-
Cancer Therapeutics Insights
Targeting Blockage of STAT3 in Hepatocellular CarcinomaCells
Augments NK Cell Functions via ReverseHepatocellular
Carcinoma–Induced Immune Suppression
Xiaoxia Sun1, Qiangjun Sui1, Cai Zhang1, Zhigang Tian1,2, and
Jian Zhang1
AbstractSTAT3 is an important transcriptional factor for cell
growth, differentiation, and apoptosis. Although evi-
dence suggests a positive role for STAT3 in cancer, the
inhibitory effects of tumor STAT3 on natural killer (NK)
cell functions in human hepatocellular carcinoma are unclear. In
this study, we found that blocking STAT3 in
hepatocellular carcinoma cells enhanced NK-cell antitumor
function. In the case of STAT3-blocked hepatocel-
lular carcinoma cells, NKG2D ligandswere upregulated, which
promoted recognition byNK cells. Importantly,
the cytokine profile of hepatocellular carcinoma cells was
altered; in particular, TGF-b and interleukin 10
(IL-10)expressionwas reduced, and type I interferon (IFN)was
induced, thus facilitatingNK-cell activation. Indeed, the
cytotoxicity of NK cells treated with supernatant from
STAT3-blocked hepatocellular carcinoma cells was
augmented, with a concomitant elevation of molecules associated
with NK cytolysis. Further experiments
confirmed that the recovery of NK cells depended on the
downregulation of TGF-b and upregulation of type IIFN derived from
STAT3-blocked hepatocellular carcinoma cells. These findings
demonstrated a pivotal role
for STAT3 in hepatocellular carcinoma-mediated NK-cell
dysfunction, and highlighted the importance of
STAT3 blockade for hepatocellular carcinoma immunotherapy, which
could restore NK-cell cytotoxicity in
addition to its direct influence on tumor cells. Mol Cancer
Ther; 12(12); 2885–96. �2013 AACR.
IntroductionThe tumor microenvironment is a highly complex
milieu, consisting of cancer cells, stromal tissue (immunecells,
fibroblasts, myofibroblasts, cytokines, and vasculartissue), as
well as the surrounding extracellular matrix (1).These components
not only promote tumor progressionand metastasis, but also induce
immunosuppression andprotect tumor cells fromhost immuneattack
(2).Asamajorbarrier to cancer progression, the immune system
controlstumor elimination; however, it is modulated by factors
inthe tumor microenvironment. Accumulating evidenceshows that
natural killer (NK) cells play a critical role intumor
immunosurveillance and act as the first line of thedefense against
tumor cells. Enhanced NK-cell infiltrationof the tumor site will
recruit host immune defense factorsand inhibit tumorigenesis.
Nevertheless, in many patients
with tumors, such as breast, lung, colorectal, and livercancers,
both peripheral blood NK cells and tumor-asso-ciatedNKcells exhibit
an alteredphenotype andprofounddefects in their degranulation
ability and interferon gam-ma (IFN-g) production (3–5). Inaddition,
it is indicated thatimmunomodulatory factors in the tumor
microenviron-ment, including interleukin 4 (IL-4), IL-10, TGF-b,
andindoleamine 2,3-dioxygenase (IDO), contribute to thedecreased
expression of stimulatory receptors (NKG2D,DNAM-1, NKp30, and
NKp44) and impaired secretion ofcytotoxic proteins (perforin,
granzymes) and IFN-g byNKcells, which limit the NK-cell–mediated
antitumor effectandpromote tumor immune evasion (1, 6, 7).
Therefore, anideal antitumor immunotherapy should improve thetumor
microenvironment and augment NK-cell functionswhen applied in the
clinic.
STAT3 is a key transcriptional regulator of genes thatcontrol
cell growth and differentiation, including bcl-xl,cyclinD1,
c-myc,mcl-1,VEGF, IL-10, TGF-b and survivin (8–10). Constitutive
activation of STAT3 has been detected ina number of human primary
tumors and cancer cell lines,indicating that inhibition of
activated STAT3 would sup-press proliferation and induce apoptosis
of cancer cells(11–13). In addition, aberrant activation of STAT3
con-tributes to tumor immune evasion via restraining antitu-mor
immunity (14–16). Inactivation of STAT3 signaling inhematopoietic
cell-specific conditional knockout (CKO)mice (17),or by the use of
pharmacologic inhibitors, suchas JSI-124 (18) or CPA-7 (17),
resulted in enhanced
Authors' Affiliation: 1Institute of Immunopharmacology &
Immunothera-py, School of Pharmaceutical Sciences, Shandong
University, Shandong;and 2School of Life Sciences, University of
Science and Technology ofChina, Anhui, China
Note: Supplementary data for this article are available at
Molecular CancerTherapeutics Online
(http://mct.aacrjournals.org/).
Corresponding Author: Jian Zhang, Institute of
Immunopharmacology &Immunotherapy, School of Pharmaceutical
Sciences, Shandong Univer-sity, 44 Wenhua West Road, Jinan 250012,
China. Phone: 86-531-8838-3781; Fax: 86-531-8838-3782; E-mail:
[email protected]
doi: 10.1158/1535-7163.MCT-12-1087
�2013 American Association for Cancer Research.
MolecularCancer
Therapeutics
www.aacrjournals.org 2885
on April 2, 2021. © 2013 American Association for Cancer
Research. mct.aacrjournals.org Downloaded from
Published OnlineFirst October 9, 2013; DOI:
10.1158/1535-7163.MCT-12-1087
http://mct.aacrjournals.org/
-
antitumor immune responses through the activation ofvarious
immune cells such as dendritic cells (DCs) andcytotoxic T
lymphocyte (CTL), and inactivation of regu-latory T cell (Tregs;
refs. 17, 18).
It has been reported that the inactivation of STAT3 intumor
cells favored NK-cell- and T-cell–mediated anti-tumor immune
responses (19–21); however, whether NKcell function can be
augmented by blocking STAT3 inhuman hepatocellular carcinoma cells
has not been fullyinvestigated. In this study, we focus on
NK-cell–medi-ated antitumor immune responses to
STAT3-blockedhepatocellular carcinoma cells. The results showed
thatblocking STAT3 enhanced NK-cell antitumor functionnot only by
promoting recognition by NK cells, but alsoby elevating the
expression of molecules associated withNK-cell cytolysis.
Furthermore, both TGF-b reductionand type I IFN production by
STAT3-blocked hepato-cellular carcinoma cells were responsible for
NK-cellactivation. Collectively, our studies demonstrated a
piv-otal role for STAT3 aberrant signaling in
hepatocellularcarcinoma-mediated NK cell dysfunction; in
addition,blocking STAT3 in hepatocellular carcinoma cells notonly
promoted hepatocellular carcinoma apoptosis, butalso improved NK
cell cytotoxicity, which limit/preventtumor cell escape from
antitumor immunity.
Materials and MethodsCell lines and primary cell cultures
Hepatocellular carcinoma cell lines HepG2 (The CellBank of Type
Culture Collection of the Chinese Academyof Sciences, TCHu 72),
H7402 (Institute of Basic MedicalSciences, Shandong Academy of
Medical Science, China),and PLC/PRF/5 (kindly supplied by Dr. Qu
Xianjun,Department of Pharmacology, Shandong University)were grown
in RPMI-1640 medium (GIBCO/BRL) sup-plemented with 10% FBS. The
human NK cell line NKLwas generously provided by Dr. Jin Boquan
(FourthMilitary Medical University, China) and cultured
inRPMI-1640, containing 10% FBS and 100 U/mL rhIL-2(Changsheng,
Changchun, China). The human NK cellline NK-92 was purchased from
the American Type Cul-ture Collection (CRL-2407) and maintained in
a-Mini-mum Essential Medium (MEM; GIBCO/BRL) supple-mented with
12.5% horse serum (GIBCO), 12.5% FBS,100 U/mL rhIL-2, 0.1 mmol/L
b-mercaptoethanol and0.02 mmol/L folic acid. All cells were
cultured in ahumidified incubator with 5% CO2 at 37
�C according tothe vendor’s instruction, and used within 6
months afterreceipt or resuscitation. Healthy peripheral blood
mono-nuclear cells (PBMC) were isolated by Ficoll densitygradient
centrifugation and cultured in RPMI-1640 con-taining 10% FBS and
100 U/mL rhIL-2. Informed consentwas provided by all participants
enrolled in this study.
STAT3 decoy/scramble oligodeoxynucleotideUsing phosphorothioate
chemistry, sense and antisense
strands of STAT3decoy or scramble oligodeoxynucleotide
(ODN) were synthesized by the Expedite Nucleic AcidSynthesis
System (Takara Biotechnology). STAT3 decoyODN sequence was
50-CATTTCCCGTAAATC-30, 30-GTAAAGGGCATTTAG-50 and scramble ODN
sequencewas 50-CATCTTGCCAATATC-30, 30-GTAGAACGGT-TATAG-50. The
sense and antisense strands wereannealed and purified by
high-performance liquid chro-matography (HPLC; ref. 22).
Cell transfectionTransient transfections were carried out with
Lipofec-
tamine 2000 (Invitrogen) according to the
manufacturer’sinstructions. One day before transfection, human
livercancer cells were seeded to ensure greater than 90%confluence.
After being washed with PBS, cells weretransfected with
Lipofectamine 2000/decoy ODN orLipofectamine 2000/scramble ODN. For
transfection ofoligonucleotides, 25 nmol/L STAT3 decoy or
scrambleODN were used, with an ODN (mg) to Lipofectamine2000 (mL)
ratio of 1: 2.5.
NK cells treated with the supernatant from tumorcells
After transfection with ODN for 6 hours, the
transfec-tionmediumwas removed, and hepatocellular carcinomacells
were washed with 1� PBS to remove residual ODN.Hepatocellular
carcinoma cells were then plated at adensity of 1 � 104 cells/well
in 96-well plates, and cul-tured in fresh medium (containing 10%
FBS) for a further24 hours. Supernatantswere collected and cells
anddebrisremoved by centrifugation. NK cells were cultured in
thesoluble supernatant for 12 hours with ratio of supernatantto
medium (v/v) of 1:1.
Flow cytometryThe phenotype of cells was analyzed by flow
cytome-
try. The fluorescence-conjugated antibodies used in thisstudy
are described in Supplementary Table S1. Cellswere incubated with
fluorescence-conjugated antibodiesfor 30 minutes at 4�C. For
detection of intracellular cyto-kines, cells were stimulated for 4
hours with monensin(6 mmol/L) and ionomycin (1 mg/mL; Sigma
ChemicalCo.) in a 37�C, 5%CO2 incubator, and thenwashed, fixed,and
permeabilized. Cells were then stained with a satu-rating amount of
the fluorescence-conjugated antibodiesfor 1 hour at 4�C. After
washing with PBS, stained cellswere acquired using a FACSCalibur
system (BD Bios-ciences) and analyzed with WinMDI 2.0 software.
RNA isolation and quantitative real-time PCRQuantitative
real-time PCR (qRT-PCR) was performed
according to the manufacturer’s instructions. Briefly, totalRNA
was isolated using the TRIzol Reagent (Invitrogen).Thequality
andconcentrationof
theRNAweredeterminedbyspectrophotometricmeasurementof theA260/A280
ratio.cDNA was synthesized using the M-MLV reverse tran-scriptase
(Invitrogen). qRT-PCR was performed usingthe SYBR Green Real-time
PCR Kit (TOYOBO) on an iQ5
Sun et al.
Mol Cancer Ther; 12(12) December 2013 Molecular Cancer
Therapeutics2886
on April 2, 2021. © 2013 American Association for Cancer
Research. mct.aacrjournals.org Downloaded from
Published OnlineFirst October 9, 2013; DOI:
10.1158/1535-7163.MCT-12-1087
http://mct.aacrjournals.org/
-
Real-Time PCR detection system (Bio-Rad). RelativemRNA
expression levels of the gene of interest were cal-
culatedusing the2�DDCt method. Theprimers aredescribedin
Supplementary Table S2.
Luciferase reporter gene assayFor the reporter gene assay,
hepatocellular carcinoma
cells were plated at a density of 1� 104 cells per well in
96-well plates (Costar), and transiently cotransfected
withpGL3-TGFb1-Promoter-Luciferase or
pGL3-IL10-Promot-er-Luciferase (200 ng/well) and STAT3-decoy or
scrambleODN in the presence of Lipofectamine 2000, whereas
theRenilla expression vector pRLTK (20 ng/well, Promega)was
cotransfected to normalize the transfection efficiency.After 24
hours, cells were washed, lysed, and a dual-GloLuciferase assay
system (Promega) was used to determineluciferase activity according
to themanufacturer’s instruc-tions. The ratio of firefly and
Renilla luciferase activityassociated with pGL3-TK-Luciferase
transfection was setas 1.
ELISATwenty-four hours after transfection with decoy ODN
or scramble ODN, the supernatant from human livercancer cell
culture was collected and centrifuged toremove cells and debris.
The concentrations of TGF-b,IL-8, and IL-10 were determined using
ELISA kits (ExCellBiology, Inc). To determine the level of IFN-g
produced byNK cells, NK-92 cells were incubated both in the
presenceand absence of the supernatant from hepatocellular
car-cinoma cells. After 12 hours, these cells were
centrifuged,resuspended in fresh medium, and cultured for 48
hoursbefore the concentration of IFN-g was measured using anELISA
kit.
Cytokines and blocking antibodiesAnti-hULBP3 monoclonal antibody
(mAb; 30 mg/mL;
R&D systems, Inc.), recombinant human TGF-b1 (2.5ng/mL,
PeproTech), anti-human LAP (TGF-b1) anti-body (2 mg/mL, R&D
systems, Inc.), IFN-a and IFN-b(400 U/mL, Changsheng Life Sciences
Ltd), and anti-IFNAR mAb (10 mg/mL, PBL) were used in
neutraliza-tion assays.
Cell viability assayHepatocellular carcinoma cells treated in
the presence
or absence of ODN were plated at a density of 1 � 104cells/well
in 96-well plates. After 4 hours, NK-92,NKL, orprimaryNKcellswere
added into the plates at different E:T values. At the 12-hour
time-point, 10 mL (5 mg/mL)MTT (Sigma) was added and incubated for
another 4hours. After centrifugation, 100 mL of the
supernatantswere removed from each well, and 100 mL of 10%
SDSsolution was added to dissolve the formazan crystals
andincubated at 37�C, in 5% CO2 overnight. Then, the absor-bance at
570 to 630 nmwas determined using aMicroplateAutoreader
(Bio-Rad).
Cell cytotoxicity assayCell cytotoxicity againstHepG2 cellswas
evaluated in a
4-hourCFSE/7-AADflowcytometry assay.After labelingwith
5-6-carboxyfluo-rescein diacetate succinimidyl ester(CSFE,
Beyotime) for 15 minutes at 37�C, CFSE-labeledhepatocellular
carcinoma cells were washed with medi-um and seeded in
completemedium for adherent culture.Then, NK-92 cells were added
with E:T at 5:1. Hepato-cellular carcinoma cells were incubated
alone to measurebasal cell death. After 4 hours, cells were
collected andwashed twice with 1� PBS and incubated with
7-aminoactinomycin D (7-AAD, KeyGEN BioTECH) for 15 min-utes at
room temperature in darkness. Acquisition wasperformed with a
FACSCalibur system (BD Biosciences).The following formulawas used
to calculate specific lysis.Ratio ¼ % CFSEþ 7-AADþ/% CFSEþ,%
specific lysis ¼(the ratio of sample � ratio of basal) � 100
Statistical analysisAll data are presented as the means � SD of
three or
more independent experiments. Statistical analysis wasperformed
using a paired Student t test. P < 0.05 wasconsidered
statistically significant.
ResultsBlocking STAT3 in hepatocellular carcinomaaugmented the
susceptibility of hepatocellularcarcinoma to NK-cell cytolysis
We have previously shown that blocking STAT3 sup-pressed the
growth and promoted the apoptosis of hepa-tocellular carcinomacells
(13). The resistance of tumor cellsto NK cell cytolysis contributed
to antitumor immunesuppression. Therefore, we investigated the
effects ofblocking activated STAT3 on the sensitivity of
hepatocel-lular carcinoma cells to NK cell cytolysis. In Fig. 1A
and B,human NK cell lines NK-92 and NKL as well as
humanPBMCswereusedaseffector cells, respectively.Comparedwith
theLipo-Ctrl (Lipofectamine reagent control) groups,the viability
of decoy ODN-treated hepatocellular carci-noma cells exposed toNK
cells for 12 hourswas decreasedsignificantly, with minimum
viability of 41% to 18%. Inaddition, NK-92-cell–mediated specific
cell lysis againsthepatocellular carcinoma cells was evaluated in a
4-hourCFSE/7-AAD flow cytometry assay. As shown in Fig. 1C,NK-92
cell cytotoxicity against decoy ODN-treatedhepatocellular carcinoma
was significantly higher at5:1 E:T ratios in comparison with
Lipo-Ctrl groups. Incontrast, scramble ODN treatment did not
influence theviability or sensitivity of hepatocellular carcinoma
toNK cell lysis. These data indicated that blocking STAT3in
hepatocellular carcinoma enhanced the sensitivity ofhepatocellular
carcinoma cells to NK cell–mediatedcytolysis.
The NKG2D ligands expressed on hepatocellularcarcinoma cellswere
upregulated by blocking STAT3
As a result of "genomic stress", NKG2D ligands, whichare rarely
detectable on the surface of healthy cells and
Blocking STAT3 in Hepatocellular Carcinoma Cells Restored NK
Activation
www.aacrjournals.org Mol Cancer Ther; 12(12) December 2013
2887
on April 2, 2021. © 2013 American Association for Cancer
Research. mct.aacrjournals.org Downloaded from
Published OnlineFirst October 9, 2013; DOI:
10.1158/1535-7163.MCT-12-1087
http://mct.aacrjournals.org/
-
tissues, are upregulated in tumor cell lines and tumorissues,
rendering the tumor cell susceptible to NK cellcytotoxicity (23).
However, cancer cells have also devel-oped strategies to evade
NKG2D-mediated responses.Cumulative evidence indicates that the
expression levelsof NKG2D ligands on tumor cells are important in
NK-cell–mediated antitumor effects. Therefore, we examinedchanges
in the expression of NKG2D ligands in STAT3-blocked hepatocellular
carcinoma cells. As shown in Fig.2A, blockingSTAT3caused
significantupregulation in theexpression levels of ULBPs,
especially ULBP3. Althoughthe experimental findings showed that
STAT3 directlyinteracts with the MICA promoter to repress MICA
tran-scription in colorectal cancer (21), there were no
obviouschanges in the levels of MICA/B in HepG2 cells treatedwith
STAT3-decoy ODN compared with that in Lipo-Ctrlcells. To further
determine the function of NKG2Dligands, an anti-hULBP3 mAb (40
mg/mL) was used toblock ULBP3 expressed on HepG2 cells, and the
viabilityof these tumor cells in the presence of NK cells
wasmeasured (Fig. 2B). ULBP3 blockade significantlyenhanced the
viability of STAT3-blocked hepatocellularcarcinoma cells,
suggesting that upregulation of NKG2Dligands, especiallyULBP3,
contributed to the sensitivity ofSTAT3-blocked hepatocellular
carcinoma cells to NK-cellcytotoxicity.
Blocking STAT3 reversed hepatocellular carcinoma-induced immune
suppression on NK cells.
Impairment of NK cell function contributes to antitu-mor immune
inefficiency in patients with hepatocellularcarcinoma (5), which is
partly mediated by immunosup-pressive factors secreted by tumor
cells. Then, to clarifywhether hepatocellular carcinoma-mediated NK
cell dys-function could be recovered by blocking STAT3, humanNK
cells were incubated with the supernatants fromhuman hepatocellular
carcinoma cells treated with orwithout STAT3-decoy ODN for 12
hours, and the influ-ence of these NK cells on the viability of
hepatocellularcarcinoma cells was then examined. As shown in Fig.
3A,after incubation with supernatants from the Lipo-Ctrl orscramble
ODN-treated hepatocellular carcinoma cells,NK cell–mediated
anti-hepatocellular carcinoma effectwas suppressed compared with
that of untreated NKcells (Medium); however, it was enhanced when
NK cellswere incubated with the supernatant from
hepatocellularcarcinoma cells treated with STAT3 decoy ODN,
withenhanced levels exceeding those mediated by untreatedNK cells
(Medium). Similar results were observed whenhuman PBMCswere used as
effector cells. The viability ofHepG2 cells exposed to healthy
PBMCs was 48.95% �0.32%, which was decreased to 30.37% � 0.53% in
thepresence of PBMCs treated with the supernatant from
020406080
100
5:1 2.5:1 1:1
Via
bili
ty (
%)
E:T
**
**
100
020406080
5:1 2.5:1 1:1
Via
bili
ty (
%)
E:T
****
*
0
20406080
100
5:1 2.5:1 1:1
Via
bili
ty (
%)
E:T
****
*
H7402
020406080
100
5:1 2.5:1 1:1
viab
ility
(%
)
E:T
**
*NK
L
020406080
100
5:1 2.5:1 1:1vi
abili
ty (
%)
E:T
****
*
HepG2
NK
-92
A
Lipo-CtrlScrambleDecoy
0
10
20
30
40
HepG2 H7402 PLC/PRF/5
% S
pec
ific
lysi
s ****
**
C
020406080
100
5:1 2.5:1 1:1
Via
bili
ty (
%)
E:T
**
***
PLC/PRF/5 Lipo-CtrlScrambleDecoy
hPBMC
0
20
40
60
80
100
50:1 25:1 12.5:1
Via
bili
ty (
%)
E:T
***
Lipo-CtrlScrambleDecoy
B
Figure 1. Blocking STAT3 in hepatocellular carcinoma augmented
the susceptibility of hepatocellular carcinoma cells to NK cell
cytolysis. As describedin the Materials and Methods section,
hepatocellular carcinoma cells were transfected with STAT3 decoy
ODN, scramble ODN, or Lipofectaminereagent control (Lipo-Ctrl) and
used as target cells. The viabilities of these cells in the
presenceofNK-92/NKL cells (A) andhumanPBMCs (B)were detected byMTT
assay with different E:T ratios. C, the sensitivities of these
cells to NK-92 cell cytolysis were evaluated in a 4-hour CFSE/7-AAD
flow cytometryassay with an E:T ratio of 5:1. Data are
representative of three independent experiments, and statistical
significance was determined as �, P < 0.05 and��, P < 0.01
compared with the Lipo-Ctrl.
Sun et al.
Mol Cancer Ther; 12(12) December 2013 Molecular Cancer
Therapeutics2888
on April 2, 2021. © 2013 American Association for Cancer
Research. mct.aacrjournals.org Downloaded from
Published OnlineFirst October 9, 2013; DOI:
10.1158/1535-7163.MCT-12-1087
http://mct.aacrjournals.org/
-
STAT3-blocked HepG2 cells (Fig. 3B). These observationswere
confirmed using NSC74859, a chemical probe inhib-itor of STAT3
activity to block STAT3 in hepatocellularcarcinoma cells
(Supplementary Fig. S1). These findingsindicated that blocking
STAT3 in hepatocellular carcino-ma reversed hepatocellular
carcinoma-mediated immune
suppression and promoted NK-cell–mediated antitumoreffects.
NK cells were activated by the products secreted bySTAT3-blocked
hepatocellular carcinoma cells
NK cell cytolysis is determined by the balance betweeninhibitory
and activatory signals. When activated, NKcells attack a target via
manymechanisms, such as releaseof cytolytic granules (e.g.,
perforin and granzymes), anti-body-dependent cell-mediated
cytotoxicity (ADCC), pro-duction and secretion of cytokines (e.g.,
IFN-g , TNF-a;ref. 24), as well as evasion of "self" by
downregulation ofMHC-I molecules (25). In order to investigate the
char-acteristics of NK cells treated as described, total mRNA
ofNK-92 cells was extracted and the molecules
associatedwithNKactivationwere analyzedbyqRT-PCR.As shownin Fig. 4A
and Supplementary Fig. S2A, compared withthe Lipo-Ctrl and Scramble
groups, we found that mRNAlevels ofmolecules related to cytolysis,
includingNKG2D,IFN-g , FasL, perforin, granzyme B, and TNF-a,
wereincreased obviously in NK-92 cells incubated with
thesupernatant from STAT3-blocked hepatocellular carcino-ma cells,
as well the protein levels detected by ELISAand flow cytometric
analysis (Fig. 4B and 4C and Sup-plementary Fig. S2B). However, the
protein level ofNKG2A was decreased in NK-92 cells by treatmentwith
the supernatant from STAT3 decoy ODN-treatedhepatocellular
carcinoma cells (Fig. 4C and Supplemen-tary Fig. S2B). Furthermore,
as shown in Fig. 4D, similarresults were observed when NK-92 cells
were replacedwith human primary CD3�CD56þ PBMCs, and thelevels of
activation-related molecules—CD69, NKG2D,and perforin—on these
cells increased obviously com-pared with the Lipo-Ctrl and Scramble
groups. Thesefindings demonstrated that NK cells were activated
asSTAT3 in hepatocellular carcinoma cells was blocked,and
implicated that the products secreted by STAT3-blocked
hepatocellular carcinoma cells in the underly-ing mechanisms.
The cytokine profile of hepatocellular carcinomacells was
changed by blocking STAT3
In the tumor microenvironment, tumor cells suppressimmune
surveillance by secreting proinflammatory cyto-kines and
immunosuppressive cytokines (26, 27). There-fore, we aimed to
determine the effects of blocking STAT3on the cytokine profile of
hepatocellular carcinoma cells.As shown in Fig. 5A, in
STAT3-blocked hepatocellularcarcinoma cells, the mRNAs levels of
inflammatory cyto-kines, including IL-6, -8, -18, -17, and -23, as
well as theimmunosuppressive factor TGF-b were
downregulated,whereas the mRNA levels of IFN-a, -b, and -g
wereupregulated. In the tumor microenvironment, IL-8, IL-10, and
TGF-b are known to function as immunosuppres-sive factors and exert
positive effects on tumor cells (28,29). ELISA analysis of the
protein levels showed lowerlevels of IL-8, IL-10, andTGF-b inHepG2
cells treatedwiththe STAT3 decoy ODN (Fig. 5B). Similar results
were also
Lipo-CtrlScrambleDecoy
0
5
10
15
20
25
30
% G
ated
*
*
**
Lipo-CtrlDecoy+IgG Decoy+a-ULBP3
ULBP-2ULBP-1 ULBP-3
0
20
40
60
80
2.5:15:1
**
NS
**** *
NS
Via
bili
ty (
%)
Scramble
E:T
ULBP-1
Eve
nts
Eve
nts
Eve
nts
ULBP-2 ULBP-3
FasMICA/B100 101
100 101 100 101 100 101
100 101
A
B
Lipo-CtrlScrambleDecoy
010
5E
ven
ts0
121
Eve
nts
012
8
010
5
010
5
Figure 2. The NKG2D ligands expressed on hepatocellular
carcinomacells were upregulated by blocking STAT3. A, twenty-four
hours aftertransfection with STAT3-decoy ODN, scramble ODN, or
Lipofectaminereagent control (Lipo-Ctrl), the expression of NKG2D
ligands and Fas onHepG2 cells was determined by flow cytometric
analysis. Data arerepresentative of three independent experiments,
and statisticalsignificance was determined as P < 0.05 (�) and P
< 0.01 (��) comparedwith the Lipo-Ctrl. B, STAT3 decoy
ODN-treated HepG2 cells wereincubated with anti-hULBP3 mAb (30
mg/mL; Decoyþa-ULBP3) orimmunoglobulin G (DecoyþIgG) for 1 hour
before assay of the viability inthepresenceofNK-92cells. Data are
representativeof three independentexperiments; statistical
significance was determined as P < 0.05 (�) andP < 0.01 (��)
compared with the indicated group. NS, no significance.
Blocking STAT3 in Hepatocellular Carcinoma Cells Restored NK
Activation
www.aacrjournals.org Mol Cancer Ther; 12(12) December 2013
2889
on April 2, 2021. © 2013 American Association for Cancer
Research. mct.aacrjournals.org Downloaded from
Published OnlineFirst October 9, 2013; DOI:
10.1158/1535-7163.MCT-12-1087
http://mct.aacrjournals.org/
-
found in the other human hepatocellular cell lines H7402and
PLC/PRF/5.
TGF-b1 reduction and type I IFN production bySTAT3-blocked
hepatocellular carcinoma cells werecritical for NK-cell
activation
As a multifunctional cytokine, TGF-b plays importantroles in
tumor immune evasion (30). It is overproduced inthe serum of
patients with cancer and is linked withreduced NK-cell activity
(3). Furthermore, IL-10, whichis also associated with tumor
malignancy via immuneescape, is involved in the phosphorylation of
JAK2/STAT3 (8). In order to determine whether the productionof
TGF-b1 and IL-10 was responsible for the suppression
in NK-cell function, first, the regulatory effects of STAT3on
the transcriptional activities of the TGF-b1 and IL-10promoters
were identified by a dual-luciferase assay (Fig.6Aand6B).Next,NK-92
cellswere incubatedwithHepG2cell supernatant with or without a
TGF-b1 neutralizingantibody, and the effects on the viability of
HepG2 cellswere evaluated. The results showed that the viability
ofHepG2 cells in the presence of NK-92 cells incubatedwithHepG2
cell supernatant was 52.42% � 1.09%, which wasdecreased to 40.63% �
0.55% in the presence of theneutralizing TGF-b1 (Fig. 6C). The
viability of HepG2cells in the presence of NK-92 cells incubated
with thesupernatant from STAT3 decoy ODN-treated hepatocel-lular
carcinoma cells was 22.05% � 2.62%, which was
PL
C/P
RF
/5
0
20
40
60
80
100
2.5:15:1 1:1
MediumLipo-CtrlScrambleDecoy
NKL
Via
bili
ty (
%)
E:T
**
0
20
40
60
80
100
2.5:15:1 1:1
Via
bili
ty (
%)
E:T
**
0
20
40
60
80
100
2.5:15:1 1:1
H74
02
Via
bili
ty (
%)
E:T
**
****
0
20
40
60
80
100
2.5:15:1 1:1
Via
bili
ty (
%)
E:T
**
****
0
20
40
60
80
100
2.5:15:1 1:1
Via
bili
ty (
%)
E:T
**
0
20
40
60
80
100
2.5:15:1 1:1
NK-92
Hep
G2
Via
bili
ty (
%)
E:T
****
*
A
100
0
20
40
60
80
12.5:125:150:1
****
**
B hPBMC MediumLipo-CtrlScrambleDecoy
E:T
Via
bili
ty (
%)
Figure 3. Blocking STAT3 inhepatocellular carcinoma
cellstriggered the antitumor effects ofNK cells. A and B,
afterhepatocellular carcinoma cellswere transfected with
STAT3-decoy ODN, scramble ODN, orLipofectamine reagent
(Lipo-Ctrl)as described in the Materials andMethods section, the
supernatantswere collected and used to cultureNK-92/NKL cells (A)
or humanPBMCs (B) for 12 hours. Theinhibitory effect of theseNK
cells orPBMCs on the viability of untreatedhepatocellular carcinoma
wasanalyzed by MTT assay. Data arerepresentative of
threeindependent experiments;statistical significance wasdetermined
as P < 0.05 (�) andP < 0.01 (��) compared with
theLipo-Ctrl.
Sun et al.
Mol Cancer Ther; 12(12) December 2013 Molecular Cancer
Therapeutics2890
on April 2, 2021. © 2013 American Association for Cancer
Research. mct.aacrjournals.org Downloaded from
Published OnlineFirst October 9, 2013; DOI:
10.1158/1535-7163.MCT-12-1087
http://mct.aacrjournals.org/
-
reversed by the addition of TGF-b1. In addition, flowcytometry
assay with CFSE/7-AAD was employed toevaluate NK-92-cell–mediated
specific cell lysis againsthepatocellular carcinoma cells, and
consistent findingswere observed (Fig. 6D). In contrast, scramble
ODN treat-mentdidnot influence theNK-92 cell cytolysis.
Thesedataindicated that TGF-b1 acted as an important factor in
thehepatocellular carcinoma-mediated immunosuppressiveeffects on NK
cells.Nevertheless, as shown in Fig. 6C and 6D, the anti-
hepatocellular carcinoma effects of NK cells treated
withhepatocellular carcinoma supernatant containing anti-TGFb1 mAb
was still lower than that of NK-92 cellsincubated with supernatant
from the STAT3 decoyODN-treated hepatocellular carcinoma cells. We
specu-lated that some activating cytokines are induced in
tumorcells by STAT3 blockade, which stimulated the
cytolyticcapacity of NK cells. The levels of IFN-a, -b, and -g
wereupregulated (Fig. 5A), similarly as the p-STAT1 levels
andtranscriptional activity of STAT1 in HepG2 cells
beingincreasedby STAT3decoyODN(Fig. 7A). Investigation of
the role of interferons by antibody neutralization showedthat
anti-IFNAR mAb suppressed the anti-hepatocellularcarcinoma
activities of NK cells induced by the superna-tant from
STAT3-blocked hepatocellular carcinoma cells(Fig. 7B and 7C).
Concomitant downregulation of mole-cules associated with NK
activation was observed (Fig.7D). Furthermore, the cytolytic
activity of NK cells wasrestored by the addition of type I IFN
added to thesupernatant of hepatocellular carcinoma cells
(from18.24% � 1.1% to 33.57% � 1.65%; Fig. 7C). These
resultsdemonstrated that both TGF-b1 reduction and type I
IFNproduction were essential for blocking STAT3-inducedNK-cell
activation.
DiscussionAs an oncogene, aberrant STAT3 activation has been
detected in many tumors, controlling differentiation,
pro-liferation, survival, angiogenesis, and immune functionduring
tumorigenesis (15). Furthermore, increasing evi-dence demonstrates
that tumor-induced abnormalities in
C D CD3- CD56+ PBMCNK-92
NKG2A NKG2D perforin GZM B0
20
40
60
80
100
% G
ated
**
****
CD69 NKG2A NKG2D Perforin IFN-g0
20
40
60
80
100%
Gat
ed
***
* *
**
A BLipo-CtrlScrambleDecoy
0
0.5
1.0
1.5
2.0
2.5
3.0
NKG2A NKG2D FasL GraB IFN-g perforin
****
**
**
****
**
No
rmal
ized
fo
ld e
xpre
ssio
n Lipo-CtrlScrambleDecoyMedium
10
20
30
40
50
60
70
H7402HepG2 PLC/PRF/5
IFN
-g (p
g/m
L)
**
**
**
NKG2D NKG2DCD69Granzyme B PerforinNKG2A NKG2A Perforin
MediumLipo-CtrlScrambleDecoy
MediumLipo-CtrlScrambleDecoy
Lipo-CtrlScrambleDecoyMedium
Lipo-CtrlScrambleDecoyMedium
100 101 100 101 100 101 102 100 101 102 100 101 102 100 101 102
100 101 102 103 100 101 102 103
Eve
nts
Eve
nts
0
1
28
0
1
28
0
1
28
0
1
28
0
40
0
40
0
40
0
47
Eve
nts
Eve
nts
Eve
nts
Eve
nts
Eve
nts
Eve
nts
Figure 4. NK cells were activated by the products secreted by
STAT3-blocked hepatocellular carcinoma cells. The supernatants,
collected from HepG2 cellstreated with STAT3-decoy ODN, scramble
ODN, or Lipofectamine reagent (Lipo-Ctrl) as described in the
Materials and Methods section, were used toincubate NK-92 cells for
12 hours. The molecules associated with NK-cell cytolysis were
analyzed by qRT-PCR (A), ELISA (B), and flow cytometry (C).D, human
PBMCs were incubated with the supernatants from HepG2 cells treated
as described earlier. Twelve hours later, the levels of molecules
related toNK-cell activation and cytolysis were analyzed by flow
cytometry. The histogram represents statistical analysis of the
percentage of positive cells. Data arerepresentative of three
independent experiments; statistical significance was determined as
P < 0.05 (�) and P < 0.01 (��) compared with the
Lipo-Ctrl.
Blocking STAT3 in Hepatocellular Carcinoma Cells Restored NK
Activation
www.aacrjournals.org Mol Cancer Ther; 12(12) December 2013
2891
on April 2, 2021. © 2013 American Association for Cancer
Research. mct.aacrjournals.org Downloaded from
Published OnlineFirst October 9, 2013; DOI:
10.1158/1535-7163.MCT-12-1087
http://mct.aacrjournals.org/
-
the immune system promote tumor progression andimmune evasion.
Therefore, a better identification andunderstanding of the
factor(s) involved in this abnormal-ity is the crux of this
problem.
In a previous study, we observed that blocking STAT3activation
efficiently inhibited the growth of hepatocellu-lar carcinoma cells
(13). Nevertheless, aberrant activationof STAT3 is closely related
with immunosuppression andtumor evasion (14–16). Therefore, in this
study, we aimedto determine whether targeted blockade of STAT3
inhuman hepatocellular carcinoma cells would break hepa-tocellular
carcinoma-mediated antitumor immune sup-pression and improve the
function of NK cells.
Decoy ODNs, specifically designed to compete withendogenous
cis-elements of the target gene, have beenproposed as a useful
approach to block the function oftranscription factors. Decoy
methods exhibit severalattractive advantages over other therapeutic
strategies.First, decoyODNs, as small DNAmolecules, can be
easilydelivered to target tissues and transfected into cells,
anddirectly abrogate the activated transcript factors. Second,a
growing number of transcription factors and their pro-moter
sequences have been characterized, which makespotential drug
targets plentiful and readily identifiable.Third, the synthesis,
storage, and transportation of decoyODNs are much simpler than
other approaches. More-over, decoy ODN strategies have been widely
used inthe laboratory and clinical studies of many diseases(11, 13,
19, 31). In this study, by fluorescence microscopy(Supplementary
Materials and Methods), the subcellulardistribution of STAT3 decoy
ODN in HepG2 cells was
shown to be mainly located in nucleus and lasted atleast 48
hours after transfection (Supplementary Fig.S3). In addition, this
was observed when using scrambleODN.
Our results showed that blocking STAT3 in hepatocel-lular
carcinoma augmented the susceptibility of hepato-cellular carcinoma
to NK-cell–mediated cytolysis (Fig. 1).The activatory and
inhibitory receptors present on NKcells are triggered during target
cell recognition andinduce a positive or a negative cell signaling
pathway,respectively; in addition, the balance between
activatoryand inhibitory signals determines the activation of
NKcells. When activated, NK cells eliminate their targetthrough the
release of cytotoxic enzymes (perforin, gran-zymes, granulysin)
and/or soluble factors (chemokinesand inflammatory cytokines),
which, in turn, recruit and/or activate other effectors, such as
promoting macro-phages phagocytosis and CD8þ T-cell cytotoxicity
(24).In contrast, inhibitory receptors on NK cells limit exces-sive
activation and regulate immune responses. Thesemolecules are the
targets of evasion strategies exerted bytumors. For instance, in a
breast tumor, impaired NK-cellfunction correlated with decreased
activatory NK cellreceptors (such as NKp30, NKG2D, DNAM-1, and
CD16)and increased inhibitory receptors (such as NKG2A) (3).In the
present study, we found that NK cells were acti-vated by incubation
with the supernatant from STAT3-blocked hepatocellular carcinoma
cells and accompaniedby enhancement of perforin, granzymes, IFN-g ,
NKG2D,and CD69, whereas the inhibitory receptor (NKG2A)
wasdecreased (Fig. 3 and 4).
A
No
rmal
ized
fo
ld
exp
ress
ion ** **
0
0.8
1.6
2.4
IL-6 IL-8 IL-18 IL-17 IL-23 TGF-ββ IFN-α IFN-β IFN-g
****
** * **
**
HepG2 Lipo-CtrlScrambleDecoy
H7402
020406080
600700
800900
pg
/mL
TGF-b1 IL-8 IL-10*
**
**
PLC/PRF/5
0
10
200400600800
pg
/mL
TGF-b1 IL-8 IL-10
**
**
*
B
TGF-b1 IL-8 IL-100
10203040
600
700
800
900
pg
/mL
*
**
**
HepG2
No
rmal
ized
fo
ld
exp
ress
ion **
**
0
1.0
2.0
3.0
IL-6 IL-8 IL-18 IL-17 IL-23 TGF-β IFN-α IFN-β IFN-g
* *** * **
**H7402 Lipo-Ctrl
ScrambleDecoy
PLC/PRF/5
**
No
rmal
ized
fo
ld
exp
ress
ion
**
0
1.0
2.0
3.0
4.0
IL-8 IL-18 IL-17 IL-23 TGF-β IFN-α IFN-β IFN-g****** **
**
*
Lipo-CtrlScrambleDecoy
Lipo-CtrlScrambleDecoy
Lipo-CtrlScrambleDecoy
Lipo-CtrlScrambleDecoy
Figure 5. The cytokine profile inhepatocellular carcinoma cells
waschanged by STAT3 decoy ODN.After transfection with STAT3decoy
ODN, scramble ODN, orLipofectamine reagent (Lipo-Ctrl)for 24 hours,
HepG2, H7402, orPLC/PRF/5 cells were harvestedand the levels of
tumor-relatedinflammatory cytokines wereexamined by qRT-PCR (A)
andELISA (B) analysis. Data arerepresentative of threeindependent
experiments;statistical significance wasdetermined as P < 0.05
(�) andP < 0.01 (��) compared withthe Lipo-Ctrl.
Sun et al.
Mol Cancer Ther; 12(12) December 2013 Molecular Cancer
Therapeutics2892
on April 2, 2021. © 2013 American Association for Cancer
Research. mct.aacrjournals.org Downloaded from
Published OnlineFirst October 9, 2013; DOI:
10.1158/1535-7163.MCT-12-1087
http://mct.aacrjournals.org/
-
The recognition between activatory receptors of NKcells and
ligands expressed on cancer cells is an importantmechanism by which
NK cells eliminate targets. Amongthe major activatory receptors
expressed on NK cells,NKG2D is implicated in the surveillance of
viral infectionand cancer. Despite DNA damage or heat-shock
proteins,which would actively upregulate the levels of NKG2Dligands
on cancer cells, many tumors have been shown toinhibit NKG2D ligand
expression by the production ofimmunomodulatory cytokines (21,
32–34). Hilpert andcolleagues demonstrated that the antitumor
activitymediated byNK cells was greatly dependent on the
levelsofNKG2D ligands expressed on the surface of tumor
cells,including MICA/B and ULBP1-3 (35). In this study, we
found that the expression levels of ULBPs, especiallyULBP3, were
obviously upregulated in hepatocellularcarcinoma cells by STAT3
blockade (Fig. 2A), which wasfavorable for NK-cell recognition and
enhanced the sus-ceptibility of hepatocellular carcinoma to NK-cell
cytoly-sis. Furthermore, our results showed that the
cytolyticactivity of NK cells against decoy ODN-treated
hepato-cellular carcinoma cells was suppressed by blocking
theexpression of ULBP3 (Fig. 2B). Although Bedel and col-leagues
demonstrated that STAT3 modulated MICAexpression in cancer cells
(21), the expression of MICA/B in hepatocellular carcinoma was not
changed signifi-cantly by blocking STAT3. It can be speculated that
this isbecauseNKG2D ligands are stress-response genes andare
0
Lu
ci a
ctiv
ity
fold
ind
uct
ion
0
0.2
0.4
0.6
0.8
1.0
1.2
Ctrl TGF-b1
**
ScrambleDecoy
A
Lu
ci a
ctiv
ity
fold
in
du
ctio
n
0
0.5
1.0
1.5
2.0
2.5
3.0 **
Ctrl IL-10
ScrambleDecoy
B%
Sp
ecif
ic ly
sis
10
20
30
40
50 ***
**
NS
D
C Medium IgG+BSA Lipo-Ctrl
Decoy +TGF-b1Scr+a-TGF-b1
Decoy Scr
*
Via
bili
ty (
%)
0
20
40
60
80
5:1 2.5:1
*
**** * *
E:T
Figure 6. TGF-b1 reduction by STAT3 decoy ODN-treated
hepatocellular carcinoma cells were critical for NK-cell
activation. A and B, HepG2 cells werecotransfected with
pGL3-TGFb1-Promoter-Luciferase, pGL3-IL10-Promoter-Luciferase, or
pGL3-TK-Luciferase and STAT3-decoy or scramble ODN in thepresence
of Lipofectamine 2000, whereas the Renilla expression vector pRLTK
was cotransfected to normalize the transfection efficiency. After
24 hours,luciferase activity of lysed cells was measured using a
dual-Glo Luciferase assay system. The ratio of firefly and Renilla
luciferase activity associatedwith pGL3-TK-Luciferase transfection
was set as 1. C and D, NK-92 cells were cultured with the
supernatants from HepG2 cells supplemented with orwithout
anti-TGF-b1 mAb (2 mg/mL) and the supernatants from STAT3 decoy
ODN-treated HepG2 cells in the presence or absence of TGF-b1 (2.5
ng/mL)for 12 hours. Subsequently, the inhibitory effect of these
NK-92 cells on HepG2 cell viability was detected by MTT assay (C),
and the specific lysis ofthese NK-92 cells against HepG2 cells was
detected by a 4-hour CFSE/7-AAD flow cytometry assay with E: T at
5:1 (D). Medium, NK cells cultured in ana-MEM without any
treatment; IgG þBSA, NK cells treated with IgG and BSA to exclude
the effects of IgG þBSA; Lipo-Ctrl, NK cells cultured with
thesupernatants fromLipofectamine reagent-treatedHepG2 cells; Scr,
NKcells culturedwith the supernatants from scrambleODN-treatedHepG2
cells; Scrþa-TGFb1, NK cells culturedwith the supernatants from
scramble ODN-treated HepG2 cells supplementedwith anti-TGF-b1mAb;
Decoy, NK cells culturedwiththe supernatants from STAT3 decoy
ODN-treated HepG2 cells; DecoyþTGFb1, NK cells cultured with the
supernatants from STAT3 decoy ODN-treatedHepG2 cells supplemented
with TGF-b1. Statistical significance was determined as P < 0.05
(�) and P < 0.01 (��); NS, no significance.
Blocking STAT3 in Hepatocellular Carcinoma Cells Restored NK
Activation
www.aacrjournals.org Mol Cancer Ther; 12(12) December 2013
2893
on April 2, 2021. © 2013 American Association for Cancer
Research. mct.aacrjournals.org Downloaded from
Published OnlineFirst October 9, 2013; DOI:
10.1158/1535-7163.MCT-12-1087
http://mct.aacrjournals.org/
-
upregulated by stimuli such as DNA damage and heat-shock. Thus,
the high basic level of MICA/B in hepato-cellular carcinoma may be
the result of the oncogenicprocess itself. These data confirmed
that blocking STAT3in hepatocellular carcinoma cells could reverse
the hepa-tocellular carcinoma-mediated inhibitory effects on
NKcells through augmenting the expression of activatoryreceptors on
NK cells and corresponding ligands onhepatocellular carcinoma
cells, resulting in the enhance-ment of NK cell–antitumor
functions.
Many factors regulated by STAT3 in tumor cells par-ticipate in
promoting malignant cell proliferation, migra-tion, and invasion.
In STAT3-targeted genes, IL-6, IL-8, IL-10, VEGF, and TGF-b disturb
antitumor immune surveil-lance and result in tumor immune escape
(8, 10, 28).Therefore, we tried to identify which cytokines are
majorfactors involved in the regulation of NK-cell
cytolyticactivity during STAT3 decoy ODN treatment. First, wefound
that the cytokine profile of hepatocellular carcino-ma cells was
changed by blocking STAT3, with the reduc-tion of inflammatory
cytokines (IL-6, -8, -10, -17, and -23)and the immunosuppressive
factor TGF-b, whereas stim-ulatory IFN was increased (Fig. 5).
Cumulative studiesshowed thatmodulatory cytokines from the
tumormicro-environment are responsible for the resistance of
tumorcells to immune effectors (15). The immunosuppressivecytokine
TGF-b and IL-10 play key roles in tumor immune
evasion. The levels of TGF-b are often elevated in theserum of
patients with cancer, which is responsible forsuppressing the
activation of NK cells and cytotoxic Tcells, as well as the
differentiation of regulatory T cells(1, 31). By blocking STAT3 in
hepatocellular carcinomacells, we found that levels of both TGF-b1
and IL-10 weredecreased at the protein andmRNA levels (Fig. 5),
accom-panied by downregulation of TGF-b1 and IL-10 transcrip-tional
activities (Fig. 6A and 6B). In addition, it has beenreported that
activated STAT3 directly binds to the IL-6promoter to positively
regulate IL-6 expression (36),which is in accord with our
observation that IL-6 mRNAwas downregulated by blocking STAT3. In
contrast, as astimulator of innate and adaptive immune
responses,the mRNA of IFNs in hepatocellular carcinoma wasincreased
by STAT3 blockade (Fig. 5A). Type I IFNactivates NK, DCs, CD4 and
CD8 T cells, plays a centralrole in the process of antitumor immune
responses andinduces STAT1 activation (37, 38). In addition,
IFN-a/bhas been considered as a strategy for tumor
immuno-chemotherapy through upregulation of NKG2D ligandsand MHC
class I, leading to the growth inhibition oftumor cells (39). Our
data presented in Fig. 7A, showsthat p-STAT1 was upregulated by
STAT3 decoy treat-ment in HepG2 cells, and the cytolytic activity
of NKcells was inhibited by anti-IFNR neutralizing antibody(Fig. 7B
and 7C).
*
Lu
ci a
ctiv
ity
fold
ind
uct
ion
0
1
2
3
Ctrl STAT1
ScrambleDecoy
b-Actin
p-STAT1
STAT1
A
DB
C **
10
20
30
40
50 *** **
NS
% S
pec
ific
lysi
s
0
Medium IgG+BSAScr Scr+IFNDecoy Decoy+a-IFNR
Lipo-Ctrl
**
Via
bili
ty (
%)
E:T
0
20
40
60
80
5:1 2.5:1
******
**
*****
** **
% G
ated
0
15
30
45
60
75
90
NKG2D NKG2A CD69 GranzymeB Perforin
*******
*** **
**
** ****
****** **
Medium IgG+BSA ScrScr+IFN Decoy Decoy+a-IFNR
Lipo-Ctrl
Figure 7. Type I IFN production by STAT3 decoyODN-treated
hepatocellular carcinoma cells was critical for NK-cell activation.
A,Western blotting analysis ofphospho-STAT1 in lysates prepared
fromHepG2 transfected with STAT3 decoy ODN, scramble ODN, or
Lipofectamine reagent (Lipo-Ctrl) for 24 hours (left).Luciferase
reporter assay was used to determine the transcriptional activity
of STAT1 with STAT3-decoy or scramble ODN transfection (right).B to
D, NK-92 cells were cultured with the supernatants from scramble
ODN-treated-HepG2 cells supplemented with or without IFN-a (400
U/mL) þ IFN-b(400 U/mL; ScrþIFN and Scr, respectively), and the
supernatants from STAT3 decoy ODN-treated HepG2 cells in the
presence or absence of anti-IFNAR(10 mg/mL; Decoyþa-IFNR and Decoy,
respectively) for 12 hours. Subsequently, the suppressive effect of
these NK-92 cells on HepG2 cell viability (B)and the specific lysis
of NK-92 cells against HepG2 cellswere detected (C), and
expressions of themolecules associatedwithNK-cell cytolysis were
analyzedby flow cytometry (D). Data are representative of three
independent experiments; statistical significance was determined as
P < 0.05 (�) and P < 0.01 (��)compared with the indicated
groups; NS, no significance.
Sun et al.
Mol Cancer Ther; 12(12) December 2013 Molecular Cancer
Therapeutics2894
on April 2, 2021. © 2013 American Association for Cancer
Research. mct.aacrjournals.org Downloaded from
Published OnlineFirst October 9, 2013; DOI:
10.1158/1535-7163.MCT-12-1087
http://mct.aacrjournals.org/
-
In conclusion, this study demonstrated that hepato-cellular
carcinoma cells create a multifaceted immuno-suppressive
microenvironment that suppresses NK-cell functions and,
furthermore, that blocking STAT3in hepatocellular carcinoma cells
restored NK-cell–
mediated antitumor efficiency. As shown in Fig. 8,TGF-b
reduction and type I IFN production were bothcritical to NK-cell
activation, and the upregulation ofNKG2D ligands in hepatocellular
carcinoma cellsfacilitated NK-cell recognition and activation.
Thesefindings indicate that targeted blockade of STAT3
inhepatocellular carcinoma cells not only promotes tumorcell
apoptosis directly but also released
hepatocellularcarcinoma-induced antitumor immune suppression ofNK
cells indirectly, suggesting a new anti-hepatocellu-lar carcinoma
immunotherapy.
Disclosure of Potential Conflicts of InterestNo potential
conflicts of interest were disclosed.
Authors' ContributionsConception and design: X. Sun, Q. Sui, C.
Zhang, Z. Tian, J. ZhangDevelopment of methodology: X. Sun, Q. Sui,
J. ZhangAcquisition of data (provided animals, acquired and managed
patients,provided facilities, etc.): X. Sun, Q. Sui, J.
ZhangAnalysis and interpretation of data (e.g., statistical
analysis, biostatis-tics, computational analysis): X. Sun, Q. Sui,
C. Zhang, J. ZhangWriting, review, and/or revisionof
themanuscript:X. Sun,Q. Sui, Z. Tian,J. ZhangAdministrative,
technical, or material support (i.e., reporting or orga-nizing
data, constructing databases): X. Sun, Q. Sui, C. Zhang, J.
ZhangStudy supervision: Z. Tian
AcknowledgmentsThe authors thank the editor and three anonymous
referees for their
constructive advice and comments to improve this work.
Grant SupportThisworkwasfinancially supported by theNatural
Science Foundation
of China (grants no. 81172789, 30972692, and 30628014; to J.
Zhang).The costs of publication of this article were defrayed in
part by the
payment of page charges. This article must therefore be hereby
markedadvertisement in accordance with 18 U.S.C. Section 1734
solely to indicatethis fact.
Received November 15, 2012; revised July 29, 2013; accepted
August 14,2013; published OnlineFirst October 9, 2013.
References1. Leonardi GC, Candido S, CervelloM,Nicolosi D, Raiti
F, Travali S, et al.
The tumor microenvironment in hepatocellular carcinoma (review).
IntJ Oncol 2012;40:1733–47.
2. Ito K, Stannard K, Gabutero E, Clark AM, Neo SY, Onturk S, et
al.Galectin-1 as a potent target for cancer therapy: role in
thetumor microenvironment. Cancer metastasis reviews
2012;31:763–78.
3. Mamessier E, Sylvain A, Thibult ML, Houvenaeghel G,
Jacquemier J,Castellano R, et al. Human breast cancer cells enhance
self toleranceby promoting evasion from NK cell antitumor immunity.
J Clin Invest2011;121:3609–22.
4. Ihara S, Kida H, Arase H, Tripathi LP, Chen YA, Kimura T, et
al.Inhibitory roles of signal transducer and activator of
transcription 3 inantitumor immunity during carcinogen-induced lung
tumorigenesis.Cancer Res 2012;72:2990–9.
5. Cai L, Zhang Z, Zhou L, Wang H, Fu J, Zhang S, et al.
Functionalimpairment in circulating and intrahepatic NK cells and
relative mech-anism in hepatocellular carcinoma patients. Clin
Immunol 2008;129:428–37.
6. Castriconi R, Cantoni C, Della Chiesa M, Vitale M, Marcenaro
E,Conte R, et al. Transforming growth factor beta 1 inhibits
expres-sion of NKp30 and NKG2D receptors: consequences for the
NK-mediated killing of dendritic cells. Proc Natl Acad Sci U S
A2003;100:4120–5.
7. Pietra G, Manzini C, Rivara S, Vitale M, Cantoni C, Petretto
A, et al.Melanoma cells inhibit natural killer cell function by
modulating theexpression of activating receptors and cytolytic
activity. Cancer Res2012;72:1407–15.
8. Gupta M, Han JJ, Stenson M, Maurer M, Wellik L, Hu G, et al.
Elevatedserum IL-10 levels in diffuse large B-cell lymphoma: a
mechanism ofaberrant JAK2 activation. Blood 2012;119:2844–53.
9. Zhang X, Yue P, Page BD, Li T, Zhao W, Namanja AT, et al.
Orallybioavailable small-molecule inhibitor of transcription factor
Stat3regresses human breast and lung cancer xenografts. Proc Natl
AcadSci U S A 2012;109:9623–8.
10. Luwor RB, Baradaran B, Taylor LE, Iaria J, Nheu TV, Amiry N,
et al.Targeting Stat3 and Smad7 to restore TGF-beta cytostatic
regulationof tumor cells in vitro and in vivo. Oncogene
2013;32:2433–41.
11. ZhangX, Zhang J,WangL,WeiH, Tian Z. Therapeutic effects of
STAT3decoy oligodeoxynucleotide on human lung cancer in xenograft
mice.BMC Cancer 2007;7:149.
12. Su Y, Li G, Zhang X, Gu J, Zhang C, Tian Z, et al. JSI-124
inhibitsglioblastoma multiforme cell proliferation through G(2)/M
cell cyclearrest and apoptosis augment. Cancer Biol Ther
2008;7:1243–9.
NK
HCC
p p
STAT3
Apoptosis
TGF-b, IL-10
NKG2A
Anti-HCC effects
PerforinGZMIFN-g …
IFN-a /bMHC I
+
+
++
+
+-
Immunosuppresive molecules
Cytolysis
+
STAT3 decoy ODN
Figure 8. Targeted blockade of STAT3 in hepatocellular carcinoma
(HCC)cells augmented NK cell function. Blocking STAT3 in
hepatocellularcarcinoma induced tumor apoptosis directly while NK
cell antitumorfunction was enhanced indirectly. As shown,
expression of NKG2Dligands by hepatocellular carcinoma cells was
upregulated with STAT3blockade, which promoted the recognition by
NK cells and enhanced thesusceptibility of hepatocellular carcinoma
cells to NK cell cytolysis.Furthermore, the cytokine profile of
hepatocellular carcinoma cells wasaltered, resulting in the
reduction of immunosuppressive molecules,especially TGF-b and
IL-10, and the induction of type I IFN. All thesechanges resulted
in the release of the hepatocellular
carcinoma-inducedimmunosuppressive status, and the cytotoxicity of
NK cells againsthepatocellular carcinoma cells was enhanced.
Blocking STAT3 in Hepatocellular Carcinoma Cells Restored NK
Activation
www.aacrjournals.org Mol Cancer Ther; 12(12) December 2013
2895
on April 2, 2021. © 2013 American Association for Cancer
Research. mct.aacrjournals.org Downloaded from
Published OnlineFirst October 9, 2013; DOI:
10.1158/1535-7163.MCT-12-1087
http://mct.aacrjournals.org/
-
13. Sun X, Zhang J, Wang L, Tian Z. Growth inhibition of human
hepato-cellular carcinoma cells by blocking STAT3 activation with
decoy-ODN. Cancer Lett 2008;262:201–13.
14. Wang T, Niu G, Kortylewski M, Burdelya L, Shain K, Zhang S,
et al.Regulation of the innate and adaptive immune responses by
Stat-3signaling in tumor cells. Nature Med 2004;10:48–54.
15. YuH,KortylewskiM, Pardoll D. Crosstalk between cancer and
immunecells: role of STAT3 in the tumour microenvironment. Nat Rev
Immunol2007;7:41–51.
16. FujitaM, Zhu X, Sasaki K, UedaR, LowKL, Pollack IF, et al.
Inhibition ofSTAT3 promotes the efficacy of adoptive transfer
therapy using type-1CTLs by modulation of the immunological
microenvironment in amurine intracranial glioma. J Immunol
2008;180:2089–98.
17. Kortylewski M, Kujawski M, Wang T, Wei S, Zhang S,
Pilon-ThomasS, et al. Inhibiting Stat3 signaling in the
hematopoietic systemelicits multicomponent antitumor immunity.
Nature Med 2005;11:1314–21.
18. Nefedova Y, Cheng P, Gilkes D, BlaskovichM, Beg AA, Sebti
SM, et al.Activation of dendritic cells via inhibition of
Jak2/STAT3 signaling.J Immunol 2005;175:4338–46.
19. Kong LY, Gelbard A, Wei J, Reina-Ortiz C, Wang Y, Yang EC,
et al.Inhibition of p-STAT3 enhances IFN-alpha efficacy against
metastaticmelanoma in a murine model. Clin Cancer Res
2010;16:2550–61.
20. TkachM, Coria L, Rosemblit C, RivasMA, Proietti CJ, Diaz
FlaqueMC,et al. Targeting Stat3 induces senescence in tumor cells
and elicitsprophylactic and therapeutic immune responsesagainst
breast cancergrowth mediated by NK cells and CD4þ T cells. J
Immunol 2012;189:1162–72.
21. Bedel R, Thiery-Vuillemin A, Grandclement C, Balland J,
Remy-MartinJP, Kantelip B, et al. Novel role for STAT3 in
transcriptional regulationof NK immune cell targeting receptor MICA
on cancer cells. CancerRes 2011;71:1615–26.
22. Leong PL, Andrews GA, Johnson DE, Dyer KF, Xi S, Mai JC, et
al.Targeted inhibition of Stat3 with a decoy oligonucleotide
abrogateshead and neck cancer cell growth. Proc Natl Acad Sci U S A
2003;100:4138–43.
23. Nausch N, Cerwenka A. NKG2D ligands in tumor immunity.
Oncogene2008;27:5944–58.
24. Vivier E, Tomasello E, Baratin M, Walzer T, Ugolini S.
Functions ofnatural killer cells. Nat Immunol 2008;9:503–10.
25. Yokoyama WM, Kim S. How do natural killer cells find self to
achievetolerance? Immunity 2006;24:249–57.
26. CoulouarnC,CorluA,GlaiseD,Guenon I,
ThorgeirssonSS,ClementB.Hepatocyte-stellate cell cross-talk in the
liver engenders a permissiveinflammatory microenvironment that
drives progression in hepatocel-lular carcinoma. Cancer Res
2012;72:2533–42.
27. He W, Liu Q, Wang L, Chen W, Li N, Cao X. TLR4 signaling
promotesimmune escape of human lung cancer cells by inducing
immunosup-
pressive cytokines and apoptosis resistance. Molecular
Immunol2007;44:2850–9.
28. Teruel A, Romero M, Cacalano NA, Head C, Jewett A.
Potentialcontribution of naive immune effectors to oral tumor
resistance: rolein synergistic induction of VEGF, IL-6, and IL-8
secretion. CancerImmunol, Immunother 2008;57:359–66.
29. LopezMN,PesceB, KurteM, PerezC, SegalG, Roa J, et al. A
syntheticpeptide homologous to IL-10 functional domain induces
monocytedifferentiation to TGF-beta þtolerogenic dendritic cells.
Immunobiol-ogy 2011;216:1117–26.
30. Donkor MK, Sarkar A, Li MO. Tgf-beta1 produced by activated
CD4(þ)T Cells Antagonizes T Cell Surveillance of Tumor
Development.Oncoimmunology 2012;1:162–71.
31. Shimizu H, Nakagami H, Tsukamoto I, Morita S, Kunugiza Y,
Tomita T,et al. NFkappaB decoy oligodeoxynucleotides ameliorates
osteopo-rosis through inhibition of activation and differentiation
of osteoclasts.Gene ther 2006;13:933–41.
32. Guerra N, Tan YX, Joncker NT, Choy A, Gallardo F, Xiong N,
et al.NKG2D-deficient mice are defective in tumor surveillance in
models ofspontaneous malignancy. Immunity 2008;28:571–80.
33. Zocchi MR, Catellani S, Canevali P, Tavella S, Garuti A,
Villaggio B,et al. High ERp5/ADAM10 expression in lymph node
microenviron-ment and impaired NKG2D ligands recognition in Hodgkin
lympho-mas. Blood 2012;119:1479–89.
34. Nam JS, TerabeM, Mamura M, KangMJ, Chae H, Stuelten C, et
al. Ananti-transforming growth factor beta antibody suppresses
metastasisvia cooperative effects on multiple cell compartments.
Cancer Res2008;68:3835–43.
35. Hilpert J, Grosse-Hovest L, Grunebach F, Buechele C,
Nuebling T,Raum T, et al. Comprehensive analysis of NKG2D ligand
expressionand release in leukemia: implications for NKG2D-mediated
NK cellresponses. J Immunol 2012;189:1360–71.
36. Yoon S, Woo SU, Kang JH, Kim K, Kwon MH, Park S, et al.
STAT3transcriptional factor activated by reactive oxygen species
induces IL6in starvation-induced autophagy of cancer cells.
Autophagy 2010;6:1125–38.
37. Diamond MS, Kinder M, Matsushita H, Mashayekhi M, Dunn
GP,Archambault JM, et al. Type I interferon is selectively required
bydendritic cells for immune rejection of tumors. J Exp Med
2011;208:1989–2003.
38. Hervas-StubbsS,Perez-Gracia JL,RouzautA, SanmamedMF, LeBonA,
Melero I. Direct effects of type I interferons on cells of the
immunesystem. Clin Cancer Res 2011;17:2619–27.
39. Khallouf H, Marten A, Serba S, Teichgraber V, Buchler MW,
Jager D,et al. 5-Fluorouracil and interferon-alpha
immunochemotherapyenhances immunogenicity of murine pancreatic
cancer through upre-gulation of NKG2D ligands and MHC class I. J
Immunother 2012;35:245–53.
Sun et al.
Mol Cancer Ther; 12(12) December 2013 Molecular Cancer
Therapeutics2896
on April 2, 2021. © 2013 American Association for Cancer
Research. mct.aacrjournals.org Downloaded from
Published OnlineFirst October 9, 2013; DOI:
10.1158/1535-7163.MCT-12-1087
http://mct.aacrjournals.org/
-
2013;12:2885-2896. Published OnlineFirst October 9, 2013.Mol
Cancer Ther Xiaoxia Sun, Qiangjun Sui, Cai Zhang, et al. Induced
Immune Suppression−
Augments NK Cell Functions via Reverse Hepatocellular Carcinoma
Targeting Blockage of STAT3 in Hepatocellular Carcinoma Cells
Updated version
10.1158/1535-7163.MCT-12-1087doi:
Access the most recent version of this article at:
Material
Supplementary
http://mct.aacrjournals.org/content/suppl/2013/10/09/1535-7163.MCT-12-1087.DC1
Access the most recent supplemental material at:
Cited articles
http://mct.aacrjournals.org/content/12/12/2885.full#ref-list-1
This article cites 39 articles, 17 of which you can access for
free at:
Citing articles
http://mct.aacrjournals.org/content/12/12/2885.full#related-urls
This article has been cited by 2 HighWire-hosted articles.
Access the articles at:
E-mail alerts related to this article or journal.Sign up to
receive free email-alerts
Subscriptions
Reprints and
[email protected]
To order reprints of this article or to subscribe to the
journal, contact the AACR Publications Department at
Permissions
Rightslink site. Click on "Request Permissions" which will take
you to the Copyright Clearance Center's (CCC)
.http://mct.aacrjournals.org/content/12/12/2885To request
permission to re-use all or part of this article, use this link
on April 2, 2021. © 2013 American Association for Cancer
Research. mct.aacrjournals.org Downloaded from
Published OnlineFirst October 9, 2013; DOI:
10.1158/1535-7163.MCT-12-1087
http://mct.aacrjournals.org/lookup/doi/10.1158/1535-7163.MCT-12-1087http://mct.aacrjournals.org/content/suppl/2013/10/09/1535-7163.MCT-12-1087.DC1http://mct.aacrjournals.org/content/12/12/2885.full#ref-list-1http://mct.aacrjournals.org/content/12/12/2885.full#related-urlshttp://mct.aacrjournals.org/cgi/alertsmailto:[email protected]://mct.aacrjournals.org/content/12/12/2885http://mct.aacrjournals.org/