Overexpression of Kinesins Mediates Docetaxel Resistance in Breast Cancer Cells

Published OnlineFirst September 29, 2009; DOI: 10.1158/0008-5472.CAN-09-1224

Experimental Therapeutics, Molecular Targets, and Chemical Biology

Sarmishtha De,1 Rocky Cipriano,2 Mark W. Jackson,2 and George R. Stark1,3

Departments of 1Genetics and 2Pathology, Case Western Reserve University, Case Comprehensive Cancer Center, and 3Department of Molecular Genetics, Lerner Research Institute, Cleveland Clinic, Cleveland, Ohio

Abstract Our laboratory has developed a novel method, called VBIM, for Resistance to chemotherapy remains a major barrier to the validation-based insertional mutagenesis, for inserting strong successful treatment of cancer. To understand mechanisms cytomegalovirus (CMV) promoters approximately randomly into underlying docetaxel resistance in breast cancer, we used an the genomes of mammalian cells in tissue culture. In principle, the insertional mutagenesis strategy to identify proteins whose method is capable of identifying any protein or RNA whose overexpression confers resistance. A strong promoter was overexpression in a mutant cell line confers a selectable phenotype inserted approximately randomly into the genomes of tumor- (10). The inserted promoter can be removed or silenced at will, to derived breast cancer cells, using a novel lentiviral vector. We prove that its function is responsible for the altered phenotype. We isolated a docetaxel-resistant clone in which the level of the used the VBIM method to make the novel discovery that overexpression of the kinesin KIFC3 confers resistance to docetaxel kinesin KIFC3 was elevated. When KIFC3 or the additional kinesins KIFC1, KIF1A, or KIF5A were overexpressed in the in breast cancer cell lines. breast cancer cell lines MDA-MB231 and MDA-MB 468, the cells The kinesin motor proteins, found in all cells, attach to became more resistant to docetaxel. The binding of kinesins to microtubules and move along them to transport cellular cargoes. microtubules opposes the stabilizing effect of docetaxel that Kinesins are also key participants in chromosomal and spindle prevents cytokinesis and leads to apoptosis. Our finding that movements during mitosis and meiosis (11, 12). They typically kinesins can mediate docetaxel resistance might lead to novel include two large globular heads that attach to microtubules, a therapeutic approaches in which kinesin inhibitors are paired central coiled region, and a light-chain region that connects the kinesin to the intracellular component to be moved (13). The 45 with taxanes. [Cancer Res 2009;69(20):8035–42] murine and human kinesin genes have been classified into three major types, based on the positions of their motor domains: amino- Introduction terminal (N), middle (M), and carboxy-terminal (C; ref. 14). There Systemic chemotherapy of breast cancer has improved over the are 39 N-kinesins, three M-kinesins, and three C-kinesins, KIFC1, past few decades with the introduction of new drugs, and the KIFC2, and KIFC3 (14). KIFC1 (15) is expressed primarily in optimization of numerous combination regimens (1). An important proliferating tissues and cell lines, and KIFC2 (15, 16) is expressed advance has been the use of docetaxel and paclitaxel, which bind to specifically in neural tissues. In contrast, KIFC3 is apparently the h-tubulin subunits of microtubules, inhibiting the rate of ubiquitous (14, 17). KIFC3 has minus end–directed microtubule exchange of free and bound tubulin (2). This inhibition disrupts motor activity and functions in Golgi positioning and integration mitotic spindle formation, inhibiting cell division and leading to (18) and also in apical transport in epithelial cells (19). Studies of cell death (3). Although up to half of the patients treated with null mice revealed that KIFC3 is dispensable for normal docetaxel achieve a clinical response, their tumors may develop development and reproduction (17). resistance subsequently (3). Several different mechanisms account In addition to KIFC3, we find that overexpression of a second for the taxane resistance observed in human tumors and tumor COOH-terminal kinesin, KIFC1, or of the NH2-terminal kinesins cell lines, including overexpression of the multidrug transporter KIF5A and KIF1A, all mediate the resistance of breast cancer cells P-glycoprotein (4), altered metabolism of the taxane, decreased to docetaxel. These findings suggest that different types of kinesins, sensitivity to death-inducing stimuli (5), alterations in microtubule probably including some of the 41 family members not tested in dynamics, and altered binding of the taxane to its microtubule our study, might also function as mediators of docetaxel resistance target (6, 7). Candidate proteins that might mediate taxane in cancer. resistance through their inclusion in microtubules or their association with these structures include hIII-tubulin (2), the microtubule-associated proteins MAP4 and Tau, and the microtu- Materials and Methods bule-destabilizing phosphoprotein Stathmin (8, 9). Analysis of the Cells and reagents. MDA MB 231 cells and MDA MB 468 cells, kindly levels of these proteins might be clinically useful in predicting provided by Dr. John Pink (Case Western Reserve University, Cleveland, OH), responses to taxanes, but none have proven to be specific targets were grown in a humidified atmosphere containing 5% CO2 in RPMI 1640, for increasing the efficacy of taxane-based therapy in breast cancer. supplemented with 5% fetal bovine serum (HyClone Laboratories, Inc.). Docetaxel and Vincristine were from Sigma. Anti-KIFC3 was from Proteintech Group, Inc. Anti-KIFC1 and anti-KIF5A were from Abcam, a Note: National Center for Bioinformatics Information’s auxiliary Web site—Gene Inc. Anti–pan-actin and anti– -tubulin were from Neomarkers and Santa Expression Omnibus (http://www.ncbi.nlm.nih.gov/geo/), Affymetrix Netaffx Analysis Cruz Biotechnology, respectively. Full-length cDNAs for KIFC1, KIF1A, and Center (http://affymetrix.com/analysis/index.affx). KIF5A were from Open Biosystems. Requests for reprints: George R. Stark, Research Institute, 9500 Euclid Avenue, Plasmids. The three VBIM vectors (pVBIM SD1, SD2, and SD3), Cleveland, OH 44195. Phone: 216-444-3900; Fax: 216-444-3279; E-mail: [email protected]. I2009 American Association for Cancer Research. constructed on a lentiviral backbone, represent three different reading doi:10.1158/0008-5472.CAN-09-1224 frames (10). Cre recombinase (20) was expressed from a construct in the www.aacrjournals.org 8035 Cancer Res 2009; 69: (20). October 15, 2009

Cancer Research pBabePuro vector. To construct KIFC3 cDNA, total RNA was isolated from puro or pBabe Cre recombinase. The cells were selected with puromycin for mutant cells by using a Qiagen kit according to the protocol provided by the 2 wk to generate stable pools. The Cre and vector control cells were plated manufacturer. Reverse transcription was performed using oligo DT20 primers at 300,000 cells per well in six-well plates and treated with docetaxel (10 or with the SuperScript III First-Strand Synthesis System kit and protocol 12 nmol/L) for 24 h. After 1 week, the cells were stained with methylene blue. (Invitrogen). Standard PCR reactions were performed by using a forward Mapping the inserts and detection of overexpressed molecules. To primer from the VBIM vector and a KIFC3-specific reverse primer with a stop identify insertion sites, inverse PCR was performed with EcoR1- and Mfe1- codon. The PCR product was cloned into the pCR2.1 vector (Invitrogen). The digested genomic DNA from mutant clone SD1.10, which was circularized DNA was digested with EcoR1 and recloned into the retroviral vector pLPCX. by self-ligation and subjected to nested PCR (10). The PCR product was KIFC1, KIF5A, and KIF1A were also recloned into the pLPCX vector. cloned into pCR2.1, sequenced, and aligned to the human genome database. Lentiviral and retroviral transduction. Lentiviruses were produced by The insertion in the KIFC3 gene was confirmed by reverse transcription- transient transfection of 293T cells with the plasmids pVBIM-SD1, pVBIM- PCR (RT-PCR), using the vector-derived forward primer and a KIFC3- SD2, and pVBIM-SD3 along with a second-generation packaging construct derived reverse primer. Western analysis was performed using anti-KIFC3. (pCMV-dR8.74) and pMD2G both provided by D. Trono, University of Cell survival assay. Cells (300,000) were plated in each well of six-well Geneva, Geneva, Switzerland (21). plates, allowed to attach overnight, and treated with docetaxel or vincristine Generation of a VBIM-mutated cell library and selection of for 24 h. After 7 to 10 d, the cells were lysed with 0.5 mol/LNaOH and docetaxel resistant clones. MDA-MB 231 cells were plated at 10,000 cells diluted 50-fold before the A260 was measured, as an indication of the total per well in 48-well plates and infected with pVBIM SD1, SD2, and SD3 amount of nucleic acid. The fraction of surviving cells was calculated (20 pools for each), with 12 pools left uninfected. The cells were grown for relative to untreated controls. 2 d and plated in six-well plates. When the cells attained 80% confluency, Immunoblotting and mRNA expression. Western analyses were 300,000 cells were replated in each well of six-well plates and allowed to performed as described before (21). The membranes were probed with attach overnight. The cells were treated with 10 nmol/Ldocetaxel and kept antibodies against KIFC3, KIFC1, and KIF5A or anti–pan-actin, followed by under selection for 24 h. incubation with anti-mouse or anti-rabbit secondary antibody-horseradish Validation of docetaxel-resistant clones. Colonies surviving in peroxidase conjugates (Hoffman-La Roche). Chemiluminescence was devel- docetaxel were expanded and infected with the retroviral vectors pBabe oped by using an enhanced chemiluminescence kit (Perkin-Elmer). To

Figure 1. Identification and characterization of docetaxel-resistant SD1.10 cells. A, MDA-MB 231 breast cancer cells were infected with the three VBIM insertional mutagenesis viruses and selected in docetaxel (Doce). The docetaxel-resistant mutant clone SD1.10 was infected with control vector (pBabePuro) or with a vector encoding Cre recombinase. The cells remaining after docetaxel treatment were stained with methylene blue. B, to identify the responsible gene, a PCR fragment from mutant SD1.10 was cloned, sequenced, and aligned to the human genome database. The insertion maps to the KIFC3 gene, and the VBIM integration was in the third intron of this gene, in the sense direction. C, overexpression of KIFC3 mRNA in SD 1.10 mutant cells. RT-PCR was performed with RNAs from parental cells, the SD1.10 mutant infected with control vector, and SD1.10 cells expressing Cre. Overexpression of KIFC3 mRNA was observed in the mutant at an early (22nd) PCR cycle, and this expression was completely reversed by Cre. Glyceraldehyde-3-phosphate dehydrogenase (GAPDH) was used as a control. D, overexpression of the KIFC3 protein in SD1.10 cells. Western analysis was performed with anti-KIFC3 antibody. Anti–pan-actin was used as a loading control. Two independent experiments gave very similar results.

Cancer Res 2009; 69: (20). October 15, 2009 8036 www.aacrjournals.org

Docetaxel Resistance Due to Kinesin Overexpression

Figure 2. KIFC3 is responsible for docetaxel resistance. A, KIFC3 cDNA was used to infect naı¨ve MDA-MB 231 cells and the level of expression was determined by Western analysis, using total cell lysates. Anti–pan-actin was used as a control. The experiment was repeated twice with very similar results. B, survival of cells expressing KIFC3 or empty vector, with or without docetaxel treatment. The cells were lysed with 0.5 mol/L NaOH and A260 was measured as an indication of the total amount of nucleic acid. The fraction of surviving cells was determined by normalizing the data from docetaxel-treated cells to untreated controls. Data shown represent means F SD of three experiments, in which each measurement was performed in triplicate. C, overexpression of KIFC3 in individual clones. D, cell survival assay with three individual clones, performed as in B. Columns, means of three experiments, and each measurement was performed in triplicate; bars, SD. *, P < 0.05; **, P < 0.005. determine the mRNA expression level, total cellular RNA was extracted using sensitive groups. Gene probe set identifiers and annotations for members of a Qiagen kit. The RT reaction was performed using the SuperScript III First- the family of kinesin genes were found in the Affymetrix Netaffx Analysis Strand Synthesis System, and PCR was done using KIF1A-specific primers. Center5 and exported as tab delimited text. These identifiers were Measurement of free or polymerized tubulin. Cells were plated in associated with the respective signal profile records found in the Affymetrix 60-mm plates (40 Â 104 cells/plate) and allowed to attach overnight. The results, using MS Access database manager. Gross profiles were then cells were treated with docetaxel for 24 h or left untreated. After 2 d, free or inspected to see if there were any consistent differences in expression of polymerized tubulin was extracted, using a standard protocol (22). Briefly, members of the kinesin family of genes between docetaxel-resistant and the cells were washed with PBS at 37jC, and then with 1 mLof Buffer A -sensitive samples. F P [containing 0.1 mol/LMES (pH 6.75), 1 mmol/LMgSO 4, 2 mmol/LEGTA, Statistical analysis. Values were expressed as means SD. values 0.1 mmol/LEDTA, and 4 mol/Lglycerol], which prevents microtubule were based on the paired t test and the significance was set at 0.05 (marked depolymerization. The soluble proteins were then extracted at 37jCfor with an asterisk wherever data are statistically significant). 5 min in a buffer (Buffer A plus 0.1% v/v Triton X-100 and protease inhibitors) and the extracts were centrifuged for 5 min at 13,000 rpm. The pellet fraction and lysed cells in the culture dish were dissolved in a buffer Results containing 25 mmol/LTris (pH 6.8), plus 0.5% SDS. Total proteins from the Isolation of a docetaxel-resistant clone. Using three varieties free or polymerized tubulin fractions were analyzed by the Western method, of VBIM lentiviral vectors, intended to allow expression in all three a with anti– -tubulin. The intensity of each band was measured by using NIH reading frames (10), we inserted the strong CMV promoter software (image J) and the percentages of free and polymerized tubulin in approximately randomly into the genomes of docetaxel-sensitive, each cell extract were quantified. Apoptosis assay. Apoptosis was measured by flow cytometry, using the estrogen receptor–negative MDA-MB 231 breast cancer cells. A Annexin V FITC Apoptosis Detection kit (Calbiochem) according to the total of 60 pools (10,000 virally infected cells per pool; 20 each with manufacturer’s protocol. pVBIM SD1, SD2, and SD3) were selected with 10 nmol/Ldocetaxel. Analysis of gene expression in breast cancers. The ‘‘Cel’’ files for After selection, 8 of the 60 pools yielded resistant colonies; there Affymetrix Hg U95-A v2 arrays used by Chang and colleagues (ref. 23; series were no colonies in 12 pools of uninfected control cells. Cells from GSE349 and GSE350) were downloaded from the National Center for each positive pool were infected with a lentivirus encoding Cre or Bioinformatics Information’s auxiliary Web site Gene Expression Omnibus with the control pBabePuro virus, treated with docetaxel and 4 (GEO) , and the signals were analyzed by the Gene Expression and grown to test for reversion. Clone SD1.10 was validated in this way Genotyping Facility of the Case Comprehensive Cancer Center, Case (Fig. 1A) but the other seven clones were not validated, showing Western Reserve University, Cleveland Ohio, using Affymetrix proprietary that their resistance was not caused by promoter-driven events. software, Gene Chip Operating Software. All samples were scaled to 500 in KIFC3 accordance with the recommendations of the vendor (Affymetrix). Signal Identification of as the gene responsible for docetaxel data were exported as excel spreadsheets and sorted into resistant and resistance. A PCR fragment from mutant SD1.10 was cloned and

4 http://www.ncbi.nlm.nih.gov/geo/ 5 http://affymetrix.com/analysis/index.affx www.aacrjournals.org 8037 Cancer Res 2009; 69: (20). October 15, 2009

Cancer Research

results indicate that the overexpression of full-length KIFC3 protein is responsible for the mutant phenotype of SD1.10 cells. High expression of KIFC3 increases resistance to docetaxel in naı¨ve cells. To confirm KIFC3 as a mediator of docetaxel resistance, a full-length KIFC3 cDNA was introduced into unmutagenized, MDA-MB 231 cells, resulting in stable pools of cells in which the KIFC3 protein is overexpressed (Fig. 2A). We noted two bands in western analyses of KIFC3 protein (Fig. 2A and C), perhaps due to posttranslational modifications. These cells were significantly more resistant to 4 or 6 nmol/Ldocetaxel than were the controls (Fig. 2B). The mixed population of cells expressing KIFC3 contains individual cells with different levels of expression, causing some cells in the population to be more resistant than others. We isolated individual clones with different levels of KIFC3 (Fig. 2C) and analyzed their resistance to docetaxel (Fig. 2D). Clones K1 and K3, which have substantially higher levels of KIFC3 protein than clone K2, were correspondingly more resistant to docetaxel. These results confirm the role of KIFC3 in docetaxel resistance in MDA-MB 231 breast cancer cells in vitro and also show that the level of KIFC3 protein correlates with the level of resistance. Overexpression of KIFC1, KIF1A, or KIF5A increases resistance to docetaxel. In a previous study (23), core biopsy samples were taken from 24 primary breast tumors before treatment and the response of the tumors to neoadjuvant docetaxel was assessed by analysis of RNA extracted from biopsy samples. We used the Cel files for these Affymetrix Hg U95-Av2 arrays (series GSE349 and GSE350), from the National Center for Bioinformatics Information’s auxiliary Web site GEO4 to analyze the signals. In these data, kinesins KIFC1, KIF1A, and KIF5A were slightly overexpressed in docetaxel-resistant tumor samples, suggesting that these kinesins might have a role in mediating resistance in vivo. Therefore, we determined whether KIFC1, KIF5A, and KIF1A could mediate docetaxel resistance in our in vitro system, by generating stable MDA-MB 231 cell lines expressing them (Fig. 3). Upon exposure to 4 or 6 nmol/Ldocetaxel, there was a significant increase in the percentage of surviving KIFC1-expressing cells compared with control cells infected with vector alone (Fig. 3A). Increased cell survival was also observed in cells expressing KIF5A (Fig. 3B) and KIF1A (Fig. 3C). We used the same approach with another estrogen receptor–negative breast carcinoma cell line, MDA-MB 468, Figure 3. Increased expression of kinesins KIFC1, KIF5A, or KIF1A increases docetaxel resistance in MDA-MB 231 cells. A, B, and C, cell survival assays were creating cell pools overexpressing the four kinesins (Fig. 4) and performed as in Fig. 2B in cells infected with empty vector or with retroviruses analyzing the resistance to docetaxel. Cell survival was significantly encoding KIFC1, KIF5A, or KIF1A, after treatment with docetaxel. The fraction of higher in KIFC3 overexpressing cells compared with vector control surviving cells was determined by normalizing the data from docetaxel-treated A cells to untreated controls. Columns, means of three experiments, and each cells (Fig. 4 ), showing that the effect of KIFC3 in docetaxel measurement was performed in triplicate; bars, SD. Western analyses were resistance is not restricted to a single cell type. Similarly, we created performed with anti-KIFC1 and anti-KIF5A, or the mRNA level was determined MDA-MB 468 cells overexpressing the other three kinesins. Upon in KIF1A-expressing cells (insets). *, P < 0.05; **, P < 0.005. treatment with docetaxel, cells expressing KIFC1 (Fig. 4B), KIF5A (Fig. 4C), or KIF1A (Fig. 4D) survived much better than control cells, sequenced. The insertion maps to the KIFC3 gene, located on indicating that four different kinesins can mediate docetaxel human chromosome 16q13-q21. The full coding sequence of 2,061 resistance in two different breast cancer cell lines. bp of KIFC3 translates to 687 amino acids (24). The insertion in Kinesins KIFC3, KIFC1, and KIF5A oppose the microtubule- SD1.10 mapped to the third intron of the gene (Fig. 1B). Hoang and stabilizing effect of docetaxel. Taxanes stabilize microtubules, colleagues (24) showed that the translational start site is within exon thereby inhibiting cell processes that depend upon microtubule 5, and therefore, the viral integration in the third intron leads to the dynamics (25). To learn more about how KIFC3 confers resistance overexpression of the complete coding sequence of KIFC3. Using a to docetaxel, we determined whether it affects microtubules by 5¶ VBIM-specific primer and a 3¶ gene-specific primer, analysis by measuring the amounts of free tubulin in MDA-MB 231 cells RT-PCR showed the overexpression of the VBIM-specific KIFC3 overexpressing this kinesin, with or without docetaxel treatment. mRNA in mutant SD1.10 and Cre recombinase eliminated this In response to 4 nmol/Ldocetaxel, the amount of free tubulin was expression (Fig. 1C). The level of KIFC3 protein was significantly substantially decreased in a pool of cells infected with empty vector higher in mutant SD1.10 than in parental cells (Fig. 1D). These (Fig. 5A), and at 6 nmol/Ldocetaxel, no free tubulin was detected.

Cancer Res 2009; 69: (20). October 15, 2009 8038 www.aacrjournals.org

Docetaxel Resistance Due to Kinesin Overexpression

However, in a pool of cells expressing KIFC3, the amount of free substantially higher in MDA-MB 231 and MDA-MB 468 cells tubulin was decreased much less. Therefore, KIFC3 expression compared with MCF7 and T47D cells (Fig. 6B). Cell survival assays opposes the stabilization of microtubules that results from were performed to determine whether increased levels of docetaxel treatment. In addition to KIFC3, we also observed a endogenous microtubule depolymerization correlate with doce- substantial increase in the level of free tubulin in MDA-MB 231 taxel resistance. Consistently, survival was greater in MDA-MB cells overexpressing KIFC1 (Fig. 5B) or KIF5A (Fig. 5C), compared 231 and MDA MB 468 cells by 2.5- to 3-fold, compared with MCF7 with the controls, upon docetaxel treatment. However, treatment and T47D cells, following treatment with 3 or 4 nmol/Ldocetaxel with docetaxel had no effect on free tubulin pools in cells (Fig. 6C). Furthermore, when cells were exposed to the microtubule expressing KIF1A (Fig. 5D). These results suggest that kinesins depolymerizing Vinca alkaloid Vincristine (Fig. 6D), we observed KIFC3, KIFC1, and KIF5A oppose the microtubule-stabilizing effect increased resistance in cells with elevated polymerized tubulin of docetaxel. However, KIF1A increases resistance to docetaxel by a (MCF7 and T47D), compared with those with elevated free tubulin different mechanism, which may also be used by the other kinesins (MDA-MB 231 and MDA-MB 468). These results suggest that the (see the Discussion). relative levels of free and microtubule-bound tubulin may be useful KIFC3 inhibits docetaxel-mediated apoptosis. Because as markers of docetaxel resistance and vincristine sensitivity in taxane-mediated disruption of microtubules ultimately drives cells breast cancer. to apoptosis, we examined whether KIFC3 expression affects apoptosis in response to docetaxel in breast cancer cells. Pools of MDA-MB 231 cells overexpressing KIFC3, or control cells, were Discussion exposed to different concentrations of docetaxel for 24 hours, and In cancer cells, resistance to chemotherapy is often due to the after 8 days, an apoptosis assay was performed. After exposure to overexpression of a specific protein, often achieved through docetaxel, the percentage of apoptotic cells was significantly higher amplification of the corresponding gene. Earlier reports described in control cells than in cells expressing KIFC3 (Fig. 6A). several different function-based screens to discover cellular genes Microtubule depolymerization contributes to docetaxel whose inactivation or overexpression yielded phenotypes of resistance in breast cancer cells. Free tubulin levels were interest (26–28). We developed a novel validation-based insertional

Figure 4. Kinesin overexpression increases docetaxel resistance in MDA-MB 468 cells. The cells were infected with retroviral vectors encoding four different kinesins or with empty vector. Cell survival assays were performed as in Fig. 2B in cells expressing KIFC3 (A), KIFC1 (B), KIF5A (C), or KIF1A (D). Insets, the protein levels of KIFC3, KIFC1, and KIF5A and the mRNA level of KIF1A. Columns, means of two experiments, and each measurement was performed in triplicate; bars, SD. *, P < 0.05; **, P < 0.005; ***, P < 0.0005. www.aacrjournals.org 8039 Cancer Res 2009; 69: (20). October 15, 2009

Cancer Research

Figure 5. Kinesins KIFC3, KIFC1, and KIF5A increase the pools of free tubulin in MDA-MB 231 cells treated with docetaxel. A, cells overexpressing KIFC3 were treated with docetaxel for 24 h. After 2 d, free and polymerized (Pol) tubulin was extracted and Western analysis was performed with anti–a-tubulin. The percentages of free and polymerized tubulin in each cell extract were quantified. All of the experiments above were repeated twice, with very similar results. B, C, and D, cells expressing KIFC1, KIF5A, or KIF1A were treated with docetaxel. Free and polymerized (Pol) tubulin was extracted and Western analysis was performed with anti–a-tubulin. The percentages of free and polymerized tubulin in each cell extract were quantified. The results shown represent two independent experiments. mutagenesis method to generate mutant cells lines through the volume) or sensitive (<25% residual tumor volume) to docetaxel overexpression of a protein due to promoter insertion (10). treatment. There were two probe sets for KIFC3 gene on the Here, we have used this method to discover that the overexpression affymetrix Hg U95-Av2 chip. One probe set showed a very slight of the kinesin KIFC3 confers resistance to docetaxel in breast increase (1.2-fold) in docetaxel-resistant compared with docetaxel- cancer cells. Several studies show that the cellular target of taxanes sensitive tumors and the other showed a decrease of similar is h-tubulin within microtubules (29), and taxane resistance is magnitude (0.8-fold). We noted that the kinesins KIFC1, KIF1A, and thought to involve alterations in microtubule structure and/or KIF5A were slightly overexpressed in docetaxel-resistant tumors function (30). Our discovery that overexpression of the kinesin compared with docetaxel-sensitive tumors, and thus, we investi- KIFC3 mediates resistance to docetaxel in breast cancer cells gated the possible role of these three kinesins in docetaxel represents a previously unknown mechanism. Overexpression of resistance in breast cancer cell lines, finding that overexpression of KIFC3 increases the survival of two estrogen receptor–negative these kinesins indeed does mediate docetaxel resistance in breast metastatic breast cancer cell lines, MDA-MB 231, and MDA-MB cancer cell lines. KIFC1, a minus end–directed, microtubule- 468, upon treatment with docetaxel. We analyzed these particular dependent molecular motor, is required for bipolar spindle cell lines because docetaxel is used to treat estrogen receptor– formation. It is very similar to the kinesins ncd and kar 3, and is negative breast cancer patients (25, 31). reported to be involved in mitosis (34). KIF1A is an anterograde Several studies (23, 32, 33) have used microarrays to study drug- motor protein that transports membranous organelles along resistance mechanisms in breast cancer. Chang and colleagues (23) axonal microtubules. It may play a critical role in the development analyzed gene expression in core biopsy samples of primary breast of axonal neuropathies resulting from impaired axonal transport cancers taken from patients resistant (>25% residual tumor (13). A recent study showed that KIF1A is up-regulated in

Cancer Res 2009; 69: (20). October 15, 2009 8040 www.aacrjournals.org

Docetaxel Resistance Due to Kinesin Overexpression endometrial cancer (35). KIF5A is expressed in nerve tissues (36) are needed to establish this point clearly. Inhibition of microtubule and transports vesicles and membranous organelles (37). KIF5A dynamics by Taxol induces persistent modification of biological and KIF1A are not expressed detectably in the cell lines we have processes and signaling pathways, which ultimately lead to cell analyzed but might be expressed ectopically in specific breast death through the accumulation of signals that finally reach the cancers due, for example, to gene amplification. threshold for the onset of apoptosis (29). Microtubules are dynamic structures with lengths that are tightly The breast cancer cell lines MDA-MB 231 and MDA-MB 468, regulated. KinI or M-kinesins (38, 39) stimulate microtubule with higher levels of free tubulin, showed increased resistance depolymerization in an ATP-dependent fashion (40, 41). The Kin when challenged with docetaxel compared with MCF7 and T47D I kinesin HsMCAK (KNSL6) is overexpressed in human cancers (42, cells, which have lower levels. Furthermore, opposite effects were 43) and the hamster homologue can depolymerize paclitaxel- noticed when the cells were exposed to vincristine, which is a stabilized microtubules (44). We have found that overexpression of clinically used microtubule-destabilizing agent. These results are KIFC3, KIFC1, or KIF5A increase the pools of free tubulin in breast supported by the previous findings that increased levels of cancer cells treated with docetaxel, suggesting that these kinesins polymerized tubulin is associated with increased microtubule may antagonize the effect of docetaxel at least in part by promoting stability, decreased sensitivity to microtubule-depolymerizing the dissociation of tubulin from microtubules. This result supports drugs such as the Vinca alkaloids, and increased sensitivity the previous deduction that the expression of endogenous to the microtubule-polymerizing drug Taxol (47). Therefore, microtubule depolymerizing factors might favor the development the level of microtubule depolymerization in tumors may be of Taxol resistance (45). Our failure to observe microtubule a predictive factor for clinical response to taxane-based depolymerizing activity in cells overexpressing KIF1A is consistent chemotherapy. with the previous finding by Ogawa and colleagues (46) that KIF1A In conclusion, we have shown that overexpression of four does not have this activity. It is likely that KIF1A opposes the different kinesins can mediate docetaxel resistance in breast cancer inhibition of tubulin-microtubule association-dissociation kinetics cell lines and have determined that three of them increase the that is caused by docetaxel rather than affecting the equilibrium fraction of soluble tubulin in cells. Resistance to docetaxel is a between free and microtubule-bound tubulin (2), and the other complex phenomenon that can be orchestrated by a number of kinesins are likely to have the same effect. Kinetic measurements different mechanisms. Our increased understanding of the full

Figure 6. Effects of kinesins on responses to docetaxel or vincristine. A, KIFC3 inhibits docetaxel-mediated apoptosis, **, P < 0.005; ***, P < 0.0005 (vector versus KIFC3). B, relative amounts of free and microtubule-bound tubulin in breast cancer cell lines, correlated with sensitivity to docetaxel. MDA-MB 231, MDA-MB 468, MCF7, and T47D cells were analyzed. C, survival assays were performed on cells treated with docetaxel or untreated. *, P < 0.05 (MDA MB 231 and 468 versus MCF7). D, survival assays were performed on cells treated with vincristine (vincris) or untreated. *, P < 0.05; ***, P < 0.0005 (MDA MB 468 versus MCF7). Columns, means of two experiments, and each measurement was performed in triplicate; bars, SD. www.aacrjournals.org 8041 Cancer Res 2009; 69: (20). October 15, 2009

Cancer Research range of mechanisms contributing to taxane resistance will Disclosure of Potential Conflicts of Interest hopefully enable the development of therapies tailored to the No potential conflicts of interest were disclosed. needs of individual patients. In particular, it will be important to assess the association of kinesin overexpression in breast cancer specimens as a function of progression-free survival in patients Acknowledgments treated with a docetaxel-containing regimen, with the goal of developing kinesins as new prognostic and diagnostic marker. Received 4/2/09; revised 8/4/09; accepted 8/10/09; published OnlineFirst 9/29/09. The costs of publication of this article were defrayed in part by the payment of page Eventually, it may be possible to develop inhibitors of any kinesins charges. This article must therefore be hereby marked advertisement in accordance that are responsible for taxane resistance in tumors, with the hope with 18 U.S.C. Section 1734 solely to indicate this fact. that such inhibitors may synergize with taxanes by interfering with We thank John Pink, John Patton, and Patrick Leahy (Case Western Reserve University) for the assistance and advice, the Flow Cytometry and Gene Expression the resistance mechanism. Inhibitors of some kinesins are already Array Core Facilities of the Case Comprehensive Cancer Center, and the Case known (48). Comprehensive Cancer Center for the seed funding.

References 18. Xu Y, Takeda S, Nakata T, Noda Y, Tanaka Y, adjuvant therapy: derived and validated in two Hirokawa N. Role of KIFC3 motor protein in population-based cohorts. Breast Cancer Res 2005;7: 1. Pusztai L, Rouzier R, Wagner P, Symmans WF. Golgi positioning and integration. J Cell Biol 2002;158: R953–64. Individualized chemotherapy treatment for breast 293–303. 34. Kuriyama R, Kofron M, Essner R, et al. Character- cancer: is it necessary? Is it feasible? Drug Resist Updat 19. Noda Y, Okada Y, Saito N, et al. KIFC3, a microtubule ization of a minus end-directed kinesin-like motor 2004;7:325–31. minus end-directed motor for the apical transport of protein from cultured mammalian cells. J Cell Biol 1995; 2. Kamath K, Wilson L, Cabral F, Jordan MA. hIII-tubulin annexin XIIIb-associated Triton-insoluble membranes. 129:1049–59. induces paclitaxel resistance in association with re- J Cell Biol 2001;155:77–88. 35. Wong YF, Cheung TH, Lo KWK, et al. Identifica- duced effects on microtubule dynamic instability. J Biol 20. Kandel ES, Lu T, Wan Y, Agarwal MK, Jackson MW, tion of molecular markers and signaling pathway in Chem 2005;280:12902–7. Stark GR. Mutagenesis by reversible promoter insertion endometrial cancer in Hong Kong Chinese women by 3. Shalli K, Brown I, Heys SD, Schofield AC. Alterations of to study the activation of NF-nB. Proc Natl Acad Sci U S genome-wide gene expression profiling. Oncogene h-tubulin isotypes in breast cancer cells resistant to A 2005;102:6425–30. 2007;26:1971–82. docetaxel. FASEB J 2005;19:1299–301. 21. Kan CE, Patton JT, Stark GR, Jackson MW. p53- 36. Hirokawa N. Kinesin and dynein superfamily proteins 4. Gottesman MM. Mechanisms of cancer drug resis- mediated growth suppression in response to nutlin-3 in and the mechanism of organelle transport. Science 1998; tance. Annu Rev Med 2002;53:615–7. cyclin D1-transformed cells occurs independently of 279:519–26. 5. Blagosklonny MV, Fojo T. Molecular effects of p21. Cancer Res 2007;67:9862–68. 37. Reid E, Kloos M, Koch AA, et al. A kinesin heavy paclitaxel: myths and reality (a critical review). Int J 22. Ren Y, Liu W, Jiang H, Jiang Q, Feng J. Selective chain (KIF5A) Mutation in hereditary spastic paraplegia Cancer 1999;83:151–6. vulnerability of dopaminergic neurons to microtubule (SPG10). Am J Hum Genet 2002;71:1189–94. 6. Dumontet C, Sikic BI. Mechanisms of action of and depolymerization. J Biol Chem 2005;280:34105–12. 38. Helenius J, Brouhard G, Kalaidzidis Y, Diez S, Howard resistance to antitubulin agents: microtubule dynamics, 23. Chang JC, Wooten EC, Tsimelzon A, et al. Gene J. The depolymerizing kinesin MCAK uses lattice drug transport, and cell death. J Clin Oncol 1999;17: expression profiling for the prediction of therapeutic diffusion to rapidly target microtubule ends. Nature 1061–70. response to docetaxel in patients with breast cancer. 2006;441:115–9. 7. Drukman S, Kavallaris M. Microtubule alterations and Lancet 2003;362:362–9. 39. Miki H, Okada Y, Hirokawa N. Analysis of the kinesin resistance to tubulin-binding agents (review). Int J 24. Hoang EH, Whitehead JL, Dose´ AC, Burnside B. superfamily: insights into structure and function. Oncol 2002;21:621–8. cloning of a novel C-terminal kinesin (KIFC3) that maps Trends Cell Biol 2005;15:467–76. 8. Galletti E, Magnani M, Renzulli ML, Botta M. Paclitaxel to human chromosome 16q13–21 and thus is a 40. Hunter AW, Caplow M, Coy DL, et al. The kinesin- and docetaxel resistance: Molecular mechanisms and candidate gene for Bardet-Biedl syndrome. Genomics related protein MCAK is a microtubule depolymerase development of new generation taxanes. Chem Med 1998;52:219–22. that forms an ATP-hydrolyzing complex at microtubule Chem 2007;2:920–42. 25. McGrogan BT, Gilmartin B, Carney DN, McCann A. ends. Mol Cell 2003;11:445–57. 9. McGrogan BT, Gilmartin B, Carney DN, McCann A. Taxanes, microtubules and chemoresistant breast can- 41. Desai A, Verma S, Mitchison TJ, Walczak CE. Kin I Taxanes, microtubules and chemoresistant breast cancer. Biochimica et Biophysica Acta (BBA)—Reviews on kinesins are microtubule-destabilizing enzymes. Cell cer. Biochim Biophys Acta 2008;1785:96–132. Cancer 2008;1785:96–132. 1999;96:69–78. 10. Lu T, Jackson MW, Singhi AD, et al. Validation-based 26. Kimchi A. Antisense libraries to isolate tumor 42. Scanlan MJ, Welt S, Gordon CM, et al. Cancer-related insertional mutagenesis identifies lysine demethylase suppressor genes. Methods Mol Biol 2003;222:399–412. serological recognition of human colon cancer: identi- FBXL11 as a negative regulator of NFnB. Proc Natl Acad 27. Berns K, Hijmans EM, Mullenders J, et al. A large- fication of potential diagnostic and immunotherapeutic Sci U S A. In press 2009. scale RNAi screen in human cells identifies new targets. Cancer Res 2002;62:4041–7. 11. Hirokawa N, Noda Y, Okada Y. Kinesin and dynein components of the p53 pathway. Nature 2004;428: 43. Perou CM, Jeffrey SS, van de Rijn M, et al. Distinctive superfamily proteins in organelle transport and cell 431–37. gene expression patterns in human mammary epithelial division. Curr Opin Cell Biol 1998;10:60–73. 28. Lih CJ, Wei W, Cohen SN. Txr1: a transcriptional cells and breast cancers. Proc Natl Acad Sci U S A 1999; 12. Sharp DJ, Rogers GC, Scholey JM. Microtubule regulator of thrombospondin-1 that modulates cellular 96:9212–7. motors in mitosis. Nature (London) 2000;407:41–7. sensitivity to taxanes. Genes Dev 2006;20:2082–95. 44. Maney T, Wagenbach M, Wordeman L. Molecular 13. Hirokawa N, Noda Y. Intracellular transport and 29. Mollinedo F, Gajate C. Microtubules, microtubule- dissection of the microtubule depolymerizing activity of kinesin superfamily proteins, KIFs: structure, function, interfering agents and apoptosis. Apoptosis 2003;8:413–50. mitotic centromere-associated kinesin. J Biol Chem and dynamics. Physiol Rev 2008;88:1089–118. 30. Gonc¸alvesA, Braguer D, Kamath K, et al. Resistance 2001;276:34753–8. 14. Miki H, Setou M, Kaneshiro K, Hirokawa N. All to Taxol in lung cancer cells associated with increased 45. Jordan MA, Wilson L. Microtubules as a target for kinesin superfamily protein, KIF, genes in mouse and microtubule dynamics. Proc Natl Acad Sci U S A 2001;98: anticancer drugs. Nat Rev Cancer 2004;4:253–65. human. Proc Natl Acad Sci U S A 2001;98:7004–11. 11737–42. 46. Ogawa T, Nitta R, Okada Y, Hirokawa N. A 15. Saito N, Okada Y, Noda Y, Kinoshita Y, Kondo S, 31. Henderson IC, Berry DA, Demetri GD, et al. common mechanism for microtubule destabilizers-M Hirokawa N. KIFC2 is a novel neuronspecific C-terminal Improved outcomes from adding sequential paclitaxel type kinesins stabilize curling of the protofilament type kinesin superfamily motor for dendritic transport but not from escalating doxorubicin dose in an using the class-specific neck and loops. Cell 2004;116: of multivesicular body-like organelles. Neuron 1997;18: adjuvant chemotherapy regimen for patients with 591–602. 425–38. node-positive primary breast cancer. J Clin Oncol 47. Hari M, Wang Y, Veeraraghavan S, Cabral F. 16. Hanlon DW, Yang Z, Goldstein LS. Characterization 2003;21:976–83. Mutations in a- and h-tubulin that stabilize micro- of KIFC2, a neuronal kinesin superfamily member in 32. Kudoh K, Ramanna M, Ravatn R, et al. Monitoring tubules and confer resistance to colcemid and vinblas- mouse. Neuron 1997;18:439–51. the expression profiles of doxorubicin-induced and tine. Mol Cancer Ther 2003;2:597–605. 17. Yang Z, Xia C, Roberts EA, Bush K, Nigam SK, doxorubicin-resistant cancer cells by cDNA microarray. 48. Mu¨ller C, Gross D, Sarli V, et al. Inhibitors of kinesin Goldstein LS. Molecular cloning and functional analysis Cancer Res 2000;60:4161–6. Eg5: antiproliferative activity of monastrol analogues of mouse C-terminal kinesin motor KifC3. Mol Cell Biol 33. Pawitan Y, Bjohle J, Amler L, et al. Gene expression against human glioblastoma cells. Cancer Chemother 2001;21:765–70. profiling spares early breast cancer patients from Pharmacol 2007;59:157–64.

Cancer Res 2009; 69: (20). October 15, 2009 8042 www.aacrjournals.org

Overexpression of Kinesins Mediates Docetaxel Resistance in Breast Cancer Cells

Sarmishtha De, Rocky Cipriano, Mark W. Jackson, et al.

Cancer Res 2009;69:8035-8042. Published OnlineFirst September 29, 2009.

Updated version Access the most recent version of this article at: doi:10.1158/0008-5472.CAN-09-1224

Cited articles This article cites 47 articles, 18 of which you can access for free at: http://cancerres.aacrjournals.org/content/69/20/8035.full#ref-list-1

Citing articles This article has been cited by 17 HighWire-hosted articles. Access the articles at: http://cancerres.aacrjournals.org/content/69/20/8035.full#related-urls

E-mail alerts Sign up to receive free email-alerts related to this article or journal.

Reprints and To order reprints of this article or to subscribe to the journal, contact the AACR Publications Subscriptions Department at [email protected].

Permissions To request permission to re-use all or part of this article, use this link http://cancerres.aacrjournals.org/content/69/20/8035. Click on "Request Permissions" which will take you to the Copyright Clearance Center's (CCC) Rightslink site.