doxorubicin [
19], the lack of cross-resistance to doxorubicin in MCF-7
TXT cells suggests that the ABCB1 overexpression alone may not efficiently mediate the efflux of doxorubicin in this selected MCF-7 cell line. Moreover, the overexpression of ABCC1, but not ABCB1 in MCF-7
DOX cells suggests that the resistance to doxorubicin is associated with high expression of the ABCC1 protein and lack of overexpression of ABCB1 does not block the ability of cells to acquire the resistance to doxorubicin. The lack of cross-resistance to taxanes in MCF-7
DOX cells also support the previous finding that ABCC1 may be more efficient in mediating the efflux of doxorubicin, but not taxane [
19,
28]. Given that the expression of ABCG2 protein was similar amongst MCF-7CC, MCF-7TXT, and MCF-7DOX cells, our data also suggest that resistance to either taxanes or doxorubicin is unrelated to ABCG2 expression. Our observation that treatment of the above cell lines with docetaxel for 24 h did not alter the expression and localization of these ABC proteins (Figure
3) also suggests that docetaxel itself cannot induce the expression of ABC transporter proteins.
It is also likely that the overexpression of any one of the above ABC proteins is not sufficient to confer the resistance to chemotherapy and other mechanisms are responsible for the resistance. Indeed, as we showed previously, while the drug resistance is related to the expression of drug transporters and the drug accumulations in the cells, drug transporter inhibitors are insufficient to fully restore sensitivity of MCF-7
DOX cells to doxorubicin or MCF-7
TXT cells to docetaxel [
25]. A series of experiments are performed in this study to provide insight into the additional mechanisms underlying resistance to taxanes in these cells.
We show in this study that the acquired resistance of MCF-7
TXT cells to taxanes as revealed in cytotoxicity assay (Figure
1) is also associated with resistance to taxane-induced apoptosis as revealed by quantification of chromosome condensation (Figure
4). While some studies suggest that MCF-7 cells are unable to go apoptosis due to the deletion of caspase-3 gene and thus the lack of caspase-3 protein, other studies indicate that MCF-7 cells are able to go apoptosis due to the existence of caspase-3 independent apoptotic pathways [
30‐
36]. We found in the cytotoxicity assay that the total of surviving cells follow a two-phase decrease with the increase of the concentration of docetaxel and paclitaxel (Figure
1). A sharp decrease at the low dose range (< 1 μM) and a further decrease at the high dose rage (> 10 μM) following a plateau at the middle dose range (1-10 μM). This observation is consistent with our previous report [
37]. We further showed in this study that taxane treatment arrested the cells at M phase at the dosage lower than 1 μM, which eventually leads to cell apoptosis (Figure
4A). The first phase of decrease in cell population (Figure
1) is coincident with the increase of cell apoptosis induced by the treatment of docetaxel and paclitaxel (Figure
4), which suggest that taxane-induced apoptosis is likely responsible for the reduction of cell population at the doses less than 1 μM. On the other hand, treatment of the cells with doxorubicin did not arrest the cells at M phase and did not induce significant cell apoptosis (Figure
5). It has been reported that doxorubicin interacts with topoisomerases I and II [
10] leading to DNA damage followed by G1 and G2 growth arrest, which has been proposed to correlate with tumor response and patient’s outcome [
1,
11]. The few apoptotic cells induced by doxorubicin observed under a fluorescence microscope could be due to the detaching of the apoptotic cells from the coverslip.
The interaction between taxanes and the microtubules stabilises microtubules by reducing depolymerisation. Stabilization of microtubules by taxane binding prevents normal formation of mitotic spindles [
6]. This leads to chronic activation of the spindle assembly checkpoint (SAC), which in turn leads to mitotic arrest [
7]. Extended mitotic arrest eventually leads to cell death [
8]. We studied the effects of taxanes on the formation of microtubules and the mitotic spindles in MCF-7
TXT, MCF-7
DOX and MCF-7
CC cells by indirect immunofluorescence (Figure
4A). We showed that the abnormal mitotic spindles induced by taxane treatment are accompanied by the arrest of cells at M phase and the initiation of cell apoptosis (nuclear condensation). While 10 - 100 nM of docetaxel and paclitaxel induced significant mitotic spindle disruption and M phase arrest in MCF-7
CC and MCF-7
DOX cells, ten times higher concentrations of docetaxel and paclitaxel are needed to induce a similar level of mitotic spindle disruption and M phase arrest in MCF-7
TXT cells. Thus, our data clearly suggests that the acquired resistance to taxanes in MCF-7
TXT cells is due to the resistance to taxane-induced mitotic spindle disruption and M phase arrest.
We also examined microtubule dynamics in MCF-7
TXT cells (Figure
7, Additional file
2: Figure S1, Additional file
4: Figure S2, Additional file
5: Figure S3 and Additional file
3: Video S4-10). We showed that in the absence of docetaxel treatment the microtubule dynamics are robust in both MCF-7
TXT and MCF-7
CC cells, but the microtubule dynamics are weaker in MCF-7
TXT cells than that in MCF-7
CC cells. Moreover, microtubule dynamics are greatly more insensitive to docetaxel in MCF-7
TXT calls than in MCF-7
CC cells. For example, treatment with 0.5 μM docetaxel only slightly reduces the microtubule dynamics in MCF-7
TXT cells, but significantly reduced both the shortening and extending rate of microtubules in MCF-7
CC cells. Our findings suggest that the resistant MCF-7
TXT cells have unique microtubule dynamics that are likely unrelated to the overexpression of ABC transporters. The insensitivity of microtubules to docetaxel treatment in MCF-7
TXT cells may be partially the reason that docetaxel is less effective in inducing the M-phase arrest and the apoptosis in MCF-7
TXT cells in comparison to MCF-7
CC cells. This unique microtubule dynamics may contribute to the resistance to docetaxel.
Although taxane-binding sites on microtubules are only present in assembled tubulin [
38], the stabilized microtubules are not able to extend without depolymerisation. It is interesting to note that 0.5 μM docetaxel induces very high level M-phase arrest at MCF-7
TXT cells (Figures
1 and
4), but does not significantly reduce the microtubule dynamics in MCF-7
TXT cells (Figure
7). Similarly, 100 nM of docetaxel induces very high level M-phase arrest at MCF-7
CC cells (Figures
1 and
4), but does not significantly reduce the microtubule dynamics in MCF-7
CC cells (Figure
7). The reason for this could be the duration of the treatment. We assayed the M-phase arrest and cell apoptosis following the treatment for 24 h, but we assayed the microtubule dynamics only after the treatment for 1 h. Besides, multiple mechanisms are involved in the acquired resistance to docetaxel, microtubule dynamics is just one of these mechanisms.
Finally, we showed that the all four β-tubulin isoforms are expressed in the three MCF-7 cell lines. While the relative expression levels of the four β-tubulin isoforms are very similar between MCF-7
DOX and MCF-7
CC cells, the relative expression levels of the β-tubulin isoforms are quite different in MCF-7
TXT cells (Figure
8). MCF-7
TXT cells have relatively higher β2- and β4-tubulin expression and relatively lower β3-tubulin expression level (Figure
8). These results suggest that the expression level of various β-tubulin isoforms is related to the microtubule dynamics of the MCF-7 cells in response to docetaxel treatment. The expression levels of various tubulin isoforms have been linked to the resistance to taxanes in breast cancers. It has been reported that both β3- and β4-tubulin are overexpressed in a MCF-7 cell line selected for resistant to paclitaxel under increased paclitaxel concentration [
21]. The overexpression of β3-tubulin induces paclitaxel resistance by reducing the ability of paclitaxel to suppress microtubule dynamics [
20]. It is also reported that mRNA levels of β2-, β3- and β4-tubulin are significantly upregulated in paclitaxel- and docetaxel-resistant MCF-7 cells [
23]. The MCF-7
TXT cell line used in this research is selected for resistance to docetaxel, but showed similar resistance to paclitaxel. While we also found that MCF-7
TXT cells have higher β2- and β4-tubulin expression than MCF-7
CC and MCF-7
DOX cells, we actually showed that the β3-tubulin expression level is lower. The significant difference in the expression levels of various β-tubulin isoforms suggest that the composition of β-tubulin in the formation of microtubules may contribute to the microtubule dynamics and its response to taxane treatment, which could be part of the mechanisms underlying the acquired resistance to taxanes in breast cancer cells.
We also examined the localization of these tubulin isoforms. We showed that the localization pattern of the various β-tubulin isoforms in MCF-7
TXT cells is different from that of MCF-7
CC and MCF-7
DOX cells (Figure
8B). While we did not know how the different subcellular distribution of these β-tubulin isoforms affects its response to docetaxel treatment, it is possible that the relative composition of various β-tubulin isoforms in microtubules and their formation pattern may play a role in determining microtubule dynamics and sensitivity of microtubules to docetaxel treatment.
As shown from this study, multiple mechanisms are likely involved in the acquired drug resistance. Besides discussed above, many other proteins and mechanisms may also be involved in the acquired drug resistance. It has been shown that extracellular matrix proteins, apoptosis related proteins, cytokine and growth factor signaling proteins, and many other proteins are overexpressed in the selected MCF-7 cells resistant to taxanes and other cancer drugs [
29,
39,
40]. It is interesting to find out in the future study whether and how these different mechanisms regulate drug-resistance independently or coordinately.
In conclusion, our results suggest the presence of multiple mechanisms for acquired drug resistance to taxanes. Prolonged exposure to taxanes may result in the selection of the breast cancer cells that overexpress certain drug resistance proteins, such as ABCB1 in MCF-7TXT cells, which will lower the taxane level inside the cells and thus contribute to the resistance to taxanes. Prolonged exposure to taxanes may also result in the selection of the breast cancer cells that have differential expression of various β-tubulin isoforms, such as higher β-2 and β-4 and lower β-3 tubulin in MCF-7TXT cells. In addition, the relative composition of various β-tubulin isoforms within the microtubules and the specific distribution of these β-tubulin isoforms along the microtubules may determine the dynamics of the microtubules and its sensitivity to taxane treatment. For example, in MCF-7TXT cells, the distinct distribution of the β-tubulin isoforms can be related to the weak microtubule dynamics and its insensitivity to taxane treatment.
Authors’ contributions
HW: Participating in the design of the project, performing most of the experiments including cell culture, cytotoxicity assay, immunoblotting, immunofluorescence, and living image, and participating in the data analysis and manuscript writing. TV: Performing experiments including cell culture and immunoblotting, and participating in data analysis. AH: Performing experiments including cell culture, cytotoxicity assay, and immunoblotting, and participating in data analysis. SL: Performing experiments including cell culture, cytotoxicity assay, and immunoblotting, and participating in data analysis. XC: Performing experiments including cell culture, cytotoxicity assay, immunoblotting, immunofluorescence, and living image, and participating in data analysis. AMP: Providing drug-resistance cell lines, participating in the design of the project, and participating in the writing of the manuscript. DNB: Participating in the design of the project, data analysis and the writing of the manuscript. ZW: Participating in the design of the project, performing some experiments including living image and immunofluorescence, and participating in data analysis and the writing of the manuscript. All authors read and approved the final manuscript.