Laurent Selek

Translation of the ecological trap concept to glioma therapy: the cancer cell trap concept

Viewing tumors as ecosystems offers the opportunity to consider how ecological concepts can be translated to novel therapeutic perspectives. The ecological trap concept emerged approximately half a century ago when it was observed that animals can prefer an environment of low quality for survival over other available environments of higher quality. The presence of such a trap can drive a local population to extinction. The cancer cell trap concept is the translation of the ecological trap into glioma therapy. It exploits and diverts the invasive potential of glioma cells by guiding their migration towards specific locations where a local therapy can be delivered efficiently. This illustrates how an ecological concept can change therapeutic obstacles into therapeutic tools.

Imaging and histological characterization of a human brain xenograft in pig: The first induced glioma model in a large animal

The prognosis of glioblastoma remains poor despite significant improvement in cytoreductive surgery, external irradiation and new approach of systemic treatment as antiangiogenic therapy. One of the issues is the low concentration in the infiltrated parenchyma of therapeutic agent administered intravenously mainly due to the blood-brain barrier. An intracerebral injection is advocated to overpass this barrier, this kind of administration need a low flow and continuous injection. The development of sophisticated implanted devices for convection-enhanced delivery is a mandatory step to have a controlled released of a therapeutic agent in glioblastoma treatment. Before testing such a device in a clinical trial a serious preclinical studies are required, in order to test it in realistic conditions we have develop the first induced high grade glioma model in a non-rodent animal: the pig. 21 pigs have been implanted in the parietal lobe with human glioblastoma cell lineage under a chemical immunosuppression by ciclosporine. A MRI follow up was then realized. 15 pigs have been implanted with U87MG, 14 have presented a macroscopic significant tumor, with radiological and anatomapathological characteristics of high grade glioma. 6 pigs were implanted with G6, stem-like cells tumors of glioblastoma, 1 pig develops a macroscopic tumor. This is the first reproducible glioma model in a large animal described, it open the way to preclinical studies to test implanted devices in anatomic realistic conditions, without the ethical issues of a primate use.

A Micro-Silicon Chip for in Vivo Cerebral Imprint in Monkey

Access to cerebral tissue is essential to better understand the molecular mechanisms associated with neurodegenerative diseases. In this study, we present, for the first time, a new tool designed to obtain molecular and cellular cerebral imprints in the striatum of anesthetized monkeys. The imprint is obtained during a spatially controlled interaction of a chemically modified micro-silicon chip with the brain tissue. Scanning electron and immunofluorescence microscopies showed homogeneous capture of cerebral tissue. Nano-liquid chromatography-tandem mass spectrometry (nano-LC-MS/MS) analysis of proteins harvested on the chip allowed the identification of 1158 different species of proteins. The gene expression profiles of mRNA extracted from the imprint tool showed great similarity to those obtained via the gold standard approach, which is based on post-mortem sections of the same nucleus. Functional analysis of the harvested molecules confirmed the spatially controlled capture of striatal proteins implicated in dopaminergic regulation. Finally, the behavioral monitoring and histological results establish the safety of obtaining repeated cerebral imprints in striatal regions. These results demonstrate the ability of our imprint tool to explore the molecular content of deep brain regions in vivo. They open the way to the molecular exploration of brain in animal models of neurological diseases and will provide complementary information to current data mainly restricted to post-mortem samples.

Existence of tumor-derived endothelial cells suggests an additional role for endothelial-to-mesenchymal transition in tumor progression.

Dear Editor, The idea that cancer cells can acquire a new phenotype and participate in the formation of blood vessels has been proposed at least half a century ago by Willis. In recent years, this view received some support from findings demonstrating presence of cancer cells in blood vessels. Now, two recent reports published in the International Journal of Cancer provide evidence that cancer cells can acquire some endothelial features when incorporated in tumor blood vessels. This capability of cancer cells to acquire an endothelial-like phenotype and to form mosaic vessel walls is of tremendous interest with regard to tumor cell progression as endothelial cells are capable to acquire mesenchymal traits through a process named endothelial-to-mesenchymal transition (EndMT). EndMT is a special kind of epithelial mesenchymal transition that occurs physiologically in the developing heart and occurs pathologically in some fibrotic diseases and cancer. In cancer, EndMT of nontransformed endothelial cells is a nonexclusive source of cancer-associated fibroblasts (CAFs). The possibility that cancer endothelial-like cells that have been incorporated into vessels also undergo EndMT and participate to tumor progression can now be considered. The mesenchymal drift of cancer cells is critical for tumor progression. Currently, this transition is reported to occur through a process named epithelial-to-mesenchymal transition (EMT), a key developmental program activated during tumorigenesis. The importance of cancer cell mesenchymal transition in tumor progression motivates the model proposed herein in which EndMT acts as an additional driver of tumor progression. In this model, interactions of endotheliallike cancer cells (EndCC) with endothelial cells, blood components and inflammatory signals induce the transition of EndCC toward a mesenchymal phenotype. This process is comparable to the EndMT that occurs during embryonic vascular development and cardiac fibrosis or to the EndMT observed in tumors when endothelial cells generate CAFs. This EndMT generates mesenchymal cancer cells (MCC) with migratory and invasive properties that protrude at the outer surface of tumor vessels and invade tumor microenvironment or alternatively form perivascular satellitosis, a hallmark of tumors such as glioma (Fig. 1). The release and dissemination of these EndCC and MCC in the tumor environment is further enhanced following vessel disruption either during tumor vascular remodeling or antiangiogenic therapies. Release of MCC in the lumen vessel could also occur and be an additional mechanism participating to metastasis. In this latter case, it is notable that cancer cells do not necessarily need to penetrate vessels for entering blood circulation as the EndCC is already integrated in the tumor neovasculature. The interest of this model is that it gives clues for some unresolved points or paradoxes. For example, (i) EndMT is one possible mechanism for the mesenchymal drift observed during progression of nonepithelial tumors such as glioma; (ii) EndMT generates a population of cancer stem cells expected to express cell markers different from those found in the primitive cancer stem cell population. This provides an issue to some conflicting results reporting the emergence of heterogenous populations of CSCs in solid tumor such as glioma; (iii) CD133, a brain cancer stem cell marker regulated by hypoxia, is also paradoxically expressed by cancer stem cells in the normoxic perivascular space. This paradox can be solved if we consider that endothelial precursor cells are known to express CD133. Endothelial microenvironment cues is a possible explanation for the presence of CD133 positive cancer stem cells both in normoxic perivascular areas and hypoxic regions and (iv) it has been demonstrated that EMT can generate cells with properties of stem cells such as self-renewal. Hence, cancer cells integrated into blood vessels and undergoing EndMT could reacquire a stem cell phenotype. Regeneration of the stem cell features by EndMT provides an explanation for data reporting that cancer stem cell phenotype can fluctuate. In addition, the model provides new insights into the correlation between tumor aggressiveness and neoangiogenesis. In this model, tumor neoangiogenesis is not only a mandatory step for the delivery of O2 and nutrients but also provides the microenvironment (cell–cell interactions between endothelial and cancer cells and blood components) involved in the mesenchymal transdifferentiation process. This model also has therapeutic consequences. In vascular endothelium, the EndMT is triggered by TGF-b and inhibited by BMP-7. This identifies these pathways as potential therapeutic targets in the control of the tumoral EndMT. On the other hand, antiangiogenic therapies, by disrupting tumor vasculature, are expected to release the blood vessels-interacting CSC and suppress blood vessels migratory tracks observed in glioma. This provides one explanation for the increased cancer cells dissemination in Le tt er to th e E di to r

A three-dimensional (3D) representation of pericardial cavity based on computed tomography (CT).

PurposeThe 3D modeling of human anatomy is more and more often used in medical education and in computer-augmented medicine. The lack of a 3D model of the pericardium has led us to its implementation.MethodsThe pericardium was reconstructed from a CT scan recording of a young, healthy subject. The anonymous CT scan data were blindly reviewed and interpreted by two independent radiologists. Stage one consisted in reconstructing the entire heart with the main afferent and efferent vessels. As the pericardial layers cannot be observed only with the CT scan, the second stage was to draw its reflection line following the most frequent model of pericardium defined in one of our prior studies. Afterwards, the epicardium had to be milled to finally create a pericardial sac area.ResultsFirstly, a model of one normal heart was reconstructed. Secondly, parietal and visceral layers of the pericardium have been achieved from the representation of their line of reflection. A short video shows recesses and sinuses and particularly, the transverse sinus crossed by a virtual object.ConclusionsThe resulting model is subject to certain limits, including reproducibility linked to the operator, individual anatomical variation, and scanner resolution but it represents a pericardial pouch true to its more common anatomical morphology. It offers a very precise educational tool. It must be considered as the first step of an automatic segmentation and reconstruction process to modelize normal and pathological pericardium. This is also the first step before a 3D dynamic model, synchronized with heartbeats.

Biodiversity as a barrier to glioma cell invasion

Gliomas are extremely aggressive and lethal forms of brain cancer. Unlike many other cancer types, glioma cells rarely metastasize. They spread throughout the brain and invasiveness of glioma cells is a major cause of therapeutic failure. In plant ecosystem, biodiversity acts locally as a barrier to ecological invasion. By analogy, we hypothesize that the low cell diversity of differentiated tissues, a counterpart of their functional specificity, opens the way to local cancer cell invasion. Seeding the brain tumor microenvironment with heterogeneous cell populations could be a mean to limit cancer cell invasion by enhancing cell biodiversity.

Operation and Chemotherapy: Prognostic Factors for Lung Cancer With One Synchronous Metastasis

BACKGROUND Stage IV non-small cell lung cancer (NSCLC) is considered incurable; however, some patients with only few metastases may benefit from treatment with a curative intent. We aimed to identify the prognostic factors for stage IV NSCLC with synchronous solitary M1. METHODS A database constructed from our weekly multidisciplinary thoracic oncology meetings was retrospectively screened from 1993 to 2012. Consecutive patients with NSCLC stages I to IV were included. RESULTS Of the 6,760 patients found, 4,832 patients were studied. Among the 1,592 patients (33%) with stage IV NSCLC, 109 (7%) had a synchronous solitary M1. Metastasis involved the brain in 64% of patients. Median overall survival was significantly longer in synchronous solitary M1 than in other stage IV (18.9 months, interquartile range [IQR]: 9.9 to 34.6 months versus 6.1 months, IQR: 2.3 to 13.7 months], respectively, p < 10-4). Among patients with synchronous solitary M1, 90 (83%) received a local treatment with curative intent at the primary and metastatic sites. Factors independently associated with survival were age older than 63 years (hazard ratio [HR] 1.63, 95% confidence interval [CI]: 1.01 to 2.63), Performance status of 3 or 4 (HR 7.91, 95% CI: 2.23 to 28.03), use of chemotherapy (HR 0.38, 95% CI: 0.23 to 0.64), and operation conducted at both sites (HR 0.35, 95% CI: 0.19 to 0.65). CONCLUSIONS Synchronous solitary M1 treated with chemotherapy and operation at both sites resulted in better survival. Survival of NSCLC with synchronous solitary M1 was more similar to stage III than other stage IV NSCLCs. The eighth TNM classification takes this into account by distinguishing between stages M1b and M1c.