European Journal of Cancer (2014) 50, 2328– 2335

A v a i l a b l e a t w w w . s c i e nc e d i r e c t . c o m

ScienceDirect

jour na l homepage : www.e jcancer . com

Combination of microtubule associated protein-tauand b-tubulin III predicts chemosensitivity of paclitaxelin patients with advanced gastric cancer

http://dx.doi.org/10.1016/j.ejca.2014.06.017

0959-8049/� 2014 Elsevier Ltd. All rights reserved.

⇑ Corresponding author: Address: Fu-Cheng Road 52, Hai-Dian District, Beijing 100142, China. Tel./fax: +86 10 88196561.E-mail address: lin100@medmail.com.cn (L. Shen).

1 These authors contributed equally to this study.

Jingwei Yu a,1, Jing Gao a,1, Zhihao Lu a,1, Jifang Gong a, Yanyan Li a, Bin Dong b,Zhongwu Li b, Xiaotian Zhang a, Lin Shen a,⇑

a Department of Gastrointestinal Oncology, Key Laboratory of Carcinogenesis and Translational Research (Ministry of Education), Peking

University Cancer Hospital and Institute, Beijing, Chinab Department of Pathology, Key Laboratory of Carcinogenesis and Translational Research (Ministry of Education), Peking University

Cancer Hospital and Institute, Beijing, China

Received 25 February 2014; received in revised form 16 May 2014; accepted 23 June 2014Available online 10 July 2014

KEYWORDS

MAP-tauTUBB3PaclitaxelGastric cancer

Abstract Aim: To investigate the role of microtubule associated protein-tau (MAP-tau) andb-tubulin III (TUBB3) in predicting the chemosensitivity of paclitaxel in patients withadvanced gastric cancer (GC).Methods: MAP-tau and TUBB3 expressions were detected using immunohistochemistry in244 advanced GC patients prior to chemotherapy. The associations of MAP-tau and TUBB3expressions with paclitaxel sensitivity were assessed using in vitro and in vivo xenograftanalysis.Results: A total of 149 patients receiving paclitaxel plus capecitabine (cohort 1) and 95patients receiving cisplatin plus capecitabine (cohort 2) were included in this study. In cohort1, the clinical benefit rate (CBR), median progression-free survival (PFS) and overall survival(OS) were found to be significantly higher in patients with low levels of MAP-tau and TUBB3expressions and were significantly higher than those in patients with high levels of MAP-tauand TUBB3 expressions (CBR: 72.2% versus 35.9%; PFS: 271 versus 102 days; OS: 394 versus173 days; all P < 0.05). This was not observed in cohort 2. In in vitro studies, the sensitivity ofpaclitaxel in human gastric cancer cells was inversely correlated with the expression levels ofMAP-tau and TUBB3, as in in vivo animal xenografts.

J. Yu et al. / European Journal of Cancer 50 (2014) 2328–2335 2329

Conclusions: The combination of MAP-tau and TUBB3 was found to predict chemosensitiv-ity to paclitaxel in gastric cancer in vitro and in vivo. This merits further study and may helpguide individual therapy.

� 2014 Elsevier Ltd. All rights reserved.

1. Introduction

Taxane is a microtubule-stabilising agent with broadrange of activity in both haematological and solidtumours, including gastric cancer (GC) tumours [1].Although it has potent antitumour activity, primaryand acquired resistance to taxane is not rare in clinicalsituations. This limits its utility [2–4]. According to pub-lished reports, resistance to taxane involves multiplemechanisms, including increased drug efflux, tubulinmutation, altered microtubule dynamics and impairedcell death signalling [5]. Because the target of taxane ismicrotubulin, attention has been paid to tubulin ormicrotubule-associated protein in investigations of themechanisms underlying taxane resistance.

A subtype of tubulin, class III beta-tubulin (b-tubulinIII, TUBB3) has been reported to be overexpressed in sev-eral types of cancer cells with paclitaxel resistance [6–9].TUBB3 expression has also been found to be inverselycorrelated with the clinical outcomes of breast cancer,ovarian cancer and non-small-cell lung cancer in patientsreceiving taxane-containing chemotherapy [10–12], aswell as in recurrent and metastatic gastric cancer [13]. Inrecent years, microtubule-associated protein-tau (MAP-tau) has been investigated with respect to the mechanismunderlying paclitaxel resistance. MAP-tau functions pri-marily by enabling tubulin assembly and microtubule sta-bilisation and may bind to the paclitaxel-binding site onthe inner surface of the microtubule [3]. Several studieshave suggested that breast cancer patients with low levelsof MAP-tau expression could benefit from paclitaxel ther-apy [14,15]. High levels of MAP-tau expression have beenfound to be critical to the chemoresistance of gastric can-cer to paclitaxel [11,16]. However, according to the resultsof the famous NSABP-B 28 randomised clinical trial, nosignificant interaction exists between MAP-tau expres-sion and the benefits of paclitaxel [17].

The present study was conducted because paclitaxel-containing chemotherapy is now widely used in locallyadvanced and metastatic gastric cancer patients [18].The role of MAP-tau combined with TUBB3 in predict-ing chemosensitivity to paclitaxel in advanced gastriccancer patients was investigated in vitro and in vivo.

2. Materials and methods

2.1. Patients and sample collection

This study included 149 advanced GC patients receiv-ing first-line paclitaxel plus capecitabine (cohort 1,

n = 149) and 95 advanced GC patients receiving first-line cisplatin plus capecitabine (cohort 2, n = 95) inthe Department of Gastrointestinal Oncology of PekingUniversity Cancer Hospital from December 2006 toDecember 2010. All patients were pathologically con-firmed to have locally advanced or metastatic gastriccancer. Patients in cohort 1 were from a published phaseII prospective study (Chinese Clinical Trial Registry:ChiCTR-TRC-07000014) [19]. They served as the exper-imental group. Patients in cohort 2 were collected con-tinuously during the same period and served as thecontrol group. All patients had formalin-fixed and par-affin-embedded (FFPE) tumour sections made beforechemotherapy, and provided written informed consentfor their tissues to be used in future research. This studywas approved by the Medical Ethics Committee of Pek-ing University Cancer Hospital and was performedaccording to the Declaration of Helsinki Principles.

2.2. Evaluation of clinical response

Clinical response was evaluated using computedtomography (CT) scans every two cycles according toRECIST 1.0 criteria. Responses were defined as com-plete response (CR), partial response (PR), stable dis-ease (SD) or progressive disease (PD) and werereviewed from the electronic medical records. In orderto determine the exact relationship between biomarkersand sensitivity of paclitaxel, patients with SD weredivided into two groups according to the electronic med-ical records: shrunken SD (tumour reduction <30%) andenlarged SD (tumour increase <20%). Patients were alsodivided into two groups according to the extent of ben-efit that they gleaned from chemotherapy: clinical bene-fit group (including patients with CR, PR, and shrunkenSD) and non-clinical benefit group (including patientswith PD and enlarged SD). Clinical benefit rate (CBR)was calculated as follows: (CR + PR + shrunken SD)/total patients. Progression-free survival (PFS) andoverall survival (OS) were calculated from the first dayof chemotherapy to disease progression and death fromany cause, respectively.

2.3. Immunohistochemistry staining of MAP-tau and

TUBB3

FFPE tumour sections with 4 lm thick were deparaf-finised in xylene and hydrated in graded alcohols,followed by retrieval in 0.01 M citrate buffer (pH 6.0)

2330 J. Yu et al. / European Journal of Cancer 50 (2014) 2328–2335

and endogenous peroxidase treatment. Sections werethen incubated with MAP-tau (dilution: 1:50; SantaCruz Biotechnology, Santa Cruz, CA, United States)and TUBB3 (dilution: 1:500; Abcam, United States)monoclonal antibody for 60 min, respectively. Signalproduction involved general type IgG-HRP Polymer(Beijing CoWin Biotech Co., Ltd.) and diaminobenzi-dine substrate. Each experiment included a negativecontrol. Sections were scored using two independentprofessional pathologists from the pathology depart-ment of the present hospital without any knowledge ofthis study. Staining was graded as in a previous study[6,18]. Briefly, intensity of staining (1, weak; 2, moder-ate; 3, strong) and relative numbers of cells stained werecalculated (1: 0–10%; 2: 11–50%; 3: 51–100%). Theexpression levels were considered high or low based onthe median staining score (intensity score plus percent-age score).

2.4. Gastric cancer cell lines

Seven GC cell lines (AGS, BGC-823, HGC-27,MGC-803, MKN45, NCI-N87 and SGC-7901) wereprovided by professor You-yong LV (Peking UniversityCancer Hospital and Institute). Cells were cultured inRPMI 1640 medium (Gibco BRL, Rockville, MD,United States) supplemented with 10% foetal bovineserum (Gibco BRL) and incubated in a humidified37 �C incubator supplemented with 5% CO2.

2.5. Plasmid construction and screening of stable

transfected cells

Human MAP-tau expression plasmid pEGFP-tau wasprovided by Dr. Kenneth Kosik (University of Califor-nia, Santa Barbara, CA, United States). Cells were trans-fected with pEGFP-tau or control vector pEGFP-C1using lipofectamine 2000 (Invitrogen). After 48 h of trans-fection, cells were selected in medium containing 400 lg/mL G418 for about 6 weeks. Cells with stable expressionof MAP-tau were used in subsequent experiments.

2.6. Lentivirus package and infection

Lentiviral shRNA of TUBB3 with a titre of1 � 109 TU/mL was purchased from Shanghai Gene-Pharma Co., Ltd, China. Before infection, cell mediumwas replaced with fresh complete medium and then len-tivirus was added. After 48 h of incubation, cells wereharvested and sorted by flow cytometry for the contin-uing study.

2.7. Western blot

Total protein was extracted from cell pellets usingCytoBuster Protein Extraction Reagent (Merck

Millipore, Darmstadt, Germany). Protein concentrationwas measured using the DC protein assay method pub-lished by Bradford (Bio-Rad, Hercules, CA, UnitedStates), and about 20 micrograms of protein from eachsample was separated on 12% SDS-PAGE. The nitrocel-lulose membrane (GE Healthcare, Piscataway, NJ,United States) was incubated with primary antibody at4 �C overnight after transfer and secondary antibodyat room temperature for 1–2 h. Proteins were visualisedusing ECL Plus Western Blotting Detection Reagents(GE Healthcare).

2.8. Cell viability assay

Cells were seeded in 96-well plates (5000/well), andpaclitaxel was added to cells the next day. Afterincubation for 36–48 h, cell viability was analysed usingCellTiter 96� AQueous One Solution Cell ProliferationAssay (MTS assay; Promega, Madison, WI, UnitedStates) according to the manufacturer’s instructions.The absorbance was measured at 490 nm using aspectrophotometer.

2.9. Xenograft model in nude mice

BGC-823 cells (1 � 106) with stable expression ofMAP-tau or downregulation of TUBB3 were suspendedin 0.1 mL phosphate buffered saline and injected subcu-taneously into the right oxter of 6-week-old femaleBALB/c athymic nu/nu mice (5 mice/group; Vital River,China). When tumour volumes reached about 50 mm3,mice were injected intraperitoneally with paclitaxel(10 mg/kg body weight) twice a week. Tumuors weremeasured twice every week and tumour volume was cal-culated as follows: V = L �W2 � 1/2 (V, volume; L,length; W, width of tumour). All animal experimentswere performed in accordance with the animal experi-mental guidelines of Peking University Cancer Hospital.

2.10. Calculation of sample size

According to previous results, the sample size wascalculated based on the following hypothesis: The rela-tive numbers of patients with low levels of MAP-tauand TUBB3 expressions were about 25% and the CBRwas about 60%. The relative numbers of patients withhigh levels of MAP-tau and TUBB3 expressions wereabout 25%, and the CBR was about 30%. The calculatedsample size was 82 (41 per group) if power = 0.8,a = 0.05, R = 1:1, P1 = 0.6, and P2 = 0.3.

2.11. Statistical analysis

SPSS 18.0 software was used to perform statisticalanalysis. The chi-square test was used to analyse therelationships between clinicopathological characteristics

J. Yu et al. / European Journal of Cancer 50 (2014) 2328–2335 2331

and levels of MAP-tau/TUBB3 expression. TheKaplan–Meier survival curve and log-rank test wereused to describe progression-free survival and overallsurvival. Univariate and multivariate Cox regressionmodels were used to analyse prognostic factors. The dif-ferences in MAP-tau or TUBB3 expression betweenpatients with different characteristics were determinedusing the Mann–Whitney U test. Repeated-measuredanalysis of variance was used to evaluate the differencesin tumour growth between two groups of mice. P < 0.05was considered statistically significant.

3. Results

3.1. Patient characteristics

From December 2006 to December 2010, a total of244 advanced gastric cancer patients were included inthis study with 149 patients receiving paclitaxel pluscapecitabine (cohort 1) and 95 patients receiving

Table 1Patient demographics and clinical characteristics.

Characteristic Cohort 1 (n =No. of patients

GenderMale 98 (65.8%)Female 51 (34.2%)

Age (years)Median (range) 59 (19–79)

LocalisationCardia 38 (25.5%)Fundus 9 (6.0%)Antrum 56 (37.6%)

Gastric body 46 (30.9%)Histological differentiation

Good–moderate 36 (24.1%)Poor 95 (63.8%)Othera 18 (12.1%)

KPSb

P80 144 (96.6%)<80 5 (3.4%)

Metastasis sitesLymph node 113 (75.8%)Peritoneum 26 (17.4%)Liver 46 (30.9%)Ovary 19 (12.8%)Otherc 15 (10.1%)

ResponseComplete response (CR) 2 (1.4%)Partial response (PR) 67 (46.5%)Shrunken stable disease (SD) 15 (10.4%)Enlarged SD 42 (29.2%)Progressive disease (PD) 18 (12.5%)NA 5

Median progression-free survival (PFS) (days) 20295% confidence interval (CI) 169.60–234.40

Median overall survival (OS) (days) 35395% CI 310.82–195.18

NA, not available.a Including mucinous, signet-ring cell.b Karnofsky Performance Status.c Including lung, adrenal glands, pelvic cavity, bone and thoracic cavity.

cisplatin plus capecitabine (cohort 2). The characteris-tics of 244 patients are presented in Table 1. There wereno significant differences in age, histological differentia-tion, Karnofsky performance status (KPS) score ormetastasis sites between the two cohorts. The lastfollow-up was March 2012. No difference in medianPFS (202 versus 170 days) or OS (353 versus 339 days)was observed between cohort 1 and cohort 2.

3.2. MAP-tau expression and clinical outcome

Among 244 patients, 135 (55.3%) showed high levelsof MAP-tau expression (Fig. 1A). There was no signifi-cant correlation between MAP-tau expression andpatient characteristics (Supplementary Table 1). Incohort 1, the CBR of patients with high levels ofMAP-tau expression was poorer than in patients withlow levels of MAP-tau expression (49.4% versus69.2%, P = 0.016). Neither median PFS (P = 0.121)nor median OS (P = 0.105) was found to be associated

149) Cohort 2 (n = 95) Total (n = 244)(%) No. of patients (%) No. of patients (%)

75 (78.9%) 173 (70.9%)20 (21.1%) 71 (29.1%)

57 (28–79) 58 (19–79)

35 (36.8%) 73 (29.9%)4 (4.2%) 13 (5.3%)26 (27.4%) 82 (33.6%)30 (31.6%) 76 (31.1%)

22 (23.2%) 58 (23.8%)66 (69.5%) 161 (66.0%)7 (7.4%) 25 (10.3%)

93 (97.9%) 237 (97.1%)2 (2.1%) 7 (2.9%)

70 (73.7%) 183 (75.0%)8 (8.4%) 34 (13.9%)39 (41.1%) 85 (34.8%)11 (11.6%) 30 (12.3%)9 (9.5%) 27 (11.1%)

1 (1.1%) 3 (1.3%)37 (40.2%) 104 (44.1%)4 (4.3%) 19(8.1%)31 (33.7%) 73 (30.9%)19 (20.7%) 37 (15.7%)3 8170 191143.29–196.71 (169.39–212.61)339 349234.17–443.83 (303.57–394.43)

Table 2Associations between microtubule associated protein-tau (MAP-tau) and b-tubulin III expressions and clinical outcomes.

Cohort 1 (n = 144) Cohort 2 (n = 92)

Low High Low High

MAP-tau expression

Clinical benefit rate (CBR) 69.2% 49.4% 40.5% 50.0%P value 0.016 0.316Median progression-free survival (PFS) (days) 240 170 146 20095% confidence interval (CI) 214.7–265.3 132.8–207.2 99.5–192.5 149.3–250.8P value 0.121 0.071Median overall survival (OS) (days) 374 319 259 39295% CI 319.7–428.4 192.9–445.1 108.7–409.3 280.6–503.4P value 0.105 0.552

TUBB3 expression

CBR 67.1% 48.5% 43.5% 47.8%P value 0.024 0.675Median PFS (days) 248 151 200 16295% CI 216.6–279.4 108.9–193.1 152.3–247.7 116.0–208.0P value 0.002 0.063Median OS (days) 385 273 327 28795% CI 317.3–452.7 203.6–342.4 344.3–509.7 234.2–443.8P value 0.030 0.079

Table 3Associations between microtubule associated protein-tau (MAP-tau) and b-tubulin III expressions and clinical outcomes.

Expressions CBR Median progression-freesurvival (PFS) (days)

Median overallsurvival (OS) (days)

Cohort 1

Low MAP-tau and low TUBB3 (n = 36) 72.2% 271 394High MAP-tau and high TUBB3 (n = 39) 35.9% 102 173P value 0.008 <0.001 0.012

Cohort 2

Low MAP-tau and low TUBB3 (n = 23) 39.1% 156 345High MAP-tau and high TUBB3 (n = 27) 51.9% 169 295P value >0.05 >0.05 >0.05

2332 J. Yu et al. / European Journal of Cancer 50 (2014) 2328–2335

with MAP-tau expression (Table 2). In cohort 2, neitherCBR nor survival was associated with MAP-tau expres-sion (Table 2).

3.3. TUBB3 expression and clinical outcome

A total of 117 patients (48.0%) in this study showedhigh levels of TUBB3 expression (Fig. 1A). There wasno significant correlation between the expression ofTUBB3 and patient characteristics (SupplementaryTable 1). In cohort 1, TUBB3 expression was inverselycorrelated with the CBR of patients (67.1% versus48.5%, P = 0.024). Patients with high levels of TUBB3expression had shorter PFS (P = 0.002) and OS(P = 0.030) than patients with low levels of TUBB3expression (Table 2). In cohort 2, no significant differ-ence in CBR or survival was found between patientswith high and low levels of TUBB3 expression (Table 2).

3.4. MAP-tau and TUBB3 expressions and clinical

outcome

The results given above show, patients with low levelsof MAP-tau or TUBB3 expression in cohort 1 were

found to benefit more from paclitaxel plus capecitabinetherapy. MAP-tau and TUBB3 were then analysedtogether. In cohort 1, the CBR (72.2% versus 35.9%,P = 0.008), median PFS (271 versus 102 days, P <0.001) and OS (394 versus 173 days, P = 0.012) inpatients with low levels of MAP-tau and TUBB3 expres-sions were higher than those in patients with high levels(Table 3, Fig. 1B). However, in cohort 2, no significantdifferences in CBR (39.1% versus 51.9%, P > 0.05),median PFS (156 versus 169 days, P > 0.05) or OS(345 versus 295 days, P > 0.05) were observed betweenpatients with different levels of expression of MAP-tauand TUBB3 (Table 3, Fig. 1C). Multivariate analysis,showed that high levels of MAP-tau and TUBB3 expres-sions were an independent factor for poor prognosis incohort 1 (P = 0.009, HR = 1.995, 95% confidenceinterval (CI) = 1.184–3.360).

3.5. MAP-tau and TUBB3 expressions and sensitivity to

paclitaxel in vitro gastric cancer cells

The results given above suggest that MAP-tau orTUBB3 expression might predict the sensitivity ofpaclitaxel. The expressions of MAP-tau and TUBB3

J. Yu et al. / European Journal of Cancer 50 (2014) 2328–2335 2333

were detected in 7 types of gastric cancer cells by Wes-tern blot. Five of these seven sets had no or very lowexpression of MAP-tau, but TUBB3 was expressed inall of them (Fig. 2A). Ectopic expression of MAP-tauin BGC-823 cells was found to induce resistance to pac-litaxel (Fig. 2B). Downregulation of TUBB3 increasedsensitivity to paclitaxel in BGC-823 cells (Fig. 2C). Nei-ther MAP-tau nor TUBB3 expression had any associa-tion with the sensitivity to cisplatin in BGC-823 cells(Supplementary Fig. S1).

3.6. Ectopic expression of MAP-tau and efficacy of

paclitaxel in in vivo xenografts

Mice harbouring BGC-823 xenografts that stablyexpressed of MAP-tau or control were treated with pac-litaxel twice a week for 3 weeks. Ectopic expression ofMAP-tau was associated with significantly less paclit-axel efficacy than in the control group (P = 0.003,Fig. 2D).

Fig. 1. Combination of microtubule associated protein-tau (MAP-tau)and b-tubulin III (TUBB3) and prediction of chemosensitivity topaclitaxel. (A) Representative staining of MAP-tau and TUBB3expression in gastric cancer tissues (magnification 100�). (B) Progres-sion-free survival (PFS) and overall survival (OS) curves of patientswith different MAP-tau and TUBB3 expressions. Patients with low

3.7. Downregulation of TUBB3 and sensitivity to

paclitaxel in in vivo xenografts

BGC-823 cells with downregulation of TUBB3 orcontrol were transplanted into the mice, and tumourvolume was measured at the time of paclitaxel injection.The growth of xenografts with downregulation ofTUBB3 was significantly slower than in the controlgroup (P = 0.031), which suggested that downregulationof TUBB3 could increase sensitivity to paclitaxel(Fig. 2E).

levels of MAP-tau and TUBB3 expression were found to benefit morefrom paclitaxel plus capecitabine than other patients. (C) Amongpatients receiving cisplatin plus capecitabine, no significant differencesin PFS or OS were found between patients with low and high levels ofMAP-tau and TUBB3 expressions (P > 0.05).

4. Discussion

In the present study, the correlations between MAP-tau and TUBB3 expression and clinical outcomes weresystematically analysed in advanced gastric cancerpatients receiving paclitaxel plus capecitabine (cohort1) or cisplatin plus capecitabine (cohort 2). Results dem-onstrated that both MAP-tau and TUBB3 could predictpaclitaxel sensitivity (Table 2). As in the present find-ings, MAP-tau and TUBB3 have been reported to beassociated with paclitaxel sensitivity in other cancers,including non-small cell lung cancer, ovarian cancerand breast cancer [10–12,14,15]. After the two markerswere combined together, the clinical outcome was signif-icantly better in patients with low levels of expression oftwo markers than in patients with high levels in cohort 1(Table 3, Fig. 1). This may provide a foundation forprospective clinical trials.

From Tables 2 and 3, we found that the CBR, PFSand OS with a combined analysis of markers(72.2%, 271 days and 394 days) were similar with those(67.1%, 248 days and 385 days) of TUBB3 alone.Moreover, from Table 2, no significant correlation was

found between MAP-tau expression and survival, whileTUBB3 expression was related with survival. This sug-gested that TUBB3 might play a prior role to determinethe sensitivity of paclitaxel, and the most possible reasonmight be TUBB3 was the direct target of paclitaxel.

As shown in Table 2, in cohort 2, no difference wasobserved between TUBB3 expression and response,however, the median PFS and OS in patients with lowlevels of TUBB3 expression were slightly higher thanthose in patients with high levels, although the differencewas not significant. It is here reported that TUBB3might play an important role in tumourigenesis,progression, chemoresistance and prognosis throughsignal transduction factors and pathways, such ashypoxia-induced factor 1a, p53, nuclear factor-jB orthe PI3K/Akt pathway [10,20–22]. As in the presentresults, TUBB3 may have functioned as a prognosticmarker in patients with non-small cell lung cancer or

Fig. 2. Microtubule associated protein-tau (MAP-tau) and b-tubulin III (TUBB3) expressions and correlations with sensitivity to paclitaxel in vitro

and in vivo. (A) MAP-tau was silent in 5 of 7 sets of gastric cancer cells, but TUBB3 was expressed in all 7 cell lines. (B) Ectopic expression of MAP-tau was found to attenuate sensitivity to paclitaxel in BGC-823 cells. *P < 0.05 (n = 3, mean ± SD). (C) Downregulation of TUBB3 was found toincrease sensitivity to paclitaxel in BGC-823 cells. *P < 0.05 (n = 3, mean ± SD). (D) After the mice were killed, tumours were tested for expressionof MAP-tau using immunohistochemistry. The growth of xenografts with stable expression of MAP-tau was faster than in the control group whenall cells received paclitaxel injection. Arrows indicate the time of paclitaxel injection (5 mice/group, mean ± SD). (E) TUBB3 expression wasverified in two groups and downregulation of TUBB3 was found to increase the efficacy of paclitaxel. Arrows indicated the time of paclitaxelinjection (5 mice/group, mean ± SD).

2334 J. Yu et al. / European Journal of Cancer 50 (2014) 2328–2335

ovarian cancer [11,23–26]. Currently, the possible mech-anisms under the prognostic function of TUBB3 remainunclear and need to be investigated further. Also, theprognostic significance of TUBB3 in advanced gastriccancer would be studied in future large samples.

Some studies have reported high levels of MAP-tauexpression to be associated with significantly decreasedrisk of disease relapse or death and a longer PFS orOS in breast cancer patients [27,28]. The correlationbetween MAP-tau expression and prognosis was exam-ined, and no difference was found between MAP-tauexpression and patient prognosis (data not shown).

Based on in vitro and in vivo assays, MAP-tau wassilent or downregulated in most gastric cancer cellsand TUBB3 was expressed in all gastric cancer cells(Fig. 2). BGC-823 cells without MAP-tau expressionbut with TUBB3 expression were selected for study.After ectopic expression of MAP-tau or downregulationof TUBB3, the growth and morphology of BGC-823

cells were not affected (data not shown). This was con-sistent with other reports [29]. Results from cells andxenografts validated that the expression of MAP-tauand TUBB3 can predict sensitivity to paclitaxel.

Whether the expression of MAP-tau and TUBB3 canchange after paclitaxel therapy begins or as the diseaseprogresses remains to be determined. Here, severalmatched tumour tissues were collected, and theexpression of MAP-tau and TUBB3 was found to besignificantly higher in tumours that progressed than ithad been when treatment began (data not shown). Thisstudy was continued and the results were promising.

5. Conclusions

The combination of MAP-tau and TUBB3 was foundto predict chemosensitivity to paclitaxel in gastric cancerin vitro and in vivo. Future studies should be conductedto validate this result and to guide individual therapy.

J. Yu et al. / European Journal of Cancer 50 (2014) 2328–2335 2335

Funding

This work was supported by Beijing MunicipalScience & Technology Commission Program(No. Z11110706730000), National Natural ScienceFoundation of China (No. 81172110), and NationalHigh Technology Research and Development Program(No. 2012AA 02A 504).

Conflict of interest statement

None declared.

Acknowledgements

The authors would like to thank Doctor Zhong-huHe (Laboratory of Genetics, Peking University CancerHospital and Institute) for calculations of sample size cal-culation and Professor You-yong Lv (Peking UniversityCancer Hospital and Institute) for the critical reading ofthis manuscript. We thank LetPub (www.letpub.com)for its linguistic assistance during the preparation of thismanuscript.

Appendix A. Supplementary data

Supplementary data associated with this article canbe found, in the online version, at http://dx.doi.org/10.1016/j.ejca.2014.06.017.

References

[1] Van Cutsem E. The treatment of advanced gastric cancer: newfindings on the activity of the taxanes. Oncologist 2004;9:9–15.

[2] Orr GA, Verdier-Pinard P, McDaid H, Horwitz SB. Mechanismsof taxol resistance related to microtubules. Oncogene2003;22:7280–95.

[3] McGrogan BT, Gilmartin B, Carney DN, McCann A. Taxanes,microtubules and chemoresistant breast cancer. Biochim BiophysActa 2008;1785:96–132.

[4] Cortes J, Baselga J. Targeting the microtubules in breast cancerbeyond taxanes: the epothilones. Oncologist 2007;12:271–80.

[5] Kavallaris M. Microtubules and resistance to tubulin-bindingagents. Nat Rev Cancer 2010;10:194–204.

[6] Seve P, Mackey J, Isaac S, et al. Class III b-tubulin expression intumor cells predicts response and outcome in patients with non-small cell lung cancer receiving paclitaxel. Mol Cancer Ther2005;4:2001–7.

[7] Kavallaris M, Kuo DYS, Burkhart CA, et al. Taxol-resistantepithelial ovarian tumors are associated with altered expression ofspecific b-tubulin isotypes. J Clin Invest 1997;100:1282–93.

[8] Ranganathan S, Benetatos CA, Colarusso PJ, Dexter DW, HudesGR. Altered b-tubulin isotype expression in paclitaxel-resistanthuman prostate carcinoma cells. Br J Cancer 1998;77:562–6.

[9] Liu B, Staren ED, Iwamura T, Appert HE, Howard JM.Mechanisms of taxotere-related drug resistance in pancreaticcarcinoma. J Surg Res 2001;99:179–86.

[10] Azuma K, Sasada T, Kawahara A, et al. Expression of ERCC1and class III beta-tubulin in non-small lung cancer patientstreated with carboplatin and paclitaxel. Lung Cancer2009;64:326–33.

[11] Ferrandina G, Zannoni GF, Martinelli E, et al. Class III beta-tubulin overexpression is a marker of poor clinical outcome inadvanced ovarian cancer patients. Clin Cancer Res2006;12:2774–9.

[12] Paradiso A, Mangia A, Chiriatti A, et al. Biomarkers predictivefor clinical efficacy of taxol-based chemotherapy in advancedbreast cancer. Ann Oncol 2005;16:14–9.

[13] Hwang JE, Hong JY, Kim K, et al. Class III b-tubulin is apredictive marker for taxane-based chemotherapy in recurrentand metastatic gastric cancer. BMC Cancer 2013;13:431–8.

[14] Rouzier R, Rajan R, Wagner P, et al. Microtubule-associatedprotein tau: a marker of paclitaxel sensitivity in breast cancer.Proc Natl Acad Sci USA 2005;102:8315–20.

[15] Hess KR, Anderson K, Symmans WF, et al. Pharmacogenomicpredictor of sensitivity to preoperative chemotherapy with pac-litaxel and fluorouracil, doxorubicin, and cyclophosphamide inbreast cancer. J Clin Oncol 2006;24:4236–44.

[16] Wu H, Huang M, Lu M, et al. Regulation of microtubule-associated protein tau (MAPT) by miR-34c-5p determines thechemosensitivity of gastric cancer to paclitaxel. Cancer Chemo-ther Pharmacol 2013;71:1159–71.

[17] Mimori K, Sadanaga N, Yoshikawa Y, et al. Reduced tauexpression in gastric cancer can identify candidates for successfulpaclitaxel treatment. Br J Cancer 2006;94:1894–7.

[18] Pusztai L, Jeong JH, Gong Y, et al. Evaluation of microtubule-associated protein-Tau expression as a prognostic and predictivemarker in the NSABP-B 28 randomized clinical trial. J Clin Oncol2009;27:4287–92.

[19] Sakamoto J, Matsui T, Kodera Y. Paclitaxel chemotherapy forthe treatment of gastric cancer. Gastric Cancer 2009;12:69–78.

[20] Gong J, Hu B, Zhang X, et al. The multicenter, phase IIprospective study of paclitaxel plus capecitabine as first-linechemotherapy in advanced gastric carcinoma. Oncologist 2014[Epub ahead of print].

[21] Mariani M, Shahabi S, Sieber S, Scambia G, Ferlini C. Class IIIbeta-tubulin (TUBB3): more than a biomarker in solid tumors?Curr Mol Med 2011;11:726–31.

[22] Yu HG, Ai YW, Yu LL, et al. Phosphoinositide 3-kinase/Aktpathway plays an important role in chemoresistance of gastriccancer cells against etoposide and doxorubicin induced cell death.Int J Cancer 2008;122:433–43.

[23] Dennis K, Uittenbogaard M, Chiaramello A, Moody SA.Cloning and characterization of the 50-flanking region of the ratneuron-specific class III beta-tubulin gene. Gene 2002;294:269–77.

[24] Levallet G, Bergot E, Antoine M, et al. High TUBB3 expression,an independent prognostic marker in patients with early non-small cell lung cancer treated by preoperative chemotherapy, isregulated by K-Ras signaling pathway. Mol Cancer Ther2012;11:1203–13.

[25] Seve P, Dumontet C. Is class III beta-tubulin a predictive factor inpatients receiving tubulin-binding agents? Lancet Oncol2008;9:168–75.

[26] Koh Y, Jang B, Han SW, et al. Expression of class III beta-tubulin correlates with unfavorable survival outcome in patientswith resected non-small cell lung cancer. J Thorac Oncol2010;5:320–5.

[27] Pentheroudakis G, Kalogeras KT, Wirtz RM, et al. Geneexpression of estrogen receptor, progesterone receptor andmicrotubule-associated protein Tau in high-risk early breastcancer: a quest for molecular predictors of treatment benefit inthe context of a Hellenic Cooperative Oncology Group trial.Breast Cancer Res Treat 2009;116:131–43.

[28] Baquero MT, Lostritto K, Gustavson MD, et al. Evaluation ofprognostic and predictive value of microtubule associated proteintau in two independent cohorts. Breast Cancer Res 2011;13:R85.

[29] Kamath K, Wilson L, Cabral F, Jordan MA. Beta III-tubulininduces paclitaxel resistance in association with reduced effects onmicrotubule dynamic instability. J Biol Chem 2005;280:12902–7.