PLK1 Is Transcriptionally Activated by NF-kB during Cell Detachmentand Enhances Anoikis Resistance through Inhibiting b-CateninDegradation in Esophageal Squamous Cell Carcinoma
AbstractPurpose: To investigate the molecular mechanisms through which polo-like kinase-1 (PLK1) takes part
in anoikis resistance of esophageal squamous cell carcinoma (ESCC) cells.
Experimental Design: The role of PLK1 in cell anoikis resistance was examined by ectopic gene
expression and siRNA-mediated knockdown. Glutathione S-transferase pull-down and co-immunopreci-
pitation assays were utilized to investigate PLK1-interacting proteins. Electrophoretic mobility shift assay,
chromatin immunoprecipitation, and reporter gene assays were carried out to identify the transcription
factors responsible for PLK1 expression during anoikis resistance.
Results: We found that detachment of ESCC cells triggers the upregulation of PLK1. Elevated PLK1
expression contributes to protection against anoikis in cancer cells through the regulation of b-cateninexpression. Moreover, we showed that, through direct binding to the PLK1 promoter, the NF-kB subunit
RelA transcriptionally activates PLK1, which inhibits the ubiquitination and degradation of b-catenin.Inhibition of the NF-kB pathway restores the sensitivity of cancer cells to anoikis by downregulating
PLK1/b-catenin expression. In addition, RelA gene amplification and protein overexpression was
significantly correlated with PLK1 expression in ESCC tissues.
Conclusions: Our findings suggest that upregulation of PLK1 triggered by cell detachment is
regulated by RelA at the transcriptional level. PLK1 protects esophageal carcinoma cells from anoikis
through modulation of b-catenin protein levels by inhibiting their degradation. Taken together, this
study reveals critical mechanisms involved in the role of RelA/PLK1/b-catenin in anoikis resistance of
ESCC cells. Clin Cancer Res; 17(13); 4285–95. �2011 AACR.
Introduction
Esophageal squamous cell carcinoma (ESCC) is a devas-tating disease because it metastasizes early and is highlyresistant to conventional chemotherapy and radiation ther-apy. Characterization of the genetic alterations and thedownstream effectors that contribute to ESCC metastasiswill facilitate the development of effective therapeuticstrategies for combating the disease.The extracellular matrix (ECM) provides adhesive sup-
port to tissues and controls numerous signals that reg-ulate diverse cellular processes, such as survival, growth,
and differentiation (1). Detachment of normal epithelialcells from the ECM typically results in anoikis, which isessential for several morphogenetic and homeostatic pro-cesses, such as embryo cavitation (2), postweaning mam-mary gland regression (3), and elimination of epithelialcells shed into the intestinal lumen (4). In contrast,malignant cells are resistant to anoikis, which leads toenhanced survival after detachment from the supportingmatrix and facilitates metastasis (5). Thus, elucidating themolecular mechanisms involved in cancer-associatedanoikis resistance is critical for the development of thera-pies designed to restore the sensitivity of malignant cellsto anoikis.
Human polo-like kinase 1 (PLK1) is a highly conservedserine (Ser)/threonine (Thr) kinase that regulates a multi-tude of mitotic processes. PLK1 has been reported to beupregulated in several solid tumors, such as esophageal,breast, ovarian, prostate, and colon cancer (6–9). Inhibi-tion of PLK1 with antisense oligonucleotides, siRNA, ordominant-negative mutations leads to mitotic catastrophe,apoptosis, and tumor inhibition (10–12). However, themolecular mechanisms that underlie PLK1 overexpressionand its anti-apoptotic function in cancer cells are largelyunknown.
Authors' Affiliations: 1State Key Laboratory of Molecular Oncology and2Department of Pathology, Cancer Institute (Hospital), Peking UnionMedical College & Chinese Academy of Medical Sciences, Beijing,China
Note: Supplementary data for this article are available at Clinical CancerResearch Online (http://clincancerres.aacrjournals.org/).
In this study, we showed that upregulation of PLK1triggered by cell detachment enhances anoikis resistanceof esophageal cancer cells by inhibiting b-catenin degra-dation. Most importantly, we show that PLK1 is a directtarget of RelA. RelA binds to the PLK1 promoter andregulates PLK1 transcription, which consequently stabi-lizes the b-catenin protein by disrupting its interactionwith GSK-3b and b-TrCP. Altogether, our study revealscritical mechanisms involved in the dysregulation ofPLK1 in human tumors and the role of PLK1 in anoikisresistance.
Materials and Methods
Patients and tissue specimensTissue specimens from 157 pathologically confirmed
ESCCs and adjacent histologically normal tissues wereobtained from patients who had undergone single-stagecurative esophagectomy at the Cancer Hospital, ChineseAcademy of Medical Sciences, Beijing, China. The sampleswere obtained following written informed consent frompatients and used with the approval of the InstitutionalReview Board. Of the 157 samples, 125 and 53 sampleswere used for immunohistochemistry and for genomicreal-time PCR, respectively. Twenty-one samples wereassayed, both for RelA gene amplification and RelA proteinexpression.
phageal tumors and the corresponding normal epitheliawere created, and immunohistochemical analysis was doneas described previously (9). TMAs were incubated with ananti-RelA antibody (sc-8008; Santa Cruz Biotechnology) oran anti-PLK1 antibody (Upstate). The results were sepa-rately evaluated by 2 independent observers. For RelA, thestaining intensity was graded on the following scale: 0(negative), 1 (weak-moderate), and 2 (strong). The evalua-tion criteria for PLK1 expression have been described pre-viously (9).
Real-time PCRThe genomic copy number of RelA was assessed in
53 ESCC samples by TaqMan probe-based real-time
PCR (assay ID: Hs02229145_cn; Applied Biosystems)on a LightCycler 480 Real-Time PCR system (RocheApplied Science) according to the manufacturer’s instruc-tions. The RNase P H1 RNA gene was used as thestandard reference (Applied Biosystems). The thresholdcycle (Ct) for each gene was determined, and the aver-age of 4 independent experiments was calculated. Therelative copy number of the RelA gene normalized tothe reference and relative to the calibrator is given bythe formula 2�DDCt using the comparative Ct method(13). Gene amplification was defined as a copy number(2 � 2�DDCt) > 3.
Real-time reverse transcriptase PCR (RT-PCR) analysis ofthe PLK1 mRNA levels is described in the SupplementaryMethods.
Cell culture, plasmids, and antibodiesThe human ESCC cell line EC3 was established in culture
in our laboratory. The human ESCC cell lines KYSE450 andKYSE410 were generously provided by Dr. Y. Shimada(Kyoto University, Kyoto, Japan). Cells were cultured inDulbecco’s modified Eagle medium (DMEM) containing10% FBS. Other plasmids, siRNAs, and antibodies aredescribed in the Supplementary Methods.
Assessment of anoikisCells were trypsinized and plated in 6-well polyhydrox-
yethylmethacrylate (polyHEMA) plates (which were pre-pared by applying 1.5 mL of a 10 mg/mL solution ofpolyHEMA in ethanol onto the plate and then allowingit to dry in a tissue culture hood). After 24 hours of growthin suspension, cells were harvested for apoptosis measure-ments using the Annexin V–FITC Apoptosis Kit (Sigma) orthe Annexin V–R-phycoerythrin Apoptosis Kit (SouthernBiotech).
See Supplementary Methods for details.Electrophoretic mobility shift assay. Electrophoretic
mobility shift assay (EMSA) was conducted using theLightShift Chemiluminescent EMSA Kit (Pierce) accordingto the manufacturer’s instructions. The oligonucleotide 50-GCTGCGCAGGGCGCTCCCATGGTGCC-30 (the NF-kBbinding element is underlined) was labeled with biotin.For the competitive experiments, a 100-fold excess ofunlabeled oligonucleotide was used. For supershift experi-ments, nuclear extracts were preincubated with anti-RelAantibody (sc-372x; Santa Cruz Biotechnology).
Chromatin immunoprecipitationChromatin immunoprecipitation (ChIP) assays were
conducted using the Magna ChIP G Kit (Millipore) accord-ing to the manufacturer’s instructions. The target proteinwas immunoprecipitated with either 2 mg of anti-RelApolyclonal antibody (ab7970; Abcam) or rabbit immuno-globulin G (IgG) as a negative control. The primersequences are provided in Supplementary Table S2.
Translational Relevance
Our work revealed that the RelA–PLK1–b-cateninpathway plays an important role in anoikis resistancein esophageal squamous cell carcinoma (ESCC) cells.Furthermore, we found RelA gene amplification and co-overexpression of RelA and PLK1 proteins in esophagealcancer cell samples. These results suggested that theRelA–PLK1–b-catenin pathway might be a potentialtarget of therapies designed to restore the sensitivityof ESCC cells to anoikis.
Lin et al.
Clin Cancer Res; 17(13) July 1, 2011 Clinical Cancer Research4286
Luciferase reporter assayESCC cells were transiently transfected with 0.4 mg of luci-
ferase reporter plasmid pGL3-Basic or expression plasmids(NF-kB-Luc, PLK1-WT-Luc, or PLK1-Mut-Luc). To correct forvariations in transfection efficiency and cell number, 0.4 ngof the pRL-SV40 vector encoding the Renilla luciferase genedriven by the cytomegalovirus promoter (Promega) wascotransfected in each experiment. Twenty-four hours post-transfection, cell lysates were prepared, and both firefly andRenilla luciferase activities were quantified using the Dual-Luciferase Reporter Assay System (Promega). The data arepresented as the ratio of firefly to Renilla luciferase activity.
Statistical analysisThe correlation between the expression levels of RelA
and PLK1 was analyzed using the Spearman rank cor-
relation test. The correlation between RelA protein expres-sion and gene amplification was analyzed by Pearson’sc2 test. Other statistical analyses were conducted usingthe Student’s t-test. Statistical significance was defined atP < 0.05.
Results
Detachment-induced upregulation of PLK1 inhibitsanoikis of ESCC cells
In our previous study, we showed that several ESCC celllines are relatively resistant to anoikis (14, 15). In thisstudy, we found that detachment of ESCC cells upregulatedboth the mRNA and protein levels of PLK1 (Fig. 1B; Sup-plementary Fig. S1). In contrast, PLK1 expression wasdownregulated in the nonmalignant intestinal epithelial
Figure 1. Upregulation of PLK1during cell detachmentsuppresses anoikis of ESCC cells.A, cells were cultured onpolyHEMA-coated dishes for24 hours and subjected to flowcytometry (FCM) analysis byAnnexin V–FITC/propidium iodidestaining. B, detached KYSE450,EC3, and IEC-6 cells werecultured for the indicated timesand assayed for PLK1 expressionby Western blot analysis.Extracellular signal-regulatedkinase (ERK) was used as aloading control. C, ESCC or IEC-6cells were transiently or stablytransfected with PLK1 or treatedwith BI-2536 before suspension,as indicated. Anoikis-inducedcells were then subjected to FCManalysis by Annexin V–R-phycoerythrin/propidium iodidestaining. Columns, mean; Errorbars, SEM (n ¼ 3); *, P < 0.05;**, P < 0.01.
A B100
90
80
70
60
50
40
30
20
10
0KYSE450
KYSE450 KYSE450
KYSE450
Apo
ptot
ic c
ell p
erce
nt (
100%
)
C100
90
80
70
60
50
40
30
20
10
0
Apo
ptot
ic c
ell p
erce
nt (
100%
)
EC3
EC3
Vecto
r
PLK1
Vecto
r
PLK1
Vecto
r
PLK1
Vecto
r
PLK1
DMSO
BI 253
6
EC3 stable
EC3
IEC-6
IEC-6
IEC-6
PLK1
ERK
PLK1
ERK
PLK1
ERK
0 h 3 h 6 h 10 h
0 h 2.5 h 5 h 10 h7.5 h
0 h 2.5 h 5 h 10 h7.5 h
NF-kB Promotes Anoikis Resistance via PLK1/b-Catenin
www.aacrjournals.org Clin Cancer Res; 17(13) July 1, 2011 4287
cell line IEC-6, which is highly susceptible to anoikis(Fig. 1A and B).
On the basis of the above observations, we hypo-thesized that changes in PLK1 expression level might beresponsible for the susceptibility of ESCC cells to anoikis.To address this hypothesis, we first measured the ability ofexogenous PLK1 to inhibit anoikis of esophageal cancercells. Transient expression of PLK1 significantly promoted
the survival of EC3, KYSE450, and IEC-6 cells cultured insuspension; similar results were obtained when PLK1was stably expressed in these cells (Fig. 1C). To validatethe anti-anoikis role of PLK1, we silenced endogenousPLK1 using specific siRNAs. As shown in Figure 2F andG (the first 3 groups), both siRNAs significantly sup-pressed PLK1 expression, which led to an increase inapoptotic cells upon detachment from the ECM. Similar
Figure 2. PLK1 regulates anoikis resistance through b-catenin in ESCC cells. A, after SDS-PAGE, bands present only in the PBD eluents were excisedand digested with trypsin and subjected to LC/MS-MS. Amino acid sequences of 2 peptides corresponding to human b-catenin were identified. B, total 1 mgKYSE450 and KYSE410 cell lysates were prepared and immunoprecipitated with mouse monoclonal anti-PLK1 antibody or mouse IgG (as control).The immunocomplexes were resolved by 12% SDS-PAGE and immunoblotted with the indicated antibodies. After 48 hours of treatment by indicated siRNAsor plasmids, KYSE450 cells cultured in monolayer (C) and (D), or in suspension (E), for 16 hours were assayed for b-catenin, phosphorylated b-catenin,and PLK1 expression byWestern blot analysis. b-Actin and ERKwere used as loading controls. F, knockdown (KD) of PLK1 increased anoikis cells and forcedtransient expression of b-catenin markedly led to a decrease of apoptotic cells in PLK1-KD cells in anoikis assay. Columns, mean; error bars, SEM (n ¼ 3);* and #, P < 0.05 as compared with the corresponding PLK1-KD groups. G, after transfection for 48 hours, cells of each group were lysed andsubjected to Western blot analysis with Flag or PLK1 antibodies.
Lin et al.
Clin Cancer Res; 17(13) July 1, 2011 Clinical Cancer Research4288
results were noted when these cells were treated with acommonly used PLK1 inhibitor, BI-2536 (Fig. 1C). Thesedata suggest that PLK1 is capable of inhibiting anoikis if itsexpression is induced in detached ESCC cells.
PLK1 regulates anoikis resistance of esophagealcancer cells through b-cateninThe polo-box domain of PLK1 was previously shown to
mediate substrate recognition and targeting of PLK1 (16).To understand the molecular mechanisms by which PLK1inhibits anoikis, glutathione S-transferase (GST) pull-downassays were utilized to identify PLK1-interacting proteins.Peptide mixtures extracted from the GST pull-down assayswere separated using reverse-phase high-performanceliquid chromatography (HPLC; Supplementary Fig. S2)and identified by tandem mass spectrometry (MS-MS).Two peptides (SAIVHLINYQDDAELATR and NEGTA-TYAAAVLFR) weighing approximately 90 kDa were char-acterized, and the Mascot database search revealed that the2 sequences corresponded to amino acids 134 to 151 and648 to 661 of human b-catenin, respectively. Co-immu-noprecipitation (Co-IP) experiments showed that b-cate-nin binds to endogenous PLK1 in esophageal cancer cells(Fig. 2A and B).b-Catenin has been reported to promote anoikis resis-
tance (17), and we confirmed that knockdown of b-cateninresulted in decreased survival of ESCC cells upon detach-ment (Supplementary Fig. S3). Interestingly, in bothattached and detached cells, depletion of PLK1 or transfec-tion of a kinase-dead PLK1 mutant (PLK1/K82M) led toreductions in b-catenin protein levels (Fig. 2C–E). In con-trast, knockdown of b-catenin did not affect PLK1 expres-sion (data not shown). Ectopic overexpression of b-catenin
in PLK1-depleted cells partially restored their resistance toanoikis (Fig. 2F and G). Taken together, these data suggestthat b-catenin contributes to PLK1-mediated anoikis resis-tance in ESCC cells. It is worth noting that the b-cateninmRNA levels did not change in PLK1-depleted cells (Sup-plementary Fig. S4), indicating that posttranslational mod-ification of b-catenin was affected by PLK1 depletion.
PLK1 depletion decreases b-catenin protein levels viathe proteasomal degradation pathway
Ubiquitination and proteasomal degradation of b-cate-nin is initiated by GSK-3b phosphorylation, and phospho-Ser/Thr residues of b-catenin are targets for b-TrCPand specific components of the ubiquitination apparatus(18–20). We examined the phosphorylation status ofb-catenin with an antibody that detected phospho-Ser37of b-catenin. PLK1 depletion led to an increase in phospho-b-catenin (Fig. 2D and E), which is consistent with thedownregulation of b-catenin. No change was observed inGSK–3b and b-Trcp expression levels or activity followingPLK1 depletion (data not shown).
Treatment of cells with the proteasome inhibitor MG-132 restored b-catenin levels in PLK1-depleted cells (Sup-plementary Fig. S5), indicating that downregulation ofb-catenin in PLK1-depleted cells is proteasome depen-dent. Immunoprecipitation of cell lysates with an anti-b-catenin antibody revealed that b-catenin in PLK1-RNAicells was ubiquitinated andmigrated as a smear of proteinbands with slower mobilities (Fig. 3A, lane 2). In addi-tion, greater amounts of GSK-3b and b-Trcp co-immunoprecipitated with b-catenin (Fig. 3A and B). Incontrast, b-catenin ubiquitination was undetectable incontrol cells (Fig. 3A, lane 1). These results indicate that
Figure 3. RNAi of PLK1 enhancedubiquitination of b-catenin. A,ESCC cells from PLK1-KD1 groupor control were treated withproteasomal inhibitor MG-132.Cell lysates were prepared andsubjected to immunoprecipitation(IP) with anti-b-catenin antibody.The level of ubiquitin/GSK-3b–attached b-catenin wasdetected by Western blot analysiswith ubiquitin and GSK-3bantibody, respectively. B,immunoprecipitates weresubjected to SDS-PAGE followedby immunoblot analysis with theindicated antibodies. Equalamounts of cellular proteinextracts were used for each IPreaction. b-Actin was used as aloading control. SN, supernatant.
Ubiquitinated
+
–+
–
++
Nonsilencing
Nonsil
encin
g
MG-132
170 kD
100 kD
55 kD
42 kD
42 kD
SN
IP
67 kD
PLK1-KD1
PLK1-
KD1
IgG IgG
PLK1
GSK-3β
GSK-3β
-β-cateninβ-ca
teni
n
SN
IPβ-
cate
nin
β-Actin
PLK1
β-Actin
β-Trcp
A B
NF-kB Promotes Anoikis Resistance via PLK1/b-Catenin
www.aacrjournals.org Clin Cancer Res; 17(13) July 1, 2011 4289
PLK1 depletion enhances the interaction between b-cate-nin and GSK-3b/b-Trcp and, thereby, promotes the degra-dation of b-catenin.
RelA subunit of NF-kB upregulates PLK1 upon celldetachment
Because PLK1-dependent pathways play an importantrole in the regulation of anoikis resistance, we explored themechanisms underlying the transcriptional activation ofPLK1 in detached ESCC cells.
Searching of online bioinformatics database (http://www.cbrc.jp/research/db/TFSEARCH.html) revealed apotential NF-kB–binding element (designated PLK1-NE)at �93 base pairs (bp) upstream of the transcription-initiation site within the PLK1 promoter region. As a firstapproach to investigate whether PLK1 is a transcriptionaltarget of NF-kB, an EMSA was conducted with a 21-bp
probe that encompassed the predicted PLK1-NE. As shownin Figure 4A, PLK1-NE showed strong binding activityto nuclear proteins and was supershifted by an anti-body against RelA, a subunit of the NF-kB complex. More-over, ChIP assays showed that RelA directly bound tothe �151/þ94 region of the PLK1 promoter, both inattached and detached cells. This interaction was signi-ficantly inhibited if the tumor cells were incubated withpyrrolidine dithiocarbamate (PDTC), a selective NF-kBinhibitor, indicating that the binding was specific (Fig. 4B).
We subsequently tested whether RelA regulated PLK1transcription in suspended cells using a luciferase reporterassay. Consistent with the increase in PLK1 mRNA andprotein levels upon cell detachment, PLK1 promoter activ-ity was upregulated in suspended cells. More importantly,deletion of the PLK1-NE sequence abrogated PLK1 pro-moter activity. These results clearly indicate that PLK1-NE
A
C
BLoading
Supershift
1: NF-κB Hot Probe2: NF-κB Hot Probe + NE3: NF-κB Hot Probe + NE + NF-κB Cold Probe4: NF-κB Hot Probe + NE + RelA Antibody
0 h0
1
2
3
4
5
Rel
ativ
e lu
cife
rase
act
ivity 6
2.5 h 7.5 h 10 h5 h
EC3
KYSE450
Attached
– – – – +
In-p
ut
RIgG
RelA RelA RelA
Detached
Shift
PLK1
BECN1
GAPDH
PDTC
–151 +94
+94–151
NF-κB binding siteLuciferase
Luciferase
GGGCGCTCCC
PLK1-MUT Luc
PLK1-WT Luc
PLK1
BECN1
GAPDH
Free Probe
1 2 3 4
Figure 4. PLK1 is a transcription target of RelA. A, EMSA was conducted using nuclear extracts (NE) from EC3 cells and PLK1-NE labeled with biotin(Hot Probe). Competitive assays were conducted by adding 100-fold excess cold probe. Supershift analysis was conducted using anti-RelA antibody. B, afterEC3 and KYSE450 cells were fixed and sonicated, cellular DNA–protein complex was immunoprecipitated by anti-RelA antibody or anti-rabbit IgG.BECN1 and GAPDH promoter were used as a positive control and a negative control, respectively. PCR products were resolved on a 1.5% agarose gel.C, deletion of PLK1-NE (GGGCGCTCCC) prevented NF-kB–mediated activation of PLK1 transcription activity during cell detachment. After 24 hours oftransfection with the plasmids, detached EC3 cells were cultured for the indicated times, and the luciferase activity was measured as described in Materialsand Methods. Each experiment was done in triplicate wells and repeated 3 times. Columns, mean; error bars, SEM; *, P < 0.05.
Lin et al.
Clin Cancer Res; 17(13) July 1, 2011 Clinical Cancer Research4290
plays an important role in RelA-mediated promoter activa-tion during cell detachment (Fig. 4C).
RelA modulates anoikis resistance through thePLK1/b-catenin pathwayTo further explore the role of RelA in the regulation of
PLK1 expression and anoikis resistance, 2 selective NF-kBinhibitors, PDTC and caffeic acid phenethyl ester (CAPE),were used to block the NF-kB signaling pathway. Bothcompounds significantly inhibited NF-kB–dependent tran-scription (Supplementary Fig. S6A). Notably, treatment ofdetached ESCC cells with PDTC or CAPE not only sup-pressed PLK1 promoter activity but also caused a significant
downregulation of both PLK1 mRNA and protein levels(Fig. 5A and B). This, in turn, led to a reduction of b-cateninprotein and enhanced anoikis of ESCC cells in a dose-dependent manner (Fig. 5A–D). It has been reported thatRelA polyubiquitination was induced at the lysine 195residue, and this ubiquitination event is critical for thedegradation of RelA and termination of NF-kB activation(21). More importantly, ectopic expression of RelA K195Rmutant, which is resistant to polyubiquitination and degra-dation, led to upregulation of both PLK1 mRNA andprotein in suspended cells (Fig. 5E and F) but not inattached ESCC cells (data not shown). In addition, wefound that NF-kB–dependent transcriptional activity was
Figure 5. RelA regulatesPLK1/b-catenin pathway A,KYSE450 and (B) EC3 cells werecultured in suspension andtreated with 50 mmol/L of PDTC,50 mmol/L of CAPE, or DMSO for20 hours. Cells were assayed forb-catenin and PLK1 expression byWestern blot analysis (left); PLK1mRNA level was measured byreal-time RT-PCR (middle); PLK1promoter activity was measuredby luciferase reporter assay (right).C and D, KYSE450 cells werecultured in suspension and treatedwith 50 and 200 mmol/L of PDTCor DMSO for 20 hours. Cells werethen subjected to FCM analysis orassayed for b-catenin and PLK1expression by Western blotanalysis. E and F, after 24 hours oftransfection with the plasmids,detached cells were cultured for16 hours and then assayed forPLK1 mRNA and proteinexpression. EV, empty vector;K195R, Flag-RelA K195R;Columns, mean; error bars, SEM(n ¼ 3); *, P < 0.05; **, P < 0.01;***, P < 0.01.
Paren
tal
DMSO
Paren
tal
DMSO
50 μm
ol/L
200
μmol/
L
PDTCCAPE
ParentalRel
ativ
e m
RN
A e
xpre
ssio
n (%
)
0 Rel
ativ
e Iu
cife
rase
act
ivity
0
5
10
15
20
25
20
40
60
80
100
120
140
DMSO PDTC
PLK1PLK1-WT Luc
mRNA
CAPE DMSO PDTC CAPE
ParentalRel
ativ
e m
RN
A e
xpre
ssio
n (%
)
0 Rel
ativ
e Iu
cife
rase
act
ivity
0123456789
10
20
40
60
80
100
120
DMSO PDTC
PLK1PLK1-WT Luc
mRNA
CAPE
Parental DMSO 50 μmol/L 200 μmol/L
DMSO PDTC
PDTC
PDTC
EV0
0.5
1.0
1.5
2.0
2.5
3.0
3.5
Rel
ativ
e P
LK1
mR
NA
exp
ress
ion
0
10
20
30
40
50
60
70
80
90
100
Apo
ptot
ic c
ell (
%)
K195R EV K195R
CAPE
β-Catenin
β-Actin
PLK1
Paren
tal
DMSO
PDTCCAPE
β-Catenin
β-Actin
PLK1
β-Catenin
ERK
EC3
D
A
B
C
E F
EV K195R EV K195R
KYSE450
EC3
EC3
KYSE450
KYSE450
RelA K195R
ERK
PLK1
PLK1
NF-kB Promotes Anoikis Resistance via PLK1/b-Catenin
www.aacrjournals.org Clin Cancer Res; 17(13) July 1, 2011 4291
induced upon cell detachment (Supplementary Fig. S6C).These data suggest that RelA promotes anoikis resistance,mainly, by regulating PLK1/b-catenin expression. Interest-ingly, we also found that PLK1 promoter activity andexpression were diminished upon PDTC treatment ofattached cells (Supplementary Fig. S6B and D), indicatingthat RelA is also able to regulate PLK1 transcription inattached cells, but this process requires additional factors.
Overexpression of RelA due to gene amplification ispositively correlated with PLK1 dysregulation inESCC tissues
To extend our findings in vivo, we determined whetherthere is a correlation between the expression of RelA andPLK1 in resected human esophageal cancer specimens. Inhistologically normal tissues, RelA was undetectable inmost epithelial cells. In contrast, esophageal tumor cellsexhibited moderate (30.4%; 38 of 125) to intense (35.2%;44 of 125) staining with an anti-RelA antibody (Fig. 6A).More importantly, RelA overexpression correlated withhigh levels of PLK1 expression. In 43 tumor sampleswithout RelA expression, 74.4% (32 of 43) also showedPLK1 negative expression, and only 2.3% (1 of 43)expressed high levels of PLK1. However, in 44 tumor tissueswith RelA overexpression, 54.5% (24 of 44) exhibitedmoderate to intense expression of PLK1 (Fig. 6C; Supple-mentary Table S1).
RelA is located on chromosome 11q in region 11q13,and gene amplification in this region has frequentlybeen observed in human tumors, including those of theesophagus. To determine whether overexpression of RelAis associated with gene amplification, we used quantita-tive genomic real-time PCR to analyze RelA gene copynumbers. Amplification of RelA gene was observed in35.8% (19 of 53) of ESCC patients, and the copy numbersof RelA gene were relatively high in tumors that overex-pressed RelA protein, suggesting that gene amplificationwas responsible for RelA overexpression in some part ofESCC patients (Fig. 6B; Supplementary Table S2).
Discussion
Recent studies have indicated that detachment of epithe-lial cells triggers not only pro-apoptotic but also anti-apoptotic signals, and the equilibrium between these sig-nals regulates anoikis (22–24). In suspended carcinomacells, survival signals typically prevail, and anoikis induc-tion is blocked, ultimately leading to tumor invasion andmetastasis. Several anti-anoikis signals in cancer cells havebeen identified. For example, we and other researchersfound that the activation of oncoproteins, including CTTN(14), calreticulin (15), Ras (25, 26), and b-catenin (17),suppressed anoikis of various types of cancer cells. In thepresent study, we found that detachment of ESCC cells
A BCase-20Tumor Normal
C Case-112RelA PLK1
Case-14
ESCC samples0123456789
10
Rel
A c
opy
num
ber
11121314
RelA PLK1
Figure 6. RelA amplification andoverexpression is correlated withPLK1 upregulation in humanESCC tissues. A, sample caseindicating that RelA isoverexpressed in ESCC tissues byimmunohistochemical staining onthe tissuemicroarray. Top, originalmagnification, �100; bottom,�400 magnification of tissues inblack frames. B, real-time PCRdetection of genomic copynumber of RelA in 53 ESCCspecimens. Thick, horizontal linerepresents the cutoff value forgene amplification whennormalized by the reference gene.C, correlation between RelA andPLK1 expression in ESCC tissues.Top, a representative specimenwith both negative RelA/PLK1staining (left) and another withboth strong positive RelA/PLK1staining. Original magnification,�100; bottom, �400magnification of tissues in blackframes.
Lin et al.
Clin Cancer Res; 17(13) July 1, 2011 Clinical Cancer Research4292
triggered the upregulation of PLK1, a critical anti-apoptoticprotein. Overexpression of PLK1 blocked anoikis indetached cells, and inhibition or depletion of PLK1restored the sensitivity of ESCC cells to anoikis. Further-more, it was noticed that PLK1 depletion did not affect theexpression level of CTTN and calreticulin (SupplementaryFig. S7), indicating that these proteins still can inhibitanoikis in PLK1-depleted cells. This is consistent withour results that a subset of PLK1-depleted cells still survivedafter 24 hours of being cultured in suspension. Based onour previous findings that PLK1 overexpression inhibits themitochondrial apoptotic pathway in ESCC cells, upregula-tion of PLK1 induced by cell detachment probably shiftsthe balance between the life and death signals towardsurvival and suppression of anoikis. However, the apop-tosis-related molecules responsible for this process areunknown as yet and require further investigation.Results from the GST pull-down and Co-IP assays suggest
that b-catenin interacts with PLK1 in ESCC cells. Theobservation that PLK1 modulated b-catenin protein levelsin attached and detached cells led to the hypothesis thatb-catenin could be involved in the PLK1-dependent reg-ulation of anoikis resistance. Ectopic expression of b-cate-nin rescued ESCC cells from detachment-inducedapoptosis due to PLK1 knockdown, indicating that b-cate-nin is an integral component and downstream signalingmolecule in the PLK1-dependent anoikis-resistance path-way. The fact that PLK1 depletion did not affect b-catenin atthe mRNA level but downregulated b-catenin at the proteinlevel led us to speculate that posttranslational stabilizationof b-catenin was regulated by PLK1. Indeed, the interac-tions between b-catenin andGSK-3b/b-Trcpwere enhancedupon PLK1 knockdown, which promoted the ubiquitina-tion and degradation of b-catenin. The importance ofregulating b-catenin stability in cancer cells is supportedby findings that GSK-3b-phosphorylation sites in b-cateninare commonly mutated in human colorectal cancers andother malignancies (27, 28). Furthermore, b-catenin isabnormally highly expressed in ESCC cells that overexpressend-binding protein 1 (EB1) and frequently rearranged inadvanced T-cell lymphomas-1 (FRAT1), indicating thatother events contribute to the dysregulation of b-cateninexpression (29, 30). Thus, the interaction between PLK1and b-catenin and the regulation of b-catenin by PLK1identified in our study represents another mechanism bywhich b-catenin degradation is controlled in ESCC.Arai and colleagues previously found that Ser-718 of
b-catenin was specifically phosphorylated in M-phase byPLK1 (31). However, the study did not show that b-cateninprotein degradation was regulated by PLK1, suggesting thatthis regulation is limited to certain types of tumor cells. Wetransfected PLK1 into ESCC cells and found that PLK1 wasnecessary, but not sufficient, for maintaining high proteinlevels of b-catenin. One possibility is that, in ESCC cells,PLK1 is already overexpressed (9) and it blocks proteaso-mal degradation of b-catenin; therefore, ectopic expressionof PLK1 is somewhat redundant for this process and is notable to make the b-catenin protein more stable. Intrigu-
ingly, ectopic expression of a kinase-dead PLK1 mutant ledto b-catenin downregulation, indicating that PLK1 kinaseactivity is required for b-catenin stabilization. Additionalstudies are required to clarify whether the regulation ofb-catenin protein degradation is mediated by direct phos-phorylation by PLK1 or through cooperation with otherunidentified cellular signals.
The NF-kB family of transcription factors regulates abroad spectrum of biological responses, but its pro- andanti-apoptotic effects appear to be cell-type and contextdependent (32–34). RelA has been shown to induce p53-dependent apoptosis, but it also efficiently counteractsTNF-a-induced apoptosis (35), suggesting that RelA pro-motes the expression of both anti- and proapoptotic mole-cules. In the present study, we showed the role of RelA inanoikis resistance of ESCC cells, which is consistent withprevious reports that NF-kB delays anoikis of intestinalepithelial cells (24). Our results from EMSA, ChIP, andreporter assays suggest that PLK1 is a transcriptional targetof RelA and that upregulation of PLK1 induced by celldetachment is RelA dependent. Using 2 different highlyselective inhibitors of the NF-kB pathway, we showed thatRelA protected ESCC cells from anoikis by modulatingPLK1 and b-catenin expression. It is worth noticing thatPDTC or CAPE treatment suppressed PLK1 promoter activ-ity more strongly than deletion of PLK1-NE, suggesting thatNF-kB has other binding sites in PLK1 promoter region oreffectors downstream of the NF-kB pathway regulate PLK1transcriptional activity. Other regulators of the NF-kB path-way, such as lipopolysaccharide (LPS), TNF-a, or interleu-kin (IL)-1b, had no effect on PLK1 promoter activity(data not shown). Thus, the regulation of PLK1 by RelA,identified in our study, is a novel mechanism of thecomplex NF-kB signaling network. Interestingly, recentstudies have suggested that, through regulating NEMOand IkB kinase b, PLK1 inhibits NF-kB transcriptionalactivation induced by TNF-a or IL-1b, but not by RelA(36, 37). Altogether, these results suggest that there is nonegative feedback loop between RelA and PLK1 regulation.
In the present study, we found that elevated proteinlevels of RelA were present in 65.6% of tumor samples,indicating that alterations in RelA expression are frequentevents in ESCC. Analyses of both genomic copy numberand protein expression of RelA in the same cases showedthat gene amplification was one of the mechanisms under-lying RelA overexpression. Our recent findings that only37% of specimens with PLK1 protein upregulation exhib-ited gene amplification (9) suggest that transcriptionalactivation could be another mechanism of PLK1 overex-pression in some esophageal carcinomas. Notably, a sig-nificant correlation between RelA and PLK1 overexpressionwas observed in ESCC specimens. In view of our findingthat RelA is also required for PLK1 expression in attachedcells, these data indicate that RelA is an important tran-scriptional regulator of PLK1 and is responsible for PLK1overexpression in ESCC tissues.
In summary, we showed that upregulation of PLK1triggered by cell detachment is regulated by RelA at the
NF-kB Promotes Anoikis Resistance via PLK1/b-Catenin
www.aacrjournals.org Clin Cancer Res; 17(13) July 1, 2011 4293
transcriptional level. Induction of PLK1 protects esopha-geal carcinoma cells from anoikis through modulationof b-catenin protein levels by inhibiting their degrad-ation. Altogether, our study reveals novel mechanismsthat underlie PLK1 overexpression in ESCC and the roleof PLK1 in anoikis resistance.
Disclosure of Potential Conflicts of Interest
No potential conflicts of interest were disclosed.
Grant Support
This study was supported by the National Natural Science Foundation(grant nos. 30971482 and 81021061), the Special Public Health Fund(grant no. 200902002-4), and the National Hightech R&D Program ofChina (grant no. 2009AA022706).
The costs of publication of this article were defrayed in part by thepayment of page charges. This article must therefore be hereby markedadvertisement in accordance with 18 U.S.C. Section 1734 solely to indicatethis fact.
Received December 7, 2010; revised April 27, 2011; accepted May 11,2011; published OnlineFirst May 24, 2011.
References1. Hynes RO. Cell adhesion: old and new questions. Trends Cell Biol
1999;9:M33–7.2. Coucouvanis E, Martin GR. Signals for death and survival: a two-step
mechanism for cavitation in the vertebrateembryo.Cell 1995;83:279–87.3. Strange R, Li F, Saurer S, Burkhardt A, Friis RR. Apoptotic cell death
and tissue remodelling during mouse mammary gland involution.Development 1992;115:49–58.
4. Ikeda H, Suzuki Y, Suzuki M, Koike M, Tamura J, Tong J, et al.Apoptosis is a major mode of cell death caused by ischaemia andischaemia/reperfusion injury to the rat intestinal epithelium. Gut1998;42:530–7.
5. Hanahan D, Weinberg RA. The hallmarks of cancer. Cell 2000;100:57–70.
6. Winkles JA, Alberts GF. Differential regulation of polo-like kinase 1, 2,3, and 4 gene expression in mammalian cells and tissues. Oncogene2005;24:260–6.
7. Weichert W, Schmidt M, Gekeler V, Denkert C, Stephan C, Jung K,et al. Polo-like kinase 1 is overexpressed in prostate cancer and linkedto higher tumor grades. Prostate 2004;60:240–5.
8. Weichert W, Kristiansen G, Schmidt M, Gekeler V, Noske A, NiesporekS, et al. Polo-like kinase 1 expression is a prognostic factor in humancolon cancer. World J Gastroenterol 2005;11:5644–50.
9. Feng YB, Lin DC, Shi ZZ, Wang XC, Shen XM, Zhang Y, et al.Overexpression of PLK1 is associated with poor survival by inhibitingapoptosis via enhancement of survivin level in esophageal squamouscell carcinoma. Int J Cancer 2009;124:578–88.
10. Spankuch-Schmitt B, Bereiter-Hahn J, Kaufmann M, Strebhardt K.Effect of RNA silencing of polo-like kinase-1 (PLK1) on apoptosis andspindle formation in human cancer cells. J Natl Cancer Inst2002;94:1863–77.
11. Spankuch-Schmitt B, Wolf G, Solbach C, Loibl S, Knecht R, Stegm€ul-ler M, et al. Downregulation of human polo-like kinase activity byantisense oligonucleotides induces growth inhibition in cancer cells.Oncogene 2002;21:3162–71.
12. Cogswell JP, Brown CE, Bisi JE, Neill SD. Dominant-negative polo-like kinase 1 induces mitotic catastrophe independent of cdc25Cfunction. Cell Growth Differ 2000;11:615–23.
13. WangG, Brennan C, RookM,Wolfe JL, Leo C, Chin L, et al. Balanced-PCR amplification allows unbiased identification of genomic copychanges in minute cell and tissue samples. Nucleic Acids Res 2004;32:e76.
14. LuoML, Shen XM, Zhang Y,Wei F, Xu X, Cai Y, et al. Amplification andoverexpression of CTTN (EMS1) contribute to the metastasis ofesophageal squamous cell carcinoma by promoting cell migrationand anoikis resistance. Cancer Res 2006;66:11690–9.
15. Du XL, Yang H, Liu SG, Luo ML, Hao JJ, Zhang Y, et al. Calreticulinpromotes cell motility and enhances resistance to anoikis throughSTAT3-CTTN-Akt pathway in esophageal squamous cell carcinoma.Oncogene 2009;28:3714–22.
16. Elia AE, Rellos P, Haire LF, Chao JW, Ivins FJ, Hoepker K, et al. Themolecular basis for phospho dependent substrate targeting andregulation of Plks by the Polo-box domain. Cell 2003;115:83–95.
anoikis, and radiation-induced cell cycle arrest. J Cell Biol 1999;146:855–68.
18. Hart M, Concordet JP, Lassot I, Albert I, del los Santos R, Durand H,et al. The F-box protein beta-TrCP associates with phosphorylatedbeta-catenin and regulates its activity in the cell. Curr Biol 1999;9:207–10.
20. Winston JT, Strack P, Beer-Romero P, Chu CY, Elledge SJ, HarperJW. The SCFbeta-TRCP-ubiquitin ligase complex associates speci-fically with phosphorylated destruction motifs in IkappaBalpha andbeta-catenin and stimulates IkappaBalpha ubiquitination in vitro.GenesDev 1999;13:270–83.
21. Fan Y, Mao R, Zhao Y, Yu Y, Sun W, Song P, et al. Tumor necrosisfactor-alpha induces RelA degradation via ubiquitination at lysine 195to prevent excessive nuclear factor-kappaB activation. J Biol Chem2009;284:29290–7.
22. Loza-Coll MA, Perera S, Shi W, Filmus J. A transient increase inthe activity of Src-family kinases induced by cell detachmentdelays anoikis of intestinal epithelial cells. Oncogene 2005;24:1727–37.
23. Wang P, Valentijn AJ, Gilmore AP, Streuli CH. Early events in theanoikis program occur in the absence of caspase activation. J BiolChem 2003;278:19917–25.
24. Yan SR, Joseph RR, Rosen K, ReginatoMJ, Jackson A, Allaire N, et al.Activation of NF-kappaB following detachment delays apoptosis inintestinal epithelial cells. Oncogene 2005;24:6482–91.
25. Liu Z, Li H, Derouet M, Berezkin A, Sasazuki T, Shirasawa S, et al.Oncogenic Ras inhibits anoikis of intestinal epithelial cells by pre-venting the release of a mitochondrial pro-apoptotic protein Omi/HtrA2 into the cytoplasm. J Biol Chem 2006;281:14738–47.
26. Liu Z, Li H, Derouet M, Filmus J, LaCasse EC, Korneluk RG, et al. rasOncogene triggers up-regulation of cIAP2 and XIAP in intestinalepithelial cells: epidermal growth factor receptor-dependent and-independent mechanisms of ras-induced transformation. J BiolChem 2005;280:37383–92.
27. Morin PJ, Sparks AB, Korinek V, Barker N, Clevers H, Vogelstein B,et al. Activation of beta-catenin-Tcf signaling in colon cancer bymutations in beta-catenin or APC. Science 1997;275:1787–90.
28. Rubinfeld B, Robbins P, El-Gamil M, Albert I, Porfiri E, Polakis P.Stabilization of beta-catenin by genetic defects inmelanoma cell lines.Science 1997;275:1790–2.
29. Wang Y, Zhou X, Zhu H, Liu S, Zhou C, Zhang G, et al. Overexpressionof EB1 in human esophageal squamous cell carcinoma (ESCC) maypromote cellular growth by activating beta-catenin/TCF pathway.Oncogene 2005;24:6637–45.
30. Wang Y, Liu S, Zhu H, Zhang W, Zhang G, Zhou X, et al. FRAT1overexpression leads to aberrant activation of beta-catenin/TCF path-way in esophageal squamous cell carcinoma. Intl J Cancer 2008;123:561–8.
31. Arai T, Haze K, Iimura-Morita Y, Machida T, Iida M, Tanaka K, et al.Identification of beta-catenin as a novel substrate of Polo-like kinase1. Cell Cycle 2008;7:3556–63.
Lin et al.
Clin Cancer Res; 17(13) July 1, 2011 Clinical Cancer Research4294
32. Chen YH, Lu Y, De Plaen IG, Wang LY, Tan XD. Transcription factorNF-kappaB signals antianoikic function of trefoil factor 3 on intestinalepithelial cells. Biochem Biophysic Res Commun 2000;274:576–82.
33. Gill JS, Windebank AJ. Ceramide initiates NFkappaB-mediatedcaspase activation in neuronal apoptosis. Neurobiol Dis 2000;7:448–61.
34. Yang CH, Murti A, Pfeffer SR, Kim JG, Donner DB, Pfeffer LM.Interferon alpha/beta promotes cell survival by activating nuclearfactor kappa B through phosphatidylinositol 3-kinase and Akt. J BiolChem 2001;276:13756–61.
35. Ryan KM, O’Prey J, Vousden KH. Loss of nuclear factor-kappaB istumor promoting but does not substitute for loss of p53. Cancer Res2004;64:4415–8.
36. ZhangW,Wang J, Zhang Y, Yuan Y, GuanW, Jin C, et al. The scaffoldprotein TANK/I-TRAF inhibits NF-kappaB activation by recruitingpolo-like kinase 1. Mol Biol Cell 21:2500–13.
37. Higashimoto T, Chan N, Lee YK, Zandi E. Regulation of I(kappa)Bkinase complex by phosphorylation of (gamma)-binding domain of I(kappa)B kinase (beta) by Polo-like kinase 1. J Biol Chem 2008;283:35354–67.
NF-kB Promotes Anoikis Resistance via PLK1/b-Catenin
www.aacrjournals.org Clin Cancer Res; 17(13) July 1, 2011 4295
