S22
european journal of cancer 48, suppl. 5 (2012) S20–S24
90 New Targeted Agents in Development for Cancer Treatment J. Soria1 , C.P. Massard1 . 1 Institut Gustave Roussy, Medical Oncology/Lung Unit, Villejuif, France The treatment of certain cancers has been revolutionised by a greater knowledge in cancer biology and the isolation of molecular targets implicated in cancer biology. New treatments able to interfere with these different pathways have been developed with success such as imatinib in GIST, trastuzumab in HER2 amplified breast cancer or erlotinib/gefitinib in EGFR mutated lung cancer. During the last decade, oncologists are moving from a pathology-based therapy to a new molecular classification and targeted-oriented therapies, and significantly altered pathways were discovered in different tumor types. BRAF mutation is very common in malignant melanoma (MM), and vemurafenib (BRAF inhibitor) has shown a dramatic antitumor activity in patients with advanced V600E BRAF mutated MM. In castration resistant prostate cancer patients, Abiraterone acetate, an androgen biosynthesis inhibitor, was tested in patients with CRPC pre-treated with docetaxel in a phase III trial with demonstration of an overall survival benefit, confirming that CRPC remains hormone-driven, even in advanced stages of the disease. Recently presence of a translocation of the ALK gene have been described in a small portion of patient corresponding to 5% of adenocarcinomas, and this translocation is associated to a high rate of response to ALK receptor kinase inhibitors (crizotinib). Other therapies targeting FGFR1, MET, MEK, PI3K, etc, are in clinical testing. Moreover, novel immunotherapeutic strategies (anti-CTLA4 and vaccine) showed overall survival improvement in melanoma and prostate cancer. Recent developments in high-throughput technologies for gene sequencing, comparative genomic hybridization. have revealed many known molecular abnormities and described some new alterations with clinical interest. Several prospective trials are ongoing to treat patients with advanced cancer according to the existence of oncogenic driver events (mutations/translocations etc.), and to demonstrate that the development of personalized medicine with a focus on novel targeted therapies will challenge the one-size-fits-all approach to the treatment of cancers. 91 Proffered Paper: Testing Individualised Treatment Strategies in Preclinical Models of Pancreatic Cancer M. Pajic1 , D.K. Chang1 , K.S. Kassahn2 , J. Wu1 , M. Cowley1 , N. Waddell2 , A. Johns1 , Australian Pancreatic Cancer Genome Initiative3 , S.M. Grimmond2 , A.V. Biankin1 . 1 Garvan Institute of Medical Research, Cancer Research Program, Sydney NSW, Australia, 2 University of Queensland, Institute for Molecular Biosciences, St Lucia QLD, Australia, 3 www.garvan.org.au/apgi, NSW, Australia Background: Pancreatic cancer (PC) is the fourth leading cause of cancer death in Western societies with an overall 5-year survival rate of less than 5%. There are few therapeutic options for patients with PC, and new insights into the pathogenesis of this lethal disease are urgently needed. Methods: Recent advances in nucleic acid sequencing technology have made it feasible to rapidly, and exhaustively sequence an entire genome, including large scale efforts such as the International Cancer Genome Consortium (ICGC). Integrating companion biomarker discovery through next generation sequencing with therapeutic development at the pre-clinical stage creates the opportunity to identify candidate biomarkers early, significantly facilitating both therapeutic and biomarker development in PC. Results: Our cohort of primary xenografts generated directly from resected human PC is extensively characterised (genome, transcriptome, epigenome and phosphoproteome) as part of the ICGC, and histologically recapitulates the human disease, including a significant stromal component. Since these xenografts retain the genomic signature of primary PC, are generated from individual patients, and are renewable, multiple treatments can be simultaneously examined, providing the opportunity to test proposed personalised medicine strategies, and facilitate novel therapeutic development linked with companion biomarker discovery that are currently intractable in a clinical trial setting. In an interim integrative analysis of DNA copy number and whole genome aberrations in a 100 pancreatic tumors, we have identified several molecular phenotypes that have now been examined in the preclinical models and have the potential to inform decisions on optimal treatment strategies for individual tumours. Conclusion: We demonstrate how integrative analysis of multidimensional data encompassing genomic sequence, copy number variation, gene expression and methylation can be utilized to develop new therapeutic strategies that demonstrate efficacy in model systems. This approach may be used to define biologically and clinically relevant cancer phenotypes.
Tuesday 10 July 2012
92 Personalised Oncology − a Worldwide Challenge for Drug Development, Epidemiology, and Public Health T. Tursz1 . 1 Institut Gustave Roussy, Honorary Director General, Villejuif, France The accumulating proofs of efficacy of targeted drugs, mostly in highly selected cancer patient populations, and the dissemination of modern gene analysis technologies, like new generation high throughput sequencing, are clearly going to deeply modify the current practice of clinical trials in oncology, as well as the standards of care for drug allocation to cancer patients in daily practice, at least for advanced or relapsing diseases. No Cancer Institute, no Pharma, no national healthcare system can alone face all facets of such revolutionary changes in medical practice. New strategies, new partnerships and new alliances between Academy, Pharma Industry, private and public stakeholders, and regulation authorities are urgently needed. This presentation will emphasize a limited number of key and complex questions, which could hardly be addressed without worldwide initiatives and partnerships. 1. The whole current knowledge in cancer epidemiology is based upon purel morphological tumor classifications, obviously mixing and confusing various different diseases with distinct genetic patterns. Our strategies for the identification of cancer risk factors, for cancer prevention and early diagnosis should be revisited, especially in some quite common tumor types worldwide (lung, breast, melanoma). 2. New anti-cancer drug development is an enormously long, complex and costly process, leading to major inequalities among patients for access to innovation, depending mostly of geographic origins and socio-economic levels. This issue becomes even more complex when about 900 new targeted agents are under development, any of them expected to be of therapeutic interest in only a small still undetermined subset of patients. Furthermore, examples are now accumulating showing that the most promising of those drugs will have a long-term effect only when given in combinations, in attempts to target efficiently cellular key pathways at several levels (or to target several parallel pathways) in order to avoid subclonal selection of additional molecular events leading to tumor resistance (anti-B-RAF in melanomas, resistance to Herceptin in HER-2 positive breast cancers. Such data illustrate how important would be to initiate biology-driven multiple drug combinations early in drug development, overcoming the fact that relevant drugs are commonly developed by different currently competing firms. 3. The enormous financial burden linked to new anti-cancer drug development, in the global context of huge constraints on healthcare budgets worldwide represents the major hurdle for the future of personalised medicine. The need for health-economy studies, evaluating the more appropriate strategies to deliver promptly the right drug to the right patient, but also to avoid the allocation of the wrong treatment to the wrong patient is highly necessary. 4. New ethical and legal issues are going to appear, as soon as patients will realize that not to freeze their tumor materials or not to perform in due time proper molecular analysis clearly result in a loss of chances for better survival. The EUROCAN Platform should be a necessary forum for European cancer institutes, pharmas and patients to elaborate together the appropriate strategies to face these emerging challenges for modern oncology.
Tuesday 10 July 2012
10:15−12:00
Symposium
Metastases and EMT 93 MicroRNAs, EMT and Cancer Stem Cells ¨ T. Brabletz1 . 1 Universitatsklinikum Freiburg, Freiburg, Germany We have shown, that in particular tumor cells at the invasive front undergo an epithelial-mesenchymal transition (EMT) and aberrantly express EMTassociated transcriptional repressors, like ZEB1. The amount of such cells strongly correlates with metastasis formation and poor clinical outcome. Strikingly, metastases show again a differentiated phenotype, indicating a mesenchymal-epithelial re-transition (MET) and a support a regulatory role of the tumor environment for malignant tumor progression. We described that the EMT-activator ZEB1 is a crucial promoter of metastasis and demonstrated that ZEB1 inhibits expression of cell polarity factors and the microRNA-200 family, whose members are strong inducers of epithelial differentiation. These results indicate that ZEB1 triggers a microRNA-mediated feedback-loop, which stabilizes EMT and promotes dissemination of cancer cells. Moreover we detected that in addition ZEB1 is necessary for the tumor initiating capacity of pancreatic and colorectal cancer cells. ZEB1 inhibits expression of miR-200c, miR-203 and miR-183, which cooperate to suppress
Symposia
european journal of cancer 48, suppl. 5 (2012) S20–S24
expression of stem cell factors, as demonstrated for the polycomb repressor Bmi1. We propose that ZEB1 links EMT-activation and stemness-maintenance by suppressing stemness-inhibiting microRNAs and thereby is a promoter of mobile, migrating cancer stem cells. Notably, these cells also acquired a drugresistance phenotype. Thus, targeting the ZEB1–miR-200 feedback loop might be a promising treatment option for fatal tumors, such as pancreatic cancer. 94 Regulation of Epithelial Plasticity and Metastasis of Breast Carcinoma by Lysyl Oxidase-like 2 (LOXL2) 1
1 1
A. Cano , G. Moreno-Bueno . Universidad Autonoma de Madrid/ Instituto de Investigaciones Biomedicas "Alberto Sols" CSIC-UAM, Biochemistry/ Cancer Biology, Madrid, Spain Background: Lysyl oxidase-like 2 (LOXL2), a member of the lysyl oxidase (LOX) family, has been previously shown to interact with Snail1 promoting Snail1 stabilization and epithelial-mesenchymal transition (EMT) [1]. Our previous studies identified LOXL2 as a poor prognosis marker in larynx squamous cell carcinomas inducing malignant transformation by both Snail1-dependent and -independent mechanisms and revealed that LOXL2 overexpression is associated to poor prognosis of N0 breast carcinomas [2]. In the present study we have investigated further the implication of LOXL2 in breast carcinoma tumors. Material and Methods: A combination of transcriptomics and immunohistochemistry analyses for LOXL2 in a series of grade 3 IDC breast carcinomas (n = 195) have been carried out. Breast tumors samples were from the archives of the Pathology Departments of Hospital La Paz, Madrid, and MD Anderson Cancer Center Madrid. LOXL2 expression analysis were performed at mRNA and protein level in a collection of human breast carcinoma cell lines, representative of different breast tumor subtypes (luminal; ErbB2+; and triple negative: ER−/PR−/ErbB2−). Loss of function studies by shRNA have been have been carried out in basal carcinoma cells (MDA-MB231; BT549) and non-tumorigenic breast basal cells (HBL100) followed by phenotypic and molecular characterization and studies of tumorigenity and metastasis. Results: Expression profiling studies of breast carcinomas grade 3 IDC tumors (n = 58) identified a molecular signature for basal-like breast tumors (BBS: Basal-like Breast Signature), a subtype of highly aggressive triple negative breast tumors. Basal-like breast carcinomas have been previously described as prone to suffer epithelial-mesenchymal transition (EMT) [3,4]. Interestingly, LOXL2 was included in the BBS as one of the up-regulated genes and indeed LOXL2 expression was confirmed to be restricted to basal-like tumors by qRT-PCR. Immunohistochemical analyses in an extended series of grade 3 IDC (n = 195) showed that increased LOXL2 expression with an intense cytoplasmic/perinuclear pattern significantly correlated with the subgroup of metastatic basal-like breast tumors. In agreement with those findings, LOXL2 is expressed with the intense perinuclear pattern in breast carcinoma cell lines of basal-like B/mesenchymal phenotype and in non-tumorigenic basal cells, but it is completely absent in carcinoma cells of the luminal and ErbB2 subtypes. Silencing of LOXL2 in basal-like breast cells by shRNA induces a dramatic reversion of the phenotype leading to mesenchymalepithelial transition (MET), associated to reduced cell motility and invasion of basal-like cells. Furthermore, LOXL2 knockdown reduced the tumor growth potential and completely suppressed the lung metastatic colonization of highly aggressive MDA-MB-231 cells. Remarkably, the MET-like process induced by LOXL2 knockdown in basal-like cells is independent of E-cadherin and Snail1 modulation, but involves the upregulation and reorganization of tight junctions and cell polarity complexes at the apico-lateral membrane and downregulation of FAK activation. Interestingly, LOXL2 is able to repress tight junction (Claudin1) and cell polarity components (Lgl2) at transcriptional level by a mechanism independent of Snail1 and LOXL2 catalytic activity [5]. Conclusions: These results indicate that intracellular LOXL2 is required for metastatic dissemination of basal-like breast tumors. The new uncovered role of LOXL2 in the negative modulation of cell polarity and the epithelial phenotype may unveil new targets for anti-metastatic therapies of highly aggressive basal-like tumors. Reference(s) [1] Peinado, H. et al. EMBO J, 24: 3446−58 (2005). [2] Peinado, H., Moreno-Bueno, G. et al. Cancer Res, 68: 4541−50 (2008). [3] Sarrio, ´ D. et al. Cancer Res, 68: 989−97 (2008). [4] Taube, J.H. et al., Proc Natl Acad Sci USA, 107: 15449−54 (2010). [5] Moreno-Bueno, G. et al. EMBO Mol Med, 3: 528−44 (2011).
S23
95 Proffered Paper: EMT Inducers Catalyse Malignant Transformation of Mammary Epithelial Cells and Drive Tumorigenesis Towards Claudin-low Tumours in Transgenic Mice A.P. Morel1 , G.W. Hinkal1 , C. Thomas1 , F. Fauvet1 , S. Courtois-Cox1 , A. Wierinckx1 , M. Devouassoux-Shisheboran1 , D. Spicer2 , S. Ansieau1 , ´ A. Puisieux1 . 1 Centre de Recherche en Cancerologie de Lyon (CRCL), Tumor Escape, Lyon, France, 2 Main Medical Center Research Institute, Scarborough, USA The malignant transformation of human epithelial cells is generally described as a multistep process resulting from the accumulation of five to seven rate-limiting changes. Here, we challenge this dogma and demonstrate that this number is radically reduced when cells undergo an epithelialmesenchymal transition (EMT). EMT is an embryonic transdifferentiation process that consists of the conversion of polarized epithelial cells into motile mesenchymal ones. EMT-inducing transcription factors, including Twist and Zeb proteins, are aberrantly expressed in multiple tumor types and are known to favor the metastatic dissemination process. We demonstrate that, beyond their prometastatic potential, EMT inducers also act as potent drivers of tumorigenesis. We indeed show that the Twist1 EMT-inducing transcription factor promotes breast and skin cancer development in vivo in cooperation with the K-RasG12D oncoprotein. Importantly, in the model of breast tumorigenesis, transgene expression in differentiated mammary epithelial cells leads to the development of undifferentiated tumors exhibiting all the characteristic features of the claudin-low subtype. These observations challenge the concept that this tumor subtype specifically arises from transformation of an early epithelial precursor with inherent stemness properties and metaplastic features. Consistently, oncogenic cooperation assays performed in human mammary epithelial cells with Twist or Zeb EMT-inducers in combination with H-RasG12V generate transformed cell lines displaying all characteristics of claudin-low tumors including mesenchymal features, undifferentiated traits, and stem-cell-like properties. EMT might thus drive the development of claudin-low tumors by exhibiting a dual role in cell transformation and dedifferentiation. In other terms, the claudin-low tumor subtype of breast cancers might thus constitute a first example of human adult malignancies driven by aberrant reactivation of an embryonic transdifferentiation program. 96 Identification of EMTed Carcinoma and Potential Therapeutic Strategies J. Thiery1 , W.J. Sim1 , A. Ng1 , K. Chua2 , R. Huang2 , S. Mori2 , T. Tan2 , J.Y. Pierga3 , F.C. Bidart3 . 1 Institute of Molecular and Cell Biology (IMCB), 138673, Singapore, 2 Cancer Science Institute, NUS, 117456, Singapore, 3 Institut Curie, Paris, France Epithelial mesenchymal transition (EMT) is a major process controlling multiple events during development. Mesenchyme appeared as a transient state in diploblasts more than 600 million years ago. Since then, EMT has been conserved throughout evolution to control morphogenetic events, such as the formation of the three primary germ layers during gastrulation. Most interestingly, signal transduction pathways have been remarkably conserved in many different species. EMT is likely to operate during the early stages of carcinoma invasion that lead to blood or lymph vessel intravasation. Mesenchymal-like carcinoma cells undergo a mesenchymal to epithelial transition in distant sites from the primary tumor and eventually become macrometastatic. We have characterized bone marrow micrometastases from breast cancer patients and found that the detection of micrometastatic carcinoma cells was associated with poorer distant metastasis-free survival, local relapse-free survival, and overall survival. Thus, bone marrow micrometastases should become a very useful prognostic indicator for relapse, and an excellent surrogate marker for patient’s response to treatment. The mesenchymal-like state of carcinoma confers stemness, protection from cell death, escape from immune response and, most importantly, resistance to conventional and targeted therapies. Current strategies based on the EMT concept are aimed at designing new therapeutic approaches that interfere with the plasticity of carcinoma cells. We have designed a robust EMT score for ovarian and breast carcinoma cell lines and tumors, with the ultimate goal to stratify individual cancer patients. In parallel, our laboratory has devised a high-content, high-throughput screen for EMT. Several combinations of drugs have been shown to selectively inhibit EMT. This strategy may be used to interfere with tumor progression, particularly in ovarian carcinomas that have acquired resistance to conventional therapies.