erbb receptor signaling promotes ependymoma cell...

Advances in Brief

ERBB Receptor Signaling Promotes Ependymoma Cell Proliferationand Represents a Potential Novel Therapeutic Targetfor This Disease1

Richard J. Gilbertson,2 Lyndsay Bentley,Roberto Hernan, Teemu T. Junttila,Adrian J. Frank, Hannu Haapasalo,Michele Connelly, Cynthia Wetmore,Tom Curran, Klaus Elenius, andDavid W. EllisonDepartment of Developmental Neurobiology, St. Jude Children’sResearch Hospital, Memphis, Tennessee 38105 [R. J. G., L. B., R. H.,A. J. F., M. C., C. W., T. C.]; MediCity Research Laboratories andDepartment of Medical Biochemistry and Molecular Biology [T. T. J.,K. E.] and Turku Graduate School of Biomedical Sciences [T. T. J.],University of Turku, FIN-20520 Turku, Finland; Department ofPathology, Tampere University Hospital, FIN-33521 Tampere,Finland [H. H.]; Northern Institute for Cancer Research, The MedicalSchool, University of Newcastle upon Tyne, Newcastle upon TyneNE2 4HH, United Kingdom [D. W. E.]; and Department ofNeuropathology, Newcastle General Hospital, Newcastle Upon TyneNE4 6BE, United Kingdom [D. W. E.]

AbstractPurpose: This study was designed to investigate the

biological and therapeutic significance of ERBB1, ERBB2,ERBB3, and ERBB4 in childhood ependymoma.

Experimental Design: The expression frequency andclinical significance of ERBB1–4 was analyzed in a largecohort of pediatric ependymoma (n � 121) using immuno-histochemistry, Western blotting, and reverse transcription-PCR analysis. Histological markers of anaplasia (necrosis,microvascular proliferation, and Ki-67 proliferative index)were also determined. Functional assessment of ERBB-dependent cell signaling and proliferation, in addition to noveltherapeutic inhibition of these processes, was conducted usingshort-term cultures of human ependymoma cells.

Results: Coexpression of ERBB2 and ERBB4 was iden-tified in over 75% of tumors. High-level coexpression ofthese receptors was significantly related to tumor prolifer-ative activity [P < 0.05; Ki-67 labeling index (LI)] and, in

combined survival analysis of clinical (degree of surgicalresection) and molecular (ERBB2/ERBB4 expression statusand Ki-67 LI) factors, enabled a greater resolution of patientprognosis than any individual variable alone. Ligand-depen-dent activation of ERBB receptor-signaling in culturedependymoma cells resulted in AKT phosphorylation andcellular proliferation that was significantly blocked in adose-dependent manner using WAY-177820, a novel inhib-itor of ERBB2 tyrosine kinase activity.

Conclusions: This study suggests that ERBB receptorsignaling results in aggressive disease behavior in ependy-moma by promoting tumor cell proliferation. An analysis ofERBB2 and ERBB4 expression, in association with Ki-67 LIand the degree of surgical resection, may provide an accu-rate tool for assessing disease risk in children with thisdisease. In addition, these receptors may serve as a target fornovel therapeutic approaches in ependymoma.

IntroductionLess than one-half of all children receiving contemporary

treatment for ependymoma, the third most common intracranialtumor of childhood, will be disease free at 5 years (1–4).Clinical outcome is significantly better for patients whose tu-mors are totally resected (1–6). However, attempts to defineoptimum adjuvant treatment have been hindered by the rela-tively low incidence of the disease and the subsequent paucity ofdefinitive clinical trials (1). A lack of consensus regarding theprognostic value of histological grading and poor understandingof disease biology has compounded this situation further (1, 5,6). Substantial improvements in clinical management are un-likely to be achieved in the absence of reliable morphologicaland molecular predictors of disease behavior.

The WHO classification of nervous system tumors recog-nizes four major subgroups of ependymal tumor: subependy-moma (WHO grade I), myxopapillary ependymoma (WHOgrade I), ependymoma (WHO grade II), and anaplastic ependy-moma (WHO grade III). The last of these is generally believedto demonstrate more aggressive clinical behavior. However, thediagnosis of anaplasia has proved subjective, and its prognosticsignificance remains controversial (1–7). The identification ofmolecular genetic abnormalities responsible for promoting ag-gressive disease behavior would provide a more accurate meansof predicting disease risk in ependymoma. Such abnormalitiesmay also prove useful as novel therapeutic targets. However,although non-random chromosomal alterations have been de-scribed in ependymoma (8–13), their clinical and molecularsignificance remains unclear.

In this study, we investigated the expression and clinicalsignificance of the RTK I3 family in childhood ependymoma.

Received 3/11/02; revised 6/5/02; accepted 6/18/02.The costs of publication of this article were defrayed in part by thepayment of page charges. This article must therefore be hereby markedadvertisement in accordance with 18 U.S.C. Section 1734 solely toindicate this fact.1 Supported in part by the United Kingdom Medical Research Council,Wessex Cancer Trust, Cancer Center (CORE) Support Grant CA 21765,American Lebanese Syrian Associated Charities (ALSAC), Academy ofFinland, Sigrid Juselius Foundation, Turku University Central Hospital,University Foundation of Turku, and the Aurator Biomed Fund.2 To whom requests for reprints should be addressed, at Department ofDevelopmental Neurobiology, St. Jude Children’s Research Hospital,332 North Lauderdale Street, Memphis, TN 38105. Phone: (901) 495-3913; Fax: (901) 495-2270; E-mail: [email protected].

3054 Vol. 8, 3054–3064, October 2002 Clinical Cancer Research

Research. on September 9, 2019. © 2002 American Association for Cancerclincancerres.aacrjournals.org Downloaded from

http://clincancerres.aacrjournals.org/

The RTK I family includes the EGFR (also known as ERBB1),and the ERBB2, ERBB3, and ERBB4 receptors (14, 15). Af-finity-based, hierarchical interaction between these receptorsand their numerous ligands results in the formation of a networkof transmembrane homo- and heterodimers. Signaling via thisnetwork has been implicated in the development of both normaland malignant tissues within the nervous system (16–19), in-cluding the pediatric brain tumor medulloblastoma (20–23).Here, we show that the ERBB2 and ERBB4 receptors arefrequently coexpressed in childhood ependymoma and are as-sociated with increased tumor-proliferative capacity and a poorclinical outcome. In addition, we demonstrate that activation ofthe ERBB signaling network promotes ependymoma cell pro-liferation in vitro and that this can be blocked using a novelinhibitor of ERBB2 tyrosine kinase activity.

Patients and MethodsPatients and Tumor Material. Two populations of chil-

dren (�18 years) with ependymoma were used in the study.Group A included 99 newly diagnosed consecutive patients withposterior fossa ependymoma (Table 1). Patients were treated insix (Bristol, n � 17; Edinburgh, n � 27; London, n � 20;Newcastle, n � 16; Oxford, n � 8; Southampton, n � 11)United Kingdom Children’s Cancer Study Group centers, be-tween 1972 and 1998. A total of 89 cases were diagnosed after1980. Formalin-fixed paraffin-embedded material obtained atprimary surgery was available from each case. In addition, fiveof these cases had samples of relapsed tumor available that wasobtained during surgical resection of progressive disease. Pro-tocols for the fixation and histological processing of tumorsamples were identical between study centers. The diagnosis ofependymoma was reviewed for each case by a single neuropa-thologist (D. W. E.). Subependymomas and other variants ofclassic ependymoma were excluded from the study.

The presence of tumor anaplasia was also assessed duringthis review. In an attempt to make this more objective, threediagnostic features of anaplasia were scored independently.These included the presence or absence of necrosis and thepresence or absence of MVP. The third feature, the degree oftumor cell proliferation, was assessed by Ki-67 LI as describedin “Immunohistochemistry.” The extent of primary surgicalresection was determined using the surgeon’s operative notesand, when available, postoperative computed tomography andmagnetic resonance brain imaging. Survival for all of the pa-tients was recorded from the date of diagnosis to the time ofdeath or last follow-up. Median survival for the whole popula-tion was 52 months (range, 2–312 months).

Group B included 15 posterior fossa, 4 supratentorial, and1 spinal tumor(s), from St. Jude Children’s Research Hospital,Memphis, TN, and 2 posterior fossa cases from Tampere Uni-versity Hospital, Tampere, Finland. Fresh frozen tumor material

obtained at primary surgery was available from each case. All ofthe patients were diagnosed and treated between 1985 and 1998(Table 1). Diagnosis was confirmed by Drs. Ashley Hill andJesse Jenkins, Department of Pathology, St. Jude Children’s

3 The abbreviations used are: RTK I, receptor tyrosine kinase type 1; LI,labeling index; MVP, microvascular proliferation; EGF, epidermalgrowth factor; EGFR, EGF receptor; IHC, immunohistochemical; FBS,fetal bovine serum; NRG, neuregulin; GFAP, glial fibrillary acidicprotein; RT-PCR, reverse transcription-PCR.

Table 1 Clinicopathological variables of patients from study groupsA and Ba

Variable

Group A Group B

No. ofcases

% ofpopulation

No. ofcases

% ofpopulation

Age�3 yr 58 58.6 11 50.03–10 yr 29 29.3 5 22.7�10 yr 12 12.1 6 27.3

SexMale 60 60.6 12 54.5Female 39 39.4 10 45.5

SurgeryTotal resection 40 40.4 12 54.5Partial resection 55 55.6 5 22.7N/Kb 4 4.0 5 22.7

Adjuvant therapyRadio 57 57.6 6 27.3Chemo 11 11.1 10 45.4Radio � Chemo 16 16.2 6 27.3None 11 11.1 0 0.0N/K 4 4.0 0 0.0

Primary sitePosterior fossa 99 100 17 77.3Cerebral 0 0.0 4 18.2Spinal 0 0.0 1 4.5

HistologyMVP 46 46.4Necrosis 64 64.6Grade II 15 68.2Grade III 7 31.8

Ki-67 (LI)�25% 63 63.6�25% 30 30.3N/D 6 6.1

ErbB protein expressionc

ErbB1 21/60 35.0 0/20 0.0ErbB2 46/59 78.0 19/20 95.0ErbB3 55/59 93.2 1/20 5.0ErbB4 48/48 100 14/20 70.0

ErbB mRNA expressionErbB1 16/16 21.4 (0.3–190)d

ErbB2 16/16 30.1 (2.3–271)ErbB3 12/16 0.06 (0–0.23)ErbB4 JM-a 16/16 12.0 (0.1–88)ErbB4 JM-b 10/16 0.27 (0–4.19)ErbB4 CYT1 16/16 4.1 (0.05–18)ErbB4 CYT2 16/16 2.2 (0.06–6.2)

StatusDOD 52 52.5 5 22.7AWD 1 4.6ADF 47 47.5 16 72.7

a Clinicopathological variables for patients within the two popula-tions employed in the study.

b N/K, not known; N/D, not done; DOD, dead of disease; AWD,alive with disease; ADF, alive disease-free.

c ErbB receptor protein expression for groups A and B refers to thenumber of cases demonstrating immunoreactivity on immunohisto-chemistry and Western blotting, respectively.

d The second column of values for ErbB mRNA expression bytumors derived from Group B denotes the mean transcript expressionvalue with the expression range shown in parentheses.

3055Clinical Cancer Research



Research Hospital, and one of us (H. H.). In addition to fresh-frozen tumor material, total RNA extracted from normal fetalbrain (14 weeks) and normal frontal cortex (75 years) wasavailable for analysis.

Ependymoma Cell Explants. We analyzed cells ex-planted from two ependymomas treated at St. Jude Children’sResearch Hospital. SJEp1 cells were derived from a recurrentposterior fossa ependymoma demonstrating classic histologyexcised from a 9-year-old female. SJEp2 cells were explantedfrom a primary ependymoma demonstrating anaplastic focitaken from a 2-year-old female. Tumor tissue was minced,plated on BD BioCoat Poly-Lysine Cellware (BD Biosciences,Bedford, MA) in DMEM supplemented with 10% FBS, 1 mM

sodium pyruvate, and 4 mM glutamine (BioWhittaker, Walkers-ville, MD) and cultured in a 37°C-humidified incubator gassedwith 10% CO2 in air. Once confluent, cells were subcultured1:4. Cells were evaluated by immunofluorescence for expres-sion of GFAP (anti-GFAP; DAKO) and vimentin (anti-vimen-tin; Biomeda, Foster City, CA).

Immunohistochemistry. Expression of Ki-67 and thefour ERBB receptors was determined by IHC in 5-�m sectionsof formalin-fixed paraffin-embedded tumor material derivedfrom patients in Group A as described previously (20, 24).Primary antibodies included, anti-Ki67 antibody (DAKO,Carpinteria, CA), and NCL-EGFR, NCL-CB11 (against ERBB1and ERBB2, respectively; Novocastra, Newcastle, United King-dom) and c-17/ERBB3 and c-18/ERBB4 (against ERBB3 andERBB4, respectively; Santa Cruz Biotechnology, Santa Cruz,CA). Sections used in IHC were adjacent to those used forhistological review. Ki-67 LI was calculated using a methoddescribed elsewhere (24). ERBB receptor expression was scoredas described previously using a scale from 0 to 100% (10%increments) based on the estimated percentage of immunoposi-tive tumor cells (20).

Extraction of Total Protein, RNA, and DNA. Totalprotein, RNA, and high-molecular-weight DNA was extractedfrom Group B samples using the TriStar reagent as described bythe manufacturer (Hybaid, Franklin, MA). Total protein wasextracted from cultured ependymoma cells by washing cells inice-cold PBS and scraping into ice-cold lysis buffer [50 mM

HEPES (pH 7.5), 100 mM NaCl, 1 mM EGTA, 1% Triton X-100,20 mM NaF, 1 mM sodium orthovanadate, 1 mM p-nitrophenylphosphate, and 10 �g/ml each: trypsin inhibitor, aprotinin, leu-peptin, and pepstatin (Sigma Chemicals, St. Louis, MO)].

Western Blotting. Total protein lysate (50 �g/sample)was analyzed by Western blotting using standard techniques(20). Blots were probed with the same primary antibodies usedin ERBB receptor IHC analysis. Protein extracted from theA431 (ERBB1-expressing), SKBR3 (ERBB2- and ERBB3-expressing), and MHH-MED-1 (ERBB4-expressing; Ref. 22)cell lines were used as positive protein-expressing controls.Blots were reprobed with a monoclonal antibody for �-actin(Sigma Chemicals) to control for protein loading.

TaqMan Analysis of ERBB Receptor mRNA Levels inEpendymoma Primary Tumors. Sufficient total RNA wasavailable from 16 cases derived from Group B and both normalbrain samples. This was reverse transcribed using the Super-Script preamplification system (Life Technologies, Inc., Rock-ville, MD), used in TaqMan quantitative RT-PCR analysis, to

determine the expression level of each ERBB receptor mRNAincluding the JM-a, JM-b, CYT-1, and CYT-2 ERBB4 splicevariants. The design and evaluation of ERBB mRNA quantita-tion by TaqMan analysis will be described in detail elsewhere.4

The oligonucleotide primers used were as follows (probes areshown in parentheses): ERBB1, 5�-ccacctgtgccatccaaact-3�, 5�-ggcgatggacgggatctt-3�, (5�-ccaggtcttgaaggctgtccaacgaat-3�);ERBB2, 5�-agccttgccccatcaactg-3�, 5�-aatgccaaccaccgcaga-3�, (5�-ccactcctgtgtggacctggatgaca-3�); ERBB3, 5�-ccctgcca-tgagaactgcac-3�, 5�-tcactgtcaaagccattgtcagat-3�, (5�-gtttgtc-ctaaacagtcttgaagctctggtctttt-3�); ERBB4JM-a, 5�-ccacccatgc-catccaaa-3�, 5�-ccaattactccagctgcaatca-3�, (5�-atggacgggcaat-tccactttacca-3�); ERBB4 JM-b, 5�-ccacccatgccatccaaa-3�, 5�-ccaattactccagctgcaatca-3�, (5�-ctcaagtattgaagactgcatcggcctgat-3�);ERBB4-CYT-1, 5�-caacatcccacctcccatctatac-3�, 5�-acactccttgttcag-cagcaaa-3�, (5�-tgaaattggacacagccctcctcctg-3�); ERBB4-CYT-2,5�-caacatcccacctcccatctatac-3�, 5�-acactccttgttcagcagcaaa-3�, (5�-aattgactcgaataggaaccagtttgtataccgagat-3�); and �-actin, 5�-atctg-gcaccacaccttctacaat-3�, 5�-ccgtcaccggagtccatca-3�, (5�-tgacccagat-catgtttgagaccttcaacac-3�). Overview of the procedure is as follows.PCR was performed using 300 nM primer, 200 nM (50 nM forCYT-2) probe, 12.5 �l of TaqMan universal PCR Master mix (PEBiosystems, Foster City, CA), and 0.5-�l template cDNA in 25-�ltotal volume. Thermal cycling (ABI PRISM 7700 Sequence De-tector; PE Biosystems) started with initial steps of 2 min at 50°Cand 10 min at 95°C, followed by 40 cycles of 15 s at 95°C and 1min at 60°C (63°C when ERBB4 CYT-2 was probed). CT values(the cycle number in which the signal exceeds the fixed thresholdvalue set above the background) were determined. cDNA qualitywas first confirmed in each case by RT-PCR analysis of �-actininternal control expression. Expression of each ERBB transcriptwas then determined and recorded as the percentage expressionrelative to the internal (�-actin) control RNA. All of the analyseswere conducted in triplicate.

Southern Blotting. High-molecular-weight DNA (10�g/sample), extracted from ependymoma tumor samples andnormal peripheral WBCs (diploid control), was digested withEcoRI overnight at 37°C and was subject to Southern blottinganalysis as described in detail previously (22). Membranes wereprobed with a [32P]dCTP-labeled 1.7-kb fragment of the humanERBB2 coding region for 16 h at 68°C. After washing in 2�SSC � 0.1% SDS and 0.5� SSC � 0.1% SDS, the membranewas analyzed by PhosphorImaging.

[3H]Thymidine Incorporation Assay. SJEp1 or SJEp2cells at passage 2 and 3 were seeded into each well (3 � 103

cells/well) of a 96-well plate in DMEM supplemented with 10%FBS and were allowed to adhere over night. The medium wasthen replaced with serum-free DMEM, and incubation pro-ceeded for 12 h. Ten �l of PBS, EGF, or NRG1� (both 10 ng/mlfinal concentration) were added and incubated with cells for12 h, followed by a 4-h long pulse of [methyl-3H]thymidine(0.1�Ci/well). Cells were then harvested onto Unifilter GF/C96-well microfilter plates (Packard BioScience, Meriden, CT),and incorporated radioactivity was measured by scintillation(TopCount NXT microplate counter; Packard BioScience). The

4 T. T. Junttila et al., manuscript in preparation.

3056 ERBB Signaling in Ependymoma



ability of the novel ERBB2 tyrosine kinase inhibitor WAY-177820 (gift of Dr. S. Rabindran, Wyeth-Ayerst Pharmaceuti-cals, St. Davids, PA) to block EGF stimulated [methyl-3H]thymidine incorporation was performed in an identical manner,except cells were incubated in the presence of the inhibitor(1.8–1800 nM) or of the vehicle for only 2 h before the additionof EGF. WAY-177820 is a potent and selective inhibitor ofERBB2 kinase activity in cell-free systems (IC50, 70 nM) andhas been shown to preferentially inhibit cell proliferation inERBB2-transfected 3T3 cells and ERBB2-overexpressing hu-man breast cancer cell lines (SKBr-3 and BT474; IC50, 2–4 nM;Ref. 25). WAY-177820 was dissolved in DMSO and stored inaliquots at 20°C before use. Final DMSO concentration in allof the experiments was maintained at �0.1%.

WAY-177820 Inhibition of EGF-induced ERBB2,ERK1/2, and AKT Phosphorylation. SJEp1 or SJEp2 cellswere grown to 80% confluence at passage 2 and 3 in DMEMsupplemented with 10% FBS before incubation in serum-freeDMEM for 14 h. Cells were then preincubated in 180 nM

WAY-177820 or DMSO for only 2 h before the addition of EGFor an equal volume of PBS. After 10 min, cells were lysed, andtotal protein lysates were analyzed by Western blotting. Acti-vation of ERBB2 receptor, ERK1/2, and AKT signaling in thepresence or absence of the inhibitor was detected by phospho-specific Western blotting analysis. Western blotting was per-formed as described above using antibodies raised againstY1248 of ERBB2 (Upstate Biotechnology, Lake Placid, NY)Y204 of ERK1/2 (Santa Cruz Biotechnology) and Ser473 ofAKT (New England Biolabs, MA). After probing with phospho-specific antibodies, all of the blots were stripped (62.5 mM Tris,2% SDS, 100 mM 2�-mercaptoethanol) for 30 min at 50°C andreprobed with antibodies against ERBB2 (NCL-CB11; Novo-castra), ERK1 (Santa Cruz Biotechnology), and AKT (NewEngland Biolabs).

Statistical Analysis. Relationships among clinicopatho-logical variables were assessed using standard statistical tech-niques including Fisher’s (2 � 2 comparisons) and Mann-Whitney tests. Univariate survival analysis was performed byconstruction of Kaplan-Meier survival curves and the log-ranktest (26). Because we have previously demonstrated that a Ki-67LI of �25% confers a significantly worse prognosis than lowerscores (24), patients were divided into two groups based on thiscutoff. Retrospective analyses may be biased by the inclusion ofpatients treated over broad time frames; therefore, survivalanalyses were restricted to patients diagnosed after 1980(n � 89).

ResultsThe ERBB2 and ERBB4 Receptors Are Frequently

Coexpressed in Childhood Ependymoma. The availabilityof tumor material allowed an assessment of ERBB1, ERBB2,ERBB3, and ERBB4 protein expression by IHC in 60 (61%), 59(60%), 59 (60%), and 48 (48%) Group A cases, respectively.Total protein and RNA was available from 20 and 16 Group Bcases, respectively. The results of expression analyses are sum-marized in Table 1.

ERBB4 immunoreactivity was observed in 100% of GroupA samples examined with a median percentage tumor cell

expression rate of 60% (Fig. 1). The majority of Group A casesalso coexpressed ERBB2. This receptor was detected in 78%(46 of 59) of tumors with a median percentage tumor cellexpression rate of 30% (Fig. 1). Similar high ERBB2 andERBB4 coexpression levels were detected in Group B tumors.Nineteen of 20 and 14 of 20 samples demonstrated readilydetectable levels of ERBB2 and ERBB4 by Western blotting,respectively (Fig. 2). In keeping with these data, all of the GroupB tumors that were analyzed expressed the ERBB4JM-a mRNAat levels higher than that detected in normal brain (Fig. 2).Tumor expression of the ERBB4JM-b variant proved much lesscommon and was observed at similar levels to that displayed bynormal control tissue. All analyzed Group B cases also ex-pressed the ERBB2 mRNA (Fig. 2). In the majority, this wasdetected at levels greater than that observed in normal brain. Inaddition, a significant positive relationship was observed be-tween the expression levels of ERBB2 and ERBB4 JM-a tran-scripts in Group B samples (r2 , 0.92; P � 0.0001; Fig. 3A).These data indicate that ERBB2 and ERBB4 are frequentlyco-overexpressed in childhood ependymoma and that this maybe governed by a common mechanism(s).

In contrast, ERBB1 and ERBB3 protein and mRNA wereabsent or present at low levels in ependymoma (Figs. 1 and 2).Four Group B cases (Fig. 2, F9, F12–14) did demonstraterelatively high ERBB1 transcript levels. This did not appear tobe translated into detectable protein (Fig. 2).

Elevated Expression of ERBB2 Is Not Mediated byGene Amplification in Ependymoma. ERBB2 overexpres-sion in human cancer commonly results from gene amplifica-tion, although overexpression from an nonamplified locus isalso well described. High-level expression of ERBB4 in humantumors invariably occurs in the absence of gene amplification(14, 15, 22). The close correlation between ERBB2 and ERBB4levels observed in Group B strongly suggests that mechanismsother than gene amplification are important in maintaining ele-vated ERBB2 expression in ependymoma. Southern blottinganalysis of six samples from Group B, including two with highERBB2 mRNA and protein expression (Fig. 3B, F13 and F14)confirmed this, demonstrating diploid ERBB2 copy numberonly.

High ERBB2 and ERBB4 Receptor Coexpression Lev-els Are Associated with Elevated Tumor Proliferative Activ-ity in Childhood Ependymoma. We next sought to deter-mine the significance of ERBB2 and ERBB4 coexpression inGroup A cases of ependymoma. First, we analyzed the relation-ship between receptor expression and histological anaplasia. Toincrease the objectivity of our assessment of anaplasia, weanalyzed independently the importance of three of its charac-teristic features (MVP, necrosis, and mitotic activity). MVP andnecrosis were detected in 46 (46.4%) and 64 (64.6%) casesrespectively (Table 1). The median and mean Ki-67 LI forGroup A tumors was 16.8% (range, 0.1–71.3%) and 20.5 15.4% (mean SD), respectively. Ki-67 LIs were significantlyhigher in tumors with elevated ERBB2 and ERBB4 coexpres-sion (�50% tumor cell receptor immunoreactivity; P � 0.02;Fig. 4A). In contrast, neither the presence of necrosis nor MVPwas related to the expression of any ERBB receptor or tumorKi-67 LI. Therefore, elevated ERBB2 and ERBB4 receptor




expression may promote tumor cell proliferation but not necro-sis or MVP in childhood ependymoma.

Evidence that tumor proliferation, but not necrosis orMVP, is more indicative of an anaplastic phenotype, was pro-vided by analysis of paired primary and relapsed tumors fromGroup A. Although a significant increase in Ki-67 LI wasobserved with disease progression (P � 0.01, Fig. 4 b), nosignificant difference in the presence of MVP or necrosis wasseen between primary and relapse samples. Insufficient materialwas available to assess potential differences in ERBB receptorexpression between primary and relapsed disease.

ERBB2 and ERBB4 Coexpression and Ki-67 LI Im-prove the Accuracy of Clinical Risk Assessment in Child-hood Ependymoma. Having demonstrated that ERBB2 andERBB4 coexpression levels and Ki-67 LI are significantly re-lated in ependymoma, we next evaluated the importance of thesevariables for patient survival. First, we analyzed the prognosticsignificance of all clinicopathological variables including pa-tient age (�3 versus �3 years), sex, degree of surgical resection(total versus partial), adjuvant therapy, tumor necrosis, MVP,Ki-67 LI, and ERBB receptor expression (all receptors inde-pendently and in combination) for Group A patients. To mini-

Fig. 1 A, immunohistochemical staining of thefour ERBB receptors in tumor samples derivedfrom Group A. B, graph illustrating the expressionfrequency of the four ERBB receptors detected byimmunohistochemical analysis of tumor samplesderived from Group A. Box extends from the 25thto the 75th percentile, with a line at the median.Bars, extension to the outlying values.




mize the bias inherent in survival analyses of patients treatedover broad time frames (Group A, 1972–1998), analysis wasrestricted to patients diagnosed after 1980 (n � 89).

In agreement with the existing literature, subtotal tumorresection (1–6) and high Ki-67 LI (24, 27–29) were signifi-cantly associated with a worse prognosis (Fig. 5, A and B). Thislatter finding is also in keeping with our observation that Ki-67LI increased with disease progression (Fig. 4B). A young age atdiagnosis (�3 years) also resulted in a poorer survival, althoughthis effect was marginal (Fig. 5C). ERBB2 and ERBB4 coex-pression data were available for only 39 cases diagnosed after1980. Although the impact of these receptors on patient survivaldid not reach statistical significance, a clear trend toward aworse outcome was seen in patients whose tumors coexpressedelevated levels of the ERBB2 and ERBB4 receptors (�50%

immunoreactivity for both receptors, Fig. 5D). The 10-yearoverall survival rate for patients with elevated ERBB2/ERBB4-coexpressing disease (20.6 12.8%, mean SE) was less thanone-half that of patients with lower levels of receptor expression(44.5 10.1%, mean SE). Therefore, we determined whethera combination of Ki-67 LI and ERBB2 and ERBB4 coexpres-sion levels might afford a more accurate assessment of diseaserisk. For this analysis, patients were divided into two groups.High-molecular-risk (Ki-67 LI �25% or high ERBB2 andERBB4 coexpression; n � 41) or low-molecular-risk (neitherKi-67 LI �25% nor high ERBB2 and ERBB4 coexpression;n � 21). This risk assessment had a prognostic value that wassuperior to that of either variable alone and was almost equal tothat associated with the degree of tumor resection (Fig. 5E).Finally, we recently demonstrated the value of combining clin-

Fig. 2 Western blotting and TaqMan RT-PCR analysis of ERBB receptor protein andmRNA expression respectively by the 20(F1–F20) Group B patients diagnosed andtreated at St. Jude Children’s Research Hos-pital. Primary tumor sites include: P, poste-rior fossa; S, spinal; C, cerebral cortex. Thefirst lane of each Western blot (WB) includesa positive control sample (see “Patients andMethods”). The second lane is blank. Bargraphs underneath the Western blot of therespective receptor, the results of TaqManRT-PCR analysis of each transcript by theependymoma primary tumors illustrate thepercentage expression of the specific ERBBmRNA (open bars) relative to internal actincontrol. The first two lanes (filled bars) in-clude results from an analysis of normal fron-tal cortex and fetal control brain, respec-tively.




ical and molecular risk markers, including ERBB2 expressionand the degree of surgical resection, in disease-risk stratificationof children with medulloblastoma (23). Therefore, we nextanalyzed whether a combination of molecular and clinical (de-gree of surgical resection) risk factors may also improve theaccuracy of disease risk assessment in ependymoma. For thisfinal analysis, patients were divided into three groups. High-risk(subtotal resection and high-molecular-risk, n � 23), interme-diate-risk (either total resection and high-molecular-risk or sub-total resection and low-molecular-risk, n � 31) and low-risk(total resection and low-molecular-risk, n � 8). This analysisprovided the greatest degree of survival discrimination (P �0.0001; Fig. 5F). Of particular note, this stratification identifiedpatients with an especially poor clinical outcome. Of 23 patientsclassified as “high-risk,” 19 have died with a median survival ofonly 2 years.

RTK I Signaling Promotes Second Messenger Activa-tion and Proliferation of Ependymoma Cells in Vitro ThatCan Be Blocked with an Inhibitor of ERBB2 Tyrosine Ki-nase Activity. To further investigate the ability of ERBBreceptor signaling to support ependymoma cell proliferation, weanalyzed the impact of exogenous ligand treatment on the pro-liferation of SJEp1 and SJEp2 cells in vitro. These explants wereestablished from histologically proven ependymoma, and bothexpressed the ependymoma-associated antigens GFAP and vi-mentin (Fig. 6A). Both SJEp1 and SJEp2 also expressed readilydetectable levels of ERBB1, ERBB2, and ERBB3 (Fig. 6B).However, ERBB4 protein expression was not detected. None-theless, we sought to characterize ERBB signaling in these cellsby activating ERBB2 in the context of an ERBB1/ERBB2

heterodimer using EGF or an ERBB3/ERBB2 heterodimer us-ing NRG1�. Exogenous EGF induced a significant and approx-imately equal proliferative response in both cell types (Fig. 6C).In contrast, although NRG1� significantly stimulated the pro-liferation of SJEp2 cells, a more limited effect was observed inSJEp1 cells.

We next investigated the ability of the ERBB tyrosinekinase inhibitor WAY-177820 to abrogate ERBB receptor sig-naling and proliferation in SJEp1 and SJEp2 cells (Fig. 7).Because both explants responded significantly and equally toEGF, this ligand was used to stimulate cells in inhibitor assays.Treatment of EGF-stimulated SJEp1 and SJEp2 cells withWAY-177820, significantly inhibited their proliferation in adose-dependent manner, with the maximum inhibition occurringat �180 nM. In addition, whereas EGF treatment induced sig-nificant phosphorylation of Y1248ERBB2 and Ser473Akt inboth cell types, this was markedly inhibited by exposure of cellsto 180 nM of WAY-177820. These data support the hypothesisthat RTK I signaling via ERBB2, stimulates the proliferation ofependymoma cells through pathways that include AKT. Fur-thermore, inhibitors of this pathway may represent a noveltherapeutic mechanism for ependymoma.

DiscussionThe management of pediatric ependymoma is unlikely to

improve without a greater understanding of disease biology.Such knowledge may provide a more accurate means of deter-mining disease risk. Although surgery has an established role inthe management and clinical outcome of ependymoma (1–6),the value of adjuvant chemo- and radiotherapy is less clear (1).Therefore, a more accurate means of disease risk stratification

Fig. 3 A, graph illustrating the significant correlation between ERBB2and ERBB4-JMa transcript expression levels measured by TaqManRT-PCR in primary ependymoma samples derived from Group B (r2 �0.92; P � 0.0001). B, Southern blot analysis of ERBB2 using 10 �g eachof DNA extracted from normal peripheral WBCs (Control), and sixependymoma samples derived from Group B. Numbers prefixed F,tumor sample numbers (correspond to those in Fig. 2).

Fig. 4 A, bar graph illustrating the significantly higher mean (SE)Ki-67 LI in Group A tumors with elevated coexpression of ERBB2 andERBB4 relative to those with lower receptor expression levels. B, graphillustrating the significant increase in Ki-67 LI observed in primary andpaired relapse tumor material obtained from five cases in Group A.




would allow better evaluation of the efficacy of these conven-tional therapies. Furthermore, tumorigenic signal pathways mayin turn serve as cellular targets for novel therapeutic approaches.The anti-ERBB2 monoclonal antibody Herceptin (30) and theBCR-ABL inhibitor STI571 (31) provide precedents for thisapproach.

In this comprehensive study of RTK I expression in child-hood ependymoma, we have identified coexpression of ERBB2and ERBB4 to be a frequent event in this disease, affecting over70% of cases. Our quantitative RT-PCR analysis demonstratedthat in the great majority of cases, this expression level is greaterthan that found in normal brain. Our study also provides evi-dence that this expression is significant for the biology ofependymoma. First, we show that tumors co-overexpressingERBB2 and ERBB4 demonstrate significantly elevated prolif-erative activity. Therefore, signaling via ERBB2 homodimers orERBB2/ERBB4 heterodimers may deregulate normal cell cycle

and proliferative control in this disease. Indeed, our in vitro dataindicate that signaling via ERBB2 is pro-proliferative inependymoma. High tumor proliferative activity is one of anumber of features associated with anaplasia (5). This histolog-ical diagnosis is based on a constellation of morphologicalfeatures that include tumor necrosis and MVP in addition toproliferative activity. The individual survival significance ofeach of these factors is unknown. This may explain, at least inpart, the controversy surrounding the prognostic value ofanaplasia. Here, we demonstrate that, although elevated ERBB2and ERBB4 coexpression is significantly related to Ki-67 LI,neither of these variables is associated with tumor necrosis orMVP. Furthermore, we show that, although Ki-67 LI increaseswith disease progression, no difference in necrosis or MVP canbe detected between primary and relapsed tumor samples. Fi-nally, our survival analyses indicate that, although Ki-67 LIsignificantly relates to patient outcome, no prognostic signifi-

Fig. 5 Prognostic significance ofclinicopathological factors in Group Apatients. Kaplan-Meier survival curvesand the associated log-rank statistic areshown for: A, degree of surgical resec-tion; B, Ki-67 LI; C, age at diagnosis;D, ERBB2/ERBB4 coexpression sta-tus; E, molecular risk [High-risk;Ki-67 LI �25% or elevated ERBB2/ERBB4 coexpression; and Low-risk,neither Ki-67 LI �25% nor elevatedERBB2/ERBB4 coexpression]; F,combined molecular and clinical riskstratification [High-risk, molecularhigh-risk and subtotal tumor resection;intermediate-risk (Inter-risk), molecu-lar high-risk and total tumor resectionor molecular low-risk and subtotal tu-mor resection; Low-risk, molecularlow-risk and total tumor resection].




cance is associated with tumor necrosis or MVP. Therefore, wepropose that Ki-67 LI, but not necrosis or MVP, is a reliablemarker of anaplastic, aggressive ependymoma. Furthermore, ourdata suggest that ERBB2 and ERBB4 receptor signaling may

contribute to an aggressive tumor phenotype by mediating tu-mor proliferative activity.

The results of this study are intriguing in the light of ourrecent analyses of RTK I expression in medulloblastoma, asecond posterior fossa tumor of childhood. In these studies, wealso show ERBB2 and ERBB4 co-overexpression to be a fre-quent event in medulloblastoma (20, 22, 23), to significantlyassociate with increased proliferative potential (21), and topredict for poor clinical outcome (20, 23). ERBB2 and ERBB4receptor cooperative signaling is crucial for normal hind braindevelopment (16–19). Therefore, deregulation of these recep-tors may be important for tumorigenesis in a variety of cellularbackgrounds within the developing posterior fossa.

In addition to proliferation, signaling via the ERBB4 re-ceptor is associated with a diverse number of cellular responsesthat include AKT-mediated cell survival and chemotaxis (32,

Fig. 6 In vitro analyses of ERBB receptor expression and proliferativeresponse to ERBB-specific ligands EGF and NRG1�. A, immunofluo-rescence microscopy of GFAP and vimentin expression by the SJEp1and SJEp2 cells. B, Western blotting analysis characterizing expressionof the four ERBB receptors by SJEp1 and SJEp2 cells. C, proliferativeresponse of SJEp1 and SJEp2 cells to 10 ng/ml of exogenous EGF orNRG1� measured by [methyl-3H]thymidine incorporation.

Fig. 7 The ERBB2 tyrosine kinase inhibitor WAY-1EGF blocks EGF-induced AKT activation and proliferation in SJEp1 and SJEp2 cells. A,phospho(p)-specific Western blotting analysis of ERBB2, AKT, andERK1/2, after treatment of ependymoma cells with (�) or without ()EGF in the presence (�) or absence () of 180 nM WAY-177820. Blotswere reprobed for total ERBB2, AKT, and ERK1. B, growth inhibitioncurve illustrating the ability of increasing concentrations of WAY-177820 to inhibit EGF-stimulated DNA synthesis by the two ependy-moma cell lines. Maximum inhibition for both cell lines was achieved at�180 nM.




33). The exact cell response elicited by ERBB4 signaling isdetermined in part by the expression of alternative receptorisoforms. These isoforms are generated by RNA splicing of thejuxtamembrane (JM-a to d isoforms; Refs. 22, 34) and cytoplas-mic (CYT-1 and -2 isoforms; Refs. 35, 36) regions of theERBB4 primary transcript. Heterodimerization of the ERBB4receptor with other RTK I family members may also modulatethe signal response (14, 15, 37, 38). This is particularly true ofERBB2, which significantly enhances signal potency. Here, wehave demonstrated that ERBB4 expression in ependymomaincludes both JM-a/CYT-1 and JM-a/CYT-2 forms, whereasJM-b variants are much less common. The juxtamembraneregion of ERBB4 appears to dictate the sensitivity of the JMisoforms to receptor ectodomain proteolytic cleavage (34). TheJM-a variant is reported to be more sensitive to such processing.However, the overall effect that this exerts on receptor functionremains to be determined. The CYT-1 sequence includes aphosphoinositide 3-kinase binding site (35, 36) that retains thecapacity to induce cell proliferation and protection from apo-ptosis and to increase chemotaxis. CYT-2, which lacks phos-phoinositide 3-kinase binding supports only ligand-inducedproliferation (32). The identification of both CYT forms inependymoma suggests that ERBB4 signaling in this disease maymediate a broad range of cellular responses, all of which maycontribute to malignant behavior.

The results of this study highlight two important areas forfuture clinical evaluation. To our knowledge, this is the firstreport indicating that a combination of molecular and clinicalmarkers provides an accurate means of predicting prognosis inependymoma. We propose that this should be studied in a largerseries of contemporarily treated patients. Limited patient num-bers in the present study prevented us from accounting forpatient age in our combined molecular and clinical survivalanalyses. Therefore, a larger prospective study would allow theimportance of this and other clinical covariables to also beassessed. Finally, this study suggests that ERBB2 receptor-targeted therapies may represent a useful novel therapeuticapproach for childhood ependymoma. The ependymoma cellexplants described in this study expressed ERBB1, ERBB2, andERBB3 but not the ERBB4 receptor. Therefore the culture andanalysis of ERBB4-expressing ependymoma cells will be im-portant for the future analysis of RTK I signaling biology inependymoma.

AcknowledgmentsWe thank Dr. Fred Behm and Mike Jaynes (St. Jude Children’s

Research Hospital, Memphis, TN), Dr. Seth Love (Frenchay Hospital,Bristol, England), and Dr. Brian Harding (Great Ormond Street Hospi-tal, London, England) for supplying tumor material, and Phil Steart andJean Buontempo (Southampton General Hospital) and William Mac-Meekin (Newcastle General Hospital) for excellent technical assistance.

References1. Bouffet, E., Perilongo, G., Canete, A., and Massimino, M. Intracra-nial ependymomas in children: a critical review of prognostic factorsand a plea for cooperation. Med. Pediatr. Oncol., 30: 319–331, 1998.2. Robertson, P. L., Zeltzer, P. M., Boyett, J. M., Rorke, L. B., Allen,J. C., Geyer, J. R., Stanley, P., Li, H., Albright, A. L., McGuire-Cullen,P., Finlay, J. L., Stevens, K. R. Jr., Milstein, J. M., Packer, R. J., and

Wisoff, J. Survival and prognostic factors following radiation therapyand chemotherapy for ependymomas in children: a report of the Chil-dren’s Cancer Study Group. J. Neurosurg., 88: 695–703, 1998.

3. Horn. B., Heideman, R., Geyer, R., Pollack, I., Packer, R., Goldwein,J., Tomita, T., Schomberg, P., Ater, J., Luchtman-Jones, L., Rivlin, K.,Lamborn, K., Prados, M., Bollen, A., Berger, M., Dahl, G., McNeil, E.,Patterson, K., Shaw, D., Kubalik, M., and Russo, C. A multi-institu-tional retrospective study of intracranial ependymoma in children: iden-tification of risk factors. J. Pediatr. Hematol. Oncol., 21: 203–211, 1999.

4. Grill. J., Le Deley, M. C., Gambarelli, D., Raquin, M. A., Couanet,D., Pierre-Kahn, A., Habrand, J. L., Doz, F., Frappaz, D., Gentet, J. C.,Edan, C., Chastagner, P., and Kalifa, C.; French Society of PediatricOncology. Postoperative chemotherapy without irradiation for ependy-moma in children under 5 years of age: a multicenter trial of the FrenchSociety of Pediatric Oncology. J. Clin. Oncol., 19: 1288–1296, 2001.

5. Wiestler, O. D., Schiffer, D., Coons, S. W., Prayson, R. A., andRosenblum, M. K. Ependymoma. In: P. Kleihues and W. K. Cavenee(eds.), World Health Organisation Classification of Tumors. Pathologyand Genetics of Tumors of the Nervous System, pp. 72–77. Lyon,France: IARC Press, 2000.

6. Schiffer, D., and Giordana, M. T. Prognosis of ependymoma. ChildsNerv. Syst., 14: 357–361, 1998.

7. Needle, M. N., Goldwein, J. W., Grass, J., Cnaan, A., Bergman, I.,Molloy, P., Sutton, L., Zhao, H., Garvin, J. H., Jr., and Phillips. P. C.Adjuvant chemotherapy for the treatment of intracranial ependymomaof childhood. Cancer (Phila.), 80: 341–347, 1997.

8. Bijlsma, E. K., Voesten, A. M., Bijleveld, E. H., Troost, D., West-erveld, A., Merel, P., Thomas, G., and Hulsebos, T. J. Molecularanalysis of genetic changes in ependymomas. Genes ChromosomesCancer, 13: 272–277, 1995.

9. von Haken, M. S., White, E. C., Daneshvar-Shyesther, L., Sih, S.,Choi, E., Kalra, R., and Cogen, P. H. Molecular genetic analysis ofchromosome arm 17p and chromosome arm 22q DNA sequences insporadic pediatric ependymomas. Genes Chromosomes Cancer, 17:37–44, 1996.

10. Mazewski, C., Soukup, S., Ballard, E., Gotwals, B., and Lampkin,B. Karyotype studies in 18 ependymomas with literature review of 107cases. Cancer Genet Cytogenet., 113: 1–8, 1999.

11. Hirose, Y., Aldape, K., Bollen, A., James, C. D., Brat, D., Lamborn,K., Berger, M., and Feuerstein, B. G. Chromosomal abnormalitiessubdivide ependymal tumors into clinically relevant groups. Am. J.Pathol., 158: 1137–1143, 2001.12. Lamszus, K., Lachenmayer, L., Heinemann, U., Kluwe, L., Finckh,U., Hoppner, W., Stavrou, D., Fillbrandt, R., and Westphal, M. Molec-ular genetic alterations on chromosomes 11 and 22 in ependymomas.Int. J. Cancer., 91: 803–808, 2001.13. Carter M., Nicholson J., Ross F., Crolla J., Allibone R., Balaji V.,Perry R., Walker D., Gilbertson R., and Ellison D. Genetic abnormali-ties detected in histological variants of ependymoma by comparativegenomic hybridisation. Br. J. Cancer, in press, 2002.14. Olayioye, M., A., Neve, R. M., Lane, H. A., and Hynes, N. E. TheERBB signaling network: receptor heterodimerization in developmentand cancer. EMBO J., 19: 3159–3167, 2000.15. Yarden, Y., and Sliwkowski, M. X. Untangling the ERBB signal-ling network. Nat. Rev. Mol. Cell Biol., 2: 127–137, 2001.16. Burden, S., and Yarden, Y. Neuregulins and their receptors: aversatile signaling module in organogenesis and oncogenesis. Neuron,18: 847–855, 1997.17. Lambert de Rouvroit, C., and Goffinet, A. M. A new view of earlycortical development. Biochem. Pharmacol., 56: 1403–1409, 1998.18. Buonno, A., and Fischbach, G. D. Neuregulin and ERBB receptorsignaling pathways in the nervous system. Curr. Opin. Neurobiol., 11:287–296, 2001.19. Gerecke., K. M., Wyss, J. M., Karavanova, I., Buonno, A., andCarroll, S. L. ERBB transmembrane tyrosine kinase receptors are dif-ferentially expressed throughout the adult central nervous system.J. Comp. Neurol., 433: 86–100, 2001.




20. Gilbertson, R. J., Perry, R. H., Kelly, P. J., Pearson, A. D., andLunec, J. Prognostic significance of HER2 and HER4 coexpression inchildhood medulloblastoma. Cancer Res., 57: 3272–3280, 1997.21. Gilbertson, R. J., Jaros, E. A. B., Perry, R. H., and Pearson, A. D. J.Percentage mitotic index: a new prognostic factor for childhood medul-loblastoma. Eur. J. Cancer, 33: 609–615, 1997.22. Gilbertson, R. J., Hernan, R., Pietsch, T., Pinto, L., Scotting, P.,Allibone, R., Ellison, D., Perry, R. H., Pearson, A. D., and Lunec, J.Novel ERBB4 receptor spliced variants are frequently expressed inchildhood medulloblastoma. Genes Chromosomes Cancer, 31: 288–294, 2001.23. Gilbertson, R. J., Hernan, R., Wickramasinghe, C., Balaji, V., Hunt,D., Jones-Wallace, D., Crolla, J., Perry, R. H., Lunec, J., Pearson, A. D.,and Ellison, D. Molecular stratification of disease risk in medulloblas-toma. Br. J. Cancer, 85: 705–712, 2001.24. Bennetto, L., Foreman, B., Harding, R., Hayward, R., Ironside, J.,Love, S., and Ellison, D. Ki-67 immunolabelling index is a prognosticindicator in childhood posterior fossa ependymomas. Neuropath. Appl.Neurobiol., 24: 434–440, 1998.25. Rabindran, S. K., Discafani, C. M., Floyd, M. B., Hallett, W. A.,Johnson, B. D., Nilakantan, R., Overbeek, E., Reich, M. F., Rosfjord, E.,Shen, R., Shi, S., Tsou, H-W., Wang, Y-F., and Wissner A. Antitumoractivity of compound 820 in xenografts: an orally active irreversibleinhibitor of the Her-2 receptor tyrosine kinase. Proc. Am. Assoc. CancerRes., 43: 1001, 2002.26. Peto, R., Pike, M. C., Armitage, P., Breslow, N. E., Cox, D. R.,Howard, S. V., Mantel, N., McPherson, K., Peto, J., and Smith P. G.Design and analysis of randomized clinical trials requiring prolongedobservation of each patient, II: analysis and examples. Br. J. Cancer, 35:1–39, 1977.27. Ritter, A. M., Hess, K. R., McLendon, R. E., and Langford, L. A.Ependymomas: MIB-1 proliferation index and survival. J. Neurooncol.,40: 51–57, 1998.28. Rushing, E. J., Brown, D. F., Hladik, C. L., Risser, R. C., Mickey,B. E., and White, C. L. Correlation of bcl-2, p53 and MIB-1 expressionwith ependymoma grade and subtype. Mod. Pathol., 11: 464–470,1998.29. Prayson, R. Clinicopathological study of 61 patients with ependy-moma including MIB-1 immunohistochemistry. Ann. Diagn. Pathol., 3:11–18, 1999.

30. Cobleigh, M. A., Vogel, C. L., Tripathy, D., Robert, N. J., Scholl,S., Fehrenbacher, L., Wolter, J. M., Paton, V., Shak, S., Lieberman, G.,and Slamon, D. J. Multinational study of the efficacy and safety ofhumanized anti-HER2 monoclonal antibody in women who have HER2-overexpressing metastatic breast cancer that has progressed after chem-otherapy for metastatic disease. J. Clin. Oncol., 17: 2639–2648, 1999.

31. Druker, B. J., Talpaz, M., Resta, D. J., Peng, B., Buchdunger, E.,Ford, J. M., Lydon, N. B., Kantarjian, H., Capdeville, R., Ohno-Jones,S., and Sawyers, C. L. Efficacy and safety of a specific inhibitor of theBCR-ABL tyrosine kinase in chronic myeloid leukemia. N. Engl.J. Med., 344: 1031–1037, 2001.32. Kainulainen, V., Sundvall, M., Maatta, J. A., Santiestevan, E.,Klagsbrun, M., and Elenius, K. A natural ERBB4 isoform that does notactivate phosphoinositide 3-kinase mediates proliferation but not sur-vival or chemotaxis. J. Biol. Chem., 275: 8641–8649, 2000.33. Flores, A. I., Mallon, B. S., Matsui, T., Ogawa, W., Rosenzweig, A.,Okamoto, T., and Macklin, W. B. Akt-mediated survival of oligoden-drocytes induced by neuregulins. J. Neurosci., 20: 7622–7630, 2000.34. Elenius, K., Corfas, G., Paul, S., Choi, C. J., Rio, C., Plowman,G. D., and Klagsbrun, M. A novel juxtamembrane domain isoform ofHER4/ERBB4. Isoform-specific tissue distribution and differentialprocessing in response to phorbol ester. J. Biol. Chem., 272: 26761–26768, 1997.35. Sawyer, C., Hiles, I., Page, M., Crompton, M., and Dean, C. TwoERBB-4 transcripts are expressed in normal breast and in most breastcancers. Oncogene, 17: 919–924, 1998.36. Elenius, K., Choi, C. J., Paul, S., Santiestevan, E., Nishi, E., andKlagsbrun, M. Characterization of a naturally occurring ERBB4 isoformthat does not bind or activate phosphatidyl inositol 3-kinase. Oncogene,18: 2607–2615, 1999.37. Zhang, K., Sun, J., Liu, N., Wen, D., Chang, D., Thomason, A., andYoshinaga, S. K. Transformation of NIH 3T3 cells by HER3 or HER4receptors requires the presence of HER1 or HER2. J. Biol. Chem., 271:3884–3890, 1996.38. Pinkas-Kramarski, R., Shelly, M., Guarino, B. C., Wang, L. M.,Lyass, L., Alroy, I., Alimandi, M., Kuo, A., Moyer, J. D., Lavi, S.,Eisenstein, M., Ratzkin, B. J., Seger, R., Bacus, S. S., Pierce, J. H.,Andrews, G. C., Yarden, Y., and Alamandi M. ERBB tyrosine kinasesand the two neuregulin families constitute a ligand-receptor network.Mol. Cell. Biol., 18: 6090–6101, 1998.




2002;8:3054-3064. Clin Cancer Res Richard J. Gilbertson, Lyndsay Bentley, Roberto Hernan, et al. Target for This DiseaseProliferation and Represents a Potential Novel Therapeutic ERBB Receptor Signaling Promotes Ependymoma Cell

Updated version

http://clincancerres.aacrjournals.org/content/8/10/3054

Access the most recent version of this article at:

Cited articles

http://clincancerres.aacrjournals.org/content/8/10/3054.full#ref-list-1

This article cites 35 articles, 9 of which you can access for free at:

Citing articles

http://clincancerres.aacrjournals.org/content/8/10/3054.full#related-urls

This article has been cited by 19 HighWire-hosted articles. Access the articles at:

E-mail alerts related to this article or journal.Sign up to receive free email-alerts

Subscriptions

Reprints and

[email protected] at

To order reprints of this article or to subscribe to the journal, contact the AACR Publications

Permissions

Rightslink site. Click on "Request Permissions" which will take you to the Copyright Clearance Center's (CCC)

.http://clincancerres.aacrjournals.org/content/8/10/3054To request permission to re-use all or part of this article, use this link



http://clincancerres.aacrjournals.org/content/8/10/3054.full#ref-list-1

http://clincancerres.aacrjournals.org/content/8/10/3054.full#related-urls

http://clincancerres.aacrjournals.org/cgi/alerts

mailto:[email protected]



erbb receptor signaling promotes ependymoma cell...

Documents