Proteomics-Derived Biomarker Panel Improves ... - Home | Clinical Cancer … · Ovarian cancer affects 1.27% of females in the United States and remains a difﬁcult disease to treat

Precision Medicine and Imaging

Proteomics-Derived Biomarker Panel ImprovesDiagnostic Precision to Classify Endometrioid andHigh-grade Serous Ovarian CarcinomaDylan Z. Dieters-Castator1,2, Peter F. Rambau3, Linda E. Kelemen4, Gabrielle M. Siegers2,Gilles A. Lajoie5, Lynne-Marie Postovit1,2,6, and Martin K€obel7

Abstract

Purpose: Ovarian carcinomas are a group of distinct dis-eases classified by histotypes. As histotype-specific treatmentevolves, accurate classification will become critical for optimalprecision medicine approaches.

Experimental Design: To uncover differences between thetwomost common histotypes, high-grade serous (HGSC) andendometrioid carcinoma, we performed label-free quantita-tive proteomics on freshly frozen tumor tissues (HGSC,n¼10;endometrioid carcinoma, n ¼ 10). Eight candidate proteinbiomarkers specific to endometrioid carcinomawere validatedby IHC using tissue microarrays representing 361 cases ofeither endometrioid carcinoma or HGSC.

Results: More than 500 proteins were differentiallyexpressed (P < 0.05) between endometrioid carcinoma andHGSC tumor proteomes. A ranked set of 106 proteins wassufficient to correctly discriminate 90% of samples. IHC

validated KIAA1324 as the most discriminatory novel bio-marker for endometrioid carcinoma. An 8-marker panel wasfound to exhibit superior performance for discriminatingendometrioid carcinoma from HGSC compared with thecurrent standard of WT1 plus TP53 alone, improving theclassification rate for HGSC from 90.7% to 99.2%. Endo-metrioid carcinoma–specific diagnostic markers such asPLCB1, KIAA1324, and SCGB2A1 were also significantlyassociated with favorable prognosis within endometrioidcarcinoma suggesting biological heterogeneity within thishistotype. Pathway analysis of proteomic data revealeddifferences between endometrioid carcinoma and HGSCpertaining to estrogen and interferon signalling.

Conclusions: In summary, these findings support the use ofmulti-marker panels for the differential diagnosis of difficultcases resembling endometrioid carcinoma and HGSC.

IntroductionOvarian cancer affects 1.27% of females in the United States

and remains a difficult disease to treat with an overall 5-yearsurvival of only 46% (1). Epithelial ovarian cancer (ovariancarcinoma) is a biologically heterogeneous group of diseases

classified into five major histotypes, which are in descendingorder of frequency: high-grade serous (HGSC), endometrioid,clear cell (CCC), low-grade serous, and mucinous carcinoma (2).These histotypes differ with respect to their underlying molecularalterations and outcomes. Patients with endometrioid carcinomahave the best outcome with a 5-year survival of more than 80%,whereas patients with HGSC have the lowest with 40% (3).Histotypes also differ regarding response to therapy; thus, targetedtherapies formanaging patients stratified by histotype have begunto emerge (4). The ability to distinguish HGSC from endome-trioid carcinoma has significant ramifications for choosing theappropriate predictive test and genetic counseling. Clinicallyapproved PARP inhibitors are a promising treatment option forBRCA1/2 defective or chemosensitive HGSC (5), and the anti-angiogenic therapy, bevacizumab (Avastin), slightly increasesmedian progression-free survival in HGSC of the mesenchymalor proliferative molecular subtypes (6). In contrast, hormonaltherapy is a treatment option for low-stage endometrioid carci-noma and immunotherapy is emerging as an option for meta-static endometrioid carcinoma that are associated with defects inDNA mismatch repair genes (7–10). HGSC should be tested forBRCA1/2 germline mutations and endometrioid carcinoma formismatch repair deficiency and underlying Lynch syndrome.

Current histotype diagnosis consists of a combination of mor-phologic assessment by conventional light microscope with theintegration of IHC biomarkers, and this can achieve a precision of>90% (11, 12). The current IHC marker standard is the combi-nation of the presence of WT1 and mutant-type TP53, which ishighly specific but not entirely sensitive for HGSC (11, 13).

1Department of Anatomy and Cell Biology,Western University, London, Ontario,Canada. 2Department of Oncology University of Alberta, Edmonton, Alberta,Canada. 3Pathology Department, Catholic University of Health and AlliedSciences-Bugando, Mwanza, Tanzania. 4Department of Public Health Sciences,Medical University of SouthCarolina, Charleston, SouthCarolina. 5Department ofBiochemistry, Western University, London, Ontario, Canada. 6Department ofObstetrics and Gynecology, University of Alberta, Edmonton, Alberta, Canada.7Department of Pathology and Laboratory Medicine, University of Calgary,Foothills Medical Center, Calgary, Alberta, Canada.

Note: Supplementary data for this article are available at Clinical CancerResearch Online (http://clincancerres.aacrjournals.org/).

Corresponding Authors: Lynne-Marie Postovit, Western University and Uni-versity of Alberta, 114th St. and 87th Ave., Edmonton, Alberta T6G 2E1, Canada.Phone: 780-492-6342; E-mail: [email protected]; Gilles A. Lajoie, Don RixProtein Identification Facility, Department of Biochemistry, Western University,London, Ontario N6A 5C1, Canada. E-mail: [email protected]; and Martin K€obel,Department of Pathology and Laboratory Medicine, University of Calgary andCalgary Laboratory Services, 1304 S29 Street NW, Calgary, Alberta 2TN 2T9,Canada. E-mail: [email protected]

Clin Cancer Res 2019;25:4309–19

doi: 10.1158/1078-0432.CCR-18-3818

�2019 American Association for Cancer Research.

ClinicalCancerResearch

www.aacrjournals.org 4309

on June 16, 2020. © 2019 American Association for Cancer Research. clincancerres.aacrjournals.org Downloaded from

Published OnlineFirst April 12, 2019; DOI: 10.1158/1078-0432.CCR-18-3818

http://crossmark.crossref.org/dialog/?doi=10.1158/1078-0432.CCR-18-3818&domain=pdf&date_stamp=2019-6-27

http://clincancerres.aacrjournals.org/

Therefore, in a proportion of ovarian carcinomas cases, discrim-ination between HGSC and endometrioid carcinoma (mostlygrade 2 and 3) remains challenging based on similar morphologyand overlappingWT1 and TP53 expression (10, 14–16). Current-ly, the next best markers to distinguish endometrioid carcinomafrom HSGC are progesterone receptor (PGR), which suffersfrom limited sensitivity (81%) and low specificity (56%), andcatenin beta-1 (CTNNB1), which has only 50% sensitivity despite>95% specificity (11, 15). Thus, there is a need for better bio-markers that positively define endometrioid carcinoma.

Mass spectrometry (MS)-based proteomics is a powerful andunbiased technique for characterizing complex biological sys-tems (17). For instance, a large scale study of 169 HGSC tumorsby the Clinical Proteomics Tumour Analysis Consortium corre-lated The Cancer Genome Atlas (TCGA) data with protein expres-sion to identify novel signatures and pathways associated withsurvival (18). In another study, profiling ovarian (cancer) andfallopian tube cell line proteomes revealed three distinct groups(epithelial, clear cell, and mesenchymal) and a 67-protein signa-ture, which segregated patients withHGSC (TCGA) into good andpoor overall survival (19). In addition, multiplexed quantitativeproteomic analysis of formalin-fixed, paraffin-embedded HGSC,CCC, and endometrioid carcinoma tumors identified cystathio-nine g-lyase as a novel marker for CCC (20).

In an attempt to elucidate histotype-specific markers of endo-metrioid carcinoma, we undertook a MS-based proteomicsapproach using fresh frozen tumour samples from patients withHGSC and endometrioid carcinoma. Selected candidates werevalidated by IHC on a cohort of 361 (HGSC and endometrioidcarcinoma) ovarian tumors to assess their clinical potential fordiscriminating these histotypes.

Materials and MethodsPatient samples

All experiments were conducted in accordance with the prin-ciples of the Declaration of Helsinki. Fresh frozen tumor tissuewas obtained from the Alberta Cancer Research Biorepositoryfollowing approval by the Human Research Ethics Boardof Alberta (HREBA CC-16-555 and CC-16-0913). Ten recentchemo-naive HGSC and 10 endometrioid carcinoma were select-ed on the basis of the diagnosis assessed by routine pathology.

These cases showed prototypical features by pathology, as well asthe expected clinical features (Supplementary Table S1).

Protein extraction from fresh frozen tumorsTo prepare samples for LC-MS/MS, tumor cores were par-

tially thawed on ice and a section corresponding to approxi-mately 100 mg was removed with a razor blade. Tumor sectionswere immediately wrapped in tinfoil and submerged in liquidnitrogen for approximately 10 minutes. Cryopreserved tumorpieces wrapped in tinfoil were pulverized into a fine powder byhitting with a mallet approximately 3–5 times. One mL of lysisbuffer containing 8 mol/L urea, 50 mmol/L ammonium bicar-bonate (ABC), 10 mmol/L dithiothreitol (DTT), and 2% sodi-um dodecyl sulfate was added directly to the dissociated tumorsample on tinfoil and carefully transferred into a 1.5 mLmicrofuge tube. Tumor samples were sonicated (�20 � 0.5seconds pulses; Level 1) with a probe sonicator (Fisher Scien-tific Sonic Dismembrator Model 100) on ice to break upresidual tissue chunks and reduce viscosity. Lysates were clar-ified for 5 minutes at 14,000� g to remove insoluble debris andquantified using a Pierce 660nm Protein Assay with IonicDetergent Compatibility Reagent (Thermo Fisher Scientific)before storing at �80�C until future use.

Chloroform/methanol protein precipitationA 100 mg aliquot of tumor lysate was reduced in 10 mmol/L

DTT for 30 minutes and alkylated in 100 mmol/L iodoacetamidefor 30 minutes at room temperature in the dark. Proteins wereprecipitated in chloroform/methanol in 1.5 mL microfuge tubesaccording to Wessel and Fl€ugge (21). Briefly, samples in lysisbuffer were topped up to 150 mLwith 50mmol/L ABC thenmixedwith ice-cold methanol (600 mL) followed by addition of ice-coldchloroform (150 mL) and vortexed thoroughly. An additionalvolume (450 mL) of 4�C distilled water was added followed byvortexing and centrifugation at 14,000� g for 5 minutes. The topaqueous/methanol phase was carefully removed to avoid dis-turbing the precipitated protein interphase. A second 450 mLvolume of cold methanol was added to each sample followedby vigorous vortexing and centrifugation at 14, 000 � g for 5minutes. The remaining chloroform/methanol supernatant wasdiscarded and the precipitated protein pellet was left to air dry in afume hood.

On-pellet in-solution digestionOn-pellet in-solution protein digestion was performed simi-

larly to Duan and colleagues (22). Briefly, precipitated proteinpellets were reconstituted in 100 mL of 50 mmol/L ABC (pH 8)and sonicated (�3 � 0.5 seconds pulses) to break up the pellet.LysC (Wako Chemicals) and Mass Spec Grade Trypsin/LysC Mix(Promega) were added to protein samples at 1:100 and 1:50ratios, respectively. Protein digestion was carried out at 37�C on aThermoMixer C (Eppendorf) at 300 rpm overnight (�18 hours).The next day an additional volume of trypsin/LysC mix (1:100ratio) was added to each sample and mixed at 1,400 rpm. After 4hours, digests were acidified to pH 3–4 with 10 mL of 10% formicacid and centrifuged at 14,000 � g to pellet insoluble materialprior to LC-MS/MS.

LC-MS/MSDigests were initially analyzed using an M-Class NanoAquity

UPLC System (Waters) connected to an Orbitrap Elite Mass

Translational Relevance

Precisionmedicine approaches in ovarian cancer require theaccurate diagnosis of histotypes. In particular, there is a criticalneed to distinguish endometrioid from high-grade serous(HGSC) carcinomas, which differ greatly in outcomes, geneticpredisposition, and optimal treatment approaches. Thisapproach, involving the IHC detection of WT1 and TP53,performs well in discriminating endometrioid carcinomafrom HGSC; however, up to 10% of tumors are still misclas-sified. Herein we used proteomics to discover biomarkers thatmay distinguish endometrioid carcinoma from HGSC, andthen used IHC to validate these using a contemporarily clas-sified cohort of endometrioid carcinoma and HGSC samples.This pipeline revealed eight biomarkers that can be used toincrease the accuracy of current diagnostic methods.

Dieters-Castator et al.

Clin Cancer Res; 25(14) July 15, 2019 Clinical Cancer Research4310




Spectrometer (Thermo Fisher Scientific) and then, following IHCvalidation, subsequently analyzed on a Q Exactive Plus (QEþ)Mass Spectrometer (Thermo Fisher Scientific). Buffer A consistedof water/0.1% formic acid and buffer B consisted of acetonitrile/0.1% formic acid. Peptides (�1 mg measured using a Piercebicinchoninic acid assay) were initially loaded onto an ACQUITYUPLC M-class symmetry C18 trap column (5 mm, 180 mm �20 mm) and trapped for 6 minutes at a flow rate of 5mL/minat 99% A/1% B. Peptides were separated on an ACQUITYUPLC M-class peptide BEH C18 column (130 Å, 1.7 mm, 75 mm� 250 mm) operating at a flow rate of 300 nL/min at 35�Cusing a nonlinear gradient consisting of 1%–7% B over 1minute,7%–23% B over 179 minutes (134 for QEþ), and 23%–35% Bover 60 minutes (45 for QEþ) before increasing to 95% B andwashing. MS acquisition settings are provided in SupplementaryTable S2. The MS proteomics data have been deposited to theProteomeXchange Consortium via the PRIDE partner repositorywith the dataset identifier PXD012998 (23, 24).

Data analysisMSfileswere searched inMaxQuant (1.5.8.3) using theHuman

UniProt database (reviewed only; updatedMay 2017with 42,183entries; refs. 25, 26). Missed cleavages were set to three andcysteine carbamidomethylation was set as a fixed modification.Oxidation (M), N-terminal acetylation (protein), and deamida-tion (NQ) were set as variable amino acid modifications (max-imum number of modifications per peptide ¼ 5) and all othersettings were left as default. Protein and peptide FDR were set to0.01 (1%) and the decoy database was set to revert. The "matchbetween runs" feature was utilized to maximize proteome cov-erage and label-free quantification (LFQ; ref. 27). Datasets wereloaded into Perseus (version 1.5.5.3 and 1.6.2.1) and proteins"only identified by site" or "matched to reverse" (decoy) databasehits were removed (28). Proteins detected (razor þ unique pep-tides) and quantified (log2-transformed LFQ intensities) in eachtumor sample are provided in Supplementary Tables S3 and S4.Protein identificationswith quantitative values in�2 samples in aleast one group (HGSC or endometrioid carcinoma) wereretained for downstream analysis unless specified elsewhere.Missing values were imputed using a width of 0.3 and down shiftof 1.8. Gene ontology (GO) cellular component analysis andpathway annotation were performed using PantherDB (release13.1) andMetascape (version 3.0), respectively (29, 30). Gene setenrichment analysis (GSEA)were performed inGSEA v3.0 (BroadInstitute, Cambridge, MA) with all settings left default (exceptminimum gene set size of 10) and the Molecular SignatureDatabase (MSigDB) v6.2 collections: canonical pathways, hall-mark, oncogenic signatures, and GO biological processes (31).Support vector machine analysis and Pearson correlation heat-mapswere produced using the Bioconductor packages geNetClas-sifier and complexHeatmap, respectively (32, 33). The Estimationof STromal and Immune cells in MAlignant Tumours usingExpression data (ESTIMATE) score in R was used as a surrogateto assess epithelial (tumor) cell purity (34). Gene lists belongingto each of the fourHGSCmolecular subtypes and z-scored proteinexpression data were used to cluster HGSC samples in this studyand the CPTAC (19, 35).

IHCThe validation samples were from patients diagnosed with

HGSC and endometrioid carcinoma between 1978 and 2010

identified from the Alberta Cancer Registry in Canada. Detailsof histology review and tissue microarray construction, as well assubsequent biomarker-based histotyping, were described previ-ously (11, 36), resulting in 172 HGS and 189 endometrioidcarcinoma (HREBA CC-16-161; Supplementary Table S1). IHCstaining was performed on 4-mm sections from tissuemicroarraysrepresenting each case by up to three 0.6-mm cores using a DAKOOmnis platform within the Anatomical Pathology Research Lab-oratory at Calgary Laboratory Services in the Department ofPathology and Laboratory Medicine at the University of Calgary(Calgary, Alberta, Canada). Detailed assay and antibody infor-mation, as well as scoring cutoffs are provided in SupplementaryTable S5.

Statistical analysisTwo-sided, two sample Student t tests (P < 0.05) were used to

identify differentially expressed proteins. ROC-AUCs were calcu-lated from proteomics data in GraphPad Prism (Version 6.01).Fisher exact test was used to determine statistically significantdifferences (P < 0.05) between the proportion of endometrioidcarcinoma and HGSC tumors staining positive and negative foreach marker. Nominal logistic regression and Kaplan–Meiersurvival analysis were performed in JMP v13 (SAS).

ResultsGlobal proteomic analysis of HGSC and endometrioidcarcinoma

To identify proteins elevated in or exclusive to endometrioidcarcinoma versus HGSC, we performed label-free quantitativeproteomics on unfractionated digests from freshly frozen tumors(n ¼ 10 endometrioid carcinoma and n ¼ 10 HGSC samples;Supplementary Fig. S1A). Patient characteristics for each subtypeare listed in Supplementary Table S1. On average, approximately4,700 (unique) proteins (3,143–5,101 proteins; SupplementaryFig. S1B) were identified in each sample (�98% contained LFQintensities) and 6,360 proteins were detected in total (Supple-mentary Fig. S1C). Most proteins detected were of cytoplasmicorigin (cell part) or belonging to macromolecular protein com-plexes and organelles according to PantherDB (SupplementaryFig. S1D; ref. 29). Within the Perseus computational platform,approximately 8% of the quantified proteins within each samplewere similarly annotated with the cellular component (GO slim)term "extracellular matrix" or "extracellular region" and theseproteins accounted for a mean 37% of the total MS signalquantified (Supplementary Fig. S2A and S2B; ref. 28). Interest-ingly, the number of quantified proteins per sample were nega-tively correlated (r ¼ �0.965; Supplementary Fig. S2C) with thecumulativeMS signal (abundance) of these extracellular proteins.Analysis using the ESTIMATE tool revealed minor variations instromal and immune scores, and epithelial (tumor) purities(0.67–0.87; Supplementary Fig. S2D–S2F; ref. 34).

We next compared HGSC and endometrioid carcinoma sam-ples using unsupervised hierarchical clustering and principalcomponent analysis (PCA) on a filtered list of approximately5,648 proteins (minimum �3 LFQ intensities in a least onegroup). Although 98% of these proteins were detected in atleast one sample within each histotype (Fig. 1A) and tumorproteomes correlated regardless of histotype (SupplementaryFig. S3), most endometrioid carcinoma and HGSC samples(except HGSC-1 and 9) clustered independently based on PCA

Precision Biomarker Set for Ovarian Cancer Classification

www.aacrjournals.org Clin Cancer Res; 25(14) July 15, 2019 4311




(Fig. 1B). Furthermore, 537 proteins were found to be differen-tially expressed (two-sample t test; P < 0.05; Fig. 1C; Supplemen-tary Table S6). Among the differentially expressed proteins, weidentified known highly expressed markers in endometrioidcarcinoma, for example PGR (4.20 log2 fold change), MMP7(1.45 log2 fold change), and CTNNB1 (0.86 log2 fold change)andhighly expressed inHGSC, for example TP53 (�2.06 log2 foldchange), MSLN (�2.67 log2 fold change), IGF2 (�1.62 log2 foldchange), and CDKN2A (�2.57 log2 fold change), which sup-ported the validity of our approach (2, 37).

Selecting endometrioid carcinoma enriched proteins for IHCvalidation

To filter candidates of endometrioid carcinoma for validationby IHC,wefirst tabulated proteinswithpeptide evidence in�80%of tumor samples from one and �20% of samples from theother histotype (Supplementary Table S7). Although this strategyidentified 15 and seven proteins that were largely specific toendometrioid carcinoma and HGSC tumors, respectively, it maynot capture proteins with high differential expression. Therefore,we analyzed our proteomics data using the R package geNetClas-sifier to rank proteins with the greatest classification power inan unbiased fashion (32). In total, 106 proteins exceeded the

posterior probability cutoff (�0.95) and were used in training thesupport vector machine (Supplementary Table S8). The lowesterror rate achieved by geNetClassifier was 0.1 (10%) and corre-sponded to 69 proteins. Interestingly, the top two ranked pro-teins, MUC5B (mucin-5B) and PIGR (polymeric immunoglobu-lin receptor), were not identified by our initial filtering strategywhereas PGRwas identifiedbyboth approaches.We subsequentlyperformed unsupervised hierarchical clustering utilizing Pearsoncorrelation coefficients calculated from LFQ intensities restrictedto the top 106 proteins ranked by geNetClassifier. This clusteringapproach segregated HGSC and endometrioid carcinoma sam-ples relativelywell but revealed a third central cluster comprisedofthree endometrioid carcinoma and two HGSC samples (Fig. 2A).In agreement with our earlier findings, some tumor proteomes,HGSC-1 and 9 in particular, more closely resembled the opposinghistotype suggesting close biological relation.

Validating diagnostic markers for endometrioid carcinomaOn the basis of availability of specific antibodies for IHC

validation, we selected six endometrioid carcinoma candidateswith large log2 fold changes (MUC5B and PIGR) and/or highspecificity (PLCB1, PAM, KIAA1324, and SCGB2A1). ROC-AUCsfor the selected candidates for endometrioid carcinoma ranged

Figure 1.

Comparing endometrioid carcinoma (EC) and HGSC tumour proteomes.A, Venn diagram reveals high overlap between proteins expressed in HGSC andendometrioid carcinoma subtypes. B, PCA illustrating moderate clustering of tumor samples within each subtype. C, Volcano plot of log2 fold changes in LFQintensities (endometrioid carcinoma vs. HGSC) reveals a large number of differentially expressed proteins. Proteins significantly elevated in endometrioidcarcinoma or HGSC are colored in red and blue, respectively. Several of the top differentially expressed proteins are labeled in black. Only protein entriescontaining LFQ intensities in at least three of 10 samples for either subtype (HGSC or endometrioid carcinoma) were retained for comparisons.






from 0.82 to 0.99when calculated using LFQ intensities (Fig. 2B).Validation by IHC was performed on 361 independent tumorsections (189 endometrioid carcinoma and 172 HGSC) scoredusing a three-tier system (Fig. 3A; Supplementary Figs. S4 andS5). For comparison, we also performed IHC on standarddiagnostic histotype-specific markers of endometrioid carcino-ma (PGR and CTNNB1) and HGSC (WT1 and TP53). For allendometrioid carcinoma candidates, we confirmed that expres-sion was significantly greater in endometrioid carcinoma com-pared with HGSC (Supplementary Fig. S5). Of note, KIAA1324,an estrogen-induced gene (alternative name EIG121; ref. 38),was the only marker that exhibited performance similar to thatof PGR with a sensitivity of 88.5% and specificity of approx-imately 53% (Fig. 3B). Although PLCB1 and PAM yielded thehighest ROC-AUC values in our proteomics dataset (0.99 and0.97, respectively), these proteins were among the least sensi-tive markers by IHC (62.94 and 56.90%, respectively; Figs. 2Band 3B). SCGB2A1 (mammaglobin-B) was the least specificmarker (32.59%), which was unexpected given that its expres-sion, like KIAA1324, was relatively exclusive to endometrioidcarcinoma versus HGSC samples in our proteomics analyses(Fig. 3B; Supplementary Table S7). WT1 and TP53, as antici-pated, performed exceptionally well with sensitivities of 97.6%and 92.2%, and specificities of 91.4 and 88.7% for HGSC,respectively (Fig. 3B).

Given that individual markers did not outperform knownstandards such as PGR, we evaluated multi-marker combinationsusing nominal logistic regression. Our six endometrioid carcino-ma candidates (PLCB1, PAM, KIAA1324, SCGB2A1,MUC5B, andPIGR) yielded a ROC-AUC of 0.90016 albeit 32 of 166 (19.3%)endometrioid carcinoma and 22 of 135 (16.3%) HGSC casesremained misclassified (Fig. 4A–F). The standard clinical markercombination, WT1 plus TP53 (Model 2), achieved a ROC-AUC of0.9915 that corresponded to only 1 of 184 (0.5%) endometrioidcarcinomabut still 15of 162 (9.3%)HGSC cases that couldnot beclassified accurately (Fig. 4B–F). Adding four endometrioid car-cinoma candidates (KIAA1324, PAM, PIGR, and SCGB2A1) toModel 2 improved ROC-AUCs to 0.99594 (Model 3; Fig. 4C).Adding known experimentalmarkers PGRandCTNNB1 toModel2 achieved a similar ROC-AUC of 0.99670 (Model 4; Fig. 4D),whichwas further improved by addition of the four endometrioidcarcinoma candidates (KIAA1324, PAM, PIGR, and SCGB2A1) to

achieve a ROC-AUC of 0.99751 (Model 5; Fig. 4E). The final8-marker Model 5 could not classify four cases [3/170 (1.8%)endometrioid carcinoma and 1/130 (0.8%) HGSC)] and thusrepresents an improvement when compared with WT1 and TP53alone (Fig. 4F).

Next, we evaluated prognostic associations between markersfor endometrioid carcinoma and patient survival. Interestingly,PLCB1 [HR ¼ 0.38; 95% confidence interval (CI), 0.18–0.78],KIAA1324 (HR, 0.23; 95% CI, 0.11–0.51), and SCGB2A1 (HR,0.35; 95%CI, 0.18–0.75)were associatedwith favorable outcomein endometrioid carcinomawith log-rank P< 0.0001, 0.0003, and0.0120, respectively (Fig. 5).

Deep proteome coverage obtained through complementaryMSapproaches reveals differences in endometrioid carcinoma andHGSC biology

To explore biological pathways in detail that differ betweenendometrioid carcinoma and HGSC, we acquired a second pro-teomics dataset using a more sensitive Q Exactive Plus massspectrometer and the same unfractionated samples. For reference,using the same filtering criteria, approximately 26% and approx-imately 32%more proteins were identified (8,041 vs. 6,360) andretained (7,452 vs. 5,648) for quantitative comparisons (endo-metrioid carcinoma vs. HGSC) with this instrument, respectively(Supplementary Fig. S8A and S8B; Supplementary Table S9). PCAappeared unaffected (near identical components 1 and 2) by theadditional depth and, importantly, all candidates we validated,except for PLCB1, were significantly elevated in endometrioidcarcinoma, thus indicating minimal technical variability betweeneach instrument (Supplementary Fig. S8C and S8D). In addition,our discovery set of HGSC samples appeared to exhibit proteinexpression patterns resembling each of the four previouslyreported molecular subtypes (35), and this observation wasfurther corroborated by clustering our HGSC proteomes with the169 HGSC samples from the CPTAC (Supplementary Fig. S8;ref. 19).

The differentially expressed proteins and complete proteomes(endometrioid carcinoma and HGSC) from our second proteo-mics dataset were analyzed using Metascape (www.metascape.org) and GSEA to identify overrepresented and enriched path-ways/processes, respectively (30, 31). In endometrioid carcino-ma, terms associated with trafficking and localization, estrogen

Figure 2.

Protein selection for downstreamvalidation of endometrioidcarcinoma markers. Proteomes withLFQ intensities from each tumorsample were analyzed using theBioconductor packagegeNetClassifier in R. A,Unsupervised hierarchicalclustering of tumor samplesaccording to Pearson correlationcoefficients using 106 proteins witha posterior probability�0.95.B, ROC-AUC of endometrioidcarcinoma targets selected forvalidation by IHC.





www.metascape.org

www.metascape.org


signalling, and fatty acid metabolism appeared consistently over-represented or enriched (Fig. 6A; Supplementary Tables S10 andS11). In agreement with the IHC, CTNNB1 activation(BCAT.100_UP.V1_UP) was also significant in endometrioidcarcinoma (normalized enriched score ¼ 1.64; FDR q-value ¼0.24). Alternatively, immune/antiviral responses, mitotic andcell morphologic processes, and androgen receptor signalling(PID_AR_PATHWAY) were predominately matched to HGSC

(Fig. 6B; Supplementary Tables S12 and S13). Followingcloser inspection, key differences between endometrioid carcino-ma and HSGC could be illustrated by heatmap using normal-ized (z-scored) expression values of core proteins enriched inestrogen response late (Hallmark) and antiviral mechanism byIFN-stimulated genes (Reactome) gene sets (Fig. 6D). Indeed,known estrogen-induced genes were highly expressed in endo-metrioid carcinoma, whereas the E1 (UBA7), E2 (UBE2L6), and

Figure 3.

IHC staining and performance of endometrioid carcinomamarkers.A, Images (200�magnification; scale bar,�140 mm) taken from endometrioid carcinoma andHGSC tumor sections showing absent and present staining for PLCB1, PAM, KIAA1324, SCGB2A1, MUC5B, PIGR, PGR, WT1, nuclear, and membranous staining forCTNNB1, and wild-type andmutant-type staining for TP53. B, Corresponding sensitivity and specificity values of each target for detecting endometrioidcarcinoma or HGSC (TP53 andWT1 only) using a cut-off score�1. PGR and KIAA1324 were the best performing endometrioid carcinomamarkers with relativelygood sensitivity but intermediate specificity.






E3 (HERC5) ubiquitin ligases that mediate ISGylation of manytargets (MX1, MX2, IFT1, UBE2E1, STAT1, DDX58, EIF2AK2,and UBE2L6) via conjugation of the ubiquitin-like modifierIFN-stimulated gene 15 (ISG15) were central to the IFNresponse in HGSC (39). The Ubl carboxyl-terminal hydrolase18 (USP18) that removes ISGylation was also enriched inHGSC.

DiscussionThrough robust label-free quantitative proteomics and IHC

analyses, we identified and validated an eight-marker panel,which could improve the classification rate for HGSC from90.7% (standard model consisting of WT1 plus TP53; ref. 11) to99.2%. This improvement greatly decreases the risk of under-treatment of patients with aggressive HGSC increasing eligibilityfor targeted therapies (PARP inhibitors) and triaging to theappropriate hereditary screening programs (BRCA1/2 hereditarybreast/ovarian cancer as opposed to Lynch syndrome screening).

In the era of precision medicine, wherein accurate diagnosis is ofutmost importance, our discovery is of high clinical utility.

Using our approach, wewere able to identify a number of novel(KIAA1324, SCGB2A1, MUC5B, PLCB1, PAM, and PIGR) andexisting (PR, CTNNB1, WT1, and TP53) histotype-specific bio-markers for the differential diagnosis of endometrioid carcinomaand HGSC. Although novel candidates effectively discriminatedendometrioid carcinoma and HGSC proteomes, KIAA1324 wasthe only individual biomarker that achieved goodperformance byIHC. The performance of individualmarkers in IHCwas generallyworse than that predicted by the proteomics data, which wasmostly due to a lack of specificity for endometrioid carcinomausing the most widely clinically accepted cutoff for ancillarydiagnostic IHC markers (absence vs. presence of expression).SCGB2A1 (mammaglobin-B), for example, is primarily expressedin the breast, uterus, and salivary gland (40), and greater mRNAexpression levels have been reported in endometrioid comparedwith serous ovarian cancer samples (41, 42). We also foundSCGB2A1 to be one of the top differentially expressed proteins

Figure 4.

A combination of markers with improved diagnostic potential. A–E, ROC curves and AUC values for different combinations of markers for classifyingendometrioid carcinoma (EC) from HGSC. A panel of eight proteins (Model 5) exhibited superior performance over the standard HGSCmarkers WT1 and TP53.F, Number of misclassified cases associated with each model.






in our study, but IHC demonstrated only 34% specificity forendometrioid carcinoma. This suggests that unlike IHC, lowabundance proteins such as transcription factors remain difficultto detect using unfractionated (single-shot) proteomic analysis.Unfortunately, due to lack of quality antibodies, we could notvalidate additional targets with high specificity for endometrioidcarcinoma (e.g., PPM1H); hence the targeted development of

quality antibodies may be warranted. Alternatively, our resultsindicate a multi-marker model for endometrioid carcinoma clas-sification is a promising strategy given the lack of validating of asingle defining endometrioid carcinoma marker. This notion isalso supported by recent work byMartinez-Garcia and colleagueswho employed targeted proteomics on uterine aspirates frompatients with and without endometrial cancer, a disease sharingsimilarities with endometrioid carcinoma (43–45). Of note,SCGB2A1, CTNNB1, and PIGR were significantly elevated inpatients with endometrial cancer compared benign conditions,and among others were the best single performing markers fordiscriminating endometrioid and serous histotypes in the endo-metrium as well (45, 46). As proteomics technologies and pro-tocols become more robust, easy to use and economical, forexample, in the area of plasma biomarkers, there is potential fordirect integration into current clinical work flows (47). However,current IHC-based classification is the most efficient first-lineapproach, which is superior compared with targeted sequencingapproaches. Although histotype-specific molecular differencesbetween HGSC and endometrioid carcinoma exist, even theircombined sensitivities aremuch less than that of IHCmarkers thatis, many cases are not informative. For example, BRCA1/2 muta-tions are only present in about 25% of HGSC, and ARID1Amutations in about 35% of endometrioid carcinoma.

Beyond our biomarker discovery and validationwork, pathwayanalysis revealed interesting biological differences between endo-metrioid carcinoma and HGSC, for example, with estrogen sig-nalling associated with the former and interferon signalling withthe latter. This finding was also corroborated by a previous studyby Hughes and colleagues (20). The importance of hormonereceptor signalling for endometrioid carcinoma is well estab-lished and ESR1 and PGR are validated as prognostic markersfor endometrioid carcinoma (7). KIAA1324, an estrogen-inducedgene, was highly coexpressed with PGR and was associated withfavorable prognosis in endometrioid carcinoma. Therefore,KIAA1324 could be used as a complementary diagnostic, prog-nostic, and predictive marker for an active hormone axis inendometrioid carcinoma by identifying patients that may benefitfromhormonal therapy. Interestingly, two other diagnostic endo-metrioid carcinomamarkers (PLCB1 and SCGB2A1) also showedfavorable prognostic associations in endometrioid carcinomasuggesting that thesemarkers segregate abiologicallymore aggres-sive subset within endometrioid carcinoma that does not expresstypical endometrioid carcinoma markers. Indeed, SCGB2A1expression is associated with favorable prognosis in endometrialcarcinomas and longer recurrence-free survival in ovarian carci-nomas (41, 48). Lack of these markers might warrant moreaggressive therapy, similar to that used for HGSC.

The enrichment of key regulators and targets of IFN responsesessential to innate immunity and antiviral defense in HGSCproteomes is interesting given that IFN signalling is commonlyassociated with antitumoral effects, including cell-cycle arrest,apoptosis, and immune responses (49). Among these proteinswas the ubiquitin-like modifier ISG15, which was previouslydetected in 86% of HGSC and associated with increased overallsurvival (50). STAT1, which mediates type I (IFNa/b) and II(IFNg) IFN responses, was also elevated in HGSC, and previouslyassociated with increased progression-free survival, response tochemotherapy and CD8þ T-cell infiltration in HGSC (51, 52).Paradoxically but interestingly, recent evidence implicatesIFN/JAK/STAT signaling in resistance to immune checkpoint

Figure 5.

Survival curves and prognostic value of endometrioid carcinoma candidates.Three (PLCB1, KIAA1324, and SCGB2A1) of the six candidates validated byIHC were associated with good prognosis (increased overall survival) inpatients with endometrioid carcinoma.






Figure 6.

Pathway analysis reveals characteristics of endometrioid carcinoma and HGSC. Proteins significantly elevated in endometrioid carcinoma (A) or HGSC (B)samples were analyzed in Metascape, and the top 20 significant overrepresented pathways are shown. C, Heatmap comparing similarities and differencesbetween highly significant process and pathways. D, Heatmap showing normalized (z-scored) expression of core proteins matching to estrogen response late orantiviral mechanism by IFN-stimulated genes.






blockade, which might explain why immune checkpointblockade is ineffective in HGSC (53).

There are several potential limitations of our proteomic anal-ysis that have implications for candidate selection. First, thesample size of the discovery analysis was relatively small. How-ever, our goal was to identify highly discriminatory diagnosticbiomarkers that work in most samples all the time, therefore,more cases in the discovery cohort would not have yielded adifferent result. We observed two HGSC samples with proteomesoverlapping with endometrioid carcinoma, indicating a relativelyclose biological relation between these tumor types. Hence,finding a single discriminatory marker is unlikely. However, webelieve that the potentially confounding effect of those twoHGSCsamples is small because we focused our discovery on definingendometrioid carcinomabiomarkers. Our focus on endometrioidcarcinoma markers also diminishes the role of intertumoralheterogeneity in HGSC, namely molecular subtypes. While weshowed reasonable representation of molecular subtypes in thediscovery cohort, the size of validation almost certainly confirms arepresentation of all HGSC molecular subtypes. Of note, weselected for tumor intrinsic biomarkers (i.e., proteins expressedin epithelial tumor cells), whereas molecular subtype is mainly areflection of gene expression differences within the microenvi-ronment. Additional validationusing even larger sample sizeswillbe important to confirm our proteomics findings. For example,our marker panel can be applied to the large OTTA consortium,with known molecular subtypes, to refine histotype-specific out-come associations (37). Third, some putative diagnostic proteinsmay not exhibit high differential expression but may differpredominately in cellular compartmentalization, mutation sta-tus, andposttranslationalmodifications (e.g., CTNNB1andTP53;ref. 13). Yet we identified CTNNB1 and TP53 as differentiallyexpressed proteins, which supports the validity of our approach.Fourth, the preponderance of stromal, immune, and other celltypes in the microenvironment may mask differences in lowabundance proteins between endometrioid carcinoma andHGSC. Although our approach yielded the desired outcome toimprove upon classification of HGSC and endometrioid carcino-ma, we would make the following recommendation for futureproteomic studies: To increase the sensitivity for identification oftumor intrinsic protein markers, enrichment techniques such as

laser capture microdissection, single-cell approaches, posttrans-lational modification analysis, and/or subcellular fractionationmay reveal additional candidates.

In conclusion, we have discovered and validated a biomarkerpanel that can improve the ability to discriminate between endo-metrioid carcinoma and HSGC. Application of this protocolcould substantially improve the classification of endometrioidcarcinoma and HGSC and could direct more precise histotype-specific treatment options.

Disclosure of Potential Conflicts of InterestNo potential conflicts of interest were disclosed.

Authors' ContributionsConception and design: L.-M. Postovit, M. K€obelDevelopment of methodology: D.Z. Dieters-Castator, G.A. Lajoie, L.-M.Postovit, M. K€obelAcquisition of data (provided animals, acquired and managed patients,provided facilities, etc.): D.Z. Dieters-Castator, P.F. Rambau, L.E. Kelemen,G.A. Lajoie, L.-M. Postovit, M. K€obelAnalysis and interpretation of data (e.g., statistical analysis, biostatistics,computational analysis): D.Z. Dieters-Castator, P.F. Rambau, L.E. Kelemen,G.A. Lajoie, L.-M. Postovit, M. K€obelWriting, review, and/or revision of the manuscript: D.Z. Dieters-Castator,P.F. Rambau, L.E. Kelemen, G.M. Siegers, G.A. Lajoie, L.-M. Postovit, M. K€obelStudy supervision: L.-M. Postovit

AcknowledgmentsWe thank Paula Pittock for technical support and Dr. Miljan Kuljanin for

feedback on the article. This work was supported by the Sawin-Bladwin Chair inOvarian Cancer Research and the Dr. Anthony Noujaim Oncology Chairawarded to L.-M. Postovit by theWomenandChildrenHealthResearch Instituteand the Alberta Cancer Foundation, respectively.We thank Shuhong Liu, YoungOu, and Deon Richards for immunohistochemical stains, and Thomas Kryton,BFA, digital imaging specialist for Alberta Public Lab, for scanning the tissuemicroarrays. The study was supported by Alberta Public Lab internal researchsupport RS17-608 and an NSERC Operating Grant awarded to G. A. Lajoie.

The costs of publication of this articlewere defrayed inpart by the payment ofpage charges. This article must therefore be hereby marked advertisement inaccordance with 18 U.S.C. Section 1734 solely to indicate this fact.

Received November 23, 2018; revised February 18, 2019; accepted April 8,2019; published first April 12, 2019.

References1. Noone AM, Howlader N, Krapcho M, Miller D, Brest A, Yu M, et al.

SEER cancer statistics review, 1975–2015, NCI. Bethesda, MD: 2017.Available from: https://seer.cancer.gov/archive/csr/1975_2015/.

2. K€obelM, Kalloger SE, BoydN,McKinney S,Mehl E, Palmer C, et al.Ovariancarcinoma subtypes are different diseases: implications for biomarkerstudies. PLoS Med 2008;5:e232.

3. Le Page C, Rahimi K, K€obel M, Tonin PN, Meunier L, Portelance L, et al.Characteristics and outcome of the COEURCanadian validation cohort forovarian cancer biomarkers. BMC Cancer 2018;18:347.

4. Banerjee S, Kaye SB. New strategies in the treatment of ovarian cancer:current clinical perspectives and future potential. Clin Cancer Res 2013;19:961–8.

5. Vaughan S, Coward JI, Bast RC, Berchuck A, Berek JS, Brenton JD, et al.Rethinking ovarian cancer: recommendations for improving outcomes.Nat Rev Cancer 2011;11:719–25.

6. Kommoss S,Winterhoff B, Oberg AL, KonecnyGE,Wang C, Riska SM, et al.Bevacizumab may differentially improve ovarian cancer outcome inpatients with proliferative and mesenchymal molecular subtypes.Clin Cancer Res 2017;23:3794–801.

7. Sieh W, K€obel M, Longacre TA, Bowtell DD, deFazio A, Goodman MT,et al. Hormone-receptor expression and ovarian cancer survival: anOvarian Tumor Tissue Analysis consortium study. Lancet Oncol2013;14:853–62.

8. Westdorp H, Fennemann FL, Weren RDA, Bisseling TM, Ligtenberg MJL,Figdor CG, et al. Opportunities for immunotherapy in microsatelliteinstable colorectal cancer. Cancer Immunol Immunother 2016;65:1249–59.

9. Rambau PF, Duggan MA, Ghatage P, Warfa K, Steed H, Perrier R,et al. Significant frequency of MSH2/MSH6 abnormality in ovarianendometrioid carcinoma supports histotype-specific Lynch syn-drome screening in ovarian carcinomas. Histopathology 2016;69:288–97.

10. Assem H, Rambau PF, Lee S, Ogilvie T, Sienko A, Kelemen LE, et al. High-grade endometrioid carcinomaof theovary: a clinicopathologic studyof 30cases. Am J Surg Pathol 2018;42:534–44.

11. K€obel M, Rahimi K, Rambau PF, Naugler C, Le Page C, Meunier L, et al. Animmunohistochemical algorithm for ovarian carcinoma typing. Int JGyne-col Pathol 2016;35:430–41.





https://seer.cancer.gov/archive/csr/1975_2015/


12. K€obel M, Luo L, Grevers X, Lee S, Brooks-Wilson A, Gilks CB, et al.Ovarian carcinoma histotype: strengths and limitations of integratingmorphology with immunohistochemical predictions. Int J GynecolPathol 2018 Jun 12 [Epub ahead of print].

13. K€obelM, Piskorz AM, Lee S, Lui S, LePage C,Marass F, et al. Optimized p53immunohistochemistry is an accurate predictor of TP53 mutation inovarian carcinoma. J Pathol Clin Res 2016;2:247–58.

14. Schwartz DR, Kardia SLR, Shedden KA, Kuick R,Michailidis G, Taylor JMG,et al. Gene expression in ovarian cancer reflects both morphology andbiological behavior, distinguishing clear cell from other poor-prognosisovarian carcinomas. Cancer Res 2002;62:4722–9.

15. Madore J, Ren F, Filali-Mouhim A, Sanchez L, K€obel M, Tonin PN, et al.Characterization of the molecular differences between ovarian endome-trioid carcinoma and ovarian serous carcinoma. J Pathol 2010;220:392–400.

16. K€obel M, Bak J, Bertelsen BI, Carpen O, Grove A, Hansen ES, et al. Ovariancarcinoma histotype determination is highly reproducible, and isimproved through the use of immunohistochemistry. Histopathology2014;64:1004–13.

17. Hughes CS, Postovit LM, Lajoie GA. Matrigel: a complex protein mixturerequired for optimal growth of cell culture. Proteomics 2010;10:1886–90.

18. Zhang H, Liu T, Zhang Z, Payne SH, Zhang B, McDermott JE, et al.Integrated proteogenomic characterization of human high-grade serousovarian cancer. Cell 2016;166:755–65.

19. Coscia F, Watters KM, Curtis M, Eckert MA, Chiang CY, Tyanova S, et al.Integrative proteomic profiling of ovarian cancer cell lines reveals precursorcell associated proteins and functional status.NatCommun2016;7:12645.

20. Hughes CS, McConechy MK, Cochrane DR, Nazeran T, Karnezis AN,HuntsmanDG, et al.Quantitative profiling of single formalinfixed tumoursections: proteomics for translational research. Sci Rep 2016;6:34949.

21. Wessel D, Fl€ugge UI. A method for the quantitative recovery of protein indilute solution in the presence of detergents and lipids. Anal Biochem1984;138:141–3.

22. Duan X, Young R, Straubinger RM, Page B, Cao J, Wang H, et al. Astraightforward and highly efficient precipitation/on-pellet digestion pro-cedure coupledwith a long gradient nano-LC separation and orbitrapmassspectrometry for label-free expression profiling of the swine heart mito-chondrial proteome. J Proteome Res 2009;8:2838–50.

23. DeutschEW,CsordasA, SunZ, JarnuczakA, Perez-Riverol Y, Ternent T, et al.The ProteomeXchange consortium in 2017: supporting the cultural changein proteomics public data deposition. Nucleic Acids Res 2017;45:D1100–6.

24. Mayer G, Bernal-LlinaresM, Audain E, Ylmaz S , Pfeuffer J, Uszkoreit J, et al.The PRIDE database and related tools and resources in 2019: improvingsupport for quantification data. Nucleic Acids Res 2018;47:D442–50.

25. Cox J, Mann M. MaxQuant enables high peptide identification rates,individualized p.p.b.-range mass accuracies and proteome-wide proteinquantification. Nat Biotechnol 2008;26:1367–72.

26. UniProt Consortium. UniProt: a hub for protein information.Nucleic Acids Res 2015;43:D204–12.

27. Cox J,HeinMY, LuberCA, Paron I,NagarajN,MannM.Accurate proteome-wide label-free quantification by delayed normalization and maximalpeptide ratio extraction, termed MaxLFQ. Mol Cell Proteomics 2014;13:2513–26.

28. Tyanova S, Temu T, Sinitcyn P, Carlson A, Hein MY, Geiger T, et al.The Perseus computational platform for comprehensive analysis of(prote)omics data. Nat Methods 2016;13:731–40.

29. Mi H, Muruganujan A, Thomas PD. PANTHER in 2013: modeling theevolution of gene function, and other gene attributes, in the context ofphylogenetic trees. Nucleic Acids Res 2013;41:377–86.

30. Tripathi S, Pohl MO, Zhou Y, Rodriguez-Frandsen A, Wang G, Stein DA,et al. Meta- and orthogonal integration of influenza "oMICs" data defines arole for UBR4 in virus budding. Cell Host Microbe 2015;18:723–35.

31. SubramanianA, TamayoP,Mootha VK,Mukherjee S, Ebert BL,GilletteMA,et al. Gene set enrichment analysis: a knowledge-based approach forinterpreting genome-wide expression profiles. Proc Natl Acad Sci 2005;102:15545–50.

32. Aibar S, Fontanillo C, Droste C, Roson-Burgo B, Campos-Laborie FJ,Hernandez-Rivas JM, et al. Analyse multiple disease subtypes and build

associated gene networks using genome-wide expression profiles.BMC Genomics 2015;16:S3.

33. Gu Z, Eils R, Schlesner M. Complex heatmaps reveal patterns and correla-tions inmultidimensional genomic data. Bioinformatics 2016;32:2847–9.

34. Yoshihara K, Shahmoradgoli M, Martínez E, Vegesna R, Kim H,Torres-Garcia W, et al. Inferring tumour purity and stromal and immunecell admixture from expression data. Nat Commun 2013;4:2612.

35. Leong HS, Galletta L, Etemadmoghadam D, George J, K€obel M, Ramus SJ,et al. Efficientmolecular subtype classificationof high-grade serous ovariancancer. J Pathol 2015;236:272–7.

36. Kelemen LE, Warren GW, Koziak JM, K€obel M, Steed H. Smoking maymodify the association between neoadjuvant chemotherapy and survivalfrom ovarian cancer. Gynecol Oncol 2016;140:124–30.

37. Rambau PF, Vierkant RA, Intermaggio MP, Kelemen LE, Goodman MT,Herpel E, et al. Association of p16 expression with prognosis varies acrossovarian carcinoma histotypes: an Ovarian Tumor Tissue Analysis consor-tium study. J Pathol Clin Res 2018;4:250–61.

38. Deng L, Broaddus RR, McCampbell A, Shipley GL, Loose DS, Stancel GM,et al. Identification of a novel estrogen-regulated gene, EIG121, induced byhormone replacement therapy and differentially expressed in type I andtype II endometrial cancer. Clin Cancer Res 2005;11:8258–64.

39. Jeon YJ, Yoo HM, Chung CH. ISG15 and immune diseases.Biochim Biophys Acta 2010;1802:485–96.

40. Becker RM, DarrowC, Zimonjic DB, Popescu NC,WatsonMA, Fleming TP.Identification ofmammaglobin B, a novelmember of the uteroglobin genefamily. Genomics 1998;54:70–8.

41. Tassi RA, Calza S, Ravaggi A, Bignotti E, Odicino FE, Tognon G, et al.Mammaglobin B is an independent prognosticmarker in epithelial ovariancancer and its expression is associated with reduced risk of disease recur-rence. BMC Cancer 2009;9:1–12.

42. Bellone S, Tassi R, Betti M, English D, Cocco E, Gasparrini S, et al.Mammaglobin B (SCGB2A1) is a novel tumour antigen highly differen-tially expressed in all major histological types of ovarian cancer: implica-tions for ovarian cancer immunotherapy. Br J Cancer 2013;109:462–71.

43. Cancer Genome Atlas Research Network. Integrated genomic analyses ofovarian carcinoma. Nature 2011;474:609–15.

44. Getz G, Gabriel SB, Cibulskis K, Lander E, Sivachenko A, Sougnez C, et al.Integrated genomic characterization of endometrial carcinoma. Nature2013;497:67–73.

45. Martinez-Garcia E, Lesur A, Devis L, Campos A, Cabrera S, van Oostrum J,et al. Development of a sequential workflow based on LC-PRM for theverification of endometrial cancer protein biomarkers in uterine aspiratesamples. Oncotarget 2016;7:53102–15.

46. Martinez-Garcia E, Lesur A, Devis L, Cabrera S, Matias-Guiu X, HirschfeldM, et al. Targeted proteomics identifies proteomic signatures in liquidbiopsies of the endometrium to diagnose endometrial cancer and assist inthe prediction of the optimal surgical treatment. Clin Cancer Res 2017;23:6458–67.

47. Geyer PE, Holdt LM, Teupser D, Mann M. Revisiting biomarker discoveryby plasma proteomics. Mol Syst Biol 2017;13:942.

48. Tassi RA, Bignotti E, Falchetti M, Calza S, Ravaggi A, Rossi E, et al.Mammaglobin B expression in human endometrial cancer. Int J GynecolCancer 2008;18:1090–6.

49. Parker BS, Rautela J, Hertzog PJ. Antitumour actions of interferons: impli-cations for cancer therapy. Nat Rev Cancer 2016;16:131–44.

50. Darb-Esfahani S, Sinn BV, Rudl M, Sehouli J, Braicu I, Dietel M, et al.Interferon-stimulated gene, 15 kDa (ISG15) in ovarian high-grade serouscarcinoma: prognostic impact and link to NF-kB pathway. Int J GynecolPathol 2014;33:16–22.

51. Au KK, Le Page C, Ren R, Meunier L, Cl�ement I, Tyrishkin K, et al. STAT1-associated intratumoural TH1 immunity predicts chemotherapy resistancein high-grade serous ovarian cancer. J Pathol Clin Res 2016;2:259–70.

52. Koti M, Siu A, Cl�ement I, Bidarimath M, Turashvili G, Edwards A, et al. Adistinct pre-existing inflammatory tumourmicroenvironment is associatedwith chemotherapy resistance in high-grade serous epithelial ovariancancer. Br J Cancer 2015;112:1215–22.

53. Benci JL, Xu B, Qiu Y, Wu TJ, Dada H, Twyman-Saint Victor C, et al. Tumorinterferon signaling regulates a multigenic resistance program to immunecheckpoint blockade. Cell 2016;167:1540–54.






2019;25:4309-4319. Published OnlineFirst April 12, 2019.Clin Cancer Res Dylan Z. Dieters-Castator, Peter F. Rambau, Linda E. Kelemen, et al. CarcinomaPrecision to Classify Endometrioid and High-grade Serous Ovarian Proteomics-Derived Biomarker Panel Improves Diagnostic

Updated version

10.1158/1078-0432.CCR-18-3818doi:

Access the most recent version of this article at:

Material

Supplementary

http://clincancerres.aacrjournals.org/content/suppl/2019/04/12/1078-0432.CCR-18-3818.DC1

Access the most recent supplemental material at:

Cited articles

http://clincancerres.aacrjournals.org/content/25/14/4309.full#ref-list-1

This article cites 51 articles, 9 of which you can access for free at:

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

Subscriptions

Reprints and

[email protected]

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

Permissions

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

.http://clincancerres.aacrjournals.org/content/25/14/4309To request permission to re-use all or part of this article, use this link



http://clincancerres.aacrjournals.org/lookup/doi/10.1158/1078-0432.CCR-18-3818

http://clincancerres.aacrjournals.org/content/suppl/2019/04/12/1078-0432.CCR-18-3818.DC1

http://clincancerres.aacrjournals.org/content/25/14/4309.full#ref-list-1

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

mailto:[email protected]

http://clincancerres.aacrjournals.org/content/25/14/4309


Documents

Proteomics-Derived Biomarker Panel Improves ... - Home | Clinical Cancer … · Ovarian cancer affects 1.27% of females in the United States and remains a difﬁcult disease to treat