GeneExpressionProfilesAssociatedwithResponseto ...treatment and outcome of ovarian cancer. This topic has been the subject of intense research, and previous studies on chemoresistance

Gene Expression Profiles Associated with Response toChemotherapy in Epithelial Ovarian CancersAmir A. Jazaeri,1Christopher S. Awtrey,4 Gadisetti V.R. Chandramouli,1 Yao Eric Chuang,2

Javed Khan,3 Christos Sotiriou,1Olga Aprelikova,1Cindy J. Yee,4 Kristin K. Zorn,6

MichaelJ. Birrer,6 J. Carl Barrett,1andJeff Boyd4,5

Abstract Purpose:The goal of this study was to determine whether distinct gene expression profiles areassociatedwith intrinsic and/or acquired chemoresistance in epithelial ovarian carcinoma.Experimental Design: Gene expression profiles were generated from 21primary chemosensi-tive tumors and 24 primary chemoresistant tumors using cDNA-basedmicroarrays. Gene expres-sion profiles of both groups of primary tumors were then compared with those of 15 ovariancarcinomas obtained following platinum-based chemotherapy (‘‘postchemotherapy’’ tumors).A theme discovery tool was used to identify functional categories of genes involved in drugresistance.Results:Comparisonof primary chemosensitive and chemoresistant tumors revealed differentialexpression of 85 genes (P < 0.001). Comparison of gene expression profiles of primary chemo-sensitive tumors andpostchemotherapy tumors revealedmore robust differenceswith 760 genesdifferentiating the two groups (P < 0.001). In contrast, only 230 genes were differentiallyexpressedbetweenprimary chemoresistant and postchemotherapy groups (P < 0.001).Commonto both gene lists were178 genes representing transcripts differentially expressed between post-chemotherapy tumors and all primary tumors irrespective of intrinsic chemosensitivity. The geneexpression profile of postchemotherapy tumors compared with that of primary tumors revealedstatistically significant overrepresentation of genes encoding extracellular matrix ^ relatedproteins.Conclusions:These data show that gene expression profiling can discriminate primary chemo-resistant fromprimary chemosensitive ovarian cancers. Gene expressionprofiles were also identi-fied that correlatewith states of intrinsic and acquired chemoresistance and that represent targetsfor future investigation and potential therapeutic interventions.

Platinum-based combination chemotherapy is the standardfirst-line treatment for advanced-stage epithelial ovarian carci-noma. For the f75% of patients diagnosed with advanced-stage disease, 20% to 30% progress on or rapidly become

resistant to this treatment and subsequently show low responserates to other second-line agents (1–3). Early identification ofthis group of patients could lead to their enrollment in clinicaltrials or treatment with other experimental therapeutics becausestandard treatment affords them little benefit. Among initiallychemosensitive patients, the vast majority will eventuallyrelapse. Thus, chemoresistance may be present at the outsetof treatment (intrinsic resistance) or may develop duringtreatment (acquired resistance). In practice, ovarian cancersare considered ‘‘platinum sensitive’’ if the clinical progressionfree interval is >6 months, and evidence suggests that the longerthis interval, the higher the subsequent response rates toadditional chemotherapy (4–6).

Understanding the biological mechanisms underlying che-moresistance is of utmost importance for improving thetreatment and outcome of ovarian cancer. This topic has beenthe subject of intense research, and previous studies onchemoresistance in ovarian cancer have investigated potentialinvolvement of molecules involved in drug transport, apopto-sis, DNA repair, and detoxification pathways (7–11). Much ofthis research has been done using cell culture models and farfewer data are available on the relevance of these studies to, andbiomarkers and potential mechanisms of drug resistance for,clinical samples.

Cancer Therapy: Preclinical

Authors’ Affiliations: 1Laboratory of Biosystems and Cancer and 2RadiationBiology Branch, Center for Cancer Research, and 3Pediatric Oncology Branch,National Cancer Institute, Bethesda, Maryland; 4Gynecology and Breast ResearchLaboratory, Department of Surgery, and 5Department of Medicine, MemorialSloan-Kettering Cancer Center, New York, New York; and 6Department of Cell andCancer Biology, Center for Cancer Research, National Cancer Institute, Rockville,MarylandReceived12/27/04; revised 5/31/05; accepted 6/16/05.Grant support: NIH grant U01CA88175 (J. Boyd) and Gynecologic CancerFoundation, National Cancer Institute Gynecologic Oncology Fellowship Program(A.A. Jazaeri and K.K. Zorn).The costs of publication of this article were defrayed in part by the payment of pagecharges.This article must therefore be hereby marked advertisement in accordancewith18 U.S.C. Section1734 solely to indicate this fact.Note: Supplementary data for this article are available at Clinical Cancer ResearchOnline (http://clincancerres.aacrjournals.org/).Requests for reprints: Jeff Boyd, Department of Surgery, Memorial Sloan-Kettering Cancer Center, Box 201, 1275 York Avenue, NewYork, NY10021. Phone:212-639-8608; Fax: 212-717-3538; E-mail: [email protected].

F2005 American Association for Cancer Research.doi:10.1158/1078-0432.CCR-04-2682

www.aacrjournals.orgClin Cancer Res 2005;11(17) September1, 2005 6300

Research. on August 10, 2020. © 2005 American Association for Cancerclincancerres.aacrjournals.org Downloaded from

http://clincancerres.aacrjournals.org/

The availability of new high-throughput screening techniqueshas allowed for more global investigations of molecular profilesassociated with chemoresistance. In the present study, cDNAmicroarrays were used to investigate gene expression patterns

associated with both intrinsic and acquired chemoresistance inovarian cancer. The first aim of this investigation was to deter-mine if intrinsically chemoresistant and chemosensitive tumorscould be distinguished based on their gene expression profiles.

Table1. Clinicopathologic features of primary ovarian cancer cases

Case no. Age* Histology Stage Grade Chemotherapy DFEc

Primary chemoresistantR1 49 Clear cell IIIC 3 Carboplatin/paclitaxel 0R2 61 Endometrioid IIIC 3 Carboplatin/paclitaxel 0R3 61 Serous IV 3 Carboplatin/paclitaxel 0R4 49 Serous IIIC 2 Carboplatin/paclitaxel 0R5 77 Serous IIIC NAb Carboplatin/paclitaxel 0R6 49 Serous IIIC 2 Carboplatin/paclitaxel 0R7 55 Serous IV 3 Carboplatin/paclitaxel 0R8 56 Serous IIIC 2-3 Carboplatin/paclitaxel 0R9 61 Serous IIIC 3 Carboplatin/paclitaxel 0R10 73 Serous IV 3 Carboplatin/paclitaxel 0R11 78 Serous IIIC 3 Carboplatin/paclitaxel 0R12 71 Serous IV 2 Carboplatin/paclitaxel 0R13 52 Serous IV 2 Carboplatin/paclitaxel 0R14 64 Serous IIIA 1 Carboplatin/paclitaxel 0R15 49 Serous IIIC 2 Carboplatin/paclitaxel 0R16 47 Serous IV 3 Carboplatin/paclitaxel 0R17 66 Serous IIIC 3 Cisplatin/cyclophosphamide 0R18 69 Serous IV 2 Cisplatin/cyclophosphamide 0R19 44 Serous IV 2 HDx cisplatin/cyclophosphamide 0R20 35 Serous IV 2 Cisplatin/cyclophosphamide 0R21 67 Serous IIIC 2-3 Cisplatin/cyclophosphamide 0R22 63 Endometrioid IIIC 3 Carboplatin/paclitaxel 0R23 76 Serous IIIC 3 Carboplatin 0R24 65 Serous IIIC 2 Carboplatin/paclitaxel 0

Primary chemosensitiveS1 51 Serous IV 2 Carboplatin/paclitaxel 35S2 45 Endometrioid IIIC 2-3 HDCarboplatin/paclitaxel 41S3 65 Mixed IIIC 3 Carboplatin/paclitaxel 13S4 77 Serous IIIC 2-3 Carboplatin/paclitaxel 31S5 53 Serous IIIC 3 Carboplatin/paclitaxel 32S6 71 Serous IIIC 3 Carboplatin/paclitaxel 30S7 55 Serous IIIC 3 Carboplatin/paclitaxel 19S8 78 Endometrioid IIIC 2-3 Carboplatin/paclitaxel 24S9 46 Endometrioid IIIC 3 Carboplatin/paclitaxel 21S10 69 Serous IIIC 2 Carboplatin/paclitaxel 14S11 54 Serous IV 1 Carboplatin/docetaxel 18S12 53 Serous IIIC 2 Carboplatin/paclitaxel 18S13 77 Serous IIIB 3 Carboplatin/paclitaxel 21S14 56 Serous IIC 2 Carboplatin/paclitaxel 16S15 44 Serous IIIC 2 Carboplatin/paclitaxel 16S16 59 Carcinoma IIIC 3 Carboplatin/paclitaxel 16S17 44 Serous IIIC NA Carboplatin/paclitaxel 13S18 69 Serous IIIC 2 Carboplatin/paclitaxel 21S19 70 Carcinoma IIIC 3 Carboplatin/paclitaxel 36S20 74 Serous IIIC 2 Carboplatin/paclitaxel 21S21 57 Serous IIIC 2 Carboplatin/paclitaxel 18

*Age at time of surgery.cDisease-free interval following completion of chemotherapy (months).bNot available.xHigh-dose chemotherapy with bone marrow transplant.

Molecular Profile of Drug Resistance in Ovarian Cancer

www.aacrjournals.org Clin Cancer Res 2005;11(17) September1, 20056301



For this part of the investigation, a case was classified asintrinsic chemoresistant based on persistent or recurrent diseasewithin 6 months of initiating first-line platinum-based combi-nation chemotherapy. Chemosensitive tumors were classifiedas such based on a complete response to chemotherapy and aplatinum-free interval of z13 months. These conservativeclinical criteria for defining platinum sensitivity and resistancewere employed to exclude tumors with intermediate levels ofresistance. In the second part of the investigation, geneexpression profiles of tumors obtained following chemotherapy(‘‘postchemotherapy’’ samples) were compared with those ofthe chemosensitive and chemoresistant primary tumors. Thepostchemotherapy group consisted of tumors from ninepatients treated with neoadjuvant chemotherapy who subse-quently underwent an interval cytoreductive surgery and fromsix patients who had residual cancer present at the time ofsecond-look surgery following chemotherapy.

Gene expression profiles of these postchemotherapy tumorswere compared with those of our primary (i.e., chemonaive)chemosensitive and chemoresistant tumors. The rationale forthis approach was 2-fold. First, after each cycle of cytotoxicchemotherapy, the ‘‘log kill’’ effect leads to a significant reductionin the number of tumor cells that are sensitive to theadministered therapy (12, 13). Second, tumor cells that survivethe treatment are likely to experience changes in gene expressionthat allow them to withstand the selective pressure of the drugsused. Hence, tumor samples obtained shortly following chemo-therapy are enriched in resistant clones and are likely to displaythe molecular signature associated with chemoresistance.

Materials and Methods

Tissue specimens. This study was approved by the Memorial Sloan-Kettering Cancer Center Institutional Review Board. All tumor samples

were obtained at the time of surgery, frozen in liquid nitrogen, andstored at �80jC until use. Information on treatment and response wasobtained from patient chart review. Intrinsically chemoresistant tumorswere defined as those associated with persistent or recurrent diseasewithin 6 months of the initiation of first-line platinum-basedcombination chemotherapy. Chemosensitive tumors were defined asthose with a complete response to chemotherapy and a platinum-freeinterval of z13 months. The postchemotherapy group is composed ofpatients who had either surgical debulking following chemotherapy(i.e., neoadjuvant chemotherapy) or residual tumor at the time of asecond-look procedure. All of these tumor samples were obtainedwithin 6 weeks of the last cycle of chemotherapy. Ovarian tissues fromtwo postmenopausal women obtained at the time of salpingoophor-ectomy for benign indications were used for comparative purposes.

RNA preparation and cDNA microarray analysis. Isolation of RNA

was done using the RNeasy column (Qiagen, Valencia, CA) according

to the manufacturer’s instructions. The integrity of RNA was verified bydenaturing gel electrophoresis. Total RNA was linearly amplified using a

modification of the Eberwine procedure (14). Briefly, total RNA wasreverse transcribed by using a 63-nucleotide synthetic primer contain-

ing the T7 RNA polymerase binding site [5V-GGCCAGTGAATTGTAA-

TACGACTCACTATAGGGA-GGCGG(T)24-3V]. Second-strand cDNAsynthesis (producing double-stranded cDNA) was done with RNase

H, Escherichia coli DNA polymerase I, and E. coli DNA ligase(Invitrogen, Carlsbad, CA). After cDNA was blunt ended with T4

DNA polymerase (Invitrogen), purification was accomplished by

phenol/chloroform/isoamyl alcohol extraction and ammonium ace-tate/ethanol precipitation. The double-stranded cDNA was then

transcribed using T7 polymerase (T7 Megascript kit, Ambion, Austin,TX), yielding amplified antisense RNA that was purified using RNeasy

mini-columns. Commercially available pooled total RNA from 10

different human cell lines (Stratagene, La Jolla, CA) was amplified andused as the reference for cDNA microarray experiments.

Investigation of gene expression differences between primary chemo-sensitive and chemoresistant tumors was done using two separatecDNA microarrays to maximize the number of genes screened. The twocDNA microarrays contained 32,448 and 7,585 features each for acombined total of 40,033 transcripts. The comparison between the

Table 2. Clinicopathologic features of postchemotherapy ovarian cancer cases

Case no. Age* Histology Stage Grade Clinical hx Chemotherapy

PC1 57 Serous IIIC 2 NC/IDc Carboplatin/paclitaxel� 3PC2 74 Serous IIIB 3 NC/ID Carboplatin/paclitaxel� 3PC3 59 Serous IIIC 3 NC/ID Carboplatin/paclitaxel� 3PC4 77 Serous IIIC 2 NC/ID Carboplatin/paclitaxel� 6PC5 73 Serous IIIC 2 NC/ID Carboplatin/paclitaxel� 6PC6 68 Undifferentiated IV 3 NC/ID Carboplatin/paclitaxel� 6PC7 22 Serous IIIC 1 NC/ID Carboplatin/paclitaxel� 6PC8 71 Serous IV 3 NC/ID Carboplatin/paclitaxel� 6PC9 71 Serous IIIC 3 NC/ID Carboplatin/paclitaxel� 6PC10 45 Endometrioid IIIC 2 PSLb HDxCarboplatin/paclitaxelPC11 56 Serous IIIC 3 PSL Carboplatin/paclitaxel� 6PC12 58 mixed IIIC 3 PSL Carboplatin/paclitaxel� 6PC13 66 Serous IIIC 3 PSL Carboplatin/paclitaxel� 6PC14 42 Serous IIIC 3 PSL Carboplatin/paclitaxel� 6PC15 28 Serous IIIC 2 PSL Carboplatin/paclitaxel� 6

*Age at time of surgery.cNeoadjuvant chemotherapy (as indicated) followed by interval debulking surgery.bPositive second-look surgery.xHigh-dose chemotherapy as indicated with bone marrow transplant.





postchemotherapy samples and the primary tumors was done using the7,585-feature cDNA microarray. All cDNA microarrays were manufac-

tured at the National Cancer Institute. Amplified RNA (4 Ag) wasreverse transcribed and directly labeled using cyanine 5–conjugated

dUTP (tumor RNA) or cyanine 3–conjugated dUTP (pooled referenceRNA). Hybridization was done in the presence of 5� SSC and 25%

formamide for 14 to 16 hours at 42jC. Slides were washed, dried, and

scanned using an Axon Instruments 4000a laser scanner. A detailedprotocol for RNA amplification as well as cDNA probe labeling and

hybridization is available at http://nciarray.nci.nih.gov/reference/(under ‘‘Alternative Methods and Protocols’’). Genepix software

(Molecular Devices, Sunnyvale, CA) was used to analyze the raw data

that were then uploaded to a relational database maintained by theCenter for Information Technology at the NIH.

Data analysis. The log expression ratios for the spots on each arraywere normalized by subtracting the median log ratio for the same array.Data were filtered to exclude spots with size <25 Am, intensity less thantwice background or <500 units in both red and green channels, andany flagged or missing spots. In addition, any features found to bemissing or flagged in >10% of the arrays were not included in theanalysis. The Genepix software assigns intensity levels in arbitrary unitswith a range between 0 and 65,535. For reference, typical medianbackground on arrays used in this investigation were 250, and themedian probe signal was f4,500 (after background subtraction).Statistical comparison between tumors groups was done using the ‘‘BRBArray Tools’’ software (http://linus.nci.nih.gov/BRB-ArrayTools.html),consisting of a modified F test with P < 0.001 considered significant.This stringent P was selected in lieu of the Bonferroni correction formultiple comparisons, which was deemed excessively restrictive. Genelists were interrogated using EASE software (15) to identify possibleoverrepresentation of genes belonging to the same biological orfunctional class.

Class prediction was done using a compound covariate predictor tool

available as part of the BRB Array Tools software. This tool creates a

multivariate predictor for one of two classes to each sample. The genes

included in the multivariate predictor were those that were univariatelysignificant at the selected significance cutoff of P < 0.0001. The

multivariate predictor is a weighed linear combination of log ratios for

genes that are univariately significant. The weight consists of the

univariate t statistics for comparing the classes. A leave-one-outapproach was then employed to test the ability of the compound

covariate predictor to assign chemosensitive or chemoresistant class to

each sample. A permutation test was used to assess the significance of

our cross-validated error rate. The random permutations test the null

hypothesis that there are no systematic differences in gene expressionprofiles of the chemoresistant and chemosensitive tumors. This

assumption can be tested by randomly permuting labels among the

gene expression profiles and determining what proportion of the

permuted data sets have a misclassification error rate less than or equal

to the observed error. This rate serves as the achieved significance levelin a test against the null hypothesis. Detailed information about the

compound covariate predictor and the permutation test for significance

is provided by the Biometric Research Branch, National Cancer Institute

and is available at http://www.healthsystem.virginia.edu/internet/obgyn/supplemental-figure.pdf.

Immunohistochemical analyses. Immunohistochemical staining wasdone on 5 Am sections of formalin-fixed, paraffin-embedded ovariantumor specimens. After deparaffinization, sections were pretreated withsteam for 20 minutes in citrate buffer (pH 6.0). Slides were stained withprimary antibodies against Ki-67 (mouse monoclonal clone MIB-1;1:75 dilution; DAKOCytomation, Carpinteria, CA), proliferating cellnuclear antigen (PCNA; mouse monoclonal clone PC10; ready-to-use;DAKOCytomation), and cathepsin D (CTSD; rabbit polyclonal; ready-to-use; DAKOCytomation). Staining was done on a DAKO Autostainerusing the LSAB2 kit (DAKOCytomation) consisting of biotinylatedanti-mouse and anti-rabbit ready-to-use secondary antibodies, strepta-vidin-horseradish peroxidase, and the chromogen diaminobenzidine.

The slides were counterstained with methyl green, dehydrated, andmounted. Quantitative scoring was done as the product of two stainingcharacteristics, the percentage of immunopositive cells and theintensity of the staining. Slides were examined microscopically at20� power, and five separate areas of each tumor specimen wereexamined, with f100 cells per area analyzed. Scoring for the numberof immunopositive cells was accomplished by assigning 0% to 25% as1, 26% to 50% as 2, 51% to 75% as 3, and 76% to 100% as 4. Intensitywas scored as 1 to 3. The final score consisted of the product of theimmunopositive score (averaged over the five areas) and the intensityscore (averaged over the five areas). Statistical analysis was accom-plished using Student’s t test.

Results

Gene expression differences between primary chemosensitiveand chemoresistant tumors. The clinicopathologic features ofthe samples used in this investigation are presented in Table 1(primary cancers) and Table 2 (postchemotherapy cancers).Tumors in both chemosensitive and chemoresistant groupswere predominantly advanced stage, high grade (grade 2 or 3),and of serous histology. All patients were treated with platinum-based chemotherapy. For the purposes of this investigation, allchemoresistant tumors were from patients with persistent orrecurrent disease present within 6 months of the initiation offirst-line platinum-based combination chemotherapy. Themedian disease-free interval for the chemosensitive sampleswas 21 months (range, 13-41 months). The median age for thechemoresistant group was 61 (range, 35-77), and the medianage for the chemosensitive group was 57 (range, 44-78).

The comparison of gene expression profiles of 21 primarychemosensitive and 24 primary chemoresistant ovarian cancersrevealed 85 transcripts with expression levels significantlydifferent between the two groups (P < 0.001; Table 3). Thedifference in geometric mean expression levels for all of thesetranscripts was V2-fold. The ability of the nine most significantlydifferentially expressed genes (P < 0.0001) to predict clinicalresponse was tested using a leave-one-out prediction model. Thisanalysis revealed an accuracy of 77.8% in correctly classifyingrefractory and responsive tumors. After 5,000 random permuta-tions, the likelihood of these nine genes differentiating the twogroups with equal or higher accuracy by chance (i.e., the nullhypothesis) was calculated as P = 0.018. These data showed that,although statistically significant differences at the mRNA levelexisted between chemosensitive and chemoresistant primarytumors, the magnitude of these differences was modest inprimary tumor samples obtained before chemotherapy.

To determine whether the inclusion of a few tumors ofnonserous histologies (i.e., clear cell or endometrioid) orcomparison of tumors that did not all receive taxane-containingchemotherapy regimens biased the results, an unsupervisedclustering analysis was also done. The resulting dendrogramrevealed no segregation based on tumor histology, tumor grade,or type of platinum-based chemotherapy received. Thesesupplementary data (Fig. S1) may be viewed at http://www.healthsystem.virginia.edu/internet/obgyn/documents/Supplemental-figures.pdf.

Semiquantitative immunohistochemical analyses were usedto determine if the gene expression differences between primarychemoresistant and chemosensitive ovarian cancers wereassociated with significant protein expression differences. Theprotein product of the CTSD gene was selected for this purpose





because of the availability of a commercial antibody and priorstudies implicating this protease in the pathogenesis of severalcancers, including ovarian and breast (16, 17). Using asemiquantitative immunohistochemical scoring system, che-mosensitive samples displayed significantly higher CTSDprotein expression than the chemoresistant samples (Fig. 1A).Because CTSD expression correlates with high proliferationstates in several tumors (17–19), the expression of Ki-67 andPCNA was also analyzed in this set of tumors usingsemiquantitative immunohistochemistry (Fig. 1B and C). Both

markers showed significantly higher expression in the chemo-sensitive primary tumors. These results suggest that the higherexpression of CTSD in the chemosensitive tumors correlateswith a higher proliferative state that may in turn render themmore sensitive to cytotoxic chemotherapy. Notably, CTSDexpression was also significantly higher in the primary chemo-sensitive group compared with the postchemotherapy group(two-tailed t test; P = 0.0008), consistent with the hypothesisthat the postchemotherapy tumors are enriched for chemo-resistant clones.

Table 3. Genes differentially expressedbetweenprimary chemoresistant andprimary chemosensitive ovariancancers(P < 0.001)

IMAGE clone* UniGene Gene Description Fold differencec P

366971 Hs.156346 TOP2A Topoisomerase (DNA) IIa170-kDa 1.81 0.0008810263 Hs.335798 RHPN2 Rhophilin, RhoGTPase-binding protein 2 1.70 0.00006435303 Hs.397426 KIAA4146 KIAA1416 protein 1.58 0.0006753464 Hs.33540 LOC389677 Similar to RIKEN cDNA 3000004N20 1.57 0.00004742581 Hs.42173 C6orf107 Chromosome 6 open reading frame107 1.57 0.00002450653 Hs.84063 BCL2L11 BCL2-like11 (apoptosis facilitator) 1.57 0.0009129345 Hs.173946 PAPD1 PAP-associated domain containing1 1.57 0.00004589967 Hs.301431 ZNF71 Zinc finger protein 71 (Cos26) 1.56 0.0002725223 Hs.119563 PSME4 Proteasome (prosome, macropain) activator subunit 4 1.55 0.0008452963 Hs.321390 CUGBP1 CUG triplet repeat, RNA-binding protein1 1.54 0.000371622 Hs.496511 PRKCI Protein kinase C, i 1.52 0.0007757383 Hs.59236 ANKRD27 Ankyrin repeat domain 27 (VPS9 domain) 1.52 0.0001220658 Hs.440394 MSH2 MutShomologue 2, colon cancer, nonpolyposis type1 (E. coli) 1.49 0.00031470530 Hs.435788 NCOA6 Nuclear receptor coactivator 6 1.47 0.0004261567 Hs.25812 NBS1 Nijmegen breakage syndrome1 (nibrin) 1.47 0.00006592928 Hs.173001 KIAA1221 KIAA1221protein 1.47 0.00003839048 Hs.156682 IGSF4 Immunoglobulin superfamily, member 4 1.47 0.0003773203 Hs.43627 SOX12 SRY (sex-determining regionY) box12 1.42 0.0009452423 Hs.326392 SOS1 Son of sevenless homologue1 (Drosophila) 1.42 0.00006306568 Hs.415997 COL6A1 Collagen, typeVI, a1 1.42 0.00002855563 Hs.306251 ERBB3 v-erb-b2 erythroblastic leukemia viral oncogene homologue 3 1.42 0.0009128159 Hs.170472 TPR Translocated promoter region (to activatedMEToncogene) 1.41 0.0002343490 Hs.381189 CBX3 Chromobox homologue 3 (HP1chomologue, Drosophila) 1.41 0.000451737 Hs.437224 RBBP8 Retinoblastoma-binding protein 8 1.38 0.00003247240 Hs.8258 C19orf13 Chromosome19 open reading frame13 1.38 0.0005950667 Hs.36761 HRASLS HRAS-like suppressor 1.37 0.0001264117 Hs.343475 CTSD Cathepsin D (lysosomal aspartyl protease) 1.35 0.0006741790 Hs.7942 AFTIPHILIN Aftiphilin protein 1.34 0.0004490414 Hs.386198 EML4 Echinodermmicrotubule-associated protein-like 4 1.34 0.001230235 Hs.215766 GTPBP4 GTP-binding protein 4 1.34 0.0001853066 Hs.5719 CNAP1 Chromosome condensation-related SMC-associated protein1 1.33 0.0002950092 Hs.405144 SFRS3 Splicing factor, arginine/serine ^ rich 3 1.32 0.0002859857 Hs.442787 ZNF148 Zinc finger protein148 (pHZ-52) 1.32 0.0007447569 Hs.282901 RNPC2 RNA-binding region (RNP1, RRM) containing 2 1.32 0.0001344942 Hs.109299 PPFIA3 Protein tyrosine phosphatase, interacting protein (liprin), a3 1.32 0.0009767419 Hs.274351 ZDHHC9 Zinc finger, DHHC domain-containing 9 1.32 0.0006882434 Hs.362996 KIAA0779 KIAA0779 protein 1.31 0.00005283329 Hs.252451 SEMA3A Semaphorin 3A 1.31 0.0007359184 Hs.439523 PRKR Protein kinase, IFN-inducible 1.30 0.0006180156 Hs.123654 PCF11 Pre-mRNA cleavage complex II protein Pcf11 1.30 0.0002180156 Hs.128959 EST 1.30 0.0002

(Continued on the following page)





Gene expression differences between primary and postchemo-therapy tumors. The overview of these comparisons is depictedin Fig. 2A. This analysis revealed that 759 genes differentiatedthe postchemotherapy samples from the primary chemo-sensitive tumors (P < 0.001). In contrast, only 229 genes weredifferentially expressed between postchemotherapy and primarychemoresistant groups (P < 0.001), suggesting smaller differ-ences between the molecular profiles of the latter two groups. Acomparison of the two gene lists revealed 178 genes that werecommon to both and thus represented those genes that were

differentially expressed between postchemotherapy samplesand primary tumors irrespective of intrinsic chemoresistance.The magnitude and direction of change in the expression levelsof these 178 genes were very similar in the two comparisons(postchemotherapy versus primary chemoresistant and post-chemotherapy versus primary chemosensitive; Fig. 2B). Apartial list of these genes and their expression levels in primarytumors and postchemotherapy samples is presented in Fig. 3A.Notably, genes encoding several oxidizing enzymes, includingADH1B, ADH1C , and ALDH2 , showed higher expression in the


740707 Hs.269902 KIAA0494 KIAA0494 protein 1.28 0.0006

884892 Hs.7838 MKRN1 Makorin, ring finger protein,1 1.27 0.00002

826135 Hs.367811 STK38 Serine/threonine kinase 38 1.26 0.0007

838359 Hs.77890 GUCY1B3 Guanylate cyclase1, soluble, b3 1.26 0.00008

878174 Hs.388164 FADS2 Fatty acid desaturase 2 1.26 0.0006

837864 Hs.94262 DD5 Progestin-induced protein 1.26 0.0008753914 Hs.512235 ITPR2 Inositol1,4,5-triphosphate receptor, type 2 1.26 0.0006825197 Hs.131168 SEP1 Strand-exchange protein1 1.25 0.0006815772 Hs.369284 C20orf6 Chromosome 20 open reading frame 6 1.24 0.0009754654 Hs.118964 p66a P66 a 1.23 0.0002384018 Hs.411300 WBP4 WWdomain-binding protein 4 (formin-binding protein 21) 1.23 0.00002320834 Hs.386404 UBE4B Ubiquitination factor E4B (UFD2 homologue, yeast) 1.22 0.000773596 Hs.35086 USP1 Ubiquitin-specific protease1 1.19 0.0003743727 Hs.389638 FLJ41501 Clone BRTHA2006975 0.86 0.0006277134 Hs.93836 CIP98 CASK-interacting protein CIP98 0.85 0.0009433289 Hs.117331 TREML1 Triggering receptor expressed onmyeloid cells-like1 0.84 0.0009811999 Hs.6455 RUVBL2 RuvB-like 2 (E. coli) 0.83 0.00031292893 Hs.125785 LOC149018 Hypothetical LOC149018, mRNA 0.83 0.0009813735 Hs.410314 PCDH16 Protocadherin16 dachsous-like (Drosophila) 0.81 0.0006814271 Hs.18885 CGI-116 CGI-116 protein 0.79 0.0004399563 Hs.120332 EST 0.78 0.0004814129 Hs.283716 MSCP Mitochondrial solute carrier protein 0.77 0.00002490484 Hs.356349 ZNF145 Zinc finger protein145 (Kruppel-like, expressed in PML) 0.77 0.000747043 Hs.154138 CHI3L2 Chitinase 3^ like 2 0.73 0.001470035 Hs.14060 PROK1 Prokineticin1 0.71 0.00072571195 Hs.406683 RPS15 Ribosomal protein S15 0.71 0.0003303109 Hs.123464 P2RY5 Purinergic receptor P2Y, G-protein coupled, 5 0.70 0.0005431231 Hs.381870 EFEMP2 Epidermal growth factor ^ containing fibulin-like ECMprotein 2 0.70 0.0006280950 Hs.422340 SRI Sorcin 0.66 0.0001453183 Hs.301302 SCAM-1 Vinexin b (SH3-containing adaptor molecule-1) 0.66 0.0004490556 Hs.26518 TM4SF7 Transmembrane 4 superfamily member 7 0.63 0.0001502499 Hs.75835 PMM1 Phosphomannomutase1 0.61 0.0001344720 Hs.81994 GYPC Glycophorin C (Gerbich blood group) 0.60 0.00032559389 Hs.150833 C4A Complement component 4A 0.58 0.00082784073 Hs.322431 NEUROD2 Neurogenic differentiation 2 0.54 0.0005898305 Hs.439671 NBL1 Neuroblastoma, suppression of tumorigenicity1 0.54 0.00002

756372 Hs.37682 RARRES2 Retinoic acid receptor responder (tazarotene-induced) 2 0.52 0.0003

1474174 Hs.367877 MMP2 Matrix metalloproteinase-2 (72-kDa type IVcollagenase) 0.50 0.00004

462939 Hs.526933 EST 0.48 0.001

823851 Hs.469463 AEBP1 AE-binding protein1 0.41 0.0007

*IMAGE IntegratedMolecularAnalysis of Genomes and their Expression Consortium clone (http://madb.nci.nih.gov/CR_query.shtml).cFold difference in geometric means of chemosensitive tumors (numerator) compared with chemoresistant tumors (denominator).

Table 3. Genes differentially expressedbetweenprimary chemoresistant andprimary chemosensitive ovarian cancers(P < 0.001) (Cont’d)






postchemotherapy samples. Also showing higher expression inpostchemotherapy samples were DOC1, CAV1, DUSP1,ITM2A, DCN , and KLF4 , all of which have been reported tobe down-regulated in ovarian cancer when compared withnormal ovary (20–22). In contrast, TOP1, TOP2A , and ZWINThave been reported to be overexpressed in ovarian cancer (23,24) but were lower expressed in the postchemotherapy samplescompared with the primary tumors.

Only 51 genes uniquely differentiated the primary chemo-resistant samples from the postchemotherapy group. Incontrast, 581 genes uniquely differentiated the primary chemo-sensitive samples from the postchemotherapy group (Fig. 2A).In addition, the latter comparison resulted in higher magnitudedifferences in expression as indicated by 51 of the 581 genesshowing z2-fold difference in their geometric mean expression.The comparison of the primary chemoresistant group with thepostchemotherapy samples revealed 13 genes that showed atleast a 1.5-fold difference in mean expression levels (Fig. 3D).Furthermore, only one gene, SPP1 (also known as osteopon-tin), had a >2-fold change in expression and was higherexpressed in the primary chemoresistant group. This gene hasbeen implicated previously in ovarian cancer (25) and has beenproposed to represent a diagnostic biomarker for ovariancancer (26). Interestingly, the postchemotherapy samplesexpressed higher levels of CDKN1C (p57KIP2) and ADAMTS1,both of which have been shown to function as negativeregulators of proliferation (27, 28). However, these levels werestill far lower than those observed in normal postmenopausalovarian samples (Fig. 3D).

In the comparison of the postchemotherapy tumors with theprimary chemosensitive samples, 41 genes showed at least2-fold higher expression and 10 genes showed at least a 2-foldlower expression in the postchemotherapy tumors (Fig. 3B

and 3C). When the expression levels of these 51 differentiallyexpressed genes in normal postmenopausal ovary were graphedalong with the tumor expression levels, an interesting andunexpected pattern emerged. The expression profile of thepostchemotherapy samples resembled that of normal ovariantissue to a much greater degree than that of the primarychemosensitive tumors. These data are consistent with a modelin which the postchemotherapy samples show a partial ‘‘returnto normal’’ or ‘‘low proliferative state’’ molecular expressionprofile with respect to this set of genes. One notable exceptionto the overall similarities between postchemotherapy andnormal ovarian gene expression pattern was in the expressionof CYR61 . This gene has been implicated in angiogenesis (29)and chemoresistance (30) and was expressed at a significantlyhigher level in postchemotherapy samples compared with bothprimary chemosensitive and normal ovarian samples (Fig. 3A).In addition, several of the genes higher expressed in thepostchemotherapy group were noted to be components of theextracellular matrix (ECM) or involved in its remodeling. Thisimpression was more formally investigated by using EASEsoftware (15) to analyze the biological categories within thisgene list. This analysis confirmed the statistically significant(P < 0.05 after Bonferroni correction) overrepresentation ofgenes involved in ECM among the genes differentiating thepostchemotherapy and primary chemosensitive tumors. Onehypothesis that may be derived from this observation is thatstromal-epithelial interactions or the ECM per se may beinvolved in acquired chemoresistance in ovarian cancer.

Discussion

These data suggest that gene expression patterns in primary(pretreatment) ovarian cancers can discriminate intrinsically

Fig. 1. Semiquantitativeimmunohistochemical analyses of CTSD,Ki-67, and PCNA expression.Representative photomicrographs ofprimary chemosensitive (A) and primarychemoresistant (B) ovarian cancersimmunostained for CTSDat�40magnification.Results of immunohistochemical scoring forCTSD (C), Ki-67 (D), and PCNA (E).Columns , mean score; bars , SD.





chemoresistant from chemosensitive tumors. An accurateassessment of predictors of response to chemotherapy is bestaccomplished through a prospective clinical trial involvingadequate numbers of patients. However, conducting such a trialrequires the identification of potential molecular predictors ofresponse in ovarian cancer. The data derived from screening>40,000 transcripts provide candidate targets for such anevaluation.

Although highly statistically significant, the magnitude of themean expression differences between chemosensitive andchemoresistant groups were modest for the discriminating geneset. The observed magnitude of expression differences is likely

to be an underestimation for three reasons that are notmutually exclusive. First, according to the Goldie-Coldmanhypothesis, chemoresistance is believed to result from a clonalselection process driven by the acquisition of drug resistancemutations (31). Thus, in primary tumors, only a small per-centage of the cells are likely to possess a chemoresistant phe-notype and the associated molecular changes. This results in a‘‘dilution’’ of the observed gene expression differences whensuch tumors are compared with chemosensitive cancers.Second, the comparison of gene expression profiles in primarytumors evaluates only intrinsic chemoresistance. It is likely thata substantial component of clinical chemoresistance is biolog-ically acquired and is therefore only manifested followingexposure to chemotherapeutic agents. In support of thishypothesis, the greatest differences in gene expression wereobserved between postchemotherapy samples and primarychemosensitive tumors. Finally, as is the case with most inves-tigations involving cDNA microarrays, the primary expressiondata are in the form of logarithmic intensity ratios. Secondarydata, such as average expression levels for genes within a group,are derived by calculating the geometric rather than the arith-metic means of logarithmic intensity ratios, resulting in smallervalues and smaller apparent differences.

The higher expression of CTSD in chemosensitive tumors, asdetermined by immunohistochemistry, shows that small differ-ences in geometric mean expression as determined by cDNAmicroarrays may be associated with substantially greater differ-ences in protein expression. Although one previous reportfailed to show prognostic value of CTSD expression in ovariancancer (32), most other studies show low CTSD expressionto be an adverse prognostic indicator in ovarian cancer(16, 33, 34). In view of our findings, the prognostic value ofCTSD may be related to its higher expression in intrinsicallychemosensitive tumors. Consistent with this hypothesis, CTSDhas been implicated in p53-depenedent apoptosis followingDNA damage induced by drugs and g-irradiation (35). Furtherinvestigations are needed to evaluate a possible causalrelationship between these observations and to better definethe relationship between CTSD and chemosensitivity. Thechemosensitive tumors also showed higher expression of theproliferative markers PCNA and Ki-67. In agreement with thisobservation, a previous morphologic study found a highlysignificant correlation between proliferation and mitoticindices and the presence of apoptotic bodies in primary ovariancancers (36). Thus, higher rate of proliferation may contributeto the chemosensitive nature of these tumors by predisposingthem to undergo apoptosis following chemotherapy.

This investigation also revealed substantially different geneexpression between primary ovarian cancers and tumorsamples obtained following chemotherapy. The postchemo-therapy tumors are difficult to classify, based on customaryclinical criteria, as either chemoresistant (cancer progression onchemotherapy or recurrence within 6 months of completingchemotherapy) or chemosensitive (disease-free interval of at

Fig. 2. Gene expression differences between postchemotherapy and primarychemosensitive or primary chemoresistant tumors.A, an overview of the number ofdifferentially expressed genes (P < 0.001), with the top circle representingpostchemotherapy versus primary chemosensitive samples, the bottom circlerepresenting postchemotherapy versus primary chemoresistant samples, and theoverlap region representing genes differentially expressed betweenpostchemotherapy samples and primary tumors irrespective of intrinsicchemosensitivity. B, changes in the magnitude and direction of the178 genes thatdiscriminated the postchemotherapy samples from all primary tumors (P < 0.001).For each gene, the fold difference between postchemotherapy and primarychemoresistant tumors (X axis) is plotted against the fold difference betweenpostchemotherapy and primary chemosensitive tumors (Yaxis).Values less than1.0reflect higher expression in the primary tumors. r, Pearson correlation coefficient.

Fig. 3. Specific genes with statistically significant differential expression between postchemotherapy and primary tumors (P < 0.001).The corresponding expression valuesfromnormal postmenopausal ovaries are shown for comparison. A, genes that were differentially expressed between postchemotherapy tumors and both groups of primarytumors.The top 50 genes are shown; the full list may be found in the Supplementary Data. B, genes that were differentially higher expressed byz2-fold (n = 41) in thepostchemotherapy comparedwith primary chemosensitive tumors. Genes encoding ECM-related proteins are shown in bold type. C, genes that were differentially higherexpressedbyz2-fold (n = 10) inprimary chemosensitive comparedwithpostchemotherapy tumors.D, genes differentially expressedbetween thepostchemotherapy tumorsand the primary chemoresistant tumors.

—!









least 12 months) as they fit neither clinical criterion. They alleventually became resistant to chemotherapy. Our assertion isthat these tumor samples represent a state of enrichment inchemoresistant clones, as these are tumors that have survivedthree to six cycles of chemotherapy. A reasonable hypothesis isthat the gene expression profile of these postchemotherapysamples is likely to include molecular changes associated withacquired chemoresistance, as these samples were obtainedwithin a few weeks of completing three to six cycles ofchemotherapy. Consistent with this hypothesis, fewer andsmaller magnitude gene expression differences were observedbetween postchemotherapy and primary chemoresistant sam-ples. However, the data also suggest that intrinsic and acquiredchemoresistance are likely to manifest through nonoverlappingmolecular pathways (Fig. 4). This was evident by the lack ofsignificant overlap (some genes are part of both lists) betweenthe gene list differentiating primary chemosensitive andchemoresistant groups and the list differentiating each groupfrom the postchemotherapy samples. In comparing the primarytumors with postchemotherapy samples, three separate lists ofdifferentially expressed genes were generated. Two lists identi-fied genes that uniquely differentiated primary chemoresistantand chemosensitive tumors from the postchemotherapysamples and one that included genes that discriminated thelatter from both former groups. All three lists contain genes thathave been implicated previously in tumorigenesis and providetargets for prospective investigations of acquired chemoresist-ance in ovarian cancers.

Finally, certain patterns of expression deserve special consi-deration. First, several ECM-related genes revealed differentialexpression between postchemotherapy and primary tumors,with all being higher expressed the postchemotherapy tumors.This functional category of genes was significantly overrepre-

sented in the comparison between postchemotherapy andprimary tumors even after rigorous statistical correction formultiple comparisons. Such a finding is consistent with arecent study of ovarian cancer chemoresistance in vitro, wherea group of several ECM components was found to be up-regulated in platinum-resistant cells (37). These genesincluded DCN and COL6A3 , two of the ECM-related genesthat were found to be higher expressed in the postchemo-therapy group in this study.

Another member of this group found to be overexpressed inpostchemotherapy tumors was SPARC (also known as osteo-nectin), a matricellular glycoprotein involved in angiogenesis,cell adhesion, and ECM turnover (38). This gene is also up-regulated following chemotherapy in breast cancer (39) andhas antiproliferative and tumor suppressor function in ovarian(40, 41) and breast (42) cancer cells. Furthermore, SPARCstimulates matrix metalloproteinase-2 expression in othertissues (43, 44) and may account for the observed higherexpression of matrix metalloproteinase-2 in the postchemo-therapy tumors. Another related ECM gene, SPARCL1 (alsoknown as hevin, SC1, and MAST-9), was significantly higherexpressed in the postchemotherapy samples compared withboth groups of primary tumors, is down-regulated in severalcancers, and has a negative effect on cell proliferation (45).

The higher expression of these and other antiproliferativegenes (e.g., KLF4 and CAV1 ; refs. 21, 46) in the postchemo-therapy samples, as well as similarities in the gene expressionprofiles of the postchemotherapy tumors and normal post-menopausal ovaries, support the concept that a decreasedproliferative state may be involved in the development ofacquired chemoresistance. Furthermore, given that the primarychemoresistant tumors exhibited significantly lower Ki-67,PCNA, and CTSD protein expression compared with thechemosensitive samples, decreased proliferation may also bea contributing feature to intrinsic chemoresistance. Senescent orslow-growing cells may be more tolerant of cytotoxic chemo-therapy, thus allowing more time for selection of advantageousmutations and development of resistant clones. Consistent withsuch a hypothesis, the majority of postchemotherapy samplesexhibit a decreased mitotic index compared with the preche-motherapy sample of the same tumor (47).

This investigation represents an initial effort to discoverpotentially important molecular mediators of intrinsic andacquired chemoresistance in ovarian cancer in vivo . It is criticalto further investigate the molecular basis for chemoresistanceprospectively in a large cohort with prechemotherapy andpostchemotherapy sampling from the same patients. Thisinvestigation provides preliminary data that may inform suchfuture studies and perhaps clinical protocols.

Acknowledgments

We thank Drs. Edison T. Liu and William J. Hoskins for their vital roles inobtaining the Director’s Challenge grant from the National Cancer Institute thatmade this work possible.

Fig. 4. Diagrammatic depiction of chemoresistance in primary andpostchemotherapy samples. Smaller differences in gene expression observedbetween primary chemosensitive and chemoresistant tumors are likely the result oflow relative abundance of intrinsically chemoresistant clones as predicted by theGoldie-Colemanhypothesis. Chemotherapy results in a reduction in the number ofchemosensitive and an enrichment of chemoresistant clones in thepostchemotherapy samples. In addition, chemotherapy is likely to induce additionalgenetic changes contributing to acquired chemoresistance.The combination ofthese effects is likely to be responsible for the robust differences in gene expressionobserved between primary and postchemotherapy samples.

References1. Chi DS, Sabbatini P. Advanced ovarian cancer. CurrTreat Options Oncol 2000;1:139^46.

2. Salom E, Almeida Z, Mirhashemi R. Management of

recurrent ovarian cancer: evidence-based decisions.Curr Opin Oncol 2002;14:519^27.

3. McGuireWP, Markman M. Primary ovarian cancer

chemotherapy: current standards of care. Br JCancer2003;89:S3^8.

4. Markman M, Hoskins W. Responses to salvage





chemotherapy in ovarian cancer: a critical need forprecise definitions of the treated population [editori-al]. J Clin Oncol 1992;10:513^4.

5. Alberts DS. Treatment of refractory and recurrentovarian cancer. Semin Oncol1999;26:8^14.

6. Spriggs D. Optimal sequencing in the treatment ofrecurrent ovarian cancer. Gynecol Oncol 2003;90:S39^44.

7. Johnson SW, Ozols RF, HamiltonTC. Mechanisms ofdrug resistance in ovarian cancer. Cancer 1993;71:644^9.

8. Aebi S, Kurdihaidar B, Gordon R, et al. Loss of DNAmismatch repair in acquired resistance to cisplatin.Cancer Res1996;56:3087^90.

9. Coukos G, Rubin SC. Chemotherapy resistance inovarian cancer: new molecular perspectives. ObstetGynecol1998;91:783^92.

10.Vasey PA. Resistance to chemotherapy in advancedovarian cancer: mechanisms and current strategies. BrJCancer 2003;89:S23^8.

11. Agarwal R, Kaye SB. Ovarian cancer: strategies forovercoming resistance to chemotherapy. Nat RevCancer 2003;3:502^16.

12. Skipper HE. Kinetics of mammary tumor cellgrowth and implications for therapy. Cancer 1971;28:1479^99.

13. Norton L, Simon R, Brereton HD, Bogden AE.Predicting the course of Gompertzian growth. Nature1976;264:542^5.

14.Van Gelder RN, von Zastrow ME,Yool A, DementWC, BarchasJD, EberwineJH. Amplified RNAsynthe-sized from limited quantities of heterogeneous cDNA.Proc Natl Acad Sci US A1990;87:1663^7.

15. Hosack DA, Dennis G, Sherman BT, Lane HC,Lempicki RA. Identifying biological themeswithin listsof genes with EASE. Genome Biol 2003;4:R70.

16. LoschA, SchindlM, Kohlberger P, et al. CathepsinDin ovarian cancer: prognostic value and correlationwithp53 expression andmicrovessel density.GynecolOncol 2004;92:545^52.

17. Ioachim E,Tsanou E, Briasoulis E, et al.Clinicopatho-logical study of the expression of hsp27, pS2, cathep-sin D and metallothionein in primary invasive breastcancer. Breast 2003;12:111^9.

18. Glondu M, Liaudet-Coopman E, Derocq D, Platet N,Rochefort H, Garcia M. Down-regulation of cathepsin-D expression by antisense gene transfer inhibits tumorgrowth and experimental lung metastasis of humanbreast cancer cells. Oncogene 2002;21:5127^34.

19. Ioachim E, Charchanti A, Stavropoulos N,Athanassiou E, Bafa M, Agnantis NJ. Expression ofcathepsin Dinurothelial carcinomaof the urinaryblad-der: an immunohistochemical study including correla-tions with extracellular matrix components, CD44,p53, Rb, c-erbB-2 and the proliferation indices. Anti-cancer Res 2002;22:3383^8.

20. Mok SC,Wong KK, Chan RK, et al. Molecular clon-ing of differentially expressed genes inhuman epitheli-al ovarian cancer. Gynecol Oncol1994;52:247^52.

21.Wiechen K, Diatchenko L, Agoulnik A, et al. Caveo-lin-1 is down-regulated in human ovarian carcinomaand acts as a candidate tumor suppressor gene. AmJPathol 2001;159:1635^43.

22. Manzano RG, Montuenga LM, Dayton M, et al.CL100 expression is down-regulated in advancedepithelial ovarian cancer and its re-expressiondecreases its malignant potential. Oncogene 2002;21:4435^7.

23. van der ZeeAG, Hollema H, de Jong S, et al. P-gly-coprotein expression and DNA topoisomerase I and IIactivity in benign tumors of the ovary and inmalignanttumors of the ovary, before and after platinum/cyclo-phosphamide chemotherapy. Cancer Res 1991;51:5915^20.

24. Jazaeri AA,Yee CJ, Sotiriou C, Brantley KR, BoydJ,Liu ET. Gene expression profiles of BRCA1-linked,BRCA2-linked, and sporadic ovarian cancers. J NatlCancer Inst 2002;94:990^1000.

25.Tiniakos DG,YuH, Liapis H. Osteopontin expressionin ovarian carcinomas and tumors of low malignantpotential (LMP). Hum Pathol1998;29:1250^4.

26. Kim JH, Skates SJ, UedeT, et al. Osteopontin as apotential diagnostic biomarker for ovarian cancer.JAmMed Assoc 2002;287:1671^9.

27. Zhang P, Liegeois NJ,Wong C, et al. Altered cell dif-ferentiation and proliferation in mice lacking p57KIP2indicates a role in Beckwith-Wiedemann syndrome.Nature1997;387:151^8.

28. LuqueA, Carpizo DR, Iruela-ArispeML. ADAMTS1/METH1inhibits endothelial cell proliferation by directbinding and sequestration of VEGF165. J Biol Chem2003;278:23656^65.

29. Babic AM, Kireeva ML, Kolesnikova TV, Lau LF.CYR61, a product of a growth factor-inducible im-mediate early gene, promotes angiogenesis and tu-mor growth. Proc Natl Acad Sci U S A 1998;95:6355^60.

30.Wittig R, NesslingM,Will RD, et al. Candidate genesfor cross-resistance against DNA-damaging drugs.Cancer Res 2002;62:6698^705.

31. Goldie JH, Coldman AJ. A mathematic model forrelating the drug sensitivity of tumors to their sponta-neous mutation rate. Cancer Treat Rep 1979;63:1727^33.

32. Ferrandina G, Scambia G, Fagotti A, et al. Immunor-adiometric and immunohistochemical analysis of ca-thepsin D in ovarian cancer: lack of association withclinical outcome. BrJCancer1998;78:1645^52.

33. Scambia G, Panici PB, Ferrandina G, et al. Clinicalsignificance of cathepsin D in primary ovarian cancer.EurJCancer1994;30A:935^40.

34. Baekelandt M, Holm R, Trope CG, Nesland JM,

Kristensen GB. The significance of metastasis-relatedfactors cathepsin-D and nm23 in advanced ovariancancer. Ann Oncol1999;10:1335^41.

35.Wu GS, Saftig P, Peters C, El-Deiry WS. Potentialrole for cathepsin D in p53-dependent tumor sup-pression and chemosensitivity. Oncogene 1998;16:2177^83.

36. Brustmann H. Apoptotic bodies as a morphologicalfeature in serous ovarian carcinoma: correlation withnuclear grade, Ki-67 and mitotic indices. Pathol ResPract 2002;198:85^90.

37. Sherman-Baust CA,Weeraratna AT, Rangel LB, et al.Remodelingof the extracellularmatrix throughoverex-pression of collagenVI contributes to cisplatin resis-tance in ovarian cancer cells. Cancer Cell 2003;3:377^86.

38. BradshawAD, Sage EH. SPARC, a matricellular pro-tein that functions in cellular differentiation and tissueresponse to injury. JClin Invest 2001;107:1047^54.

39. Korn EL, McShane LM,Troendle JF, Rosenwald A,Simon R. Identifying pre-post chemotherapy differen-ces in gene expression in breast tumours: a statisticalmethod appropriate for this aim. BrJCancer 2002;83:1093^6.

40. Mok SC, ChanWY,Wong KK, Muto MG, BerkowitzRS. SPARC, an extracellular matrix proteinwith tumor-suppressing activity in human ovarian epithelial cells.Oncogene1996;12:1895^901.

41.Yiu GK, Chan WY, Ng SW, et al. SPARC (secretedprotein acidic and rich in cysteine) induces apopto-sis in ovarian cancer cells. Am J Pathol 2001;159:609^22.

42. Dhanesuan N, Sharp JA, Blick T, Price JT,Thompson EW. Doxycycline-inducible expression ofSPARC/osteonectin/BM40 in MDA-MB-231humanbreast cancer cells results in growth inhibition. BreastCancer ResTreat 2002;75:73^85.

43. Gilles C, Bassuk JA, Pulyaeva H, Sage EH, FoidartJM,Thompson EW. SPARC/osteonectin induces ma-trix metalloproteinase 2 activation in human breastcancer cell lines. Cancer Res1998;58:5529^36.

44. FujitaT, Shiba H, Sakata M, UchidaY, Nakamura S,Kurihara H. SPARC stimulates the synthesis of OPG/OCIF,MMP-2 andDNA inhumanperiodontal ligamentcells. JOral Pathol Med 2002;31:345^52.

45. Claeskens A, Ongenae N, Neefs JM, et al. Hevin isdown-regulated inmanycancers and is anegative reg-ulator of cell growth and proliferation. Br J Cancer2000;82:1123^30.

46. Chen X,Whitney EM, Gao SY,YangVW. Transcrip-tional profiling of Kruppel-like factor 4 reveals afunction in cell cycle regulation and epithelial dif-ferentiation. JMol Biol 2003;326:665^77.

47. McCluggageWG, Lyness RW, Atkinson RJ, et al.Morphological effects of chemotherapy on ovariancarcinoma. JClin Pathol 2002;55:27^31.





2005;11:6300-6310. Clin Cancer Res Amir A. Jazaeri, Christopher S. Awtrey, Gadisetti V.R. Chandramouli, et al. Chemotherapy in Epithelial Ovarian CancersGene Expression Profiles Associated with Response to

Updated version

http://clincancerres.aacrjournals.org/content/11/17/6300

Access the most recent version of this article at:

Material

Supplementary

http://clincancerres.aacrjournals.org/content/suppl/2006/01/10/11.17.6300.DC1

Access the most recent supplemental material at:

Cited articles

http://clincancerres.aacrjournals.org/content/11/17/6300.full#ref-list-1

This article cites 44 articles, 9 of which you can access for free at:

Citing articles

http://clincancerres.aacrjournals.org/content/11/17/6300.full#related-urls

This article has been cited by 26 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. (CCC)Click on "Request Permissions" which will take you to the Copyright Clearance Center's

.http://clincancerres.aacrjournals.org/content/11/17/6300To request permission to re-use all or part of this article, use this link



http://clincancerres.aacrjournals.org/content/suppl/2006/01/10/11.17.6300.DC1

http://clincancerres.aacrjournals.org/content/11/17/6300.full#ref-list-1

http://clincancerres.aacrjournals.org/content/11/17/6300.full#related-urls

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

mailto:[email protected]



Documents

GeneExpressionProfilesAssociatedwithResponseto ...treatment and outcome of ovarian cancer. This topic has been the subject of intense research, and previous studies on chemoresistance