Nous montrerons dans cet exposé quelques outils mathématiques qui permettent de quantifier l'evolution d'un cancer métastatique. Nous commencerons par présenter les résultats d'une étude clinique pilotée par un modèle mathématique développée par l'équipe du Pr Barbolosi. Nous montrerons ensuite comment contruire des indicateurs de l'état métastatique des patients.
There is great interest in understanding how the cancer stem cell population may be maintained in solid tumors. Here we show that tumor cells exhibiting stem-like properties and expression of stemness (CD133) and pluripotency markers (SOX2, NANOG, OCT4), can arise from original differentiated tumor cells, freshly isolated from human glioblastomas and which have never known any serum culture conditions. Upon activation of EGFR/ERK/EGR1 signaling, these cells transit from a more differentiated state unable to self-renew to a self-renewing and tumorigenic state, expressing NANOG and OCT4. Interestingly, expression of these pluripotency markers occurs before the cells re-entered the cell cycle, demonstrating that the cells have the capacity to change and reprogram without any cell divisions. We have further demonstrated that GBM dedifferentiation is strongly dependent on the repression of miR-199a-3p, which is expressed in differentiated GBM cells and efficiently prevents EGR1 expression. This tight EGR1/ miR-199a-3p interplay acts therefore as the “toggle switch” that orientates toward a stem-like or a differentiated state in function of the inverse proportion of these regulatory factors. Our findings constitute one more milestone in the characterization of GBM cell properties, helping to better understand how tumor cell fates and behaviors are controlled.
We would like to modelize the impact of the various forces that control the formation and maintenance of the luminal space of internal organs using an in vitro 3D-cell model. Many organs consist of a hollow cavity, named lumen, surrounded by a monolayer of epithelial cells that serves as a protective envelop and supports exchange with the outside environment. Thus, a key step of epithelial morphogenesis is the creation of the lumen. It is also of high interest in cancer biology since epithelial tumoral cells have lost the capacity to organize collectively around lumens rather forming disordered cell aggregates. Luminogenesis by hollowing proceeds through the fusion of vesicles at cell-cell contacts. The small nascent lumens grow through extension, coalescence and enlargement, coordinated with cell division and cell-cell junctions plasticity, to give rise to a single central lumen. The spherical form of the lumen depends on hydrostatic pressure, electrical repulsion of highly charged molecules and most importantly the balance of the tension forces at the cell periphery and on the luminal surface. The tension forces are imposed by the various inter-connected acto-myosin cytoskeletons linked to the cell-cell and cell-matrix junctions. In the recent years, we have demonstrated the role of our favorite protein in the formation of normal lumen. We also reported that its exogenous expression is capable of instructing cancer cells to form lumen again. Our data indicate that it contributes luminogenesis in part by controlling apical contractility. The formation of epithelial cysts with central lumens is easily recapitulated in vitro and can be followed by microscopy. It is also possible to alter tension forces using drugs or genetic approaches and to evaluate intracellular tension forces by imaging techniques. I will present the biological tools that we can provide to help modelize the contribution of the tension forces that regulate luminogenesis at different stages and control the final shape of the single central lumen.
The Hedgehog (Hh) signaling pathway controls cell differentiation and proliferation. It plays a crucial role during embryonic development and, in adulthood, in stem cell homeostasis and tissue regeneration. However, Hh signaling is also involved in cancer development, progression, and metastasis. Indeed, aberrant activation of Hh signaling has been identified in many aggressive cancers such as breast, lung, colorectal, ovarian, pancreatic cancers, melanoma or multiple myeloma, in particular in cells exhibiting resistance to chemotherapeutic agents such as cancer stem cells (CSC). In many of these cancers, the Hh morphogen, a 20 kDa peptide, is over-produced and, by interacting with its receptor Patched, activates the Hh pathway and consequently CSCs proliferation. Preventing Hh/Patched interaction is a promising strategy to inhibit cancer stem cells proliferation. Moreover, we have shown that Patched has a drug efflux activity and may contribute to the resistance of CSCs to chemotherapeutic agents making Patched a new target for anti-cancer therapy (Bidet et al. 2012, Fiorini et al. 2015, patent WO2012-080630, Hasanovic et al, in press).
Patched being a 12 transmembrane domains protein highly difficult to produce and crystallize making 3D structure determination a real challenge, we propose to use mathematics and informatics tools to perform a structural model of Patched allowing the study of the mechanisms of 1/ Hh binding to its receptor Patched in order to restore control upon Hh signaling and prevent cancer cells proliferation, 2/ Patched drug efflux in order to increase the effectiveness of chemotherapeutic treatments on CSCs and reduce the risk of metastasis and recurrence, and 3/ Patched proton transport activity which allows drug efflux.
During the last decade, intense efforts have been made to demonstrate that ion channels participate to cancer development. Three main questions remain to be answered at present: 1) What are the molecular links between ion channels and cancer cell phenotype? 2) How are these membrane proteins regulated in cancer tissue? 3) What is the link between functional heterogeneity and electrical plasticity in tumor tissue?
Recently, we demonstrated that SigmaR1, a stress activated chaperone protein, participates in the dialog between the extracellular matrix (ECM) and colorectal cancer (CRC) cells by triggering a signaling platform between hERG K+ channels and integrins. We found that this ion channel platform is associated with angiogenesis and metastasis formation in vivo (Crottès et al., Cancer Res., 2016). In another study, we discovered that KCNQ1, a K+ channel regulating water absorption in the gut, locks the [E-cadherin:beta-catenin] complex at junction adherens, participating to colorectal epithelium integrity. We also found that this channel is negatively regulated by the Wnt/beta-catenin pathway, a pathway chronically activated in colorectal cancers. This inhibition of KCNQ1 expression underlies epithelial to mesenchymal transition, an early step of metastasis formation. Interestingly, low expression of KCNQ1 in patients is associated with poor survival (Rapetti-Mauss et al., PNAS, 2017).
Altogether, these data suggest the existence of an electrical plasticity of cancer cells which can be modulated by the tumor microenvironment. This idea was further explored in pancreatic adenocarcinoma (PDAC), an epithelial cancer presenting a prominent stromal compartment (desmoplasia) which increases drug resistance and tumor cell aggressiveness. We found that stroma triggers EMT in cancer cells, a process which is associated to the regulation of Ca2+-dependent ion channels. We elaborated a model to understand how electrical plasticity participates to the functional heterogeneity of PDAC.
Our project proposes to explore, through bioinformatics approaches. the therapeutic potential of the ion channels involved in the progression of pancreatic ductal adenocarcinoma (PDAC), and most specifically in the crosstalk between stromal and cancer epithelial cell compartments.
The master regulators of electrical plasticity are expected to belong to the set of molecules exhibiting an altered expression between control PDAC cells and CAF-stimultated cells (criterion 1). However, this criterion alone is not sufficient because it will favor an independent selection of the most varying molecules without taking into account their synergy and their implication in the process under study. An effective selection method should take into account criterion 1, but also use the involvement of candidates in epithelial-mesenchymal plasticity (criterion 2) and their common implication in important biological pathways (criterion 3).
We envision to tackle this problem as a multi-objective feature selection problem by using evolutionary computation (EC). On the basis of properly defied evaluation functions (which measure the goodness of the selected features) for all the criteria, we propose a system combining EC and a classification algorithm to highlight both the set of priority targets to be studied and the relationships between targets.