Project 1: Early Detection of Epithelial Ovarian Cancer

Robert C. Bast, Jr., M. D.
Project Leader

Karen Lu, M.D.
Co-Project Leader


The goal of SPORE Project 1 is to develop effective strategies for early detection of ovarian cancer in women at average risk for the disease. Given the prevalence of ovarian cancer among postmenopausal women, a diagnostic strategy must be moderately sensitive (>75%), but highly specific (>99.6%) to achieve a positive predictive value of 10%, i.e., 10 laparotomies for each case of ovarian cancer detected. One of the most promising approaches to early detection of this neoplasm is to use rising values of a serum marker such as CA 125 to prompt the performance of transvaginal sonography (TVS). Patients with abnormal TVS or a sufficiently rapid rise in CA 125 undergo exploratory surgery. During the last grant period, we have evaluated this strategy and found a positive predictive value of 30%. If one is to pursue a two-stage strategy for early detection, the initial stage must be optimally sensitive. No single marker is likely to be adequately sensitive and multiple markers may be required to detect the full spectrum of ovarian cancers. Simple addition of multiple markers may increase sensitivity, but generally decreases specificity, posing a particular problem in a disease with the prevalence of ovarian cancer. We have identified a panel of four biomarkers (CA125, HE4, CEA and sVCAM-1) that detect 87% of early stage disease at 98% specificity.

When ovarian cancer is detected in stage I, up to 90% of patients can be cured with currently available cytoreductive surgery and combination chemotherapy, whereas advanced disease (stage III-IV) can be cured in less than 30% of cases. Only 20-25% of patients present in early stage. Detection of a larger fraction of women with early stage or pre-clinical ovarian cancer could impact substantially on the rate of cure.

Specific Aims

  • Aim 1: To evaluate the specificity and positive predictive value of an algorithm for early detection of ovarian cancer based on a panel of serum biomarkers measured annually. Multiple biomarkers promise to increase sensitivity for early stage and preclinical disease, provided that specificity is maintained, permitting performance of TVS in a small fraction of participants. While evaluating sensitivity would require a very large trial, we can test the hypothesis that rising values of multiple biomarkers will prompt the referral of no more than 2% of women for TVS and the overall strategy will achieve a positive predictive value >10%.
  • Aim 2: To develop a “point of service” test for multiple serum biomarkers using a lab-on-a-chip sensor system that is performed on blood from a finger-stick. Utilization of screening could be improved if venipuncture were not required and serum analysis could be performed in a sufficiently short time to permit TVS on the same visit. We will test the hypothesis that measurement of the levels of the four biomarkers with a nanobiochip (a complete Lab-on-a-Chip system) will correlate with standard laboratory assays.
  • Aim 3: To identify a multi-marker panel of serum autoantibodies for early detection of ovarian cancer.  Small volumes of ovarian cancer may not release sufficient antigen to be detected, but could evoke an antibody response. Proteins that are mutated or overexpressed due to gene amplification may provide particularly useful targets. Autoantibodies have been detected in sera from 90% of ovarian cancer patients. Using microarrays of human proteins expressed in human cells, we will test the hypothesis that elevated autoantibody titers against wild type and mutated proteins associated with ovarian cancers will improve detection of ovarian cancers that do not elevate antigen levels in serum.

Project 2: Targeting Dll4-Notch Signaling in Ovarian Cancer

Anil K. Sood, M.D.
Co-Project Leader

Robert L. Coleman, M.D.
Co-Project Leader


The progressive growth of primary ovarian cancer and metastasis is dependent on development of an adequate blood supply (angiogenesis). Vascular endothelial growth factor (VEGF) plays a critical role in angiogenesis and consequent ovarian cancer growth and progression. VEGF blockade has shown promise in human studies. Our pre-clinical and clinical results from the prior funding period demonstrate that a novel approach (high-affinity VEGF decoy receptor, VEGF-Trap or aflibercept) for VEGF blockade was highly effective in combination with taxane chemotherapy. However, despite initial responses, most patients eventually develop tumor progression resulting in their demise, mainly due to the development of drug resistance. The Dll4 (delta-like ligand 4)/Notch signaling pathway has recently been shown to play an important role in angiogenesis including vessel maturation, pericyte recruitment, branching and cell differentiation, proliferation, survival and apoptosis. Our preliminary data indicate that when Dll4 inhibition using a monoclonal antibody (Dll4-mAb or REGN421) was coupled with VEGF inhibition (aflibercept), this combination strikingly reduced tumor burden and ascites, suggesting that this anti-angiogenesis regimen holds promise as a novel therapeutic modality. However, the mechanisms of its potency are not fully understood. We hypothesize that the Dll4/Notch signaling pathway plays an important role in reducing the efficacy of anti-VEGF monotherapy and targeting both VEGF and Dll4/Notch signaling pathway will enhance anti-angiogenic therapy.

There are limited options for treating advanced/recurrent ovarian cancer. Anti-angiogenic therapies appear to be one of the most promising for treatment strategies for ovarian and other cancers; however, despite initial responses, most patients eventually develop progressive disease. Dll4/Notch signaling is increased in response to elevated VEGF levels and Dll4 expression is increased in the tumor vasculature. Dll4 blockade, especially in combination with anti-VEGF therapy, appears to be a highly promising approach.

Specific Aims

  • Aim 1. To determine the biological and functional significance of Dll4/Notch signaling in ovarian carcinoma.
    Hypotheses: 1) Dll4/Notch signaling downregulates VEGFR2 on endothelial cells via epigenetic mechanisms; 2) REGN421 blocks Notch activation in endothelial cells and deregulates angiogenesis, producing dysfunctional tumor vessels and hypoxia.
  • Aim 2. To assess the biological effects of dual Dll4 and VEGF targeting in well-characterized orthotopic models of ovarian carcinoma.
    Hypotheses: 1) Dll4 blockade with REGN421 will enhance the in vivo efficacy of anti-VEGF therapy; 2) HIF1α inhibition in combination with VEGF and Dll4 inhibition will result in greater anti-tumor efficacy.
  • Aim 3. To characterize the toxicity and biologically active dose of anti-Dll4 therapy (REGN421) in ovarian cancer patients.
    Hypotheses: 1) REGN421 can be safely administered at biologically relevant doses; 2) REGN421 therapy will be characterized by unique pharmacokinetic parameters and will alter vascular function as measured by non-invasive imaging, and confirmed by tumor biopsy in the expansion phase. 

Project 3 (2010-2015): Personalized Therapy for Low-Grade Serous Carcinoma of the Ovary

David M. Gershenson, M.D.
Co-Project Leader (Clinical)

Kwong-Kwok Wong, Ph.D.
Co-Project Leader (Basic)

Samuel C. Mok, Ph.D.
Co-Project Leader (Basic)


Advanced stage low-grade ovarian serous cancer (OSC) represents approximately 10% of all advanced stage ovarian serous carcinomas. While the 5-year survival rates of patients with the disease is significantly higher than those with high-grade OSC, most patients with low-grade OSC eventually succumb to their disease. Recent genetic and genomic studies have shown that low-grade OSC and high-grade OSC have different pathogenetic pathways. Furthermore, recent clinical evidence has indicated that low-grade OSC is a unique disease and is less sensitive to standard chemotherapy. In spite of differences in histology and clinical outcomes, patients with low-grade or high-grade OSC are currently treated with the treatments, which are not that effective in low-grade OSC. Thus, new therapeutic strategies and novel molecular targets are desperately needed to improve the outcome of this patient cohort. Recent studies have demonstrated that the MAP kinase pathway is activated predominately in low-grade OSC because of activating KRAS/BRAF mutations or other unknown mechanisms suggesting that molecules targeting the MAP kinase pathway are promising therapeutics. In collaboration with the Gynecologic Oncology Group, we recently initiated a phase II clinical trial (GOG 0239) targeting the activated MAP kinase pathway in recurrent low-grade OSC using the MEK inhibitor, AZD6244, as a single agent. Preliminary analysis showed promising results with a number of responders in an initial data analysis. In addition, using microarray analysis on microdissected tumor samples, we identified activation of the IGF1-AKT pathway predominately in low-grade OSC. 

Furthermore, we recently showed that FOXO3a is a common target of PI3K/AKT, MEK/ERK, IKK signaling, and its localization and expression may be a predictor for AZD6244 sensitivity in low-grade ovarian cancer cells. Based on the results from these studies, we hypothesize that multiple signaling pathways are activated in low-grade OSC, and molecular markers identified through comprehensive genomic and proteomic analyses of responders and non-responders to pathway-targeted inhibitors can be used to predict drug responses and stratify patients for future clinical trials of pathway-targeted regimens. To test these hypotheses, we propose (1) to identify genomic and proteomic predictors of anti-tumor efficacy of the MEK inhibitor AZD6244 using GOG 0239 specimens; (2) to investigate the functional role of FOXO3a in conferring resistance to inhibitors of PI3K/AKT and MEK/ERK kinases in low-grade ovarian cancer cells; (3) to investigate the IGF1-PI3K pathway as a potential therapeutic target for low-grade OSC; and (4) to develop clinical trials involving novel combined targeted agent approaches. We believe that our studies will help us to develop individualized treatment regimens that will have lower toxicity and higher efficacy, for women with low-grade OSC.

Specific Aims

  • Aim 1: To characterize the toxicity and biologically active dose of anti-Dll4 therapy (REGN421) in ovarian cancer patients.
    Hypotheses: 1) REGN421 can be safely administered at biologically relevant doses, (2) REGN421 therapy will be characterized by unique pharmacokinetic parameters and will alter vascular function as measured by non-invasive imaging, and confirmed by tumor biopsy in the expansion phase.
  • Aim 2: To evaluate IGF1 as a therapeutic target for low-grade OSC. We will test the hypothesis that IGF1 stimulates proliferation of low-grade serous ovarian cancer cell lines by activating the PI3K/AKT and/or MAP kinase pathways through interaction with the IGF1R receptor. Further, we will test the hypothesis that 1) a combination of a MEK inhibitor (AZD6244) and an AKT inhibitor (MK2206) will exert greater inhibition of xenograft growth than either alone and 2) addition of an IGF1R inhibitor will provide additional activity in cell culture and against xenografts of low-grade ovarian cancer.
  • Aim 3: To develop novel clinical trials that target relevant pathways in low-grade OSC. We will test the hypothesis that enhanced response to dual inhibition will be observed in ovarian cancer patients whose tumors exhibit activation of the Ras-MAP and the PI3K/AKT signaling pathways. We will first initiate a clinical trial using an AKT inhibitor, MK-2206, as a single agent to investigate its efficacy in recurrent low-grade OSC. We will then initiate a 2nd clinical trial to determine whether a MEK inhibitor (AZD6244) and an AKT inhibitor (MK-2206) exhibit additive or synergistic anti-tumor activity for recurrent low-grade patients. Based on these results and the pre-clinical studies from aims 1, 2 and aim 3, we will explore other trial design strategies for the combination study as well as a possible trial with an IGFR1 inhibitor.

Project 4: Personalized Therapy for High-Grade Ovarian Cancer: Targeting PI3Kness and BRCAness

Gordon Mills, M.D., Ph.D.
Project Leader

Robert Coleman, M.D.
Co-Project Leader


The aim of this proposal is to identify and validate biomarkers that will, for the first time, enable individualization of therapy in ovarian cancer. This proposal will thus include the execution of an innovative phase II clinical trial that will facilitate the validation of novel biomarkers that predict the clinical efficacy of targeted therapies in individual women with ovarian cancer. We will target two biologic processes that we and others have established as playing critical roles in the pathogenesis of epithelial ovarian cancer: (i) activation of the phosphatidylinositide-3-kinase (PI3K/AKT/mTOR) pathway ('PI3Kness'), and (ii) deficient BRCA1/2-mediated homologous recombination (HR) ('BRCAness'). This proposal will build on the successful phase I trial targeting the PI3K signaling pathway in ovarian cancer thay we executed in the previous SPORE funding period. It will also build on our new data indicating that somatic mutations and loss of BRCA1 and BRCA2 function are significantly more common than previously thought in ovarian cancer and should predict sensitivity to poly (ADP-ribose) polymerase (PARP) inhibitors (PARPi) that exhibit synthetic lethality with BRCA1/2 dysfunction. This proposal will bring together: 1. the SouthWest Oncology Group (SWOG) to facilitate execution of the phase II trial, 2. Astra Zeneca to provide access to the novel therapies olaparib (PARPi) and AZD8055 (PI3K pathway inhibitor), and 3. Myriad Genetics, Inc.

The successful execution of this study will contribute to: 1) the implementation of novel therapies with clinical utility, and 2) progress towards individualization of treatment for women with ovarian cancer. As a result, we expect that the successful execution of this research proposal will lead to improved therapy and outcomes for women with ovarian cancer as well as define paradigms and trial designs that can be applied broadly to increase the rate of successful implementation of targeted therapy in ovarian cancer.

Specific Aims

  • Aim 1:
    A. To determine whether ‘PI3Kness’ predicts responsiveness to PI3K pathway inhibitors in cell lines and ovarian cancer xenografts.
    B. To determine whether ‘PI3Kness’ predicts outcome in ovarian cancer patients treated with surgery and platinum/paclitaxel-based chemotherapy.
  • Aim 2:
    A. To determine whether ‘BRCAness’ predicts responsiveness to PARP inhibitors in cell lines and ovarian cancer xenografts.
    B. To determine whether “BRCAness” predicts outcome in ovarian cancer patients treated with surgery and platinum/paclitaxel-based chemotherapy.
  • Aim 3: To determine whether ‘PI3Kness’ and ‘BRCAness’ predict response to targeting the PI3K/AKT/mTOR pathway and PARP, respectively, in a phase II clinical trial.

Project 5 (2010-2015): Gene-modified MSCs for Control of Ovarian Cancer

Michael Andreeff, M.D., Ph.D.
Co-Project Leader

Frank Marini, Ph.D.
Co-Project Leader

Muzaffar Qazilbash, M.D.
Co-Project Leader


The interaction between ovarian tumors and stroma is critical for survival: Tumors consist of a mixture of neoplastic and nonneoplastic cells in variable proportions and tumor growth is maintained by a dynamic interplay between neoplastic and host cells [1,2,3]. A small number of ovarian cancers originate from germ cells associated with the ovarian follicle or from the ovarian stroma, but the majority (>95%) originate from epithelial cells that cover the ovarian surface (OSE) or that line inclusions cysts. Most tumors develop from inclusion cysts within the stroma [4,5] where mesenchymal-epithelial cell interactions between stromal cells and OSE are thought to be particularly important for the onset of the cancer [6,7]. Both OSE and stroma contribute to the extracellular matrix (ECM) that separates the two cell types. Typically, an ovarian tumor creates a local solid tumor intra-abdominally, to which a clear border between tumor cells and host stroma can be observed [7,8]. These stromal tissues provide structural support for the malignant cells, influence vasculogenesis, and modulate the phenotypic behavior and aggressiveness of the cancer. In response, the tumor provides growth factors, cytokines, and cellular signals that continually initiate new stromal reactions and recruit new cells into the microenvironment to further support tumor growth [9,9a]. It is not fully understood how stroma influences the neoplastic cells, but there is evidence for involvement of soluble paracrine factors, extracellular matrix formation, and direct cell-to-cell interaction [10]. Consequently, stromal-epithelial interactions are critical in the growth of ovarian tumors.

Specific Aims

  • Aim 1: To elucidate the role of MSC in ovarian tumor stroma formation and development. 
    1.1 Identify and phenotypically characterize MSC populations from humans and mice. 
    1.2 Investigate biodistribution of IP and IV administered MSC in tumor bearing xenograft/syngeneic animal models. 
    1.3. To identify the optimal dose of MSC which provides maximal engraftment in tumors. 
    1.4 Identify factors critical for the recruitment and preferential engraftment of MSC in ovarian tumors, and to determine the contribution of CD44 receptor in mediating interactions between MSC and ovarian cancer cells. We will test the hypothesis that CD44 is a key receptor in mediated MSC migration and interaction in the tumor microenvironment.
  • Aim 2: To test the hypothesis that gene-modified MSC-IFNB engraft in ovarian tumors and exert an antitumor effect. 
    2.1. Determine the efficacy of MSC-INFB by the IP and IV route.
    2.2 Evaluate the effects of chemotherapy treatment on homing/recruitment of MSC to ovarian tumors with the goal of optimizing MSC-chemotherapy combinations, and to optimize the anti-tumor effects of gene-modified MSC by using optimized chemo scheduling + optimized number of MSC (aims 1.2, 2.1, 2.2) on established xenografts and syngeneic ovarian tumor models. We will test the hypothesis that chemotherapeutic treatment of ovarian tumors results in increased tumor-associated inflammatory mediators (which we have demonstrated to be MSC attractants), resulting in increased MSC engraftment and subsequent potentiation of tumor-growth control exerted by MSC-IFNB.
  • Aim 3: To prepare and generate the requisite preclinical data to file an IND for the use of autologous MSC, engineered to produce interferon-beta (IFNB in patients with ovarian cancer), for the treatment of refractory ovarian carcinoma. 
    3.1 Conduct toxicity studies with MSC-IFNB in preclinical models to determine side effects of locally produced IFNB and determine the biologically effective dose of MSC-IFNB. 
    3.2 Upscale and optimize the production of human autologous MSC in collaboration with the Department of Stem Cell Transplantation GMP-Cell Production facility and to develop SOPs for the clinical use of gene modified MSC. 
    3.3 Conduct a phase I trial in patients with ovarian cancer with gene-modified MSC-IFNB.


Core A: AdministrativeCore

David Gershenson, M.D.

Robert C. Bast, Jr., M.D.


The Administrative Core consists of two co-directors (Drs. Robert Bast and David Gershenson), an administrator (Dr. Charlotte Clarke), a financial analyst (Ms. Audrey Jones) and a grants educator/facilitator and advocate coordinator (Ms. Nancy Hubener). The Core provides leadership and coordinates the activities of the Executive Committee that includes all Program co-leaders and Core co-directors, the Internal Advisory Committee, the External Advisory Committee, and the SPORE Advocates. It interfaces with the Multidisciplinary Gynecologic Oncology Program Steering Committee, the Blanton-Davis Ovarian Cancer Research Program, the Clinical Trials and Data Management Group, the Gynecologic Oncology Tumor Bank and the CCSG. During the current grant period, Dr. Bast and the Administrative Core has taken the lead in bringing together the 12 funded SPOREs and their advocates at MDACC. In the next cycle we will enhance support for junior faculty in Career Development Program and investigators in the Developmental Research Program and tighten financial management with the aid of new personnel. 


  • To oversee all SPORE activities, including projects and core resources
  • To provide administrative support for the developmental research and career development programs
  • To convene all meetings of the SPORE, including the Administrative Core, the Executive Committee and scientific meetings
  • To convene all meetings of the Internal, External and Advocate Advisory Committees
  • To coordinate data quality-control and quality-assurance issues in conjunction with the Biostatistics Core
  • To monitor and oversee all fiscal and budgetary issues
  • To interface closely with the other oversight committees related to ovarian cancer research at our institution, including the Gynecological Oncology Tumor Bank Oversight Committee, the Executive Committee of the Blanton-Davis Ovarian Cancer Research Program and the Multidisciplinary Program Steering Committee
  • To coordinate research with other ovarian SPOREs and other SPOREs, by distributing materials, electronic communications and progress reports
  • To communicate with the NCI project officer and other staff to prepare all required reports and publications. The NCI project officer will be promptly notified of important developments that affect the management of the SPORE either positively or negatively
  • To assure compliance with all general, governmental and NCI regulations and requirements
  • To establish and implement policies for recruitment for women and minorities.

Core B: Biostatistics, Bioinformatics and Systems Biology (BBSB) Core

Keith Baggerly, Ph.D.


The goal of the Biostatistics Core is to provide statistical support for the projects, assist in the design of new studies and projects and work with the InterSPORE Bioinformatics Committee on the definition of common data elements related to the SPOREs and on the construction of a common Ovarian SPORE tissue data base.


  • To provide the statistics and data analysis required by the projects and cores to achieve their specific aims. The core will provide expertise and computational facilities to perform the statistical analyses required by each project.
  • To assist in the design and implementation of new trials and studies arising from the ongoing research of the SPORE. As progress is made in each project, new experiments and trials will be developed. The Biostatistics Core will lead the design of such proposed trials. If ongoing studies need to be modified, the core will participate in the redesign to insure that the statistical properties of the trial are maintained.
  • To provide guidance to the projects and cores in data management issues such as data entry and retrieval, quality assurance, security and data backups. All the investigators involved in the SPORE have experience with the software required to enter, maintain and retrieve the data produced by their projects. As necessary, the core will provide assistance in the data management of each project. The core will provide expertise in the selection of software and in data design, audit and backup.
  • To facilitate the exchange of information and data between the components of the SPORE by providing assistance in networking, email, data translation and electronic file exchange. Success of the SPORE depends upon a close cooperation among the projects and cores. The Biostatistics Core will facilitate the exchange of data and information between projects. The core will assist in electronic transfer and in the selection of data items and format to ensure data compatibility.

Core C: Pathology Core

Russell Broaddus, M.D., Ph.D.

Jinsong Liu, M.D., Ph.D.


The individual research projects that make up this Ovarian Cancer SPORE application require the procurement, processing and analysis of histopathological material from patients with ovarian cancer and benign ovarian diseases. The research projects have needs for frozen and formalin-fixed, paraffin-embedded samples of tumor and normal tissue. The Pathology Core augments the already established MD Anderson Cancer Center Gynecological Tumor Bank and the P30-sponsored MD Anderson Cancer Center Centralized Tissue Repository with supporting database and intranet access. The core provides tissue acquisition by experienced gynecological pathologists to assure high-quality tissues for the investigators participating in this SPORE as well as investigators of other SPOREs. The goal of the Pathology Core is to provide frozen tissue, paraffin-embedded tissue and histopathological expertise related to the specific needs for the research projects in this SPORE.


  • To maintain a frozen and paraffin-embedded tissue repository of ovarian cancer, benign ovarian processes and normal ovary. The primary tissue source is operative and biopsy specimens submitted to the Department of Pathology at MD Anderson Cancer Center.
  • To provide pathological review for all clinical specimens utilized in the SPORE projects and to provide histopathological technical services as necessary. Such technical services include immunohistochemistry, in situ hybridization, creation of specific tissue microarray slides and microdissection of tissue sections.
  • To establish a blood/urine/ascites fluid repository from patients undergoing surgery for ovarian cancer and benign ovarian processes. These fluids provide the resources for the systemic testing of putative prognostic and diagnostic markers derived from tissue-based expression array and CGH experiments.
  • To create and maintain a SPORE database for all samples collected at MD Anderson Cancer Center. This SPORE database is a virtual tissue repository that is electronically shared by all SPORE investigators.