Fax: (203) 785-2850
Stathmin and tubulin expression and survival of ovarian cancer patients receiving platinum treatment with and without paclitaxel
Article first published online: 25 MAR 2009
Copyright © 2009 American Cancer Society
Volume 115, Issue 11, pages 2453–2463, 1 June 2009
How to Cite
Su, D., Smith, S. M., Preti, M., Schwartz, P., Rutherford, T. J., Menato, G., Danese, S., Ma, S., Yu, H. and Katsaros, D. (2009), Stathmin and tubulin expression and survival of ovarian cancer patients receiving platinum treatment with and without paclitaxel. Cancer, 115: 2453–2463. doi: 10.1002/cncr.24282
- Issue published online: 20 MAY 2009
- Article first published online: 25 MAR 2009
- Manuscript Accepted: 18 NOV 2008
- Manuscript Revised: 5 NOV 2008
- Manuscript Received: 25 AUG 2008
- ovarian cancer;
Paclitaxel interacts with microtubules to exert therapeutic effects. Molecules that affect microtubule activity, such as βIII-tubulin and stathmin, may interfere with the treatment. In this study, the authors analyzed βIII-tubulin and stathmin expression in ovarian tumors and examined their associations with treatment response and patient survival.
The study included 178 patients with epithelial ovarian cancer who underwent cytoreductive surgery followed by platinum-based chemotherapy; of these patients, 75 also received paclitaxel. Fresh tumor samples that were collected at surgery were analyzed for messenger RNA expression of βIII-tubulin and stathmin using real-time polymerase chain reaction analysis. Associations of these molecules with treatment response, disease progression, and overall survival were evaluated.
High stathmin expression was associated with worse disease progression-free and overall survival compared with low stathmin expression. This association was independent of patient age, disease stage, tumor grade, histology, and residual tumor size and was observed in patients who received platinum plus paclitaxel, but not in patients who received platinum without paclitaxel, suggesting that stathmin expression in tumor tissue may interfere with paclitaxel treatment. Similar effects were not observed for βIII-tubulin, although high βIII-tubulin expression was associated with disease progression among patients who received platinum without paclitaxel. No associations were observed between treatment response and tubulin or stathmin expression. Expression levels of βIII-tubulin and stathmin were correlated significantly.
High stathmin expression predicted an unfavorable prognosis in patients with ovarian cancer who received paclitaxel and platinum chemotherapy. This finding supports the possibility that stathmin may interfere with paclitaxel treatment, leading to a poor prognosis for patients with ovarian cancer. Cancer 2009. © 2009 American Cancer Society.
Standard treatment for advanced ovarian cancer includes cytoreductive surgery followed by chemotherapy with paclitaxel and platinum.1 Although many patients respond to the treatment initially, most patients still experience disease progression.2, 3 Intrinsic and acquired drug resistance contribute to poor clinical outcomes; more research is needed to elucidate their mechanisms. Proposed mechanisms for tumor resistance to platinum and paclitaxel include molecular and genetic changes in DNA repair genes,4 multidrug resistance genes,5, 6 apoptotic regulatory genes,7, 8 and microtubule stability and dynamics.9, 10 Recently, several studies have indicated that excision repair cross-complementation group 1 (ERCC1), a DNA repair gene, is a predictive marker for platinum-based chemotherapy.11, 12 Our previous research also produced similar results among patients with ovarian cancer who received platinum without paclitaxel. However, for patients who received both platinum and paclitaxel, the effect of ERCC1 was not observed.13 Thus, molecular and genetic factors that may influence patient response to paclitaxel also should be considered and investigated.
Microtubules are dynamic α/β-tubulin heterodimers that function in cell division, intracellular transport, maintenance of cell shape, and cellular motility.14 Paclitaxel induces mitotic arrest and cell death when it binds and stabilizes β-tubulin in microtubules.15 Identification of the paclitaxel binding site in β-tubulin has led to the hypothesis that structural alterations in this protein may contribute to paclitaxel resistance.16 Over expression of 1 isotype, βIII-tubulin, has been associated with paclitaxel resistance in mammalian cell lines17 and with poor survival in patients with ovarian cancer.18 Stathmin is another protein that may play a role in paclitaxel resistance, because it regulates microtubule dynamics by preventing tubulin polymerization and promoting microtubule destabilization and disassembly.14 High stathmin expression has been observed in paclitaxel-resistant ovarian cancer cell lines19 and has been associated with tumor progression and poor prognosis in hepatocellular carcinoma20 and oral squamous cell carcinoma.21 However, few clinical studies have examined the role of stathmin expression in the treatment and survival of patients with ovarian cancer. In the current study, we evaluated βIII-tubulin and stathmin messenger RNA (mRNA) expression in ovarian tumors that were collected from patients who received platinum-based chemotherapy with or without paclitaxel. We speculated that high levels of βIII-tubulin and stathmin expression would adversely affect patient response to chemotherapy involving paclitaxel and would be associated with poor survival outcomes.
MATERIALS AND METHODS
A clinical study of ovarian cancer was conducted in the Gynecologic Oncology Unit at the University of Turin in Italy between October 1991 and February 2000. From a consecutive series of 212 patients who were diagnosed with primary epithelial ovarian cancer, we identified 178 patients who received postoperative platinum-based chemotherapy with or without paclitaxel. The median age of patients at surgery was 57.4 years (range, 26-82 years). Thirty-three of 178 patients (18.5%) were diagnosed with stage I disease, 12 patients (6.7%) had stage II disease, 120 patients (67.4%) had stage III disease, and 13 patients (7.3%) had stage IV disease. Disease staging was classified according to International Federation of Gynecology and Obstetrics criteria.22 On the basis of World Health Organization (WHO) criteria,23 histologic tumor types included serous, endometrioid, mucinous, clear cell, and other epithelial tumors. For data analysis, tumor histotypes were grouped into serous (43.3%; n = 77) and nonserous (56.7%; n = 101). Most patients (68%; n = 121) had grade 3 tumors (poorly differentiated), and a few patients had grade 2 tumors (19.1%; n = 34) and grade 1 tumors (12.9%; n = 23). Debulking results, which were evaluated after surgical cytoreduction, were optimal for 87 patients (48.9%) and suboptimal (residual lesions >1 cm) for 91 patients (51.1%). This study was approved by an ethical review committee at the University of Turin, and all patients provided consent to participate.
After surgery, all patients received platinum-based chemotherapy: 75 patients also received paclitaxel (42.1%), whereas 103 patients were treated without paclitaxel (57.9%). For patients who were treated between 1991 and 1995, cisplatin 75 mg/m2 and cyclophosphamide 750 mg/m2 were used every 3 weeks for 6 cycles as the standard of care except for patients who had allergic reactions or medical contraindications and patients who were on experimental protocols. From 1996 to 2001, after the introduction of paclitaxel as a standard of care, the treatment was changed to paclitaxel 175 mg/m2 (3 hours infusion) and carboplatinum at an area under the concentration time curve of 5 or 6, every 3 weeks for 6 cycles. Patient response to chemotherapy was evaluated 1 month after the last cycle of treatment through clinical examination, imaging, and serum CA-125. For measurable disease, treatment response was assessed according to WHO criteria.24 A complete response (CR) required the complete disappearance of all measurable lesions, whereas a partial response (PR) was defined as a reduction ≥50% in the size of measurable lesions. Stable disease (SD) was defined as a decrease <50% or an increase ≤25% in the size of measurable lesions, and progressive disease (PD) was defined as either an increase >25% in the size of measurable lesions or the appearance of new lesions. For nonmeasurable disease, progression was defined as a doubling of CA-125 from the upper limit of normal.25 For data analysis, patients who had a PR, SD, and PD were grouped together as poor responders and were compared with patients who had a CR. Of the 178 patients in the study, 71.9% (n = 128) had a CR to treatment, and 27% (n = 48) had poor response, which included 36 with PRs, 4 patients with SD, and 8 patients with PD. Two patients (1.1%) had no information on treatment response. Patients were followed from surgery through June 2001 for survival outcome. The median overall survival, which we defined as the time between the date of surgery and the date of either death or last contact, was 37.5 months (range, 0.6-114.1 months). The median disease progression-free survival from the date of surgery to the date of first local recurrence, distant metastasis, death without recurrence, or last contact was 19.8 months (range, 0.6-108.8 months). During follow-up, 92 patients (51.7%) experienced disease progression; among them, 67 patients died, and 25 patients survived for the duration of the study. Of the 86 patients (48.3%) without disease progression, 18 patients died without remission, and 68 patients survived for the duration of the study.
Fresh tumor samples were collected during surgery. The specimens were snap-frozen in liquid nitrogen and stored at −80°C until analysis. Representative samples from each tissue specimen were examined in frozen section by 2 pathologists to confirm tumor content; the content of tumor cells in these specimens ranged between 80% and 90%. RNeasy Mini Kits (QIAGEN Inc., Valencia, Calif) were used to extract total RNA from 20 mg of tissue that was pulverized manually in liquid nitrogen. RNA (5 ng) was reverse transcribed to complementary DNA (cDNA) using SuperScript First-Strand Synthesis System for reverse transcriptase-polymerase chain reaction (PCR) (Invitrogen Corp., Carlsbad, Calif). Quantitative real-time PCR was performed to determine levels of stathmin and βIII-tubulin mRNA expression in each tumor sample by using glyceraldehyde 3-phosphate dehydrogenase (GAPDH) as an endogenous control. Plasmids that contained cloned sequences of stathmin, βIII-tubulin, and GAPDH were used as the standards for quantitative PCR. The primers for cloned sequences were designed online (available at: http://www.idtdna.com accessed on March 11, 2009) and were ordered from Integrated DNA Technologies, Inc. (Coralville, Iowa). The primer sequences were: AAA TGG CTG CCA AAC TGG AAC GT (stathmin forward), GCT TCA GTC TCG TCA GCA GGG TC (stathmin reverse), ACA GCA GCT ACT TCG TGG AGT GGA TC (βIII-tubulin forward), GTC TTC GTA CAT CTC GCC CTC TTC C (βIII-tubulin reverse), GAA GGT GAA GGT CGG AGT C (GAPDH forward), and GAA GAT GGT GAT GGG ATT TC (GAPDH reverse). Purified PCR products (3 μL) were ligated to pGEM-T Easy Vector and incubated at 4°C overnight (Promega Corp., Madison, Wis). The ligation products (5 mL) were transfected into 50 μL of subcloning efficiency DH5 alpha competent Escherichia coli cells (Invitrogen Corp.). After overnight incubation at 37°C, plasmid DNA was purified using the QIAprep MiniPrep Kit (Qiagen Inc., Valencia, Calif), and the purified plasmid was digested by EcoR1 to confirm the size of the inserted DNA. After measuring the concentration of purified DNA, gene copy numbers were calculated according to the following formula: (concentration × volume)/(size of DNA insert × 660) × (6.02 × 1023). Cloned DNA templates were serially diluted to contain copy numbers ranging from 1.0 × 108 to 1.0 × 1012 per μL.
Quantitative real-time PCR was carried out using the ABI 7500 real-time PCR system (Applied Biosystems Inc., Foster City, Calif). The PCR reaction solution (25 μL) contained 5 ng/μL cDNA, 12.5 μL Power SYBR Green PCR Master Mix (Applied Biosystems Inc.), and a pair of primers at a final concentration of 5 μM for stathmin, βIII-tubulin, or GAPDH. Thermal cycling conditions consisted of an initial 10-minute denature at 95°C and 40 cycles of annealing at 60°C for 1 minute and denaturing at 95°C for 15 seconds. Dissociation curve analysis was performed after PCR to confirm the size of PCR products. All samples were tested in duplicate along with negative controls and serially diluted standards. On the basis of the standard curve, SDS software calculated the copy number for stathmin, βIII-tubulin, and GAPDH. The ratios of stathmin or βIII-tubulin to GAPDH were used to determine the expression index (EI). Examples of PCR results are shown in Figures 1 through 3.
Stathmin and βIII-tubulin mRNA expression levels were analyzed both as continuous (EI) and categorical variables. For categorical analysis, the expression data were grouped into low, middle, and high categories based on tertile distributions. Associations of molecular markers with clinical variables were analyzed using the Spearman correlation coefficient and the Wilcoxon rank-sum test for continuous variables and using the chi-square test for categorical variables. Survival analysis was performed for progression-free and overall survival using Cox proportional hazards regression to adjust for clinical and pathologic variables. Kaplan-Meier survival curves were constructed to illustrate survival differences. Data analysis was performed using SPSS software (version 11.0; SPSS Inc., Chicago, Ill).
Table 1 illustrates stathmin and βIII-tubulin expression in relation to patient age at surgery, disease stage, tumor grade, and residual tumor size. Stathmin and βIII-tubulin expression was not correlated with these clinical features, but high stathmin expression was associated with high βIII-tubulin expression (P < .001). Stathmin expression in ovarian tumors was greater than βIII-tubulin in ovarian tumors (1.29 vs 0.06) (Table 2). Like the results shown in Table 1, stathmin and βIII-tubulin expression did not differ by disease stage (stage I/II vs stage III/IV) or tumor grade (grade 1/2 vs grade 3) (Table 2). In addition, expression levels did not differ according to the presence or absence of residual tumor or according to debulking results (optimal vs suboptimal) (Table 2). Stathmin expression, however, was greater in nonserous tumors than in serous tumors (2.01 vs 0.90; P = .007), but no difference between histotypes was observed for βIII-tubulin expression (Table 2).
|Variable||Correlation Coefficient (P Value)|
|Stathmin Expression||βIII-Tubulin Expression|
|Age||0.04 (.570)||−0.05 (.506)|
|Disease stage||−0.05 (.479)||0.06 (.454)|
|Tumor grade||−0.03 (.677)||−0.01 (.905)|
|Residual tumor size||−0.00 (.989)||0.10 (.176)|
|Stathmin Expression||βIII-Tubulin Expression|
|All patients||178||1.29 (0.00-463.00)||0.06 (0.00-16.33)|
|I-II||45||1.36 (0.00-64.00)||0.05 (0.00-2.32)|
|III-IV||133||1.27 (0.00-463.00)||1.27 (0.00-16.33)|
|1-2||57||1.41 (0.00-463.00)||0.06 (0.00-12.00)|
|3||121||1.27 (0.00-109.00)||0.06 (0.00-16.33)|
|Nonserous||101||2.01 (0.00-463.00)||0.07 (0.00-16.33)|
|Serous||77||0.90 (0.00-109.00)||0.04 (0.00-5.65)|
|Residual tumor, n=178|
|No residual||70||1.23 (0.00-64.00)||0.05 (0.00-2.82)|
|Residual||108||1.33 (0.00-463.00)||0.07 (0.00-16.33)|
|Debulking results, n=178|
|Optimal||87||1.16 (0.00-64.00)||0.05 (0.00-16.33)|
|Suboptimal||91||1.41 (0.00-463.00)||0.06 (0.00-12.00)|
|Treatment response, n=176‡|
|Complete||128||1.28 (0.00-463.00)||0.05 (0.00-16.33)|
|Poor||48||1.25 (0.00-109.00)||0.07 (0.00-4.50)|
Patients who had a CR versus a poor response to treatment did not have different stathmin or βIII-tubulin expression levels (Table 2). Survival analysis, however, indicated that different levels of stathmin expression were associated with disease progression-free survival and overall survival (Table 3). Patients who had high stathmin levels had an elevated risk of disease progression and death compared with patients who had low stathmin expression levels. The hazard ratio (HR) was 2.43 (95% confidence interval [CI], 1.41-4.19) for disease progression and 2.77 (95% CI, 1.53-5.04) for death, and the trend linking increasing stathmin levels with risk also was statistically significant for both outcomes (P ≤ .001) (Table 3). Furthermore, an analysis stratified by treatment indicated that these associations applied only to patients who received both platinum and paclitaxel. The strength of the association was substantial in these patients; the HRs reached as high as 4.87 (95% CI, 1.69-14.05) and 7.81 (95% CI, 2.32-26.25), respectively (Table 3). Similar survival differences were not observed for patients who received platinum without paclitaxel. These analyses were adjusted for patient age, disease stage, tumor grade, histology, and residual tumor size. Kaplan-Meier survival curves indicated that platinum and paclitaxel-treated patients who had medium or high stathmin expression had worse progression-free survival (Fig. 4, top) and overall survival (Fig. 5, top) than patients who had low expression, but no differences were observed for patients who received platinum without paclitaxel (Fig. 4, bottom; Fig. 5, bottom).
|Expression Category||Platinum With or Without Paclitaxel n=178||Platinum Without Paclitaxel n=103||Platinum With Paclitaxel n=75|
|HR||95% CI||P||HR||95% CI||P||HR||95% CI||P|
|High stathmin expression||2.77†||1.53-5.04†||<.001†||1.65||0.81-3.35||.169||7.81†||2.32-26.25†||<.001†|
Although βIII-tubulin expression also was associated with survival outcome, the association was observed only for disease progression-free survival and not for overall survival. Compared with low expression, patients with high βIII-tubulin expression had an elevated risk of disease progression (HR, 1.82; 95% CI, 1.07-3.10), and the trend also was significant (P = .026) (Table 3). The risk for death was slightly greater for patients who had high βIII-tubulin levels, but the difference was not significant (P = .279). Analysis stratified by treatment demonstrated that the association between βIII-tubulin expression and disease progression was significant only for patients who received platinum without paclitaxel. For patients who received both drugs, no significant difference was observed in survival between those with high versus low βIII-tubulin expression. For overall survival, no significant association was observed in either group of patients, although patients who received platinum without paclitaxel tended to have a greater risk of death if they had tumors with high βIII-tubulin expression (P = .081 for trend) (Table 3).
Recently, much attention has been given to the connection between microtubule dynamics and paclitaxel resistance, underscoring the importance of proteins that interact with microtubules, including stathmin and βIII-tubulin.9, 10 Previous studies linked elevated levels of stathmin to paclitaxel resistance in ovarian,19 breast,26 and lung cancer cell lines,10 but few investigated the role of stathmin expression in paclitaxel resistance in vivo. In the current study, we examined the association between stathmin and clinical outcome in patients with ovarian cancer who received platinum-based chemotherapy with and without paclitaxel. It is noteworthy that, in our study, stathmin expression was not related to treatment response but was associated with disease progression and overall survival. After adjusting for patient age, disease stage, tumor grade and histology, and residual tumor status, the risk of disease progression and death was elevated significantly in patients who had high stathmin expression, and the association was more evident in patients who received both platinum and paclitaxel, but not in patients who received platinum without paclitaxel. This discrepancy in association by paclitaxel treatment suggests that this drug's therapeutic effect may be affected by the level of stathmin expression.
Patient response to cancer treatment has been considered an early prognostic indicator, based on the assumption that patients who initially respond well to treatment will have better survival outcomes.27 However, early favorable response to treatment does not always translate into long-term survival benefits; patients with similar response to initial treatment can experience different survival over time. The relation between treatment response and survival also varies from study to study because of inconsistent clinical definitions and protocols for treatment and evaluation of response.28 In the current study, treatment response was evaluated 1 month after patients received chemotherapy. It is possible that the treatment response was assessed too soon to be relevant to survival outcomes. Additional complexity in clinical and pathologic variations among patients also may have contributed to our inability to detect associations between stathmin and treatment response, although stathmin was related to survival outcomes.
Although adding paclitaxel to platinum-based chemotherapy as first-line treatment has improved patient survival, the improvement is limited. Some patients who initially respond to the treatment still develop drug resistance and eventually die of the disease. Thus, understanding the molecular mechanisms that are responsible for paclitaxel resistance is critical to further improving the efficacy of chemotherapy. Paclitaxel is an antitumor agent which is believed to exert its cytotoxic effect through stabilizing microtubules. Paclitaxel can enhance the polymerization of tubulin and bind directly to β-tubulin to disturb the formation of mitotic apparatus, leading to mitotic arrest and cells death.15 Conversely, stathmin is a microtubule-destabilizing protein that regulates microtubule dynamics by preventing tubulin polymerization and promoting microtubule disassembly during cell cycle progression.14, 29 Thus, up-regulation of stathmin expression may disturb microtubule stabilization and thereby may overcome the effect of paclitaxel on microtubule polymerization. In addition to its role in microtubule dynamics, stathmin is involved in the regulation of cell proliferation, differentiation, and cell motility, which also may affect tumor progression.30, 31
Our findings suggest that stathmin mRNA expression is a strong predictor of ovarian cancer survival after standard platinum and paclitaxel treatment. This finding may be explained by several plausible mechanisms, and paclitaxel resistance is 1 of the possibilities. Previously, it was demonstrated that 2 paclitaxel-resistant ovarian cancer cell lines had stathmin overexpression, implicating stathmin involvement in tumor response to paclitaxel treatment.19 In addition, overexpression of stathmin in breast cancer cell lines decreased microtubule polymerization and caused cell cycle arrest at G2 phase, markedly reducing the binding ability of paclitaxel to microtubules.26 In the same breast cancer cells, inhibition of stathmin by RNA interference induced microtubule polymerization, promoted cell cycle progression from G2 to M phase, and increased cell sensitivity to paclitaxel, suggesting the involvement of stathmin in the efficacy of antimicrotubule therapies.32 Recent experiments demonstrated that stathmin suppression plus taxol exposure had stronger inhibitory effects than either 1 of the factors on prostate cancer cell proliferation and angiogenesis, presenting further evidence that stathmin may interfere with the therapeutic effect of taxol on cancer cells.33, 34 Our observation of an association between stathmin expression and ovarian cancer survival supports the notion that stathmin may be involved in tumor resistance to paclitaxel therapy. Disease traits, such as tumor aggressiveness and metastatic potential, may affect stathmin expression. Although we did not detect a significant association between stathmin expression and clinical or pathologic factors, previous studies have linked stathmin expression to tumor grade, clinical stage, disease progression, and poor prognosis in breast and lung cancers, suggesting that the effect of stathmin may be more global than a specific interaction with paclitaxel would suggest.35, 36
In the current study, we also observed that βIII-tubulin expression was associated with ovarian cancer survival, but the association was limited to progression-free survival among patients who received platinum without paclitaxel. Although we predicted that βIII-tubulin expression would be associated with survival for all patients, it is interesting to note that inconsistent associations also were observed in previous studies. Previously, βIII-tubulin expression was identified as a prognostic factor for patients with various forms of cancer. The overexpression of βIII-tubulin has been associated with poor survival in patients with gastric cancer,37 breast cancer,38 nonsmall cell lung cancer,39 and ovarian cancer.18 However, a recent study of patients with cervical cancer was unable to detect any significant association between βIII-tubulin expression and treatment response or survival.40 Prior research also has linked βIII-tubulin overexpression with paclitaxel resistance in patients with breast cancer,41 nonsmall cell lung cancer,42, 43 and ovarian cancer.44 In vitro studies have suggested that high βIII-tubulin expression may contribute to paclitaxel resistance by reducing microtubule assembly and preventing paclitaxel from suppressing microtubule dynamics.9, 45 Although βIII-tubulin overexpression has been associated with paclitaxel resistance in prostate cancer cells46 and ovarian cancer tumors,47 no association could be detected in ovarian cancer xenografts.17 These conflicting findings and the null results observed in our study suggest that the role of βIII-tubulin expression in paclitaxel resistance, treatment response, and survival should be examined further in vitro and in vivo.
Although the underlying mechanism remains unknown, our study has suggested that stathmin mRNA expression is an important prognostic indicator for patients with ovarian cancer who are treated with paclitaxel and platinum chemotherapy. In this study, high levels of stathmin were associated significantly with shorter disease progression-free and overall survival. These findings suggest that stathmin may interfere with platinum and paclitaxel treatment, and they highlight the clinical importance of investigating the effect of this molecule on paclitaxel treatment in patients with ovarian cancer.
Conflict of Interest Disclosures
Dan Su was supported in part by grant 2005C33020 from the Science and Technology Department of Zhejiang Province, China.
Dionyssios Katsaros was supported in part by a Regione Piemonte grant from Progetto Ricerca Sanitaria Finalizzata, 2008.
- 2Carboplatin plus paclitaxel as first-line chemotherapy in previously untreated advanced ovarian cancer. German AGO Study Group Ovarian Cancer. Arbeitsgemeinschaft Gynakologische Onkologie. Semin Oncol. 1997; 24(4 suppl 11 ): S11-28-S11-33., , , et al.
- 3Paclitaxel plus carboplatin versus paclitaxel plus alternating carboplatin and cisplatin for initial treatment of advanced ovarian cancer: long-term efficacy results: a Hellenic Cooperative Oncology Group (HeCOG) study. Ann Oncol. 2005; 16: 1116-1122., , , et al.
- 15Taxol (paclitaxel): mechanisms of action. Ann Oncol. 1994; 5( suppl 6): S3-S6..
- 20Stathmin overexpression cooperates with p53 mutation and osteopontin overexpression, and is associated with tumour progression, early recurrence, and poor prognosis in hepatocellular carcinoma. J Pathol. 2006; 209: 549-558., , , et al.
- 42Expression of class III beta tubulin in non-small cell lung cancer is correlated with resistance to taxane chemotherapy [serial online]. Bull Cancer. 2005; 92: E25-E30., , , et al.