Class III β-tubulin expression and in vitro resistance to microtubule targeting agents

Background: Class III β-tubulin overexpression is a marker of resistance to microtubule disruptors in vitro, in vivo and in the clinic for many cancers, including breast cancer. The aims of this study were to develop a new model of class III β-tubulin expression, avoiding the toxicity associated with chronic overexpression of class III β-tubulin, and study the efficacy of a panel of clinical and pre-clinical drugs in this model. Methods: MCF-7 (ER+ve) and MDA-MB-231 (ER−ve) were either transfected with pALTER-TUBB3 or siRNA-tubb3 and 24 h later exposed to test compounds for a further 96 h for proliferation studies. RT–PCR and immunoblotting were used to monitor the changes in class III β-tubulin mRNA and protein expression. Results: The model allowed for subtle changes in class III β-tubulin expression to be achieved, which had no direct effect on the viability of the cells. Class III β-tubulin overexpression conferred resistance to paclitaxel and vinorelbine, whereas downregulation of class III β-tubulin rendered cells more sensitive to these two drugs. The efficacy of the colchicine-site binding agents, 2-MeOE2, colchicine, STX140, ENMD1198 and STX243 was unaffected by the changes in class III β-tubulin expression. Conclusion: These data indicate that the effect of class III β-tubulin overexpression may depend on where the drug’s binding site is located on the tubulin. Therefore, this study highlights for the first time the potential key role of targeting the colchicine-binding site, to develop new treatment modalities for taxane-refractory breast cancer.

The high percentage of non-responders and failures following initial responses to taxanes (taxotere and paclitaxel) highlights the critical role played by drug resistance mechanisms in breast cancer progression. Although many mechanisms associated with resistance have been proposed for taxane-refractory cells, only two have been found in the clinic to date: the overexpression of ABC proteins such as the P-glycoprotein (MDR1) and alterations in tubulin-isoform expression (Correnti et al, 1995;Linn et al, 1995;Kavallaris et al, 1997;Mozzetti et al, 2005;Paradiso et al., 2005). Mutations within tubulin at the paclitaxel-binding site have been identified (Giannakakou et al, 1997), which were thought to be associated with taxane resistance. However, subsequent studies showed that these early findings resulted from the amplification of tubulin pseudogenes (Noguchi, 2006). Further studies have failed to confirm the presence of tubulin point mutations in patients with lung or ovarian cancers resistant to therapy (Sale et al, 2002;Tsurutani et al, 2002;Lamendola et al, 2003). Microtubuletargeting agents bind to the b-tubulin subunit of the a/b-tubulin heterodimers that assemble to form microtubules. In humans, at least seven distinct b-tubulin isotypes have been reported. Altered patterns of expression are seen in cancer. In ovarian cancer patients class III b-tubulin mRNA is significantly upregulated in the taxane-resistant tumours compared with biopsy samples from untreated tumours (Kavallaris et al, 1997). In a separate study, class III b-tubulin mRNA and protein were shown to be significantly upregulated in a sub-set of paclitaxel resistant ovarian cancer patients . In breast cancer high expression of class III b-tubulin is a predictive biomarker of clinical paclitaxel resistance (Tommasi et al, 2007). In gastric cancer patients, whose tumours were positive for class III b-tubulin expression were significantly less likely to respond to docetaxel-based chemotherapy, 16.7 vs 64.3% response rate, respectively. Therefore, class III b-tubulin can be regarded as a predictive marker for the clinical response to docetaxel-based chemotherapy in gastric cancer (Urano et al, 2006). The ongoing development of new agents, not sensitive to class III b-tubulin overexpression, may provide new treatment options for treatmentrefractory cancer.
In this study, the effects of seven microtubule targeting agents on the proliferation of MCF-7 and MDA-MB-231 breast carcinoma cells over-or underexpressing class III b-tubulin were investigated. Five of the agents, 2-methoxyoestradiol-3,17-O, O-bis-sulfamate (STX140), 2-ethyloestradiol-3,17-O,O-bis-sulfamate (STX243), 2-methoxyoestradiol (Panzem), colchicine and ENMD1198 are known to target the colchicine-binding site on tubulin, the other two target the taxane-binding site (paclitaxel) and the vinca-binding site (vinorelbine). These seven agents represent the currently used drugs in the clinic (paclitaxel and vinorelbine), those drugs that are currently in/have recently completed phase I trials (2-MeOE2 and ENMD1198) and a new generation of orally bioavailable compounds which are in advanced pre-clinical development (STX140 and STX243).

Class III b-tubulin cloning
The protein-expression vector (pALTER-TUBB3) was prepared by inserting the full-length human class III b-tubulin cDNA fragment from pOTB7-TUBB3 (LGC Promochem) into pALTER Max (Promega, Southampton, UK). The construct was checked by digestion and sequencing.
Transfection MCF-7 and MDA-MB-231 cells were transfected with the wildtype class III b-tubulin (pALTER-TUBB3) or with the class III b-tubulin siGenome Smartpool (siRNA-tubb3, ref. M-020099-03, Dharmacon, Cramlington, UK) using the Amaxa nucleofection technology (Amaxa GmbH, Koeln, Germany). The nucleofaction protocol for plasmid transfection was optimised using a GFPconstruct with efficacies of 47 and 70% being achieved after 24 h transfection in MCF-7 and MDA-MB-231 cells, respectively. siRNA nucleofection was carried out using a pre-optimised protocol (Mhaidat et al, 2008). Briefly, the cells were trypsinised and washed with PBS. The required number of cells (2 Â 10 6 cells per nucleofection) were harvested, resuspended in nucleofactor solution (100 ml per nucleofection) and mixed with 2 mg plasmid or 4 ml of siRNA (20 mM). Cells were then electroporated in the Amaxa nucleofector and transferred to fresh pre-warmed medium. The cells were incubated for 1 h at 371C before being seeded in 96-well plates for proliferation assay or in T-25 flasks for protein or RNA extraction. Transfection with the empty pALTER-Max (mock transfection) and a non-targeting siRNA (siGENOME Non-targeting siRNA, ref. D-001210-01, Dharmacon) had no significant effect on the proliferation of either cell line (data not shown).

Proliferation assay
The cells were seeded into 96-well culture plates at 4000 cells per well, incubated at 371C for 24 h and exposed to compounds for 96 h at 371C. Cell proliferation was measured by adding 10 ml AlamarBlue reagent (BioSource International, Camarillo, CA, USA). The viable cells are counted with a spectrofluorometric microtiter-well plate reader Fluostar optima (BMG Labtech GmbH, Offenburg, Germany) at 544 nm (excitation) and 590 nm (emission). IC 50 values were obtained from a graph of percent proliferation vs inhibitor concentration and calculated using Prism (Graphpad Software, La Jolla, CA, USA).

Real time RT -PCR
Cells were lysed with 350 ml of RNeasy lysis buffer (Qiagen, Ulm, Germany) plus 1% b-mercaptoethanol (Promega) and homogenised with the QIAshredder columns (Qiagen). The RNA was isolated from the homogenate using the RNeasy kit (Qiagen).
A 5 mg aliquot of each RNA sample was reverse transcribed using the First-Strand cDNA Synthesis kit (GE Healthcare). Reverse-transcription -PCRs were carried out in a Rotor Gene Immunohistochemistry MCF-7 and MDA-MB-231 cells grown on poly-L-lysine-coated coverslips were exposed to compounds for 24 h. The cells were fixed in methanol at À201C for 10 min, and then in ice-cold acetone for 30 s. Coverslips were rinsed and then rehydrated with PBS. To visualise the microtubules, cells were incubated for 1 h with an a-tubulin-specific FITC-conjugated antibody (ref. F2168, clone DM1A; Sigma) in PBS containing 1% BSA. Cells were washed twice with PBS and counterstained with Hoechst 33342 at 1 mg ml À1 (Sigma) to stain the nuclei. The coverslips were mounted on slides using DePex mounting medium, observed using a Zeiss inverted microscope ( Â 200) and analysed with the Axiovision imaging system.

Statistics
In vitro experiments were carried out in triplicate and data presented are representative of three or more independent experiments. All errors shown are the mean±s.d. The Student's t-test was used to assess significance. and was 2.1±0.74-fold higher than the control level 5 days after transfection (Po0.05, Figure 2B) relative to MDA-MB-231 mock-transfected cells. Transfection with siRNA-tubb3 blocked the class III b-tubulin mRNA expression in both cell lines ( Figure  2A and B). Class III b-tubulin mRNA levels were reduced by 50 ± 16% in MCF-7-situbb3 cells 24 h after transfection (Po0.05) and by 46 ± 16% 5 days after transfection (Po0.05). Class III b-tubulin mRNA levels were also reduced by 36±15% in MDA-MB-231-situbb3 cells 24 h after transfection (Po0.01), but were not significantly different from control cells 5 days after transfection (90±47%).

RESULTS
To ensure changes in mRNA level correlated with protein expression, MCF-7 and MDA-MB-231 cells were transfected with either pALTER-TUBB3 or siRNA-tubb3 and the proteins were extracted 24 h and 5 days after transfection. Equal protein amounts were analysed by SDS -PAGE and immunoblotting with the human anti-class III b-tubulin or the human anti-class I -V b-tubulin. Staining with Ponceau S and immunoblotting with an antia-tubulin was used to ensure even protein loading and transfer. Class III b-tubulin protein expression was higher in both cell lines transfected with the pALTER-TUBB3 vector ( Figure 2B and C, lane 2) and was lower in cells transfected with the class III b-tubulin siRNA ( Figure 2B and C, lane 3) compared with the control cells ( Figure 2B and C, lane 1). These changes in protein expression were still observed 5 days after transfections ( Figure 2B    Proliferation assays MCF-7 and MDA-MB-231 transfected cells were seeded at 4000 cells per well, incubated for 4 days and the percentage of proliferation was monitored by using the Alamar blue reagent. Class III b-tubulin overexpression or silencing did not significantly affect the viability of the cells (the proliferation of MCF-7-TUBB3, MCF-7-situbb3, MDA-MB-231-TUBB3 and MDA-MB-231-situbb3 cells was 105, 101, 97 and 98% respectively, the proliferation rate of control cells, ns). The cytotoxicity efficacy of STX140, STX243, 2-MeOE2, ENMD1198 and colchicine was not significantly affected by overexpression or silencing of class III b-tubulin (Figures 3  and 4, graphs A, B, D, E and G), and the IC 50 values of these compounds were not significantly different from IC 50 values calculated in control cells (Table 1a and b). Although MDA-MB-231 cells underexpressing the class III b-tubulin were more sensitive to colchicine than the mock-transfected cells when exposed to high concentrations (0.1 -10 mM), the IC 50 values in both type of cells were not significantly different (IC 50 ¼ 37 nM in MDA-MB-231-mock and IC 50 ¼ 29 nM in MDA-MB-231-situbb3, Table 1b). Class III b-tubulin expression altered the sensitivity to paclitaxel in both cell lines ( Figures 3C and 4C). MCF-7 and MDA-MB-231 cells overexpressing class III b-tubulin were the less sensitive to paclitaxel, with IC 50 values of 0.1 nM and 4 pM, respectively (Table  1a and b). These IC 50 values were significantly higher (Po0.01) than those obtained in untreated MCF-7 and MDA-MB-231 cells (7.2 and 2 pM, respectively). In contrast, MCF-7 and MDA-MB-231 cells underexpressing class III b-tubulin were found to be the most sensitive to paclitaxel. Paclitaxel's IC 50 value was significantly reduced in MCF-7-situbb3 cells (IC 50 ¼ 4.8 pM, Po0.05) compared to the IC 50 value in control cells. Despite reducing the paclitaxel concentration to 0.1 pM, no IC 50 values could be obtained for the MDA-MB-231 cells underexpressing class III b-tubulin as the inhibition of growth was greater than 50% (47%, Figure 4C). At higher concentrations (1 and 10 pM), the class III b-tubulin silencing rendered MDA-MB-231 cells significantly more sensitive to paclitaxel than the control cells (Po0.05 and Po0.01, respectively, Figure 4C).
The modification of the b-tubulin expression profile also altered the efficacy of vinorelbine (Figures 3F and 4F). When MCF-7 and MDA-MB-231 cells were transfected with pALTER-TUBB3, the IC 50 values for vinorelbine were higher than the IC 50 values in the

Microtubule structure
To investigate whether class III b-tubulin resistance is reflected at the cytoskeleton structure level, cells overexpressing class III b-tubulin were exposed to STX140, paclitaxel and colchicine for 48 h at concentrations close to IC 50 values obtained in control cells. Nuclear and cytoskeletal morphology was then examined by immuno-fluorescence microscopy. The untreated cells showed typical nuclear and cytoskeleton structures, with long and dense microtubules extending throughout the cytoplasm. No significant difference was observed in the morphology of both untreated cell lines overexpressing class III b-tubulin compared with control cells (Figure 5). STX140 and colchicine caused microtubule disruption with bundles of microtubules being observed and overexpression of class III b-tubulin did not affect the morphological changes in either cell line in response to these compounds. In the nontransfected MCF-7 control cells paclitaxel caused the virtual destruction of the normal microtubule structure with only very small and dense bundles of microtubules being visible. Furthermore, many abnormal nuclei were observed in comparison to untreated cells and those exposed to either STX140 or colchicine. Morphological changes were seen in the MCF-7 class III b-tubulin overexpressing cells in response to paclitaxel, with shortened but unusually highly defined microtubules and many abnormal nuclei present. The unusual microtubule structures seen in the MCF-7 class III b-tubulin overexpressing cells in response to paclitaxel were not observed in their MDA-231 counterparts which far more closely resemble the non-transfected MDA-231 control cells after paclitaxel exposure. Although tubulin III transfection renders both cells lines significantly more resistant to paclitaxel, effects on microtubule structure were seen in this study as the concentration of paclitaxel used was still cytotoxic. Further studies are planned to understand the unusual structures observed in the MCF-7 class III b-tubulin overexpressing cells.

DISCUSSION
Tubulin-binding agents constitute an important class of compound in chemotherapy, but their clinical success has been compromised by the emergence of drug resistance, which is derived from several mechanisms that can be de novo or acquired. Tubulin-binding agents target the microtubule network and two different resistance mechanisms have been associated to the taxane-resistant phenotype in cancer patients: the overexpression of ATP-binding cassette family of proteins such as P-glycoprotein and alterations in tubulin-isoform expression. Previous studies showed that some compounds such as STX140, are able to inhibit the proliferation of resistant MCF-7dox cells (P-gly þ ve) by arresting them in the G2/M phase (Suzuki et al, 2003) and to block the growth of MCF-7dox tumours . This study is the first to report the effects of changes in the tubulinisoform expression profile on the efficacy of new tubulin-binding agents, including Panzem (2-MeOE2), ENMD1198 (currently in phase I clinical trials) and STX140 in MCF-7 and MDA-MB-231 cells.
Identifying the role of class III b-tubulin in drug resistance is difficult because of the complexity of tubulin auto-regulation. Expression of b-tubulin isotypes is tightly regulated through co-translational degradation of b-tubulin mRNAs in response to an increase in the level of soluble tubulin (Yen et al, 1988;Theodorakis and Cleveland, 1992). This may explain why overexpression of class III b-tubulin failed to confer resistance to paclitaxel, estramustine and vinblastine, among other drugs tested in prostate cells (Ranganathan et al, 2001). In addition, too high expression of class III b-tubulin in cells can be toxic, further complicating this methodology (Hari et al, 2003). To negate this problem Hari et al (2003) inducibly overexpressed class III b-tubulin in CHO cells, this decreased the paclitaxel suppression of microtubule dynamics and caused a limited resistance to paclitaxel (Kamath et al, 2005). Recently, cervical carcinoma HeLa cells were stably transfected with class III b-tubulin  and were used to test taccalonolides, which were shown not to be susceptible to class III b-tubulin resistance (Risinger et al, 2008). In this study, class III b-tubulin protein levels was only 1.5-fold increased in cells transfected with pALTER-TUBB3. This may have protected the cells against the toxicity of high expression and enabled them to grow with a similar proliferation rate than control cells. However, the overexpression of class III b-tubulin was significant and this degree of overexpression was sufficient to confer a resistance to paclitaxel. These data are supported by the relatively small increase (five-fold) seen in patients with docetaxelrefractory breast cancer (Hasegawa et al, 2003). This study demonstrates that a variation in the b-tubulin expression pattern in two breast cancer cell lines does not affect the efficacies of STX140, STX243, 2-MeOE2, ENMD1198 and colchicine, whereas the efficacy of vinorelbine and paclitaxel are altered by both class III b-tubulin overexpression and silencing. These results correlate with previous studies using siRNA approaches which showed that class III b-tubulin induces resistance to paclitaxel, vincristine and DNA-damaging agents in non-small cell lung carcinoma (Kavallaris et al, 1999;Gan et al, 2007).
The key difference between class I and class III b-tubulin is an amino-acid substitution leading to a different three dimensional conformation. Class III b-tubulin, which presents only 92% similarity with the other b-tubulin isoforms, bears an Arg 277 instead of the Ser 277 present in class I b-tubulin. Ser 277 and Arg 278 are essential for the stable binding of paclitaxel to its binding site in class I b-tubulin (Lowe et al, 2001). In contrast, the structure of the loop of the class III b-tubulin is altered by the substitution of Ser 277 with Arg and thus, prevents a stable paclitaxel binding . It has been shown that Class III b-tubulin also generates more dynamic microtubules and counteracts proassembling activity of taxanes at the plus ends of microtubules (Kamath et al, 2005), thereby making microtubules more resistant to the stabilisation of microtubule dynamics operated by taxanes (Hari et al, 2003;Kamath et al, 2005;Gan et al, 2007). In contrast, STX140, STX243, 2-MeOE2, ENMD1198 and colchicine are structurally different to paclitaxel; they bind to the colchicine site and they have a conformation, which may allow the formation of a stable complex with class III b-tubulin. Moreover, these agents depolymerise microtubules, whereas paclitaxel stabilises them. This may be a great advantage when class III b-tubulin is overexpressed and the microtubule dynamic is amplified. However, vinorelbine, which is also a microtubule-destabilising agent but binds to the Vinca alkaloid binding site, has its efficacy reduced by the overexpression of class III b-tubulin in vitro in breast carcinoma cells (as shown in this study) and in small lung cancer cells (Gan et al, 2007) and in vivo (Seve et al, 2005). Therefore, class III b-tubulin resistance might be binding-site related and agents binding to the colchicine-binding site might avoid this resistance. This theory is supported by a recent study, wherein the authors showed that HeLa cells overexpressing class III b-tubulin are less sensitive to paclitaxel, docetaxel, epothilone B and vinblastine. Furthermore, the efficacy of the taccalonolides, which do not interact with microtubules despite having paclitaxellike effects in cells, is unaffected by class III b-tubulin expression (Risinger et al, 2008).
In summary, this study shows that microtubule disrupting agents binding to the colchicine-binding site circumvent class III b-tubulin resistance, whereas agents binding to the taxol-or the Vinca-binding sites have their efficacies altered by changes in class III b-tubulin expression. Class III b-tubulin is associated with drug resistance and poor prognosis in patients (Hasegawa et al, 2003;Mozzetti et al, 2005;Seve et al, 2005;Urano et al, 2006;Tommasi et al, 2007) and it is a predictive factor of event-free and overall survival (Seve and Dumontet, 2008). Moreover, class III b-tubulin is involved in the sensitivity of both taxanes and Vinca alkaloids, but it also has a role in chemosensitivity of DNA-damaging agents (Gan et al, 2007). This study therefore highlights the importance of developing new agents binding to the colchicine-binding site, such as STX140 and STX243 as an alternative to taxanes for refractory cancer treatment.