-
C
Fi
PSa*dM
Bmppbtsi(qcatBHtwHcwcatetigFcu
Hvudkt
GASTROENTEROLOGY 2004;127:1110–1122
LINICAL–LIVER, PANCREAS, AND BILIARY TRACT
unctional Consequences of Frizzled-7 Receptor Overexpressionn
Human Hepatocellular Carcinoma
HILIPPE MERLE,*,‡ SUZANNE DE LA MONTE,* MIRAN KIM,* MARC
HERRMANN,*HINJI TANAKA,§ ANNETTE VON DEM BUSSCHE,* MICHAEL C. KEW,�
CHRISTIAN TREPO,‡
nd JACK R. WANDS*The Liver Research Center, Department of
Medicine and Pathology, Brown Medical School, Providence, Rhode
Island; ‡Inserm U271, Viruses Hépatites et Pathologies Associées,
Lyon, France; §Department of Surgery II, Kyushu University,
Fukuoka, Japan; and �Department ofedicine, Witwatersrand University
Medical School, Johannesburg, South Africa
�uofimfWot
peangptrbsttaTurtl
ptntr
ackground & Aims: The molecular pathogenesis of hu-an
hepatocellular carcinoma (HCC) is understoodoorly. In some tumors,
activation of the Wnt/�-cateninathway as a result of �-catenin gene
mutations haseen found. However, in many other HCCs, activation
ofhe Wnt/�-catenin pathway has been shown in the ab-ence of such
mutations. Methods: We previously havedentified the upstream human
Frizzled-7 receptorFZD7) gene of this pathway. In the present
study, auantitative real-time reverse-transcription polymerasehain
reaction (RT-PCR) assay for FZD7 was developednd overexpression of
FZD7 was detected in 90% ofumors, most of which were related to
chronic hepatitisvirus infection. FZD7 also was overexpressed in
the 6CC cell lines tested and functional analysis showedhat FZD7
messenger RNA (mRNA) levels correlatedith enhanced cellular
motility. Results: Transfection ofCC cells with dominant-negative
mutant constructs en-oding a C-terminally truncated FZD7 protein
decreasedild-type �-catenin protein accumulation and reducedell
motility. More importantly, we observed �-cateninccumulation in
human HCC tumors containing the wild-ype �-catenin gene in the
context of high-level FZD7xpression. Conclusions: These
observations suggesthat the Wnt/�-catenin signal transduction
pathway isnvolved much more commonly in the molecular patho-enesis
of HCC than previously recognized becauseZD7 overexpression
occurred early in the disease pro-ess, stabilized wild-type
�-catenin levels, and contrib-ted to enhanced tumor cell
migration.
epatocellular carcinoma (HCC) is the major pri-mary malignant
tumor of the liver. Although
iral causative factors have been identified, the molec-lar
mechanisms that contribute to tumor progressionuring
hepatocarcinogenesis remain largely un-nown.1 The previous finding
of inappropriate activa-ion of the Wnt/�-catenin pathway resulting
from
-catenin gene mutations has provided clues towardnderstanding
this process.2 The upstream receptorsf this signal transduction
pathway have been identi-ed recently. The Frizzled family is
composed of 10 orore 7-transmembrane proteins that act as
receptors
or Wnt proteins, which serve as the ligands.3 Thent/Frizzled
signaling network influences diverse bi-
logical processes ranging from cell fate determinationo cell
motility and proliferation.4
The �-catenin molecule is an important multifactorialrotein. It
is involved in cell– cell adhesion by strength-ning the linkage
between cadherin and �-catenin to thectin cytoskeleton.5 In the
absence of Wnt/Frizzled sig-aling, �-catenin is phosphorylated by
interactions withlycogen synthase kinase (GSK)-3�, and forms a
com-lex with axin and the adenomatous polyposis coli pro-ein (APC);
subsequently, �-catenin is targeted for deg-adation by the
ubiquitin-proteasome system.6 In contrast,inding of a Wnt ligand to
its Frizzled receptor will lead totabilization of intracellular
�-catenin through the inhibi-ion of GSK-3� enzymatic activity.
There is subsequentranslocation of the �-catenin protein into the
nucleus inssociation with high mobility group domain factors such
ascf/Lef. This complex is associated with
transcriptionalp-regulation of growth regulatory and cell
migration–elated genes.7 In Drosophila, the Wnt/Frizzled signal
isransduced by the small G-protein RhoA, which in turn canead to
changes involving the reorganization of the cytoskel-
Abbreviations used in this paper: APC, adenomatous polyposis
colirotein; 18SrRNA, 18S ribosomal RNA; FZD7, human Frizzled-7
recep-or gene; FZD7-�T�C, human Frizzled-7 receptor gene with a
C-termi-al truncation; GFP, green fluorescence protein; GSK,
glycogen syn-hase kinase; RT-PCR, reverse-transcription polymerase
chaineaction.
© 2004 by the American Gastroenterological
Association0016-5085/04/$30.00
doi:10.1053/j.gastro.2004.07.009
-
ecznCea
ici16aoA9b8iAdytaob
hmoeo(stpt(hurattsheomvl
SfyT2huophmimMs(
unpTaewPLFsalgeTCC5CpmTct
coeeatt
October 2004 FRIZZLED GENE OVEREXPRESSION IN HCC 1111
ton through the activation of downstream kinases such as-Jun
NH2-terminal kinase. Finally, studies performed inebra fish and
Xenopus showed that the Wnt/Frizzled sig-aling can increase
intracellular calcium via phospholipase�, leading to downstream
activation of the Ca2�-sensitivenzymes, Ca2�-calmodulin–dependent
protein kinase IInd protein kinase C.8
Cytoplasmic and/or nuclear accumulation of �-catenins a frequent
event in various human tumors includingolorectal, lung, breast,
cervix, skin, and liver.9–17 Dur-ng hepatocarcinogenesis, it has
been reported that 8%–3% of HCCs have �-catenin mutation(s),
whereas 33%–9% of these tumors display aberrant �-catenin
proteinccumulation in the cytoplasm and/or nucleus.18–20 Amongther
candidate genes involved in stabilization of �-catenin,xin1 was
found to be mutated in a small fraction (7%–%) of HCCs,21,22
whereas APC mutations were found toe exceptionally rare.23 Thus, it
is estimated that in 35%–0% of HCCs, the aberrant �-catenin
cellular accumulations not associated with a mutation affecting the
�-catenin,xin1, or APC genes of the Wnt/�-catenin signal
trans-uction cascade, and the explanation for this observation
haset to be determined. Although it was reported recentlyhat the
p53 gene mutation may contribute to aberrantccumulation of
wild-type �-catenin,24 it is likely thatther components of this
signal transduction cascade maye involved as well.
The expression of Wnt ligands and Frizzled receptorsave been
found in esophageal, colon, and skin carcino-as.25–27 However, the
expression and/or functional role
f any Frizzled gene family member has not yet beenxplored during
human hepatocarcinogenesis. We previ-usly identified and cloned the
Frizzled-7 receptorFZD7), also called Frizzled-E3, and showed
overexpres-ion in human esophageal carcinoma resulting in
activa-ion of downstream APC/�-catenin signals.26 In theresent
study, we have developed a quantitative real-ime
reverse-transcription polymerase chain reactionRT-PCR) method to
measure the expression of FZD7 inuman HCC cell lines, as well as
tumors and adjacentninvolved liver. The functional association of
FZD7eceptor expression with cellular motility was explorednd
experiments were performed to determine its rela-ionship with
�-catenin stabilization. This report showshe FZD7 gene commonly is
overexpressed at the mes-enger RNA (mRNA) and protein level in most
ofepatitis B virus–related HCCs. In this context, there isnhanced
cell motility and migration. Most importantly,verexpression appears
to be an early event during theultistep process of hepatocyte
transformation and pro-
ides novel molecular targets for therapy of this seriousiver
disease.
Materials and Methods
HCC Tumors and Cell Lines
HCC tumor tissues were obtained from Taiwan andouth Africa.
Individuals from Taiwan had undergone surgeryor resection of tumor
(n � 16). Ages ranged from 36 to 64ears. All were men with
hepatitis B virus–related disease.here were 14 samples from South
Africa. Ages ranged from8 to 57 years. All were men. Ten were
hepatitis B virus, 1epatitis C virus, 2 related to hemochromatosis,
and 1 was ofnknown cause. There was a matched, uninvolved,
nontumor-us liver sample for all 30 HCCs. Cirrhosis and/or fibrosis
wasresent in 85%. Huh7, Focus, HepaRG,28 Hep3B, and HepG2epatoma
cell lines were grown in minimum essential Eagleedium (Mediatech,
Herndon, VA), and PLC/PRF/5 was grown
n Dulbecco’s modified Eagle medium (Mediatech), supple-ented
with 10% (vol/vol) fetal calf serum (Sigma, St. Louis,O), 1�
minimum essential medium nonessential amino acid
olution (Sigma), and 1% (vol/vol)
penicillin/streptomycinSigma).
Real-Time RT-PCR Assay
The copy number of FZD7 mRNAs was quantified innknown samples by
measuring the Ct value followed byormalization to 18S ribosomal RNA
(18SrRNA) after com-arison with a standard curve for both FZD7 and
18SrRNA.his ratio of FZD7/18SrRNA subsequently was normalized
tocalibrator (mean value obtained from normal livers) and
xpressed as relative abundance of FZD7 mRNA. Standardsere
prepared with 10-fold dilutions of the correspondingCR products
cloned into the pCR 2.1 Vector (Invitrogen,ife Technologies,
Rockville, MD). The specificity of theZD7 and 18SrRNA inserts was
provided by sequence analy-is. Serial dilution of FZD7- and
18SrRNA-plasmids wasliquoted and stored at �20°C until use. Primers
were se-ected by using the Primer3 website
(http://www-enome.wi.mit.edu/cgi-bin/primer/primer3_www.cgi).
Prim-rs for FZD7 and 18SrRNA, respectively, (Invitrogen,
Lifeechnologies) were as follows: (1) FZD7,
5=-GCCGCTTCTAC-ACAGACT-3= (forward) and 5=-TTCATACCGCAGTCTC-CC-3=
(reverse) to yield a 54-bp amplicon; (2) human
18SrRNA,=-GGACACGGACAGGATTGACA-3= (forward) and
5=-AC-CACGGAATCGAGAAAGA-3= (reverse) to give a 50-bp am-licon. We
conducted Nucleotide-Nucleotide Basic Local Align-ent Search Tool
searches against database of Expressed Sequenceags and the
nonredundant set of GenBank database sequences toonfirm the gene
specificity of the nucleotide sequences chosen forhe primers to
confirm the lack of DNA polymorphism.
Total RNA was extracted from liver specimens and HCCell lines by
using TRIzol reagent (Invitrogen, Life Technol-gies). The quality
of the RNA samples was determined bylectrophoresis through agarose
gels and staining withthidium bromide; the 18S and 28S rRNA bands
were visu-lized under ultraviolet light. A total amount of 250 ng
ofotal RNA was treated with DNase-I, RNase-free, and
reverseranscribed with random hexamers and the AMV reverse
tran-
http:\www-genome.wi.mit.edu/cgi-bin/primer/primer3_www.cgihttp:\www-genome.wi.mit.edu/cgi-bin/primer/primer3_www.cgi
-
soiHMeacfmrc1
meuveMsbbsp3pbpbmcTtpAp(Eae
amgcvvopf
I8csy
ttapJpob2cs(aspva
tptflsiHlmts
brNKsa4isbMTBsavB
1112 MERLE ET AL. GASTROENTEROLOGY Vol. 127, No. 4
criptase (all from Roche Diagnostics Corporation, Indianap-lis,
IN). All PCR reactions were performed using an iCyclerQ Multi-Color
Real Time PCR Detection System (Bio-Rad,ercules, CA) with a mix
composed of 1� SYBR Green PCRaster Mix (Applied Biosystems, Foster
City, CA) 500 nmol/L
ach primer, and 5 ng complementary DNA (cDNA) (equiv-lent total
RNA) from unknown samples. The thermal cyclingonditions comprised
an initial step at 95°C for 10 minutes,ollowed by 40 cycles at 95°C
for 15 seconds, and 60°C for 1
inute. Experiments were performed in duplicate. Each PCRun
included FZD7 and 18S standard curves, a nontemplateontrol, and the
unknown cDNAs analyzed for FZD7 and8SrRNA copy numbers.
Cell Motility Assay
A luminescence-based assay was used to evaluate cellotility and
migration. This assay assesses nonmigrated, migrat-
d-adherent, and migrated nonadherent hepatoma cells
throughncoated polyvinylpyrrolidone-free polycarbonate filters, as
pre-iously described,29 in chambers partitioned with 13-mm
diam-ter, 8-�mol/L pore polycarbonate membranes (Osmonics,
Inc.,innetonka, MN). After 72 hours of growth in 1% fetal calf
erum, cells were resuspended in serum-free medium with soy-ean
trypsin inhibitor (Sigma) at 0.5 mg/mL. In the lower cham-er,
collagen-I (Sigma) was diluted to 100 �g/mL in 200 �L oferum-free
medium.30 After assembly, 1 � 105 of the resus-ended cells were
added to the upper chamber and incubated forhours in a CO2
humidified incubator to allow cell migration to
roceed. The cells remaining on the upper surface of the mem-rane
(nonmigrated) were harvested with a sterile cotton swab andlaced
into a well containing adenosine triphosphate (ATP) lysisuffer
(Packard Instrument Company, Meriden, CT). To harvestigrated
adherent cells, the membrane (devoid of nonmigrated
ells) was placed into another well containing ATP lysis
buffer.he migrated nonadherent cells were harvested by
resuspending
he cells in the lower compartment of the blind chamber andlacing
them into a third well containing ATP lysis buffer.TPLite substrate
was added to each well and luminescent countser second were
measured in a TopCount Microplate readerPackard Instrument
Company). The results were analyzed usingxcel (Microsoft
Corporation, Seattle, WA) to calculate percent-ges of nonmigrated,
migrated-adherent, and migrated nonadher-nt cells in each
assay.
Transfection and Retroviral TransductionStudies
Because the ectodomain of Frizzled receptors functions asnatural
antagonist of Frizzled-mediated signal transduction, autant cDNA
with a C-terminal truncation (FZD7-�T�C) was
enerated as previously described.26 The cDNA FZD7-�T�CDNA was
subcloned into a pcDNA3 mammalian expressionector (Invitrogen) and
certified for protein expression by an initro translation
reaction.26 The in vitro effect of stable expressionf FZD7-�T�C was
examined in human hepatoma cells. ThecDNA3/FZD7-�T�C and
pcDNA3/empty vectors were trans-ected by electroporation (250
V/15-ms pulse with Gene Pulser
I; BioRad, Hercules, CA). Stable transfectants were selected
with00 �g/mL of Geneticin (Gibco, Carlsbad, CA) added to theulture
medium. Expression of the FZD7-�T�C gene was as-essed by
quantitative real-time RT-PCR and Western blot anal-sis.
Two different FZD7 truncated mutants either with dele-ions in
the intracellular domain alone (FZD7-�C) or in bothhe intracellular
and transmembrane domains (FZD7-�T�C),s well as a green
fluorescence protein (GFP) control, wererepared by PCR using the
Pfu-polymerase (Stratagene, Laolla, CA). Constructs were cloned
into the lentiviralLenti6/V5 directional TOPO vector (Invitrogen)
downstreamf the cytomegalovirus promoter. All constructs were
verifiedy sequence analysis of both strands. Virions were produced
in93FT cells, and viral stocks were frozen at �80°C. Hepatomaells
were transduced at a multiplicity of infection of 5, andtable
clones were selected in the presence of blasticidin4 �g/mL).
Quantitative real-time RT-PCR and Western blotnalysis were used to
detect gene expression. A mutant con-truct of �-catenin with
biologic activity (�N/�C) was pre-ared by PCR using the
Pfu-polymerase (Stratagene) as pre-iously described31 and after
transfection; HCC motility wasssessed 72 hours later.
We performed transient transfection assays to assess for
Tcfranscriptional activity. In brief, cells were seeded in
6-welllates at a density of 3 � 105 cells/well the day
beforeransfection. Transfections were performed with either TOP-ash
(Tcf Reporter Plasmid) or FOPflash (mutant Tcf bindingites) and
�-galactosidase– expressing plasmid in triplicate us-ng
Lipofectamine 2000 (Life Technologies). Focus, Huh7, andep3B cells
were harvested 48 hours after transfection and
uciferase activity was measured by a luminometer
TopCounticroplate reader (Packard Instrument Company).
Transfec-
ion efficiency was normalized by measurement of �-galacto-idase
activity. The experiment was performed 3 times.
Immunoprecipitation Studies
Transfected cells were washed with cold phosphate-uffered
saline, and pelleted at 2000 rpm for 5 minutes andesuspended in 500
�L of solubilization buffer (136 mmol/LaCl, 2.7 mmol/L KCl, 12
mmol/L Na2HPO4, 1.8 mmol/LH2PO4, pH 7.4, 1% NP-40 with protease
inhibitors). Ly-
ates were incubated for 3 hours at 4°C on a rocker platformnd
then microcentrifuged at 13,000 rpm for 15 minutes at°C to
pellet-out cell debris. Lysates were immunoprecipitatedn a 1.0-mL
total volume with 500 �L of lysate, 480 �L ofolubilization buffer,
and 20 �L of goat anti-human V5 anti-odies coupled to agarose beads
(Bethyl Laboratories, Inc.,ontgomery, TX) for 12 hours at 4°C with
constant shaking.he beads were washed twice with the solubilization
buffer.ound proteins were eluted by boiling for 5 minutes in
odium dodecyl sulfate sample buffer. The proteins were sep-rated
on 10% polyacrylamide gel and transferred to a poly-inylidene
difluoride membrane (NEN Life Science Products,oston, MA).
-
cpcmisBsp
fpu5TcS
nCpPtdw1aMesFae
tp
FfcFkfdm(lposs
October 2004 FRIZZLED GENE OVEREXPRESSION IN HCC 1113
Protein Extraction and Western BlotAnalysis
For whole-protein extraction, hepatoma cells at 50%onfluence
were homogenized in lysis buffer (30 mmol/L Tris,H 7.5, 150 mmol/L
NaCl, 1% NP-40, 0.5% Na deoxy-holate, 0.1% sodium dodecyl sulfate,
10% glycerol, and 2mol/L ethylenediaminetetraacetic acid) with
protease inhib-
tors (Roche Molecular Biochemicals, Indianapolis, IN)
andonicated. Protein concentration was determined with theCA
Protein Assay Kit (Pierce, Rockford, IL) using bovine
erum albumin as standard. Subcellular fractionations
wereerformed as previously described.32
Aliquots of proteins were resolved on sodium dodecyl
sul-ate–polyacrylamide gel electrophoresis and transferred
ontoolyvinylidene difluoride membranes (NEN Life Science Prod-cts)
by electroblotting. The membranes were blocked with% nonfat dry
milk in Tris-buffered saline containing 0.1%ween 20 and then probed
with a mouse monoclonal anti-�-atenin antibody diluted at 1:500
(Transduction Laboratories,an Diego, CA), or a mouse monoclonal
antiproliferating cell
igure 1. Real-time PCR assayor frizzled-7 gene. �, Standardurves
for (A) 18SrRNA and (B)ZD7. �, Mean values of un-nown HCC tumor
samples per-ormed in duplicate. The stan-ard curves show a 5 order
ofagnitude linear dynamic range.
C) The FZD7 mRNA steady-stateevels in HCC cell lines ex-ressed
as relative abundancef FZD7 mRNA. The value repre-ents the mean �
SD from 3eparate experiments.
uclear antigen antibody diluted at 1:1000 (Oncogene
Science,ambridge, MA). A specific rabbit polyclonal antibody
wasrepared against a human FZD7 peptide QNTSDGSGG-GGGPTAYPTAPYLPD,
amino acids 163–187 (NCBI Pro-ein database accession no.
BAA_34668), and used at 1:5000ilution. Each primary antibody was
followed by incubationith a secondary horseradish-peroxidase
antibody diluted:10,000 and then revealed with the
chemiluminescence im-ging Western Lightning (PerkinElmer Life
Sciences, Boston,A). The specificity of the antigen-antibody
interaction was
stablished by absorption of FZD7 immunoreactivity withpecific
and not with nonrelevant peptides, as well as a lack ofZD7 protein
detection with addition of second antibodylone (data not shown).
All of the blots were standardized forqual protein loading by
Ponceau S red staining.
Sequencing
PCR amplification of �-catenin exon-3, which con-ains the 4
potential sites for phosphorylation by GSK-3�, waserformed on the
cDNAs derived from each tumor sample
-
uetaeTu
cIw
tafctbscstwrdt(f
eFlttAgHlrHp
sc
oonsbUlcanmc
Fmhvrwmm�
1114 MERLE ET AL. GASTROENTEROLOGY Vol. 127, No. 4
sing a �-catenin exon-2 forward primer, and a �-cateninxon-4
reverse primer as previously described.33 After resolu-ion of the
PCR products by 2% agarose gel electrophoresisnd visualization with
ethidium bromide, PCR products werexcised and cloned into the pCR
2.1 Vector (Invitrogen, Lifeechnologies). Sequencing was performed
in both directionssing T7 forward and M13 reverse primers.
Statistical Analysis
The dependent or independent t tests were used forontinuous data
with StatView software (version 5.0; SASnstitute Inc., Cary, NC).
Tests were considered significanthen P values were 0.05.
Results
Real-Time RT-PCR Assay for FZD7
We used random hexamers for the initial reverse-ranscription
step. The amplification was performed withset of primers generating
a small amplicon size (�50 bp)
or both FZD7 and 18SrRNA. Figure 1 shows the standardurves for
FZD7 and 18SrRNA and shows the values ob-ained for unknown tumor
samples. A linear relationshipetween the Ct and the log of the
starting copy number washown (R2 � 0.99). The efficiency of the
reaction (E),alculated by the formula, E � 101/m � 1, where m is
thelope of the standard curve,34 ranged from 90% to 100% inhe
assays performed at different times. The same featuresith respect
to the dynamic range and efficiency of the
eactions were observed with serial dilutions of cDNAserived from
the Huh7 cell line, or from human HCCumors when using 10–0.625 ng
of equivalent total RNAdata not shown). Sensitivity was 10 copies
per reactionor FZD7 and 100 copies for 18SrRNA.
With this assay in hand, we measured FZD7 genexpression in
HepG2, Hep3B, HepaRG, PLC/PRF/5,ocus, and Huh7; values obtained on
4 normal adultiver tissues served as a control and the mean value
ofhese normal livers served as a calibrator for normaliza-ion of
FZD7 mRNA levels in other unknown samples.ll cell lines were found
to express the FZD7 mRNAene at higher levels than in normal liver
(Figure 1C).uh7, PLC/PRF/5, and Focus cells exhibited the
highest
evels of FZD7 mRNA with ratios of 354, 116, and 81,espectively,
compared with normal livers. In contrast,ep3B, HepaRG, and HepG2
had lower levels of ex-
ression with ratios of 15, 11, and 9, respectively.
Correlation of Steady-State FZD7 mRNALevels and HCC Motility
Based on the observation that the Wnt-Frizzledignal transduction
pathway is known to be involved inell motility and migration,27 we
examined the motility
f HepG2, Hep3B, Focus, and Huh7 cells in the contextf FZD7 gene
expression. In these experiments, we choseot to use fetal calf
serum as chemoattractant29 becauseerum growth factors may interfere
with �-catenin sta-ilization independent of the Wnt-Frizzled
signals.35,36
nder the experimental conditions of using soluble col-agen-I as
a chemoattractant, we observed a significantorrelation between FZD7
steady-state mRNA levelsnd percent of total migrated cells (both
adherent andonadherent) (Figures 2A and B). The highest FZD7RNA
levels and motility rates were found in Huh7
ells.
igure 2. Measurement of HCC cell motility. (A) The percent of
non-igrated and total migrated (migrated-adherent, and migrated
nonad-erent) cells were evaluated by a luminescent-based assay.
Thealues (percent of cells) are expressed as mean � SD from 6
sepa-ate measurements. t test, *P 0.01, **P 0.001 as comparedith
values obtained with HepG2 cells. (B) Correlation between FZD7RNA
steady-state levels and the percent of total migrated cells
(bothigrated-adherent and migrated-nonadherent). Z test of
correlation, P0.02.
-
nccaagmmcOectn
tl3a
FraelcmfeaF
FFctcwLppteotWh�i(wstttEVaiaw
October 2004 FRIZZLED GENE OVEREXPRESSION IN HCC 1115
Subcellular Localization of �-Catenin
To explore the hypothesis that Wnt-Frizzled sig-al may act
through the canonical �-catenin pathway toontrol cellular motility,
the �-catenin status in HCCell lines was evaluated. We found that
Huh7, Focus,nd Hep3B cells have a homozygous wild-type �-catenins
described previously.24 HepG2 cells have a heterozy-ous deletion in
exon-3 and make both wild-type andutant �-catenin proteins.37 It is
known that �-cateninay be bound to either the cytosolic
GSK3�/Axin/APC
omplex or membrane-linked with E-cadherin or c-Met.35
n signaling by Wnt through its Frizzled receptor, dishev-led
subsequently interacts with the GSK3�/Axin/APC/�-atenin complex,
causing dephosphorylation and dissocia-ion and �-catenin from the
complex, followed by itsuclear translocation. Thus, nuclear vs.
cytosolic localiza-
igure 3. Western blot analysis of total cellular �-catenin and
Tcfeporter assay. (A) There is no correlation of proliferating cell
nuclearntigen with FZD7 mRNA. (B) However, high levels of FZD7
genexpression are associated with nuclear accumulation of
�-catenin;
ower FZD7 expression levels were associated with both nuclear
andytoplasmic accumulation in HepG2 and Hep3B cells. (C)
Measure-ent of Tcf transcriptional activity. These HCC cell lines
were cotrans-
ected with either TOPflash (�) or FOPflash (□) and
�-galactosidase–xpressing plasmid. Note the high level of
Tcf-mediated transcriptionctivity in all 3 cell lines. There was a
general correlation betweenZD7 levels and Tcf transcriptional
activity.
ion of wild-type �-catenin was explored in these HCC cellines.
The �-catenin and protein levels are shown in FigureA. It was of
interest that 2 HCC cell lines, namely Huh7nd Focus, with the
highest FZD7 gene expression, had a
igure 4. Construction and expression of the FZD7 mutant
proteins. (A)ZD7-FL, wild-type full-length FZD7 protein; FZD7-�C,
FZD7 protein trun-ated in the intracellular domain; FZD7-�T�C,
secreted FZD7 proteinruncated in both the transmembrane and
intracellular domains; CRD,ystein-rich domain. The Wnt-ligand
interacts with the CRD of FZD7 andith the epidermal growth factor
repeat (EGFR) regions 1 and 2 ofRP5/6 as represented by grey boxes.
(B) Expression in a heterogenousopulation of Huh7 cells stably
transfected with pcDNA3/FZD7-�T�C orcDNA3/empty vector as a
control. Expression was assessed by quan-itative real-time RT-PCR
with either a first set of primers located in thextracellular
domain, or a second set located in the intracellular domainf FZD7
receptor. Expression levels are normalized to the values ob-ained
with pcDNA3. (C) Expression of FZD7-�T�C as assessed byestern blot
analysis with the monoclonal anti-M2 flag tag antibody in
aeterogenous population of Huh7 cells stably transfected with
FZD7-T�C or pcDNA3/empty vector. (D) Expression of various forms of
FZD7
n Huh7 cells. Cells were stably transfected with pLenti6/V5-GFP,
clone-4GFP-C4) served as a negative control; Huh7 cells were stably
transfectedith pLenti6/V5-FZD7-�C, clones 1 to 8 (�C-C1–C8); Huh7
cells tran-iently transduced with pLenti6/V5-FZD7-�C (�C-Tr). Huh7
cells stablyransduced with pLenti6/V5-FZD7-FL, clone-1 (FL-C1)
served as a posi-ive control for RT-PCR assessment with sets of
primers targeting eitherhe extracellular domain or the
intracellular domain of FZD7 receptor.xpression levels are
normalized to the values observed with pLenti6/5-GFP-C4. (E)
Expression of FZD7-�C as assessed by Western blotnalysis with
rabbit polyclonal anti-human FZD7 antibody with or
withoutmmunoprecipitation with goat anti-human V5 antibodies
coupled togarose beads. Experiments were performed in Huh7 cells
transducedith pLenti6/V5-FZD7-�C and pLenti6/V5-GFP-C4 as a
negative control.
-
snaa(
bcccctif
wdetFmhechdawtoamlbthdchhFbfeiF
eF
seocs�htdmt
FscHFtccF(oFmevsc
1116 MERLE ET AL. GASTROENTEROLOGY Vol. 127, No. 4
triking nuclear subcellular localization of wild-type �-cate-in
(Figure 3B). Furthermore, the levels of FZD7 mRNAppear not to
correlate with the cellular proliferation index,s assessed by
levels of proliferating cell nuclear antigenFigure 3A).
In addition, Huh7, Focus, and Hep3B showed highasal Tcf
transcriptional activity and there was a generalorrelation between
the level of activity and FZD7 re-eptor expression (Figure 3C).
These results also areonsistent with the levels of �-catenin found
within theell and its nuclear localization. These findings
suggesthat activation of the canonical Wnt/�-catenin pathways
involved in the multistep process of hepatocyte trans-ormation.
Expression of a Dominant-Negative FZD7Receptor Mutants Inhibits
Wild-Type �-Catenin Accumulation and Motility of HCCCells
We assessed the motility and accumulation ofild-type �-catenin
in human HCC cell lines afterown-regulation of the Wnt-FZD7 signal
after ectopicxpression of 2 types of dominant-negative FZD7 mu-ant
receptors: (1) the FZD7-�T�C represents a secretedZD7 receptor in
which both intracellular and trans-embrane domains have been
truncated; this construct
as been shown to inhibit Wnt signaling in humansophageal
carcinoma cell lines26; and (2) the FZD7-�Construct in which the
transmembrane FZD7 receptoras been truncated in the intracellular
domain only asepicted in Figure 4A. Because FZD7-�T�C receptorcts
as a soluble Wnt ligand binding protein, studiesere performed on a
heterogenous population of stably
ransfected cells. HCC cells were selected in the presencef
geneticin and subsequently evaluated for motility andccumulation of
wild-type �-catenin. The FZD7-�Cutant receptor does not act as a
soluble ligand but is
inked closely to the membrane with its 7 transmem-rane domains.
The FZD7-�C cDNA was cloned intohe lentiviral vector
pLenti6/V5-D-TOPO to allow for aigh transduction efficiency in both
dividing and non-ividing cells. Preliminary experiments with a
similarlyloned GFP construct indicated that more than 80% ofuman
HCC cells will abundantly express GFP from 72ours to 3 weeks
posttransduction. Ectopic expression ofZD7-�C and FZD7-�T�C was
assessed by Westernlot and quantitative real-time RT-PCR by using
dif-erent sets of primers targeting specifically either
thextracellular or the intracellular domain of FZD7, allow-ng
differential assessment of the ectopic expression ofZD7-�C and
FZD7-�T�C by comparison with the
ndogenous expression of the wild-type full-lengthZD7 receptor
(Figure 4B–E).The secreted FZD7-�T�C ectodomain has been
hown previously to be a functional antagonist of endog-nous FZD7
signaling by suppression of the interactionf APC with �-catenin in
the KYSE150 esophagealarcinoma cell line.26 We now show that
ectopic expres-ion of FZD7-�T�C or FZD7-�C causes a decrease
of-catenin protein levels in all HCC cell lines with aomozygous
wild-type �-catenin gene, presumablyhrough down-regulation of the
Wnt-FZD7 signal trans-uction cascade (Figure 5A, B). In contrast,
�-cateninRNA steady-state levels were not altered by transfec-
ion of the dominant-negative mutant constructs as as-
igure 5. �-catenin protein levels in HCC cells after ectopic
expres-ion of a dominant-negative FZD7-�T�C mutant
receptor–expressingonstruct. (A) Levels of �-catenin protein in
Huh7, Focus, Hep3B, andepG2 HCC cell lines transfected with the
secreted form of theZD7-�T�C receptor. (B) Level of �-catenin
protein in Huh7 cells afterransduction with the transmembrane
FZD7-�C mutant receptorloned into pLenti6/V5. (C) Motility of Huh7
cells toward solubleollagen-I under conditions of transient
transfection with pLenti6/V5-ZD7-�T�C, pLenti6/V5-FZD7-�C, or
pLenti6/V5-GFP as a control.D) Motility measurements of
representative clones under conditionsf stable integration and
expression of: C1 � FZD7-�T�C, C5 �ZD7-�C, and C6 � FZD7-�C. The
percent of nonmigrated and totaligrated (migrated-adherent and
migrated nonadherent) cells werevaluated by a luminescent-based
assay after 3 hours at 37°C. Thealues (percent of cells), expressed
as mean � SD, are from 6eparate measurements. t test, *P 0.05, **P
0.01 whenompared with GFP-negative control.
-
snsh
ttWs
aicImti
FmpmmmcvGTnDebem�r�
FmmNwbaa�faatme
October 2004 FRIZZLED GENE OVEREXPRESSION IN HCC 1117
essed by real-time RT-PCR (data not shown). However,o effects of
these mutant receptor constructs were ob-erved on �-catenin levels
in HepG2 cells, which have aeterozygous �-catenin deletion (Figure
5A).
If FZD7 potentially is involved in cellular motility viahe
canonical �-catenin pathway, it becomes importanto evaluate the
impact of down-regulation of the
nt-FZD7 signal on motility of Huh7 cells.33 Wehowed that ectopic
expression of the transmembrane
igure 6. Schematic representation of �-catenin constructs and
cellotility assay. (A) Approximate location of functional domains:
black,rotein instability; right-leaning hatch, N-terminal
transactivation do-ain; left-leaning hatch, C-terminal
transactivation domain; grey, ar-adillo repeats, protein-protein
interaction. (B) Motility measure-ents of FZD7-�C-C6 and GFP-C4
blasticidin-selected clonal Huh7
ell populations initially transduced with pLenti6/V5-D-TOPO
lentiviralectors expressing either the FZD7-�C negative-dominant
mutant orFP as control, and cotransduced once again with a
pLenti6/V5-D-OPO lentiviral vector expressing a biologically active
�N/�C �-cate-in mutant or GFP to keep constant the total amounts of
plasmidNA. The percent of nonmigrated and total migrated
(migrated-adher-nt and migrated nonadherent) cells were evaluated
by a luminescent-ased assay after 3 hours at 37°C. The values
(percent of cells)xpressed as mean � SD were derived from 6
separate measure-ents. Note the restoration of cell motility by the
�N/�C mutant-catenin in the setting of inhibition of the signal
transduction of theeceptor level with stable expression of the
dominant-negative FZD7C-C6 receptor mutant protein.
nchored FZD7-�C mutant receptor was very effectiven reducing the
number of motile Huh7 cells underonditions of transient expression
as shown in Figure 5C.n addition, this construct was more efficient
in reducingotility than the secreted FZD7-�T�C mutant recep-
or. These observations were confirmed further by strik-ng
inhibitory effects on cell motility by 2 independent
igure 7. (A) Quantitative real-time RT-PCR assessment of FZD7RNA
levels in human HCC tumors (T) and the corresponding peritu-orous
liver parenchyma (PT), derived from Taiwan and South
Africa.onparametric paired test, *P 0.0001; paired t test, P �
0.0187,hen comparing levels in tumor with peritumoral areas. (B)
Westernlot analysis of FZD7 receptor protein expression in HCC
tumors (T)nd the corresponding peritumorous areas (PT), as well as
in Huh7nd HepG2 human hepatoma cell lines. (C) Western blot
analysis of-catenin protein accumulation in cytosolic (C) or
nuclear (N) enriched
ractions from 2 HCC tumors and their corresponding
peritumoralreas compared with 2 normal liver samples. Both
peritumoral areand tumor overexpress FZD7 mRNA as shown by values
listed belowhe Western blots and expressed as relative abundance of
FZD7RNA. Each tumor and peritumor region had a wild-type
�-cateninxon-3 as assessed by PCR and sequencing.
-
s(p(
camehppv(tgbcvtIdpcmmcc
tlTtpaPtRAbF(2wvmt
e
edaagdHnpttFfi
T
A
N
T
1118 MERLE ET AL. GASTROENTEROLOGY Vol. 127, No. 4
table clones expressing either FZD7-�C at high levelsFZD7-�C-C5,
and FZD7-�C-C6, respectively), com-ared with a FZD7-�T�C–
expressing clonal cell lineFZD7-�T�C-C1) as shown in Figure 5D.
To establish a clear link between regulation of theanonical
Wnt/�-catenin pathway and the FZD7-medi-ted alteration of cell
motility in Huh7 cells, experi-ents were performed to restore cell
motility by ectopic
xpression of a mutant �-catenin construct in Huh7 cellsarboring
a low-motility phenotype caused by overex-ression of a
dominant-negative FZD7 receptor mutantrotein. We generated a
pLenti6/V5-D-TOPO lenti-iral vector expressing the �N/�C �-catenin
mutantFigure 6A) that had been shown previously to
exhibitransactivating properties on Lef/Tcf regulated targetenes in
HEK cells.31 The FZD7-�C-C6 and GFP-C4lasticidin-selected clonal
Huh7 cell populations wereotransduced by the �N/�C �-catenin mutant
lenti-irus construct or GFP as a control to keep constanthe total
amounts of plasmid DNA within Huh7 cells.n this context, we
assessed the motility of cotrans-uced Huh7 cells and showed that
the low-motilityhenotype induced by stable FZD7-�C expressionould
be reversed by ectopic expression of the �N/�Cutant �-catenin
protein (Figure 6B). These experi-ents further support a role for
activation of the
anonical Wnt/�-catenin pathway during hepatic on-ogenesis in
cells overexpressing FZD7 receptors.
FZD7 mRNA is Overexpressed in HumanHCC Tumors
Of importance was to extend our in vitro observa-ions into the
realm of human biology and disease. Normaliver showed low FZD7 mRNA
expression (Figure 7A andable 1). Among 30 paired samples tested
from 2 regions of
he world, 27 (90%) displayed significant FZD7 overex-ression in
comparison with corresponding peritumorousreas (nonparametric
paired test, P 0.0001; paired t test,� 0.0187) (Figure 7A). These
findings were confirmed at
he protein level by Western blot analysis (Figure 7B).esults
were similar between tumors derived from Southfrica and Taiwan with
an average increase of 12-foldetween tumor and peritumorous areas.
High levels ofZD7 mRNAs as defined by a value above the cut-off
level
mean of normal liver � 3 SD, � � 0.01) were observed in9 tumors
(97%) and 23 peritumoral areas (77%) comparedith completely normal
adult liver controls. These obser-ations led us to believe that
FZD7 mRNA up-regulationay be an early event occurring in the
preneoplastic peri-
umorous area of the liver.Similar to our in vitro observations
in HCC cell lines, we
xplored �-catenin levels in comparison with FZD7 gene
xpression in human tumors. It was found that, among 5ifferent
HCC-expressing FZD7 mRNA at 100-foldbove the mean of normal livers,
4 of the 5 HCC weressociated with a homozygous wild-type �-catenin
exon-3ene. However, 1 tumor (HCC #28) had a heterozygouseletion of
1 �-catenin allele similar to that found in theepG2 cell line.
Western blot analysis of cytosolic and
uclear enriched fractions showed that wild-type �-cateninrotein
also accumulates in HCC and peritumorous areas inhe context of
increased FZD7 mRNA expression and inhe absence of �-catenin exon-3
mutations as shown inigure 7C. The findings in human tumors,
therefore, con-rm our in vitro observations in HCC cell lines.
able 1. FZD7 mRNA Steady-State Levels Assessed byQuantitative
Real-Time RT-PCR in Paired SamplesIncluding HCC Tumor Versus the
CorrespondingPeritumorous Liver
Geographicarea
HCCsample,
no. Cause
FZD7 mRNAlevels
RatioT/pTT pT
bnormal liverTaiwan 1 HBV 18.5 7.3 2.5
2 HBV 46.5 6.9 6.73 HBV 11.6 5.8 2.04 HBV 53.1 7.6 7.05 HBV 60.7
9.5 6.46 HBV 16.0 3.3 4.87 HBV 31.3 14.2 2.28 HBV 17.1 18.5 0.99
HBV 9.5 13.1 0.7
10 HBV 19.6 6.5 3.011 HBV 134.2 19.6 6.812 HBV 657.1 10.2 64.413
HBV 38.2 18.5 2.114 HBV 9.5 3.6 2.615 HBV 84.0 30.9 2.716 HBV 38.2
6.9 5.5
South Africa 17 HBV 442.9 33.5 13.218 HBV 8.4 0.7 12.019 HBV 4.0
0.4 10.020 Hemochromatosis 0.7 4.4 0.221 HBV 33.5 6.2 5.422 HBV
31.3 10.2 3.123 HBV 4.7 1.1 4.324 HBV 7.3 0.4 18.325 HCV 29.1 3.3
8.826 HBV 14.9 2.2 6.827 HBV 33.1 4.7 7.028 Unknown 296.4 2.2
134.729 Hemochromatosis 14.9 8.4 1.830 HBV 20.7 1.8 11.5
ormal liverTaiwan 31 0.4
32 0.7South Africa 33 1.8
34 1.1
, tumor; pT, peritumorous; HBV, hepatitis B virus.
-
PuobrbehauafgthHthdMchbol
HaoAaqntptlttsonwbaF
Fe
laapmatoWecmp
cATqweaaesd(�wsFpnthbtFmchsF��ecW
vlca
October 2004 FRIZZLED GENE OVEREXPRESSION IN HCC 1119
Discussion
We have provided direct evidence by real-time RT-CR that FZD7
receptor gene expression commonly isp-regulated and is an early
event during the developmentf HCCs. We also have developed a rabbit
polyclonal anti-ody to a unique peptide of FZD7 to confirm that
thiseceptor also is up-regulated at the protein level as well asy
Western blot analysis. The biologic consequences of thisvent are
the stabilization of wild-type �-catenin and en-anced tumor cell
migration. Defining the genetic alter-tions associated with
hepatocarcinogenesis is essential fornderstanding the progression
of this disease. In this regard,ctivation of cellular oncogenes
(i.e., c-myc)38 and growthactors (i.e., insulin-like growth factor
II and transformingrowth factor type �)39–41 in association with
mutations ofumor suppressor genes (i.e., P53, RB, and
IGFIIR)40,42,43
ave been described in human and animal models of HCC.owever,
these genetic events are present in 50% of HCC
umors and usually occur late in the multistep process
ofepatocarcinogenesis and not in precancerous lesions ofysplasia,
cirrhosis, and noncirrhotic chronic hepatitis.ore recently,
aberrant accumulation of the potential on-
ogenic �-catenin protein, caused by mutations of the gene,as
been found in both human and murine HCC tumors,ut such mutations
are unusual (15% of HCCs) and of lateccurrence because they are
absent in dysplasia, cirrhosis, oriver fibrosis.2,44
Additional studies have revealed that in 35%–80% ofCCs, aberrant
accumulation of �-catenin is not associ-
ted with mutations of �-catenin, Axin1, or APC genesf this Wnt
inducible signal transduction pathway.18–21,45
lthough p53 gene mutations could contribute, in part, toberrant
accumulation of wild-type �-catenin, the fre-uency of gene
mutations is relatively low. The level ofuclear and cytoplasmic
�-catenin accumulation in suchumors therefore remains
unexplained.24 We explored theossibility that other members of the
Wnt/�-catenin signalransduction pathway could contribute to
aberrant accumu-ation of wild-type �-catenin in HCC cells. During
initia-ion of the Wnt signaling cascade, binding of a Wnt ligando
its target, the Frizzled receptor, may lead either to
thetabilization of intracellular �-catenin protein7 or activationf
downstream molecules such as c-Jun NH2-terminal ki-ase, and protein
kinase C.8 However, if Frizzled receptorsere overexpressed, this
signal transduction pathway mighte constitutively activated. Human
HCC cell lines, as wells tumors, were evaluated and the biologic
consequences ofZD7 overexpression were assessed.All HCC cell lines
were found to overexpress the
ZD7 gene at different levels. These observations werextended to
human disease. Indeed, analysis of human
iver tumors showed frequent FZD7 gene up-regulations compared
with normal liver and adjacent uninvolvedreas. The general finding
of significant overexpression inreneoplastic peritumoral tissue as
compared with nor-al liver suggests that up-regulation of this gene
may be
n early event in hepatocarcinogenesis. Once
completeransformation has occurred, higher levels of FZD7
werebserved. These results imply that activation of the
nt/�-catenin signaling owing to FZD7 receptor over-xpression
alone or possibly in association with LRP-oreceptor and Wnt ligand
expression or overexpression,ay be one of the major early events of
the stepwise
rocess leading to hepatocyte transformation.Activation of the
canonical Wnt pathway results in ac-
umulation of the �-catenin–free pool in the cytoplasm.fter
forming a transcriptional transactivator complex withcF/Lef,7 there
is translocation to the nucleus. Subse-uently, transactivation of
genes involved in cell migrationill occur. We found that high
levels of FZD7 mRNA
xpression were associated almost exclusively with nuclearnd/or
cytoplasmic accumulation of �-catenin in HCC linesnd human tumors
with the wild-type gene. However, as anxception, we observed
up-regulation of FZD7 gene expres-ion in the context of a
heterozygous �-catenin exon-3 geneeletion in one HCC tumor (HCC
#28) and cell lineHepG2). It is possible that an internal deletion
of the-catenin gene abolishes phosphorylation sites (localizedithin
exon-3) and abrogates the canonical Wnt-Frizzled
ignaling via the mutated �-catenin allele.46 In this setting,ZD7
may signal through either in the canonical �-cateninathway via the
wild-type �-catenin allele, or may act in aoncanonical
�-catenin–independent pathway involvinghe activation of protein
kinase C.8 More importantly, weave shown that down-regulation of
Wnt-FZD7 signalingy ectopic expression of different
dominant-negative mu-ants of FZD7 receptors (i.e., the
transmembraneZD7-�C or the secreted FZD7-�T�C) is associated
witharkedly reduced accumulation of �-catenin protein in the
ytoplasm and nucleus of HCC cell lines that display aomozygous
wild-type �-catenin gene. However, our ob-ervations on the
inefficiency of dominant-negative mutantZD7 receptors to reduce
accumulation of the wild-type-catenin protein in the environment of
a heterozygously-catenin–mutated gene, as we have shown in HepG2
cells,mphasizes the potential of cooperation between noncanoni-al
and �-catenin–independent pathways for mediating thent-FZD7
signaling, however, further studies are required.The Wnt/Frizzled
signaling network influences di-
erse biological processes.4 The �-catenin protein be-ongs to a
family of structural proteins that includesatenins and cadherins.
By forming a membrane-associ-ted complex, these proteins mediate
adhesion and are
-
etgcmtwcnwdmtTemorstyrbhtsttcmoatho
pbrcsrewTcotwmgi
lFa
csitsCsh�
1
1
1
1
1
1
1120 MERLE ET AL. GASTROENTEROLOGY Vol. 127, No. 4
ssential for the processes of cellular motility and
migra-ion.33,47 Therefore, we investigated the role of FZD7ene
expression in HCC cell motility. An ATP lumines-ence-based assay
was used to quantify directional cellotility values in chemotaxis
chambers and we observed
hat the magnitude of steady-state FZD7 mRNA levelsas correlated
strongly with enhanced motility of HCC
ells. To further characterize the role of Wnt-FZD7 sig-al on
cell motility, experiments showed that interferingith the Wnt-FZD7
interaction by ectopic expression ofominant-negative mutant FZD7
receptors led toarked reduction of HCC cell migration and occurred
in
he context of a homozygous wild-type �-catenin gene.hese results
support previous findings on the inhibitoryffect of natural
secreted Frizzled-related proteins on theotility of human glioma
cells.48 Of interest was the
bservation that the transmembrane FZD7-�C mutanteceptor was more
effective than the secreted FZD7-�T�Coluble receptor with respect
to inhibiting Huh7 cell mo-ility and migration. It is likely that
Huh7 cells can secreteet unidentified Wnt ligand for binding to the
FZD7eceptor. The secreted FZD7-�T�C mutant receptor maye less
efficient in binding, or could be overwhelmed by theigh number of
secreted Wnt ligand molecules. In contrast,he transmembrane FZD7-�C
mutant receptor may beaturated and efficient because it has the
capability to bindo the LRP5 co-receptor to allow for optimal
transmission ofhe Wnt signal. Finally, our experiments suggest that
theanonical �-catenin pathway may be involved in the FZD7-ediated
regulation of Huh7 cell motility. Indeed, we have
bserved that ectopic expression of a �-catenin mutant wasble to
restore high levels of motility in Huh7 cells wherehe
Wnt/�-catenin–mediated signal transduction cascadeas been inhibited
at the receptor level by stable expressionf a dominant-negative
FZD7-�C construct.
In summary, deregulation of the Wnt-APC-�-cateninathway is found
in a number of human (colorectal, lung,reast, cervix, melanoma, and
HCC) tumors.9–17 Mosteports have established that this deregulation
may beaused by �-catenin gene mutation.2 Surprisingly, ourtudy
shows that over 90% of human HCCs have up-egulation of the FZD7
receptor gene and this phenom-non was associated functionally with
stabilization ofild-type �-catenin and enhanced HCC cell
motility.herefore, enhanced FZD7 gene expression is the mostommon
and one of the earliest genetic abnormalitiesbserved thus far in
HCC and probably is responsible forhe �-catenin accumulation in
human HCC tumorsithout �-catenin, axin, or APC mutations.
Molecularechanisms that may lead to up-regulation of the FZD7
ene are unknown. There are various considerations thatnclude (1)
paracrine or autocrine induction by Wnt
igands, (2) gene amplification, and (3) demethylation ofZD7 gene
promoter sequences. Further investigationsre required to test these
possibilities.
Although we have emphasized the canonical Wnt/�-atenin signaling
pathway in this report, Wnt/Frizzledignals may activate at least 2
other intracellular signal-ng pathways, including the planar cell
polarity pathwayhat signals through the small GTPase Rho, and
anotherignaling cascade that activates isoforms of protein
kinase.27,49 In the future, we will need to focus on under-
tanding these pathways as well in the context of aomozygous
wild-type �-catenin gene or in tumors with-catenin mutations and
deletions.
References1. Feitelson MA, Sun B, Tufan NLS, Liu J, Pan J, Lian
Z. Genetic
mechanisms of hepatocarcinogenesis. Oncogene
2002;21:2593–2604.
2. de la Coste A, Romagnolo B, Billuart P, Renard CA, Buendia
MA,Soubrane O, Fabre M, Chelly J, Beldjord C, Kahn A, Perret
C.Somatic mutations of the �-catenin gene are frequent in mouseand
human hepatocellular carcinoma. Proc Natl Acad Sci U S
A1998;95:8847–8851.
3. Bhanot P, Brink M, Samos CH, Hsieh JC, Wang Y, Macke
JP,Andrew D, Nathans J, Nusse R. A new member of the frizzledfamily
from Drosophila functions as a Wingless receptor.
Nature1996;6588:225–230.
4. Jones SE, Jomary C. Secreted Frizzled-related proteins:
searchingfor relationships and patterns. Bioessays
2002;24:811–820.
5. Aberle H, Schwartz H, Kemler R. Cadherin-catenin complex:
pro-tein interactions and their implications for cadherin
function.J Cell Biochem 1996;61:514–523.
6. Polakis P. The adenomatous polyposis coli (APC) tumor
suppres-sor. Biochim Biophys Acta 1997;1332:F127–F147.
7. He TC, Sparks AB, Rago C, Hermeking H, Zawel L, Da Costa
LT,Morin PJ, Vogelstein B, Kinzler KW. Identification of c-MYC as
atarget of the APC pathway. Science 1998;5382:1438–1441.
8. Kuhl M, Sheldahl LC, Park M, Miller JR, Moon RT. The
Wnt/Ca2�
pathway. A new vertebrate Wnt signaling pathway takes
shape.Trends Genet 2000;16:279–283.
9. Candidus S, Bischoff P, Becker KF, Hofler H. No evidence
formutations in the alpha- and beta-catenin genes in human
gastricand breast carcinomas. Cancer Res 1996;56:49–52.
0. Chung DC. The genetic basis of colorectal cancer: insights
intocritical pathways of tumorigenesis. Gastroenterology
2000;119:854–865.
1. Hommura F, Furuuchi K, Yamazaki K, Ogura S, Kinoshita
I,Shimizu M, Moriuchi T, Katoh H, Nishimura M, Dosaka-Akita
H.Increased expression of beta-catenin predicts better prognosis
innon-small cell lung carcinomas. Cancer 2002;94:752–758.
2. Inagawa S, Itabashi M, Adachi S, Kawamoto T, Hori M,
Shimazaki J,Yoshimi F, Fukao K. Expression and prognostic roles of
�-catenin inhepatocellular carcinoma: correlation with tumor
progression andpostoperative survival. Clin Cancer Res
2002;8:450–456.
3. Johnsson M, Borg A, Nilbert M, Andersson T. Comments
oninvolvement of adenomatous polyposis coli (APC) �-catenin
sig-naling in human breast cancer. Eur J Cancer
2000;36:242–248.
4. Lin SY, Xia W, Wang JC, Kwong KY, Spohn B, Wen Y, Pestell
RG,Hung MC. �-catenin, a novel prognostic marker for breast
cancer:its roles in cyclin D1 expression and cancer progression.
ProcNatl Acad Sci U S A 2000;7:4262–4266.
5. Park WS, Oh RR, Park JY, Kim PJ, Shin MS, Lee JH, Kim HS,
LeeSH, Kim SY, Park YG, An WG, Kim HS, Jang JJ, Yoo NJ, Lee JY.
-
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
October 2004 FRIZZLED GENE OVEREXPRESSION IN HCC 1121
Nuclear localization of �-catenin is an important prognostic
factorin hepatoblastoma. J Pathol 2001;193:483–490.
6. Rimm DL, Caca K, Hu G, Harrison FB, Fearon ER. Frequent
nuclear/cytoplasmic localization of �-catenin without exon 3
mutations inmalignant melanoma. Am J Pathol 1999;154:325–329.
7. Ueda M, Gemmill RM, Wset J, Winn R, Sugita M, Tanaka M,
UekiM, Drabkin HA. Mutations of the beta- and gamma-catenin
genesare uncommon in human lung, breast, kidney, cervical and
ovar-ian carcinomas. Br J Cancer 2001;85:64–68.
8. Devereux TR, Stern MC, Flake GP, Yu MC, Zhang ZQ, London
SJ,Taylor JA. CTNNB1 mutations and beta-catenin protein
accumu-lation in human hepatocellular carcinomas associated with
highexposure to aflatoxin B1. Mol Carcinog 2001;31:68–73.
9. Hsu HC, Jeng YM, Mao TL, Chu JS, Lai PL, Peng SY.
�-cateninmutations are associated with a subset of low-stage
hepatocel-lular carcinoma negative for hepatitis B virus and with
favorableprognosis. Am J Pathol 2000;157:763–770.
0. Wong CM, Fan ST, Ng IO. �-catenin mutation and
overexpressionin hepatocellular carcinoma: clinicopathologic and
prognostic sig-nificance. Cancer 2001;92:136–145.
1. Laurent-Puig P, Legoix P, Bluteau O, Belghiti J, Franco D,
Binot F,Mones G, Thomas G, Bioulac-Sage P, Zucman-Rossi J.
Geneticalterations associated with hepatocellular carcinomas define
dis-tinct pathways of hepatocarcinogenesis. Gastroenterology
2001;120:1763–1773.
2. Satoh S, Daigo Y, Furukawa Y, Kato T, Miwa N, Nishiwaki
T,Kawasoe T, Ishiguro H, Fujita M, Tokino T, Sasaki Y, Imaoka
S,Murata M, Shimano T, Yamaoka Y, Nakamura Y. AXIN1 muta-tions in
hepatocellular carcinomas, and growth suppression incancer cells by
virus-mediated transfer of AXIN1. Nat Genet2000;24:245–250.
3. Huang H, Fuji H, Sankila A, Mahler-Araujo BM, Matsuda M,
Catho-mas G, Ohgaki H. �-catenin mutations are frequent in
humanhepatocellular carcinomas associated with hepatitis C virus
in-fection. Am J Pathol 1999;155:1795–1801.
4. Cagatay T, Ozturk M. P53 mutation as a source of aberrant
�-cate-nin accumulation in cancer cells. Oncogene
2002;21:7971–7980.
5. Holcombe RF, Marsh JL, Waterman ML, Lin F, Milovanovic
T,Truong T. Expression of Wnt ligands and Frizzled receptors
incolonic mucosa and in colon carcinoma. Mol. Pathol
2002;55:220–226.
6. Tanaka S, Akiyoshi T, Mori M, Wands JR, Sugimachi K. A
novelfrizzled gene identified in human esophageal carcinoma
medi-ates APC/�-catenin signals. Proc Natl Acad Sci U S A
1998;95:10164–10169.
7. Weeraratna AT, Jiang Y, Hostetter G, Rosenblatt K, Duray P,
BittnerM, Trent JM. Wnt5a signaling directly affects cell motility
and inva-sion of metastatic melanoma. Cancer Cell
2002;3:279–288.
8. Gripon P, Rumin S, Urban S, Le Seyec J, Glaise D, Cannie
I,Guyomard C, Lucas J, Trepo C, Guguen-Guillouzo C. Infection of
ahuman hepatoma cell line by hepatitis B virus. Proc Natl Acad SciU
S A 2002;24:15655–15660.
9. de la Monte SM, Lahousse SA, Carter J, Wands JR. ATP
lumines-cence-based motility-invasion assay. Biotechniques
2002;33:98–100.
0. Carloni V, Mazzocca A, Pantaleo P, Cordella C, Laffi G,
Gentilini P.The integrin, �6�1, is necessary for the
matrix-dependent acti-vation of FAK and MAP kinase and the
migration of humanhepatocarcinoma cells. Hepatology
2001;34:42–49.
1. Aoki M, Sobek V, Maslyar D, Hecht A, Vogt PK. Oncogenic
trans-formation by �-catenin: deletion analysis and
characterization ofselected target genes. Oncogene
2002;21:6983–6991.
2. Maloney JA, Tsygankova O, Szot A, Yang L, Li Q, Williamson
JR.Differential translocation of protein kinase C isozymes by
phorbolesters, EGF, and ANG II in rat liver WB cells. Cell Physiol
1998;274:974–982.
3. Muller T, Choidas A, Reichmann E, Ullrich A. Phosphorylation
andfree pool of beta-catenin are regulated by tyrosine kinases
andtyrosine phosphatases during epithelial cell migration. J
BiolChem 1999;15:10173–10183.
4. Bieche I, Laurendeau I, Tozlu S, Olivi M, Vidaud D, Lidereau
R,Vidaud M. Quantitation of MYC gene expression in sporadicbreast
tumors with a real-time reverse transcription-PCR assay.Cancer Res
1999;59:2759–2765.
5. Monga SPS, Mars WM, Pediaditakis P, Bell A, Mule K, Bowen
WC,Wang X, Zarnegar R, Michalopoulos GK. Hepatocyte growth
factorinduces Wnt-independent nuclear translocation of �-catenin
afterMet-�-catenin dissociation in hepatocytes. Cancer Res
2002;62:2064–2071.
6. Satyamoorthy K, Li G, Vaidya B, Patel D, Herlyn M.
Insulin-likegrowth factor-1 induces survival and growth of
biologicallyearly melanoma cells through both the mitogen-activated
pro-tein kinase and �-catenin pathways. Cancer Res
2001;61:7318–7324.
7. Carruba G, Cervello M, Micell MD, Farruggio R, Notarbartolo
M,Virruso L, Giannitrapani L, Gambino R, Montalto G, CastagnettaL.
Truncated form of b-catenin and reduced expression of wild-type
catenins feature HepG2 human liver cancer cells. N Y AcadSci
1999;886:212–216.
8. Zhang XK, Huang DP, Qiu DK, Chiu JF. The expression of c-myc
andc-N-ras in human cirrhotic livers, hepatocellular carcinomas and
livertissue surrounding the tumors. Oncogene 1990;5:909–914.
9. Cariani E, Lasserre C, Seurin D, Hamelin B, Kemeny F, Franco
D,Czech MP, Ullrich A, Brechot C. Differential expression of
insulin-likegrowth factor II mRNA in human primary liver cancers,
benign livertumors, and liver cirrhosis. Cancer Res
1988;48:6844–6849.
0. de Souza AT, Hankins GR, Washington MK, Orton TC, Jirtle
RL.M6P/IGF2R gene is mutated in human hepatocellular carcino-mas
with loss of heterozygosity. Nat Genet 1995;11:447–449.
1. Yeh YC, Tsai JF, Chuang LY, Yeh HW, Tsai JH, Florine DL, Tam
JP.Elevation of transforming growth factor alpha and its
relationshipto the epidermal growth factor and alpha-fetoprotein
levelsin patients with hepatocellular carcinoma. Cancer Res
1987;47:896–901.
2. Hsu IC, Metcalf RA, Sun T, Welsh JA, Wang NJ, Harris
CC.Mutational hotspot in the p53 gene in human
hepatocellularcarcinomas. Nature 1991;350:427–428.
3. Zhang X, Xu HJ, Murakami Y, Sachse R, Yashima K, Hirohashi S,
HuSX, Benedict WF, Sekiya T. Deletions of chromosome 13q,
muta-tions in retinoblastoma 1, and retinoblastoma protein state in
hu-man hepatocellular carcinoma. Cancer Res 1994;54:4177–4182.
4. Devereux TR, Anna CH, Foley JF, White CM, Sills RC, Barrett
JC.Mutation of beta-catenin is an early event in chemically
inducedmouse hepatocellular carcinogenesis. Oncogene
1999;18:4726–4733.
5. Clevers H. Axin and hepatocellular carcinomas. Nat Genet
2000;24:206–208.
6. Noort MV, Meeldijk J, van der Zee R, Destree O, Clevers H.
Wntsignaling controls the phosphorylation status of �-catenin. J
BiolChem 2002;20:17901–17905.
7. Liu D, el-Hariry I, Karayiannakis AJ, Wilding J, Chinery R,
Kmiot W,McCrea PD, Gullick WJ, Pignatelli M. Phosphorylation of
�-cateninand epidermal growth factor receptor by intestinal trefoil
factor.Lab Invest 1997;6:557–563.
8. Roth W, Wild-Bode C, Platten M, Grimmel C, Melkonyan
HS,Dichgans J, Weller M. Secreted Frizzled-related proteins
inhibitmotility and promote growth of human malignant glioma
cells.Oncogene 2000;37:4210–4220.
9. Cheon SS, Cheah AY, Turley S, Nadesan P, Poon R, Clevers
H,Alman BA. �-catenin stabilization dysregulates mesenchymal
cellproliferation, motility, and invasiveness and causes
aggressive
-
R
I6
c(0F
1122 MERLE ET AL. GASTROENTEROLOGY Vol. 127, No. 4
fibromatosis and hyperplastic cutaneous wounds. Proc Natl
AcadSci U S A 2002;10:6973–6978.
Received January 12, 2004. Accepted July 1, 2004.Address
requests for reprints to: Jack R. Wands, M.D., The Liveresearch
Center, 55 Claverick Street, 4th Floor, Providence, Rhode
sland 02903. e-mail: [email protected]; fax: (401)
444-194.Supported in part by a fellowship from Association Pour la
Re-
herche sur le Cancer (France), and National Institutes of
HealthBethesda, MD) grants CA-35711, AA-02666, NCRR COBRE-15578-2,
and Massachusetts Institute of Technology, National
Scienceoundation, Biotechnology Process Engineering Center.
-
本文献由“学霸图书馆-文献云下载”收集自网络,仅供学习交流使用。
学霸图书馆(www.xuebalib.com)是一个“整合众多图书馆数据库资源,
提供一站式文献检索和下载服务”的24 小时在线不限IP
图书馆。
图书馆致力于便利、促进学习与科研,提供最强文献下载服务。
图书馆导航:
图书馆首页 文献云下载 图书馆入口 外文数据库大全 疑难文献辅助工具
http://www.xuebalib.com/cloud/http://www.xuebalib.com/http://www.xuebalib.com/cloud/http://www.xuebalib.com/http://www.xuebalib.com/vip.htmlhttp://www.xuebalib.com/db.phphttp://www.xuebalib.com/zixun/2014-08-15/44.htmlhttp://www.xuebalib.com/
Functional Consequences of Frizzled-7 Receptor Overexpression in
Human Hepatocellular CarcinomaMaterials and MethodsHCC Tumors and
Cell LinesReal-Time RT-PCR AssayCell Motility AssayTransfection and
Retroviral Transduction StudiesImmunoprecipitation StudiesProtein
Extraction and Western Blot AnalysisSequencingStatistical
Analysis
ResultsReal-Time RT-PCR Assay for FZD7Correlation of
Steady-State FZD7 mRNA Levels and HCC MotilitySubcellular
Localization of -CateninExpression of a Dominant-Negative FZD7
Receptor Mutants Inhibits Wild-Type -Catenin Accumulation and
Motility of HCC CellsFZD7 mRNA is Overexpressed in Human HCC
Tumors
DiscussionReferences