Microarrayanalysis of Endometrial Carcinomas and Mixed Mullerian

Imaging, Diagnosis, Prognosis

Microarray Analysis of Endometrial Carcinomas and Mixed Mullerian Tumors Reveals Distinct Gene Expression Profiles Associated with Different Histologic Types of Uterine Cancer G. Larry Maxwell,1, 2 G.V.R. Chandramouli,2 Lou Dainty,1TracyJ. Litzi,2 Andrew Berchuck,3 J. Carl Barrett,2 and John I. Risinger 2

Abstract Previous studies using cDNA microarray have indicated that distinct gene expression profiles characterize endometrioid and papillary serous carcinomas of the endometrium. Molecular studies have observed that mixed mullerian tumors, characterized by both carcinomatous and sarcomatous components, share features that are characteristic of endometrial carcinomas. The objective of this analysis was to more precisely define gene expression patterns that distinguish endometrioid and papillary serous histologies of endometrial carcinoma and mixed mullerian tumors of the uterus. One hundred nineteen pathologically confirmed uterine cancer samples were studied (66 endometrioid, 24 papillary serous, and 29 mixed mullerian tumors). Gene expressions were analyzed using the Affymetrix Human Genome Arrays U133A and U133BGenechip set. Unsu- pervised analysis revealed distinct global gene expression patterns of endometrioid, papillary serous, mixed mullerian tumors, and normal tissues as grossly separated clusters. Two-sample t tests comparing endometrioid and papillary serous, endometrioid and mixed mullerian tumor, and papillary serous and mixed mullerian tumor pairs identified 1,055, 5,212, and 1,208 differentially expressed genes at P < 0.001, respectively. These data revealed that distinct patterns of gene expression characterize various histologic types of uterine cancer. Gene expression profiles for select genes were confirmed using quantitative PCR. An understanding of the molecular heterogeneity of various histologic types of endometrial cancer has the potential to lead to better individualization of treatment in the future.

The American Cancer Society estimated that f40,100 new estrogenic milieu (3). These tumors display a high incidence of cases of cancer of the uterine corpus will be diagnosed during alterations in the PTEN tumor suppressor gene (4–6) as well as 2004 and f6,800 women are expected to die of their disease defects in mismatch repair that results in microsatellite during that time (1). The majority of endometrial cancers are instability (7, 8). In contrast, type II endometrial cancers are carcinomas, which may be characterized as type I or type II on more often poorly differentiated, at advanced stage at the time the basis of both clinical presentation as well as histopathologic of diagnosis, and are nonendometrioid in histology (3). These variables (2). Type I endometrial carcinomas are usually tumors rarely, if ever, contain PTEN mutations or microsatellite endometrioid in histology, well-differentiated, present with instability (9) but are more likely to be characterized by p53 early-stage disease, and are often associated with a hyper- mutation and widespread aneuploidy (10–12). Although the majority of uterine cancers are carcinomas that arise from the endometrial lining, f2% to 4% of uterine cancers are sarcomas that arise in the smooth muscle of the Authors’ Affiliations: 1Walter Reed Army Medical Center, Washington, DC; uterine wall (1). The majority of uterine sarcomas are classified 2Laboratory of Biosystems and Cancer, National Cancer Institute, Bethesda, as mixed mullerian tumors, which contain both carcinomatous Maryland; and 3Department of Obstetrics and Gynecology/Division of and sarcomatous elements. Chemotherapeutics for mixed Gynecologic Oncology, Duke University, Durham, North Carolina mullerian tumors have traditionally been similar to those Received 9/28/04; revised 1/25/05; accepted 2/2/05. Grant support: Department of Defense Peer Reviewed Medical Research effective in the treatment of other types of soft tissue sarcomas. Program, award no. DAMD17-02-1-0183. There is, however, molecular evidence [i.e., X-chromosome The costs of publication of this article were defrayed in part by the payment of page activation experiments (13, 14), allelotyping studies (15), and charges. This article must therefore be hereby marked advertisement in accordance mutation analysis (16)] to suggest that the carcinomatous with 18 U.S.C. Section 1734 solely to indicate this fact. component of mixed mullerian tumors is the cell type of origin Note: Presented at the 35th Annual Meeting of the Society of Gynecologic Oncologists, San Diego, 2004. The opinions or assertions contained herein are the and that the sarcomatous component is derived from the private views of the authors and are not to be construed as official or as reflecting the carcinoma through metaplastic transformation or from a stem views of the Department of the Army or the Department of Defense. cell that undergoes divergent differentiation (17, 18). The Requests for reprints: G. Larry Maxwell, Division of Gynecologic Oncology, association of mixed mullerian tumors with obesity, exogenous Walter Reed Army Medical Center, 6900 Georgia Avenue,Washington, DC 20307. Phone: 202-782-8512; Fax: 202-782-9278; E-mail: george.maxwell@ estrogen use, and tamoxifen suggests clinical similarities with na.amedd.army.mil. endometrioid endometrial carcinomas (19, 20). However, F 2005 American Association for Cancer Research. unlike most endometrioid carcinomas, mixed mullerian tumors

ClinCancerRes2005;11(11)June1,2005 4056 www.aacrjournals.org Downloaded from clincancerres.aacrjournals.org on September 25, 2021. © 2005 American Association for Cancer Research. Endometrial Cancer, Microarray

Fig. 1. A, unsupervised analysis using multidimensional scaling based on the overall gene expression in endometrioid (green) and papillary serous (blue)using 1-correlation as distance metric of 18.4 K transcripts detected in at least 50% of the arrays. B, differentially expressed genes between 66 endometrioid carcinomas and 24 papillary serous carcinomas.Twenty-five most up-regulated and 25 most down-regulated genes at P < 0.001.Each sample in the heat map is labeled histology, stage of disease, and coded tumor number.The heat map was color coded using red for up-regulation from normal endometria and green for down-regulation.

www.aacrjournals.org 4057 Clin Cancer Res 2005;11(11) June 1, 2005 Downloaded from clincancerres.aacrjournals.org on September 25, 2021. © 2005 American Association for Cancer Research. Imaging, Diagnosis, Prognosis are aggressive with a prognosis similar to papillary serous 4 of 136 arrays have abnormally high scatter that have correlation adenocarcinomas that are associated with poor outcome. coefficients smaller than 0.85. All the arrays having correlation Despite the clinical features that mixed mullerian tumors can coefficients <0.85 were excluded from further study. The genes were share with endometrial adenocarcinomas, little is known filtered from the remaining arrays using detection P value reported by MAS5. The genes having P > 0.065 in 95% of the arrays were regarding the molecular features that distinguish uterine eliminated and all other signals were included for statistical carcinomas from sarcomas. comparisons of classes. Our group has previously used cDNA microarray to examine For multidimensional scaling (computed by Partek Pro Discover the gene expression profiles of different histologic types of software build 5, Partek, Inc., St. Charles, MO), the genes included were endometrial adenocarcinoma (21). The results of our initial at P < 0.065 in at least 50% of the arrays. Statistical calculations were analysis suggested that the gene expression profile for endome- done using logarithmic values of normalized signals. trioid, clear cell, and papillary serous endometrial cancers are Binary class comparison was done on individual comparisons of distinct, and we identified several additional pathways impor- different histologic groups using BRB Array tools software (BRB tant in the development of endometrial cancer. We have Array tools ver. 3.0c, Richard Simon, Amy Peng, Biometric research hypothesized that the gene expression profiles of mixed branch, National Cancer Institute, NIH, http://linus.nci.nih.gov/BRB- ArrayTools.html). Differentially expressed genes were identified by mullerian tumors are also distinct from both common types parametric Student’s t tests on genes having at least 50% or more of uterine adenocarcinoma. The aim of this study is to present a present calls. In each of the comparisons, genes differentially expressed more comprehensive genomic analysis of uterine cancer to above 2-fold were clustered by the similarity of their expression profiles. better characterize the molecular expression profiles of different Hierarchical clustering was done on logarithmic values of expressions histologic types of uterine cancer. Elucidation of these using 1 À q as distance metric (16). The heat map was color-coded, molecular expression signatures may be useful in predicting using red for up-regulation from normal endometria and green for the clinical behavior of uterine cancers as well as identifying down-regulation. All the statistical calculations were done on the candidate cellular pathways that can be targets for future logarithmic values of signals to the base 2. therapeutics. Validation of gene expression using quantitative PCR. The expressions of genes chosen for validation were determined by multiplex PCR using TaqMan Gene Expression Assays purchased from Applied Materials and Methods Biosystems (Foster City, CA) with h-actin as reference. Samples were run on the ABI Prism 7700 Sequence Detection System according to Tissue specimens. Flash-frozen cancer specimens were obtained manufacturer’s suggested protocols. The relative quantitation, using from 119 patients undergoing surgery for uterine cancer at Duke the comparative CT method, was calculated for each sample. The University Medical Center. These included 66 endometrioid, 24 weighted average of the mean ratios of each histologic group was papillary serous, and 29 mixed mullerian tumors of low and high presented with the SE of mean values as error bars. grades. All tissues were collected under an Institutional Review Board–approved protocol at Duke University Medical Center. Speci- Results mens were harvested by pathologists using gross specimens within 30 minutes of specimen removal at the time of surgery. Each uterine Unsupervised analysis including all three histologies sug- tumor was then frozen until the time of the analysis. Tissue gested different global expression patterns associated with each specimens were evaluated by H&E to confirm that the specimen to of these groups. We subsequently chose to perform three be analyzed contained at least 50% or greater cancer cells. During preparation of the specimens for analysis, care was taken to separate comparisons (endometrioid versus papillary serous, macroscopically dissect the cancer away from any adjacent myome- mixed mullerian tumor versus papillary serous, and mixed trium. Tissue samples were subjected to RNA isolation using TRIzol mullerian tumor versus endometrioid) to better discriminate followed by an additional level of purification with the RNeasy kit differences in gene expression patterns between histologic types (Qiagen, Valencia, CA). RNA was successfully extracted from each of of uterine cancer. the cancer specimens and 10 of the 15 normal endometrium samples. Endometrioid versus papillary serous carcinoma. Multidimen- The integrity of each of the RNA samples was confirmed using ensional scaling on all the genes having 50% present calls denaturing gel electrophoresis (22). suggested that the gene expression of endometrioid and Gene expression analysis. The gene expressions were assessed using papillary serous carcinomas were different, further supporting the Affymetrix human genome U133A and B Genechips (45,000 gene the paradigm that these two types of endometrial cancer transcripts covering 28,473 UniGene clusters). Approximately 5 Ag total RNA from each sample were labeled using high yield transcript develop in part via different pathways (Fig. 1A). In a supervised labeling kit (Enzo Life Sciences Inc., Farmingdale, NY) and labeled comparison of 66 endometrioid and 24 papillary serous RNAs were hybridized, washed, and scanned according to manu- carcinomas, 1,055 genes were found to be differentially facturer’s specifications (Affymetrix, Inc., Santa Clara, CA). Affymetrix expressed at F test P < 0.001, of which 151 of genes had at Microarray Suite 5.0 software (MAS5) was used to estimate transcript least at 2-fold change. The tumor to normal expression ratios of signal levels from scanned images (Affymetrix) by one-step Tukey’s 25 most up-regulated and 25 most down-regulated genes are biweight algorithm. The probe annotations of HG-U133 chips and shown as heat map in Fig. 1B. Examples of genes that were MAS5 statistical algorithms are available at Affymetrix website (http:// notably associated with a >2-fold papillary serous/endome- www.affymetrix.com). The signals on each array were normalized to trioid expression ratio included IGF2, PTGS1 (COX1), and a trimmed mean value of 500, excluding lowest 2% and highest 2% p16, whereas genes with a >2-fold endometrioid/papillary of the signals. An Affymetrix probe set representing a unique Genbank sequence is referred as a probe or gene hereafter for serous carcinoma expression ratio included TFF3, FOXA2, and convenience. To verify any errors in the expressions caused by image MSX2. defects, the correlation coefficient of each array to an idealized Mixed mullerian tumor versus endometrioid carcinoma. distribution was determined where the idealized distribution is Grossly separated clusters of global expression profiles were mean of all arrays. Visual inspection of scatter plots revealed that noted in the unsupervised comparison of mixed mullerian

Clin Cancer Res 2005;11(11)June1,2005 4058 www.aacrjournals.org Downloaded from clincancerres.aacrjournals.org on September 25, 2021. © 2005 American Association for Cancer Research. Endometrial Cancer, Microarray

tumor and endometrioid adenocarcinoma (Fig. 2A). To 66 endometrioid carcinomas, which revealed 5,212 genes at identify transcripts that are responsible for this delineation, F test P < 0.001, including 1,132 genes that were differentially we did supervised analysis of 29 mixed mullerian tumors and expressed by at least 2-fold and 122 genes that were

Fig. 2. A, unsupervised analysis using multidimensional scaling based on the overall gene expression in endometrioid carcinomas (red) and mixed mullerian tumor (blue). B, genes differentially expressed between 66 endometrioid carcinomas (E) and 29 mixed mullerian tumors (MMT) of the uterus.Twenty-five most up-regulated and 25 most down-regulated genes at P < 0.001.Each sample in the heat map is labeled histology, stage of disease, and coded tumor number.The heat map was color-coded using red for up-regulation from normal endometria and green for down-regulation.

www.aacrjournals.org 4059 Clin Cancer Res 2005;11(11) June 1, 2005 Downloaded from clincancerres.aacrjournals.org on September 25, 2021. © 2005 American Association for Cancer Research. Imaging, Diagnosis, Prognosis

Clin Cancer Res 2005;11(11) June 1, 2005 4060 www.aacrjournals.org Downloaded from clincancerres.aacrjournals.org on September 25, 2021. © 2005 American Association for Cancer Research. Endometrial Cancer, Microarray p------Fig. 3. A, unsupervised analysis using multidimensional scaling based on the overall gene expression in mixed mullerian tumor (red) and papillary serous carcinoma (blue). B, list of 25 highest and 25 lowest differentially expressed genes (at least 2-fold) for 29 mixed mullerian tumors and 24 papillary serous (PS)carcinomas (P < 0.001).Each sample in the heat map is labeled histology, stage of disease, and coded tumor number.The heat map was color coded using red for up-regulation from normal endometria and green for down-regulation. differentially expressed by at least 5-fold. The tumor to normal serous; endometrioid versus mixed mullerian tumor, and expression ratios of 25 most up-regulated and 25 most down- mixed mullerian tumor versus papillary serous. Using this regulated genes are shown as heat map in Fig. 2B. Greater approach, we identified 66 genes associated with endome- expression of IGF2 and lower expression of MUC1, SCGB2A1, trioid cancers present on both the endometrioid versus HOXB6, and TFF3 was observed in mixed mullerian tumor papillary serous carcinoma and endometrioid versus mixed specimens when compared with endometrioid carcinomas mullerian tumor list and consistent with a endometrioid (Fig. 2B). To further examine the differences between mixed carcinoma profile (Table 1); 21 genes associated with papil- mullerian tumors and endometrioid carcinomas, we examined lary serous cancers that were present on both the papillary the global expressions using several class prediction modeling serous versus endometrioid carcinoma list and the papillary programs. serous carcinoma versus mixed mullerian tumor list and as- Mixed mullerian tumor versus and papillary serous carcinoma. sociated with a papillary serous carcinoma profile (Table 2); Unsupervised analysis using multidimensional scaling showed and 361 genes associated with mixed mullerian tumors that that the genomic expression profiles of mixed mullerian tumors were present on both the mixed mullerian tumor versus and papillary serous carcinomas also clustered according to endometrioid carcinoma and the mixed mullerian tumor histologic type (Fig. 3A). Supervised analysis of 29 mixed versus papillary serous carcinoma lists and reflective of a mullerian tumors and 24 papillary serous tumors revealed mixed mullerian tumor profile (data not shown). The com- 1,208 genes at F test P < 0.001, of which 509 genes were plete cross-referenced lists of differentially expressed genes differentially expressed by at least 2-fold. The heat map of (P < 0.001) for each of the three comparisons are posted tumor to normal expression ratios of 25 most up-regulated on http://home.ccr.cancer.gov/lbc/risinger/pubs/hist2.asp and 25 most down-regulated genes is shown in Fig. 3B. and may provide gene candidates that seem to be associated Quantitative PCR analysis. We further evaluated the with a specific type of uterine cancer. Expression Analysis expression of six of these genes (MUC1, SCGB2A1, HOXB6, Systematic Explorer software was used to provide an PTGS1, TFF3, and IGF2) in each of the histologic groups analysis of gene ontology among the cross-referenced gene using real-time quantitative PCR to validate the results list for each of the three profiles: endometrioid, papillary obtained from the array analysis. Each of the six genes serous, and mixed mullerian tumor. The complete data revealed patterns of statistically significant differences in gene reflective of this analysis are also provided electronically expression that were consistent with the microarray analysis. (http://home.ccr.cancer.gov/lbc/risinger/pubs/hist2.asp) and Several of these genes were differentially expressed in two of further support the paradigm that gene expression associat- three comparisons and, therefore, have expression profiles ed with different types of uterine cancer seems to be reflective of all three groups in these instances (Fig. 4). unique. Cross-referencing of gene lists. In a final effort to find genes that are associated with a specific histologic variant of uterine Discussion cancer, we identified genes that were differentially expressed at least 2-fold (P < 0.001) on at least two of three gene lists for Previous data from our group (21) have suggested that the the following comparisons: endometrioid versus papillary gene expression patterns associated with different epithelial

Fig. 4. Microarray expression and quantitative PCR (TaqMan) expression analysis of six selected genes differentially expressed between mixed mullerian tumors, endometrial carcinoma, and papillary serous carcinoma.

www.aacrjournals.org4061 Clin Cancer Res 2005;11(11) June 1, 2005 Downloaded from clincancerres.aacrjournals.org on September 25, 2021. © 2005 American Association for Cancer Research. Imaging, Diagnosis, Prognosis

Ta b l e 1. Differentially expressed genes characteristic of endometrioid serous carcinomas

E/PS E/MMT UniGene Gene Map Description 8.027 19.501 Hs.104696 KIAA1324 1p13.3-p13.2 Maba1 6.279 5.010 Hs.420036 GAD1 2q31 Glutamate decarboxylase1 (brain, 67 kDa) 5.819 19.348 Hs.115838 16p13.11 Homo sapiens cDNA FLJ44282 fis, cloneTRACH2003516 5.745 9.057 Hs.82961 TFF3 21q22.3 Trefoil factor 3 (intestinal) 5.603 16.175 Hs.226391 AGR2 7p21.3 Anterior gradient 2 homologue (Xenopus laevis) 4.989 8.530 Hs.155651 FOXA2 20p11 Forkhead box A2 4.627 8.343 Hs.48403 FLJ10847 17p11.2 Hypothetical protein FLJ10847 4.204 4.452 Hs.46452 SCGB2A2 11q13 Secretoglobin, family 2A, member 2 3.807 8.202 Hs.24879 PPAP2C 19p13 Phosphatidic acid phosphatase type 2C 3.784 8.290 Hs.145807 TMC5 16p13.11 Transmembrane channel ^ like 5 3.692 5.524 Hs.104696 KIAA1324 1p13.3-p13.2 Maba1 3.617 3.865 Hs.99348 DLX5 7q22 Distal-less homeo box 5 3.406 7.984 Hs.200286 DKFZp547I048 1p31.1 Hypothetical protein DKFZp547I048 3.223 6.239 Hs.6168 KIAA0703 16q24.1 KIAA0703 gene product 3.185 4.521 Hs.158857 RASSF6 4q21.21 Ras association (RalGDS/AF-6) domain family 6 3.172 4.746 Hs.445072 ARGBP2 4q35.1 Arg/Abl-interacting protein ArgBP2 3.107 3.655 Hs.89404 MSX2 5q34-q35 Msh homeo box homologue 2 (Drosophila) 2.633 4.250 Hs.512682 CEACAM1 19q13.2 Carcinoembryonic antigen ^ related cell adhesion molecule 1 2.589 3.198 Hs.288240 IL20RA 6q22.33-q23.1 Interleukin 20 receptor 2.572 8.822 Hs.116992 HGD 3q21-q23 Homogentisate1,2-dioxygenase (homogentisate oxidase) 2.547 2.755 Hs.211587 PLA2G4A 1q25 PhospholipaseA2, group IVA (cytosolic, calcium-dependent) 2.487 2.187 Hs.306278 CD44 11p13 CD44 antigen (homing function and Indian blood group system) 2.470 2.941 Hs.156880 4 Homo sapiens , clone IMAGE:4791597, mRNA 2.458 3.778 Hs.519137 4 Homo sapiens cDNA FLJ37284 fis, clone BRAMY2013590 2.456 4.519 Hs.432615 Homo sapiens transcribed sequence 2.428 2.310 Hs.11713 ELF5 11p13 - p12 E74-like factor 5 (Ets domain transcription factor) 2.420 3.915 Hs.145807 TMC5 16p13.11 Transmembrane channel ^ like 5 2.383 2.065 Hs.348802 NMA 10p12.3-p11.2 Putative transmembrane protein 2.265 2.581 Hs.306278 CD44 11p13 CD44 antigen (homing function and Indian blood group system) 2.218 8.166 Hs.433197 ASRGL1 11q12.3 Asparaginase like1 2.215 2.271 Hs.153952 NT5E 6q14-q21 5V-Nucleotidase, ecto (CD73) 2.184 2.901 Hs.79414 PDEF 6p21.3 Prostate epithelium ^ specific Ets transcription factor 2.124 3.088 Hs.439760 CYP4X1 1p33 Likely orthologue of rat cytochrome P450 4X1 2.121 2.453 Hs.293685 Homo sapiens transcribed sequence

types of uterine cancer are distinct. Only one other group has number and the Affymetrix platform that enabled us to previously reported the details of expression profiles among evaluate f45,000 gene transcripts covering 28,473 UniGene different histologic types of endometrial cancer (23). In this clusters in contrast to 9,726 clones corresponding to 6,386 cDNA microarray analysis, Moreno-Bueno et al. (23) different genes used by the other investigators. Although we identified only 66 genes that were differentially expressed did obtain similar results for several of the transcripts by at least 2-fold (P V 0.05) between endometrioid and previously reported (i.e., BUB1, CCNB2, MYC; ref. 23), we nonendometrioid cancers. In our current microarray analysis, did not find that STK15 was significantly overexpressed we detected 160 genes that were differentially expressed (at least 2-fold increased expression at P < 0.001) among the among endometrioid versus papillary serous cancers despite papillary serous carcinomas when compared with the more stringent statistical criteria (P < 0.001). Our more endometrioid tumors (23). Our different results may be inclusive list is most likely reflective of an increased sample reflective of the fact that we chose to not include clear cell

ClinCancerRes2005;11(11)June1,2005 4062 www.aacrjournals.org Downloaded from clincancerres.aacrjournals.org on September 25, 2021. © 2005 American Association for Cancer Research. Endometrial Cancer, Microarray

Ta b l e 1. Differentially expressed genes characteristic of endometrioid serous carcinomas (Cont’d)

E/PS E/MMT UniGene Gene Map Description 2.096 3.305 Hs.518542 HGD 3q21-q23 Homogentisate1,2-dioxygenase (homogentisate oxidase) 2.073 4.648 Hs.194710 GCNT3 15q21.3 Glucosaminyl (N-acetyl) transferase 3, mucin type 0.485 0.411 Hs.183650 CRABP2 1q21.3 Cellular retinoic acid binding protein 2 0.479 0.498 Hs.12844 EGFL6 Xp22 EGF-like domain, multiple 6 0.472 0.485 Hs.308628 SIAT8D 5q21 Sialyltransferase 8D (-2, 8-polysialyltransferase) 0.465 0.384 Hs.298646 PRO2000 8q24.13 PRO2000 protein 0.454 0.451 Hs.414407 HEC 18p11.31 Highly expressed in cancer, rich in leucine heptad repeats 0.450 0.279 Hs.336224 TMEFF1 9q31 Transmembrane protein with EGF-like 0.444 0.469 Hs.137047 FLJ25157 3p24.2 Hypothetical protein FLJ25157 0.443 0.193 Hs.76118 UCHL1 4p14 Ubiquitin carboxyl-terminal esterase L1 (ubiquitin thiolesterase) 0.441 0.426 Hs.445098 FLJ20354 1p31.2 Hypothetical protein FLJ20354 0.437 0.381 Hs.169840 TTK 6q13-q21 TTK protein kinase 0.430 0.364 Hs.125124 EPHB2 1p 3 6 .1-p 3 5 E p h B2 0.423 0.349 Hs.79440 IMP-3 7p11 IGF-II mRNA-binding protein 3 0.407 0.315 Hs.250687 TRPC1 3q22-q24 Transient receptor potential cation channel, subfamily C 0.384 0.421 Hs.444096 Human full-length cDNA 5V-end of clone CS0DK007YB08 0.382 0.382 Hs.65366 KIAA1495 Xq24 KIAA1495 protein 0.370 0.435 Hs.132246 SLC38A1 12q13.11 Solute carrier family 38, member1 0.358 0.311 Hs.444096 Human full-length cDNA 5V-end of clone CS0DK007YB08 0.346 0.264 Hs.512414 NR3C1 5q31 Nuclear receptor subfamily 3C1 (glucocorticoid receptor) 0.337 0.322 Hs.30299 IMP-2 3q28 IGF-II mRNA-binding protein 2 0.308 0.199 Hs.43577 ATP8B2 1q22 ATPase, class I, type 8B, member 2 0.306 0.298 Hs.444096 Human full-length cDNA 5V-end of clone CS0DK007YB08 0.302 0.135 Hs.126085 LRRN1 3p26.2 Leucine-rich repeat neuronal 1 0.276 0.210 Hs.14968 PLAG1 8q12 Pleiomorphic adenoma gene1 0.259 0.104 Hs.349109 IGF2 11p15.5 Insulin-like growth factor 2 (somatomedin A) 2.042 2.679 Hs.387367 CYP39A1 6p21.1-p11.2 Cytochrome P450, family 39, subfamilyA, polypeptide 1 2.035 3.538 Hs.184510 SFN 1p35.3 Stratifin 2.031 2.162 Hs.302738 SLC26A2 5q31-q34 Solute carrier family 26 (sulfate transporter), member 2 2.031 2.372 Hs.105547 NPDC1 9q34.3 Neural proliferation, differentiation and control, 1 2.030 3.485 Hs.28491 SAT Xp22.1 Spermidine/spermine N1-acetyltransferase 2.017 2.669 Hs.302738 SLC26A2 5q31-q34 Solute carrier family 26 (sulfate transporter), member 2

Abbreviations: E, endometrioid carcinoma; PS, papillary serous carcinoma; MMT, mixed mullerian tumors.

cases with papillary serous carcinomas in a comparative five of the nine cases of nonendometrioid cancer available analysis of endometrioid and nonendometrioid carcinomas. for fluorescence in situ hybridization analysis; the histologic Both our group (21) and other investigators (24) have type of positive cases was not reported (23) and it is possible independently determined that the expression profile of clear that the major proportion were clear cell. When we did cell carcinomas is distinct from that of papillary serous and supervised analysis on a predominantly advanced group of endometrioid histologic types among cases of endometrial endometrioid and papillary serous carcinomas that were cancer. Although Moreno-Beuno et al. (23) determined that matched for stage and grade, we obtained a slightly shorter the expression profiles of clear cell cases and papillary serous list of differentially expressed genes (data not shown). These cases were similar, the number of cases and the array findings suggest that although endometrioid and papillary platform used may have prohibited detection. Nevertheless, serous carcinomas are distinct, they may also share genetic the investigators determined that STK15 was amplified in alterations that are common to both types of endometrial

www.aacrjournals.org4063 Clin Cancer Res 2005;11(11) June 1, 2005 Downloaded from clincancerres.aacrjournals.org on September 25, 2021. © 2005 American Association for Cancer Research. Imaging, Diagnosis, Prognosis carcinoma, especially when matched for other clinical tigations comparing these types of tumors to normal prognostic factors. endometrium is warranted to determine whether COX-1 In our validation of genes differentially expressed between inhibitors might have a role in the prevention of these types endometrioid and papillary serous carcinomas, we noted that of endometrial cancer. the expression of cyclooxygenase I (PTGS1) was increased IGF2 was noted to be overexpressed in the analysis of mixed among the papillary serous adenocarcinomas when compared mullerian tumors (P < 0.001) when compared with both with endometrioid cancer or mixed mullerian tumors (Fig. 3). endometrioid and papillary serous endometrial cancers (Fig. 4). Cyclooxygenase I (PTGS1) is constitutively expressed in most Although the data are limited, several studies have suggested tissues in the body, whereas cyclooxygenase II (PTGS2)is that IGF2 is associated with sarcomas of the uterus. In vivo induced in response to certain stimuli. Both isoforms result analysis of the SK-UT-1 cell line, derived from a uterine mixed in production of prostaglandins, some of which have been mesodermal tumor, has revealed increased binding of IGF2 implicated in carcinogenesis (i.e., PGE-3 and 6-keto PGF1a) compared with insulin and IGF-I (28). In addition, IGF2 was and angiogenesis (25). COX-2 overexpression has been found to have a stimulatory effect on the growth of these cells, observed in endometrial adenocarcinomas (26) and its ex- whereas IGF-I had no effect (29). Finally, loss of imprinting pression may be associated with parameters of aggressiveness. associated with overexpression has been reported in association The only report that evaluated COX-1 in endometrial cancer with both leiomyosarcoma and mixed mullerian tumor of the did not report histologic type in association with their uterus (30). Together with our findings, the evidence suggests results but found COX-1 expression to be negligible (27). that IGF2 may be overexpressed among mixed mullerian In vitro studies have indicated that AKT induces COX-2 tumors of the uterus. expression in mutated PTEN endometrioid endometrial Two additional genes previously described in association cancer cells. Although these studies suggest an association with soft tissue sarcomas were also noted to be differentially between COX-2 and endometrioid endometrial cancer, there expressed between the mixed mullerian tumors and uterine have not been any reports evaluating COX-1 or COX-2 carcinomas. In the supervised analysis of both the mixed expression in papillary serous cancers. Our findings would mullerian tumor versus endometrioid and the mixed mullerian suggest that papillary serous adenocarcinomas of the endo- tumor versus papillary serous carcinoma, we observed up- metrium overexpress COX-1, indicating that further inves- regulation of SNAIL2, which induces epithelial-mesenchymal

Ta b l e 2 . Differentially expressed genes characteristic of papillary serous carcinoma

PS/E PS/MMT Gene UniGene Map Description 7.12 7 9.804 VGLL1 Hs.9030 Xq26.3 Vestigial like1 (Drosophila) 4.860 8.945 Hs.147613 Homo sapiens transcribed sequences 4.292 7.194 Hs.372225 Homo sapiens transcribed sequences 3.757 3.366 NPR1 Hs.438864 1q21-q22 Natriuretic peptide receptorA/guanylate cyclaseA (atrionatriuretic peptide receptorA) 3.254 3.357 LOC2210 02 Hs.125293 10q11.21 CG4853 gene product 3.187 2.580 FAM20A Hs.144633 17q24.3 Family with sequence similarity 20, memberA 3.113 2.533 ESTs,weakly similar toJC5314 CDC28/cdc2-like kinase associating arginine-serine cyclophilin 2.854 3.476 WNT7A Hs.72290 3p25 Wingless-type mouse mammary tumor virus integration site family, member 7A 2.789 3.011 THRB Hs.203213 3p24.3 Thyroid hormone receptor h [erythroblastic leukemia viral (v-erb-a) oncogene homologue 2, avian] 2.545 2.704 Hs.391828 Homo sapiens transcribed sequence with moderate similarity to protein NP_060265.1 2.379 3.736 PTGS1 Hs.88474 9q32-q33.3 Prostaglandin endoperoxide synthase1 (prostaglandin G/H synthase and cyclooxygenase) 2.363 2.013 LU Hs.155048 19q13.2 Lutheran blood group (Auberger Bantigen included) 2.315 2.977 PTGS1 Hs.88474 9q32-q33.3 Prostaglandin endoperoxide synthase1 (prostaglandin G/H synthase and cyclooxygenase) 2.277 2.748 PTGS1 Hs.88474 9q32-q33.3 Prostaglandin endoperoxide synthase1 (prostaglandin G/H synthase and cyclooxygenase) 2.255 2.010 CDH6 Hs.32963 5p15.1-p14 Cadherin 6, type 2, K-cadherin (fetal kidney) 2.240 3.697 Hs.201441 13 Homo sapiens cDNA FLJ11076 fis, clone PLACE1005077. 2.187 3.340 HUMPPA Hs.78358 17q25.2 Paraneoplastic antigen 2.178 2.477 KIAA1554 Hs.195642 17q25.3 KIAA1554 protein 2.097 2.479 Hs.152422 Homo sapiens cDNA FLJ27210 fis, clone SYN03494 2.023 2.317 Hs.124776 Homo sapiens mRNA; cDNA DKFZp564N1116 (from clone DKFZp564N1116) 0.480 0.396 FGFR1 Hs.748 8p11.2-p11.1 Fibroblast growth factor receptor1 (fms-related tyrosine kinase 2, Pfeiffer syndrome) 0.449 0.468 KIAA1025 Hs.435249 12q24.22 KIAA1025 protein

Clin Cancer Res 2005;11(11)June1,2005 4064 www.aacrjournals.org Downloaded from clincancerres.aacrjournals.org on September 25, 2021. © 2005 American Association for Cancer Research. Endometrial Cancer, Microarray

transition, cell spreading, and cell separation in vitro (31). Investigators have previously suggested that mixed mulle- SNAIL2 is also a direct repressor of the tumor suppressor gene rian tumors are characterized by molecular features that are E-cadherin, which also encodes a cell-to-cell adhesion molecule. more consistent with a carcinoma than a sarcoma. Many Increases in SNAIL2 can result in loss of adhesiveness associated have subsequently advocated the use of chemotherapeutics with reduction in E-cadherin leading to increased invasiveness for mixed mullerian tumors that have been traditionally used (32). Reduction of E-cadherin has been observed in analysis of in the treatment of uterine papillary serous carcinomas, soft tissue sarcomas (33), but there is limited evidence instead of regimens commonly used in the treatment of soft regarding E-cadherin expression in mixed mullerian tumor tissue sarcomas (18). Although mixed mullerian tumors are (34). Although our group has previously noted cadherin associated with a poor outcome that is characteristic of mutation in association with endometrial carcinomas (35), uterine papillary serous carcinomas, our findings suggest that there have been no prior reports by our group or others the majority of mixed mullerian tumors have gene expression regarding E-cadherin expression or SNAIL2 in uterine sarcomas. profiles that are distinct from both common histologic types In the analysis of mixed mullerian tumor versus either of common endometrial carcinomas and may optimally endometrioid or papillary serous carcinoma, there was a benefit from therapies that target the unique molecular limited number of mixed mullerian tumors that seemed to profile characteristic of these tumors. have a gene expression intermediate between the mixed The purpose of the current study was to identify genes that mullerian tumors and either type of carcinoma (Figs. 2B were differentially expressed between types of endometrial and 3B). These tumors did not seem to differ in terms of cancer, not those that distinguish normal endometrium from stage from the other mixed mullerian tumors that were more endometrial cancer subtypes. Comparison of endometrial cancer distinct in gene expression. It is possible that these cases may to normal endometrium is a complex undertaking and would have had less carcinomatous component comprising the require careful selection of normal samples with consideration mixed mullerian tumor. There are no prior reports that given to age, menopausal status, and stage of the menstrual cycle. quantify the proportion of carcinomatous elements that In conclusion, data from our group has previously typically comprise most uterine mixed mullerian tumors. suggested that the gene expression patterns associated with In the absence of this type of data, we did not choose to different histologic types of uterine cancer are distinct. Using dissect the tumors to guarantee a set proportion of a robust microarray platform that queried over 28,000 carcinomatous and sarcomatous components. Similarly, a UniGene clusters in combination with a large set of uterine small subset of endometrioid carcinomas seemed to have a cancer specimens, we have determined that the gene gene expression profile that was somewhat similar to that of expression of endometrioid and papillary serous carcinomas the mixed mullerian tumors (Fig. 2B). These tumors also did as well as mixed mullerian tumors seem to be distinct, not seem to be more advanced in stage or grade compared further supporting the paradigm that different histologic with the other endometrioid carcinomas that were more types of uterine cancer may develop in part via alternate distinct in gene expression profile. pathways.

References 1. American Cancer Society. Cancer facts and figure; 11. Kohler MF, Berchuck A, Davidoff AM, et al. Overex- 20. Fotiou S, Hatjieleftheriou G, Kyrousis G, et al. Long 2003. pression and mutation of p53 in endometrial carcino- term tamoxifen treatment: a possible etiological factor 2. Deligdisch L, Holinka CF. Endometrial carcinoma: ma. Cancer Res 1992;52:1622 ^ 7. in the development of uterine carcinosarcoma: two two diseases? Cancer Detect Prev 1987;10:237 ^ 12. Kohler MF, Carney P, Dodge R, et al. p53 overex- case reports and review of the literature. Anticancer 46, 198. pression in advanced-stage endometrial adenocarci- Res 2000;20:2015^ 20. 3. Berchuck A, Boyd J. Molecular basis of endometrial noma. AmJ Obstet Gynecol 1996;175:1246 ^ 52. 21. Risinger JI, Maxwell GL, Chandramouli GV, et al. cancer. Cancer 1995;76:2034 ^ 40. 13. Wada H, Enomoto T, Fujita M, et al. Molecular evi- Microarray analysis reveals distinct gene expression 4. Kong D, Suzuki A, Zou TT, Sakurada A, et al. PTEN1 is dence that most but not all carcinocarcinomas of the profiles among different histologic types of endome- frequently mutated in primary endometrial carcinomas. uterus are combination tumours. Cancer Res 1997; trial cancer. Cancer Res 2003;63:6 ^ 11. Nat Genet 1997;17:143 ^ 4. 57:5379 ^ 85. 22. Brown T, Mackey K. Analysis of RNA by Northern 5. RisingerJI, Hayes AK, Berchuck A, Barrett JC. PTEN/ 14. Thompson L, Change B, Barsky SH. Monoclonal ori- and slot blot hybridization. In: Ausubel FM, Brent R, MMAC1 mutations in endometrial cancers. Cancer gins of malignant mixed tumors (carcinosarcomas). Kingston RE, Moore DD, Seidman JG, Struhl K, edi- Res 1997;57:4736 ^ 8. Evidence for a divergent histogenesis. Am J Surg tors. Current protocols in molecular biology. NewYork: 6. Risinger JI, Hayes K, Maxwell GL, Barrett C, Pathol 1996;20:277 ^ 85. JohnWiley & Sons, Inc.; 1997. p. 4.9.2 ^4. Berchuck A. PTEN mutation in endometrial cancers is 15. Kounelis S, Jones MW, Papadaki H, et al. Carcino- 23. Moreno-Bueno G, Sanchez-Estevez C, Cassia R, associated with favorable clinical and pathologic char- sarcoma (malignant mixed mullerian tumors) of the fe- et al. Differential gene expression profile in endome- acteristics. Clin Cancer Res 1998;4:3005 ^ 10. male genital tract: comparative molecular analysis of trioid and nonendometrioid endometrial carcinoma: 7. Risinger JI, Berchuck A, Kohler MF, Watson P, epithelial and mesenchymal components. Hum Pathol STK15 is frequently overexpressed and amplified in Lynch HT, Boyd J. Genetic instability of microsatel- 1998;29:82^ 7. nonendometrioid carcinomas. Cancer Res 2003;63: lites in endometrial carcinoma. Cancer Res 1993;53: 16. Abeln EC, Smit VTHB,Wessels JW, et al. Molecular 5697 ^ 702. 5100 ^ 3. genetic evidence for the conversion hypothesis of the 24. Zorn KK, Awtrey CS, Gardner GJ, Barrett JC, 8. Maxwell GL, Risinger JI, Alvarez AA, Barrett JC, origin of malignant mullerian tumours. J Pathol 1997; Boyd JA, Birrer MJ. Gene expression profiles of se- Berchuck A. Microsatellite instability is associated 183:424 ^ 31. rous, endometrioid, and clear cell subtypes of ovar- with favorable survival in endometrial carcinomas. 17. McCluggage WG. Uterine carcinosarcomas (malig- ian and endometrial cancer. Gynecol Oncol 2003; Obstet Gynecol 2001;97:417^ 22. nant mixed Mullerian tumors) are metaplastic carcino- 88:abstract 90. 9. Maxwell GL, Risinger JI, Hayes KA, et al. Racial dis- mas. Int J Gynecol Cancer 2002;12:687 ^ 90. 25. SmithWL, Langenbach R.Why there are two cyclo- parity in the frequency of PTEN mutations, but not 18. McCluggage WG. Malignant biphasic uterine tumors: oxygenase isozymes. J Clin Invest 2001;107:1491 ^ 5. microsatellite instability, in advanced endometrial can- carcinosarcomas or metaplastic carcinomas. J Clin 26. Fujiwaki R, Iida K, Kanasaki H, Ozaki T, Hata K, cers. Clin Cancer Res 2000;6:2999 ^ 3005. Pathol 2002;55:321^5. Miyazaki K. Cyclooxygenase-2 expression in endome- 10. Lukes AS, Kohler MF, Pieper CF, et al. Multivariable 19. McCluggage WG, Abdulkadir M, Price JH, et al. trial cancer: correlation with microvessel count and ex- analysis of DNA ploidy, p53, and HER-2/neu as Uterine carcinosarcoma in patients receiving tamoxi- pression of vascular endothelial growth factor and prognostic factors in endometrial cancer. Cancer fen: a report of 19 cases. Int J Gynecol Cancer 2000; thymidine phophorylase. Hum Pathol 2002;33:213^ 9. 1994;73:2380^5. 37:285 ^ 7. 27. Uotila PJ, Erkkola RU, Klemi PJ. The expression of

www.aacrjournals.org4065 Clin Cancer Res 2005;11(11) June 1, 2005 Downloaded from clincancerres.aacrjournals.org on September 25, 2021. © 2005 American Association for Cancer Research. Imaging, Diagnosis, Prognosis

cyclooxygenase-1and -2 in proliferative endometrium. 30. Roy RN, Gerulath AH, Cecuttie A, Bhavnani BR. 33.YooJ, Park S, Kang CS, Kang SJ, Kim BK. Expres- Ann Med 2002;34:428 ^ 33. Loss of IGF-II imprinting in endometrial tumors: over- sion of E-cadherin and p53 proteins in human soft 28. Nagamani M, Stuart CA. Specific receptors and expression in carcinosarcoma. Cancer Lett 2000;153: tissue sarcomas. Arch Pathol Lab Med 2002;126: growth effects of insulin and insulin-like growth fac- 67 ^ 73. 33^8. tors in a human cell line derived from mixed mesoder- 31. Peinado H, Marin F, Cubillo E, et al. SNAIL and E47 34. Halachmi S, DeMarzo AM, Chow NH, et al. Genetic mal tumor of the uterus. J Soc Gynecol Investig 1996; repressors of E-cadherin induce distinctly invasive and alterations in urinary bladder carcinosarcoma: evi- 3:281 ^ 8. angiogenic properties in vivo. J Cell Sci 2004;117: dence of a common clonal origin. Eur Urol 2000;37: 29. Glondemans T, Prinsen I,Van Unnik JAM, Lips CJM, 2827 ^ 39. 350^7. Otter WD, Sussenbach JS. Insulin-like growth factor 32. Katoh M, Katoh M. Identification and characteriza- 35. Risinger JI, Berchuck A, Kohler MF, Boyd J. Muta- gene expression in human smooth muscle tumors. tion of human SNAIL3 (SNAI3)geneinsilico.IntJMol tions of the E-cadherin gene in human gynecologic Cancer Res 1990;50:6689 ^95. Med 2003;11:383 ^ 8. cancers. Nat Genet 1994;7:98 ^ 102.

Clin Cancer Res 2005;11(11)June1,2005 4066 www.aacrjournals.org Downloaded from clincancerres.aacrjournals.org on September 25, 2021. © 2005 American Association for Cancer Research. Microarray Analysis of Endometrial Carcinomas and Mixed Mullerian Tumors Reveals Distinct Gene Expression Profiles Associated with Different Histologic Types of Uterine Cancer

G. Larry Maxwell, G.V.R. Chandramouli, Lou Dainty, et al.

Clin Cancer Res 2005;11:4056-4066.

Updated version Access the most recent version of this article at: http://clincancerres.aacrjournals.org/content/11/11/4056

Cited articles This article cites 31 articles, 11 of which you can access for free at: http://clincancerres.aacrjournals.org/content/11/11/4056.full#ref-list-1

Citing articles This article has been cited by 3 HighWire-hosted articles. Access the articles at: http://clincancerres.aacrjournals.org/content/11/11/4056.full#related-urls

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

Reprints and To order reprints of this article or to subscribe to the journal, contact the AACR Publications Subscriptions Department at [email protected].

Permissions To request permission to re-use all or part of this article, use this link http://clincancerres.aacrjournals.org/content/11/11/4056. Click on "Request Permissions" which will take you to the Copyright Clearance Center's (CCC) Rightslink site.