Title: Ovarian Cancer Biomarkers

Field of the invention

[0001] The present application relates to materials and methods for diagnosing and detecting ovarian cancer.

Background of the invention [0002] Epithelial ovarian carcinoma accounts for over 15,000 deaths per year in the United States of America ¹ and over 100,000 worldwide ² making it the most lethal gynecological malignancy. One of the most common symptoms of the disease is the accumulation of ascites fluid in the abdominal cavity. The mechanisms by which ascites fluid is formed are complex and consist of lymphatic obstruction, activation of mesothelial cells by the malignant metastatic process and increased vascular permeability derived by the secretion of vascular endothelial growth factors and other cytokines ^3-7 .

[0003] Ascites fluid represents the local microenvironment that is secreted by ovarian tumors and contains various cell types including malignant cells shed from the tumor, as well as soluble growth factors ^8-10 . There is evidence that malignant ascites can stimulate the growth and invasive behavior of ovarian cancer cells both in vivo and in vitro ^u and therefore, understanding the exact protein content of human ascites may provide critical information regarding ovarian tumor growth and progression. [0004] Proteomic analysis of ascites can serve as a valuable tool in the ongoing search for biomarkers of ovarian carcinoma with potential use in prognosis, chemotherapy resistance, treatment surveillance, and early detection.

[0005] Although many attempts have been made to provide molecular insight into the content of ascites, no in-depth analysis has been reported to date. Xu et al. characterized an ovarian cancer activating factor in ascites from ovarian cancer patients ¹⁴ . The authors used mass spectrometry (MS) to identify an activating factor, purified from ascites of diseased subjects. The results demonstrated that a major component of ovarian cancer activating

factor was palmitoyl lysophosphatidic acid. In another study, Gericke et al. ¹⁵ used MALDI-TOF-MS to characterize differences of transthyretin between ascitic fluid and plasma of women affected with ovarian cancer.

Summary of the invention [0006] The inventors have identified a number of biomarkers that are associated with ovarian cancer. These biomarkers, which are listed in Table 5, 6 and/or 7, can be used to detect or diagnose ovarian cancer in a subject. For example, the expression products of one or more of the biomarkers shown in Table 5, 6 and/or 7 can be used to detect or diagnose ovarian cancer in a subject. The expression products include RNA products and protein products corresponding to the biomarkers.

[0007] Accordingly, one aspect of the present application is a method of diagnosing ovarian cancer in a subject, the method comprising the steps:

(a) determining the level of an expression product of a biomarker in a test sample from a subject, wherein the biomarker comprises one or more of the biomarkers shown in Table 5, 6 and/or 7; and

(b) comparing the level of the expression product of the biomarker with a control,

wherein the difference in the level of expression product between the control and the test sample is indicative of ovarian cancer.

[0008] Another aspect of the present application is a method of diagnosing ovarian cancer in a subject, the method comprising the steps:

(a) determining the level of an expression product of a biomarker in a test sample from a subject, wherein the biomarker is selected from the group consisting of ARHGDIB, CFL1 , PFN1 , GSTP1 , S100A11 , PRDX6,

HSPE1 , LYZ, GPI, HIST2H2AA, GAPDH, HSPG2, LGALS3BP, CTSD, APOE,

IQGAPI ₁ CP, and IGLC2; and

(b) comparing the level of the expression product of the biomarker with a control, wherein the difference in the level of expression product between the control and the test sample is indicative of ovarian cancer. [0009] The method of the present application can also be used to manage the treatment of ovarian cancer and can be used to monitor the efficacy of ovarian cancer therapy.

[0010] The present application also provides kits for diagnosing ovarian cancer that include agents that can detect the expression products of a biomarker in a test sample.

[0011] Other features and advantages of the present application will become apparent from the following detailed description. It should be understood, however, that the detailed description and the specific examples while indicating preferred embodiments of the present application are given by way of illustration only, since various changes and modifications within the spirit and scope of the present application will become apparent to those skilled in the art from this detailed description.

Brief description of the drawings

[0012] The present application will now be described in relation to the drawings in which:

[0013] Figure 1 shows an analysis scheme of human ovarian cancer ascites. Approximately 1ml of ovarian cancer ascites was centrifuged to separate the soluble (S1) from the cellular fraction. The cellular fraction was consecutively extracted with a hypotonic (S2) and detergent (S3) extraction buffer. Each fraction was extensively analyzed by MudPIT and gel-enhanced LC-MS/MS (for patient 1 ; core resource). Integrated bioinformatics analysis was used to map the ascites proteome data of human plasma ¹³ and urine ¹² , as well as 59 available ovarian cancer microarrays. A limited analysis of the cellular fractions of three additional ovarian cancer ascites samples was used to confirm and prioritize biomarkers.

[0014] Figure 2 shows a schematic of the MudPIT technology.

Microcapillary chromatography columns (inner diameter 100 μm) are packed with two orthogonal resins. The first dimension consists of a strong cation exchange (SCX) resin and the second dimension of a reversed phase resin (RP). Peptide mixtures are bound to the SCX resin and columns are placed in line with a HPLC pump (high performance liquid chromatography) and directly ionized (ESI; electrospray ionization) into the mass spectrometer (MS). Collected mass spectra are searched against public protein sequence databases using powerful computer clusters. [0015] Figure 3 shows a proteomics identification strategy and number of detected proteins. (A) MS data obtained from individual ascites fractions was searched by X!Tandem and peptides passing a default log-expectation value of ≤-1 were parsed into an in-house MySQL database. A developed filter algorithm was used to calculate specific log-expectation values for individual charge state (+1 , +2, +3) of assigned peptide ions, to minimize the number of reverse peptide sequences to a user defined percentage. Only proteins identified with at least two peptide sequences were further considered for identification. A developed grouping algorithm was used to minimize the protein interference problem and only report proteins with informative peptide evidence. (B) Different user defined filter criteria are presented. The number of reverse proteins in the final dataset is blotted against the percent of total reverse/forward spectra present after applying the filter algorithm, a cut-off value of 0.2% chosen for this project. (C) Venn diagram of proteins confidently identified in each of the three biochemical fractions. (D) A comparison of proteins detected by MudPIT or gel-enhanced LC-MS/MS.

[0016] Figure 4 shows significantly enriched Gene Ontology terms

{biological process). Distribution of significantly enriched GO-terms in individual ascites fractions. Numbers were calculated by the GOFFA tool and only significant categories are reported (P-value < 0.01 , E-value > 2).

[0017] Figure 5 shows an integrated bioinformatics strategy to identify biomarker candidates. (A) Venn diagram of proteins detected in ascites, the HUPO plasma proteome ¹³ and a high quality human urine dataset ¹² . # Overlap to the initial HUPO plasma proteome ¹³ . * Overlap to the more rigorous plasma proteome dataset of States et al. ⁹¹ . (B) Proteins from the ascites resource (patient 1) were first mapped to available urine and plasma proteomes. This strategy resulted in 632 proteins found in previously published, body fluid datasets (secreted proteins), and 1646 proteins not detectable in these healthy plasma and urine resources. Next, both of these lists were mapped against 59 available ovarian cancer microarrays and only proteins showing significant changes on at least one microarray were further evaluated. As a third arm of this integrated data analysis strategy the OPHID protein interaction database was used and identified known or predicted protein interaction partners for the 72 secreted proteins also significantly changed by microarray. Finally, 3 more ovarian cancer ascites were analyzed and only considered proteins passing the above criteria, also found by mass spectrometry in all four ovarian cancer patients. This resulted in a panel of 80 ovarian cancer biomarkers for future investigation.

[0018] Figure 6 shows the visualization of 18 biomarkers. (A) These proteins were detected in all four ovarian cancer ascites and were previously reported in human body fluid proteomes (secreted proteins). Mapping to 59 ovarian cancer microarray datasets revealed significant changes on several arrays for all of these candidates. (B) Representative Western blots for candidate markers in ascites fraction S1. The ovarian cancer cell line ES-2 was used as a positive control. Different concentrations of ascites fluid (3-50 μg) were loaded on each gel.

[0019] Figure 7 shows a protein-protein interaction network for the 18 biomarkers in Figure 6. The network was generated by querying the I ² D database for the 18 proteins, which resulted in a network with 201 proteins and 201 interactions. Importantly, while all 18 have been significant on

multiple microarray studies (ovals), we highlighted those that are the most frequent (underline). We have also highlighted subnetworks that are significantly up-regulated (shaded areas). Proteins in the network highlighted as diamonds are important interconnecting proteins, and we also show their up- or down- regulation on multiple microarray studies by arrows. The color represents Gene Ontology function.

[0020] Figure 8 shows an established mouse model of ovarian cancer

(OVCA). In this model, epithelial cancer cells, which spontaneously transformed in vitro from surface ovarian epithelial cells, are transplanted under the ovarian bursa and mice develop all symptoms of ovarian cancer within 2-3 months. The epithelial cancer cell lines used for these experiments were IG 10 and IC5 ⁹² .

[0021] Figure 9 shows the expression of profilin-1 and cofilin-1 in a mouse model of OVCA. Two biomarkers (profilin-1 (PRF 1) and cofilin-1 (CFL1) were selected for further validation in a mouse model of ovarian cancer. Expression of PRF1 and CFL1 measured via Western blotting demonstrated upregulation of both biomarkers in ovarian cancer cells when compared to normal surface epithelial cells, both maintained in vitro. NOTE: IC5/a = IC5 ascites cells; and IG10/a = IG10 ascites cells. [0022] Figure 10 shows Western blot analysis of OVCA biomarkers in cell lines (mouse model). Western blotting demonstrated upregulation of profilin and cofilin biomarkers in ovarian cancer cells when compared to normal surface epithelial cells, both maintained in vitro.

[0023] Figure 11 shows the expression of profilin-1 in a mouse model of OVCA. Immunolocalization revealed elevated staining of profilin-1 in solid tumors that formed adjacent to the ovary, while ovarian surface epithelial cells in normal ovaries exhibited limited immunoreactivity.

[0024] Figure 12 shows the expression of cofilin-1 in a mouse model of

OVCA. Immunolocalization revealed elevated staining of cofilin-1 in solid

tumors which formed adjacent to the ovary, while ovarian surface epithelial cells in normal ovaries exhibited limited immunoreactivity.

[0025] Figure 13 shows results of the candidate biomarker profile.

Profilin-1 and cofilin-1 candidate biomarkers were identified from human ascites. Ascites fluid from tumor-bearing animals at the time of euthanasia exhibited upregulation of both biomarkers when compared to age matched healthy controls.

[0026] Figure 14 demonstrates the quantitation of OVCA biomarkers in serum. Western blot analysis demonstrated upregulation of cofilin-1 and profilin-1 in serum of tumor-bearing animals when compared to age matched healthy controls.

Detailed description of the invention

[0027] The present application relates to biomarkers which are associated with ovarian cancer. These biomarkers can be used to detect or diagnose ovarian cancer in a subject.

[0028] The term "biomarker" as used herein refers to a gene that is differentially expressed in individuals with ovarian cancer as compared to individuals who do not have ovarian cancer. The term "biomarker" includes one or more of the genes listed in Table 5, 6 and/or 7. The amino acid sequences of these biomarkers are known in the art. For example, the sequences may be viewed at the International Protein Index (IPI) database (http://www.ebi.ac.uk/IPI) via the IPI accession number or through Genbank or other databases. The IPI "entry names" for the biomarkers are listed in Tables 5, 6, and 7. The present application also includes variants of these biomarkers that are useful in detecting ovarian cancer.

[0029] The term "variant" as used herein includes modifications, derivatives, or chemical equivalents of the amino acid and nucleic acid sequences disclosed herein that perform substantially the same function as the polypeptides or nucleic acid molecules disclosed herein in substantially the same way. For instance, the variants have the same function of being

useful to detect ovarian cancer. In one embodiment, variants of polypeptides disclosed herein include, without limitation, conservative amino acid substitutions. Variants of polypeptides also include additions and deletions to the polypeptide sequences disclosed herein. In addition, variant nucleotide sequences and polypeptide sequences include analogs and derivatives thereof. A "conservative amino acid substitution" as used herein, is one in which one amino acid residue is replaced with another amino acid residue without abolishing the protein's desired properties.

[0030] One aspect of the present application is a method of diagnosing ovarian cancer in a subject, the method comprising the steps:

(a) determining the level of an expression product of a biomarker in a test sample from a subject, wherein the biomarker comprises one or more of the biomarkers shown in Table 5, 6 and/or 7; and

(b) comparing the level of the expression product of the biomarker with a control,

wherein the difference in the level of expression product between the control and the test sample is indicative of ovarian cancer. [0031] In a specific embodiment, the "biomarker" is selected from the group consisting of ARHGDIB (Rho GDP-dissociation inhibitor 2), CFL1 (Cofιlin-1), PFN1 (profilin-1), GSTP1 (Glutathione S-transferase P), S100A11 (Protein S100-A11), PRDX6 (Peroxiredoxin-6), HSPE1 (10 kDa heat shock protein, mitochondrial), LYZ (Lysozyme C precursor), GPI (Glucose-6- phosphate isomerase), HIST2H2AA (Histone H2A type 2-A), GAPDH (Glyceraldehyde-3-phosphate dehydrogenase), HSPG2 (Basement membrane-specific heparan sulfate proteoglycan core protein precursor), LGALS3BP (Galectin-3-binding protein precursor), CTSD (Cathepsin D precursor), APOE (Apolipoprotein E precursor), IQGAP1 (Ras GTPase- activating-like protein IQGAP1), CP (Ceruloplasmin precursor), and IGLC2 (IGLC1 protein) listed in Table 5. As noted above, the amino acid sequences of these biomarkers are known in the art and may be accessed through the

IPI database, Genbank or other databases. The IPI accession numbers for the following biomarkers are in brackets following identity of the biomarkers: ARHGDIB (IPI00003817), CFL1 (IPI00012011), PFN1 (IPI00216691), GSTP1 (IPI00219757), S100A11 (IPI00013895), PRDX6 (IPI00220301), HSPE1 (IPI00220362), LYZ (IPI00019038), GPI (IPI00027497), HIST2H2AA (IPI00216457), GAPDH (IPI00219018), HSPG2 (IPI00024284), LGALS3BP (IPI00023673), CTSD (IPI00011229), APOE (IPI00021842), IQGAP1 (IPI00009342), CP (IPI00017601), and IGLC2 (IPI00154742). The present application also includes variants of these biomarkers that are useful in detecting ovarian cancer.

[0032] Another aspect of the present application is a method of diagnosing ovarian cancer in a subject, the method comprising the steps:

(a) determining the level of an expression product of a biomarker in a test sample from a subject, wherein the biomarker is selected from the group consisting of ARHGDIB, CFL1 , PFN1 , GSTP1 , S100A11 , PRDX6, HSPE1 , LYZ, GPI, HIST2H2AA, GAPDH, HSPG2, LGALS3BP, CTSD, APOE, IQGAP1 , CP, and IGLC2; and

(b) comparing the level of the expression product of the biomarker with a control, wherein the difference in the level of expression product between the control and the test sample is indicative of ovarian cancer.

[0033] Another aspect of the present application is a method of diagnosing ovarian cancer in a subject, the method comprising the steps:

(a) determining the level of an expression product of a biomarker in a test sample from a subject, wherein the biomarker comprises CFL1 and/or

PFN1 ; and

(b) comparing the level of the expression product of the biomarker with a control,

wherein the difference in the level of expression product between the control and the test sample is indicative of ovarian cancer.

[0034] The term "subject" as used herein refers to any member of the animal kingdom, preferably a human being, that may have ovarian cancer. [0035] The term "test sample" as used herein refers to any fluid, cell or tissue sample from a subject which can be assayed for biomarker expression products, particularly genes differentially expressed in subjects with ovarian cancer. In one embodiment, the test sample ovarian tissue. In another embodiment, the test sample is an ascite from the peritoneal cavity of a subject. In a further embodiment, the test sample is the fluid fraction of the ascite or the cellular fraction of the ascite.

[0036] The term "control" refers to refers to a sample from a subject known to have ovarian cancer (positive control) or known not to have ovarian cancer (negative control). The control can be an actual sample from an individual or from a population of samples. The control can also be a predetermined threshold value for one or more of the biomarkers listed in Table 5, 6 and/or 7. In another embodiment, the controls can also be a predetermined threshold value for ARHGDIB ₁ CFL1 , PFN1 , GSTP1 , S100A11 , PRDX6, HSPE1, LYZ, GPI, HIST2H2AA, GAPDH, HSPG2, LGALS3BP, CTSD, APOE, IQGAP1 , CP, and/or IGLC2. In another embodiment, the control can also be a pre-determined threshold value for CFL1 or PFN 1.

[0037] A person skilled in the art will appreciate that the comparison between the expression of the biomarkers in the test sample and the expression of the biomarkers in the control will depend on the control used. For example, if the control is from a subject known to have ovarian cancer (positive control) and there are lower levels of expression of the biomarkers in the test sample as compared to the control, then this is not indicative of ovarian cancer in the test sample. If the control is from a subject known to have ovarian cancer (positive control) and there are similar levels of expression of the biomarkers in the test sample as compared to the control, then this is indicative of ovarian cancer in the test sample. If the control is

from a subject known to not to have ovarian cancer (negative control) and there are higher levels of expression of the biomarkers in the test sample as compared to the control, then this is indicative of ovarian cancer in the test sample. If the control is from a subject known to not to have ovarian cancer (negative control) and there are similar levels of expression of the biomarkers in the test sample as compared to the control, then this is not indicative of ovarian cancer in the test sample.

[0038] The term "differentially expressed" or "differential expression" as used herein refers to a difference in the level of expression of the biomarkers that can be assayed by measuring the level of expression of the products of the biomarkers, such as the difference in level of RNA expressed or proteins expressed of the biomarkers. In a preferred embodiment, the difference is statistically significant. The term "difference in the level of expression" refers to an increase or decrease in the measurable expression level of a given biomarker as measured by the amount of RNA and/or the amount of protein in a sample as compared with the measurable expression level of a given biomarker in a control. In one embodiment, the differential expression can be compared using the ratio of the level of expression of a given biomarker or biomarkers as compared with the expression level of the given biomarker or biomarkers of a control, wherein the ratio is not equal to 1.0. For example, an RNA or protein is differentially expressed if the ratio of the level of expression in a first sample as compared with a second sample is greater than or less than 1.0. For example, a ratio of greater than 1, 1.2, 1.5, 1.7, 2, 3, 3, 5, 10, 15, 20 or more, or a ratio less than 1 , 0.8, 0.6, 0.4, 0.2, 0.1 , 0.05, 0.001 or less. In another embodiment the differential expression is measured using p-value. For instance, when using p-value, a biomarker is identified as being differentially expressed as between a first sample and a second sample when the p-value is less than 0.1 , preferably less than 0.05, more preferably less than 0.01 , even more preferably less than 0.005, the most preferably less than 0.001.

[0039] The term "similarity in expression" as used herein means that there is no difference in the level of expression of the biomarkers between the test sample and the control.

[0040] The phrase "detecting or diagnosing" as used herein refers to a method or process of determining whether an individual has or does not have ovarian cancer or the extent of ovarian cancer.

[0041] The phrase "determining the expression of biomarkers" as used herein refers to determining or quantifying RNA or proteins expressed by the biomarkers. The term "RNA" includes mRNA transcripts, and/or specific spliced variants of mRNA. The term "RNA product of the biomarker" as used herein refers to RNA transcripts transcribed from the biomarkers and/or specific spliced variants. In the case of "protein", it refers to proteins translated from the RNA transcripts transcribed from the biomarkers. The term "protein product of the biomarker" refers to proteins translated from RNA products of the biomarkers.

[0042] In one embodiment, the level of expression of one or more biomarkers is used to diagnose ovarian cancer. In another embodiment, two or more biomarkers are used to diagnose ovarian cancer. In a further embodiment, ten or more, twenty or more, thirty or more, forty or more, fifty or more, sixty or more, seventy or more or all the biomarkers listed in Table 5, 6 and/or 7 are used to diagnose ovarian cancer.

[0043] In another embodiment, the level of expression of one or more biomarkers listed in Table 5 is used to diagnose ovarian cancer. In a further embodiment, two or more, three or more, four or more, five or more, six or more, seven or more, eight or more, nine or more, ten or more, eleven or more, twelve or more, thirteen or more, fourteen or more, fifteen or more, sixteen or more, seventeen or more, or all of the biomarkers listed in Table 5 are used to diagnose ovarian cancer.

[0044] In another embodiment, the level of expression of one or more biomarkers from ARHGDIB, CFL1 , PFN1 , GSTP1 , S100A11 , PRDX6,

HSPE1 , LYZ, GPI, HIST2H2AA, GAPDH, HSPG2, LGALS3BP, CTSD, APOE, IQGAP1 , CP, and/or IGLC2 is used to diagnose ovarian cancer. In a further embodiment, two or more, three or more, four or more, five or more, six or more, seven or more, eight or more, nine or more, ten or more, eleven or more, twelve or more, thirteen or more, fourteen or more, fifteen or more, sixteen or more, seventeen or more, or all of the biomarkers from ARHGDIB, CFL1 , PFN1 , GSTP1 , S100A11 , PRDX6, HSPE1 , LYZ, GPI, HIST2H2AA, GAPDH, HSPG2, LGALS3BP, CTSD, APOE, IQGAP1 , CP, and IGLC2 are used to diagnose ovarian cancer. [0045] A person skilled in the art will appreciate that a number of methods can be used to detect or quantify the level of RNA products of the biomarkers within a sample, including microarrays, RT-PCR (including quantitative RT-PCR), nuclease protection assays and northern blots.

[0046] In addition, a person skilled in the art will appreciate that a number of methods can be used to determine the amount of a protein product of the biomarker of the present application, including immunoassays such as Western blots, ELISA, and immunoprecipitation followed by SDS-PAGE and immunocytochemistry.

[0047] Conventional techniques of molecular biology, microbiology and recombinant DNA techniques, which are within the skill of the art. Such techniques are explained fully in the literature. See, e.g., Sambrook, Fritsch &

Maniatis, 1989, Molecular Cloning: A Laboratory Manual, Second Edition;

Oligonucleotide Synthesis (MJ. Gait, ed., 1984); Nucleic Acid Hybridization

(B.D. Harnes & S.J. Higgins, eds., 1984); A Practical Guide to Molecular Cloning (B. Perbal, 1984); and a series, Methods in Enzymology (Academic

Press, Inc.); Short Protocols In Molecular Biology, (Ausubel et al., ed., 1995).

[0048] The method of the present application can also be used in combination with traditional diagnostic techniques for ovarian cancer, which include physical examination and histology.

[0049] The method of the present application can also be used to manage the treatment of ovarian cancer and can be used to monitor the efficacy of ovarian cancer therapy.

[0050] The method of the present application can also be used to determine if a subject has acquired or will acquire chemoresistance.

[0051] The present application also provides kits for diagnosing ovarian cancer that include agents that can detect the expression products of a biomarker in a test sample.

[0052] Accordingly, the present application includes kits for diagnosing ovarian cancer, comprising detection agents that can detect the expression products of a biomarker in a test sample, wherein the biomarker comprises one or more of the biomarkers shown in Table 5, 6 and/or 7. The present application also includes kits wherein the biomarker comprises one or more of

ARHGDIB ₁ CFL1 , PFN1 , GSTP1 , S100A11 , PRDX6, HSPE1 , LYZ, GPI, HIST2H2AA, GAPDH, HSPG2, LGALS3BP, CTSD, APOE, IQGAP1 , CP, and

IGLC2. The present application also includes kits wherein the biomarker comprises CFL 1 and/or PFN 1.

[0053] A person skilled in the art will appreciate that a number of detection agents can be used to determine the expression of the biomarkers. For example, to detect RNA products of the biomarkers, probes, primers, complementary nucleotide sequences or nucleotide sequences that hybridize to the RNA products or nucleotide sequences complementary to the RNA products can be used. To detect protein products of the biomarkers, ligands or antibodies that specifically bind to the protein products can be used. [0054] The term "nucleic acid" includes DNA and RNA and can be either double stranded or single stranded.

[0055] The term "hybridize" refers to the sequence specific non- covalent binding interaction with a complementary nucleic acid. In a preferred embodiment, the hybridization is under high stringency conditions. Appropriate stringency conditions which promote hybridization are known to

those skilled in the art, or can be found in Current Protocols in Molecular Biology, John Wiley & Sons, N.Y. (1989), 6.3.1 6.3.6. For example, 6.0 x sodium chloride/sodium citrate (SSC) at about 45°C, followed by a wash of 2.0 x SSC at 5O°C may be employed. [0056] The term "primer" as used herein refers to a nucleic acid sequence, whether occurring naturally as in a purified restriction digest or produced synthetically, which is capable of acting as a point of synthesis of when placed under conditions in which synthesis of a primer extension product, which is complementary to a nucleic acid strand is induced (e.g. in the presence of nucleotides and an inducing agent such as DNA polymerase and at a suitable temperature and pH). The primer must be sufficiently long to prime the synthesis of the desired extension product in the presence of the inducing agent. The exact length of the primer will depend upon factors, including temperature, sequences of the primer and the methods used. A primer typically contains 15-25 or more nucleotides, although it can contain less. The factors involved in determining the appropriate length of primer are readily known to one of ordinary skill in the art. The term "primer" as used herein refers a set of primers which can produce a double stranded nucleic acid product complementary to a portion of the RNA products of the biomarker or sequences complementary thereof.

[0057] The term "probe" as used herein refers to a nucleic acid sequence that will hybridize to a nucleic acid target sequence. In one example, the probe hybridizes to an RNA product of the biomarker or a nucleic acid sequence complementary thereof. The length of probe depends on the hybridize conditions and the sequences of the probe and nucleic acid target sequence. In one embodiment, the probe is at least 8, 10, 15, 20, 25, 50, 75, 100, 150, 200, 250, 400, 500 or more nucleotides in length.

[0058] The term "antibody" as used herein is intended to include monoclonal antibodies, polyclonal antibodies, and chimeric antibodies. The antibody may be from recombinant sources and/or produced in transgenic animals. The term "antibody fragment" as used herein is intended to include

Fab, Fab', F(ab')2, scFv, dsFv, ds-scFv, dimers, minibodies, diabodies, and multimers thereof and bispecific antibody fragments. Antibodies can be fragmented using conventional techniques. For example, F(ab')2 fragments can be generated by treating the antibody with pepsin. The resulting F(ab')2 fragment can be treated to reduce disulfide bridges to produce Fab ¹ fragments. Papain digestion can lead to the formation of Fab fragments. Fab, Fab ¹ and F(ab')2, scFv, dsFv, ds-scFv, dimers, minibodies, diabodies, bispecific antibody fragments and other fragments can also be synthesized by recombinant techniques. [0059] Antibodies having specificity for a specific protein, such as the protein product of a biomarker, may be prepared by conventional methods. A mammal, (e.g. a mouse, hamster, or rabbit) can be immunized with an immunogenic form of the peptide which elicits an antibody response in the mammal. Techniques for conferring immunogenicity on a peptide include conjugation to carriers or other techniques well known in the art. For example, the peptide can be administered in the presence of adjuvant. The progress of immunization can be monitored by detection of antibody titers in plasma or serum. Standard ELISA or other immunoassay procedures can be used with the immunogen as antigen to assess the levels of antibodies. Following immunization, antisera can be obtained and, if desired, polyclonal antibodies isolated from the sera.

[0060] To produce monoclonal antibodies, antibody producing cells

(lymphocytes) can be harvested from an immunized animal and fused with myeloma cells by standard somatic cell fusion procedures thus immortalizing these cells and yielding hybridoma cells. Such techniques are well known in the art, (e.g. the hybridoma technique originally developed by Kohler and Milstein (Nature 256:495-497 (1975)) as well as other techniques such as the human B-cell hybridoma technique (Kozbor et a/., Immunol. Today 4:72 (1983)), the EBV-hybridoma technique to produce human monoclonal antibodies (Cole et a/., Methods Enzymol, 121 :140-67 (1986)), and screening of combinatorial antibody libraries (Huse et a/., Science 246:1275 (1989)).

Hybridoma cells can be screened immunochemically for production of antibodies specifically reactive with the peptide and the monoclonal antibodies can be isolated.

[0061] A person skilled in the art will appreciate that the detection agents can be labeled.

[0062] The label is preferably capable of producing, either directly or indirectly, a detectable signal. For example, the label may be radio-opaque or a radioisotope, such as ³ H, ¹⁴ C, ³² P, ³⁵ S, ¹²³ I, ¹²⁵ I, ¹³¹ I; a fluorescent (fluorophore) or chemiluminescent (chromophore) compound, such as fluorescein isothiocyanate, rhodamine or luciferin; an enzyme, such as alkaline phosphatase, beta-galactosidase or horseradish peroxidase; an imaging agent; or a metal ion.

[0063] The kit can also include ancillary reagents, such as buffers, vessels or controls. The kit can also include instructions for use thereof. [0064] The above disclosure generally describes the present application. A more complete understanding can be obtained by reference to the following specific examples. These examples are described solely for the purpose of illustration and are not intended to limit the scope of the present application. Changes in form and substitution of equivalents are contemplated as circumstances might suggest or render expedient. Although specific terms have been employed herein, such terms are intended in a descriptive sense and not for purposes of limitation.

[0065] The following non-limiting examples are illustrative of the present application: Examples

[0066] Epithelial ovarian carcinoma is the most lethal gynecological malignancy and disease specific biomarkers are urgently needed to improve diagnosis, prognosis and to monitor treatment efficiency. The inventors present an in-depth proteomic analysis of selected biochemical fractions of human ovarian cancer ascites. All ascites samples were analyzed by an

academic pathologist and the diagnosis of papillary serous carcinoma was confirmed. None of the patients had previously been treated with chemotherapy (Table 1). Extensive MS-based proteome profiling by multidimensional protein identification technology (MudPIT) ^16-19 and gel- enhanced LC-MS ¹² resulted in the confident identification of over 2,500 proteins.

[0067] Rigorous analysis of the MS-data was applied on the peptide and protein level to reduce the false discovery rate and to minimize protein inference by only accepting proteins with substantial peptide evidence. Data mining of this rich resource of ascites proteins revealed a large number of proteins involved in essential biological processes including immune function, metabolism, cytoskeletal dynamics and cell motility ²⁰ . Interestingly, over 600 proteins detected in this study have recently been reported in large-scale proteomics datasets of human plasma and urine ¹² ' ¹³ . To further minimize these 'secreted' proteins to a panel of candidate ovarian cancer markers, these data were mapped against 59 available ovarian cancer microarrays and to protein-protein interaction data from the Online Predicted Human Interaction Database (OPHID) ²¹ . This integrated bioinformatics data mining strategy reduced the large number of detected proteins to a short-list of 80 biomarkers. A further bioinformatics data mining strategy including several independent filters (i.e. identification in all four ovarian cancer ascites samples, previously identified in human plasma and/or urine, significant changes on several ovarian cancer microarray studies) was also employed to minimize the identified proteins. This integrative data mining resulted in the identification of 18 unique proteins that passed the selection criteria. Two of these unique proteins were selected for first phase validation to evaluate the diagnostic or prognostic value of these biomarkers. Initial biomarker validation was accomplished in an established mouse model of ovarian cancer. The presented proteomics analysis provides a framework for a more detailed analysis of ovarian cancer ascites.

Materials and Methods

Materials

[0068] Ultrapure grade urea, ammonium acetate, calcium chloride,

TRIS, HEPES and Triton-X-100 were from BioShop Canada Inc. (Burlington, ON, Canada). Ultrapure grade iodoacetamide, DTT and formic acid were obtained from Sigma. HPLC grade solvents (methanol, acetonitrile and water) were obtained from Fisher Scientific. Recombinant, proteomics grade trypsin was from Roche Diagnostics (Montreal, QC, Canada). OMIX solid phase extraction pipette tips were from Varian (Mississauga, ON, Canada).

Fractionation of human ovarian cancer ascites [0069] Ascites samples were obtained from the University Health

Network Ovarian Tissue Bank according to a protocol approved by the Institutional Research Ethics Board. Ascites were centrifuged at 2,000 rpm for 10 min. at 4°C to separate the fluid (Fraction S1) from the cellular component (S2/S3). The cells were incubated in hypotonic lysis buffer containing 1OmM HEPES, pH 7.4 for 30 min. on ice. The suspension was briefly sonicated and debris was sedimented by centrifugation at 14,000 rpm for 30 min. at 4°C. The supernatant, containing most soluble proteins was labeled Fraction S2. The cellular debris pellet was resuspended and further incubated for 30 min. at 4°C in hypotonic lysis buffer containing 1.5% Triton-X-100. Following incubation the suspension was recentrifuged at 14,000 rpm for 30 min. at 4°C. The supernatant was labeled Fraction S3.

In-solution protein digestion

[0070] An aliquot of 150 μg of total protein (determined by Bradford assay) from each fraction was precipitated overnight with 5-volumes of ice cold acetone followed by centrifugation at 14,000 rpm for 15 min. The protein pellet was solubilized in 8M urea, 2mM DTT, 5OmM Tris-HCI, pH 8.5 at 37°C for 1 hour, followed by carboxyamidomethylation with 1OmM iodoacetamide for 1 hour at 37°C in the dark. Samples were diluted with 5OmM ammonium bicarbonate, pH 8.5 to ~ 1.5M urea. Calcium chloride was added to a final concentration of 1mM and the protein mixture digested with a 1 :30 molar ratio

of recombinant, proteomics grade trypsin at 37°C overnight. The resulting peptide mixtures were solid phase-extracted with Varian OMIX cartridges (Mississauga, ON, Canada) according to the manufacturer's instructions and stored at -8O°C until further use. Multidimensional protein identification technology - MudPIT

[0071] A fully automated 9-cycle, 16-hour MudPIT procedure was set up essentially as described ^16-19 . A quaternary HPLC-pump was interfaced with a linear ion-trap mass spectrometer (LTQ Thermo Fisher Scientific, San Jose, CA) equipped with a nanoelectrospray source (Proxeon Biosystems, Odense, Denmark). A 100μm inner diameter fused silica capillary (InnovaQuartz, Phoenix, AZ) was pulled to a fine tip using a P-2000 laser puller (Sutter Instruments, Novato, CA) and packed with ~ 7cm of Jupiter™ 4μ Proteo 9θA C- _I2 reverse phase resin (Phenomenex, Torrance, CA), followed by ~5cm of Luna ^® 5μ SCX 100A strong cation exchange resin (Phenomenex, Torrance, CA). Samples were loaded manually on separate columns using an in-house pressure vessel. As peptides eluted from the microcapillary columns, they were sprayed directly into the MS. A distal 2.3 kV spray voltage was applied to the microsplitter tee (Proxeon Biosystems). The MS operated in a cycle of one full-scan mass spectrum (400-1400 m/z), followed by 6 data-dependent MS/MS spectra at 35% normalized collision energy, which was continuously repeated throughout the entire MudPIT separation. The MS functions and the HPLC solvent gradients were controlled by the Xcalibur data system (Thermo Fisher Scientific, San Jose, CA). A schematic of the MudPIT technology is displayed in Fig. 2. Gel-enhanced LC-MS analysis

[0072] For gel separations approximately 50-100 μg total protein of each ascites fraction were loaded onto 10-14.5% gradient polyacrylamide precast Tris-HCI gels (Bio-Rad), followed by Coomassie staining (Bio-Safe Coomassie stain, Bio-Rad). Each gel lane was cut into 11 slices which were subject to in-gel proteolysis, essentially as described ²² . Briefly, destained gel blocks were washed with 10OmM ammonium bicarbonate, and shrunk with

acetonitrile. Reduction and alkylation was performed with 1OmM DTT and 55mM iodoacetamide. Gel slices were washed twice with ammonium bicarbonate/acetonitrile and dried down in a speed-vac centrifuge. Each gel block was rehydrated in 12.5ng/μl trypsin in 5OmM ammonium bicarbonate at 4°C for 15 min. The solution was removed and replaced by 5OmM ammonium bicarbonate without trypsin and incubated over night at 37°C. The next day the supernatant was removed and the gel slices were extracted twice with 5% formic acid/acetonitrile (50:50). The three supernatants were combined and speed-vac concentrated to low volumes and diluted with 5% acetonitrile, 0.1% formic acid and stored at -80 °C until analyzed by LC-MS. For Gel-enhanced LC-MS analysis a 100 μm inner diameter fused silica capillary was pulled to a fine tip using a P-2000 laser puller and packed with ~ 7cm of JupiterTM 4μ Proteo 9θA Ci ₂ reverse phase resin. An aliquot of the in-gel digested peptide mixture was pressure loaded on a microcapillary column and analyzed by a 1- hour LC-MS gradient. The MS scan functions were exactly as described above.

Western Blotting

[0073] The soluble ascites fraction S1 was used for Western blotting.

Total protein was quantified by Bio-Rad Protein Assay (Bio-Rad Laboratories, Mississauga, ON). Total protein (3-50 μg) was separated by 7.5% or 16% SDS-PAGE and transferred onto nitrocellulose membranes (Bio-Rad Laboratories). Blots were probed for cofilin-1 (Abeam, Cambridge, MA), profilin-1 (Abeam), GAPDH (Abeam), and IQGAP-1 (Santa Cruz Biotechnology, Santa Cruz, CA). Immunoreactive proteins were detected with secondary HRP-coupled antibodies (Sigma) and visualized with ECL Western Blotting Detection Reagents (Amersham).

Protein identification, validation and grouping

[0074] Xcalibur raw files were converted to m/z XML using ReAdW and searched by XITandem ²³ against a human IPI protein sequence database (v3.20). To estimate and minimize our false positive rate the protein sequence database also contained every IPI protein sequence in its reversed amino

acid orientation (target-decoy strategy) as recently described ^{16 18} ' ²⁴ ' ²⁵ . Search parameters were: Parent ion δ-mass of 4 Da, fragment mass error of 0.4 Da, and complete carbaminomethyl modification of cysteine by IAA. Only peptides passing a default log-expectation value of -1 were further evaluated (see below).

[0075] A rigorous peptide quality control strategy was applied to effectively minimize false positive identifications. Briefly, a Perl-based tool was written to calculate, on the peptide level, a user-defined false positive rate. Identified peptides are binned into three charge states (+1 , +2, +3) and individual XITandem expectation values are calculated for each charge state to minimize the number of peptides mapping to decoy sequences to a user defined percentage. In the course of this project the value was set to 0.2% (total reverse spectra to total forward spectra), resulting in a low number of IPI accessions mapping to decoy sequences in the final output (<0.5%; 13 reverse proteins in 2,737 forward proteins). The calculated expectation values were: -3.22 for +1-ions, -2.17 for +2- ions, and -2.89 for +3-ions. Only fully tryptic peptides ≥7 amino acids, matching these criteria were accepted to generate the final list of identified proteins.

[0076] To minimize the significant problem of protein inference a database grouping scheme was developed and only report proteins with substantial peptide information. Briefly, the algorithm consists of several grouping steps performed on the database level and applies a heuristic approach by favoring proteins identified with the largest number of peptides. Similar strategies for the inference from peptide to protein level have been applied by the HUPO Plasma Proteome Project ²⁶ and more recently in a large-scale investigation of human adipocytes ²⁷ . This strategy includes several consecutive steps and only considers confidently identified peptides that passed our above filter criteria:

(1) Linking peptides to proteins: Peptide identifications obtained from XITandem searches (passing the above criteria) including all matching

IPI protein accessions were parsed into our MySQL database.

(2) Connecting groups: These initial assignments of peptides to proteins were arranged by combining peptide groups which identify the same proteins.

(3) Arranging groups: These second level groupings were then sorted in order to reflect number of peptides matching to a given protein. This step takes into account proteins with sequence homologies as manifested by shared peptides (protein inference problem).

(4) Minimal proteins list: In the final grouping step, proteins are arranged into "levels" favoring proteins identified with the largest number of peptides, which were placed onto the highest level.

Importantly, only proteins on this top level are reported in our final "minimal protein inference list'. Although, if highly homologous proteins were identified by this process, having an identical number of diagnostic peptides, an arbitrary choice of the protein having the lowest IPI accession number was reported in our final output (e.g. actin, cytoplasmic 1 and actin cytoplasmic 2). Proteins on a lower level within one group were only reported if they were identified by at least one additional distinct peptide.

Comparison to the HUPO plasma proteome and urine proteome datasets [0077] Comparison of the ascites proteome resource to the HUPO plasma proteome data (3,020 proteins) ¹³ and a recently published urine proteome dataset (1543 proteins) ¹² was achieved using the ProteinCenter bioinformatics software (Proxeon Biosystems, Odense, Denmark). To link these data sets, the IPI accession keys were loaded for each of the three projects into ProteinCenter. Individual data sets were then linked via accession keys in the ProteinCenter database, which contains over 30 million accessions.

Comparison to ovarian cancer microarray datasets

[0078] Identified candidate proteins (Fig. 5B), were queried against the inventors' in-house database for differentially expressed genes assembled

from several published and unpublished ovarian cancer microarray studies. Protein/gene pairs found to be significantly changed in at least one of the comparisons were selected.

Protein-Protein Interaction Analysis [0079] To further focus the selection of biomarkers, the set of 18 and

10 potential biomarkers were mapped (see Results and Discussion) into I ² D protein-protein interaction databases ²¹ . This enabled the inventors to (1) identify subnetworks that are significantly deregulated in multiple microarray ovarian cancer studies, and (2) determine additional markers that may have been removed initially due to stringent peptide-to protein filtering steps, or were absent in plasma and urine from healthy subjects. To further increase the confidence in these biomarkers, the inventors assessed whether the enrichment can be achieved by chance, as described next.

Random Network Comparisons [0080] To determine if the experimentally derived networks could be obtained by random chance, the following permutation method was employed. Eighteen (or 10) proteins were chosen at random from a pool of known secreted proteins, or the set of all known human proteins in UniProt (build 8.2). The random proteins (seeds) were then used to search the I ² D database in two passes. On the first pass, all proteins known to interact with the seeds were selected and added to the seed list. On the second pass, protein interactions from the I ² D database among any of the expanded set of seed proteins were selected and added to generate the random network. Local and global network statistics were then calculated. The process was repeated for 1000 iterations (for the 18 seeds) or 10 000 iterations (for the 10 seeds) to generate the random distribution of network properties. The Student's t test was then used to compare the properties of the experimentally determined networks against the distributions of random networks.

Integrative Data Mining

[0081] To minimize the large number of identified proteins to a feasible number of biomarkers for further evaluation a bioinformatics data mining strategy was applied. Several independent filters were included: 1) only proteins confidently identified in ovarian cancer ascites of all four patients were accepted; 2) the resource of ascites proteins was mapped onto published proteome datasets of human plasma and urine: secreted proteins; and 3) the resource of ascites proteins was compared against an in-house database of 59 ovarian cancer microarrays to identify proteins with significant changes: disease relevant proteins. This strategy resulted in 18 unique proteins which passed the selection criteria (Fig. 6 and Table 5).

First phase validation -Mouse Model of Ovarian Cancer

[0082] To fully evaluate the diagnostic or prognostic value of these biomarkers two proteins (CFL1 and PFN 1) from Figure 6 were selected for further validation. [0083] In order to overcome the limited availability of high quality, well defined human tissue and plasma samples and the high biological variability within a human population, the inventors turned to the established mouse model of ovarian cancer for initial biomarker validation (Roby et al.). In this model, epithelial cancer cells, which spontaneously transformed in vitro from surface ovarian epithelial cells, are transplanted under the ovarian bursa and mice develop all symptoms of ovarian cancer within 2-3 months. The epithelial cell lines used for these experiments were IG10 and IC5 (Fig. 8).

Western blotting — Validation for biomarkers in Mouse Model of OVCA

[0084] Cells, serum and ascites fluid from mouse model of ovarian cancer, and normal surface epithelial cells were used for Western blotting.

Total protein was quantified by BCAAssay (Pierce). 20ug of total protein for cells or 100ug of serum or ascites was loaded and was separated on gradient

(4-12%) NuPAGe BIS- Tris (Novex) and transferred onto nitrocellulose membranes (Bio-Rad Laboratories). Blots were probed for cofilin-1 (Abeam, Cambridge, MA) and profilin-1 (Abeam). Immunoreactive proteins were

detected with secondary HRP-coupled antibodies (Santa Cruz, CA) and visualized with ECL Western Blotting Detection Reagents (Amersham).

Immunolocalization — Validation for biomarkers in Mouse Model of OVCA

[0085] Tissue from solid tumors developed in the mouse model of ovarian cancer, and normal surface epithelial cells were used for immunolocalization experiments. Sections were stained with anti-profilin

(1 :200; Abeam) or anti-cofilin (1 :50, Abeam) diluted in 5% goat serum in PBS.

Slides were then held at 4°C overnight, washed in PBS, followed by a 2-hour incubation at room temperature with biotinylated anti rabbit secondary antibody (1 :200 dilution, VectaStain ABC, Vector Labs) using DAB substrate

(brown precipitate) as a final detection step, followed by counterstaining with hematoxylin.

Results and Discussion

A systems biology analysis strategy to identify putative biomarkers [0086] In this study, in-depth shot-gun proteomics was applied to generate a high confidence protein resource of ovarian cancer ascites. Although, MudPIT is an excellent technology for the identification of large numbers of proteins, it lacks the highthroughput capabilities of microarray or SELDI-TOF-MS, required to analyze large numbers of patients for useful clinical insights. In contrast, current gene chip platforms are highly automated and allow for the analysis of many patients, albeit lacking the ability to directly identify proteins, the molecular species most likely to be useful as a disease biomarker. To overcome some of these limitations the inventors applied an integrated systems biology approach, combining extensive MudPIT-based profiling, with publicly available proteome datasets, predicted protein-protein interactions in OHPID and a large number of available ovarian cancer microarray datasets. This strategy enabled the inventors to 1) confidently identify a large number of proteins in a selected number of ovarian cancer ascites 2) identify proteins previously described as secreted proteins, as these are most likely the best biomarker candidates, and 3) link these putative

candidates to changes in the expression profile of several ovarian cancer microarrays. To fully evaluate the potential of these proteins as candidate biomarkers for early detection, chemotherapy resistance or treatment efficiency, future investigations in large numbers of carefully controlled plasma samples are required. These studies require high-throughput, sensitive and quantitative methods such as ELISA (Enzyme-Linked Immunosorbent Assay) or protein arrays for targets with validated antibodies, or highly specialized MS-based methods such as Multiple Reaction Monitoring (MRM).

Biochemical fractionation of human ovarian cancer ascites [0087] As ascites are produced in large amounts and contain both tumor cells and soluble growth factors, the inventors sought to perform an in- depth proteomics analysis of human ovarian cancer ascites to guide the discovery of candidate biomarkers for ovarian cancer. Ascites were obtained from four patients with papillary serous ovarian carcinoma who were not previously treated by chemotherapy (Table 1). To minimize the protein complexity the ascites were biochemically fractionated into cellular and soluble protein fractions and independently analyzed by extensive MudPIT- based expression proteomics and gel-enhanced LC-MS (Fig. 1). To compensate for the high sample complexity and large dynamic range of protein abundance, repeat MudPIT-based analyses were performed. ¹⁶ ' ^28'30 The combination of a simple biochemical fractionation protocol, with extensive MudPIT-based and gel-enhanced LC-MS allowed the inventors to confidently detect over 2,500 proteins (Table 2) and the inventors present for the first time a comprehensive human ovarian cancer ascites resource. Protein identification

[0088] In the course of this study, a total of 30 MudPITs and 33 gel- enhanced LCMS/MS runs were recorded, resulting in the accumulation of over 3 million mass spectra. An essential process of proteomics is to rigorously filter the recorded data to minimize the number of false positive identifications. Several strategies have been presented in the literature to accomplish this challenging problem. Sophisticated statistical algorithms have

been developed ¹⁸ ' ^31-33 to assign defined statistical confidence values to peptide identifications obtained by search algorithms such as SEQUEST ³⁴ , Mascot ³⁵ , XITandem ²³ and MyriMatch ³⁶ . A common strategy to estimate the number of false positive identifications is the use of "target/deco/ databases ¹⁸ ' ²⁴ ' ^25>36 . The number of reverse "decoy" protein identifications in the final data allow for an objective estimation of false positive identifications. The whole concept is further complicated by the "protein inference" problem, a term which refers to the observation that confidently identified peptide sequences are frequently shared by multiple database entries. Several strategies have been presented dealing with protein inference in large-scale proteomics ²⁶ ' ²⁷ ' ³⁷ - ⁴⁰ .

[0089] ProteinProphet developed by Aebersold and colleagues ³⁷ , and

PROT_PROBE developed by the laboratory of John Yates ³² apply elegant statistical models to validate protein identifications based on the peptides identified by LC-MS/MS. While these algorithms provide the most reliable protein identifications, they require extensive computation and sophisticated proteomics workflows. Several other strategies have been presented in the literature. These include integrated heuristic workflows as used in the HUPO plasma proteome project ²⁶ , database grouping schemes ⁴¹ ' ⁴² , DTASelect ⁴³ , or the Protein Hit Score ³⁹ to minimize protein inference.

[0090] In the course of this project the inventors used the open-source search algorithm XITandem ²³ , as it provides a cost efficient, fast and accurate alternative to the commercially available algorithms SEQUEST ³⁴ and Mascot ³⁵ . XITandem searches were performed as described in the Experimental Procedures. Using the "target/decoy" strategy, the generated data was filtered to minimize false positive identifications (< 0.5%) and group identifications into a minimal list of proteins, reporting only a parsimonious group of proteins supported by the largest number of high confidence peptides (see Experimental Procedures). A similar strategy was recently reported by Mann and colleagues in a large-scale investigation of human adipocytes ²⁷ . Briefly, searches were performed against a human IPI protein

sequence database (v3.20) containing every protein sequence in its forward and reverse amino acid orientation. Next, a stringent multi-step filter was applied to minimize false positive identifications and only proteins represented by two or more fully tryptic peptides >7 amino acids were accepted in our final proteomics dataset. First, peptides with a default log-expectation value of ≤ - 1.0 were parsed into an in-house MySQL database. Next, identified peptides were binned into +1 , +2, and +3 charge states and an in-house algorithm was applied to calculate a user-defined false positive rate (0.2% on the level of total spectra) (Fig. 3A/B). For this study, conservative log-expectation values of -3.22 / -2.17 / -2.89 for the +1 , +2, and +3 charge states, respectively, were calculated (on the peptide level) (Fig. 3B) ³⁶ ' ⁴⁴ . As a second filter the inventors only accepted proteins identified with at least 2 peptides, resulting in an estimated false positive rate of < 0.5% (only 13 reverse proteins passed these criteria) (Table 2). [0091] In total 18,559 (patient 1 only; 22,972 for all four patients) unique peptides were identified, mapping to a total of 5,629 unique IPI accessions (>2 peptides per protein). To minimize the protein inference problem, a peptide centric database scheme was developed (see Experimental Procedures) and grouped IPI accessions to a minimal set of reported proteins, similar to a recently reported strategy ²⁷ . After applying the grouping scheme the 5,629 unique IPI accessions were grouped to a reported "minimal proteins" list of 2,299 unique IPI accessions. After removal of known contaminants (human keratins and trypsin) a total of 2,278 proteins are currently present in our ascites resource (Table 3) [Note: The 2,278 proteins are derived form the very detailed analysis of patient 1 , referred to as 'core ascites resource'. The analyses of selected cellular fractions of three more patients 2, 3, and 4 increased the total number of identified proteins to 2,737, referred to as 'extended ascites resource ^* ]. This data was used to confirm and prioritize some of our initial candidates. The entire set of 2,737 proteins is available in Table 3 (The entire RAW data recorded in this project will also be deposited to the Tranche serveή ⁴⁵ .

[0092] Comparison of the biochemical fractions revealed that the majority of the detected proteins were present in the cellular fractions (S2 and S3) and only 229 proteins were confidently detected in the soluble S1 fraction (Fig. 3C). This is not surprising as this fraction is very similar in protein composition and concentration to human plasma, containing very high concentrations of serum albumin and immunoglobulins, complicating the detection of low abundance proteins ⁴⁶ . Furthermore, MudPIT detected a larger number of proteins compared to gel-enhanced LC-MS. The reason for this increased number of protein identifications is most likely due to the larger number of MudPIT experiments (30 in total) as compared to gel-enhanced LC-MS experiments (Fig. 3D).

Functional annotation and comparison to human body fluid datasets

[0093] To obtain a systematic and functional overview of proteins present in human ovarian cancer ascites, a FDA Gene Ontology Tool termed GOFFA (Gene Ontology For Functional Analysis) ²⁰ was used. GOFFA is an intuitive tool that calculates statistically enriched GO-terms and allows the user to dynamically visualize high throughput data in the context of biological functions.

[0094] In Fig. 4, proteins confidently detected in selected ascites fractions were analyzed and significantly enriched (P-value < 0.01 ; E-value >

2) GO-terms in the category 'biological process' are presented. A large number of detected proteins could be mapped to at least one GO-term (67-

88%) and a significant number of enriched GO-terms were detected in each ascites fractions, covering a range of biologically relevant processes. As to be expected, fraction S1 was enriched in Gene Ontology terms related to 'blood, plasma and immune' related functions (69%). Several other GO categories, metabolism and death, were also significantly enriched in the soluble ascites fraction. Interestingly, the fraction of mixed cell populations (S2 and S3) was heavily enriched in 'metabolism' related GO-terms, this finding is in agreement with other reports that found malignant cells to be enriched with this biological process ^47-49 . Other enriched Gene Ontology functions were translation,

splicing, transport, and cytoskeleton/mobility. Table 4 contains the most significantly enriches Gene Ontology terms in the category "cellular component" for each of the three analyzed fraction.

Bioinformatics data mining to identify putative biomarkers [0095] Clinical biomarkers are urgently needed to improve patients diagnosis, prognosis and to monitor the effects of therapeutic treatment ⁵⁰ ' ⁵¹ . However, molecular heterogeneity of the tumors, non-cancer diseases that reduce biomarker specificity and low sensitivity of the cancer biomarker due to low concentration in early disease, are the three main pitfalls for the discovery of effective biomarkers for ovarian cancer. The most commonly used marker for ovarian cancer is CA-125 ⁵² , which is widely used for the surveillance of ovarian cancer patients whose diagnosis was previously made by histology. A downside of CA-125 based diagnosis is its low sensitivity and specificity, especially for the detection of early disease. [0096] A large number of molecules have been evaluated as possible biomarkers for ovarian cancer. These include vascular endothelial growth factor (VEGF) ⁵³⁵⁴ , matrix metalloproteinases (MMP's) ⁵⁵ ' ⁵⁶ and members of the kallikrein family of serine proteases ^57-59 . Due to concerns about low sensitivity and specificity the search for new and improved biomarkers for ovarian cancer continues.

[0097] To identify biomarker candidates all proteins confidently identified in ovarian cancer ascites of patient 1 (core resource) were used and mapped to several previously reported large-scale datasets. These included the HUPO plasma proteome ¹³ , a recently published high quality urine proteome dataset ¹² (Fig. 5A/B) and 59 available ovarian cancer microarrays (for detailed information see Tables 5, 6 and 7). Three different mining strategies were then applied to generate the list of putative biomarkers (Fig. 5B)

1) Consider only proteins found in human ovarian cancer ascites that were also found in plasma and/or urine and significantly changed on at least one ovarian cancer microarray -> secreted proteins

2) Proteins found in human ovarian cancer ascites, significantly changed on at least one ovarian cancer microarray, but not detected in healthy body fluids -> putative disease related proteins

3) Protein interacting partners of group 1) according to the OPHID database ²¹ that were also significantly changed on at least one ovarian cancer microarray and found in our ascites proteome. [0098] To verify and confirm the reproducible detection of these initial candidate proteins in human ovarian cancer ascites three more ovarian cancer ascites were analyzed {"extended resource") and only accepted proteins confidently detected by mass spectrometry in all four ascites samples. This reduced our panel of putative ovarian cancer biomarkers to a non-redundant list of 80 proteins (Tables 5, 6 and 7).

[0099] In the current study, the inventors present for the first time a high quality proteome resource of proteins detected in ovarian cancer ascites. Using extensive MudPIT-based and gel-enhanced LC-MS/MS proteomics, the inventors were able to identify over 2,500 proteins in the soluble and cellular fractions of four confirmed cases of ovarian cancer. Applying an integrated informatics platform, the inventors were able to minimize the large number of identified proteins to a small panel of proteins reproducibly detectable in ovarian cancer ascites.

[00100] As described, a bioinformatics data mining strategy was employed in order to minimize the large number of identified proteins to a feasible number of biomarkers for further evaluation. The integrative data mining strategy included several independent filters, namely, comparison to HUPO plasma and urine proteome datasets, comparison to ovarian cancer microarray datasets, and identification in all four patients with ovarian cancer.

[00101] As a result of this integration 18 unique proteins were identified that passed the selection criteria (i.e. identified in all four ovarian cancer ascites samples, previously identified in human plasma and/or urine, significant changes on several ovarian cancer microarray studies) (see Fig. 6 and Table 5).

[00102] These secreted proteins (described in further detail below) include: ARHGDIB (Rho GDP-dissociation inhibitor 2), CFL1 (Cofilin-1), PFN1 (profilin-1), GSTP1 (Glutathione S-transferase P), S100A11 (Protein S100- A11), PRDX6 (Peroxiredoxin-6), HSPE1 (10 kDa heat shock protein, mitochondrial), LYZ (Lysozyme C precursor), GPI (Glucose-6-phosphate isomerase), HIST2H2AA (Histone H2A type 2-A), GAPDH (Glyceraldehyde-3- phosphate dehydrogenase), HSPG2 (Basement membrane-specific heparan sulfate proteoglycan core protein precursor), LGALS3BP (Galectin-3-binding protein precursor), CTSD (Cathepsin D precursor), APOE (Apolipoprotein E precursor), IQGAP1 (Ras GTPase-activating-like protein IQGAP1), CP (Ceruloplasmin precursor), and IGLC2 (IGLC1 protein).

1) Secreted proteins in human piasma/urine also detected in ascites

[00103] As indicated above, a total of 18 proteins were detected in all four ascites and have previously been reported as secreted proteins in large- scale analyses of human plasma ¹³ and urine ¹² (Fig. 6 and Table 5). Comparison to 59 available ovarian cancer microarrays allowed to further prioritize these proteins as biomarkers. Systematic analysis allowed further grouping into functional categories such as cell proliferation, differentiation and apoptosis, signaling to the cytoskeleton, cell adhesion and motility, transport, metabolic processes and proteolysis.

Proteins associated with cell proliferation

[00104] S100A11 (Calgizzarin), a calcium-binding protein, regulates cell growth by inhibiting DNA synthesis ⁶⁰ . In epithelial ovarian carcinoma cells, nuclear expression of S100 protein is associated with aggressive behavior of ovarian tumors ⁶¹ . The inventors robustly detected S100A11 in all four ovarian

cancer ascites and its expression was up-regulated in ovarian cancer, as compared to normal ovary by microarray analyses (Fig. 6).

Proteins involved in cell differentiation and apoptosis

[00105] Several candidates are associated with cellular differentiation and apoptosis. These include, glutathione-S-transferase P (GSTP1), Cofilin-1 (CFL1), apolipoprotein-E (APOE), Lysozyme C (LYZ), and mitochondrial 1OkDa heat shock protein (HSPE1).

[00106] Increased expression of GSTP1 in human cancers confers some protection of malignant cells from anti-neoplastic drug treatment by preventing apoptosis and poor prognosis ⁶² ' ⁶³ . APOE is involved in lipid homeostasis and was found up-regulated in ovarian cancer samples through SAGE analysis ⁶⁴ . Chen et al. showed that APOE was absent from borderline ovarian tumors but was detected in later stage ovarian cancers ⁶⁵ . Its inhibition in an ovarian cancer cell line (OVCAR3) led to cell cycle arrest and apoptosis. Cofilin-1 , is a cytoskeletal protein related to cell migration, aggregation and differentiation. Regulation of cofilin affects morphological changes by remodeling the actin cytoskeleton associated with differentiation in cancer cells 66 and has been shown to modulate tumor cell invasion ⁶⁷ .

[00107] The robust detection of these proteins in the present study, its previous detection by mass spectrometry in urine/plasma and significant changes for each candidate protein on several ovarian cancer microarrays advocates further investigation of these proteins as potential biomarkers for ovarian cancer (Fig. 6).

Proteins involved in signaling to the cvtoskeleton [00108] Several candidate proteins, including CFL1 , Profilin-1 (PFN1), Rho GDP dissociation inhibitor 2 (ARHGDIB), Ras GTPase-activating-like protein 1 (IQGAP1), Galectin-3-binding protein (LGALS3BP), S100A11 , APOE, and 1OkDa Heat shock protein (HSPE1) are involved in intracellular signaling to the cytoskeleton.

[00109] IQGAP1 , a scaffold protein, is involved in cytoskeletal rearrangement through interaction with various proteins, including actin, calmodulin, CD44, E-cadherin, the tumor suppressor adenomatuous polyposis coli (APC), CDC42 and Rac1 ⁶⁸ ' ⁶⁹ . Over-expression of IQGAP1 can significantly increase the migratory and invasive potential of cancer cells ⁶⁸ . Additionally, IQGAP1 is expressed in endothelial cells and can bind VEGFR2, promoting endothelial cell migration and proliferation, following vascular injury ⁷⁰ , and contribute to angiogenesis ⁷¹ . The connection to invasion, migration, angiogenesis, as well as its robust detection in ovarian cancer ascites, human plasma ¹³ and urine ¹² make it an interesting candidate biomarker. To the best of our knowledge, IQGAP1 has not been proposed as a biomarker candidate for ovarian cancer (Fig. 6).

Proteins associated with cell adhesion and motility

[00110] Oncogenic signaling can stimulate cytoskeletal changes that in turn alter substrate adhesions, essential for cell spreading and motility.

Several proteins identified in the present study are associated with cellular adhesion and motility. These include ARHGDIB, LGALS3BP, Basement membrane-specific heparan sulfate proteoglycan (HSPG2), and PFN1.

Profilins are ubiquitous, 12-15kDa proteins that regulate actin polymerization by binding to and sequestering the actin monomer ⁷² . Over-expression of

PFN 1 increases endothelial cell adhesion to fibronectin ⁷³ and reduces the migration of invasive breast cancer cells ⁷⁴ . Although, decreased expression of

PFN1 has been reported in invasive and migratory breast cancer cells, in our study comparison to available microarray data revealed up-regulation of the PFN 1 transcript in several grades and stages of ovarian cancer (Fig. 6).

[00111] Interestingly, a recent proteomics study (Faca et al. PLoS ONE, 2008 Jun 18; 3(6):e2425) identified both CFL1 and PFN1 as proteins secreted into the culture media by three different human ovarian cancer cell lines (CaOV3, OVCAR3 and ES2) and purified ovarian cancer cells enriched from human ovarian cancer ascites. This provides additional information for

secreted nature of these proteins and their potential use as a diagnostic biomarker for ovarian cancer.

Proteins involved in transport, metabolic processes and proteolysis

[00112] Several identified protein are associated with transport, metabolic and proteolytic processes. These include Ceruloplasmin (CP), Glucose-6-Phosphate lsomerase (GPI), Peroxiredoxin-6 (PRX6), and Cathepsin D (CTSD). CP is plasma glycoprotein that transports copper throughout the body ⁷⁵ . Interestingly, microarray analysis has linked this gene to tumor invasion and metastasis ⁷⁶ . High serum levels of CP have been reported in many cancer patients and increase with tumor mass ⁷⁷ ' ⁷⁸ . Its transcript was up-regulated in ovarian cancer compared to normal ovary, as well as in malignant vs. benign cancer (Fig. 6) making it an interesting candidate. Serum levels of G6PI are known to be elevated in ovarian cancer patients ⁷⁹ . It has been linked to tumor invasion and metastasis ⁸⁰ ' ⁸¹ . The proteinase CTSD, has been implicated in tumor invasion, metastasis ⁸² , tumor cell proliferation, apoptosis and angiogenesis. It is abundantly detected in invasive ovarian cancer tumors and is associated with poor clinical outcome ⁸³ . In the present study Cathepsin-D was detected in ovarian cancer ascites and was previously detected in urine and plasma (Fig. 6). [00113] The inventors used Western blotting to detect the presence of four candidates in ascites fraction S1 (soluble fraction). Representative Western blots using commercially available antibodies against CFL1 , PFN1 , GAPDH ₁ and IQGAP-1 are shown in Fig. 6B.

First phase validation of secreted protein biomarkers [00114] The diagnostic or prognostic value of the biomarkers in Fig. 6 was evaluated by selecting two proteins (CFL1 and PFN 1) for further validation in a mouse model of ovarian cancer. Selection of the biomarkers was based on several objective criteria: biological annotation, number and nature of the microarray studies with significant changes and the availability of validated antibodies. The mouse model of ovarian cancer was established

using the IC5 and IG10 epithelial cancer cell lines. Several antibody-based validation experiments for CFL1 and PFN 1 in this mouse model of ovarian cancer have been completed.

[00115] Western blotting demonstrated upregulation of both CFL1 and PFN 1 biomarkers in ovarian cancer cells when compared to normal surface epithelial cells, both maintained in vitro (Figs. 9 and 10).

[00116] Immunolocalization revealed elevated staining of both CFL1 and PFN 1 biomarkers in solid tumors that formed adjacent to the ovary, while ovarian surface epithelial cells in normal ovaries exhibited limited immunoreactivity (Figs. 11 and 12).

[00117] Preliminary results from the candidate biomarker profile indicated that both PFN 1 and CFL1 were identified in human ascites. At the time of euthanasia both ascites fluid and serum from tumour-bearing animals exhibited upregulation of both PFN1 and CFL1 biomarkers when compared to serum from aged matched healthy controls (Figs. 13 and 14). Quantitation of PFN1 and CFL1 in serum was determined via Western blot analysis (Fig. 14).

2) Proteins identified in ovarian cancer ascites but not in other body fluids

[00118] These included Proliferation-associated protein 2G4 (PA2G4), a transcriptional co-repressor of the androgen receptor involved in cell growth regulation and transcriptional inhibition of some E2F1-regulated promoters;

Proteasome activator complex subunit 2 (PSME2), implicated in immunoproteasome assembly required for efficient antigen processing;

15OkDa oxygen regulated protein (HYOU1), which can trigger cytoprotective cellular mechanisms in a hypoxic environment; Lamin B1 (LMNB1) a structural component of the nuclear lamina thought to play an important role in nuclear architecture, DNA replication, and gene expression. Some evidence exists placing this protein as a potential marker for drug resistance and transformation to malignancy ⁸⁴ ; and Myeloperoxidase (MPO), which has microbicidal activity against a wide rage of organisms. Importantly, all these

proteins were detected in all four ascites samples, showed significant changes on at least one ovarian cancer microarray, or were known or predicted protein interaction partners of secreted ascites proteins (Tables 6 and 7). Protein-Protein Interaction Networks

[00119] The protein-protein interaction network for the 18 candidate biomarkers (Fig. 7) was generated by querying the I ² D database (version 1.5) (called seeds and highlighted as ovals), which resulted in a sparse network with 201 proteins and 201 interactions. Importantly, while all 18 proteins are significant on multiple microarray studies, the inventors highlighted those that are the most frequent (underlined). The inventors have also highlighted with shaded areas, subnetworks that are significantly up-regulated. Interestingly, this includes both proteins identified in human body fluids and those that were not, but were significantly deregulated on multiple microarrays. There is a strong enrichment of finding the 18 markers significantly deregulated on multiple microarrays compared to random chance, where less than half of these seed proteins are significantly deregulated (using a random secreted protein network, p ) 2.6383 * 10 ^-7 , or using a random protein network p ) 2.1179 x 10- ⁹ ). [00120] While the proteins are not highly interconnected and on average do not have a high degree, they form a large main connected subgraph. To determine if this can happen by chance, the inventors compared the resulting network to random networks generated from 18 random secreted proteins and 18 random proteins. Interestingly, both comparisons show significantly larger size of the largest connected subgraph for our network (p ) 0.000115438 for random secreted proteins; p ) 4.0451 x 10 ^-5 for the random proteins).

[00121] Proteins in the network highlighted as diamonds are proteins interconnecting seed proteins, and the inventors show their up- or down- regulation on ovarian cancer microarray studies by arrows. Interestingly, these proteins are significantly enriched for being deregulated on multiple microarray studies; no interconnecting proteins are significantly deregulated in

average in random networks (random secreted protein network p ) 4.73864 * 10 ^~11 ; random protein network p ) 2.63083 x 10 ^-54 ). The network for the 10 proteins shows similar significant results (data not shown).

[00122] Both 18- and 10-seed networks have lower average degree and are not highly interconnected compared to random networks. This could be caused by two phenomena. First, it is possible that these biomarkers represent proteins that are less studied in average (although there are exceptions such as CFL1 and PFN 1) compared to any randomly selected proteins from the current interaction network. Second, it may be possible that secreted biomarkers form a network of different topology, a network that is more planar and less interconnected, compared to signaling networks. Most importantly, neither of the two types of random networks have many seed proteins significantly deregulated on multiple arrays, and in average none of the interconnected proteins have such support. This makes our 18- and 10- protein networks much more relevant to ovarian cancer biology.

Conclusion

[00123] This study only considered the most robustly detected proteins (e.g. detected in all four patients by stringent criteria and supported by microarray analysis) as biomarkers. The current ovarian cancer biomarker of choice, mucin 16 (CA-125) was detected only in the first patient by gel- enhanced LC-MS. Metalloproteinase-9, a protein previously proposed as a biomarker for ovarian cancer ⁸⁵ , was detected in three of the four patients, rejecting it from our high-stringency panel of putative biomarkers. The entire panel of detected proteins (Table 3) is available to the public for further investigation.

[00124] Proteomics of ascites fluid could serve as a valuable tool to study the mechanisms of acquired chemoresistance. Whereas most women respond to the initial platinum-based chemotherapy, resistance to the treatment develops in 1/3 of the women during the initial treatment and in almost all cases of recurrent disease. This issue was previously studied mainly by microarray gene chip analyses ^86'88 . A limited number of proteomic

investigations have addressed the issue of chemoresistance, mainly by comparing chemosensitive to chemo-resistant cell lines and several proteins with differential expression patterns have been identified in these studies ⁸⁹ ' ⁹⁰ .

[00125] In summary, the inventors present for the first time a high-quality in-depth proteomics analysis of ovarian cancer ascites. Integrated bioinformatics analyses of this data with previously published body fluid proteomes, as well as 59 available ovarian cancer microarray datasets, allowed for prioritized data analysis, with the goal to identify biomarkers. The inventors present a panel of 80 proteins, robustly detected in all four cases of ovarian cancer ascites, including several novel markers.

[00126] The inventors also present 18 secreted proteins as biomarkers for ovarian cancer identified in all four ovarian cancer ascites samples, previously identified in human plasma and/or urine, and exhibiting significant changes on several ovarian cancer microarray studies. [00127] Further, the inventors present data evaluating the diagnostic or prognostic value of these biomarkers. Two biomarkers, CFL1 and PFN1 , were selected for validation in a mouse model of ovarian cancer. Data from antibody-based experiments demonstrated upregulation of both biomarkers in ovarian cancer cells versus controls. [00128] While the present application has been described with reference to what are presently considered to be the preferred examples, it is to be understood that the present application is not limited to the disclosed examples. To the contrary, the present application is intended to cover various modifications and equivalent arrangements included within the spirit and scope of the appended claims.

[00129] All publications, patents and patent applications are herein incorporated by reference in their entirety to the same extent as if each individual publication, patent or patent application was specifically and individually indicated to be incorporated by reference in its entirety.

Table 1

Table 2

Table 3

Table 3 (Continued)

Table 3 (Continued)

Table 3 (Continued)

Table 3 (Continued)

Table 3 (Continued)

Table 3 (Continued)

Table 3 (Continued)

Table 3 (Continued)

Table 3 (Continued)

Table 3 (Continued)

Table 3 (Continued)

Table 3 (Continued)

Table 3 (Continued)

Table 3 (Continued)

Table 3 (Continued)

Table 3 (Continued)

Table 3 (Continued)

Table 3 (Continued)

O

Table 3 (Continued)

Table 3 (Continued)

Table 3 (Continued)

Table 3 (Continued)

Table 3 (Continued)

Table 3 (Continued)

Table 3 (Continued)

Table 3 (Continued)

Table 3 (Continued)

Table 3 (Continued)

Table 3 (Continued)

Table 3 (Continued)

Table 3 (Continued)

Table 3 (Continued)

Table 3 (Continued)

Table 3 (Continued)

Table 3 (Continued)

Table 3 (Continued)

Table 3 (Continued)

Table 3 (Continued)

Table 3 (Continued)

Table 4

Presented are the most significantly enriched Gene Onology terms.

Category: cellular component - for the three biochemical fractions S1, S2, and S3

Table 4 Continued

Table 4 Continued

CO

Table 6

Table 6 (Continued)

Table 7

FULL CITATIONS FOR REFERENCES REFERRED TO IN THE SPECIFICATION

1. Jemal, A.; Siegel, R.; Ward, E.; Murray, T.; Xu, J.; Smigal, C; Thun, M. J., Cancer statistics, 2006. CA Cancer J Clin 2006, 56, (2), 106-30.

2. Parkin, D. M.; Bray, F.; Ferlay, J.; Pisani, P., Global cancer statistics, 2002. CA Cancer J Clin 2005, 55, (2), 74-108. 3. Abramov, Y.; Anteby, S. O.; Fasouliotis, S. J.; Barak, V., The role of inflammatory cytokines in Meigs' syndrome. Obstet Gynecol 2002, 99, (5 Pt 2), 917-9.

4. Davidson, B.; Reich, R.; Kopolovic, J.; Berner, A.; Nesland, J. M.; Kristensen, G. B.; Trope, C. G.; Bryne, M.; Risberg, B.; van de Putte, G.; Goldberg, I., lnterleukin-8 and vascular endothelial growth factor mRNA and protein levels are down-regulated in ovarian carcinoma cells in serous effusions. Clin Exp Metastasis 2002, 19, (2), 135-44. 5. Fang, X.; Yu, S.; Bast, R. C; Liu, S.; Xu, H. J.; Hu, S. X.; LaPushin, R.; Claret, F. X.; Aggarwal, B. B.; Lu, Y.; Mills, G. B., Mechanisms for lysophosphatidic acidinduced cytokine production in ovarian cancer cells. J Biol Chem 2004, 279, (10), 9653- 61. 6. Feldman, G. B.; Knapp, R. C; Order, S. E.; Hellman, S., The role of lymphatic obstruction in the formation of ascites in a murine ovarian carcinoma. Cancer Res 1972, 32, (8), 1663-6.

7. Zebrowski, B. K.; Yano, S.; Liu, W.; Shaheen, R. M.; Hicklin, D. J.; Putnam, J. B., Jr.; Ellis, L. M., Vascular endothelial growth factor levels and induction of permeability in malignant pleural effusions. Clin Cancer Res 1999, 5, (11), 3364-8.

8. Berchuck, A.; Carney, M., Human ovarian cancer of the surface epithelium. Biochem Pharmacol 1997, 54, (5), 541-4.

9. Bookman, M. A., Biological therapy of ovarian cancer: current directions. Semin Oncol 1998, 25, (3), 381-96. 10. Westermann, A. M.; Beijnen, J. H.; Moolenaar, W. H.; Rodenhuis, S., Growth factors in human ovarian cancer. Cancer Treat Rev 1997, 23, (2), 113-31.

11. Mills, G. B.; May, C; McGiII, M.; Roifman, C. M.; Mellors, A., A putative new growth factor in ascitic fluid from ovarian cancer patients: identification,

characterization, and mechanism of action. Cancer Res 1988, 48, (5), 1066- 71.

12. Adachi, J.; Kumar, C; Zhang, Y.; Olsen, J. V.; Mann, M., The human urinary proteome contains more than 1500 proteins including a large proportion of membranes proteins. Genome Biol 2006, 7, (9), R80.

13. Omenn, G. S.; States, D. J.; Adamski, M.; Blackwell, T. W.; Menon, R.; Hermjakob, H.; Apweiler, R.; Haab, B. B.; Simpson, R. J.; Eddes, J. S.; Kapp, E. A.; Moritz, R. L.; Chan, D. W.; Rai, A. J.; Admon, A.; Aebersold, R.; Eng, J.;

Hancock, W. S.; Hefta, S. A.; Meyer, H.; Paik, Y. K.; Yoo, J. S.; Ping, P.;

Pounds, J.; Adkins, J.; Qian, X.; Wang, R.; Wasinger, V.; Wu, C. Y.; Zhao, X.;

Zeng, R.; Archakov, A.; Tsugita, A.; Beer, I.; Pandey, A.; Pisano, M.;

Andrews, P.; Tammen, H.; Speicher, D. W.; Hanash, S. M., Overview of the HUPO Plasma Proteome Project: results from the pilot phase with 35 collaborating laboratories and multiple analytical groups, generating a core dataset of 3020 proteins and a publicly-available database. Proteomics 2005,

5, (13), 3226-45. 14. Xu, Y.; Shen, Z.; Wiper, D. W.; Wu, M.; Morton, R. E.; Elson, P.; Kennedy, A. W.; Belinson, J.; Markman, M.; Casey, G., Lysophosphatidic acid as a potential biomarker for ovarian and other gynecologic cancers. Jama 1998, 280, (8), 719-23. 15. Gericke, B.; Raila, J.; Sehouli, J.; Haebel, S.; Konsgen, D.; Mustea, A.; Schweigert, F. J., Microheterogeneity of transthyretin in serum and ascitic fluid of ovarian cancer patients. BMC Cancer 2005, 5, 133.

16. Kislinger, T.; Cox, B.; Kannan, A.; Chung, C; Hu, P.; Ignatchenko, A.; Scott, M. S.; Gramolini, A. O.; Morris, Q.; Hallett, M. T.; Rossant, J.; Hughes, T. R.; Frey, B.; Emili, A., Global survey of organ and organelle protein expression in mouse: combined proteomic and transcriptomic profiling. Cell 2006, 125, (1), 173-86. 17. Kislinger, T.; Gramolini, A. O.; Pan, Y.; Rahman, K.; MacLennan, D. H.; Emili, A., Proteome dynamics during C2C12 myoblast differentiation. MoI Cell Proteomics 2005, 4, (7), 887-901.

18. Kislinger, T.; Rahman, K.; Radulovic, D.; Cox, B.; Rossant, J.; Emili, A., PRISM, a generic large scale proteomic investigation strategy for mammals.

MoI Cell Proteomics 2003, 2, (2), 96-106.

19. Washburn, M. P.; Wolters, D.; Yates, J. R., 3rd, Large-scale analysis of the yeast proteome by multidimensional protein identification technology. Nat Biotechnol 2001 , 19, (3), 242-7.

20. Sun, H.; Fang, H.; Chen, T.; Perkins, R.; Tong, W., GOFFA: Gene Ontology For Functional Analysis - A FDA Gene Ontology Tool for Analysis of Genomic and Proteomic Data. BMC Bioinformatics 2006, 7 Suppl 2, S23. 21. Brown, K. R.; Jurisica, I., Online predicted human interaction database. Bioinformatics 2005, 21 , (9), 2076-82.

22. Shevchenko, A.; WiIm, M.; Vorm, O.; Mann, M., Mass spectrometric sequencing of proteins silver-stained polyacrylamide gels. Anal Chem 1996, 68, (5), 850-8.

23. Craig, R.; Beavis, R. C, TANDEM: matching proteins with tandem mass spectra. Bioinformatics 2004, 20, (9), 1466-7. 24. Peng, J.; Elias, J. E.; Thoreen, C. C; Licklider, L J.; Gygi, S. P., Evaluation of multidimensional chromatography coupled with tandem mass spectrometry (LC/LCMS/MS) for large-scale protein analysis: the yeast proteome. J Proteome Res 2003, 2, (1), 43-50. 25. Elias, J. E.; Gygi, S. P., Target-decoy search strategy for increased confidence in large-scale protein identifications by mass spectrometry. Nat Methods 2007, 4, (3), 207- 14.

26. Adamski, M.; Blackwell, T.; Menon, R.; Martens, L.; Hermjakob, H.; Taylor, C; Omenn, G. S.; States, D. J., Data management and preliminary data analysis in the pilot phase of the HUPO Plasma Proteome Project. Proteomics 2005, 5, (13), 3246-61.

27. Adachi, J.; Kumar, C; Zhang, Y.; Mann, M., In-depth analysis of the adipocyte proteome by mass spectrometry and bioinformatics. MoI Cell

Proteomics 2007.

28. Kislinger, T.; Gramolini, A. O.; MacLennan, D. H.; Emili, A., Multidimensional protein identification technology (MudPIT): technical overview of a profiling method optimized for the comprehensive proteomic investigation of normal and diseased heart tissue. J Am Soc Mass Spectrom

2005, 16, (8), 1207-20.

29. Koziol, J. A.; Feng, A. C; Schnitzer, J. E., Application of capture-recapture models to estimation of protein count in MudPIT experiments. Anal Chem

2006, 78, (9), 3203-7.

30. Liu, H.; Sadygov, R. G.; Yates, J. R., 3rd, A model for random sampling and estimation of relative protein abundance in shotgun proteomics. Anal Chem 2004, 76, (14), 4193-201.

31. Keller, A.; Nesvizhskii, A. I.; Kolker, E.; Aebersold, R., Empirical statistical model to estimate the accuracy of peptide identifications made by MS/MS and database search. Anal Chem 2002, 74, (20), 5383-92. 32. Sadygov, R. G.; Liu, H.; Yates, J. R., Statistical models for protein validation using tandem mass spectral data and protein amino acid sequence databases. Anal Chem 2004, 76, (6), 1664-71.

33. Sadygov, R. G.; Yates, J. R., 3rd, A hypergeometric probability model for protein identification and validation using tandem mass spectral data and protein sequence databases. Anal Chem 2003, 75, (15), 3792-8.

34. Eng, J. K.; McCormack, A. L.; Yates, J. R., An approach to correlate tandem mass-spectral data of peptides with amino-acid-sequences in a protein database. J Am Soc Mass Spectrom 1994, 11 , 976-989.

35. Perkins, D. N.; Pappin, D. J.; Creasy, D. M.; Cottrell, J. S., Probability- based protein identification by searching sequence databases using mass spectrometry data. Electrophoresis 1999, 20, (18), 3551-67.

36. Tabb, D. L.; Fernando, C. G.; Chambers, M. C ₁ MyriMatch: Highly Accurate Tandem Mass Spectral Peptide Identification by Multivariate Hypergeometric Analysis. J Proteome Res 2007, 6, (2), 654-661. 37. Nesvizhskii, A. I.; Keller, A.; Kolker, E.; Aebersold, R., A statistical model for identifying proteins by tandem mass spectrometry. Anal Chem 2003, 75, (17), 4646-58. 38. Old, W. M.; Meyer-Arendt, K.; Aveline-Wolf, L.; Pierce, K. G.; Mendoza, A.; Sevinsky, J. R.; Resing, K. A.; Ahn, N. G., Comparison of label-free methods for quantifying human proteins by shotgun proteomics. MoI Cell Proteomics 2005, 4, (10), 1487-502. 39. Park, G. W.; Kwon, K. H.; Kim, J. Y.; Lee, J. H.; Yun, S. H.; Kim, S. I.; Park, Y. M.; Cho, S. Y.; Paik, Y. K.; Yoo, J. S., Human plasma proteome analysis by reversed sequence database search and molecular weight correlation based on a bacterial proteome analysis. Proteomics 2006, 6, (4), 1121-32.

40. Resing, K. A.; Ahn, N. G., Proteomics strategies for protein identification. FEBS Lett 2005, 579, (4), 885-9.

41. Resing, K. A.; Meyer-Arendt, K.; Mendoza, A. M.; Aveline-Wolf, L. D.; Jonscher, K. R.; Pierce, K. G.; Old, W. M.; Cheung, H. T.; Russell, S.;

Wattawa, J. L.; Goehle, G. R.; Knight, R. D.; Ahn, N. G., Improving

reproducibility and sensitivity in identifying human proteins by shotgun proteomics. Anal Chem 2004, 76, (13), 3556-68.

42. Yang, X.; Dondeti, V.; Dezube, R.; Maynard, D. M.; Geer, L. Y.; Epstein, J.; Chen, X.; Markey, S. P.; Kowalak, J. A., DBParser: web-based software for shotgun proteomic data analyses. J Proteome Res 2004, 3, (5), 1002-8.

43. Tabb, D. L.; McDonald, W. H.; Yates, J. R., 3rd, DTASelect and Contrast: tools for assembling and comparing protein identifications from shotgun proteomics. J Proteome Res 2002, 1 , (1), 21-6.

44. Kapp, E. A.; Schutz, F.; Connolly, L. M.; Chakel, J. A.; Meza, J. E.; Miller, C. A.; Fenyo, D.; Eng, J. K.; Adkins, J. N.; Omenn, G. S.; Simpson, R. J., An evaluation, comparison, and accurate benchmarking of several publicly available MS/MS search algorithms: sensitivity and specificity analysis. Proteomics 2005, 5, (13), 3475-90.

45. Falkner, J. A.; Falkner, J. W.; Andrews, P. C, ProteomeCommons.org JAF: reference information and tools for proteomics. Bioinformatics 2006, 22, (5), 632-3.

46. Anderson, N. L.; Anderson, N. G., The human plasma proteome: history, character, and diagnostic prospects. MoI Cell Proteomics 2002, 1 , (11), 845- 67.

47. Gamberoni, G.; Storari, S.; Volinia, S., Finding biological process modifications in cancer tissues by mining gene expression correlations. BMC Bioinformatics 2006, 7, 6. 48. Patil, M. A.; Chua, M. S.; Pan, K. H.; Lin, R.; Lih, C. J.; Cheung, S. T.; Ho, C; Li, R.; Fan, S. T.; Cohen, S. N.; Chen, X.; So, S., An integrated data analysis approach to characterize genes highly expressed in hepatocellular carcinoma. Oncogene 2005, 24, (23), 3737-47. 49. Chen, E. I.; Hewel, J.; Krueger, J. S.; Tiraby, C; Weber, M. R.; Kralli, A.; Becker, K.; Yates, J. R., 3rd; Felding-Habermann, B., Adaptation of energy metabolism in breast cancer brain metastases. Cancer Res 2007, 67, (4), 1472-86. 50. Engwegen, J. Y.; Gast, M. C; Schellens, J. H.; Beijnen, J. H., Clinical proteomics: searching for better tumour markers with SELDI-TOF mass spectrometry. Trends Pharmacol Sci 2006, 27, (5), 251-9.

51. Gogoi, R.; Srinivasan, S.; Fishman, D. A., Progress in biomarker discovery for diagnostic testing in epithelial ovarian cancer. Expert Rev MoI Diagn 2006, 6, (4), 627- 37.

52. Bast, R. C, Jr.; Klug, T. L.; St John, E.; Jenison, E.; Niloff, J. M.; Lazarus, H.; Berkowitz, R. S.; Leavitt, T.; Griffiths, C. T.; Parker, L.; Zurawski, V. R., Jr.; Knapp, R. C, A radioimmunoassay using a monoclonal antibody to monitor the course of epithelial ovarian cancer. N Engl J Med 1983, 309, (15), 883-7.

53. Fujimoto, J.; Sakaguchi, H.; Hirose, R.; lchigo, S.; Tamaya, T., Biologic implications of the expression of vascular endothelial growth factor subtypes in ovarian carcinoma. Cancer 1998, 83, (12), 2528-33. 54. Yamamoto, S.; Konishi, I.; Mandai, M.; Kuroda, H.; Komatsu, T.; Nanbu, K.; Sakahara, H.; Mori, T., Expression of vascular endothelial growth factor (VEGF) in epithelial ovarian neoplasms: correlation with clinicopathology and patient survival, and analysis of serum VEGF levels. Br J Cancer 1997, 76, (9), 1221-7.

55. Fishman, D. A.; Bafetti, L. M.; Banionis, S.; Kearns, A. S.; Chilukuri, K.; Stack, M. S., Production of extracellular matrix-degrading proteinases by primary cultures of human epithelial ovarian carcinoma cells. Cancer 1997, 80, (8), 1457-63.

56. Moser, T. L.; Young, T. N.; Rodriguez, G. C; Pizzo, S. V.; Bast, R. C ₁ Jr.; Stack, M. S., Secretion of extracellular matrix-degrading proteinases is increased in epithelial ovarian carcinoma, lnt J Cancer 1994, 56, (4), 552-9. 57. Diamandis, E. P.; Yousef, G. M.; Soosaipillai, A. R.; Bunting, P., Human kallikrein 6 (zy me/protease M/neurosin): a new serum biomarker of ovarian carcinoma. Clin Biochem 2000, 33, (7), 579-83.

58. Dong, Y.; Kaushal, A.; Bui, L; Chu, S.; Fuller, P. J.; Nicklin, J.; Samaratunga, H.; Clements, J. A., Human kallikrein 4 (KLK4) is highly expressed in serous ovarian carcinomas. Clin Cancer Res 2001 , 7, (8), 2363- 71.

59. Luo, L. Y.; Bunting, P.; Scorilas, A.; Diamandis, E. P., Human kallikrein 10: a novel tumor marker for ovarian carcinoma? Clin Chim Acta 2001 , 306,

(1-2), 111-8.

60. Makino, E.; Sakaguchi, M.; Iwatsuki, K.; Huh, N. H., Introduction of an Nterminal peptide of S100C/A11 into human cells induces apoptotic cell death. J MoI Med 2004, 82, (9), 612-20.

61. Kikuchi, N.; Horiuchi, A.; Osada, R.; Imai, T.; Wang, C; Chen, X.; Konishi, I., Nuclear expression of S100A4 is associated with aggressive behavior of epithelial ovarian carcinoma: an important autocrine/paracrine factor in tumor progression. Cancer Sci 2006, 97, (10), 1061-9.

62. Trachte, A. L.; Suthers, S. E.; Lerner, M. R.; Hanas, J. S.; Jupe, E. R.; Sienko, A. E.; Adesina, A. M.; Lightfoot, S. A.; Brackett, D. J.; Postier, R. G., Increased expression of alpha-1 -antitrypsin, glutathione S-transferase pi and vascular endothelial growth factor in human pancreatic adenocarcinoma. Am J Surg 2002, 184, (6), 642-7; discussion 647-8.

63. Boss, E. A.; Peters, W. H.; Roelofs, H. M.; Boonstra, H.; Steegers, E. A.; Massuger, L. F., Glutathione S-transferases P1-1 and A1-1 in ovarian cyst fluids. Eur J Gynaecol Oncol 2001 , 22, (6), 427-32.

64. Hough, C. D.; Sherman-Baust, C. A.; Pizer, E. S.; Montz, F. J.; Im, D. D.; Rosenshein, N. B.; Cho, K. R.; Riggins, G. J.; Morin, P. J., Large-scale serial analysis of gene expression reveals genes differentially expressed in ovarian cancer. Cancer Res 2000, 60, (22), 6281-7.

65. Chen, Y. C; Pohl, G.; Wang, T. L.; Morin, P. J.; Risberg, B.; Kristensen, G. B.; Yu, A.; Davidson, B.; Shih Ie, M., Apolipoprotein E is required for cell proliferation and survival in ovarian cancer. Cancer Res 2005, 65, (1), 331-7. 66. Tavazoie, S. F.; Alvarez, V. A.; Ridenour, D. A.; Kwialkowski, D. J.;

Sabatini, B.

L., Regulation of neuronal morphology and function by the tumor suppressors

Tsc1 and

Tsc2. Nat Neurosci 2005, 8, (12), 1727-34.

67. Wang, W.; Mouneimne, G.; Sidani, M.; Wyckoff, J.; Chen, X.; Makris, A.; Goswami, S.; Bresnick, A. R.; Condeelis, J. S., The activity status of cofilin is directly related to invasion, intravasation, and metastasis of mammary tumors. J Cell Biol 2006, 173, (3), 395-404.

68. Mataraza, J. M.; Briggs, M. W.; Li, Z.; Entwistle, A.; Ridley, A. J.; Sacks, D. B., IQGAP1 promotes cell motility and invasion. J Biol Chem 2003, 278, (42), 41237-45. 69. Meng, X.; Krokhin, O.; Cheng, K.; Ens, W.; Wilkins, J. A., Characterization of IQGAP1 -containing complexes in NK-like cells: evidence for Rac 2 and RACK1 association during homotypic adhesion. J Proteome Res 2007, 6, (2), 744-50. 70. Yamaoka-Tojo, M.; Ushio-Fukai, M.; Hilenski, L.; Dikalov, S. I.; Chen, Y. E.; Tojo, T.; Fukai, T.; Fujimoto, M.; Patrushev, N. A.; Wang, N.; Kontos, C. D.; Bloom, G. S.; Alexander, R. W., IQGAP1 , a novel vascular endothelial growth factor receptor binding protein, is involved in reactive oxygen species- dependent endothelial migration and proliferation. Circ Res 2004, 95, (3), 276- 83.

71. Yamaoka-Tojo, M.; Tojo, T.; Kim, H. W.; Hilenski, L.; Patrushev, N. A.; Zhang, L.; Fukai, T.; Ushio-Fukai, M., IQGAP1 mediates VE-cadherin-based cell-cell contacts and VEGF signaling at adherence junctions linked to angiogenesis. Arterioscler Thromb Vase Biol 2006, 26, (9), 1991-7.

72. Babcock, G.; Rubenstein, P. A., Control of profilin and actin expression in muscle and nonmuscle cells. Cell Motil Cytoskeleton 1993, 24, (3), 179-88.

73. Moldovan, N. I.; Milliken, E. E.; Irani, K.; Chen, J.; Sohn, R. H.; Finkel, T.; Goldschmidt-Clermont, P. J., Regulation of endothelial cell adhesion by profilin. Curr Biol 1997, 7, (1), 24-30.

74. Roy, P.; Jacobson, K., Overexpression of profilin reduces the migration of invasive breast cancer cells. Cell Motil Cytoskeleton 2004, 57, (2), 84-95.

75. Luza, S. C; Speisky, H. C, Liver copper storage and transport during development: implications for cytotoxicity. Am J Clin Nutr 1996, 63, (5), 812S- 2OS. 76. Kluger, H. M.; Kluger, Y.; Gilmore-Hebert, M.; DiVito, K.; Chang, J. T.; Rodov, S.; Mironenko, O.; Kacinski, B. M.; Perkins, A. S.; Sapi, E., cDNA microarray analysis of invasive and tumorigenic phenotypes in a breast cancer model. Lab Invest 2004, 84, (3), 320-31. 77. Sergeev, A. I.; Naidich, V. I.; Bunto, T. V., [Study of the kinetics of ascitic tumor development by proton magnetic relaxation]. Dokl Akad Nauk SSSR 1978, 238, (4), 974- 6.

78. Vachtenheim, J.; Pavel, S.; Duchon, J., Dopa oxidase activity and ceruloplasmin in the sera of hamsters with melanoma. Neoplasma 1981 , 28,

(1), 59-65.

79. Yeshowardhana; Singh, V. S., Significance of serum phosphohexose isomerase, hexokinase and aldolase in carcinoma ovary. Indian J Physiol Pharmacol 1985, 29, (1), 51-4.

80. Torimura, T.; Ueno, T.; Kin, M.; Harada, R.; Nakamura, T.; Kawaguchi, T.; Harada, M.; Kumashiro, R.; Watanabe, H.; Avraham, R.; Sata, M., Autocrine motility factor enhances hepatoma cell invasion across the basement membrane through activation of betai integrins. Hepatology 2001 , 34, (1), 62- 71.

81. Haga, A., [Possibility that AMF will serve as a target molecule for the diagnosis and treatment of a metastatic neoplasm]. Yakugaku Zasshi 2005,

125, (2), 169-75.

82. Leto, G.; Gebbia, N.; Rausa, L.; Tumminello, F. M., Cathepsin D in the malignant progression of neoplastic diseases (review). Anticancer Res 1992, 12, (1), 235-40. 83. Losch, A.; Schindl, M.; Kohlberger, P.; Lahodny, J.; Breitenecker, G.; Horvat, R.; Birner, P., Cathepsin D in ovarian cancer: prognostic value and correlation with p53 expression and microvessel density. Gynecol Oncol 2004, 92, (2), 545-52.

84. de Angelis, P. M.; Fjell, B.; Kravik, K. L.; Haug, T.; Tunheim, S. H.; Reichelt, W.; Beigi, M.; Clausen, O. P.; Galteland, E.; Stokke, T., Molecular characterizations of derivatives of HCT116 colorectal cancer cells that are resistant to the chemotherapeutic agent 5-fluorouracil. lnt J Oncol 2004, 24, (5), 1279-88.

85. Ozalp, S.; Tanir, H. M.; Yalcin, O. T.; Kabukcuoglu, S.; Oner, U.; Uray, M., Prognostic value of matrix metalloproteinase-9 (gelatinase-B) expression in epithelial ovarian tumors. Eur J Gynaecol Oncol 2003, 24, (5), 417-20.

86. Bachvarov, D.; L'Esperance, S.; Popa, I.; Bachvarova, M.; Plante, M.; Tetu, B., Gene expression patterns of chemoresistant and chemosensitive serous epithelial ovarian tumors with possible predictive value in response to initial chemotherapy, lnt J Oncol 2006, 29, (4), 919-33.

87. Jazaeri, A. A.; Awtrey, C. S.; Chandramouli, G. V.; Chuang, Y. E.; Khan, J.; Sotiriou, C; Aprelikova, O.; Yee, C. J.; Zorn, K. K.; Birrer, M. J.; Barrett, J. C; Boyd, J., Gene expression profiles associated with response to chemotherapy in epithelial ovarian cancers. Clin Cancer Res 2005, 11 , (17), 6300-10.

88. Spentzos, D.; Levine, D. A.; Kolia, S.; Otu, H.; Boyd, J.; Libermann, T. A.; Cannistra, S. A., Unique gene expression profile based on pathologic response in epithelial ovarian cancer. J Clin Oncol 2005, 23, (31), 7911-8.

89. Le Moguen, K.; Lincet, H.; Deslandes, E.; Hubert-Roux, M.; Lange, C; Poulain, L.; Gauduchon, P.; Baudin, B., Comparative proteomic analysis of cisplatin sensitive IGROV1 ovarian carcinoma cell line and its resistant counterpart IGROV1-R10. Proteomics 2006, 6, (19), 5183-92.

90. Stewart, J. J.; White, J. T.; Yan, X.; Collins, S.; Drescher, C. W.; Urban, N. D.; Hood, L.; Lin, B., Proteins associated with Cisplatin resistance in ovarian cancer cells identified by quantitative proteomic technology and integrated with mRNA expression levels. MoI Cell Proteomics 2006, 5, (3), 433-43.

91. States, D. J.; Omenn, G. S.; Blackwell, T. W.; Fermin, D.; Eng, J.; Speicher, D. W.; Hanash, S. M., Challenges in deriving high-confidence

protein identifications from data gathered by a HUPO plasma proteome collaborative study. Nat Biotechnol 2006, 24,(3), 333-8.

92. Roby KF, Taylor CC, Sweetwood JP, Cheng Y, Pace JL, Tawfik O, Persons DL, Smith PG, Terranova PF. Development of a syngenic mouse model for events related to ovarian cancer. Carcinogenesis 2000, 21(4), 585- 591.