Microvesicle Proteomic Profiling of Uterine Liquid Biopsy for Ovarian Cancer Early Detection
High-grade ovarian cancer (HGOC) is the leading cause of mortality from gynecological malignancies, because of diagnosis at a metastatic stage. Current screening options fail to improve mortality because of the absence of early-stage-specific biomarkers. We postulated that a liquid biopsy, such as utero-tubal lavage (UtL), may identify localized lesions better than systemic approaches of serum/plasma analysis. Further, while mutation-based assays are challenged by the rarity of tumor DNA within nonmutated DNA, analyzing the proteomic profile, is expected to enable earlier detection, as it reveals perturbations in both the tumor as well as in its microenvironment. To attain deep proteomic coverage and overcome the high dynamic range of this body fluid, we applied our method for microvesicle proteomics to the UtL samples. Liquid biopsies from HGOC patients (n = 49) and controls (n = 127) were divided into a discovery and validation sets. Data-dependent analysis of the samples on the Q-Exactive mass spectrometer provided depth of 8578 UtL proteins in total, and on average ∼3000 proteins per sample. We used support vector machine algorithms for sample classification, and crossed three feature-selection algorithms, to construct and validate a 9-protein classifier with 70% sensitivity and 76.2% specificity. The signature correctly identified all Stage I lesions. These results demonstrate the potential power of microvesicle-based proteomic biomarkers for early cancer diagnosis. High-grade ovarian cancer (HGOC) is the leading cause of mortality from gynecological malignancies, because of diagnosis at a metastatic stage. Current screening options fail to improve mortality because of the absence of early-stage-specific biomarkers. We postulated that a liquid biopsy, such as utero-tubal lavage (UtL), may identify localized lesions better than systemic approaches of serum/plasma analysis. Further, while mutation-based assays are challenged by the rarity of tumor DNA within nonmutated DNA, analyzing the proteomic profile, is expected to enable earlier detection, as it reveals perturbations in both the tumor as well as in its microenvironment. To attain deep proteomic coverage and overcome the high dynamic range of this body fluid, we applied our method for microvesicle proteomics to the UtL samples. Liquid biopsies from HGOC patients (n = 49) and controls (n = 127) were divided into a discovery and validation sets. Data-dependent analysis of the samples on the Q-Exactive mass spectrometer provided depth of 8578 UtL proteins in total, and on average ∼3000 proteins per sample. We used support vector machine algorithms for sample classification, and crossed three feature-selection algorithms, to construct and validate a 9-protein classifier with 70% sensitivity and 76.2% specificity. The signature correctly identified all Stage I lesions. These results demonstrate the potential power of microvesicle-based proteomic biomarkers for early cancer diagnosis. Overall survival of patients with high-grade ovarian cancer (HGOC)1.Vaughan S. Coward J.I. Bast R.C. Berchuck A. Berek J.S. Brenton J.D. Coukos G. Crum C.C. Drapkin R. Etemadmoghadam D. Friedlander M. Gabra H. Kaye S.B. Lord C.J. Lengyel E. Levine D.A. McNeish I.A. Menon U. Mills G.B. Nephew K.P. Oza A.M. Sood A.K. Stronach E.A. Walczak H. Bowtell D.D. Balkwill F.R. Rethinking ovarian cancer: recommendations for improving outcomes.Nat. Rev. Cancer. 2011; 11: 719-725Crossref PubMed Scopus (930) Google Scholar correlates with disease stage at diagnosis: whereas patients with stage I disease have >90% 5-year overall survival, rates for stage IV disease are extremely low. Regrettably, ∼75% of HGOC cases are diagnosed at late-stage regardless of adherence to testing recommendations (1.Vaughan S. Coward J.I. Bast R.C. Berchuck A. Berek J.S. Brenton J.D. Coukos G. Crum C.C. Drapkin R. Etemadmoghadam D. Friedlander M. Gabra H. Kaye S.B. Lord C.J. Lengyel E. Levine D.A. McNeish I.A. Menon U. Mills G.B. Nephew K.P. Oza A.M. Sood A.K. Stronach E.A. Walczak H. Bowtell D.D. Balkwill F.R. Rethinking ovarian cancer: recommendations for improving outcomes.Nat. Rev. Cancer. 2011; 11: 719-725Crossref PubMed Scopus (930) Google Scholar). Early-detection of HGOC among high-risk population, such as germline BRCA1/2 mutation carriers, is of exceptional importance. These women are currently counseled to undergo prophylactic removal of the ovaries and fallopian tubes (risk reducing bilateral salpingo-oophorectomy, RRBSO) at age ∼40, therefore there is an urgent unmet need for a personalized risk-assignment modality to guide RRBSO timing and alleviate unnecessary morbidity of early menopause (2.Harmsen M.G. Arts-de Jong M. Hoogerbrugge N. Maas A.H.E.M. Prins J.B. Bulten J. Teerenstra S. Adang E.M.M. Piek J.M.J. van Doorn H.C. van Beurden M. Mourits M.J.E. Zweemer R.P. Gaarenstroom K.N. Slangen B.F.M. Vos M.C. van Lonkhuijzen L.R.C.W. Massuger L.F.A.G. Hermens R.P.M.G. de Hullu J.A. Early salpingectomy (TUbectomy) with delayed oophorectomy to improve quality of life as alternative for risk-reducing salpingo-oophorectomy in BRCA1/2 mutation carriers (TUBA study): a prospective non-randomised multicentre study.BMC Cancer. 2015; 15: 593-601Crossref PubMed Scopus (76) Google Scholar, 3.Arts-de Jong M. Harmsen M.G. Hoogerbrugge N. Massuger L.F. Hermens R.P. de Hullu J.A. Risk-reducing salpingectomy with delayed oophorectomy in BRCA1/2 mutation carriers: Patients' and professionals' perspectives.Gynecol. Oncol. 2015; 136: 305-310Abstract Full Text Full Text PDF PubMed Scopus (27) Google Scholar). This grim reality stems primarily from the lack of effective screening methods and of early-stage biomarkers. A multitude of biomarkers have been proposed and tested over the years, however even the most established markers, namely serum CA125 and HE4, have not proven to be effective in improving survival (4.Lu K.H. Skates S. Hernandez M.A. Bedi D. Bevers T. Leeds L. Moore R. Granai C. Harris S. Newland W. Adeyinka O. Geffen J. Deavers M.T. Sun C.C. Horick N. Fritsche H. Bast R.C. A 2-stage ovarian cancer screening strategy using the Risk of Ovarian Cancer Algorithm (ROCA) identifies early-stage incident cancers and demonstrates high positive predictive value.Cancer. 2013; 119: 3454-3461Crossref PubMed Scopus (96) Google Scholar, 5.Moore R.G. McMeekin D.S. Brown A.K. DiSilvestro P. Miller M.C. Allard W.J. Gajewski W. Kurman R. Bast Jr, R.C. Skates S.J. A novel multiple marker bioassay utilizing HE4 and CA125 for the prediction of ovarian cancer in patients with a pelvic mass.Gynecol Oncol. 2009; 112: 40-46Abstract Full Text Full Text PDF PubMed Scopus (641) Google Scholar, 6.Sölétormos G. Duffy M.J. Othman Abu Hassan S. Verheijen R.H.M. Tholander B. Bast R.C. Gaarenstroom K.N. Sturgeon C.M. Bonfrer J.M. Petersen P.H. Troonen H. CarloTorre G. Kanty Kulpa J. Tuxen M.K. Molina R. Clinical Use of Cancer Biomarkers in Epithelial Ovarian Cancer: Updated Guidelines From the European Group on Tumor Markers.Int. J. Gynecol. Cancer. 2015; 26: 43-51Crossref Scopus (163) Google Scholar, 7.Karlan B.Y. Thorpe J. Watabayashi K. Drescher C.W. Palomares M. Daly M.B. Paley P. Hillard P. Andersen M.R. Anderson G. Drapkin R. Urban N. Use of CA125 and HE4 serum markers to predict ovarian cancer in elevated-risk women.Cancer Epidemiol. Biomarkers Prev. 2014; 23: 1383-1393Crossref PubMed Scopus (28) Google Scholar, 8.Song E. Gao Y. Wu C. Shi T. Nie S. Fillmore T.L. Schepmoes A.A. Gritsenko M.A. Qian W.-J. Smith R.D. Rodland K.D. Liu T. Targeted proteomic assays for quantitation of proteins identified by proteogenomic analysis of ovarian cancer.Sci. Data. 2017; 4: 170091Crossref PubMed Scopus (17) Google Scholar). Several recent large-scale screening trials based on blood CA125-based monitoring, with or without transvaginal ultrasound, showed insignificant stage shift among high-risk population and low specificity and sensitivity (9.Jacobs I.J. Menon U. Ryan A. Gentry-Maharaj A. Burnell M. Kalsi J.K. Amso N.N. Apostolidou S. Benjamin E. Cruickshank D. Crump D.N. Davies S.K. Dawnay A. Dobbs S. Fletcher G. Ford J. Godfrey K. Gunu R. Habib M. Hallett R. Herod J. Jenkins H. Karpinskyj C. Leeson S. Lewis S.J. Liston W.R. Lopes A. Mould T. Murdoch J. Oram D. Rabideau D.J. Reynolds K. Scott I. Seif M.W. Sharma A. Singh N. Taylor J. Warburton F. Widschwendter M. Williamson K. Woolas R. Fallowfield L. McGuire A.J. Campbell S. Parmar M. Skates S.J. Ovarian cancer screening and mortality in the UK Collaborative Trial of Ovarian Cancer Screening (UKCTOCS): a randomised controlled trial.Lancet. Full Text Full Text PDF PubMed Scopus Google Scholar, E. A. C.C. L. C. D.J. B. D. B. J.D. J.M. T.L. P. G. Miller J.K. of screening on ovarian cancer the and Ovarian Cancer Screening 2011; PubMed Scopus Google Scholar, S. R. Burnell M. P. R. I. Benjamin E. Singh N. Ryan A. Liston R. Dawnay A. Ford J. Gunu R. J. Skates S.J. Menon U. I.J. Ovarian Cancer Screening of stage shift in women diagnosed with ovarian cancer of the Ovarian Cancer Screening Oncol. 2017; PubMed Scopus Google Scholar). have been proposed as and the and for early are not established K. K. A. Y. E. T. Clinical of in ovarian Cancer. 15: PubMed Scopus Google Scholar, J.D. L. Y. C. B. L. C. A.A. F. A. M.G. M. R. J. L. J. N. M. J.D. J. P. J. M. S. C. C. B. A.M. N. and of cancers with a blood PubMed Scopus Google Scholar, W. K. S.J. G. P. Crum M. D. potential for a serum for of ovarian 2017; PubMed Scopus Google Scholar). testing for most biomarkers because of with tumor results in diagnosis at the metastatic stage. body are expected to the biomarkers at an earlier disease stage. the most of from lesions that in the of the fallopian the of the fallopian to the R. Drapkin R. the of ovarian cancer PubMed Scopus Google Scholar, K. Crum C. Drapkin R. into the of ovarian cancer and its Oncol. 26: PubMed Scopus Google Scholar, E. D. N. M. R. M. S. J. L. M. A. Kurman R. R. Drapkin R. ovarian in the fallopian 2017; PubMed Scopus Google Scholar). the of the or liquid from the gynecological may markers of the lesions. Several liquid methods were primarily for DNA E. Y. N. R. E. I. F. J. I. N. B. P. R. of the for of A Oncol. 2015; PubMed Scopus Google Scholar, I. Y. R.D. M.G. N. B. of in of patients with high-grade ovarian Gynecol. 2014; PubMed Scopus Google Scholar, I. C. Y. Wu J. N. Kurman R. F. Levine D.A. R. R. N. B. of DNA from the to ovarian and 2013; Scopus Google Scholar, Y. L. C. J.D. I. K. S.K. Kurman S. J. M. J. N. L. A. M. K. B. L. Levine D.A. K. J. L. R. C. B. L. N. of liquid from the and liquid biopsies for the of and ovarian PubMed Scopus Google Scholar). of showed low sensitivity and proteomics may be to assays for the of of because it is to the of both tumor and microenvironment. The of discovery in the high dynamic range of most body of primarily the proteins from tumor and therefore To overcome this as we have we deep proteomic analysis of to of in M. P. O. Y. T. of with a novel proteomic method for 2015; Full Text Full Text PDF PubMed Scopus Google Scholar). by of the and are into body from all A. J.M. H. D. in cancer: of Full Text Full Text PDF PubMed Scopus Google Scholar, A.M. 2017; Full Text Full Text PDF PubMed Scopus Google Scholar). as a of the for the of an that be used as a screening or of in Rev. Cancer. 2017; PubMed Scopus Google Scholar). are of analysis of the potential biomarkers. the we the analysis to utero-tubal lavage samples. We the of proteomics with and a proteomic as a early HGOC samples were in with of the at and from patients gynecological HGOC or ovarian or gynecological I and with and were as well as patients with ovarian in both patients and and all three analysis in a for we UtL and of were because of the were to the discovery HGOC patients and whereas samples were as a validation (n = and in a be in and for UtL samples in the proteomic of stage stage for ovarian reducing in a UtL samples were of by in the or into the the of were into the and fallopian tubes and an average of per as M. P. O. Y. T. of with a novel proteomic method for 2015; Full Text Full Text PDF PubMed Scopus Google Scholar). the UtL samples were at for to and were at and analysis in as UtL samples were at for to by microvesicle by at for at were with and at for at were in in with at for and with for in the The were with by with to and to were with on and J. Y. M. for and sample in PubMed Scopus Google Scholar). The were in the analysis. were by using the to the Q-Exactive or Q-Exactive mass were on with with The were with A and with a of at a of over a of in a with of the from for and analysis. Full were at a of or range of with of and of or were for with an of with of or were at a of or with of or and of or to of all the samples are with for based are in the were in the and the J. M. high mass and 26: PubMed Scopus Google Scholar, J. N. A. M. A into the 2011; PubMed Scopus Google Scholar). were for the discovery (n = and the validation (n = using the were the with a of the and a of The as a and and as for the mass of the were and in the the mass for and and the and the of The to and a of per The results were with discovery of for and for The in used for and the proteins that not be based on the identified were into a were with the S. T. P. A. T. M. J. The for analysis of PubMed Scopus Google Scholar). we proteins identified in the proteins identified based on the potential and The and the from the for the discovery and validation are as analysis of the discovery on the were to proteins with in at of the samples. were by with low that a with a of and of of the The of the discovery are provided in on the of the discovery vector using to a predictive signature that the and ovarian cancer We three and I. J. S. for cancer using support vector Scopus Google Scholar). of validation on the discovery with of of of the samples as and as The of of three methods to predictive with the rates in the discovery and of the validation in the as for the discovery The of the classifier on the of the proteins in the validation and the were in on the using The of the signature proteins by the sample of the discovery by the and based on the of the the of the HGOC and were from the Tumor to tumor The were as K. Y. M.C. Crum J.A. Drapkin R. of fallopian as a for ovarian 2009; PubMed Scopus Google Scholar, K. H. S. J. Y. J. T. Y. A. K. of fallopian to in lesions of Oncol. 2014; Full Text Full Text PDF PubMed Scopus Google Scholar). from HGOC and using by to using for the are in were from the of at the with the We of cases of of patients that were and to as of for the to HGOC age = age = of the not the fallopian age = and Risk reducing bilateral salpingo-oophorectomy prophylactic removal of the ovaries and fallopian tubes because of age = of HGOC age = range were and for and on a of or in of in of of of and of of positive and positive by for or by for analysis used to the of age and with the used to the of with and used to the of with to identify early-stage biomarkers with high sensitivity and specificity liquid biopsies from the of the with deep proteomics of the samples. To the of the utero-tubal body and we the method for by and with analysis M. P. O. Y. T. of with a novel proteomic method for 2015; Full Text Full Text PDF PubMed Scopus Google Scholar). we from of UtL liquid samples by and by and were on the Q-Exactive or Q-Exactive and proteins were using the in samples from HGOC patients and controls gynecological from three analysis all in a to the and there are with the sample and analysis identified a of we markers of such as and as well as and such as and The dynamic range of of the microvesicle of with our results K. Y. M.C. Crum J.A. Drapkin R. of fallopian as a for ovarian 2009; PubMed Scopus Google the of markers CA125 HE4 and not HGOC patients and samples the of CA125 in UtL samples with a and patients and controls not we potential or in of samples for UtL and analysis analysis showed the of low the and the three and analysis samples showed an average of within and of we controls from than patients and controls from the We therefore that the and are and not We we identify proteins patients and to identify markers, we divided the into a discovery and a validation To the discovery and the validation we of in A discovery a of patients and used to construct a classifier for HGOC diagnosis. to patients from all three patients and carriers, and have of cases and analysis of samples identified a of UtL microvesicle and an average of proteins per sample To a signature of of proteins with and we tested three vector machine and The in a validation to To the on the we tested the of of in from to and and the of of sensitivity were with a 9-protein of were in the HGOC and that were in controls and showed that of the signature proteins and were = the and samples in the discovery This signature sensitivity at a specificity of and an of in the discovery The for the signature proteins The of of the signature proteins were this signature correctly all three stage HGOC cases in the discovery of of the proteins from samples better than stage HGOC samples from controls the potential of this signature in of early-stage lesions. to for of we the and in the prediction the machine identified age as a of the potential were not in the the HGOC on than the this is potential to the signature in of the that not the signature the 9-protein in a We to validate the of the proteomic signature on the validation of UtL samples (n = This sample in to of a of and an average of per sample. of the 9-protein classifier to the validation of the controls and patients sensitivity at specificity for the validation showed an of of an of as by the 9-protein because of the of early stage the in cases validation UtL samples from HGOC patients that the samples were the samples from HGOC patients at the of the the classifier 70% sensitivity and 76.2% specificity for diagnosis of the of CA125 and HE4 as markers, we with our signature predictive The of a of the and the biomarkers and in the discovery and validation sensitivity and specificity were with the 9-protein signature sensitivity of and and specificity of and in the discovery and validation The of even with sensitivity of and and specificity of and in the discovery and validation the age with the and most HGOC patients were we tested age and the signature not for all we divided the into age = age analysis of the signature with age = for of with because diagnosis of HGOC and correlates with menopause = the diagnosis with the signature prediction = the of the signature proteins not with with age = The UtL liquid samples proteins that are not by the cancer to tumor or from an in tumor markers an in mass of a and are not by markers are expected to at an and may not be for early cancer whereas may the sensitivity of signature and the of the disease at an early stage. We therefore the of the signature proteins at the by and the by We the of all signature in HGOC on an of HGOC (n = and from (n = results in with the proteomic analysis of of the The that not all are that proteins and may from the in the of microvesicle proteins and the used and methods lack of the that of the To the of signature proteins in HGOC and and the by tumor we for and proteins that were in UtL samples of HGOC on a of HGOC from women of the not the fallopian or carriers is an of the that is its in it been as cancer and a in cancer M.J. I. L. S. The of for cancer with tumor 2017; PubMed Scopus Google Scholar). been a as an I. Y. A. J. G. S. The and of tumor as an 2017; PubMed Scopus Google Scholar). with the proteomic in than of and a in a of HGOC = is a of the to be in the of and of H. Y. C. A. F. J. S. Y. T. Liu of in by and of Full Text Full Text PDF PubMed Scopus Google Scholar). it localized to the of with low in with the proteomic its in HGOC tumor with the = R. Drapkin R. the of ovarian cancer PubMed Scopus Google Scholar). We from the M. L. C. P. A. A. C. E. A. I. K. E. S. C. J. D. S. T. H. H. J. J. P. J.M. M. K. M. L. F. M. G. J. F. of the 2015; PubMed Scopus Google for the of three to in the and in HGOC tumor than in the results the of the signature proteins in HGOC with and to tumor this we the discovery of potential early markers for using microvesicle proteomics of UtL liquid of the dynamic range of this body fluid, and of of proteins per sample in to our of proteomics M. P. O. Y. T. of with a novel proteomic method for 2015; Full Text Full Text PDF PubMed Scopus Google in the we used than because we not a for this body The that all potential even in the analysis of the discovery and validation predictive of the signature We that this is the of the signature proteins into a and that such a assays and are expected to the and with used Further, the and because of the high specificity of the results M. T. proteomics and its to cancer J. 2017; Full Text Full Text PDF PubMed Scopus Google Scholar). to the of the assays have the sample and the and with E. A.M. L. Smith D. K.D. Anderson Skates S.J. of to multiple mass for proteins in Full Text Full Text PDF PubMed Scopus Google Scholar, G. D. M. to and Full Text Full Text PDF PubMed Scopus Google Scholar, Y. A. O. Andersen J. H. A and serum proteomics of the of with PubMed Scopus Google Scholar, M. S.J. K. van I. proteomics sample for mass PubMed Scopus Google Scholar). of our with such to of the signature proteins to Early diagnosis of HGOC is of to women with because are currently counseled to undergo RRBSO the age of the and the age of This from the for early-stage diagnosis and the high mortality extremely mutation that the of and HGOC is than E. D. N. M. R. M. S. J. L. M. A. Kurman R. R. Drapkin R. ovarian in the fallopian 2017; PubMed Scopus Google that early be UtL liquid biopsy, as to may localized HGOC are 9-protein classifier 70% sensitivity and specificity results of biomarkers based on gynecological liquid E. Y. N. R. E. I. F. J. I. N. B. P. R. of the for of A Oncol. 2015; PubMed Scopus Google Scholar, I. Y. R.D. M.G. N. B. of in of patients with high-grade ovarian Gynecol. 2014; PubMed Scopus Google Scholar, I. C. Y. Wu J. N. Kurman R. F. Levine D.A. R. R. N. B. of DNA from the to ovarian and 2013; Scopus Google Scholar). mutation analysis in UtL samples at a of cancer proteomics the of a in multiple within the tumor a of biomarkers. of the proteomic signature and its predictive power analysis of early-stage HGOC UtL samples or samples are with the liquid this for early of The UtL that we is a of the method E. Y. N. R. E. I. F. J. I. N. B. P. R. of the for of A Oncol. 2015; PubMed Scopus Google it for testing of women at high for women have not for such as and of the population to undergo UtL need to be with proteomic assays may be as a of for high-risk to RRBSO and To the specificity of the signature proteins to HGOC we in and using and We results for proteins and results for of the the of proteins in HGOC These results the potential of the proteomic signature as a the proteomic and validation results may from the and and the and of it is that the of proteins is not into because are not in the may be for of the within HGOC lesions. the and the proteomic as well as of liquid may be to a classifier with an sensitivity and specificity to early of HGOC in both and high-risk and enable personalized and of RRBSO in women without HGOC mass of all the samples are with for based are in the We the of the Tumor at for samples and the of the of for We for with analysis. We and for with high-grade ovarian cancer risk-reducing bilateral salpingo-oophorectomy utero-tubal lavage