Presence and role of stem cells in ovarian cancer.

Many studies report that ovarian carcinogenesis is associated with epithelial-mesenchymal transition (EMT)[ 19 , 27 ]. Based on our own experimental experience and data from the literature, we suggest that small, spherical VSEL stem cells present amongst epithelial cells in the ovarian surface epithelium induce EMT by growing into larger (diameter approximately 10-15 μm), NANOG and vimentin-positive spindle/elliptical shaped cells, and transform into mesenchymal-like stem cells (Figures 1 and 2 ); surrounding epithelial cells are not excluded from the EMT process and support it in unknown ways[ 27 ]. Different signals from tumor microenvironment might be involved in a network of interactions that activate so-called EMT programs[ 29 - 31 ]. Cancer cells are, beside epithelial cells, surrounded by a large variety of stromal cells, such as fibroblasts, myoblasts, lymphocytes and macrophages, and endothelial cells and pericytes recruited to tumor vasculature[ 32 ]. Paracrine and juxtacrine signals in such microenvironment include different members of the transforming growth factor superfamily, epidermal growth factor, fibroblast growth factor, hypoxia-inducible factor, Wnt, Notch, and many others[ 33 ]. Hypothetical model of epithelial-mesenchymal transition. Small, progenitor very small embryonic-like stem cells (diameters of about 5 μm) among the epithelial cells in the ovarian surface epithelium develop into bigger round cells (diameters of 10-15 μm) and (in)-directly undergo transformation into mesenchymal-like stem cells. EMT is a basic physiological cell reprogramming event active in tissue remodeling during embryogenesis and, later in life, during the regeneration of adult tissue in cases of injury. During EMT, cells acquire unique mesenchymal cells characteristics, like epithelial cell polarity, intracellular adhesion and loss of specific cell surface markers. Pursuant to cytoskeletal change, cells subsequently gain a mesenchymal-like phenotype. The acquisition of mesenchymal characteristics during EMT occurs progressively, where fully epithelial and mesenchymal phenotype represent just the extreme edges[ 34 ]. This plastic and dynamic process involves several intermediate states, including so-called hybrid phenotypes in which cells express epithelial and mesenchymal features[ 35 , 36 ]. Cells expressing such hybrid epithelial/mesenchymal phenotype do not only play a fundamental role in embryogenesis, but also in cancer formation and progression[ 37 , 38 ]. Epithelial trait loss and mesenchymal characteristic acquisition, such as vimentin and myosin presence, which occurs during EMT, enable invasive cellular motility which enhances tumor progression[ 39 ]. Elliptical fibroblast-like cells with stem cell characteristics and positive pluripotency and mesenchymally-derived cell markers seen in high grade ovarian serous carcinoma, invaded ovarian tissue by changing their round shape into mesenchymal–like phenotype involving elongations and protrusions[ 27 ]. By promoting high degrees of cell invasion, EMT increases tumor malignancy[ 36 , 40 ]. EMT is also considered a key step in CSC metastases[ 41 ]. Mesenchymal properties, acquired by carcinoma cells during EMT, promote invasion into the extracellular matrix and further dissemination. Transformed tumor cells spread throughout the organism, where they present a reservoir that expands and refills cancer cell populations[ 42 ]. Circulating tumor cells with a fully-mesenchymal state display lower metastatic potential compared to hybrid epithelial/mesenchymal cells undergoing partial EMT[ 43 ]. Interestingly, not all cancer cells are able to undergo EMT at the same time and even those which have activated EMT may not be sufficient for metastasis[ 44 ]. Furthermore, in cases of colorectal cancer, it has been shown that EMT occurs only in a subset of cells at the invasive front of primary carcinoma, usually associated with stromal components[ 45 ]. The presence of a major cytoskeletal component of mesenchymal cell vimentin was, further to metastatic progression, also related to primary (intrinsic) resistance or poor response to chemotherapy[ 46 , 47 ]. Tumor cells undergoing EMT gain the ability to disarm body antitumor defence, resist apoptosis and anticancer drugs[ 47 ]. Morphological and functional changes involved in EMT related processes (generation and maintenance of CSCs, tumor invasion, metastasis formation and chemoresistance) require robust reprogramming of gene expression, which is partially accomplished at the transcriptional level of gene expression and partially in gene expression’s post-transcriptional regulation[ 44 ]. A well established factor regulating EMT is the activation of the signal transducer and activator of the transcription 3 (STAT3) pathway[ 48 ]. As EMT represents the main event in terms of cancer manifestation and progression, it is becoming a promising target for therapeutic intervention[ 42 , 49 ]. In conclusion, EMT as well as its reversal, mesenchymal-epithelial transition, represents a highly dynamic process in tumor cells which may be triggered by VSEL stem cells present amongst epithelial cells in the ovarian surface epithelium layer. Cancer stemness seems to be more associated with partial-EMT phenotype than fully-driven EMT[ 34 , 44 , 50 ]. In line with this concept is the concept of CSC plasticity, which postulates that CSCs are able to switch between different states, including a non-stem state[ 51 ].

There are multiple pathways involved in promoting stem cell phenotype and chemoresistance in terms of ovarian cancer. Each pathway could potentially be therapeutically targeted with pathway inhibitors. If a therapeutic goal is to eliminate CSCs, more studies are needed to define which subpopulation of CSCs should to be targeted. Small, VSEL stem cells, proposed to be progenitor stem cells involved in ovarian cancer triggering and progression via EMT and interaction with other types of stem cells, might represent potential targets. Before introducing CSC target therapy in clinical practice, we should think about the many challenges concerning stem cell inhibition. CSCs share epitopes with normal stem cells, so treatment regimens targeting CSCs may also harm normal stem cells, which would increase drug toxicity and the risk of adverse side effects. An ideal therapeutic agent should selectively target CSCs above normal stem cells, so further biological CSC characterization, particularly the underlying mechanisms regulating their function, is required. The development of novel CSC treatments also requires a thorough understanding of the complex genomic profile of ovarian cancer, since its heterogeneity might influence treatment response[ 104 , 116 ]. Finally, only a small fraction of cells within the tumor would be eliminated by CSC target therapy, so treatment response, shrinking of the tumor bulk, may require some time to become visible. Therefore, a combined therapeutic approach with cytotoxic chemotherapy and/or other treatment regime should be introduced to reduce tumor bulk[ 117 ]. Treatment success should then be estimated in terms of tumor bulk shrinkage and the eradication of CSC populations for which identification of reliable ovarian CSC biomarkers is critical, so it can be used in clinical practice.

Successful ovarian cancer treatment depends on ovarian CSC eradication, as CSCs present the driving force of disease manifestation, progression and recurrence pursuant to conventional treatment. There are three potential targets for ovarian CSC eradication: CSC markers, CSC signaling pathways implicated in renewal and CSC niche. Due to the heterogeneous nature of ovarian cancer, there are probably more markers identifying different subpopulations of ovarian CSCs, and diverse signaling pathways involved in CSC renewal. In terms of identifying CSC specific markers and signaling pathways, and additionally exploring ovarian CSC microenvironment, cancer cell lines are helpful, but in vitro tumor formation analysis should be upgraded by analysing ovarian cancer patient tumor tissue in vivo . Selected ovarian CSC markers, signaling pathways and factors from CSC microenvironment should then be tested in clinical practice, where their expression, influence and inhibition should be correlated, not only with disease outcome, but also in terms of their influence on chemoresistance. In vitro and in vivo investigation of CSC properties and their microenvironment properties may lead to novel therapeutic regimens for ovarian cancer elimination and relapse prevention. Despite the tremendous progress made in recent years in the field of ovarian CSCs, a number of issues should be considered. Firstly, ovarian cancer presents a heterogeneous group of tumors and stem cells. A high level of heterogeneity makes identifying a target for a wide population of ovarian cancer patients difficult; to mitigate this, more phenotypic, genetic and epigenetic studies of ovarian cancer patient CSCs need to be performed. Secondly, we still don’t know if different CSC populations arise from common progenitor cells and which CSC populations are the most critical to target. We suggest a population of small, VSEL stem cells as progenitor stem cells needs be better elucidated. CSC subpopulations probably change during cancer progress, and specific CSC subpopulations might play a role at certain disease phases. Finally, do varying marker profiles signify different subpopulation of CSCs and how does CSCs microenvironment influence markers expression? These and many other questions in the field of CSCs should be answered in order to introducing personalised medicine in the treatment of ovarian cancer.

The authors would like to thank Shawn Nicholas Thomson for proofreading the article and Eva Skuk for technical preparation of figures.