全文

1. Introduction
Ovarian cancer is the most lethal of all common gynecologic malignancies, with more than 204,000 new cases and 125,000 deaths each year, accounting for 4% of all cancer cases and 4.2% of all cancer deaths in women around the world [1]. In Switzerland, recent statistical data from the Swiss Association of Cancer Registries (www.nicer.org) showed that ovarian cancer is the seventh most common cancer and the fifth cause of death from cancer in Swiss women.

Contributing to the poor prognosis of ovarian cancer is the lack of symptoms in the early stages of the disease. More than 70% of the women are diagnosed with late stage disease [International Federation of Gynecology and Obstetrics (FIGO) stage III or IV], after distant metastasis has occurred. The 5-year survival rate for women diagnosed with late stage disease is less than 20% even with extensive surgery and chemotherapy, compared to up to 90% for women diagnosed with early stage disease [2]. Therefore, detection of ovarian cancer at an early stage is critical for curative treatment interventions. Unfortunately, current diagnosis methods for the detection of early stage ovarian cancer are inadequate. Only 25% of all ovarian cancer is found at early stage [2] and [3]. CA-125, which is significantly associated with ovarian cancer, is the only serum molecule now normally used in the clinical practice. While CA-125 serum levels are increased in about 80–85% of women with advanced ovarian cancer, only 50% of patients with stage I ovarian cancer will have an elevated CA-125 level. Therefore, CA-125 is mostly considered as a useful biomarker for follow-up (e.g., monitoring of progression and regression) of patients with established ovarian cancer, but has neither sufficient sensitivity nor specificity for early detection [3].

The ideal ovarian cancer biomarker would be utilized in a widespread screening process and would enable clinicians to diagnose asymptomatic women. To be useful as clinical tests, diagnostic biomarkers must be sufficiently noninvasive and inexpensive to allow widespread applicability. A substance secreted by tumor tissues, and not secreted by normal tissues, or an tumor specific immune marker easily detectable in a body fluid are therefore examples of ideal biomarkers [4]. Currently, very few ovarian cancer biomarkers have high sensitivity for early-stage disease. Given the prevalence of ovarian cancer, strategies for early detection must have high sensitivity for early stage disease (>75%), as well as an extremely high specificity (99.6%) to attain a positive predictive value of at least 10% [2] and [3]. Such high specificity will not likely be met by use of a single screening test alone, and cannot yet be met with any existing screening paradigm. Thus, discovery of novel ovarian cancer specific molecular biomarkers/panels is emerging as an important platform toward early detection. The present review summarizes various types of ovarian cancer markers investigated at present, including gene-, protein-based and emerging ovarian cancer biomarkers (such as microRNA-, metabolite-based).

2. Gene-based ovarian cancer biomarkers
Cancers are thought to arise from genetic alterations, environmental factors and a combination of both. Malignant transformation of normal ovarian epithelial cells is caused by genetic alterations that disrupt regulation of proliferation, programmed cell death and senescence. The vast majority of ovarian tumors arise due to accumulation of genetic damage, but the specific genetic pathways for the development of epithelial ovarian tumors, borderline and malignant, are largely unknown. Considering that a close connection exists between genetic changes and ovarian tumorigenesis, it is obvious that research on gene level (including studies of inherited gene mutations, epigenetic changes and gene expression) would also provide potential ovarian cancer biomarkers. DNA- or RNA-based cancer biomarkers utilize microarrays, polymerase chain reaction (PCR), reverse transcriptase polymerase chain reaction (RT-PCR), DNA sequencing, fluorescent in situ hybridization (FISH) etc. to detect the genetic alterations occurring in the cancerous state.

2.1. Inherited gene mutations
At least 10% of all epithelial ovarian cancers (EOCs) are hereditary, with germline mutations of the breast cancer 1/2 (BRCA1 and BRCA2) tumor suppressor genes accounting for approximately 90% of cases. Most of the remaining 10% are caused by germline mutations of the DNA mismatch repair (MMR) genes, primarily hMLH1 and hMSH2, which are susceptibility genes of Lynch syndrome [5] and [6]. Research showed that both BRCA proteins participate in transcriptional regulation of gene expression as well as the recognition or repair of certain forms of DNA damage, particularly double-strand breaks. Mutations of BRCA1 and BRCA2 are mainly of the frameshift or nonsense variety [6]. hMLH1 and hMSH2 genes are involved in the most important DNA repair mechanisms and are responsible for the repair of the nucleotide mismatch during DNA replication. Mutations of these MMR genes alter the ability of the cells to repair errors normally produced during DNA replication [5]. Currently, genetic testing for mutations of the above-mentioned genes can potentially identify patients at higher risk for developing ovarian cancer, and then clinical intervention strategies can be offered to dramatically reduce the risk of ovarian cancer

2.2. Epigenetic changes
Epigenetic mechanisms including DNA methylation and histone modifications are important means of gene regulation and play essential roles in tumor initiation and progression. Measurement of the methylation status of the promoter regions of specific genes can aid early detection of cancer, determine prognosis and predict therapy responses. The identification of specific genes that are altered by epigenetic events is currently under active investigation in ovarian cancer. For example, Tumor-specific hypermethylation of at least one of a panel of six tumor suppressor gene promoters, including BRCA1, RAS association domain family protein 1A (RASSF1A), adenomatous polyposis coli (APC), p14ARF, p16INK4a, and death associated protein-kinase (DAPKinase), was found in tumor DNA obtained from 50 patients with ovarian or primary peritoneal tumors. An identical pattern of gene hypermethylation was detected in the matched serum DNA from 41 of 50 patients (82% sensitivity), including 13 of 17 cases of stage I disease. In contrast, no hypermethylation was observed in nonneoplastic tissue or serum from 40 control women (100% specificity) [7]. Most studies to date have focused on candidate gene approaches to identify hypermethylated and silenced candidate tumor suppressor genes, but there is also a growing literature on specific regions of hypomethylation in ovarian cancer. Moreover, epigenetic markers can be assayed in circulating DNA of the blood, which provides the promise of a non-invasive test [7].

2.3. Gene expression
Quantitative or semi-quantitative measurement of the expression of particular genes in serum or tumor tissue has the potential for helping tumor diagnosis. In the last decade, the field of gene expression has progressed rapidly due in large part to the development of microarray technology, which enables us to measure the expression of tens of thousands of genes in a given tissue sample through a single experiment. This high-throughput technology, when coupled with powerful data analysis software, allows to rapidly compare gene expression between normal and malignant cells and to identify genes that are differentially regulated during cancer development. Microarray data can also be used to categorize tumors on the basis of their transcriptional profile, which may provide important biological, diagnostic and prognostic information. The current state of knowledge about the potential clinical value of gene expression profiling in ovarian cancer is discussed, focusing on three main areas: distinguishing normal ovarian tissue from ovarian tumors, identifying different subtypes of ovarian cancer and identifying cancer likely to be responsive to therapy. In EOC, gene-expression profiling has been used to provide prognostic information, to predict response to first-line platinum-based chemotherapy, and to discriminate between different histological subtypes [8]. In addition to microarray technology, serial analysis of gene expression (SAGE) represents another major class of technology currently available for the quantitative analysis of gene expression in ovarian cancer. SAGE facilitates the measurement of mRNA transcripts and generates a non-biased gene expression profile of normal and pathological disease tissue. Particularly, the SAGE technique has the capacity of detecting the expression of novel transcripts allowing for the identification of previously uncharacterized genes, thus providing a unique advantage over the traditional microarray-based approach for expression profiling. In ovarian cancer, several known and novel genes whose expressions are elevated have been identified by SAGE technology. These genes included claudin 3 (CLDN3) [9], WAP four-disulfide core domain 2 (WFDC2, also known as HE4) [9], folate receptor 1 (FOLR1) [9], collagen type XVIII a1 (COL18A1) [9], cyclin D1 (CCND1) [9], FLJ12988 [9].

3. Protein-based ovarian cancer biomarkers
Although gene-based biomarkers are known to have potential for ovarian cancer, there is still no novel cancer specific biomarker in clinic. This is due to the fact that gene levels are not always linked directly to levels of proteins, the molecules that biologically do functions. Proteomics has emerged as a powerful technology to decipher biological processes. It means large-scale characterization of proteins including more complicated features like isoforms, modifications, interactions and functional structures. One of the main goals of proteomics is the identification of biomarkers for diseases from tissues and body fluids. The major proteomics technique that fundamentally supported the discovery of cancer biomarkers is MS which can determine precise mass and charge of protein, thus identity of the actual precursor proteins or protein profiles. Among several different MS-based proteomics approaches, currently, matrix-assisted laser desorption and ionization time-of-flight (MALDI-TOF) and surface-enhanced laser desorption and ionization time-of-flight (SELDI-TOF) are two of the most frequently used methods for new biomarker discovery [10].

Proteomic applications to ovarian cancer diagnosis have followed two paths [11]: one, called “proteomic pattern diagnostics” or “serum proteomic profiling”, is based on complex mass spectrometric differences between proteomic patterns of samples with and without cancer identified by bioinformatics. Many previously published studies showed that proteomic pattern analysis in ovarian cancer has the potential to be a novel, highly sensitive diagnostic tool for detection at an early stage [12]. However, with the impressive results in terms of specificity and sensitivity in ovarian cancer detection, some criticism regarding instrument reproducibility, quality control and standard operating procedures for sample collection, handling and shipping have been raised. Recently researchers have emphasized more and more on the importance of reliability and reproducibility of a MS technology in protein profiling.

An alternative or integrative proteomic approach to ovarian cancer biomarkers is its use for the identification of single, novel biomarkers and the subsequent development of new assays [11]. In recent years many promising biomarkers discovered by proteomic analysis for ovarian cancer diagnosis were published [13], [14], and [15]. Among the markers identified by proteomic analysis, some biomarkers, such as cleavage fragment of inter-alpha-trypsin inhibitor heavy chain H4 [13], have often been normal serum proteins that have undergone posttranslational modification by proteases and reflect the protease profiles of particular cancers. Some biomarkers, such as transferrin [14], are acute phase proteins and have been associated with systemic inflammation as well as other non–cancer conditions. Other biomarkers, such as the vitamin E-binding plasma protein Afamin, had decreased serum concentrations in ovarian cancer patients and could contribute independent diagnostic information to CA-125, thus establishing their potential as an adjunct marker to CA-125 [15]. However, they are all not cancer-specific markers and derived directly from the ovarian cancers. So for proteomics-based biomarkers, their significance and degree of specificity for ovarian cancer remain to be explored. Recently, there are many proteins that have been studied in the search for EOC biomarkers. Of these proteins, mesothelin, osteopontin, and HE4 have been selected by the SPORE (Specialized Program of Research Excellence) committee for their high level of sensitivity and specificity in differentiating EOC from normal ovarian epithelium [16]. But the fact is that to date no single test or modality has met the criteria (positive predictive value of 10%) for early diagnosis of ovarian cancer [2].

4. Emerging ovarian cancer biomarkers
Following biomarker discovery on gene and protein level, recently two new fields are receiving increased attention in biomarker research of cancer, including ovarian cancer: analysis of the miRNAome and of the metabolome.

4.1. MicroRNA-based ovarian cancer biomarkers
MicroRNAs (miRNAs) are approximately 22 nt non-coding RNAs, which regulate gene expression in a sequence-specific manner via translational inhibition or messenger RNA (mRNA) degradation, and thus regulate diverse biological processes including development, cell proliferation, differentiation and apoptosis. About 3% of human genes encode for miRNAs, and up to 30% of human protein coding genes may be regulated by miRNAs, unique to each cell type and to the development and differentiation stage of the cell. Accumulating evidence has revealed aberrant expression of miRNAs in cancer including ovarian cancer, suggesting that they may act as a novel class of oncogenes or tumor-suppressor genes. Given the critical pathogenic roles of miRNAs in cancer progression, characterizing the regulation of miRNAs will provide novel opportunities for the development of cancer biomarkers and/or the identification of new therapeutic targets in the foreseeable future. Recently, development of dedicated microarrays has made it possible to analyze miRNA expression profiles in different oncotypes. Because miRNA expression profiles parallel the developmental origins of tissues, and because relatively few miRNAs can be used to effectively type tissues, they are potentially superior markers than messenger RNAs for cancer diagnosis and classification [17]. In the last 5 years several miRNA expression profiles of EOC have been published, reporting a decreased expression of a substantial proportion of miRNAs as compared to normal counterpart [18]. Recently, by using a custom microarray platform to compare miRNA profiles between 69 EOC surgical specimens and 15 normal ovaries, 29 differentially expressed miRNAs were found. Among them, miR-200a, miR-200b, miR-200c and miR-141 have been shown to be overexpressed. On the other hand, miR-199a, miR-140, miR-145 and miR-125b1 were among the most down-modulated miRNAs. In addition, it is believed that miRNA signatures of ovarian tumors may also distinguish these tumors based on their histologic subtypes and low- and high-grade malignancies [18].

One aspect of miRNA biogenesis that makes them particularly attractive as a biomarker is the fact that they are maintained in a protected state in serum and plasma, thus allowing the detection of miRNA expression patterns directly from serum. Recent work founded that the miRNA profiles of circulating tumor exosomes from EOC patients closely related with miRNA expression in primary tumors and could be used to distinguish cancer patients from patients with benign ovarian disease and from normal controls, thus having potential to be diagnostic markers of ovarian cancer. In this work, circulating tumor exosomes were isolated from serum using magnetic beads and an antiEpCAM antibody, and then miRNAs were extracted, labeled and detected by microarray. The results indicated that eight diagnostic miRNAs, including miR-21, miR-141, miR-200a, miR-200c, miR-200b, miR-203, miR-205, and miR-214 were up-regulated in ovarian cancer exosomes [19]. More recently, a novel real-time PCR platform was used to detect serum miRNA, founding miR-21, miR-92, miR-93, miR-126, and miR-29a were up-regulated, while miR-155, miR-127, and miR-99b were down-regulated in serum collected from ovarian carcinoma patients compared to healthy controls. Up-regulation of miR-21, miR-92, and miR-93 in the serum of three cancer patients with normal CA-125 level suggests that miRNA may be complementary to current detection approaches [20].

Overall, supported by a growing number of findings, it has become clear that miRNAs play key roles in both normal and pathologic ovarian activities by targeting the expression of specific genes. However, until now a clear consensus on miRNA signatures associated to diagnosis, prognosis or prediction of response to therapy has not yet been reached in the case of EOC. A greater understanding of the role of miRNAs in ovarian cancer is needed and will allow for improved interventions against this devastating malignancy.

4.2. Metabolite-based ovarian cancer biomarkers
Metabolomics, an omic science in systems biology, is the global quantitative assessment of endogenous metabolites within a biological system. Metabolites result from the interaction of the system's genome with its environment; they are not merely the end product of gene expression, but form part of the regulatory system in an integrated manner. Either individually or grouped as a metabolomic profile, detection of metabolites is usually carried out in cells, tissues, or biofluids by either nuclear magnetic resonance (NMR) spectroscopy or MS. With the development of metabolic and molecular imaging technologies which enable the discrimination of metabolic markers noninvasively in vivo, metabolomics, as a translational research tool, can provide a link between the laboratory and clinic. It is also possible for the metabolome to have a multitude of uses in oncology, including the early detection and diagnosis of cancer, monitoring drug treatment response and drug toxicity [21]. In the area of ovarian cancer diagnosis, 1H NMR spectroscopy was done on serum specimens of 38 preoperative EOC patients, 12 patients with benign ovarian cysts and 51 healthy women including 32 postmenopausal and 19 pre-menopausal. The results showed that 100% sensitivity and 100% specificity for the detection of EOC at the 1H NMR regions 2.77 and 2.04 parts per million (ppm) from the origin. These findings indicated that 1H NMR metabonomic analysis of serum achieves complete separation of EOC patients from healthy controls and deserves further evaluation as a potential novel strategy for the early detection of EOC [22]. In another study, gas chromatography/time-of-flight mass spectrometry (GC–TOF MS) was used to analyze metabolite profiling of fresh frozen tumor samples from 66 invasive ovarian carcinomas and 9 borderline tumors of the ovary, showing that a statistically significant differentiation between borderline tumors and carcinomas as reflected by differences in 51 metabolites. This study indicated there is a consistent and significant change in primary metabolism of ovarian tumors, which can be detected using large-scale metabolic profiling [23].

These limited available data are encouraging and show that the potential utility of metabolomics in ovarian cancer diagnosis, but metabolomics is still in its infancy. For the future development and application of metabolomics, it will be important to prompt a full integration of metabolomics into the context of cancer research for entire analyses of molecular changes in malignant tumors.

5. Ovarian cancer biomarker panels
Given the complexity and heterogeneity of ovarian cancer, it is unlikely that a single biomarker will be able to detect all subtypes and stages of the disease with a high specificity and a high sensitivity. Many current studies show that combining several biomarkers dramatically improves sensitivity of CA-125 in ovarian cancer patients [24]. Markers have generally been analyzed only 2 or 3 at a time. The increased sensitivity achieved with markers in combination has generally been associated with a marked decrease in specificity [3]. A couple of biomarker panels have been published with adaptable sensitivity and specificity range, which might hold great potential for the detection of ovarian cancer [4]. For example, recently, a novel multiplex assay that used a panel of six serum biomarkers: leptin, prolactin, osteopontin, insulin-like growth factor II (IGF-II), macrophage inhibitory factor (MIF) and CA-125 and was studied on 362 healthy controls and 156 patients with newly diagnosed ovarian cancer (including 13 stage I cases), yielded 95.3% sensitivity and 99.4% specificity [25]. However, these data generated much controversy about experimental design and statistical analysis. Most of impressive sensitivities and specificities for biomarker panels arose from relatively small numbers of samples (especially few cases of stage I diseases) without an independent validation study. So before biomarker tests are translated for routine use, more researches, such as retrospective and prospective clinical trials, are needed to evaluate the overall clinical utility of the tests. In the future, it will still be crucial to further develop panels of biomarkers not only for early detection but also for treatment guidance of ovarian cancer.

6. Conclusion
During the last decade, with the development of high-throughput technologies in genomics and proteomics, a number of biomarkers, some part of which were listed in Table 1, have shown promises across a variety of ovarian cancer studies and also provided new insights into ovarian cancer diagnosis, but few have turned out to be useful in clinic. It remains unclear, whether a single biomarker, a panel of biomarkers, or multiplexed information will yield the most accurate approaches to ovarian cancer detection. The strategies or technologies mentioned in this review hold significant promise in discovering more robust biomarkers for diagnosis, prognosis or prediction of therapy in ovarian cancer. At present, the research on ovarian cancer biomarkers is still under way in three main aspects: One is further validation and the ongoing clinical trials of available or potential biomarkers. Another is investigation of novel more specific and sensitive ovarian cancer biomarkers with further improved technologies on different biological levels. The third is development of multiple biomarkers for generating panels to maximize the sensitivity and specificity of detection. In the future, through effective integration of various more advanced technologies and help of bioinformatics, more useful biomarkers for ovarian cancer diagnosis are likely to emerge. Furthermore, sharing of information among the scientific community will quicken the pace in the field of biomarker research from different angles.

Conflict of interest statement
No potential conflicts of interest were disclosed

Acknowledgments
We thank Mr. Reza Asadollahi, Mr. Lei Fang, Ms. Zeinab Barekati, Mrs. Corina Kohler, Mr. Ramin Radpour, Mrs. Hong Bo Chen, and Mrs. Vivian Kiefer for their kind support.

References
[1] D.M. Parkin, F. Bray, J. Ferlay, P. Pisani. Global cancer statistics, 2002. CA Cancer J Clin. 2005;55:74-108 Crossref.
[2] R.C. Bast Jr., D. Badgwell, Z. Lu, et al., New tumor markers: CA125 beyond. Int J Gynecol Cancer. 2005;15(3):274-281 Crossref.
[3] R.C. Bast Jr. Early detection of ovarian cancer: new technologies in pursuit of a disease that is neither common nor rare. Trans Am Clin Climatol Assoc. 2004;115:233-247 discussion 47–48
[4] A. Gagnon, B. Ye. Discovery and application of protein biomarkers for ovarian cancer. Curr Opin Obstet Gynecol. 2008;20:9-13 Crossref.
[5] A. Russo, V. Calo, L. Bruno, S. Rizzo, V. Bazan, G. Di Fede. Hereditary ovarian cancer. Crit Rev Oncol Hematol. 2009;69:28-44 Crossref.
[6] J. Prat, A. Ribe, A. Gallardo. Hereditary ovarian cancer. Hum Pathol. 2005;36:861-870 Crossref.
[7] I. Ibanez de Caceres, C. Battagli, M. Esteller, et al. Tumor cell-specific BRCA1 and RASSF1A hypermethylation in serum, plasma, and peritoneal fluid from ovarian cancer patients. Cancer Res. 2004;64:6476-6481
[8] P.A. Konstantinopoulos, D. Spentzos, S.A. Cannistra. Gene-expression profiling in epithelial ovarian cancer. Nat Clin Pract Oncol. 2008;5:577-587 Crossref.
[9] D.G. Peters, D.M. Kudla, J.A. Deloia, et al. Comparative gene expression analysis of ovarian carcinoma and normal ovarian epithelium by serial analysis of gene expression. Cancer Epidemiol Biomarkers Prev. 2005;14:1717-1723 Crossref.
[10] B. Zhang, Z. Barekati, C. Kohler, et al. Proteomics and biomarkers for ovarian cancer diagnosis. Ann Clin Lab Sci. 2010;40:218-225
[11] M. Plebani. Proteomics: the next revolution in laboratory medicine?. Clin Chim Acta. 2005;357:113-122 Crossref.
[12] E.F. Petricoin, A.M. Ardekani, B.A. Hitt, et al. Use of proteomic patterns in serum to identify ovarian cancer. Lancet. 2002;359:572-577 Crossref.
[13] Z. Zhang, R.C. Bast Jr., Y. Yu, et al. Three biomarkers identified from serum proteomic analysis for the detection of early stage ovarian cancer. Cancer Res. 2004;64:5882-5890 Crossref.
[14] N. Ahmed, K.T. Oliva, G. Barker, et al. Proteomic tracking of serum protein isoforms as screening biomarkers of ovarian cancer. Proteomics. 2005;5:4625-4636 Crossref.
[15] H. Dieplinger, D.P. Ankerst, A. Burges, et al. Afamin and apolipoprotein A-IV: novel protein markers for ovarian cancer. Cancer Epidemiol Biomarkers Prev. 2009;18:1127-1133 Crossref.
[16] R.G. Moore, S. Maclaughlan. Current clinical use of biomarkers for epithelial ovarian cancer. Curr Opin Oncol. 2010;22:492-497 Crossref.
[17] J. Lu, G. Getz, E.A. Miska, et al. MicroRNA expression profiles classify human cancers. Nature. 2005;435:834-838 Crossref.
[18] M.V. Iorio, R. Visone, G. Di Leva, et al. MicroRNA signatures in human ovarian cancer. Cancer Res. 2007;67:8699-8707 Crossref.
[19] D.D. Taylor, C. Gercel-Taylor. MicroRNA signatures of tumor-derived exosomes as diagnostic biomarkers of ovarian cancer. Gynecol Oncol. 2008;110:13-21 Crossref.
[20] K.E. Resnick, H. Alder, J.P. Hagan, D.L. Richardson, C.M. Croce, D.E. Cohn. The detection of differentially expressed microRNAs from the serum of ovarian cancer patients using a novel real-time PCR platform. Gynecol Oncol. 2009;112:55-59 Crossref.
[21] J.L. Spratlin, N.J. Serkova, S.G. Eckhardt. Clinical applications of metabolomics in oncology: a review. Clin Cancer Res. 2009;15:431-440 Crossref.
[22] K. Odunsi, R.M. Wollman, C.B. Ambrosone, et al. Detection of epithelial ovarian cancer using 1H NMR-based metabonomics. Int J Cancer. 2005;113:782-788 Crossref.
[23] C. Denkert, J. Budczies, T. Kind, et al. Mass spectrometry-based metabolic profiling reveals different metabolite patterns in invasive ovarian carcinomas and ovarian borderline tumors. Cancer Res. 2006;66:10795-10804 Crossref.
[24] Z.R. Yurkovetsky, F.Y. Linkov, D. E.M., A.E. Lokshin. Multiple biomarker panels for early detection of ovarian cancer. Future Oncol. 2006;2:733-741 Crossref.
[25] K. Kim, I. Visintin, A.B. Alvero, G. Mor. Development and validation of a protein-based signature for the detection of ovarian cancer. Clin Lab Med. 2009;29:47-55 Crossref.

Footnotes
Laboratory for Gynecologic Oncology, Women's Hospital/Department of Biomedicine, University of Basel, Basel, Switzerland/Hebelstrasse 20, CH 4031 Basel, Switzerland

 Corresponding author. Tel.: +41 61 3286986/2659248; fax: +41 61 2659399.