Associations Among -TrCP, an E3 Ubiquitin Ligase Receptor, -Catenin, and NF-B in Colorectal Cancer

Associations Among �-TrCP, an E3 Ubiquitin LigaseReceptor, �-Catenin, and NF-�B in Colorectal Cancer

Andrei Ougolkov, Bin Zhang, Kaname Yamashita, Vladimir Bilim, MasayoshiMai, Serge Y. Fuchs, Toshinari Minamoto

Background: The ubiquitin–proteasome pathway is impor-tant in regulating protein signaling pathways that are in-volved in tumorigenesis. �-transducin repeat–containingproteins (�-TrCP) are components of the ubiquitin ligasecomplex targeting �-catenin and I�B� for proteasomal deg-radation and are thus a negative regulator of Wnt/�-cateninsignaling and a positive regulator of NF-�B signaling. Weanalyzed expression of �-TrCP in colorectal cancers and itsassociation with types of �-catenin subcellular localization,an indirect measure of activation. Methods: Levels of�-TrCP1 mRNA and protein were measured by quantitativereverse transcription–polymerase chain reaction and immu-noblotting, respectively, in samples of tumor and normaltissues from 45 patients with colorectal cancer. Types of�-catenin activation (diffuse or invasion edge) and NF-�Bactivation were examined by immunohistochemistry. Apo-ptosis was determined by the terminal deoxynucleotidyltransferase–mediated deoxyuridine triphosphate–biotinnick-end labeling (TUNEL) assay. All statistical tests weretwo-sided. Results: Compared with the �-TrCP1 levels innormal tissues, 25 (56%) of 45 tumors had increased�-TrCP1 mRNA and protein levels. Of the 22 (49%) tumorswith �-catenin activation, 12 had the diffuse type (i.e., nu-clear accumulation throughout the tumor) and 10 had theinvasion edge type (i.e., nuclear accumulation predominantlyin the tumor cells that formed the invasion edge). Increased�-TrCP1 levels were statistically significantly associatedwith �-catenin activation (P � .023) and decreased apoptosis(P � .035). �-TrCP accumulated in the nuclei of tumor cellsthat contained increased levels of �-TrCP1 mRNA and theactive form of NF-�B. Higher levels of �-TrCP1 mRNA weredetected in primary tumors of patients who had metastases(0.960 arbitrary units, 95% confidence interval � 0.878 to1.042) than in the tumors of patients who did not (0.722arbitrary units, 95% confidence interval � 0.600 to 0.844; P� .016). Conclusion: In colorectal cancer, increased expres-sion of �-TrCP1 is associated with activation of both�-catenin and NF-�B, suggesting that the integration ofthese signaling pathways by increased �-TrCP expressionmay contribute to an inhibition of apoptosis and tumormetastasis. [J Natl Cancer Inst 2004;96:1161–70]

Diverse molecular events are integrated in the developmentand progression of colorectal cancer, a leading cause of cancerdeath worldwide (1). A complex combination of genetic, epi-genetic, and postgenetic (e.g., posttranslational) alterations areinvolved in the multistage development of colorectal cancer(2,3). Of the genetic alterations that occur in these malignancies,the best known is the mutation and activation of the K-rasoncogene, which could serve as a marker for molecular diagno-sis and risk assessment and as a molecular therapeutic target

(4,5). Other genetic alterations include mutation or loss of theadenomatous polyposis coli (APC) gene and oncogenic activa-tion of �-catenin (the product of the CTNNB1 gene) via itsstabilization (6).

�-catenin is a multifunctional protein involved in cell–celladhesion, normal embryonic development, cell differentiation,and malignant transformation (6,7). The oncogenic properties of�-catenin are associated with its function as a transcriptionfactor in the Wnt/�-catenin/T-cell factor (Tcf) signaling path-way. In normal cells, �-catenin is associated with glycogensynthase kinase 3� (GSK3�) in a complex that includes the APCprotein. GSK3� phosphorylates �-catenin, which results in APC-mediated �-catenin degradation via the ubiquitin–proteasomepathway. The binding of Wnt proteins to their receptor inhibitsGSK3� phosphorylation activity and subsequently stabilizes�-catenin. GSK3� phosphorylation activity can also be inhibitedby phosphoinositide 3 kinase (PI3K)/protein kinase B (Akt)signaling. In tumor cells, mechanisms that inhibit GSK3�-induced phosphorylation of �-catenin block its interaction withthe E3 ubiquitin ligase receptor, �-transducin repeat–containingprotein (�-TrCP), which prevents �-catenin ubiquitination anddegradation, and ultimately leads to �-catenin activation. Thesemechanisms include mutations in the phosphorylation recogni-tion motif of �-catenin, mutations in the APC gene that result infailure to recruit GSK3� to the complex, and unregulated PI3K/Akt signaling. In both normal and transformed cells, stabilized�-catenin is translocated to the nucleus, where it interacts withTcf/lymphoid enhancer factor (Lef) proteins and activates thetranscription of a number of target genes including c-myc, cyclinD1, and matrix metalloproteinase 7 (8–10). Therefore, �-catenintranslocated into the nucleus is predominantly in an active andoncogenic form (6,8–10).

Recently, several studies have reported that the nuclear ac-cumulation of �-catenin (i.e., �-catenin activation) is an earlyevent in colorectal cancers (11,12). However, little was knownregarding the clinical relevance of oncogenic �-catenin activa-tion until we determined two distinct types of �-catenin activa-tion (13,14). We found that �-catenin activation was detectablein 50%–65% of colorectal cancers and that, in 40%–50% of

Affiliations of authors: Divisions of Diagnostic Molecular Oncology (AO, BZ,TM) and Surgical Oncology (KY, MM, TM), Cancer Research Institute, Ka-nazawa University, Kanazawa, Japan; Molecular Laboratory, Division of Urol-ogy, Graduate School of Medical and Dental Sciences, Niigata University,Niigata, Japan (VB); Department of Animal Biology, School of VeterinaryMedicine, University of Pennsylvania, Philadelphia (SYF).

Correspondence to: Toshinari Minamoto, MD, PhD, Division of DiagnosticMolecular Oncology, Cancer Research Institute, Kanazawa University, 13-1 Takara-machi, Kanazawa 920-0934, Japan (e-mail: [email protected]).

See “Notes” following “References.”

DOI: 10.1093/jnci/djh219Journal of the National Cancer Institute, Vol. 96, No. 15, © Oxford UniversityPress 2004, all rights reserved.

Journal of the National Cancer Institute, Vol. 96, No. 15, August 4, 2004 ARTICLES 1161

by guest on March 21, 2016

http://jnci.oxfordjournals.org/D

ownloaded from

http://jnci.oxfordjournals.org/

colorectal cancers, �-catenin accumulated in nuclei throughoutthe tumor (referred to as the diffuse type), but in 10%–15% ofcolorectal cancers, �-catenin accumulated in nuclei only intumor cells that formed the invasion edge (referred to as theinvasion edge type). The latter type of �-catenin activation wasstrongly associated with advanced tumor stage and recurrence.Consequently, this type of oncogenic �-catenin activation is anindependent and reliable factor that can identify a subset ofpatients with colorectal cancer who are highly susceptible totumor recurrence and thus have a less favorable survival rate(13,14).

The observation of different types of �-catenin activation issupported by evidence that �-catenin is activated in most colo-rectal cancers as a result of mutational inactivation of APC andthat mutations of its gene were detected in a minority of colo-rectal cancers with wild-type APC (15,16). Recently, we foundthat the diffuse type of �-catenin activation was closely associ-ated with loss of heterozygosity in the APC loci, whereas theinvasion edge type occurred independently of APC inactivation(Ougolkov A: unpublished data). Thus, our findings indicate thatthe diffuse type and the invasion edge type of �-catenin activa-tion are early (i.e., APC inactivation–dependent) and presum-ably late (i.e., APC inactivation–independent) events, respec-tively, in colorectal carcinogenesis. It is therefore important toclarify mechanisms underlying differences in �-catenin activa-tion types to aid in the identification and development of newdiagnostic tools and molecular therapeutic targets.

Oncogenic activation of �-catenin occurs primarily as a con-sequence of its stabilization by escaping ubiquitin-mediatedproteasomal degradation, and the effects of activated �-catenindepend on a growing number of effectors and regulators (8–10).A major regulator of �-catenin stability and activity is the�-TrCP subfamily of F-box proteins, including �-TrCP1 and�-TrCP2 (17–21). Human �-TrCPs are components of theSkp1–Cul1–F-box protein complex that regulates cellular levelsof various proteins by functioning as the E3 ubiquitin–proteinligase that targets such substrates as �-catenin, IB� (the inhib-itor of nuclear factor kappaB [NF-B]), and Emi1 cell-cycleregulator protein for degradation via the ubiquitin–proteasomepathway (22,23). The gene encoding �-TrCP has been mappedto chromosome 10q24 (24). We demonstrated that, in 293T cellsharboring wild-type CTNNB1 and APC genes, �-catenin/Tcfsignaling increased levels of �-TrCP1 mRNA and protein in aTcf-dependent manner but without increased �-TrCP1 genetranscription (25). We also showed that increased induction of�-TrCP1 expression by the �-catenin/Tcf pathway resulted inaccelerated degradation of wild-type �-catenin protein, suggest-ing that a negative feedback loop may control the �-catenin/Tcfpathway under physiologic conditions (25). In addition, induc-tion of �-TrCP was associated with the activation of the NF-Bpathway, which is known to promote cell survival (25). Whether�-catenin-dependent increases in �-TrCP levels, activation ofNF-B, and inhibition of cell death are linked in colorectalcancers is unknown.

We hypothesized that �-TrCP participates in tumorigenesisby modifying not only the activation of the �-catenin signalingpathway but also the activation of other oncogenic signalingpathways. However, because little is known about �-TrCP incancer, we examined whether �-TrCP alters the distinct type ofoncogenic �-catenin activation in colorectal cancer and howthese distinct types of �-catenin activation affect the malignant

potential of the tumors and patient outcome. We also investi-gated the relationship between �-TrCP expression, NF-B, andthe incidence of apoptosis.

PATIENTS AND METHODS

Patients

From all consecutive patients who underwent surgery for theremoval of colorectal cancer in our institute between 1998 and2002, 45 patients, for whom a complete set of samples wereavailable (described below), agreed to be enrolled in the presentstudy. All patients signed written informed consent forms. TheInstitutional Review Board of Kanazawa University GraduateSchool of Medical Science approved all study protocols. Of the45 patients, 23 were men and 22 were women. The mean patientage was 68 years (range � 44–88 years). Seventeen patients hadcancer of the cecum and ascending colon, three patients hadcancer of the transverse colon, two patients had cancer of thedescending colon, nine patients had cancer of the sigmoid colon,and 14 patients had cancer of the rectum. The Japanese Classi-fication of Colorectal Carcinoma criteria (26) were used todescribe the gross appearance and histologic characteristics ofthe primary tumors. Regarding gross appearance, five tumorshad a protruding appearance (type 1), 39 tumors had localizedand expansive growth with ulceration (type 2), and one tumorhad infiltrative proliferation (type 3). On histologic examination,all tumors were classified as adenocarcinomas, of which 19 werewell differentiated, 23 were moderately differentiated, and threewere poorly differentiated (including mucinous adenocarci-noma). Carcinoma cells had invaded lymph vessels in 19 (42%)patients and invaded blood vessels in 24 (53%) patients. Ac-cording to the Tumor–Node–Metastasis (TNM) classificationsystem (27), routine clinical examination and pathology diagno-sis indicated the following: one patient had stage I disease, 26had stage II disease, and 18 had stage III disease. Follow-upexamination after surgery disclosed metastases to distant organs(i.e., liver, lung) in eight patients.

Tissue Samples and Cell Lines

Samples from normal mucosa and tumor tissues were takenfrom each patient’s fresh surgical specimen, immediately snap-frozen, and stored at 80 °C. After sampling, the surgicalspecimens were then fixed with neutral buffered 10% formalin,embedded in paraffin, and processed for routine histopathology.We used serial sections of paraffin-embedded tumors to ensurethat all studies used samples that contained the same histopatho-logic characteristics of the respective tumor. Tissue sectionswere prepared on silica-coated glass slides (DAKO Cytomation,Kyoto, Japan) for immunohistochemical analysis of �-TrCP andNF-B expression and �-catenin activation (described below).Patients with esophageal cancer or rheumatoid arthritis providedwritten informed consent to allow use of normal esophagealmucosa and synovitis tissues, respectively, which were sub-jected to routine pathology examination.

Colorectal cancer cell lines (SW480, SW620, HCT116,HT29, and LoVo) and the 293T line, a human renal epithelialcell line that expresses simian virus 40 large T antigen, wereobtained from American Type Culture Collection (Manassas,VA) and grown in the specified media. Cells were harvestedduring the exponential growth phase, subjected to centrifugation

1162 ARTICLES Journal of the National Cancer Institute, Vol. 96, No. 15, August 4, 2004



ownloaded from


to pellet the cells, and stored as cell pellets at 80 °C until use.These colorectal cancer cell lines were chosen because theycontain genetic alterations that have resulted in �-catenin accu-mulation and activation, including mutational inactivation of theAPC tumor suppressor gene in the SW480 and SW620 (codon1338: CAGGln to TAGstop codon), HT29 (codon 853: CAGGln toTAGstop codon; and codon 1555: GAAAAAACT to GAAAAAAACT), and LoVo (codon 1114: CGAArg to TAGstop codon; andcodon 1430: AAACCAT to AAACAT) cell lines, and mutationin the phosphoacceptor site of CTNNB1 (�-catenin gene) in theHCT116 cell line (codon 45: deletion of three bases and loss ofthe serine residue) (28).

Total RNA was isolated from frozen tissue samples and cellpellets by using acid guanidinium thiocyanate (RNA-Bee; Tel-Test, Friendswood, TX) and phenol–chloroform (29,30). Cellu-lar proteins were extracted from fresh surgical specimens andpellets of cell lines using a lysis buffer (CelLytic-MT, Sigma-Aldrich, St. Louis, MO). After the protein concentration in eachsample was measured by using the Bradford method with theCoomassie Protein Assay Reagent (Pierce, Rockford, IL), thetotal protein extracts were mixed with a cocktail of proteaseinhibitors (Sigma-Aldrich) according to the manufacturer’s rec-ommendation, divided into aliquots, and stored at 80 °C.

Reverse Transcription–Polymerase Chain Reaction

Expression of �-TrCP1 mRNA was quantitatively deter-mined by reverse transcription–polymerase chain reaction (RT-PCR). Complementary DNA (cDNA) was generated from 1 �gof total RNA by reverse transcription using a Reverse Transcrip-tion System Kit (Promega, Madison, WI). To quantify �-TrCP1expression in each cDNA sample, the target (i.e., �-TrCP1) wasamplified by PCR in parallel with an internal control (glyceral-dehyde-3-phosphate dehydrogenase [GAPDH]). To amplify a211-base-pair (bp) fragment of �-TrCP1 that contains the down-stream portion of exon 9 and the upstream portion of exon 10,the following set of primers was designed: upstream primer5�-ATGCAAGCGAATTCTCACAGG-3� and downstreamprimer 5�-GGAACGATCTTTGGAGCAGGT-3�. Amplificationfor �-TrCP1 was performed by using a thermal cycling programwith varying numbers of cycles, in which each cycle consisted of94 °C for 1 minute, 59 °C for 1 minute, and 72 °C for 1 minute.An expression vector containing the �-TrCP1 gene (31) wasused as a positive control for the PCR. A 598-bp GAPDHfragment was amplified following a program with varying num-bers of cycles, in which each cycle consisted of 94 °C for 30seconds, 50 °C for 30 seconds, and 72 °C for 1 minute, and usingthe forward primer 5�-CCACCCATGGCAAATTCCATGGCA-3� and the reverse primer 5�-TCTAGACGGCAGGTCAGGTCCACC-3�. All PCR products were subjected to electro-phoresis through a native 10% polyacrylamide gel and were visu-alized by staining with SYBRGreen 1 (Nippon Gene, Tokyo, Japan).

Before we amplified and quantified �-TrCP1 expression, weperformed densitometric analysis using National Institutes ofHealth (NIH) Image Program software, version 1.62 (NIH, Be-thesda, MD) to measure and compare the levels of the GAPDHPCR products amplified by varying numbers of cycles to deter-mine the number of PCR cycles that would permit us to quantifythe target product within a linear range. �-TrCP1 was thenamplified using the quantitative parameters indicated by thecomparison (Fig. 1). Finally, we adjusted the densitometric

value of �-TrCP1 mRNA in each sample to that of GAPDHamplified during the same number of cycles to determine thelevel of �-TrCP1 expression. For each sample, we repeated allsteps (from generating cDNA to measuring the PCR product bydensitometry) to confirm reproducibility. Increased expressionof �-TrCP1 was defined as the level of expression in the primarytumor that exceeded the level of constitutive expression in thematched normal tissue by more than 50%. Because no tumor hadlower levels of expression than those in the matched normaltissue, all tumors were found to have either constitutive orincreased levels of expression.

Immunoblotting and Immunoprecipitation Analysis

Protein extracts (100 �g) from tissue samples or cell pelletswere thawed and separated by sodium dodecyl sulfate–8% poly-

Fig. 1. Constitutive expression (A) and increased expression (B) of �-transducinrepeat–containing protein 1 (�-TrCP1) mRNA from two representative sets ofnormal and tumor tissue samples from patients with colorectal cancer. TotalRNA from tumor (T) and corresponding normal (N) tissues was isolated andsubjected to quantitative reverse transcription–polymerase chain reaction (RT-PCR) amplification. Glyceraldehyde-3-phosphate dehydrogenase (GAPDH) wasamplified as an internal control for RNA integrity and to normalize �-TrCP1expression. To optimize the PCR conditions, aliquots of each PCR product weretaken from the same reaction tube after 15, 18, 21, 24, and 27 cycles ofamplification. All PCR products were separated on a 10% native polyacrylamidegel (lower panels) and were visualized by staining with SYBRGreen. Densi-tometry, using NIH Image Program (version 1.62) software, was performed tomeasure the intensity of the PCR products. Arbitrary values of intensity for eachPCR product at each time point (i.e., number of cycles) were compared on alogarithmic scale (upper panels). x-axis � number of amplification cycles;y-axis � intensity of PCR product in a 2n logarithmic scale. Constitutive�-TrCP1 expression was defined as levels of expression that were identical to thelevels of �-TrCP1 expression in the normal tissue, after normalizing for levels ofGAPDH expression. Increased �-TrCP1 mRNA expression was defined as levelsof expression that were at least 50% higher than the levels of �-TrCP1 mRNAexpression in the normal tissue, after normalizing for levels of GAPDH expres-sion. Solid square � normal mucosa; solid circle � tumor sample; solid line �GAPDH expression; dotted line � �-TrCP1 expression.




ownloaded from


acrylamide gel electrophoresis, and electro-transferred to a nitro-cellulose membrane (Amersham, Buckingham, U.K.). Trans-ferred proteins were analyzed by immunoblotting with a rabbitpolyclonal antibody against �-TrCP proteins [described in (32);diluted 1:1000] or a mouse monoclonal antibody to �-catenin[epitope described in (13); diluted 1:500] (Transduction Labo-ratories, Lexington, KY). Primary antibodies were diluted in abuffer containing 0.1% Tween-20 and 1% nonfat milk in Tris-buffered saline. Blotted membranes were incubated with pri-mary antibodies overnight at 4 °C with gentle shaking. Signalswere developed using enhanced chemiluminescence (ECL; Am-ersham, Little Chalfont, Buckinghamshire, U.K.). Protein ex-tract (100 �g) from 293T cells was used as positive control fordetecting �-TrCP and �-catenin proteins.

Expression of the NF-B active form in the tumor wasconfirmed by immunoblotting with the same antibody used forimmunohistochemistry (described below). The antibody detectstwo bands: the p65 subunit of NF-B (65 kd) and the splicingvariant of p65, delta2 (63 kd). The epidermoid carcinoma cellline A431, which has constitutively active NF-B, was used asa positive control for the detection of the p65 subunit of NF-Bby immunoblotting.

To examine the association between �-TrCP and �-catenin incolorectal cancer cell lines, a 250-�g aliquot of protein extractfrom each cell line was immunoprecipitated with the antibody to�-catenin, according to the method reported previously (33).Immunoprecipitated materials were probed serially with anti-bodies to �-TrCP and �-catenin by the immunoblotting proce-dure. Protein extract (250 �g) from 293T cells was used as apositive control.

Immunohistochemistry

The distinct patterns of subcellular localization of �-cateninand localization of �-TrCP and the p65 subunit of NF-B weredetermined immunohistochemically using the avidin–biotin–peroxidase complex method in microwaved tissue sections, asdescribed (13,14). Sections were incubated with antibodies to�-catenin (diluted to 1:100 in phosphate-buffered saline [PBS]containing 5% normal horse serum), �-TrCP (diluted 1:200 inPBS containing 5% normal goat serum), or the p65 subunit ofNF-B (diluted 1:100 in PBS containing 5% normal horseserum) overnight at 4 °C in a moist chamber. We immunostainednormal esophageal mucosa from a specimen used in our previ-ous study (13,14) as a positive control for membranous expres-sion of �-catenin, and we immunostained normal colorectalmucosa adjacent to the tumor as an internal positive control. Themonoclonal antibody to the p65 subunit of NF-B (Chemicon,Temecula, CA) recognizes an epitope (CDTDDRHRIEEKRKRKT) within the nuclear localization signal for p65, the DNAbinding subunit mainly responsible for the strong gene transac-tivating potential. Synovitis tissues removed from patients withrheumatoid arthritis were fixed in 10% buffered formalin, pro-cessed for immunohistochemistry, and stained as a positivecontrol for the NF-B p65 subunit. Creating negative controlsfor the immunohistochemistry involved replacing the primaryantibodies with nonimmune mouse and rabbit immunoglobulinG1 (IgG1) (DAKO, Glostrup, Denmark).

�-catenin subcellular expression in the primary tumor wasclassified into one of three patterns: membranous expression,similar to that found in normal crypts; diffuse nuclear accumu-

lation, defined as �-catenin-positive nuclei in cells distributedthroughout the tumor; and nuclear accumulation in invasionedge, defined as �-catenin-positive nuclei only in tumor cells atthe invasion edge adjacent to the stromal tissue with membra-nous �-catenin expression in the remaining tumor cells (Fig. 2)(13,14). Because �-catenin translocated into the nucleus is pre-dominantly in an active and oncogenic form (6,8–10), we referto the nuclear accumulation of �-catenin as �-catenin activation.Two well-trained investigators (B. Zhang and K. Yamashita)who were blinded to the histopathologic characteristics of theprimary tumors or the patients’ clinical outcomes independentlyreviewed the types of �-catenin activation and the expression of�-TrCP and the NF-B p65 subunit in each tumor specimen.The investigators agreed on the expression patterns for allsamples.

Analysis of Apoptosis

Carcinoma cells undergoing apoptosis were detected in tissuesections by the terminal deoxynucleotidyl transferase–mediateddeoxyuridine triphosphate–biotin nick-end labeling (TUNEL)method (34), using the in situ apoptosis detection TUNEL kit(Takara, Shiga, Japan) according to the manufacturer’s recom-mended protocol. The frequency of apoptosis was calculated asan apoptotic index, in which the proportion of cells undergoingapoptosis was expressed as a percentage of all carcinoma cells

Fig. 2. Expression of �-transducin repeat–containing protein 1 (�-TrCP1) and�-catenin in colorectal cancers. A) Representative samples from patients withcolorectal cancer showing increased �-TrCP1 mRNA and �-TrCP protein andlevels of �-catenin protein. Levels of glyceraldehyde-3-phosphate dehydroge-nase (GAPDH) mRNA expression were similar in paired normal (N) and tumor(T) tissues for all samples. Each pair of normal and tumor tissues is from adifferent patient. Each case is labeled with its identification number at the top ofpanel A. Extracts from 293T cells were used as a positive control. Patterns of�-catenin localization, which is indicative of �-catenin activation in colorectalcancer, were determined by immunohistochemistry using a commercially avail-able anti-�-catenin antibody and the avidin–biotin–peroxidase detection method.Staining patterns were classified as diffuse type (B) or invasion edge type (C) of�-catenin activation. The diffuse type shows nuclear �-catenin accumulation inthe cells distributed throughout the tumor. The invasion edge type shows tumorcells with nuclear �-catenin accumulation only at the invasion edge adjacent tothe stromal tissue and membranous �-catenin expression in the remaining tumorcells (C). All tumors showed either the diffuse (cases 31 and 32) or the invasionedge type (cases 51, 53, 67, and 74) of �-catenin activation.




ownloaded from


observed. The apoptotic index of each primary tumor was cal-culated as the number of TUNEL-positive cells and bodies (35)per 2500 carcinoma cells counted in five randomly selectedfields in each tumor; each field was subjected to two independentcounts. The in situ apoptosis detection TUNEL kit containedtissue sections that served as a positive control.

Statistical Analysis

Differences in expression levels, as determined by densitom-etry between paired normal and tumor tissues, were analyzed bythe Wilcoxon signed rank test. Expression levels (i.e., densito-metric values) of �-TrCP1 were compared between groups ofpatients by using the Mann–Whitney U test (for comparisonbetween two unpaired groups) and the Kruskal–Wallis test witha posttest (for comparisons among three unpaired groups). As-sociations between �-TrCP1 expression and clinical and his-topathologic variables were determined using Fisher’s exact testfor 2 � 2 tables and the exact test for large tables for 2 � 3tables. Student’s t test was used to determine whether differ-ences in the apoptotic index (mean value and 95% confidenceintervals [CIs]) between tumors with different levels of �-TrCPexpression and between those with different types of �-cateninactivation were statistically significant. In all statistical analyses,a P value of less than .05 was considered statistically significant.All statistical analyses were two-sided and were computed usingStatView (version 5.0) software for Macintosh (SAS Institute,Cary, NC).

RESULTS

To determine the number of amplification cycles that wouldallow quantitative measurement and comparison of mRNA ex-pression, the quantities of RT-PCR products from GAPDH and�-TrCP1 were compared for different numbers of PCR cycles byusing densitometry. For all samples, the quantities of GAPDHand �-TrCP1 products increased logarithmically between 18 and27 cycles of amplification (Fig. 1), and all samples had similarlevels of GAPDH expression in this range of amplificationcycles. Given these preliminary results, we selected 21 cycles ofamplification for the quantitative comparison of mRNA levels inthis study.

Constitutive expression (i.e., similar expression levels be-tween the tumor and respective normal tissue) of �-TrCP1 wasobserved in 20 (44%) of 45 samples (Fig. 1, A). Increasedexpression of �-TrCP1 was observed in the other 25 (56%)samples (representative sample shown in Fig. 1, B). The meanintensity of �-TrCP1 expression level in all tumors combined(0.759 arbitrary units, 95% CI � 0.683 to 0.835 arbitrary units)was statistically significantly higher than that in the respectivenormal tissues (0.484 arbitrary units, 95% CI � 0.408 to 0.560arbitrary units; P�.001 by the Wilcoxon signed rank test).Moreover, the mean intensity level of �-TrCP1 expression intumors with increased expression (0.835 arbitrary units, 95% CI� 0.735 to 0.935 arbitrary units) was statistically significantlyhigher than that in tumors with constitutive expression (0.663arbitrary units, 95% CI � 0.559 to 0.767 arbitrary units; P �.015 by the Mann–Whitney U test). All tumors that had in-creased �-TrCP1 expression had increased or overexpressed�-TrCP protein (Fig. 2, A). No tumor lost expression of�-TrCP1 mRNA or �-TrCP protein relative to the expression inits normal counterpart.

Next, we examined the types of �-catenin activation in the 45colorectal tumor samples. Nuclear accumulation of �-cateninprotein, indicative of oncogenic activation, was detected in 22(49%) of the 45 samples. Of these 22 tumors, 12 showed thediffuse type of �-catenin activation [i.e., nuclear accumulation incells throughout the tumor (13)] (Fig. 2, B), and 10 showed theinvasion edge type of �-catenin activation [i.e., nuclear accu-mulation in cells only along the invasion edge (13)] (Fig. 2, C).Increased �-TrCP1 mRNA or �-TrCP protein expression in thetumor was associated with �-catenin activation (P � .023)(Table 1). The association was stronger for tumors with theinvasion edge type of �-catenin activation (P � .026) than fortumors with the diffuse type of �-catenin activation (P � .314).The expression level of �-TrCP1 was higher in tumors with�-catenin activation than in tumors without �-catenin activation,although the difference between the groups was not statisticallysignificant (P � .6).

In all cancer cell lines with activated �-catenin (28), over-expression of �-TrCP coincided with increased levels of�-catenin protein, which is indicative of its stabilization (Fig. 3).We have previously shown that in 293T cells, which have anintact Wnt/�-catenin signaling pathway, a negative feedbackmechanism regulates the stability and levels of �-catenin proteinthrough increases in �-TrCP protein levels (25). To determinewhether �-catenin and �-TrCP proteins directly interact witheach other, we performed co-immunoprecipitation experiments.By using an antibody to �-catenin, we could immunoprecipitate�-TrCP from 293T cell lysates but not from colorectal cancercell line lysates (Fig. 3). This observation and the associationbetween increased �-TrCP expression and �-catenin activationin the tumors (shown above) indicate that the interaction be-tween �-catenin and �-TrCP is impaired in colorectal cancer celllines and that this aberration might be responsible for the�-catenin stabilization and activation in colorectal cancer celllines and in clinical cancer cells.

We next evaluated the tumors for localization and expressionof �-TrCP protein by immunohistochemistry. Light cytoplasmic

Table 1. Expression of �-transducin repeat–containing protein 1 (�-TrCP1)mRNA in colorectal cancers with different patterns of �-catenin subcellularlocalization

�-cateninsubcellularlocalization*

No. of tumors with �-TrCP1 expression

Constitutive Increased Mean intensity (95% CI)†

Membranous 14 9 0.735 (0.629 to 0.841)Diffuse 5 7† 0.741 (0.595 to 0.887)Invasion edge 1 9† 0.836 (0.672 to 1.000)

*Immunohistochemistry was used to assess �-catenin subcellular localization.Diffuse � nuclear accumulation in cells throughout the tumor; invasion edge �nuclear accumulation in cells only along the invasion edge adjacent to the stromaand membranous staining in the rest of the tumor. Both diffuse and invasion edgetypes of nuclear accumulation indicate oncogenic �-catenin activation. CI �confidence interval.

†Total RNA was extracted and subjected to quantitative reverse transcription–polymerase chain reaction (RT-PCR). The PCR products were subjected toelectrophoresis and, after staining the gel with SYBRGreen, to densitometry.Band intensity for �-TrCP1 expression in the tumor was normalized against aninternal control gene (glyceraldehyde-3-phosphate dehydrogenase) and com-pared with �-TrCP1 expression in corresponding normal tissue. Intensity isexpressed in arbitrary units. No statistical difference in intensity for �-TrCP1expression was found among the three groups (membranous, diffuse, and inva-sion edge types). P � .023 by the exact test for large tables.




ownloaded from


staining of �-TrCP was found in non-neoplastic crypt cells andcarcinoma cells showing constitutive �-TrCP1 mRNA expres-sion (Fig. 4, A and B, respectively). By contrast, nuclear accu-mulation of �-TrCP protein was observed in 19 of 25 tumors thathad increased mRNA expression by quantitative RT-PCR (Fig.4, C), in agreement with other studies (36,37). Nuclear accumu-lation of �-TrCP was more frequently found in tumors withnuclear accumulation of �-catenin (18 of 22 tumors; P � .035),especially in tumors with the invasion edge type of �-cateninactivation (Fig. 4, C and D), than in tumors without nuclearaccumulation of �-catenin (11 of 23 tumors). Although it isknown that �-TrCP targets IB� for ubiquitination-dependentdegradation (22,23,31), and thus leads to activation of theNF-B pathway (25,38), we found nuclear localization of thep65 subunit of NF-B, which represents its active form (39), in10 of 30 tumors showing nuclear expression of �-TrCP (Fig. 4,C and E). The p65 subunit was also detected in tumor cell lysatesby immunoblotting (data not shown). Thus, nuclear co-localization of �-TrCP protein and the active form of NF-Bwas characteristic of a subset of colorectal carcinomas withoncogenic �-catenin activity and indicated that the �-catenin-dependent increase in �-TrCP expression is associated withNF-B activation in colorectal cancers.

Because an increased level of �-TrCP promotes the activationof NF-B (21), which suppresses apoptosis and promotes cellsurvival (38), we determined whether induction of �-TrCP andactivation of NF-B in colorectal cancers inhibited apoptosis.We calculated the apoptotic index for each tumor and assessedwhether there was any association between the frequency ofapoptosis and increased �-TrCP expression. The apoptotic indexwas statistically significantly lower in tumors with increased�-TrCP expression (mean apoptotic index � 0.704, 95% CI �0.492 to 0.908) than in tumors with constitutive �-TrCP expres-sion (mean apoptotic index � 1.040, 95% CI � 0.730 to 1.350;

mean difference � 0.340, 95% CI � 0.025 to 0.701; P � .035by Student’s t test) (Table 2). Although the apoptotic index wasstatistically significantly lower in tumors with �-catenin activa-tion (mean apoptotic index � 0.664, 95% CI � 0.470 to 0.858)than in tumors without �-catenin activation (mean apoptoticindex � 1.035, 95% CI � 0.739 to 1.331; P � .043), there wasno difference in the apoptotic index between tumors that showedthe diffuse type of �-catenin activation (mean apoptotic index �0.650, 95% CI � 0.322 to 0.978) and those that showed theinvasion edge type of �-catenin activation (mean apoptotic in-dex � 0.680, 95% CI � 0.480 to 0.870; P � .882).

To assess the clinical significance of our results, we examinedrelationships between �-TrCP1 expression and clinical parame-ters and the histopathologic characteristics of patients with colo-rectal cancer (Table 3). Increased �-TrCP1 expression was sta-tistically significantly associated with venous invasion bycarcinoma cells (P � .028), a characteristic that is considered arisk factor for metastasis. The expression level of �-TrCP1 in theprimary tumors of the patients who developed metastases todistant organs (0.960 arbitrary units, 95% CI � 0.878 to 1.042arbitrary units) was statistically significantly higher than in pa-tients who did not develop metastasis (0.722 arbitrary units, 95%CI � 0.600 to 0.844 arbitrary units; P � .016). This observationwas attributed to a higher incidence of increased �-TrCP1 ex-

Fig. 3. Expression of the 54-kd �-transducin repeat–containing protein (�-TrCP)and the 92-kd �-catenin protein in colorectal cancer cell lines (28), and in the humanrenal epithelial cell line 293T, which expresses the simian virus 40 large T antigenand constitutively expresses �-TrCP and �-catenin proteins (25). A 250-�g aliquotof protein extract from each cell line was immunoprecipitated with the antibody to�-catenin (IP: �-catenin). Precipitated material was subjected to electrophoresis andimmunoblotting. Immunoblots were probed sequentially with antibodies to �-TrCP(WB: �-TrCP) to detect any possible interaction between the two molecules and thenwith the antibody to �-catenin (WB: �-catenin).

Fig. 4. Expression of �-transducin repeat–containing protein (�-TrCP),�-catenin, and the p65 subunit of nuclear factor kappaB (NF-B) in normalcolorectal tissues and colorectal carcinomas (A–E). A) Weak �-TrCP expres-sion, detected by immunohistochemistry with an anti-�-TrCP antibody (32) andthe avidin–biotin–peroxidase detection method, was found in the cytoplasm ofnormal colorectal crypt epithelial cells. B) Colon adenocarcinoma showingconstitutive cytoplasmic expression of �-TrCP. Immunostaining serial sectionsfrom the same tumor showed nuclear �-TrCP expression (C), nuclear accumu-lation of �-catenin (D), and nuclear expression of the p65 subunit of NF-B (E).Representative sections are shown.




ownloaded from


pression in primary tumors that had metastasized than in tumorsthat had not. There was no statistically significant difference inthe frequency of increased �-TrCP1 expression between patientswith and without distant metastases (Table 3).

DISCUSSION

Previously, we demonstrated that the malignant potential ofcolorectal cancer and patient outcome are closely associatedwith the distinct types of oncogenic �-catenin activation (13,14).In this study, we sought to clarify whether the pattern of�-catenin activation is affected by expression of its regulator, theE3 ubiquitin ligase receptor �-TrCP. We found that, in colorec-tal cancers, increased levels of �-TrCP were associated withnuclear accumulation of �-catenin, which is indicative of itsoncogenic activation. The association was more prominent intumors showing the invasion edge type of �-catenin activationthan in tumors showing the diffuse type. These results supportour previously published data that demonstrated induction of�-TrCP by the �-catenin/Tcf signal transduction pathway (25).Co-localization of �-TrCP and �-catenin in nuclei of the sametumor cells further corroborate our earlier conclusions.

Phosphorylation of �-catenin protein within the Asp-Ser-Gly-Iso-His-Ser sequence by GSK3�, which is tethered to the com-plex with �-catenin by APC and axin, is essential for �-TrCP-mediated ubiquitination and degradation of �-catenin (17–19).Genetic alterations in axin or CTNNB1 are rarely detected incolorectal cancers, whereas genetic alterations—gene deletion

or mutation—that inactivate APC are frequently detected (2,3).Inactivation of APC abrogates phosphorylation of �-catenin andthereby prevents its interaction with �-TrCP, rendering �-cateninresistant to ubiquitination. Therefore, the strong association be-tween increased �-TrCP expression and oncogenic �-cateninactivation in the same tumors suggests that, unlike 293T cells(25), the negative feedback loop between this E3-ubiquitin li-gase receptor (�-TrCP) and its target oncoprotein (�-catenin) isdefective in colorectal cancer.

The selective and intentional degradation of proteins in theubiquitin–proteasome system is important in the turnover ofregulatory proteins (40–42). The E3 ubiquitin ligase receptorsare responsible for substrate specificity (40–42). Loss of expres-sion or function of certain types of ubiquitin ligase has beenobserved in several human diseases, both inherited and acquired,including neurologic and neoplastic diseases (43–46). Variousaberrations in the ubiquitin–proteasome system–mediated con-trol of signaling by oncogenes’ products (or oncoprotein) andtumor suppressors have been implicated in development of hu-man cancers (43,44). One such mechanism involves uncon-trolled and accelerated removal of a substrate (i.e., loss-of-function), whereas another involves the stabilization of asubstrate resulting from an inactivation of an enzyme in thesystem or from a mutation in a targeting motif in the substrate(i.e., gain-of-function) (43). Mdm2/Hdm2 functions as an E3ubiquitin ligase, and inactivation of p53 resulting from Mdm2/Hdm2-mediated degradation is an example of a loss-of-function–type aberration (47,48). Similarly, low expression levels of thecyclin-dependent kinase inhibitor p27kip1, a tumor suppressorprotein (49), have been associated with progression of a varietyof human cancers (50–55). In particular, a decrease or loss ofp27kip1 expression in colorectal cancer is associated with aggres-sive tumor behavior and unfavorable patient outcome (56,57).This decrease is attributed to ubiquitination of p27kip1 mediatedby its cognate E3 ubiquitin ligase, Skp2 (58–61), which hasrecently been reported to have oncogenic activity (62,63). Bycontrast, somatic mutations either in the APC gene or in the�-catenin phosphoacceptor sites that result in the stabilizationand subsequent oncogenic activation of �-catenin by escapingthe ubiquitin–proteasome system represent one of the best doc-umented gain-of-function–type aberrations (43).

Several proteins that are regulated via the ubiquitin-proteasome system and are involved in colorectal carcinogenesisare affected by changes in �-TrCP expression. Although �-TrCPis not a direct participant in the gain-of-function–type aberrationof the ubiquitin system associated with �-catenin activation,�-TrCP is implicated in colorectal carcinogenesis because thenegative feedback loop regulating �-catenin oncoprotein expres-sion is defective. In addition, �-TrCP targets IB� (31,64,65),the inhibitor of NF-B (66), for ubiquitination. Increased�-TrCP expression is associated with increased IB� ubiquiti-nation, which in turn results in the translocation of NF-B to thenucleus and the transactivation of antiapoptosis genes importantfor cell survival (38,67–69). Consistent with this scenario, wefound an association between increased �-TrCP expression anddecreased apoptosis in the colorectal cancers. Furthermore,�-TrCP also plays a critical role in the inducible processing thatconverts NF-B into an active form (70,71). By immunohisto-chemistry, we detected expression of the p65 subunit of NF-Band found that its nuclear localization was closely associatedwith the presence of �-TrCP and a low incidence of apoptosis in

Table 2. Comparison of apoptosis indexes in colorectal cancers with varyingexpression levels of �-transducin repeat–containing protein (�-TrCP) andpatterns of �-catenin subcellular localization

Molecular variables n*Mean apoptotic index

(95% CI)† P value‡

�-TrCP1 expression§Constitutive 20 1.040 (0.730 to 1.350)Increased 25 0.704 (0.492 to 0.908) .035

�-catenin subcellular localization�Membranous 23 1.035 (0.739 to 1.331)Diffuse 12 0.650 (0.322 to 0.978)¶ .046#Invasion edge 10 0.680 (0.480 to 0.870)¶ .026#

*Number of cancer cases in each category.†Apoptotic index was determined by the terminal deoxynucleotidyl trans-

ferase–mediated deoxyuridine triphosphate–biotin nick-end labeling (TUNEL)method (34). The apoptotic index of each primary tumor was calculated as thenumber of TUNEL-positive cells and apoptotic bodies (35) per 2500 carcinomacells counted by microscopic observation of five randomly selected fields in eachtumor (each field was subjected to two independent counts) and expressed as themean of all samples with 95% confidence intervals (CIs).

‡Determined by Student’s t test.§Constitutive expression was defined as no difference in the level of expres-

sion between the matched tumor and normal colorectal tissue. Increased expres-sion was defined as an increase in the level of expression in the tumor of at least50% over that of the corresponding normal colorectal tissue.

�Tumors that expressed �-catenin protein only on tumor cell membranes weredetermined to have a membranous type of localization. Tumors that showednuclear accumulation of �-catenin throughout were determined to have a diffusetype of localization indicative of �-catenin activation. Tumors that showednuclear accumulation of �-catenin only in the cells along the invasion edge of thetumor were determined to have an invasion edge type of �-catenin activation.

¶There was no statistical difference between the groups of diffuse and inva-sion edge types.

#P value was obtained by comparison with apoptosis index in the tumors withmembranous �-catenin localization (i.e., no �-catenin activation).




ownloaded from


tumor cells. Thus, our results suggest that stabilizing NF-B andprocessing it for activation is one of the important molecularmechanisms that may explain how increased �-TrCP functionsagainst apoptosis of tumor cells in colorectal cancer.

Inhibition of apoptosis in tumors expressing high levels of�-TrCP may contribute to their propensity to metastasize. In-deed, we noted that patients with metastases expressed higherlevels of �-TrCP than patients without metastases (Table 3).This observation is important because it demonstrates that one ofthe key molecules in the ubiquitin system that regulates theoncogenic activity of �-catenin is associated with biologic be-havior or malignant behavior and/or the potential of colorectalcancer. Such relevance to human cancer has not been previouslydemonstrated, in part because �-TrCP1 gene mutations or alter-ations in protein expression were rarely found in stomach cancer

(72), prostate cancer (73), or melanoma (74). Investigating andunderstanding the clinical implications of the pathologic alter-ations in the ubiquitin–proteasome system may lead to thedevelopment of molecular cancer therapies that target �-catenin-and/or NF-B-mediated oncogenic signaling pathways by mod-ulating the E3 family of ubiquitin ligases (75,76).

In summary, increased expression of �-TrCP participates incolorectal cancer development and progression by coupling andintegrating oncogenic �-catenin- and NF-B-dependent cell sur-vival signaling. To strengthen the clinical associations we found,multivariable analyses from larger prospective studies will berequired. Greater understanding of the involvement of �-cateninin human carcinogenesis may lead to the development of mo-lecular therapies targeting this oncogene (and its product onco-protein), its regulators, and/or its effectors.

Table 3. Expression of �-transducin repeat–containing protein 1 (�-TrCP1) mRNA in primary tumors of colorectal cancer patients

Variables

�-TrCP1 expression*

P value†

Mean arbitrary unitsof �-TrCP1 level

(95% CI)‡ P value§Constitutive Increased

Age, y�60 5 5 NS 0.704 (0.540 to 0.868) NS�60 15 20 0.777 (0.691 to 0.863)

SexMale 7 16 NS 0.771 (0.669 to 0.873) NSFemale 13 9 0.746 (0.632 to 0.860)

Tumor site�Right-side colon 12 8 NS 0.690 (0.588 to 0.792) NSLeft-side colon 4 7 0.763 (0.650 to 0.875)Rectum 4 10 0.853 (0.685 to 1.021)

Gross type¶1 3 2 NS 0.742 (0.456 to 1.028) NS2 17 22 0.756 (0.676 to 0.836)3 0 1 0.956

Histology¶WD 9 10 0.699 (0.573 to 0.825)MD 8 15 .045 0.823 (0.727 to 0.919) .034PD, Muc 3 0 0.641 (0.355 to 0.927)

Lymphatic invasionAbsent 13 13 NS 0.734 (0.628 to 0.840) NSPresent 7 12 0.793 (0.687 to 0.899)

Venous invasionAbsent 13 8 .028 0.668 (0.566 to 0.770) .020Present 7 17 0.838 (0.736 to 0.940)

Tumor stage#I 0 1 NS 0.879 NSII 13 13 0.729 (0.627 to 0.831)III 7 11 0.699 (0.677 to 0.913)

Distant metastasesAbsent 18 19 NS 0.722 (0.600 to 0.844) .016Present 2 6 0.960 (0.878 to 1.042)

*Constitutive expression was defined as no difference in the level of expression between the matched tumor and normal colorectal tissues. Increased expressionwas defined as an increase in the level of expression in tumors of at least 50% over that of the corresponding normal colorectal tissues. NS � not statisticallysignificant.

†Determined by Fisher’s exact test for 2 � 2 tables, and the exact test for large tables for 2 � 3 tables.‡�-TrCP1 mRNA levels were determined by quantitative reverse transcription–polymerase chain reaction. Band intensity for �-TrCP expression was normalized

against an internal gene (glyceraldehyde-3-phosphate dehydrogenase) and compared with expression in normal tissue. Intensity is expressed in arbitrary units as themean value of all tumors in each category with 95% confidence intervals (CIs).

§Analyzed by using the Mann–Whitney U test for comparison between two unpaired groups (age, sex, lymphatic invasion, venous invasion, and distant metastasis)and the Kruskal–Wallis test for comparisons among three unpaired groups (tumor site, gross type, histology, and tumor stage).

�Right-side colon � cecum, ascending and transverse segments; left-side colon � descending and sigmoid segments.¶Gross and histologic types of the primary tumors are described according to the Japanese Classification of Colorectal Carcinoma (26). Type 1 � protruding tumor;

type 2 � localized and ulcerating tumor; type 3 � diffusely infiltrating and ulcerating tumor; WD � well differentiated adenocarcinoma; MD � moderatelydifferentiated adenocarcinoma; PD � poorly differentiated adenocarcinoma; Muc � mucinous adenocarcinoma.

#Tumor stage at surgery was determined by routine clinical examination and pathology diagnosis of the surgical specimens according to the Tumor–Node–Metastasis (TNM) classification system (27).




ownloaded from


REFERENCES

(1) Jemal A, Thomas A, Murray T, Thun M. Cancer statistics, 2002. CACancer J Clin 2002;52:23–47.

(2) Chung DC. The genetic basis of colorectal cancer: insights into criticalpathways of tumorigenesis. Gastroenterology 2000;119:854–65.

(3) Jass JR, Whitehall VL, Young J, Leggett BA. Emerging concepts incolorectal neoplasia. Gastroenterology 2002;123:862–76.

(4) Minamoto T, Mai M, Ronai Z. K-ras mutation—early detection in molec-ular diagnosis and risk assessment of colorectal, pancreas and lung cancers.Cancer Detect Prev 2000;24:1–12.

(5) Adjei AA. Blocking oncogenic ras signaling for cancer therapy. J NatlCancer Inst 2001;93:1062–74.

(6) Polakis P. The oncogenic activation of beta-catenin. Curr Opin Genet Dev1999;9:15–21.

(7) Taipale J, Beachy PA. The Hedgehog and Wnt signalling pathways incancer. Nature 2001;411:349–54.

(8) Wong NA, Pignatelli M. Beta-catenin—a linchpin in colorectal carcino-genesis. Am J Pathol 2002;160:389–401.

(9) Minamoto T, Ougolkov A, Mai M. Detection of oncogenes in the diagnosisof cancers with active oncogenic signaling. Expert Rev Mol Diagn 2002;2:565–75.

(10) Kikuchi A. Tumor formation by genetic mutations in the components of theWnt signaling pathway. Cancer Sci 2003;94:225–9.

(11) Hugh TJ, Dillon SA, O’Dowd G, Getty B, Pignatelli M, Poston GJ, et al.Beta-catenin expression in primary and metastatic colorectal carcinoma. IntJ Cancer 1999;82:504–11.

(12) Brabletz T, Jung A, Reu S, Porzner M, Hlubek F, Kunz-Schughart LA, etal. Variable beta-catenin expression in colorectal cancers indicates tumorprogression driven by the tumor environment. Proc Natl Acad Sci U S A2001;98:10356–61.

(13) Ougolkov AV, Yamashita K, Mai M, Minamoto T. Oncogenic beta-cateninand MMP-7 (matrilysin) cosegregate in late-stage clinical colon cancer.Gastroenterology 2002;122:60–71.

(14) Zhang B, Ougolkov A, Yamashita K, Takahashi Y, Mai M, Minamoto T.Beta-catenin and ras oncogenes detect most human colorectal cancer. ClinCancer Res 2003;9:3073–9.

(15) Miyaki M, Iijima T, Kimura J, Yasuno M, Mori T, Hayashi Y, et al.Frequent mutation of beta-catenin and APC genes in primary colorectaltumors from patients with hereditary nonpolyposis colorectal cancer. Can-cer Res 1999;59:4506–9.

(16) Mirabelli-Primdahl L, Gryfe R, Kim H, Miller A, Luceri C, Dale D, et al.Beta-catenin mutations are specific for colorectal carcinomas with micro-satellite instability but occur in endometrial carcinomas irrespective ofmutator pathway. Cancer Res 1999;59:3346–51.

(17) Aberle H, Bauer A, Stappert J, Kispert A, Kemler R. Beta-catenin is atarget for the ubiquitin-proteasome pathway. EMBO J 1997;16:3797–804.

(18) Hart M, Concordet JP, Lassot I, Albert I, del los Santos R, Durand H, et al.The F-box protein beta-TrCP associates with phosphorylated beta-cateninand regulates its activity in the cell. Curr Biol 1999;9:207–10.

(19) Winston JT, Strack P, Beer-Romero P, Chu CY, Elledge SJ, Harper JW.The SCFbeta-TRCP-ubiquitin ligase complex associates specifically withphosphorylated destruction motifs in IkappaBalpha and beta-catenin andstimulates IkappaBalpha ubiquitination in vitro. Genes Dev 1999;13:270–83.

(20) Liu C, Kato Y, Zhang Z, Do VM, Yankner BA, He X. Beta-Trcp couplesbeta-catenin phosphorylation-degradation and regulates Xenopus axis for-mation. Proc Natl Acad Sci U S A 1999;96:6273–8.

(21) Fuchs SY, Spiegelman VS, Suresh Kumar KG. The many faces of beta-TrCP E3 ubiquitin ligase receptors: reflections in the magic mirror ofcancer. Oncogene 2004;23:2028–36.

(22) Maniatis T. A ubiquitin ligase complex essential for the NF-kappaB,Wnt/Wingless, and Hedgehog signaling pathways. Genes Dev 1999;13:505–10.

(23) Peters JM. Emi1 proteolysis: how SCFbeta-Trcp1 helps to activate theanaphase-promoting complex. Mol Cell 2003;11:1420–1.

(24) Chiaur DS, Murthy S, Cenciarelli C, Parks W, Loda M, Inghirami G, et al.Five human genes encoding F-box proteins: chromosome mapping andanalysis in human tumors. Cytogenet Cell Genet 2000;88:255–8.

(25) Spiegelman VS, Slaga TJ, Pagano M, Minamoto T, Ronai Z, Fuchs SY.Wnt/beta-catenin signaling induces the expression and activity of betaTrCPubiquitin ligase receptor. Mol Cell 2000;5:877–82.

(26) Japanese Society for Cancer of the Colon and Rectum. Japanese classifi-cation of colorectal carcinoma. First English ed. Tokyo (Japan): Kanehara& Co.; 1997. p. 27–66.

(27) TNM classification of malignant tumours. In: Sobin LK, Wittekind C,editors. UICC International Union Against Cancer. 5th ed. New York(NY): John Wiley & Sons; 1997. p. 66–9.

(28) Gayet J, Zhou XP, Duval A, Rolland S, Hoang JM, Cottu P, et al. Extensivecharacterization of genetic alterations in a series of human colorectal cancercell lines. Oncogene 2001;20:5025–32.

(29) Chomczynski P, Sacchi N. Single-step method of RNA isolation by acidguanidinium thiocyanate-phenol-chloroform extraction. Anal Biochem1987;162:156–9.

(30) Puissant C, Houdebine LM. An improvement of the single-step method ofRNA isolation by acid guanidinium thiocyanate-phenol-chloroform extrac-tion. Biotechniques 1990;8:148–9.

(31) Spiegelman VS, Stavropoulos P, Latres E, Pagano M, Ronai Z, Slaga TJ,et al. Induction of beta-transducin repeat-containing protein by JNK sig-naling and its role in the activation of NF-kappaB. J Biol Chem 2001;276:27152–8.

(32) Spiegelman VS, Tang W, Katoh M, Slaga TJ, Fuchs SY. Inhibition of HOSexpression and activities by Wnt pathway. Oncogene 2002;21:856–60.

(33) Harlow E, Lane D. Using antibodies: a laboratory manual. Plainview (NY):Cold Spring Harbor Laboratory Press; 1999.

(34) Tornusciolo DR, Schmidt RE, Roth KA. Simultaneous detection of TDT-mediated dUTP-biotin nick end-labeling (TUNEL)-positive cells and mul-tiple immunohistochemical markers in single tissue sections. Biotechniques1995;19:800–5.

(35) Thompson CB. Apoptosis in the pathogenesis and treatment of disease.Science 1995;267:1456–62.

(36) Lassot I, Segeral E, Berlioz-Torrent C, Durand H, Groussin L, Hai T, et al.ATF4 deregulation relies on a phosphorylation-dependent interaction withthe SCFbetaTrCP ubiquitin ligase. Mol Cell Biol 2001;21:2192–202.

(37) Davis M, Hatzubai A, Andersen JS, Ben-Shushan E, Fisher GZ, Yaron A,et al. Pseudosubstrate regulation of the SCFbeta-TrCP ubiquitin ligase byhnRNP-U. Genes Dev 2002;16:439–51.

(38) Karin M, Cao Y, Greten FR, Li ZW. NF-kappaB in cancer: from innocentbystander to major culprit. Nat Rev Cancer 2002;2:301–10.

(39) Verma IM, Stevenson J. IkappaB kinase: beginning, not the end. Proc NatlAcad Sci U S A 1997;94:11758–60.

(40) Hershko A, Ciechanover A. The ubiquitin system. Annu Rev Biochem1998;67:425–79.

(41) Hershko A, Ciechanover A, Varshavsky A. The ubiquitin system. Nat Med2000;6:1073–81.

(42) Ciechanover A, Orian A, Schwartz AL. Ubiquitin-mediated proteolysis:biological regulation via destruction. Bioessays 2000;22:442–51.

(43) Ciechanover A, Schwartz AL. Ubiquitin-mediated degradation of cellularproteins in health and disease. Hepatology 2002;35:3–6.

(44) Fuchs SY. The role of ubiquitin-proteasome pathway in oncogenic signal-ing. Cancer Biol Ther 2002;1:337–41.

(45) Shimura H, Hattori N, Kubo S, Mizuno Y, Asakawa S, Minoshima S, et al.Familial Parkinson disease gene product, perkin, is a ubiquitin-proteinligase. Nat Genet 2000;25:302–5.

(46) Jiang YH, Armstrong D, Albrecht U, Atkins CM, Noebels JL, Eichele G,et al. Mutation of the Angelman ubiquitin ligase in mice causes increasedcytoplasmic p53 and deficits of contextual learning and long-term poten-tiation. Neuron 1998;21:799–811.

(47) Daujat S, Neel H, Piette J. MDM2: life without p53. Trends Genet 2001;17:459–64.

(48) Alarcon-Vargas D, Ronai Z. p53-MDM2–the affair that never ends. Car-cinogenesis 2002;23:541–7.

(49) Reynisdottir I, Polyak K, Iavarone A, Massague J. Kip/Cip and Ink4 Cdkinhibitors cooperate to induce cell cycle arrest in response to TGF-beta.Genes Dev 1995;9:1831–45.

(50) Porter PL, Malone KE, Heagerty PJ, Alexander GM, Gatti LA, Firpo EJ, etal. Expression of cell-cycle regulators p27Kip1 and cyclin E, alone and incombination, correlate with survival in young breast cancer patients. NatMed 1997;3:222–5.




ownloaded from


(51) Catzavelos C, Bhattacharya N, Ung YC, Wilson JA, Roncari L, Sandhu C,et al. Decreased levels of the cell-cycle inhibitor 27Kip1 protein: prognos-tic implications in primary breast cancer. Nat Med 1997;3:227–30.

(52) Mori M, Mimori K, Shiraishi T, Tanaka S, Ueno H, Sugimachi K, et al. p27expression and gastric carcinoma. Nat Med 1997;3:593.

(53) Guo Y, Sklar GN, Borkowski A, Kyprianou N. Loss of the cyclin-dependent kinase inhibitor p27Kip1 protein in human prostate cancercorrelates with tumor grade. Clin Cancer Res 1997;3:2269–74.

(54) Masciullo V, Sgambato A, Pacilio C, Pucci B, Ferrandina G, Palazzo J, etal. Frequent loss of expression of the cyclin-dependent kinase inhibitor p27in epithelial ovarian cancer. Cancer Res 1999;59:3790–4.

(55) Bamberger AM, Riethdorf L, Milde-Langosch K, Bamberger CM, ThunekeI, Erdmann I, et al. Strongly reduced expression of the cell cycle inhibitorp27 in endometrial neoplasia. Virchows Arch 1999;434:423–8.

(56) Loda M, Cukor B, Tam SW, Lavin P, Fiorentino M, Draetta GF, et al.Increased proteasome-dependent degradation of the cyclin-dependent ki-nase inhibitor p27 in aggressive colorectal carcinomas. Nat Med 1997;3:231–4.

(57) Hirano K, Minamoto T. Altered expression of p53 and p27 proteins, aloneor combined, as a predictor of metastatic potential in early invasive carci-nomas of colon and rectum—a comparative clinicopathologic and molec-ular analysis. Cancer Detect Prev 2000;24:343–55.

(58) Carrano AC, Eytan E, Hershko A, Pagano M. SKP2 is required forubiquitin-mediated degradation of the CDK inhibitor p27. Nat Cell Biol1999;1:193–9.

(59) Sutterluty H, Chatelain E, Marti A, Wirbelauer C, Senften M, Muller U, etal. p45SKP2 promotes p27Kip1 degradation and induces S phase in qui-escent cells. Nat Cell Biol 1999;1:207–14.

(60) Tsvetkov LM, Yeh KH, Lee SJ, Sun H, Zhang H. p27Kip1 ubiquitinationand degradation is regulated by the SCFSkp2 complex through phosphor-ylated Thr187 in p27. Curr Biol 1999;9:661–4.

(61) Hershko D, Bornstein G, Ben-Izhak O, Carrano A, Pagano M, KrauszMM, et al. Inverse relation between levels of p27Kip1 and of itsubiquitin ligase subunit Skp2 in colorectal carcinomas. Cancer 2001;91:1745–51.

(62) Shim EH, Johnson L, Noh HL, Kim YJ, Sun H, Zeiss C, et al. Expressionof the F-box protein SKP2 induces hyperplasia, dysplasia and low-gradecarcinoma in the mouse prostate. Cancer Res 2003;63:1583–8.

(63) Shigemasa K, Gu L, O’Brien TJ, Ohama K. Skp2 overexpression is aprognostic factor in patients with ovarian adenocarcinoma. Clin Cancer Res2003;9:1756–63.

(64) Kroll M, Margottin F, Kohl A, Renard P, Durand H, Concordet JP, et al.Inducible degradation of IkappaBalpha by the proteasome requiresinteraction with the F-box protein h-betaTrCP. J Biol Chem 1999;274:7941–5.

(65) Nakayama K, Hatakeyama S, Maruyama S, Kikuchi A, Onoe K, Good RA,et al. Impaired degradation of inhibitory subunit of NF-kappaB (IkappaB)and beta-catenin as a result of targeted disruption of the beta-TrCP1 gene.Proc Natl Acad Sci U S A 2003;100:8752–7.

(66) Yaron A, Gonen H, Alkalay I, Hatzubai A, Jung S, Beyth S, et al. Inhibitionof NF-kappaB cellular function via specific targeting of the IkappaB-ubiquitin ligase. EMBO J 1997;16:6486–94.

(67) Schmid RM, Adler G. NF-kappaB/Rel/IkappaB: implications in gastroin-testinal diseases. Gastroenterology 2000;118:1208–28.

(68) Karin M, Lin A. NF-kappaB at the crossroads of life and death. NatImmunol 2002;3:221–7.

(69) Pahl HL. Activators and target genes of Rel/NF-kappaB transcriptionfactors. Oncogene 1999;18:6853–66.

(70) Fong A, Sun SC. Genetic evidence for the essential role of beta-transducinrepeat-containing protein in the inducible processing of NF-kappaB2/p100.J Biol Chem 2002;277:22111–4.

(71) Lang V, Janzen J, Fischer GZ, Soneji Y, Beinke S, Salmeron A, et al.betaTrCP-mediated proteolysis of NF-kappaB1 p105 requires phosphory-lation of p105 serines 927 and 932. Mol Cell Biol 2003;23:402–13.

(72) Saitoh T, Katoh M. Expression profiles of betaTRCP1 and betaTRCP2, andmutation analysis of betaTRCP2 in gastric cancer. Int J Oncol 2001;18:959–64.

(73) Gerstein AV, Almeida TA, Zhao G, Chess E, Shih IeM, Buhler K, et al.APC/CTNNB1 (beta-catenin) pathway alterations in human prostate can-cers. Genes Chromosomes Cancer 2002;34:9–16.

(74) Reifenberger J, Knobbe CB, Wolter M, Blaschke B, Schulte KW, PietschT, et al. Molecular genetic analysis of malignant melanomas for aberrationsof the Wnt signaling pathway genes CTNNB1, APC, ICAT and BTRC. IntJ Cancer 2002;100:549–56.

(75) Soldatenkov VA, Dritschilo A, Ronai Z, Fuchs SY. Inhibition of homo-logue of Slimb (HOS) function sensitizes human melanoma cells forapoptosis. Cancer Res 1999;59:5085–8.

(76) Zhang YW, Nakayama K, Nakayama KI, Morita I. A novel route forconnexin 43 to inhibit cell proliferation: negative regulation of S-phasekinase-associated protein (Skp2). Cancer Res 2003;63:1623–30.

NOTES

Supported by Grants-in-Aid for Scientific Research from the Japanese Min-istry of Education, Science, Sports, Technology and Culture (to T. Minamotoand M. Mai) and from the Japan Society for the Promotion of Science (to T.Minamoto, A. Ougolkov, and B. Zhang). Additional support was provided bygrants from the Naito Foundation, Tokyo, Japan, and Kanazawa Research Centerfor Advanced Technology, Kanazawa, Japan (to T. Minamoto) along with PublicHealth Service grant CA 92900 (to S. Y. Fuchs) from the National CancerInstitute, National Institutes of Health, Department of Health and HumanServices.

We thank Dr. Naofumi Mukaida (Division of Molecular Bioregulation, Ka-nazawa University Cancer Research Institute) for assistance in preparation of the�-TrCP1 plasmid, Atsuko Kaneda-Shimizu for technical assistance, and MichaelMeyer for editorial assistance.

Manuscript received October 29, 2003; revised June 3, 2004; accepted June15, 2004.




ownloaded from


Documents

Associations Among -TrCP, an E3 Ubiquitin Ligase Receptor, -Catenin, and NF-B in Colorectal Cancer