Tubulin-Targeted Drug Action : Functional Significance of Class II and Class IVb B-Tubulin in Vinca Alkaloid Sensitivity

Aberrant expression of B-tubulin isotypes is frequently described in tumor tissues and tubulin-binding agent (TBA)–resistant cell lines. There is limited understanding of the role of specific B-tubulin isotypes in cellular sensitivity to TBAs, and to gain insights into the functional role of BIIand BIVb-tubulin, we examined these isotypes in lung cancer cell lines NCI-H460 (H460) and Calu-6. Drug-treated clonogenic assays revealed that small interfering RNA–mediated knockdown of either BIIor BIVb-tubulin hypersensitized the lung cancer cell lines to Vinca alkaloids, with the effects more pronounced following BIVb-tubulin knockdown. In contrast, there was no change in paclitaxel sensitivity following knockdown of either isotype. Cell cycle analysis revealed a greater propensity for the BIIand BIVb-tubulin knockdown cells to undergo G2-M cell cycle block following 5 nmol/L vincristine treatment, with the BIVb knockdown cells being more sensitive than the BII-tubulin knockdown cells compared with control. In contrast to BII-tubulin knockdown, BIVb-tubulin knockdown cells showed a significant increase in the sub-G1 population (cell death) following treatment with both 5 and 40 nmol/L of vincristine compared with controls. Importantly, BIVb-tubulin knockdown in H460 cells caused a significant dose-dependent increase in Annexin V staining in response to vincristine but not paclitaxel. Therefore, increased sensitivity to induction of apoptosis is one mechanism underlying the Vinca alkaloid hypersensitivity. This study provides direct evidence that BIIor BIVb-tubulins have functionally distinct roles and expression of these isotypes may serve as strong predictors of Vinca alkaloid response and resistance. [Cancer Res 2008;68(23):9817–24]


Introduction
Tubulin-binding agents (TBA) disrupt microtubule dynamics by binding to the h-tubulin subunit on a/h-tubulin, inducing mitotic arrest and cell death (1).Clinically important TBAs can be subdivided into two broad groups, microtubule-stabilizing agents (e.g., paclitaxel, docetaxel, and epothilones) and microtubuledestabilizing agents (e.g., vincristine, vinblastine, and vinorelbine), all of which are used to treat a range of hematologic and solid tumors.Despite the success of TBAs in the clinic, chemoresistance can be a significant barrier to cure.Drug resistance is multifactorial and the development of resistance to TBAs has been associated with increased expression of ATP-binding cassette transporters, alterations in apoptosis pathways, and alterations in tubulin/ microtubule proteins (2,3).
Altered expression of specific h-tubulin isotypes has been reported in tumors and is associated with resistance to TBAs (2,3).Despite strong correlational evidence, the functional significance of differential expression of h-tubulin isotypes on microtubule morphology, cell cycle, and response to TBAs is not well understood.The cellular target of TBAs is h-tubulin and there are at least seven h-tubulin isotypes, encoded by multiple genes that display tissue-specific expression.For example, hI-tubulin is constitutively expressed in many tissues; hII-tubulin is found at high levels in brain and low levels in various tissues; hIII-tubulin is expressed in neurons and Sertoli cells; and hIV-tubulin consists of two isotypes, hIVa and hIVb, with hIVa restricted to brain and hIVb found in most tissues (4).The h-tubulin isotypes are highly homologous and conserved across species and differ predominately in their last 15 amino acids at the COOH-terminal end (4).The differential tissue expression and highly conserved sequences of h-tubulin isotypes across all species suggest that each isotype may provide unique characteristics that impart functional differences to microtubules.The potential role of h-tubulin isotypes is supported by evidence from in vitro studies using purified bovine brain tubulin that showed that microtubules composed of different tubulin isotypes contribute to differences in assembly properties, microtubule dynamics, and drug binding (5)(6)(7)(8).Although the in vivo significance of this remains uncertain, inherent functional differences among isotype classes could potentially alter the cellular response to TBAs.Addressing this issue is of clinical importance as it could provide a means to predict the response of tumors to TBAs.
We have recently shown that in addition to mediating sensitivity to diverse classes of TBAs, hIII-tubulin also mediates sensitivity to DNA-damaging agents by behaving as a cellular survival factor (9). Previous studies have reported altered patterns of h-tubulin isotype expression in cell lines, which have been selected for resistance to TBAs (reviewed in refs.10,11).To date, studies have largely focused on the role of hIII-tubulin in clinical resistance to paclitaxel or vinorelbine in epithelial tumors (reviewed in refs.3,12).Nevertheless, the relevance of other non-hIII-tubulin isotypes, such as hIIand hIVb-tubulin, in response to TBAs remains to be determined.
Increased expression of hII-tubulin has been associated with resistance to paclitaxel (13,14).We have previously identified decreased expression of a hII-tubulin isoform by two-dimensional gel electrophoresis and mass spectrometry in both vincristine-and vinblastine-resistant leukemia cell lines, although total hII-tubulin protein levels were unchanged (15,16).To date, however, there is limited functional evidence that altered hII-tubulin expression is a major determinant of cellular sensitivity toward paclitaxel and/or Vinca alkaloids.Stable overexpression of hII-tubulin in Chinese hamster ovary (CHO) cells and NIH 3T3 cells did not confer resistance to paclitaxel (17,18).Several clinical studies have assessed the prognostic or predictive value of hII-tubulin in breast (19,20) and ovarian cancer (21).In a study involving 41 advanced breast cancer patients, hII-tubulin expression was correlated with lower docetaxel response rates (19), whereas hII-tubulin has been found in both normal and tumor breast tissues (20).More recent investigations have shown that the ratio of hII/hV-tubulin gene expression is higher in non-small cell lung cancer (NSCLC) tumor samples compared with normal lung tissues (22), suggesting that the broad distribution of hII-tubulin in tumor tissues may potentially suggest its relevance in disease progression and cellular response to TBAs.
Increased expression of hIV-tubulin has also been described in TBA-resistant cancer cell lines (14,(23)(24)(25)(26)(27).We have also previously reported overexpression of hIVa-tubulin in ovarian tumors (14).However, in many cases, there is no clear distinction between hIVaand hIVb-tubulins about their specific role in microtubule functions and drug response, with investigation in this area being hampered by the lack of specific antibodies to distinguish hIVaand hIVb-tubulin isotypes.Similar to hII-tubulin, overexpression of hIVb-tubulin in CHO cells does not affect paclitaxel sensitivity (17).Data on the contribution of hIVb-tubulin to Vinca alkaloid response are again limited.
To understand the functional significance of hIIand hIVbtubulin isotypes in drug response, we used RNA interference (RNAi) technology to specifically knock down the expression of these isotypes in NSCLC cell lines and characterize the effects on morphology, cell cycle, and drug-induced apoptosis.

Materials and Methods
Cell culture and small interfering RNA transfection.Human NSCLC cell lines Calu-6 and H460 were maintained as previously described (9).For gene knockdown studies, cells were transfected with hII-tubulin small interfering RNA (siRNA) ON-Target plus SMARTpool reagent ( final concentration, 25 nmol/L; Dharmacon) or hIVb-tubulin siRNA ( final concentration, 100 nmol/L; Dharmacon) using Lipofectamine 2000 (Invitrogen) following the manufacturer's instructions.All experiments involved the use of a Lipofectamine only control (mock-transfected) and a nonsilencing control siRNA (negative control; Qiagen) transfected at equivalent concentrations as the target siRNA.
Cytotoxic drugs.Cytotoxic drugs were obtained as previously described (9).Vinblastine (vinblastine sulfate; David Bull Laboratories) was prepared at a stock concentration of 1.1 mmol/L in saline [0.9% (w/v) NaCl].Vinflunine (kindly provided by Dr. B. Hill, Division of Experimental Cancer Research, Pierre Fabre Research Centre, Castres, France) was solubilized in water at a stock concentration of 2 mmol/L.
Reverse transcription-PCR analysis of B-tubulin isotypes.Total RNA was isolated using Trizol reagent (Invitrogen) following the manufacturer's instructions.RNA samples were DNase treated and reverse transcribed for reverse transcription-PCR (RT-PCR) analysis using methodology and specific primers as described (14).For Hb9 (hII), H5b (hIVa), and Hb2 (hIVb) genes, semiquantitative RT-PCR-based assay involved setting up two separate PCR tubes for target (h-tubulin) and control (h 2 -microglobulin) gene sequences.PCR amplifications were performed in triplicate using two independently prepared cDNA preparations.
Western blotting.Total cellular proteins from siRNA-transfected cells (10 Ag) were resolved on 12% SDS-PAGE and electrotransferred to nitrocellulose membrane as described previously (28).Western blot analyses and antibodies used have been reported previously (9).
Immunofluorescence staining.siRNA-transfected Calu-6 cells were grown on sterile chamber slides until 60% to 70% confluent.Cells were treated with vincristine for 1 h at 10 nmol/L.Cells were then fixed and processed for staining as previously described (29).Slides were viewed and images were captured using a Zeiss Axioplan 2 Immunofluorescence Microscope (Zeiss) fitted with a cooled charge-coupled device camera and Image-Pro Plus 4.1 software (Media Cybernetics, L.P.).
Drug-treated clonogenic assay.Briefly, cells were transfected with siRNA (25-100 nmol/L) for 24 h and 600 cells (Calu-6) or 150 cells (H460) were seeded into each well of a six-well plate and allowed to attach for 4 to 6 h before drug treatment.Drug-treated clonogenic assays were performed as previously described (9).The colonies were manually counted, and the results were presented as a surviving fraction (9).Dose-response curves were plotted for each drug-cell line combination and inhibitory dose (ID 50 ) values were extrapolated from this curve using GraphPad Prism program (version 4; GraphPad Software).ID 50 was defined as the concentration of the drug required to reduce the number of colonies in drug-treated wells to 50% to that of the relevant untreated controls.The ID 50 values were the means of at least three independent experiments.
Accumulation of [ 3 H]vincristine and Annexin V-FITC staining.These assays were performed as previously described (9).
Cell cycle analysis.Distribution of DNA content in siRNA-transfected H460 cells was determined by flow cytometry (9).After 72 h of siRNA transfection, cells were exposed to vincristine at the indicated concentrations for 24 h.On the day of analysis, both adherent and floating cells were harvested, washed with PBS, and then stained with a solution containing 0.4% Triton X-100 (Sigma-Aldrich), 50 Ag/mL propidium iodide (Sigma-Aldrich), and 2 Ag/mL DNase-free RNase (Roche) for 15 min at 37jC in the dark.DNA content was measured by a FACSCalibur flow cytometer (Becton Dickinson) and analyzed using CellQuest program as described (9).
Statistical analysis.Statistical analysis was performed using the GraphPad Prism program.Results were expressed as means of at least three independent experiments F SE.A two-tailed Student's t test was used to determine the statistical differences between various experimental and control groups, with P < 0.05 considered statistically significant.

Results
Specific knockdown of BIIand BIVb-tubulins in NSCLC cell lines.To gain insight into the functional significance of hIIand hIVb-tubulins in drug response, we used the siRNA method to knock down their expression in two NSCLC cell lines, Calu-6 and H460.As shown in Fig. 1A, when transfected with specific siRNA, hII-tubulin (Hb9) mRNA expression was significantly reduced in both cell lines at the gene level, with complete knockdown of hIVb-tubulin (Hb2) mRNA expression observed in hIVb-tubulin siRNA-transfected cells.At 72 hours, Western blot analysis indicated that levels of hII-tubulin and hIV-tubulin proteins were markedly reduced in cells transfected with siRNA against hIIor hIVb-tubulin, respectively, but not in control siRNA-transfected cells (Fig. 1B).The specificity of the siRNAs was shown by probing with specific antibodies against other h-tubulin isotypes.As shown in Fig. 1C and D, knockdown of either hII-tubulin (Fig. 1C) or hIVb-tubulin (Fig. 1D) did not affect the endogenous levels of other h-tubulin isotypes.
Knockdown of either BIIor BIVb-tubulin does not affect microtubule organization but does enhance vincristineinduced microtubule depolymerization.To assess the qualitative effects of hIIand hIVb-tubulin knockdown on microtubule organization, hIIand hIVb-tubulin siRNA-transfected cells were stained with an antibody to a-tubulin.As shown in Supplementary Fig. S1, neither untreated hII-tubulin nor hIVb-tubulin knockdowns showed obvious changes to microtubule morphology.However, after incubation with 10 nmol/L vincristine for 1 hour, both hIIand hIVb-tubulin knockdowns showed marked depolymerization of the microtubule cytoskeleton compared with the control siRNA-treated cells.Microtubules were largely depolymerized and most of the cells were rounded with small remnants of microtubules occasionally observed.In contrast, no gross depolymerization of the microtubules was observed in the control siRNAtransfected cells at a comparable vincristine concentration.Thus, although the morphology of microtubules was unchanged in both hIIand hIVb-tubulin knockdowns in the absence of vincristine, low concentrations of vincristine caused extensive microtubule depolymerization in these cells compared with the controls.
Knockdown of BIIand BIVb-tubulin significantly enhances sensitivity of NSCLC cells to Vinca alkaloids.To investigate the effects of hIIand hIVb-tubulin knockdowns on the response of NSCLC cells to TBAs, drug-treated clonogenic assays were performed.Knockdown of hII-tubulin in H460 cells resulted in significant increased sensitivity to Vinca alkaloids when compared with mock-transfected cells as determined by its lower ID 50 to vincristine (P < 0.001; Table 1; Supplementary Fig. S2A), vinblastine (P < 0.01; Table 1), vinflunine (P < 0.001; Table 1), and vinorelbine (P < 0.01; Table 1).In contrast, there was no statistically significant difference in drug sensitivity between control siRNA-transfected cells and mock-transfected cells cultured in the presence of these TBAs.Interestingly, knockdown of hII-tubulin did not significantly increase paclitaxel sensitivity in H460 cells (Table 1; Supplementary Fig. S2B).Similar results were obtained with cRelative sensitivity is the fold sensitivity of the siRNA-treated cells compared with the mock.Relative sensitivity was determined by dividing the ID 50 of the mock by the ID 50 of the hIIand hIVb-tubulin knockdowns or control siRNA-treated cells.bMock-transfected cells were used as the reference.
x P < 0.001.k P < 0.01.hII-tubulin knockdown Calu-6 cells with increased sensitivity to vincristine (P < 0.001; Supplementary Fig. S2A; Supplementary Table S1) but no change in paclitaxel sensitivity compared with mock-transfected Calu-6 cells (Supplementary Fig. S2B; Supplementary Table S1).The results obtained for other Vinca alkaloids examined (including vinblastine, vinflunine, and vinorelbine) were consistent in both NSCLC cell lines, and the results for hIItubulin knockdown in Calu-6 cells are presented in Supplementary Table S1.Clonogenic survival showed that knockdown of hIVb-tubulin significantly increased the sensitivity of H460 cells to vincristine (2.7-fold at ID 50 ; P < 0.001) compared with mock-treated cells (Fig. 2A; Table 1).Interestingly, the knockdown of hIVb-tubulin expression in Calu-6 cells led to a marked drop in surviving colonies when treated with vincristine (Fig. 2A).This represents a 9.7-fold (P < 0.001) increase in sensitivity in the ID 50 concentration compared with mock-transfected cells (Supplementary Table S1).No significant change in sensitivity to vincristine was observed in the control siRNA-treated cells compared with mock-treated cells (Fig. 2A; Supplementary Table S1).Similar to hII-tubulin, knockdown of hIVb-tubulin expression did not significantly change paclitaxel sensitivity in these cells (Fig. 2B; Table 1 for H460 cells and Supplementary Table S1 for Calu-6 cells).Additional Vinca alkaloids were also tested to ascertain whether the hypersensitivity effects observed with vincristine were specific to this agent or to other members of this drug class.Consistent with hII-tubulin, knockdown of hIVb-tubulin expression resulted in enhanced sensitivity to Vinca alkaloids in H460 cells (Table 1) and Calu-6 cells (Supplementary Table S1).In summary, specific knockdown of either hIIor hIVb-tubulin expression hypersensitized both H460 and Calu-6 cells to all Vinca alkaloids tested, with greatest sensitization observed in the hIVb-tubulin knockdown.
Vinca alkaloid hypersensitivity is not due to increased drug accumulation.To determine whether changes in drug accumula-tion were contributing to drug hypersensitivity, the accumulation of [ 3 H]vincristine was measured in hIVb-tubulin knockdown cells.There was no significant difference in intracellular [ 3 H]vincristine levels between hIVb-tubulin knockdown and controls in both Calu-6 and H460 cells (Supplementary Fig. S3).These results suggest that enhanced sensitivity to vincristine in hIVb-tubulin knockdown cells is not attributable to changes in drug accumulation.
Knockdown of BIVb-tubulin affects vincristine-induced G 2 -M cell cycle arrest and enhances cell death.To determine whether silencing of specific h-tubulin isotypes affects cell cycle progression, cell cycle analysis was performed on hIIor hIVbtubulin knockdown H460 cells (Figs. 3 and 4, respectively).Knockdown of hII-tubulin alone caused a small but significant decrease in G 1 ( from 62.56 F 2.07% in control to 55.64 F 1.13% in hII-tubulin knockdown; P < 0.05; Fig. 3) and a concomitant increase in S phase ( from 16.88 F 0.10% in control to 20.41 F 0.49% in hII-tubulin knockdown; P < 0.01; Fig. 3).In contrast, the only variable altered in the hIVb-tubulin knockdown cells was a small but significant increase in sub-G 1 levels ( from 5.00 F 0.71% in control to 10.60 F 1.10% in hIVb-tubulin knockdown; P < 0.05; Fig. 4).Following 24-hour incubation with 5 nmol/L vincristine, the hII-tubulin knockdown cells had a significantly higher G 2 -M population compared with the control siRNA-treated cells (P < 0.05; Fig. 3).Both hII-tubulin siRNA-treated and control siRNA-treated cells had similar levels of sub-G 1 content compared with untreated samples.At the higher concentration of vincristine (40 nmol/L), both hII-tubulin siRNA-treated and control siRNAtreated cells showed a marked G 2 -M block, a hallmark effect of TBAs.There was a small but significant increase in the G 1 content of the hII-tubulin knockdown cells.The sub-G 1 content was, however, not significantly different compared with control-treated cells at this concentration (Fig. 3).
The hIVb-tubulin knockdown, on the other hand, had a higher sub-G 1 content when compared with the control siRNA-treated cells at 5 nmol/L vincristine (P < 0.001; Fig. 4).The G 1 population was also significantly decreased (P < 0.001; Fig. 4) with a concomitant increase in the G 2 -M population (P < 0.01; Fig. 4) when compared with control siRNA-treated cells.A greater difference was observed with 40 nmol/L vincristine, with the control siRNA-treated cells showing a marked G 2 -M block (P < 0.01; Fig. 4), whereas the hIVb-tubulin knockdown displays an increase in the sub-G 1 population reflective of cell death (P < 0.05; Fig. 4).However, the extent of changes was relatively small between 5 and 40 nmol/L vincristine, suggesting that at a concentration as low as 5 nmol/L, knockdown of hIVb-tubulin significantly induced a marked G 2 -M block and an increase in the sub-G 1 populations reflective of cell death, thereby confirming the increase in sensitivity of these cells to vincristine.
Knockdown of BIVb-tubulin increases sensitivity of cells to apoptosis in the presence of vincristine but not paclitaxel.The cell cycle data suggested that knockdown of hII-tubulin expression caused a vincristine-induced G 2 -M block and that hIVb-tubulin knockdown sensitized cells to vincristine-induced cell death.To determine if knockdown of hIIor hIVb-tubulin enhanced apoptotic cell death, Annexin V-FITC staining was performed in the siRNA-transfected cells following 48 hours of drug treatment.Consistent with the cell cycle analysis, knockdown of hII-tubulin did not significantly increase the percentage of apoptotic cells in a dose-dependent manner in the presence of vincristine or paclitaxel (Supplementary Fig. S4A and B ). Instead, treatment with vincristine inhibited cell proliferation of the hII-tubulin knockdown cells in a concentration-dependent manner (Supplementary Fig. S4C).In contrast, paclitaxel treatment did not result in growth inhibition of hII-tubulin knockdown cells (Supplementary Fig. S4D).However, results presented in Fig. 5A clearly reveal that hIVb-tubulin knockdown cells undergo vincristine-induced apoptosis at significantly higher levels than control siRNA-treated cells in a dose-dependent manner, suggesting that increased apoptosis induction might serve as one of the mechanisms involved in hypersensitivity to Vinca alkaloid treatment.To address this possibility, hIVb-tubulin knockdown cells were treated with paclitaxel followed by Annexin V-FITC staining.Consistent with the drug-treated clonogenic data showing that levels of hIVb-tubulin did not affect sensitivity to paclitaxel, the percentage of apoptotic cells was comparable between the control and hIVb-tubulin knockdown cells (Fig. 5B), suggesting that hIVb-tubulin silencing did not enhance paclitaxel-induced apoptosis.

Discussion
Differential expression of h-tubulin isotypes in tumors may provide unique characteristics that impart functional differences to microtubules and mediate response to TBAs.Herein, we show for the first time that hIVb-tubulin and, to a lesser extent,hII-tubulin mediate sensitivity to Vinca alkaloids but not the taxanes in NSCLC cells.The finding that knockdown of hIVb-tubulin expression has a marked effect on increasing Vinca sensitivity by enhancing apoptotic cell death suggests that expression of this isotype may serve as a strong predictor of vincristine response.
Increased expression of hIIand/or hIV-tubulin has been described in lung (14), ovarian (14), breast (23,27), and prostate cancer (26) cell lines selected for resistance to a variety of TBAs, including paclitaxel, docetaxel, and estramustine.However, detailed analysis of the role of these two isotypes in response to TBAs, and in particular Vinca alkaloids, in cancer cells has been scarce.Many of the above studies are correlative and there is little direct evidence that hIIand hIVb-tubulin can affect cellular sensitivity to TBAs.Despite a clinical correlation between increased hII-tubulin expression and lower response rates for docetaxel in breast cancer (19), and a higher ratio of hII/hV-tubulin in lung cancer (22), we found that neither hII-tubulin nor hIVb-tubulin influenced paclitaxel response in our NSCLC cell lines.In agreement, two previous studies using stable overexpression of hIIor hIVb-tubulin isotypes in CHO cells and the hII-tubulin isotype in NIH 3T3 cells failed to confer resistance to paclitaxel (17,18).Taken together, hIIand hIVb-tubulin do not seem to play a direct role in sensitivity and resistance to paclitaxel.
Microtubule morphology following knockdown of hIIand hIVbtubulin isotypes was not altered and is similar to the finding that overexpression of either of these isotypes in CHO cells does not affect microtubule morphology (17).However, on exposure to vincristine, microtubule morphology was consistent with increased sensitivity to microtubule disruption in both hIIand hIVb-tubulin knockdown cells compared with control siRNA-treated cells.These findings were subsequently validated with quantitative analysis, where knockdown of hIIand hIVb-tubulin significantly increased sensitivity to Vinca alkaloids (i.e., vincristine, vinblastine, vinorelbine, and vinflunine), with hIVb-tubulin knockdown exhibiting more potent effects in the drug-treated clonogenic assays than hIItubulin.The NSCLC cells used in the current study have not been selected for drug resistance and changes in sensitivity mediated by specific tubulin isotypes represent changes in intrinsic sensitivity to Vinca alkaloids and potentially predict resistance to these agents.A potential explanation for increased sensitivity to Vinca alkaloids is that knockdown of either hIIor hIVb-tubulin may lead to alterations in drug-binding affinity.However, because Vinca alkaloids (including vincristine, vinblastine, and vinorelbine) all bind to h-tubulin isotypes with similar affinity (30,31), it is unlikely that alterations in drug-binding affinity due to knockdown of either hIIor hIVb-tubulin is the underlying mechanism.
Tubulin can undergo complex autoregulation (32).To ensure the specificity of siRNA targeting hIIand hIVb-tubulin, we confirmed by Western blotting that knockdown was specific to each isotype and the cells did not compensate by altering the expression of other h-tubulin isotypes expressed in the NSCLC cells used in this study.These findings highlight that the changes in sensitivity to Vinca alkaloids following knockdown of either hIIor hIVb-tubulin are specific to these isotypes.Neuronal hIVa-tubulin and

Cancer Research
Cancer Res 2008; 68: (23).December 1, 2008 ubiquitously expressed hIVb-tubulin are encoded by the H5b and Hb2 genes, respectively, and share a high degree of amino acid homology (33).Although it is possible to specifically distinguish hIVa (H5b) and hIVb (Hb2) at the gene level, there are no commercially available antibodies to differentiate hIVaand hIVbtubulin protein expression.We found approximately 40% to 50% H5b gene reduction in Calu-6 cells transfected with hIVb-tubulin siRNA (data not shown) but no significant changes in the H460 hIVb-tubulin knockdown cells.Because both cell lines showed similar responses in drug-treated clonogenic assays, the results obtained can be attributed to hIVb-tubulin suppression rather than hIVa-tubulin.
Although hIVb-tubulin shares a high degree of amino acid homology with the neuronal-specific hIVa-tubulin isotype, these two isotypes exhibit distinct effects in drug response.In a recent study, hIVa-tubulin overexpressing CHO cells displayed increased sensitivity to colcemid, vinblastine, and nocodazole (34).Their response to paclitaxel, however, changed with levels of expression of the protein.Low hIVa-tubulin expression caused increased sensitivity to paclitaxel (34).Whereas the level of expression increased, the sensitivity to paclitaxel was lost (i.e., cells exhibited normal sensitivity to paclitaxel).Overexpression of hIVb-tubulin in the same study, however, did not affect sensitivity to colcemid, vinblastine, and paclitaxel.In contrast, our results clearly show that knockdown of hIVb-tubulin mediates sensitivity to all four Vinca alkaloids tested, including vinblastine.The differences in response to vinblastine between the two studies may be due to several factors.The hIVb-tubulin overexpression study used CHO cells and our study used NSCLC cells.The endogenous repertoire of tubulin and microtubule proteins within each of these cell types is likely to differ and may have influenced the response to vinblastine.Modulating expression of h-tubulin may lead to compensatory changes in other tubulin isotypes and in turn influence TBA response.Although we addressed this possibility in the current study, it is unclear whether this issue was examined in the hIVbtubulin overexpression study.
To address the mechanisms underlying the chemosensitivity to Vinca alkaloids following hIIand hIVb-tubulin knockdown in NSCLC cells, we examined the propensity of the cells to undergo drug-induced cell cycle arrest and apoptosis.Cell cycle analysis of h-tubulin isotype expression has previously been examined in leukemic and sarcoma cell lines (35).Both hIVaand hIVb-tubulin gene levels were found to be expressed at twice the levels in dividing cells than in resting cells.In contrast, no change in hIItubulin levels was observed.Our data showing that knockdown of hIVb-tubulin does not markedly affect the cell cycle profile compared with control cells suggest that expression of this isotype may not be playing a critical role in cell division or other isotypes may be compensating for the suppression of hIVb-tubulin.The effects of hIIor hIVb-tubulin knockdown were most noticeable following 24-hour vincristine treatment.At concentrations as low as 5 nmol/L vincristine, hII-tubulin knockdown cells display enhanced sensitivity to G 2 -M cell cycle arrest compared with controls.At the concentrations and time tested, the cells were arrested but not yet undergoing cell death.In contrast, hIVbtubulin knockdown cells were more sensitive to vincristine-induced cell death as shown by the increased sub-G 1 populations following 5 and 40 nmol/L of vincristine treatment.This finding is reflective of their heightened sensitivity to vincristine in clonogenic assays.
Interestingly, unlike hIVb-tubulin, hII-tubulin knockdown did not enhance apoptosis induction on exposure to vincristine at equivalent concentrations, suggesting that these two isotypes differ in their behavior when exposed to the drug.We found that incubation with vincristine inhibited cell proliferation of the hIItubulin knockdown cells in a concentration-dependent manner compared with increased apoptosis induction seen in the hIVbtubulin knockdown cells, thus potentially explaining the heightened sensitivity to Vinca alkaloids in these cells compared with the hII-tubulin knockdown cells.Similar effects were not observed with paclitaxel treatment (neither growth inhibition nor apoptosis induction) following knockdown of either hIIor hIVb-tubulin.
One notable aspect of our findings is that hIIand hIVb-tubulin, which display functionally distinct roles in drug response, are present in relatively small amounts.Using quantitative RT-PCR, Nicoletti and colleagues (36) showed that hI-tubulin isotype is the most abundant (88.8%), followed by hIII (7.8%) and hIVa (2%), with hII and hIVb representing only 1% and 0.2%, respectively, of all h-tubulin isotypes present in H460 cells.Manipulation of the expression of these minor isotypes produces significant changes in Vinca alkaloid response, suggesting that they are functionally distinct.
Microtubule function relies on the interaction of microtubules with other cellular proteins, and thus, specific h-tubulin isotypes may play an equally important role in modulating these interactions.The dramatic increase in vincristine sensitivity observed in hIVb-tubulin knockdown cells is of interest because this isotype is constitutively expressed in low amounts in all tissues b-Tubulin Isotypes Mediate Vinca Alkaloid Sensitivity www.aacrjournals.org(4).It is conceivable that differences in h-tubulin composition may influence the interactions of microtubules with other elements of the cytoskeletal network and hence dictate tumor cell behavior, including the cellular response to TBAs.There is high structural homology between the various h-tubulin isotypes and within the Vinca-binding domain of tubulin (amino acids 175-215; ref. 37), with hIIand hIVb-tubulin sharing 100% identity within this region (33).Therefore, differences in cell cycle and Vinca response cannot simply be explained by h-tubulin isotype structural differences.The greatest diversity occurs in the COOH-terminal region of h-tubulin, a region known to bind microtubule-associated proteins (MAP) and interact with other cellular factors (reviewed in ref. 38).It is possible that by decreasing specific h-tubulin isotypes, the ratio of specific MAP binding is altered, which in turn affects the dynamic nature of the microtubules and hence accessibility and/or efficacy of Vinca alkaloids in cells expressing hIIor hIVb-tubulin.
Collectively, our recent demonstration that hIII-tubulin mediates sensitivity to Vinca alkaloids and taxanes (9), and the current study showing that knockdown of hIIand hIVb-tubulin increases sensitivity to Vinca alkaloids but not the taxanes, proposes that each isotype is unique in terms of drug-target interactions and isotype composition of a cell affects its response to TBAs.The clinical relevance of various h-tubulin isotypes is still being unraveled.The functional significance of hIIand hIVb-tubulin in Vinca sensitivity in NSCLC cell lines shown in this study awaits further validation from large prospective clinical studies in human tumors.This information may be valuable in exploiting intrinsic differences in h-tubulin isotype composition in specific tumor types to improve therapeutic responses.

Disclosure of Potential Conflicts of Interest
M. Kavallaris and P.P. Gan: PCT filed March 2008 on data included in the article.M. Kavallaris was paid as a speaker/consultant at Pierre Fabre Vinflunine Basic Science Meeting, Paris, France, in 2007 (honorarium paid).Please note that none of the data in the submitted manuscript was discussed or presented at this meeting.

Figure 1 .
Figure1.siRNA directed against hIIand hIVb-tubulin specifically inhibits its expression.A, gene expression of hIIand hIVb-tubulin 48 h after transfection of siRNA directed against hII-tubulin (25 nmol/L) and hIVb-tubulin (100 nmol/L) in NSCLC cell lines Calu-6 and H460.Control siRNA was used at the same concentration as the target siRNA.h 2 -Microglobulin (b 2 M) was used as the internal control.B, protein expression of hIIand hIVb-tubulin by Western blotting following 72 h of siRNA transfection.Glyceraldehyde-3-phosphate dehydrogenase (GAPDH) expression was detected simultaneously as the loading control.M, mock; C, control siRNA; bII, hII-tubulin siRNA; bIVb, hIVb-tubulin siRNA.Specificity of hII-tubulin (C) and hIVb-tubulin siRNA (D ) in H460 cells as detected by Western blotting.No significant changes were observed for other h-tubulin isotypes at the protein level in hIIor hIVb-tubulin H460-silenced cells.

Figure 3 .
Figure 3. Cell cycle analysis of hII-tubulin knockdown H460 cells treated with vincristine.Cells were harvested after 24 h of drug treatment and subsequently assayed for their DNA content by flow cytometry.Representative figures are shown.Data represent mean F SE of at least three independent experiments.Bolded text is indicative of a significant difference between the control and hII-tubulin knockdown at the equivalent drug concentrations.*, P < 0.05; **, P < 0.01.

Figure 4 .
Figure 4. Cell cycle analysis of hIVb-tubulin knockdown H460 cells treated with vincristine.Cells were harvested after 24 h of drug treatment and subsequently assayed for their DNA content by flow cytometry.Representative figures are shown.Data represent mean F SE of at least three independent experiments.Bolded text is indicative of a significant difference between the control and hIVb-tubulin knockdown at the equivalent drug concentrations.*, P < 0.05; **, P < 0.01; ***, P < 0.001.