Projects

Principle investigators: Elena Ioana Braicu, Jalid Sehouli, Berlin

Epithelial ovarian cancer is the main cause of death due to gynaecological malignancies. The poor prognosis is mostly due to late diagnosis in advanced stages, FIGO III and IV, when the therapeutic options are limited. This is mainly caused by lack of screening and early detection biomarker and lack of specific symptoms.

The CA125, the so called gold standard biomarker for monitoring epithelial ovarian cancer patients, has a poor diagnosis performance, being elevated in only 50% of early stages. The new biomarker, the human epididymal protein 4 (HE4), showed increased sensitivity and specificity for the diagnosis of ovarian cancer patients, especially in premenopausal women.

The imaging, especially transvaginal ultrasound has been successfully used for the ovarian cancer diagnosis by ultrasound experts. Nevertheless they are limitation of this method due to its subjective character and required expertise.

In the BERLINER study we aim to develop a new algorithm based on biomarkers (HE4, CA125 and new biomarkers) and systematic ultrasound (according to IOTA criteria) for the early detection of epithelial ovarian cancer in pelvic mass patients. Therefore serum, plasma and urine have been prospectively collected from all patients receiving surgery for pelvic mass within the Charité Vivantes network. Together with the bio sampling for all patients anamnesis, familiar risk, clinical symptoms, ultrasound appearance, tumor pattern and histology data have been prospectively documented. This project has currently recruited more than 1200 pelvic patients.

Secondary aims are: Diagnostic role of glycosilation profile, autoantibodies against VEGF and EGF;

Identification of differences in proteomics profile in borderline tumors of the ovary, different histological subtypes of epithelial ovarian cancer and benign diseases using imaging mass spectrometry.

Prediction of surgical outcome and platinum response in ovarian cancer patients.

Principle investigators: Paul Speiser, Robert Zeillinger

Currently, state of the art differential diagnosis between malignant and benign ovarian pathologies suspicious for epithelial ovarian cancer (EOC) relies predominantly on transvaginal ultrasonography and serum cancer antigen (CA-125) measurements. The specificity of these diagnostic tools however is low, and both tests are not effective enough to reliably differentiate between benign and malignant conditions.
A promising approach for improvement of differential diagnosis of suspicious ovarian changes has been established by Paul Speiser and Robert Zeillinger (Molecular Oncology Group, Department of General Gynaecology and Gynaecologic Oncology, Medical University of Vienna, Austria). This approach includes a lavage of the uterine cavity and proximal fallopian tubes and subsequent analysis of this lavage fluid for the presence of pre-malignant and malignant cells.

For the proof of principle that tumor cells are shed from ovarian cancer and can be found in the lavages of the uterine cavity, uterine lavages were collected before a surgical intervention for suspected ovarian malignancy at our institution and at the Catholic University Leuven, Department of Obstetrics and Gynaecology. After malignancy was confirmed, genetic changes in the TP53 and KRAS gene were determined in tumor tissue. In a set of 9 epithelial ovarian cancer patients (EOC) and 1 ovarian metastases of a signet ring carcinoma, the presence of these genetic changes was examined in lavage samples, using digital droplet PCR (ddPCR). Ten genetic changes were identified in tumor tissue of these patients and 10/10 (90%) of these changes were detected in the corresponding lavage specimen too.

In a next step Bert Vogelstein (Johns Hopkins University, Baltimore, USA) studied 26 lavage samples from EOCs and the corresponding tumor tissue for the presence of somatic mutations in a panel of genes using massively parallel sequencing. Mutations were identified in 16/26 (62%) lavage samples.

These results are proof that ovarian cancer cells are shed into the fallopian tubes and uterine cavity, and can be collected through our lavage technique. The fact that ovarian cancer cells were detected with high sensitivity in the lavage of the uterine cavity and proximal tubes shows that this approach has a great potential in differential diagnosis of ovarian masses.

In the current project we will study the specificity and sensitivity of the lavage of uterine cavity and proximal tubes as a test to differentiate between malign and benign ovarian tumors. We aim to detect cells from EOCs or genetic material from those cells in the lavage. On the basis of statistical calculations patients will be entered in the study until 108 malignant tumors are diagnosed. With the arithmetic assumption that only about 20% of all patients receiving surgery for suspected ovarian cancer will finally have the malignant disease, we plan to enter 540 patients into this study. Until March 2015 6 centers (see list below) have already entered 114 cases.

MUW, Universitätsklinik f. Frauenheilkunde u. Geburtshilfe , Elisabeth Maritschnegg, Nina Pecha, Christoph Grimm, Robert Zeillinger, Paul Speiser – Vienna, Austria

Klinikum Essen Mitte, Department Gynecology, Gynecologic Oncology, Florian Heitz, Philipp Harter, Andreas du Bois – Essen, Germany

Charité University - Campus Virchow Clinic, Department of Gynecology, Vlad Vacaru, Ioana Braicu, Jalid Sehouli – Berlin, Germany

Charles University, Department of Obst.&Gyn., Jiri Bouda, Alena Bartakova – Pilsen, Czech Republic

University Medical Center Hamburg-Eppendorf, Department of Gynecology, Fabian Trillsch, Sven Mahner – Hamburg, Germany

Catholic University Leuven, Division Gynaecological Oncology, Els Van Nieuwenhuysen, Ignace Vergote – Leuven, Belgium

Principle investigators: Paul Speiser and Robert Zeillinger

Pilot study of the Lavage of the Uterine cavity for the Diagnosis of Serous Tubal Intraepithelial Carcinoma

Over the last years, increasing scientific evidence could show that a large proportion of not only familial “high grade serous carcinomas” (HGSC) develop primarily in the lining of the fallopian tube. Recent findings highlighted the malignant potential of “serous tubal intraepithelial carcinomas” (STICs) as a precursor lesion to ovarian cancer.

A promising approach for the detection of these STICs has been established by Paul Speiser and Robert Zeillinger (Molecular Oncology Group, Department of General Gynaecology and Gynaecologic Oncology, Medical University of Vienna, Austria). It includes a lavage of the uterine cavity and proximal fallopian tubes and subsequent analysis of this lavage fluid for STICs. These precursor lesions are characterized by p53 overexpression on immunohistochemistry and high Ki-67 labelling index, indicating a high proliferation index. In over 90 percent, STICs carry mutations in the TP53 tumor suppressor gene.

The Molecular Oncology Group is currently collecting lavage samples and their corresponding STIC-positive tissue samples of women who opt for risk-reducing bilateral salpingo-oophorectomy because of an increased risk of HGSC (mostly carrying a BRCA mutation).

In case of the presence of STICs on histopathology, laser microdissection is performed on the STICs and DNA isolated for TP53 mutation analysis. DNA of cells obtained from lavage specimens are analysed for the presence of the specific TP53 mutation found in the STIC.

A positive result of this study could leed to a prospective screening study, screening women with high risk for HGSC.

Principle investigators: Jalid Sehouli, Khayal Gasimli, Elena Ioana Braicu, Berlin

Cytoreductive surgery and platinum based chemotherapy are two main pillars of the multimodal treatment of ovarian cancer. Around 75% of ovarian cancer patients will present peritoneal dissemination of the disease at time of first surgery. Peritoneal carcinomatosis leads to ascites, bowel obstruction and consequently to poor prognosis. Furthermore extended peritoneal spread of ovarian cancer in terms of milliar peritoneal carcinomatosis will lead to sub-optimal tumor debulking

Intraperitoneal spread of ovarian cancer is the most typical, earliest distribution pathway. To determine objectively and quantify the extent of peritoneal metastasis, there are currently two main groups of established assessment tools in surgical oncology. The prognostic role of Peritoneal Cancer Index (PCI) was reported in multiple studies, and it represents a decision making factor for palliative or curative treatment in gastrointestinal malignancies. Nevertheless there are inconsistent data about the role of PCI in ovarian cancer. Therefore we run a retrospective analysis of 80 primary ovarian cancer patients, optimally tumor debulked within our TOC project. The results of the study showed a median PCI score of 12.0 (range 3–32). We found statistical correlations between PCI and ascites (p=0.001), surgery duration (p<0.001), and preoperative CA 125 (p=0.025). PCI was also an independent predictive factor for progression free survival, but showed no significant association with overall survival in the multivariate analysis. These data have been published in Annals of Surgical Oncology.
The validation of these results is ongoing within a prospective study within the TOC biobank. During this study both the PCI and IMO score will be evaluated for their predictive role of surgical and clinical outcome in primary epithelial ovarian cancer.

1. Gasimli K, Braicu EI, Richter R, Chekerov R, Sehouli J. Prognostic and Predictive Value of the Peritoneal Cancer Index in Primary Advanced Epithelial Ovarian Cancer Patients After Complete Cytoreductive Surgery: Study of Tumor Bank Ovarian Cancer. Ann Surg Oncol. 2015 Feb 12.
2. Sehouli J, Könsgen D, et al. ,,IMO‘‘-Intraoperatives Mapping des Ovarialkarzinoms. Zentralbl Gynakol 2003; 125: 129-135.

Principle investigators: Elena Ioana Braicu, Silvia Darb-Esfahani

High-grade serous ovarian cancer (HGSOC) is a poor-prognosis disease, and primary or acquired chemotherapy resistance leading to tumor recurrence and death remains a major clinical problem. This type of epithelial ovarian cancer is mostly encountered and it is characterized by high genomic instability. The aim of our study is to molecularly analyze primary tumor tissue for genomic, epigenetic and post-translational characteristics of tumor suppressor genes p53 and BRCA1/2, and link it to tumor environment. Almost 100% of HGSOC present a mutant p53, nevertheless different isoforms have been identified, their role being unclear. Next-generation sequencing, epigenetic qPCR, and proximity ligation assay will be used to determine the status of BRCA1/2 and p53. The impact on most abundant gene expression and potential biomarkers will be determined by image mass spectrometry and validated by real time RT-PCR and immunohistochemistry. Clinical data, including co-morbidities will be retrieved and their prognostic and predictive value will be analyzed.

HGSOC patients will usually relapse within first 3 years after the end of platinum based chemotherapy. Little is known about the 15% of patients, who are free of relapse at 5 years. We will characterize this good prognosis group, and compare it with “classical” HGSOC and platinum resistant patients.

Therefore molecular differences within these groups, will provide a better understanding of the mechanisms of resistance and metastasis. The discovery of new predictive biomarkers might lead to a personalized therapy in this good prognosis group of advanced HGSOC. This is central for development of novel predictive, and therapeutic approaches for HGSOC patients to prevent recurrence and death.

Currently we identified around 100 patients within our TOC consortium, who did not relapsed within the first 5 years of follow-up after the end of first platinum-based chemotherapy. These patients will be matched according to age and residual tumor mass with platinum resistant and “classic” HGSOC patients, who relapsed within the first 3 years.

Integrated transcriptome and proteome analysis to identify therapeutic targets within the inflammatory cytokine network of high-grade serious ovarian cancer

Principle investigators: Hagen Kulbe, Elena Ioana Braicu, Berlin

We previously showed how key pathways in cancer-related inflammation and Notch signaling are part of an autocrine malignant cell network in ovarian cancer. This network, which we named the “TNF network”, has paracrine actions within the tumor microenvironment, influencing angiogenesis and the immune cell infiltrate.

By using an unbiased bioinformatics approach, we have identified kinases, particularly CK2, JAK2 and SRC, associated with the inflammatory cytokine network, which might play a central role in sustaining the cytokine expression levels and/or mediating its effects in ovarian cancer. The purpose of this project is to discover how best to target the TNF network, focusing on the validation of CK2 in combination with inhibiting the reciprocal feedback mechanism in ovarian cancer cells.

Such a holistic approach to determine novel treatments has not been evaluated previously in high-grade serous ovarian cancer (HGSC) and could have an impact on patient survival. As other malignant cell types produce TNF, IL-6, IL8 and VEGF, the inflammatory cytokine network may be similar in other cancers and therefore such therapeutic strategies identified by these analyses widely applicable.

A schematic signalling map of the TNF network including chemokine, cytokine and growth factor signalling in the HGSC cells. Marked in colours are key signalling nodes in the network.

Development of novel systemic therapies in human Ovarian Clear Cell Carcinoma

Principle investigators: Hagen Kulbe, Elena Ioana Braicu, Silvia Darb-Esfahani, Berlin

Ovarian clear cell carcinoma (OCCC) was defined over 40 years ago as a distinct histiotype according to the World Health Organisation of epithelial ovarian cancers (EOC) and is characterized by hobnailed glycogen-containing clear cells. OCCC is also classically associated with specific clinical features such as high incidence in early stages (stage I/II, 59% to 71%) (Kennedy et al, Gynecol Oncol, 1989) especially stage Ic, unilateral large pelvic masses and paraneoplastic phenomena such as hypercalcaemia (Anglesio et al, Clin Cancer Res, 2011) and thromboembolic complications (Stone et al NEJM, 2012 in press). It has also been commonly reported that endometriosis can transform sequentially into OCCC and EOC, hence atypical endometriosis represents a premalignant state (Kobayashi et al, Int J Gynecol Cancer, 2007). One of the main challenges in the treatment of OCCC relates to its inherent platinum resistance in comparison to serous counterparts; which may be a reflection of lower proliferative index found in OCCC (Itamochi et al, Oncology, 2002).

Recent molecular studies support the hypothesis that CCC represents a biologically distinct entity. Clear cell ovarian cancer cells are more cell autonomous and present biological features of cancer initiating cells (CIC). Hence, they rely on cross-talk and increased NOTCH, TGFB, WNT, VEGF and G protein coupled receptor signaling pathways like CXCR4. Especially in transit during single cell invasion, a process driven by TGFbeta1, might this cross-talk between these pathways essential for their survival and renewal. Therefore, to study the hierarchy of these pathways and events during cancer progression might open new possibilities to target clear cell ovarian cancer cells directly.

Example of immunohistochemical characterization of OCCC A) H&E and B) HNF1ß staining

Establishment and characterization of patient derived xenografts of ovarian cancer to develop novel therapeutic approaches

Principle investigators: Elena Ioana Braicu, Konrad Klinghammer, Hagen Kulbe, Berlin

Despite concerted research efforts over the past two decades, there has been little improvement in the prognosis of epithelial ovarian cancer (EOC). Understanding the complexity and underlying molecular mechanism are required to develop novel therapeutic opportunities in EOC.

Cell lines and hence established xenografts have only a limited value for preclinical drug evaluation and biomarker analysis due to clonal selection and cellular differentiation. To understand the mechanisms of response to chemotherapy and current treatment regimes we are establishing a large panel of patient derived xenografts from ovarian cancer displaying the heterogeneity of this disease. Together with the molecular genetic characterization, these patient derived xenografts can be used to elucidate the systemic biology of the tumor and can help to develop new rational hypotheses for further clinical trials.

Furthermore using these models we evaluate new treatment options, identify predictive biomarkers and test the sensitivity of experimental drugs and treatment approaches.

PDX models are developed by directly transferring surgically resected patient tumor fragments to immune-deficient mice. The tumor will be serially passaged and extensively characterized for drug evaluation study.

Ovarian cancer classification and discrimination taking into account of the heterogeneity of tumor morphology by imaging mass spectrometry

Principle investigators: Elena Ioana Braicu, Silvia Darb-Esfahani, Oliver Klein, Berlin

The heterogeneities tumor morphology could hinder a successful investigation of the underlying mechanisms of tumor behaviour. Therefore, the relevance of in situ analysis for assess and classify subtypes of tumors, shown increase relevance in cancer research. Since, due to the lack of knowing targets and the limitation of simultaneous investigation of several targets by e.g. immunohistochemistry and in situ hybridization, the classification of tumor heterogeneity and underlying mechanisms of relapse/recurrence and therapy resistance remain elusive. Therefore imaging mass spectrometry (Fig 1.) were used to differentiate between and classify different tumor regions through direct (in situ) analysis in tumor tissue sections Imaging mass spectrometry (IMS) enables the spatial investigation of protein profiles from an examined tissue section, thus making it possible to relate and visualize changes in tissue histology to changes in the proteomic signature of ovarian tissue biopsies (Fig. 2). IMS will be performed on FFPE-tumor tissue sections by combining commonly used antigen retrieval techniques coupled with tryptic digestion of proteins in situ¹⁾.Subsequently, “bottom-up” LC-MS/MS mass spectrometry will be used to identify lead proteins characterizing spatial patterns of pathophysiological change²⁾.

Figure 1. Imaging mass spectrometry workflow enables (A) the in situ simultaneous spatial acquisition of hundreds protein/peptides signatures in tumor biopsies. (B) The protein signatures (mass spectra), will be evaluated by statistical analysis, (C) which can visualize by digital staining. Figure 2. Characteristic peptide pattern from TMA tumor of ovarian clear cell carcinomas (OCCC, red label) and borderline ovarian tumors (BOT, green label) could be (A) discriminated by probabilistic latent semantic analysis. (B) Spatial peak distribution of peptide mz / 1717 Da ± 0.125%) show (C) significantly lower spatial intensities (AUC > 0.7; p < 0.05) in the BOTs in comparison to the OCC patient.

Establishment of long-term in vitro 3D organoid culture of fallopian tube epithelium as a model to study upper genital tract pathology

Principle investigators: Mirjana Kessler

The fallopian tubes are vital parts of the female reproductive system, since it is here that fertilization and the initial stages of blastocyst development take place, prior to implantation in the endometrium. In addition, fallopian tube epithelium has been implicated as the tissue of origin of high grade serous ovarian cancer. The mucosa of the tube consists of single-layered columnar epithelium composed of secretory and ciliated cells. Previously, we have shown in ex vivo tissue culture that the genital pathogen Chlamydia trachomatis causes profound disruption of the functional epithelium through activation of paracrine Wnt signaling (Kessler et al 2012). Within the scope of this project we have established long-term (>1 year) in vitro 3D organoid cultures of human fallopian tube epithelial cells based on maintenance of adult stem cells. Supported by paracrine signaling pathways (Wnt, Notch, BMP etc) the organoids exhibit a cellular organization and differentiation which closely resembles the epithelial architecture observed in vivo, as well continuous expansion capacity. Thus, this model offers a unique system to study pathogen-host cell interactions of the genital tract, as well as investigate the initial events that contribute to transformation of fallopian tube cells into ovarian cancer.

Confocal microscopy image of 3D organoid culture of fallopian tube epithelium. Characterization was performed by immunoflourescense staining of Ki67 and epithelial marker Cadherin-1 (Cdh1) - Max Planck Institute for infection Biology.

TOC projects of the Translational Gynecological Pathology Group (Institute of Pathology, Charité)

Principle investigators: Silvia Darb-Esfahani, Elena Ioana Braicu, Wolfgang Schmitt, Berlin

Pathology projects are being performed within the TOC Network since its foundation in 2005. Formalin-fixed, paraffin-embedded (FFPE) tumor tissue from patients included in TOC is collected pro- and retrospectively. Tissue microarrays (TMAs) are constructed from FFPE blocks and comprise samples from up to 50 patients within one TMA block (Figure 1). TMAs enable the high-throughput analysis of protein expression using immunohistochemistry. As ovarian carcinoma hsitological subtypes differ strongly in terms of molecular aberrations, therapy response, and prognosis, FFPE cohorts are sytematically separated into different histotypes. To date, TMA cohorts from TOC patients are available for primary high-grade serous carcinomas (n=400), clear cell carcinomas (n=60), low-grade serous carcinomas (n=60), endometrioid carcinomas (n=50). Within the OCTIPS FP7 EU-funded collaborative project 50 pairs of matched primary and recurrent high-grade serous ovarian carcinomas were assembled to a tumor cohort. In a collaborative effort of several international TOC partners a cohort of n=80 of the rare but highly interesting subgroup of high-grade serous ovarian carcinomas with a long recurrence-free survival could be established. Primary and recurrent adult granulosa cell tumors comprise another FFPE tumor cohort from TOC (n=75). During processing of new cases for the FFPE cohorts, apart from TMA production sections for RNA and DNA isolation are prepared. FFPE cohorts are continously extended.

Figure 1. Series of TMA blocks comprising tumor tissue cores (white) and control cores (liver tissue, brown)

A prognostic impact for tumor-infiltrating lymphocytes (TILs) in ovarian carcinoma biology has been shown repeatedly, however, the role of TILs in the different histological subtypes is unclear to date, similarly to the role of certain lymphocyte subpopulations. Primary high-grade serous carcinomas are evaluated conventional histopathology, and by IHC for patterns of immune cell infiltration. Lymphocyte subpopulations are typed by IHC, using e.g. CD3, CD4, CD8, but also markers for regulatory T cells (e.g. FOXP3, RORgammaT). The analyses are being performed in cooperation with the hematopathology group in the Institute of Pathology of the Charité. Preliminary data form 200 high-grade serous carcinomas show an independent prognostic value of a high density of certain T cell subpopulations, if located within the tumor cell nests. Current analyses focus on the interaction of regulatory T cells with each other and with tumor cells. Quantification of lymphocyte subpopulations is performed using the ROI manager tool (Figure 2A, VM Scope, Berlin). In the context of the evaluation of the immunlogical microenvironment immune ceckpoint molecules are of particular interest because they constitute targets of a variety of novel therapeutics, potentially also applicable to ovarian carcioma. PD-1 and PD-L1 are two key players in this regard and are being evaluated on the protein and RNA level in several FFPE cohorts to date. PD-1 and PD-L1 are not only expressed on T lymphocytes but also on tumor cells (Figure 2B).

Figure 2. Examples of immunohistochemical characterization on FFPE section of ovarian cancer biopsies A) Screen shot from the ROI manager software: tumor tissue has been encircled in blue, intratumoral and stromal CD8 positive TILs have been labelled with green or pink dots. B) High-grade serous ovarian carcinoma with membranous and cytoplasmic expression of PD-1.

A large number of oncogenic pathways and molecular alterations bear the potential to constitue either therapeutic targets of novel agents or to comprise biomarkers for prediction of therapy response and patient survival. Examples of current FFPE projects in this context are e.g. the evaluation of HER3 expression on the protein and RNA level and its relationship to patients outcome in different histological ovarian carcinoma subytpes. APOBEC enzymes have been shown recently to be causally related to mutational patterns in several cancers types, comprising ovarian carcinoma. FOXM1 is a central regulator of several oncogenic pathways in high-grade serous carcinoma, which has been shown in the TCGA project, however, its use as a predictive biomarker is unclear to date. BRCA1/2 mutations are found in a significat subset of high-grade serous carcinomas. Current projects focus on the evaluation of BRCA1/2 mutation rates and types within extreme prognostic groups of ovarian carcinoma (see long-term survivor project) and the elucidation of mutational intratumoral heterogeneity.

Principal investigators: Ignace Vergote, Els Van Nieuwenhuysen, Leuven

Non-epithelial ovarian cancer (NEOC), including sex cord-stromal tumors (SCST) and germ-cell tumors (GCT), are an uncommon subset of ovarian cancer (OC), together accounting for 10% of all OC. The aetiology of these tumors remains largely unresolved. It is well established that tumorigenesis is the result of multiple genetic alterations driving a normal cell toward a malignant state. Much effort has been made into researching the molecular mechanisms underlying epithelial ovarian cancer, but far less is known about the genetic changes in sex cord-stromal tumors and germ cell tumors.

Our aim is to identify genetic markers characterizing ovarian germ cell tumors and sex cord stromal tumors, leading to an improved insight and treatment. For this purpose we aim to:

1. determine germline single nucleotide polymorphisms (SNPs) on the germline DNA of included patients. These germline SNPs can be linked to susceptibility for the disease.
2. identify candidate cancer-related genes. First of all, we will use whole exome sequencing on an Illumina HiSeq2000 platform in a discovery group. Secondly, we will try try to validate these findings through a candidate approach in a larger validation group.
3. undertake validation of potential biomarkers and potential therapeutic targets in OGCTs and SCSTs.
4. establish a large multi-center collection of ovarian germ cell and sex cord stromal tumor tissue and blood samples to validate the genetic changes found in our discovery set. Such a collection is so far unique, since previous studies have been performed on very small populations. Multicenter collaboration is needed to achieve this goal. The TOC biobank is contributing to this aim.