RESEARCH ARTICLE

Effects of CXCR4 gene silencing by lentivirus shRNAon proliferation of the EC9706 human esophagealcarcinoma cell line

Dao-feng Wang & Ning Lou & Miao-zhen Qiu &

Yong-bin Lin & Ying Liang

Received: 28 January 2013 /Accepted: 10 May 2013# International Society of Oncology and BioMarkers (ISOBM) 2013

Abstract CXCL12/CXCR4 has been studied as an importantbiomarker for many human malignancies, but studies arelimited for esophageal squamous cell carcinoma (ESCC). Inthis study, an effective RNAi sequence targeting the CXCR4gene was selected, a lentiviral shRNAvector was constructedto specifically silence CXCR4 expression in the EC9706ESCC cell line, and the effects of CXCR4 silencing on cellgrowth in vitro and tumour growth in nude mice were thenevaluated. The expression of CXCR4 in EC9706 was signif-icantly downregulated after transfection with a lentiviralshRNA vector. The expression of the apoptosis-related geneBcl-2 was decreased. In addition, after CXCR4 inhibition, cellgrowth was considerably inhibited, increased apoptosis in theEC9706 cells was found, the G0/G1 percentage was signifi-cantly increased, and the number of cells in S phase wasreduced. Moreover, tumour growth in nude mice wasinhibited. In conclusion, the downregulation of CXCR4 ex-pression by transfection with a lentiviral shRNA vector inESCC cells could inhibit tumour proliferation. Our data mayprovide an avenue for finding new ESCC treatments.

Keywords CXCL12 . CXCR4 . Esophageal squamous cellcarcinoma . RNA interference . Proliferation

Introduction

Esophageal squamous cell carcinoma (ESCC) is the mostcommon malignant gastro-enteric tumour and has a higherincidence than esophageal adenocarcinoma in China [1].With characteristics including rapid progression in a shortperiod of time, ESCC cells can proliferate in early stages,then invade the submucosa and muscular layers, and finallyinvade the peripheral tissue of the oesophagus, thusresulting in difficulties during radical resection. The prog-nosis of ESCC remains poor, although chemotherapeuticregimens are currently being developed.

Chemokines are a family of cytokines that regulate che-motaxis. CXCL12 is secreted by stromal cells, and a con-stant low level expression of CXCL12 is found in normaltissues. CXCR4 is the specific receptor for CXCL12, and itis widely expressed in the immune system and central neu-rons [2, 3]. It has been reported that CXCL12 and CXCR4are highly expressed in many types of malignancies, includ-ing non-small cell lung cancer [4], gastric cancer [5], andoesophageal carcinoma [6]. It was found that CXCR4 andCXCL12 were located in the cytoplasm and membrane [6].CXCL12/CXCR4 may also activate many signalling path-ways in tumour cells including mitogen-activated proteinkinases and phosphatidylinositol 3-kinase [7]. Thus, theCXCL12 and CXCR4 signalling pathway may serve as apotential therapy target.

Recently, it has been suggested that the CXCL12/CXCR4signalling pathway plays an important role in carcinogenesisand progression [8]. Inhibition of this pathway may improveprognosis. Currently, the CXCR4 inhibitors AMD3100 andTN14003 have been shown to be promising prospects

Dao-Feng Wang and Ning Lou contributed equally to this manuscript.

D.<f. Wang :N. Lou :M.<z. Qiu :Y.<b. Lin :Y. LiangState Key Laboratory of Oncology in South China, Guangzhou,People’s Republic of China

D.<f. Wang (*) :N. LouDepartment of ICU, Sun Yat-Sen University Cancer Center,651 Dong Feng Road East,Guangzhou 510060, People’s Republic of Chinae-mail: wangdf@sysucc.org.cn

M.<z. Qiu :Y. LiangDepartment of Medical Oncology, Sun Yat-Sen University CancerCenter, Guangzhou, People’s Republic of China

Y.<b. LinDepartment of Thoracic Surgery, Sun Yat-Sen University CancerCenter, Guangzhou, People’s Republic of China

Tumor Biol.DOI 10.1007/s13277-013-0858-0

experimentally [9]. However, the specificity of these inhibi-tors is poor. RNA interference (RNAi) has been widely usedas an effective tool for gene research. RNAi could reduce theexpression of target proteins through inhibition of their geneexpression at the mRNA level by double-stranded RNA. Incontrast with CXCR4 inhibitors, gene expression inhibitionby RNAi is highly specific. In this study, we used a lentivirus-based vector to transfect a CXCR4-specific short hairpin RNA(shRNA) into ESCC cells, and as a result, CXCR4 expressionwas stably inhibited by RNAi. In addition, we investigated thebiological effects of CXCR4 inhibition on ESCC.

Materials and methods

Cell culture

Human ESCC cell line EC9706 was obtained from thelaboratory of Sun Yat-sen University Cancer Center(Guangzhou, China) and maintained in DMEM medium(Gibco Biocult, Paisley, UK) supplemented with 10 % calfbovine serum (Sijiqing Biotechnology, Hangzhou, China),100 μg/mL streptomycin, and 100 U/mL penicillin (NorthChina Pharmaceutical Group Corporation, Shijiazhuang,China) at 37 °C in a humidified atmosphere of 5 % CO2 inair. When the EC9706 cells grew to 70 % confluence, theculture medium was removed and replaced with free-serummedium for 24 h. 293T cells were obtained from ShanghaiGeneChem Co. Ltd. and maintained in DMEM medium(Gibco Biocult, Paisley, UK) supplemented with 10 % calfbovine serum (Sijiqing Biotechnology, Hangzhou, China).

Construction of plasmid targeting the CXCR4 mRNA

Based on principles of shRNA design and the humanCXCR4 structure (NM_001008540.1), the preparation, syn-thesis, and a preliminary experiment for screening for apositive oligonucleotide fragment targeting CXCR4mRNA were conducted. Briefly, three plasmid vectorsencoding shRNA directed against CXCR4 mRNA wereconstructed. Of these vectors, one resulted in over 75 %CXCR4 inhibition. The shRNA targeting sequence was 5′-TTCAAGAGA-3′. The 5′-TTCTCCGAACGTGTCACGT-3′ scrambled sequence was used as a negative control thatdoes not target any known human mRNA. shRNA wassynthesised and cloned into the pGCL-GFP vector viadouble-enzyme digestion with Age I (New EnglandBiolabs, Inc. USA) and EcoR I (New England Biolabs,Inc. USA) followed by T4 ligation. The vector was thentransformed into DH5α competent cells (TaKaRaBiotechnology Co., Ltd. Japan) for AMP-resistant plasmidscreening. Positive plasmids were selected for clonalexpansion.

Transfection and detection of the transfection rate

Three groups were set up for the cell transfection: thecontrol group without the transfection, the NC-GFP-RNAi-LV group with negative sequence transfection, and theCXCR4-RNAi-LV group with lentivirus target sequenceinterference. The CXCR4-specific shRNA-encoding pGC-LV vector, pHelper 1.0 vector, and pHelp2.0 vector (Fig. 1)(Genechem, Shanghai, China) were cotransfected into the293T cell line using lipofectamine 2000 (Invitrogen,Carlsbad, CA) to produce lentivirus stocks with a negativeconstruct as a negative control. Forty-eight hours after trans-fection, the 293T cells were harvested, and the cell debriswas removed by centrifugation at 4 °C. The crude viralextract was filtered, and centrifuge enrichment wasconducted. The virus concentration was determined usinga single/uniporous tracer dilution. A total of 2 μL of humanCXCR4-RNAI-LV with a titre of 2E+09T U/ml was addedto the EC9706 cell suspension and cultured for 72 h. Theexpression of the GFP lentivirus reporter gene was observedby fluorescence microscopy. Five visual fields (200×) werechosen to count the number of the cells presenting withgreen fluorescence, which were considered to represent theproportion of the total cells in the field.

RT-PCR assay

RNAs from different groups were extracted with the Trizolreagent (Invitrogen, Carlsbad, CA, USA) and treated withDNase (Tiangen Biotech, Beijing, China). cDNA wassynthesised from 2 μg of total RNA according to the man-ufacturer’s instructions in a total volume of 25 μL(Fermentas, MD, USA). Negative control reactions wererun without reverse transcriptase. An equal volume of prod-uct was subjected to PCR. The primer sequences werespecific to an updated version of the GenBank sequences.Sequences of primers used to detect human CXCR4(NM_001008540.1) [10], Bcl-2 (NM_000633.2) [11], andβ-actin (NM_001101.3; internal control); the sequences ofprimers were shown in Table 1. The amplification condi-tions were as follows: 40 cycles of 95 °C for 5 min, 95 °Cfor 15 s, 60 °C for 1 min for CXCR4 and β-actin, and30 cycles of 94 °C for 30 s, 57.5 °C for 30 s, 72 °C for1 min for Bcl-2. Images of the RT-PCR ethidium bromide-stained agarose gels were acquired with a Cohu High-Performance CCD camera (Cohu Inc. San Diego, CA),and quantification of the bands was performed by ImageTool 3.0 (The University of Texas Health Science Center,San Antonio, TX, USA). Band intensity was expressed asrelative absorbance units. The ratio between the sampleRNA to be determined and β-actin was calculated to nor-malize for initial variations in sample concentration and as acontrol for reaction efficiency. Mean and standard deviation

Tumor Biol.

of all experiments performed were calculated after normal-ization to β-actin.

Western blotting

Seventy-two hours after transfection, the cells in each groupwere lysed in 400 μL of ice-cold buffer A. (1 mMdithiothreitol, 0.5 mM phenylmethylsulfonyl fluoride) bygentle pipetting. The cells were allowed to swell on ice for25 min, 40 μL of a 10 % solution of Nonidet P-40 (Amresco,OH, USA) was then added, and the tube was vigorouslyvortexed for 10 s. The homogenate was centrifuged for45 s in a microcentrifuge. The supernatant which containsthe cytoplasmic fraction was transferred for further valida-tion. Equal amounts of protein, which was quantified bybicinchoninic acid protein assay kit (Pierce Biotechnology,Rockford, IL, USA), were subjected to 10 % SDS-polyacrylamide gel electrophoresis and transferred topolyvinylidene difluoride membranes. The membranes wereblocked with 5 % nonfat milk in Tris-buffered saline withTween-20 [TBST; 50 mM Tris–HCl (pH 7.6),150 mMNaCl,0.1 % Tween-20] for 2 h at room temperature and incubatedwith primary antibody (anti-CXCR4, 1:1,000, Abcam,Cambridge, UK; anti-CXCL12, 1:1,000, R&D system,USA; anti bcl-2, 1:2,000, Abcam, Cambridge, UK; and β-actin, 1:2,000, Santa Cruz Biotechnology, Santa Cruz, CA,USA) in blocking buffer at 4 °C overnight. Following a washwith TBST, the CXCR4,CXCL12, and Bcl-2 membraneswere incubated with a horseradish peroxidase-conjugatedgoat anti-rabbit secondary antibody (1:4,000, Dako,

Glostrup, Denmark) for 3 h, and the β-actin membrane wasincubated with a horseradish peroxidase-conjugated rabbitanti-mouse secondary antibody (1:1,000, Dako, Glostrup,Denmark) for 3 h. The membranes were washed withTBST, and protein bands were visualized by enhancedchemiluminescence according to the manufacturer’s instruc-tions (KPL, Gaithersburg, USA). The β-actin bands wereused as a loading control. The protein quantity was analysedusing the UTHSCSA Image Tool 3.0. The formula for therelative expression of CXCR4 was as follows: CXCR4 pro-tein relative intensity ratio/loading control.

Cell proliferation assay

An EC9706 cell proliferation assay was used to evaluate cellproliferation according to the manufacturer's instructions.The ESCC cells from the three groups were seeded in 96-well plates at 1×104 cells per well and cultured for 24, 48,72, 96, 120, and 144 h. A total of 20 μL of MTT (Sigma, St.Louis, MO, USA) stock solution was added to 200 μL ofmedium in each well, the plates were incubated for 5 h at37 °C, and 150 μL of dimethyl sulfoxide was added to eachwell, the plates were incubated for 15 min at room temper-ature, and the absorbance values were read by an enzyme-linked immunosorbent assay (490 nm).

Flow cytometry assay

Three cell groups (1.0×106) were harvested bytrypsinisation, centrifuged, washed three times with ice-

Table 1 Primer sequences used in this study

Gene name Forward primer Reverse primer PCR (frag) bp

CXCR4 5′-ATGTAGACACTGGCGGAAATGG-3′ 5′-AGGTGGGGCGAAAGGAAAC-3′ 459

Bcl-2 5′-ATGTGTGTG GAGAGCGTCAA-3′ 5′-ACAGTTCCACAA AGGCATCC-3′ 136

β-actin 5′-CATCCTGCGTCTGGACCT-3′ 5′-CAGGAGGAGCAATGATCTTG-3′ 480

Fig. 1 Lentiviral vector system. a pGCL-GFR vector encoding sustained RNA expression, b pHelper 1.0 vector, encoding a viral structuralprotein, c pHelper 2.0 vector, encoding a viral capsid protein

Tumor Biol.

cold PBS, resuspended in PBS containing 10 mg/Lpropidium iodide (Sigma) and 100 mg/L RNase A (AsbioTechnology Inc, Guangzhou, China), and then incubated at25 °C in the dark for 45 min. The cell population percentagein each phase of the cell cycle was measured usingFACSstar, and the results were analysed using theCELLQUEST software (Becton Dickinson and Company,NJ, USA). Apoptosis was detected with the sub-G1 DNAcontent fraction using propidium iodide, and the experimentwas repeated three times.

TUNEL assay

Apoptosis assay was performed using Apo-Direct TUNELAssay kit (Millipore, Billerica, MA).Cells were harvestedand fixed in 4 % PFA for 60 min at 4 °C, followed by asecond fixation in 70 % (v/v) ethanol overnight at −20 °C.Cells were then treated by various reagents for a designedperiod according to the manufacturer’s instruction.Finally, cells were analysed by flow cytometry usingFACS Vantage machine (Becton Dickinson). The CellQuest software (Verity Software House) was used to an-alyse the data.

Xenograft cancer model

The study protocol was reviewed and approved by theEthics and Research Committee of the State KeyLaboratory of Oncology of South China and the Sun Yat-Sen University Cancer Center, Guangzhou, China. Femaleathymic Balb/c nude mice (Laboratory Animal Center, SunYat-sen University, Guangzhou, China) were housed in in-dividually ventilated cages at (25±1 °C) and a 12-h light/12-h dark cycle. The use of animals in this study complies withthe Guide for the Care and Use of Laboratory Animals(National Institutes of Health publication no. 86–23, revised1985). Prior to injection, 15 nude mice (age 5 weeksweighing 16.6±0.34 g) were assigned at random to threegroups with five mice per group. Different groups of cellswere washed and resuspended at 5×107 cells/mL in PBS,and 0.25 mL was then subcutaneously injected into the rightflank of nude mice. Tumour growth was measured by thetumour diameter with a Vernier calliper every 5 days begin-ning on the 5th day until the 30th. The tumour volume wascalculated as the length×width2/2, where the length andwidth are the longest and shortest axes in millimeters [12].At the end of the experiment (the 30th day), followingeuthanasia, tumours were excised, fixed in 10 % formalin,and embedded in paraffin block for immunohistochemistry(IHC) study. IHC was performed using the standardstreptavidin–biotin–peroxidase complex method. A 1:100diluted anti-CXCR4 (Abcam, Cambridge, UK) antibodywas used for CXCR4 detection.

Statistical analysis

All statistical analyses were performed by StatisticalPackage of Social Sciences 13.0 software. P value<0.05was considered to be statistically significant. All of theresults are presented as the mean±standard deviation. Alldata were statistically analysed by Student’s t test.

Results

CXCR4 and CXCL12 expression in EC9706 cells

The expression of CXCL12 and CXCR4 in EC9706 cellswas examined by Western blot (Fig. 2).

Plasmid construction and transfection

The coded target sequence of the obtained recombinantplasmids was constructed including the CXCR4-shRNArecombinant and negative control plasmids. The transfectionoutcome was observed using a fluorescence microscope72 h after experimental treatment. The transfection rate ofthe CXCR4-RNAi-LV group was (87.3±1.2)%, and it was(90.1±1.4)% for the NC-GFP-RNAi-LV group (Fig. 3).

Inhibition of CXCR4 shRNA on CXCR4 proteinand mRNA

The CXCR4 mRNA levels in the control, CXCR4-RNAi-LV, and NC-GFP-RNAi-LV groups were 107.5±5.5, 28.5±2.3, and 104.5±6.4 %, respectively (Fig. 4a). Western blotanalysis was used to determine the expression of theCXCR4 protein. The CXCR4 protein levels in the control,CXCR4-RNAi-LV, and NC-GFP-RNAi-LV groups were

0102030405060708090

100

CXCL12 CXCR4

Rat

io o

f C

XC

L12

and

CX

CR

4pr

otei

n/ β

-act

in (

%)

Fig. 2 CXCL12 and CXCR4 expression in EC9706 cells. CXCL12and CXCR4 expression was determined by Western blot. β-actin wasused as a loading control

Tumor Biol.

80.2±2.3, 21.5±1.0, and 77.5±3.6 %, respectively(Fig. 4b). These results indicate that CXCR4-RNAi-LVcould efficiently silence CXCR4 in EC9706 cells.

Knocking down CXCR4 expression inhibits cellproliferation

An MTT assay was employed to study the impact ofdownregulating CXCR4 expression on cell proliferation.The number of proliferating cells was determined bymeasuring

the optical density (OD) value at 490 nm. The cell lines in all ofthe groups were cultured for 144 h, and the OD value was thendetermined and compared among the groups. The detectionresults indicated that the speed of cell proliferation was slowerin the CXCR4-RNAi-LV than in the control and NC-GFP-RNAi-LV groups (P<0.05, Fig. 5).

Silence of CXCR4 inhibits proliferation of EC9706 cellsby cell cycle and apoptosis

Cell cycle profiles were monitored by flow cytometric anal-ysis. Compared with that of the NC-GFP-RNAi-LV andcontrol group cells, the percentage of the G0/G1 phase andapoptosis rate in the CXCR4-RNAi-LV group cells in-creased, and the percentage of S phase in the CXCR4-RNAi-LV group cells decreased. (Figs. 6 and 7 andTable 2). These indicate that the proliferation inhibitoryeffect of silencing CXCR4 may be due to G0/G1 cell cyclearrest and increased apoptosis in EC9706 cells.

Knocking down CXCR4 decreases the growth of EC9706xenograft tumour in nude mice

To further investigate the CXCR4-RNAi-LV inhibitoryeffect on tumour cell growth in vivo, we used a xeno-graft oesophageal cancer model (Figs. 8a and 9).Tumours first appeared after 5 days, and growth inhibi-tion was observed from days 0 to 30 after inoculation,which was at the observation end point. The final tumourvolume was 603.4±60.8 mm3 for the CXCR4-RNAi-LVgroup, whereas the volume in the NC-GFP-RNAi-LV andcontrol groups was 1,095.1±102.1 and 1,065.3±116.2 mm3,

Fig. 3 Representative graphsof EC9706 cells infected withthe indicated lentiviruses areshown. a Control group, bCXCR4-RNAi-LV group, cNC-GFP-RNAi-LV group

Fig. 4 CXCR4 shRNA inhibits CXCR4 gene expression in EC9706cells. a Reverse transcriptase-polymerase chain reaction (RT-PCR)analysis of CXCR4 RNA levels. b Western blot analysis of theCXCR4 protein levels in the three groups; β-actin was used as aloading control. ① Control group; ② NC-GFP-RNAi-LV group; ③CXCR4-RNAi-LV group. *P<0.01 as compared with the control andNC-GFP-RNAi-LV groups

0

0.2

0.4

0.6

0.8

1

1.2

1.4

1 2 3 4 5 6Time (day)

Opt

ical

den

sity

valu

e (4

90nm

)

control groupNC-GFP-RNAi-LV groupCXCR4-RNAi-LV group

* *

Fig. 5 MTT analysis of cell proliferation after knocking down CXCR4expression. *P<0.05 as compared with the control and NC-GFP-RNAi-LV groups

Tumor Biol.

respectively (P<0.05, Fig. 8b). The average tumour weightin the CXCR4-RNAi-LV group was 283.5±21.3 mg com-pared with 504.6±42.5 mg in NC-GFP-RNAi-LV and492.5±43.1 mg in control group, respectively (P<0.05,Fig. 8c). These data indicate that CXCR4-RNAi-LV cansuppress the tumour growth of human oesophageal cancercells in nude mice.

Down-expression of CXCR4 leads to a decrease in Bcl-2

Bcl-2 expression was decreased when CXCR4 was knockeddown by RNAi. The efficiency of the Bcl-2 proteindownregulation was analysed by RT-PCR and Western blot.Bcl-2 mRNA level in the control, NC-GFP-RNAi-LV, andCXCR4-RNAi-LV groups was 110.5±6.5, 115.6±8.5, and35.4±3.6 %, respectively (Fig. 10a). The Bcl-2 proteinexpression intensity in the control, NC-GFP-RNAi-LV, andCXCR4-RNAi-LV groups was 80.2±2.3, 77.5±3.6, and

21.5±1.0 %, respectively (Fig. 10b). Statistical analysisdemonstrated that Bcl-2 protein expression in the CXCR4-RNAi-LV cell group was significantly downregulated com-pared with that in the NC-GFP-RNAi-LV and control cellgroups (P<0.05). There was no significant difference be-tween the NC-GFP-RNAi-LV and control groups.

Discussion

Chemokines are chemotactic cytokines that cause the direct-ed migration of leukocytes, and are induced by inflamma-tory cytokines and pathogenic stimuli. CXCL12 is alsocalled stromal cell-derived factor-1, and CXCR4 is a spe-cific receptor of CXCL12. Previous studies found that theCXCL12/CXCR4 pathway participates in physiologicalprocesses such as the regulation of organic immunity[4,13] and hemopoietic stem cell transplantation [14].

Fig. 6 RNA interference targeted CXCR4 inhibits proliferation ofEC9706 cells due to the G0/G1 cell cycle arrest and the increasedapoptosis. Cell cycle profiles were obtained by staining with propidium

iodide. Cell cycle distributions displayed here were representative ofthree experiments (a control group, b NC-GFP-RNAi-LV group, cCXCR4-RNAi-LV group)

Fig. 7 Apoptotic cells were increased in CXCR4-RNAi-LV group. *P<0.05 as compared with the control and NC-GFP-RNAi-LV groups

Tumor Biol.

Recent studies revealed that there was high expres-sion of CXCL12 and CXCR4 in normal tissues andimmune cells as well as in a variety of malignanciesincluding colorectal cancer [15], breast cancer [16], andlung cancer [17]. Primary tumours with positive CXCR4

expression demonstrate higher proliferation and meta-static potential than those with negative CXCR4 expres-sion [8]. The findings of Sehgal et al. [18] indicatedthat CXCR4 overexpression occurs in glioblastoma, andproliferation was significantly depressed in glioblastomacell lines treated with a CXCR4-specific antibody.Sasaki et al. [19] reported that CXCL12/CXCR4 is inten-sively expressed in ESCC. Whether the CXCL12/CXCR4passageway does and how it influences the survival andproliferation of ESCC cells may be useful information forthe treatment and research of ESCC. Studies of ESCC havedemonstrated that the expression of the CXCL12/CXCR4signalling pathway closely correlates with ESCC prognosis[19,20]. These results suggested that CXCL12/CXCR4 mayplay a key role in ESCC.

RNAi is one of the posttranscriptional gene silencingmechanisms, which is now widely used in the research ofgene analysis and therapy. RNAi has emerged as a powerfultool to induce loss-of-function phenotypes by the posttran-scriptional silencing of gene expression. A lentivirus is aretrovirus, and lentiviral vectors can efficiently deliversi/shRNA expression cassettes into various cells withsustained expression and potent function of the encodedsiRNAs [21].

In our study, the expression of CXCL12 and CXCR4 at themRNA and protein level was confirmed by RT-PCR andWestern blot, respectively. A lentiviral shRNA expressionvector was chosen to design the RNAi effector sequencefragment targeting the CXCR4 gene to make the gene silenc-ing effects more specific, efficient, and stable. The resultsdemonstrated that the transfection of CXCR4-shRNA intoEC9706 cells significantly suppressed CXCR4 expression inWestern blot and RT-PCR assays. Lapteva et al. [22] reportedthat small interfering RNA directed against CXCR4 effective-ly downregulates CXCR4 expression in human MDA-MB-231 breast cancer cells, leading to a significant decrease inbreast cancer cell proliferation. Our previous study found thatthe level of CXCR4 is positively correlated with ESCC pro-gression and prognosis [23]. In the current study, MTT assayand xenograft tumour model revealed that lentivirus-mediatedshRNA could efficiently silence CXCR4, EC9706 cell prolif-eration was inhibited, and the EC9706 cell inhibition was dueto G0/G1 arrest in the cell cycle and apoptosis induction. Weiet al. [24] also reported that the application of a CXCR4

Table 2 Cell distributions de-termined by flow cytometry inthree group cells

*P<0.05, as compared withcontrol group and NC-GFP-RNAi-LV group

Groups Distribution in cell cycle(%) Apoptosis(%)

G0/G1 S G2/M

Control group 47.1±1.0 35.4±1.4 17.5±0.6 2.17±0.35

NC-GFP-RNAi-LV group 45.3±2.5 36.1±1.6 18.6±1.6 1.93±0.25

CXCR4-RNAi-LV group 66.9±3.5* 18.9±1.7* 14.2±1.4 11.37±2.56*

Fig. 8 CXCR4 silencing decreases EC9706 xenograft tumour growthin nude mice. a Image of tumour specimens dissected from nude micexenografted with human oesophageal cancer cells; b Tumour volumecurves were generated using data from the three different groups; cAnalysis of the tumour weight in three groups. ① Control group; ②NC-GFP-RNAi-LV group; ③ CXCR4-RNAi-LV group. *P<0.05, ascompared with control group and NC-GFP-RNAi-LV group

Tumor Biol.

monoclonal antibody could suppress the growth of acutemyelocytic leukemia cell lines HL-60, the proportion of the

cells in the proliferative phase decreased and that of cells inthe stationary phase increased.

Except for the G0/G1 cell cycle arrest, the induction ofapoptosis is another important factor in the cell proliferationchange. Our results indicated that highly efficient genetransfection and decreased CXCR4 target gene and proteinexpression inhibited the CXCL12/CXCR4 biological ef-fects. As a result, the EC9706 cell apoptotic rate was en-hanced, and at the same time, we also found that Bcl-2expression at the protein and mRNA levels was decreased.Previous studies found that CXCL12/CXCR4 signaling in-duce upregulation of Bcl-2 protein in human intrathymicCD34+ cells and granulosa cells [25,26]. We hypothesizedthat the change in the apoptotic rate could possibly beexplained by mechanisms of CXCL12/CXCR4 downstreamapoptosis-related gene silencing, such as Bcl-2. Bcl-2 is aproto-oncogene intrinsically involved in the apoptosis cas-cade. Bcl-2 protein overexpression is associated with poorprognosis and aggressive biological behaviour in carcinoma[27,28]. Further investigation is needed to determine wheth-er there are any other pathways that participate and effectchanges in cell cycle and apoptosis in ESCC cells. Studiesof ESCC have demonstrated that CXCL12/CXCR4 alsoinfluences ESCC cell invasiveness, metastatic potential,and angiogenesis, CXCR4 was highly expressed in ESCCand correlated with MMP-9, VEGF, and prognosis. Itsuggested that CXCR4 play an important role during theprogression of this disease, and there might be some regu-latory mechanism existing between CXCR4 and MMP-9/VEGF in ESCC [29,30].

In summary, lentivirus-mediated shRNA could silencethe specific CXCL12 receptor CXCR4 in the EC9706ESCC cells and inhibit the biological activity of theCXCL12/CXCR4 pathway. EC9706 cell proliferation wassignificantly inhibited in vitro and in vivo. We propose thatlentivirus-mediated shRNA directed against CXCR4 may bea novel approach for ESCC treatment, and this strategy mayprovide a new method for gene therapy in ESCC patients.

Acknowledgments We gratefully thank the staff members in theDepartment of Medical Oncology and Thoracic Surgery Oncology atSun Yat-sen University Cancer Center for their suggestion and assistance.

Conflicts of interest None

Funding This work was supported by a grant from the Natural ScienceFoundation of the Guangdong Province of China (S2011040004502).

Fig. 9 Representatives ofCXCR4 expression weredetected by IHC in xenograft(×400 objective). a Controlgroup, b NC-GFP-RNAi-LVgroup, c CXCR4-RNAi-LVgroup

Fig. 10 RNA interference inhibits Bcl-2 gene expression in EC9706cells. a Reverse transcriptase-polymerase chain reaction (RT-PCR)analysis of the Bcl-2 RNA levels. b Western blot analysis of Bcl-2protein levels in the three cell groups. β-actin was employed as aloading control. ① Control group, ② NC-GFP-RNAi-LV group, ③CXCR4-RNAi-LV group. *P<0.01 as compared with the control andNC-GFP-RNAi-LV groups

Tumor Biol.

References

1. Lu SH. Alterations of oncogenes and tumor suppressor genes inesophageal cancer in China. Mutat Res. 2000;462:343–53.

2. Aiuti A et al. Expression of CXCR4, the receptor for stromal cell-derived factor-1 on fetal and adult human lympho-hematopoieticprogenitors. Eur J Immunol. 1999;29:1823–31.

3. Jinquan T et al. CXC chemokine receptor 4 expression and stromalcell-derived factor-1 alpha-induced chemotaxis in CD4+ T lym-phocytes are regulated by interleukin-4 and interleukin-10.Immunology. 2000;99:402–10.

4. Oonakahara K, Matsuyama W, Higashimoto I, Kawabata M,Arimura K, Osame M. Stromal-derived factor-1alpha/CXCL12-CXCR 4 axis is involved in the dissemination of NSCLC cellsinto pleural space. Am J Respir Cell Mol Biol. 2004;30:671–7.

5. Yasumoto K et al. Role of the CXCL12/CXCR4 axis in peritonealcarcinomatosis of gastric cancer. Cancer Res. 2006;66:2181–7.

6. Sasaki K et al. Expression of CXCL12 and its receptor CXCR4 inesophageal squamous cell carcinoma. Oncol Rep. 2009;21:65–71.

7. Ganju RK et al. The alpha-chemokine, stromal cell-derived factor-1alpha, binds to the transmembrane G-protein-coupled CXCR-4receptor and activates multiple signal transduction pathways. J BioChem. 1998;273:23169–75.

8. Balkwill F. Cancer and the chemokine network. Nat Rev Cancer.2004;4:540–50.

9. Juarez J, Bradstock KF, Gottlieb DJ, Bendall LJ. Effects of in-hibitors of the chemokine receptor CXCR4 on acute lymphoblasticleukemia cells in vitro. Leukemia. 2003;17:1294–300.

10. Liu G et al. Critical role of SDF-1α-induced progenitor cell re-cruitment and macrophage VEGF production in the experimentalcorneal neovascularization. Mol Vis. 2011;17:2129–38.

11. Yoon O, Roh J. Downregulation of KLF4 and the Bcl-2/Bax ratioin advanced epithelial ovarian cancer. Oncol Lett. 2012;4:1033–6.

12. De Cesare M et al. Enhanced antitumour efficacy of gimatecan incombination with Bcl-2 antisense oligonucleotide in human mela-noma xenografts. Eur J Cancer. 2005;41:1213–22.

13. Peled A et al. The chemokine SDF-1 activates the integrins LFA-1,VLA-4, and VLA-5 on immature human CD34(+) cells: role intransendothelial/stromal migration and engraftment of NOD/SCIDmice. Blood. 2000;95:3289–96.

14. Lee Y et al. Enhancement of intracellular signaling associated withhematopoietic progenitor cell survival in response to SDF-1/CXCL12 in synergy with other cytokines. Blood. 2002;99:4307–17.

15. Brand S et al. CXCR4 and CXCL12 are inversely expressed incolorectal cancer cells and modulate cancer cell migration, inva-sion and MMP-9 activation. Exp Cell Res. 2005;310:117–30.

16. Müller A et al. Involvement of chemokine receptors in breastcancer metastasis. Nature. 2001;410(6824):50–6.

17. Wagner PL et al. CXCL12 and CXCR4 in adenocarcinoma of thelung: association with metastasis and survival. J ThoracCardiovasc Surg. 2009;137:615–21.

18. Sehgal A, Keener C, Boynton AL, Warrick J, Murphy GP.CXCR4, a chemokine receptor, is overexpressed in and requiredfor proliferation of glioblastoma tumor cells. J Surg Oncol.1998;69:99–104.

19. Sasaki K et al. Expression of CXCL12 and its receptor CXCR4correlates with lymph node metastasis in submucosal esophagealcancer. J Surg Oncol. 2008;97:433–8.

20. Kaifi JT et al. Tumor-cell homing to lymph nodes and bonemarrow and CXCR4 expression in esophageal cancer. J NatlCancer Inst. 2005;97:1840–7.

21. Li M, Rossi JJ. Lentiviral vector delivery of siRNA and shRNAencoding genes into cultured and primary hematopoietic cells.Methods Mol Biol. 2005;309:261–72.

22. Lapteva N, Yang AG, Sanders DE, Strube RW, Chen SY. CXCR4knockdown by small interfering RNA abrogates breast tumorgrowth in vivo. Cancer Gene Ther. 2005;12:84–9.

23. Wang DF, Lou N, Zeng CG, Zhang X, Chen FJ. Expression ofCXCL12/CXCR4 and its correlation to prognosis in esophagealsquamous cell carcinoma. Ai Zheng. 2009;28:154–8.

24. Wei L, Kong PY, Chen XH, Peng XG, Zeng DF, Chang C, et al.Effects of anti-CXCR4 monoclonal antibody on adhesion andproliferation of human acute myelocytic leukemia cell line HL-60. Ai Zheng. 2004;23:1273–7.

25. Hernández-López C, Varas A, Sacedón R, Jiménez E, Muñoz JJ,Zapata AG, et al. Stromal cell-derived factor 1/CXCR4 signaling iscritical for early human T-cell development. Blood. 2002;99:546–54.

26. Kryczek I et al. The chemokine SDF-1/CXCL12 contributes to Tlymphocyte recruitment in human pre-ovulatory follicles and co-ordinates with lymphocytes to increase granulosa cell survival andembryo quality. Am J Reprod Immunol. 2005;54:270–83.

27. Agui T, McConkey DJ, Tanigawa N. Comparative study of variousbiological parameters, including expression of survivin, betweenprimary and metastatic human colonic adenocarcinomas.Anticancer Res. 2002;22:1769–76.

28. Tsai ST, Jin YT, Leung HW, Wang ST, Tsao CJ, Su IJ. Bcl-2 andproliferating cell nuclear antigen (PCNA) expression correlateswith subsequent local recurrence in nasopharyngeal carcinomas.Anticancer Res. 1998;18:2849–54.

29. Lu CL, Ji Y, Ge D, Guo J, Ding JY. The expression of CXCR4 andits relationship with matrix metalloproteinase-9/vascular endothe-lial growth factor in esophageal squamous cell cancer. DisEsophagus. 2011;24:283–90.

30. Gros SJ, Graeff H, Drenckhan A, Kurschat N, Blessmann M,Rawnaq T, et al. CXCR4/SDF-1α-mediated chemotaxis in an in vivomodel of metastatic esophageal carcinoma. In Vivo. 2012;26:711–8.

Tumor Biol.