A personalized probabilistic approach to ovarian cancer diagnostics.

OBJECTIVE: The identification/development of a machine learning-based classifier that utilizes metabolic profiles of serum samples to accurately identify individuals with ovarian cancer. METHODS: Serum samples collected from 431 ovarian cancer patients and 133 normal women at four geographic locations were analyzed by mass spectrometry. Reliable metabolites were identified using recursive feature elimination coupled with repeated cross-validation and used to develop a consensus classifier able to distinguish cancer from non-cancer. The probabilities assigned to individuals by the model were used to create a clinical tool that assigns a likelihood that an individual patient sample is cancer or normal. RESULTS: Our consensus classification model is able to distinguish cancer from control samples with 93% accuracy. The frequency distribution of individual patient scores was used to develop a clinical tool that assigns a likelihood that an individual patient does or does not have cancer. CONCLUSIONS: An integrative approach using metabolomic profiles and machine learning-based classifiers has been employed to develop a clinical tool that assigns a probability that an individual patient does or does not have ovarian cancer. This personalized/probabilistic approach to cancer diagnostics is more clinically informative and accurate than traditional binary (yes/no) tests and represents a promising new direction in the early detection of ovarian cancer.

High throughput, label-free isolation of circulating tumor cell clusters in meshed microwells.

Extremely rare circulating tumor cell (CTC) clusters are both increasingly appreciated as highly metastatic precursors and virtually unexplored. Technologies are primarily designed to detect single CTCs and often fail to account for the fragility of clusters or to leverage cluster-specific markers for higher sensitivity. Meanwhile, the few technologies targeting CTC clusters lack scalability. Here, we introduce the Cluster-Wells, which combines the speed and practicality of membrane filtration with the sensitive and deterministic screening afforded by microfluidic chips. The >100,000 microwells in the Cluster-Wells physically arrest CTC clusters in unprocessed whole blood, gently isolating virtually all clusters at a throughput of >25 mL/h, and allow viable clusters to be retrieved from the device. Using the Cluster-Wells, we isolated CTC clusters ranging from 2 to 100+ cells from prostate and ovarian cancer patients and analyzed a subset using RNA sequencing. Routine isolation of CTC clusters will democratize research on their utility in managing cancer.

Label-free microfluidic enrichment of cancer cells from non-cancer cells in ascites.

The isolation of a patient's metastatic cancer cells is the first, enabling step toward treatment of that patient using modern personalized medicine techniques. Whereas traditional standard-of-care approaches select treatments for cancer patients based on the histological classification of cancerous tissue at the time of diagnosis, personalized medicine techniques leverage molecular and functional analysis of a patient's own cancer cells to select treatments with the highest likelihood of being effective. Unfortunately, the pure populations of cancer cells required for these analyses can be difficult to acquire, given that metastatic cancer cells typically reside in fluid containing many different cell populations. Detection and analyses of cancer cells therefore require separation from these contaminating cells. Conventional cell sorting approaches such as Fluorescence Activated Cell Sorting or Magnetic Activated Cell Sorting rely on the presence of distinct surface markers on cells of interest which may not be known nor exist for cancer applications. In this work, we present a microfluidic platform capable of label-free enrichment of tumor cells from the ascites fluid of ovarian cancer patients. This approach sorts cells based on differences in biomechanical properties, and therefore does not require any labeling or other pre-sort interference with the cells. The method is also useful in the cases when specific surface markers do not exist for cells of interest. In model ovarian cancer cell lines, the method was used to separate invasive subtypes from less invasive subtypes with an enrichment of ~ sixfold. In ascites specimens from ovarian cancer patients, we found the enrichment protocol resulted in an improved purity of P53 mutant cells indicative of the presence of ovarian cancer cells. We believe that this technology could enable the application of personalized medicine based on analysis of liquid biopsy patient specimens, such as ascites from ovarian cancer patients, for quick evaluation of metastatic disease progression and determination of patient-specific treatment.

A phase II randomized, double-blind trial of a polyvalent Vaccine-KLH conjugate (NSC 748933 IND# 14384) + OPT-821 versus OPT-821 in patients with epithelial ovarian, fallopian tube, or peritoneal cancer who are in second or third complete remission: An NRG Oncology/GOG study.

OBJECTIVE: Early-phase data have demonstrated induction of antibody responses to a polyvalent vaccine conjugate (Globo-H, GM2, MUC1-TN, TF) with adjuvant OPT-821. We sought to determine if this combination decreases the hazard of progression or death compared to OPT-821 alone in patients with ovarian cancer in second/third clinical complete remission following chemotherapy. Secondary and translational objectives were overall survival (OS), safety, and immunogenicity. METHODS: From 2010-2013, patients were randomized (1:1) to receive OPT-821±vaccine-KLH conjugate subcutaneously at weeks 1, 2, 3, 7, 11, and then every 12 weeks (total 11). Dose delay or reduction was not permitted. Patients were removed for pre-defined dose-limiting toxicity. RESULTS: Of 171 patients randomized, 170 were treated. Most had disease of serous histology (85%), stage 3 disease at diagnosis (77%), and had received 2 prior regimens (68%). 32% received >6 treatment cycles [median 6, each arm (p = 0.33)]. 77% discontinued due to progression, 4% due to toxicity, and 1 due to myeloid dysplastic syndrome (MDS). Maximum toxicities included grade 4 MDS and depression/personality change (1 each, unlikely related), as well as grade 3 gastrointestinal disorders and others (n = 21, 4 related). Lesser adverse events were injection site reactions (82%) and fever (11%). Estimated HR for progression-free survival (PFS) of the vaccine + OPT-821 to OPT-821 arm was 0.98 (95% CI: 0.71-1.36). At a median follow-up of 60 months, median OS was 47 and 46 months, respectively. CONCLUSIONS: Vaccine + OPT-821 compared to OPT-821 alone was modestly immunogenic and did not prolong PFS or OS. Multi-remission patients are a viable, well-defined population for exploring innovative consolidation and maintenance approaches. TRIAL REGISTRATION: NCT00857545.

Machine learning predicts individual cancer patient responses to therapeutic drugs with high accuracy.

Precision or personalized cancer medicine is a clinical approach that strives to customize therapies based upon the genomic profiles of individual patient tumors. Machine learning (ML) is a computational method particularly suited to the establishment of predictive models of drug response based on genomic profiles of targeted cells. We report here on the application of our previously established open-source support vector machine (SVM)-based algorithm to predict the responses of 175 individual cancer patients to a variety of standard-of-care chemotherapeutic drugs from the gene-expression profiles (RNA-seq or microarray) of individual patient tumors. The models were found to predict patient responses with >80% accuracy. The high PPV of our algorithms across multiple drugs suggests a potential clinical utility of our approach, particularly with respect to the identification of promising second-line treatments for patients failing standard-of-care first-line therapies.

Evidence for the importance of post-transcriptional regulatory changes in ovarian cancer progression and the contribution of miRNAs.

High-throughput technologies have identified significant changes in patterns of mRNA expression over cancer development but the functional significance of these changes often rests upon the assumption that observed changes in levels of mRNA accurately reflect changes in levels of their encoded proteins. We systematically compared the expression of 4436 genes on the RNA and protein levels between discrete tumor samples collected from the ovary and from the omentum of the same OC patient. The overall correlation between global changes in levels of mRNA and their encoding proteins is low (r = 0.38). The majority of differences are on the protein level with no corresponding change on the mRNA level. Indirect and direct evidence indicates that a significant fraction of the differences may be mediated by microRNAs.

Highly-accurate metabolomic detection of early-stage ovarian cancer.

High performance mass spectrometry was employed to interrogate the serum metabolome of early-stage ovarian cancer (OC) patients and age-matched control women. The resulting spectral features were used to establish a linear support vector machine (SVM) model of sixteen diagnostic metabolites that are able to identify early-stage OC with 100% accuracy in our patient cohort. The results provide evidence for the importance of lipid and fatty acid metabolism in OC and serve as the foundation of a clinically significant diagnostic test.

Glycomic analysis of membrane glycoproteins with bisecting glycosylation from ovarian cancer tissues reveals novel structures and functions.

Biomarkers capable of detecting and targeting epithelial ovarian cancer cells for diagnostics and therapeutics would be extremely valuable. Ovarian cancer is the deadliest reproductive malignancy among women in the U.S., killing over 14 000 women each year. Both the lack of presenting symptoms and high mortality rates illustrate the need for earlier diagnosis and improved treatment of this disease. The glycosyltransferase enzyme GnT-III encoded by the Mgat3 gene is responsible for the addition of GlcNAc (N-acetylglucosamine) to form bisecting N-linked glycan structures. GnT-III mRNA expression is amplified in ovarian cancer tissues compared with normal ovarian tissue. We use a lectin capture strategy coupled to nano-ESI-RPLC-MS/MS to isolate and identify the membrane glycoproteins and unique glycan structures associated with GnT-III amplification in human ovarian cancer tissues. Our data illustrate that the majority of membrane glycoproteins with bisecting glycosylation are common to both serous and endometrioid histological subtypes of ovarian cancer, and several have been reported to participate in signaling pathways such as Notch, Wnt, and TGFß.

p66Shc longevity protein regulates the proliferation of human ovarian cancer cells.

p66Shc functions as a longevity protein in murine and exhibits oxidase activity in regulating diverse biological activities. In this study, we investigated the role of p66Shc protein in regulating ovarian cancer (OCa) cell proliferation. Among three cell lines examined, the slowest growing OVCAR-3 cells have the lowest level of p66Shc protein. Transient transfection with p66Shc cDNA expression vector in OVCAR-3 cells increases cell proliferation. Conversely, knock-down of p66Shc by shRNA in rapidly growing SKOV-3 cells results in decreased cell growth. In estrogen (E2)-treated CaOV-3 cells, elevated p66Shc protein level correlates with ROS level, ErbB-2 and ERK/MAPK activation, and cell proliferation. Further, the E2-stimulated proliferation of CaOV-3 cells was blocked by antioxidants and ErbB-2 inhibitor. Additionally, in E2-stimulated cells, the tartrate-sensitive, but not the tartrate-resistant, phosphatase activity decreases; concurrently, the tyrosine phosphorylation of ErbB-2 increases. Conversely, inhibition of phosphatase activity by L(+)-tartrate treatment increases p66Shc protein level, ErbB-2 tyrosine phosphorylation, ERK/MAPK activation, and cell growth. Further, inhibition of the ERK/MAPK pathway by PD98059 blocks E2-induced ERK/MAPK activation and cell proliferation in CaOV-3 cells. Moreover, immunohistochemical analyses showed that the p66Shc protein level was significantly higher in cancerous cells than in noncancerous cells in archival OCa tissues (n = 76; P = 0.00037). These data collectively indicate that p66Shc protein plays a critical role in up-regulating OCa progression.

Ectopic over-expression of miR-429 induces mesenchymal-to-epithelial transition (MET) and increased drug sensitivity in metastasizing ovarian cancer cells.

OBJECTIVE: We recently determined that the ectopic over-expression of miR-429 and other members of the miR-200 family of microRNAs in ovarian cancer (OC) mesenchymal-like cell lines induces mesenchymal-to-epithelial transition (MET) with a concomitant increase in sensitivity to platinum drugs. We sought to determine if metastasizing OC cells isolated from an OC patient could also be induced by miR-429 to undergo MET and become sensitized to established first-line platinum-based therapies. METHODS: We established and characterized a new primary cell line (OCI-984) from free-floating OC cells isolated from the ascites fluid of an advanced stage OC patient. miR-429 was ectopically over-expressed in these cells. RESULTS: The over-expression of miR-429 in OCI-984 cells induced morphological, functional and molecular changes consistent with MET and a concomitant significant increase in the sensitivity of the converted cells to cisplatin. CONCLUSIONS: Our findings indicate that the miR-200 family of microRNAs, and miR-429 in particular, play a critical role in the functioning of OC metastasizing cells and that targeted delivery of miR-429, and perhaps other miR-200 family members, in combination with platinum-based chemotherapies may be an effective strategy in reducing OC metastasis and tumor recurrence.

Transcriptional override: a regulatory network model of indirect responses to modulations in microRNA expression.

BACKGROUND: Documented changes in levels of microRNAs (miRNA) in a variety of diseases including cancer are leading to their development as early indicators of disease, and as a potential new class of therapeutic agents. A significant hurdle to the rational application of miRNAs as therapeutics is our current inability to reliably predict the range of molecular and cellular consequences of perturbations in the levels of specific miRNAs on targeted cells. While the direct gene (mRNA) targets of individual miRNAs can be computationally predicted with reasonable degrees of accuracy, reliable predictions of the indirect molecular effects of perturbations in miRNA levels remain a major challenge in molecular systems biology. RESULTS: Changes in gene (mRNA) and miRNA expression levels between normal precursor and ovarian cancer cells isolated from patient tissue samples were measured by microarray. Expression of 31 miRNAs was significantly elevated in the cancer samples. Consistent with previous reports, the expected decrease in expression of the mRNA targets of upregulated miRNAs was observed in only 20-30% of the cancer samples. We present and provide experimental support for a network model (The Transcriptional Override Model; TOM) to account for the unexpected regulatory consequences of modulations in the expression of miRNAs on expression levels of their target mRNAs in ovarian cancer. CONCLUSIONS: The direct and indirect regulatory effects of changes in miRNA expression levels in vivo are interactive and complex but amenable to systems level modeling. Although TOM has been developed and validated within the context of ovarian cancer, it may be applicable in other biological contexts as well, including of potential future use in the rational design of miRNA-based strategies for the treatment of cancers and other diseases.

High-dose chemotherapy with autologous stem cell support as salvage therapy in recurrent gestational trophoblastic disease.

BACKGROUND: Gestational trophoblastic disease usually follows a molar pregnancy but can occur also after an abortion or a term pregnancy. In only 10% of cases will treatment be required; and usually, single-agent chemotherapy will suffice. In high-risk disease, the multiagent regimen EMA-CO is usually used; and if that fails, most oncologists will use the EMA-EP regimen. If this does not produce a remission, there is no unanimity of opinion as to how to proceed. Numerous salvage regimens are in current use, and some centers do not consider high-dose chemotherapy. CASE: A young woman presented 4 months after a normal spontaneous delivery with an elevated human chorionic gonadotropin level and multiple pulmonary metastases. She failed both the EMA-CO and EMA-EP regimens as well as additional standard chemotherapy. She was then treated with 4 separate courses of high-dose chemotherapy with autologous stem cell support, which produced a complete remission. CONCLUSION: Even patients with high-risk gestational trophoblastic disease are usually cured with standard chemotherapy. Patients who fail such treatment should be considered for high-dose chemotherapy.

Molecular profiling supports the role of epithelial-to-mesenchymal transition (EMT) in ovarian cancer metastasis.

BACKGROUND: While metastasis ranks among the most lethal of all cancer-associated processes, on the molecular level, it remains one of the least well understood. One model that has gained credibility in recent years is that metastasizing cells at least partially recapitulate the developmental process of epithelial-to-mesenchymal transition (EMT) in their transit from primary to metastatic sites. While experimentally supported by cell culture and animal model studies, the lack of unambiguous confirmatory evidence in cancer patients has led to persistent challenges to the model's relevance in humans. METHODS: Gene expression profiling (Affymetrix, U133) was carried out on 14 matched sets of primary (ovary) and metastatic (omentum) ovarian cancer (serous adenocarcinoma) patient samples. Hierarchical clustering and functional pathway algorithms were used in the data analysis. RESULTS: While histological examination reveled no morphological distinction between the matched sets of primary and metastatic samples, gene expression profiling clearly distinguished two classes of metastatic samples. One class displayed expression patterns statistically indistinguishable from primary samples isolated from the same patients while a second class displayed expression patterns significantly different from primary samples. Further analyses focusing on genes previously associated with EMT clearly distinguished the primary from metastatic samples in all but one patient. CONCLUSION: Our results are consistent with a role of EMT in most if not all ovarian cancer metastases and demonstrate that identical morphologies between primary and metastatic cancer samples is insufficient evidence to negate a role of EMT in the metastatic process.

Molecular profiling predicts the existence of two functionally distinct classes of ovarian cancer stroma.

Although stromal cell signaling has been shown to play a significant role in the progression of many cancers, relatively little is known about its importance in modulating ovarian cancer development. The purpose of this study was to investigate the process of stroma activation in human ovarian cancer by molecular analysis of matched sets of cancer and surrounding stroma tissues. RNA microarray profiling of 45 tissue samples was carried out using the Affymetrix (U133 Plus 2.0) gene expression platform. Laser capture microdissection (LCM) was employed to isolate cancer cells from the tumors of ovarian cancer patients (Cepi) and matched sets of surrounding cancer stroma (CS). For controls, ovarian surface epithelial cells (OSE) were isolated from the normal (noncancerous) ovaries and normal stroma (NS). Hierarchical clustering of the microarray data resulted in clear separations between the OSE, Cepi, NS, and CS samples. Expression patterns of genes encoding signaling molecules and compatible receptors in the CS and Cepi samples indicate the existence of two subgroups of cancer stroma (CS) with different propensities to support tumor growth. Our results indicate that functionally significant variability exists among ovarian cancer patients in the ability of the microenvironment to modulate cancer development.

Evidence for the complexity of microRNA-mediated regulation in ovarian cancer: a systems approach.

MicroRNAs (miRNAs) are short (∼22 nucleotides) regulatory RNAs that can modulate gene expression and are aberrantly expressed in many diseases including cancer. Previous studies have shown that miRNAs inhibit the translation and facilitate the degradation of their targeted messenger RNAs (mRNAs) making them attractive candidates for use in cancer therapy. However, the potential clinical utility of miRNAs in cancer therapy rests heavily upon our ability to understand and accurately predict the consequences of fluctuations in levels of miRNAs within the context of complex tumor cells. To evaluate the predictive power of current models, levels of miRNAs and their targeted mRNAs were measured in laser captured micro-dissected (LCM) ovarian cancer epithelial cells (CEPI) and compared with levels present in ovarian surface epithelial cells (OSE). We found that the predicted inverse correlation between changes in levels of miRNAs and levels of their mRNA targets held for only ∼11% of predicted target mRNAs. We demonstrate that this low inverse correlation between changes in levels of miRNAs and their target mRNAs in vivo is not merely an artifact of inaccurate miRNA target predictions but the likely consequence of indirect cellular processes that modulate the regulatory effects of miRNAs in vivo. Our findings underscore the complexities of miRNA-mediated regulation in vivo and the necessity of understanding the basis of these complexities in cancer cells before the therapeutic potential of miRNAs can be fully realized.

Rapid mass spectrometric metabolic profiling of blood sera detects ovarian cancer with high accuracy.

BACKGROUND: Ovarian cancer diagnosis is problematic because the disease is typically asymptomatic, especially at the early stages of progression and/or recurrence. We report here the integration of a new mass spectrometric technology with a novel support vector machine computational method for use in cancer diagnostics, and describe the application of the method to ovarian cancer. METHODS: We coupled a high-throughput ambient ionization technique for mass spectrometry (direct analysis in real-time mass spectrometry) to profile relative metabolite levels in sera from 44 women diagnosed with serous papillary ovarian cancer (stages I-IV) and 50 healthy women or women with benign conditions. The profiles were input to a customized functional support vector machine-based machine-learning algorithm for diagnostic classification. Performance was evaluated through a 64-30 split validation test and with a stringent series of leave-one-out cross-validations. RESULTS: The assay distinguished between the cancer and control groups with an unprecedented 99% to 100% accuracy (100% sensitivity and 100% specificity by the 64-30 split validation test; 100% sensitivity and 98% specificity by leave-one-out cross-validations). CONCLUSION: The method has significant clinical potential as a cancer diagnostic tool. Because of the extremely low prevalence of ovarian cancer in the general population (approximately 0.04%), extensive prospective testing will be required to evaluate the test's potential utility in general screening applications. However, more immediate applications might be as a diagnostic tool in higher-risk groups or to monitor cancer recurrence after therapeutic treatment. IMPACT: The ability to accurately and inexpensively diagnose ovarian cancer will have a significant positive effect on ovarian cancer treatment and outcome.

Identification of candidate biomarkers with cancer-specific glycosylation in the tissue and serum of endometrioid ovarian cancer patients by glycoproteomic analysis.

Epithelial ovarian cancer is diagnosed less than 25% of the time when the cancer is confined to the ovary, leading to 5-year survival rates of less than 30%. Therefore, there is an urgent need for early diagnostics for ovarian cancer. Our study using glycotranscriptome comparative analysis of endometrioid ovarian cancer tissue and normal ovarian tissue led to the identification of distinct differences in the transcripts of a restricted set of glycosyltransferases involved in N-linked glycosylation. Utilizing lectins that bind to glycan structures predicted to show changes, we observed differences in lectin-bound glycoproteins consistent with some of the transcript differences. In this study, we have extended our observations by the use of selected lectins to perform a targeted glycoproteomic analysis of ovarian cancer and normal ovarian tissues. Our results have identified several glycoproteins that display tumor-specific glycosylation changes. We have verified these glycosylation changes on glycoproteins from tissue using immunoprecipitation followed by lectin blot detection. The glycoproteins that were verified were then analyzed further using existing microarray data obtained from benign ovarian adenomas, borderline ovarian adenocarcinomas, and malignant ovarian adenocarcinomas. The verified glycoproteins found to be expressed above control levels in the microarray data sets were then screened for tumor-specific glycan modifications in serum from ovarian cancer patients. Results obtained from two of these glycoprotein markers, periostin and thrombospondin, have confirmed that tumor-specific glycan changes can be used to distinguish ovarian cancer patient serum from normal serum.

Selective removal of ovarian cancer cells from human ascites fluid using magnetic nanoparticles.

A majority of ovarian cancer metastases result from the shedding of malignant cells from the primary tumor into the abdominal cavity. Free-floating cancer cells in serous effusions of late-stage ovarian cancer patients may spread to internal organs, making effective treatment extremely difficult. Selective removal of ovarian cancer cells from serous fluids may abate metastasis and improve long-term prognoses. We have already shown that superparamagnetic nanoparticles conjugated to an ephrin-A1 mimetic peptide with a high affinity for the EphA2 receptor can be used to capture and remove cultured human ovarian cancer cells from the peritoneal of experimental mice. Here we demonstrate the potential clinical utility of the methodology by in vitro capture and isolation of cancer cells from the ascites fluid of ovarian cancer patients. FROM THE CLINICAL EDITOR: Ovarian cancer metastases usually are the result of shedding of malignant cells from the primary tumor into the abdominal cavity. In this paper, a novel nanotechnology-based method is demonstrated for the in vitro capture and isolation of cancer cells from the ascites fluid of ovarian cancer patients.

Gene expression profiling supports the hypothesis that human ovarian surface epithelia are multipotent and capable of serving as ovarian cancer initiating cells.

BACKGROUND: Accumulating evidence suggests that somatic stem cells undergo mutagenic transformation into cancer initiating cells. The serous subtype of ovarian adenocarcinoma in humans has been hypothesized to arise from at least two possible classes of progenitor cells: the ovarian surface epithelia (OSE) and/or an as yet undefined class of progenitor cells residing in the distal end of the fallopian tube. METHODS: Comparative gene expression profiling analyses were carried out on OSE removed from the surface of normal human ovaries and ovarian cancer epithelial cells (CEPI) isolated by laser capture micro-dissection (LCM) from human serous papillary ovarian adenocarcinomas. The results of the gene expression analyses were randomly confirmed in paraffin embedded tissues from ovarian adenocarcinoma of serous subtype and non-neoplastic ovarian tissues using immunohistochemistry. Differentially expressed genes were analyzed using gene ontology, molecular pathway, and gene set enrichment analysis algorithms. RESULTS: Consistent with multipotent capacity, genes in pathways previously associated with adult stem cell maintenance are highly expressed in ovarian surface epithelia and are not expressed or expressed at very low levels in serous ovarian adenocarcinoma. Among the over 2000 genes that are significantly differentially expressed, a number of pathways and novel pathway interactions are identified that may contribute to ovarian adenocarcinoma development. CONCLUSIONS: Our results are consistent with the hypothesis that human ovarian surface epithelia are multipotent and capable of serving as the origin of ovarian adenocarcinoma. While our findings do not rule out the possibility that ovarian cancers may also arise from other sources, they are inconsistent with claims that ovarian surface epithelia cannot serve as the origin of ovarian cancer initiating cells.