- Email: [email protected]
Approaches to elucidate cancer metabolism using zebrafish models
EACR24 Poster Sessions / European Journal of Cancer 61, Suppl. 1 (2016) S9–S218 of these CSC marker in miRNA recruitment or in miRNA processing, which remains to be explored. Finally, Tspan8 and associated integrins play a major role in TEX targeting. CSC protein markers are gathered in GEM, become jointly internalized, are engaged in exosome biogenesis including loading and are recovered in TEX. Joint target modulation is promoted by associations between CSC markers and additional TEX components including miRNA. Thus, TEX-based diagnosis and therapy should appreciate the central role of CSC markers. No conflict of interest. 352 PKCtheta controls invasion of triple-negative breast cancer by direct phosphorylation of FAK L. Chadelle1 , V. Cadamuro1 , J. Liu1 , X. Wang1 , K. Belguise1 . 1 LBCMCPUniversite´ Paul Sabatier- CNRS UMR5088, LBCMCP- Universite´ Paul Sabatier- CNRS UMR5088, Toulouse, France Background: Triple-negative breast cancers (TNBC: negative for oestrogenreceptor, progesterone-receptor and human-epidermal-growth-factor receptor 2 [HER2]) account for 15−20% of breast cancers and are characterized by an extremely aggressive phenotype. Currently, there is no efficient targeted therapy that can prevent the metastatic dissemination of TNBC. The serine threonine kinase PKCQ has recently been identified as a potential new marker of TNBC and as a strong inducer of cancer invasion. However, how PKCQ controls tumour invasion in TNBC is still unclear. Material and Methods: To study PKCQ control of invasion, we worked on several breast cancer cell lines combining biochemical and imaging approaches. Results: Here our results show that PKCQ can promote invasion by activating the focal adhesion kinase (FAK) signalling that regulates the cycle of focal contact formation and disassembly. As a consequence of this activation, PKCQ enhances the dynamic of adhesion turnover, Rho GTPases activity and protrusions formation. We also have unravelled the molecular mechanisms by which PKCQ could control FAK: PKCQ interacts with and phosphorylates FAK at S890–892–893 thus leading to its opening and activation. By using non phosphorylatable mutants of FAK, we show that the identified phosphorylation sites are essential to the PKCQ controlled invasion. Conclusion: We show that PKCQ is controlling invasion by directly phosphorylating FAK at S890–892–893. We are currently testing the in vivo relevance of this mechanism by mouse metastasis assay. Next, we will investigate the possibility of a correlation between the presence of the identified phosphorylations on patient samples and the tumour’s aggressiveness. In the long term, this study could help develop new targeted therapy for TNBC such as PKCQ specific inhibitors that would limit metastasis formation. No conflict of interest. 353 Quantitative analysis of estrogen receptors alpha and beta expression in ovarian cancer patients by flow cytometry E. Dudko1 , T. Bogush1 , S. Kolomiytsev1 , O. Rjabinina1 , A. Tjulandina2 , V. Kirsanov3 , E. Bogush3 , S. Tjulandin2 , I. Mamichev1 . 1 N.N. Blokhin Russian Cancer Research Center, Laboratory of medical chemistry, Moscow, Russian Federation, 2 N.N. Blokhin Russian Cancer Research Center, Department of clinical pharmacology and chemotherapy, Moscow, Russian Federation, 3 N.N. Blokhin Russian Cancer Research Center, Department of surgery, Moscow, Russian Federation Introduction: The results of chemotherapy of ovarian cancer is not satisfactory enough and new approaches are needed. Perspective therapy strategy implying estrogen receptors (ER) as drug target is investigated. Usually only ER alpha are examined, but recent researches had shown that the other type of ER − ER beta − is a significant pathogenetic parameter of ovarian cancer as well. The purpose of the study is estimation expression and co-expression of both types of ER − ER alpha and beta − in ovarian cancer tissue. Materials and Methods: 53 serous ovarian cancer surgical biopsy specimens were analyzed by flow cytometry. Single-cell suspensions were incubated overnight with primary anti-ER alpha (SP1, Abcam) and anti-ER beta (14C8, Abcam) and then − for 1.5 h with secondary antibodies, conjugated with DyLight650 (ab98510 and ab98729, respectively). Cell fluorescence was analyzed with FlowJo 10.0.8 software and number of specifically fluorescent cells was calculated by using Kolmogorov–Smirnov statistical approach. Three levels of ER expression were analyzed: high − ER were revealed more than in 40% of the cells; low − in 15−40%; negative − less than in 15%. Data were treated by nonparametric methods, Spearman rank correlation coefficient (rs) was used for estimation of dependence between expression of ER alpha and beta. Results and Discussion: (1) ER beta was revealed in all the patients (pts) but ER alpha − in 75% only. (2) In the pts with ER positive tumors low ER alpha level was revealed in 1.5 times often in comparison to ER beta (57% vs 38%), but for high ER alpha level, opposite, it was 3 times rarely in comparison to ER beta (20% vs 62%). (3) In 19% of the pts co-expression high level of ER alpha and ER beta was revealed. (4) Moderate correlation
S71
between expression of ER alpha and ER beta was found (rs = 0.61; 95% CI 0.38−0.74). Conclusion: It was shown that (1) ovarian cancer tumors are actually ER positive ones with higher level and frequency of ER beta expression as compared to ER alpha; (2) expression of both markers was correlated each other and (3) in 1/5 of the tumors co-expression high level of ER alpha and ER beta was revealed. We suppose that controversial information about contribution of ER in pathogenesis of ovarian cancer is because one type of ER − ER alpha only − is estimated in the tumor routinely. We believe that quantitative determination both ER alpha and ER beta in the tumor let distinguish the pts with high level of ER expression (ER alpha, ER beta or both), who needs adjuvant antiestrogen therapy similarly with the ER alpha positive breast cancer pts. Acknowledgement: Supported by RFBR grants (15-04-06991-a, 16−34-01049mol-a) and President of Russian Federation grant MK-7709.2016.7. No conflict of interest. 354 Colorectal adenocarcinoma stem cells show distinct growth patterns depending on the culture environment A. Borys1 , P. Wołkow1 . 1 Jagiellonian University Medical College, Center for Medical Genomics OMICRON, Krakow, Poland Introduction: Disease models for basic cancer research rely predominantly on established cell lines and animal studies. Lgr5, an R-spondin receptor, and a specific marker of gastrointestinal tract stem cells, enabled in vitro culture and observation of all steps of their differentiation and self-organization from single cells to “mini-organs”. Organoid cultures are currently the state-of-theart model for studies of colon, esophagus and small intestine pathophysiology. We used this method to produce cancer colonoids from stem cells, isolated from LoVo cell line. LoVo cell line, established from Dukes’ type C, grade IV, colorectal adenocarcinoma, contains, according to the literature, a significant fraction of cancer stem cells (CSCs). Materials and Methods: The characteristics of CSC derived cultures were assessed in two types of growth environments: 2D and 3D (laminin-rich extracellular matrix). Lgr5+ CSCs were isolated from amplified LoVo cells with anti-Lgr5+ magnetic beads. Proportion of living and dead cells, were measured with fluorescence/luminescence assay at several time points. Developed colonoids were treated with demethylation agent 5-azacytidine and with monoclonal antibody Cetuximab. Results and Discussion: The results show that established LoVo cell line contains small but relevant fraction of stem cells (1.5−3%), which can be isolated and cultured. CSC growth pattern and cell morphology are grossly affected by 2D or 3D culture environment. Cells kept in 2D, form cell aggregates, very similar to general LoVo population with elongated cells extending toward the inter-cellular space. Cells cultured in 3D matrix grow as compact ‘colonoids’ with almost invisible cell junctions and very smooth edges. Three-dimensional growth conditions support cell survival − there is a significantly higher proportion of living cells and lower of dead cells than in 2D. Colon cancer ‘colonoids’ show distinct response patterns to cytotoxic drugs treatment. Cells treated with 5-azacytidine show higher rate of viable cells in three-dimensional culture than in 2D conditions. LoVo cells were demonstrated to exhibit the high level of EGFR protein and therefore should be very sensitive to monoclonal EGFR targeted antibody, e.g. Cetuximab. Our study shows that colon cancer ‘colonoids’ are even more affected by Cetuximab in 3D conditions than in traditional monolayer. Conclusion: CSCs derived from LoVo cell line present distinct growth patterns with varying culture environment. In 2D, they form a monolayer of spindleshaped cells, in 3D they generate round, spherical structures with smooth edges. As organoids should develop all cell types, present in vivo in human colon, we plan to investigate cell subtypes present in LoVo ‘colonoids’. Also, patterns of demethylation induced changes in gene expression are under evaluation by microarrays. No conflict of interest. 355 Approaches to elucidate cancer metabolism using zebrafish models M.C. Mione1 , A. Kalkbrenner2 , V. Gourain1 , M. Mayrhofer1 , S. Burkart1 , C. Muhle-Goll2 , B. Luy2 . 1 Karlsruhe Institute of Technology, Institute for Toxicology and Genetics, Eggenstein-Leopoldshafen, Germany, 2 Karlsruhe Institute of Technology, Institute of Organic Chemistry, Karlsruhe, Germany Introduction: Metabolic reprogramming is a hallmark of cancer, as proliferating cells require a large amount of nutrients, energy, and biosynthetic activity to produce the macromolecular components required by the growing tumour cells. This leads to a switch from oxidative metabolism to glycolysis (Warburg 1925; 1956) to utilize metabolites, such as lactate and glutamine, as building blocks for macromolecule biosynthesis. Previously, changes in metabolism were mostly attributed to opportunistic utilization of alternative biochemical pathways; however, the systematic analysis of cancer genomes and their expression profiling have revealed a surprising abundance of genetic
S72
EACR24 Poster Sessions / European Journal of Cancer 61, Suppl. 1 (2016) S9–S218
alterations of key metabolic enzymes as early events in cancer. These are particularly interesting as enzymatic activities can be targeted with great specificity and success. To understand the consequences of altered enzymatic activities on cancer metabolism in vivo, analytical methods for quantifying metabolites, and system approaches to analyse the genetic and metabolome data in model systems are required. Materials and Methods: We have started to develop integrated maps of metabolic alterations at different stages of melanoma and glioma development using our zebrafish cancer models (Santoriello et al., 2010; Spitzner et al., 2014). We combined metabolome analysis (through NMR) and gene expression data (through RNA-Seq) at different stages of cancer development, in whole animals and dissected tumors. Results and Discussion: We identified metabolic reprogramming in the choline pathway leading to decreased activity of the trans-sulphuration pathway and S-adenosyl-methionine availability, and increased glutamine /acetate production in glioma due to extensive reprogramming of the TCA cycle. Conclusions: Changes in metabolites and enzyme expression leading to these metabolic reprogramming events were already detectable in transgenic larvae, thus suggesting that zebrafish cancer models may serve to study the evolution of metabolic reprogramming in cancer and provide an in vivo platform to screen for metabolic vulnerability in cancer. No conflict of interest. 356 Non-canonical Hedgehog/AMPK-mediated control of polyamine metabolism is required for medulloblastoma growth G. Canettieri1 , S. Coni1 , D.M. Laura1 , G. Sdruscia1 . 1 Sapienza University, Department of Molecular Medicine, Roma, Italy Introduction: Developmental Hedgehog signaling controls proliferation of cerebellar granule cell precursors (GCPs) and its aberrant activation is a leading cause of Medulloblastoma (SHH-MB, Hedgehog molecular subgroup). Treatment of SHH-MB patients with the FDA-approved Hh inhibitor Vismodegib (which targets the transducer Smoothened) has been disappointing because of the occurrence of resistance, attributed to Smo inactivating mutations or to activation of downstream effectors. Thus, it is now believed that the identification and targeting of novel downstream components represents a preferable option for this disease. Material and Methods: We used biochemical, cell biology and molecular biology approaches to characterize the molecular features of a novel mechanism involved in polyamine production upon Hh activation. The biological relevance of our findings has been evaluated in mice and patients with SHH MB. Results and Discussion: We show here that Hedgehog promotes polyamine biosynthesis in GCPs by engaging a non-canonical axis leading to ODC translation. This process is regulated by AMPK, which phosphorylates threonine 173 of CNBP in response to Hedgehog activation. Phosphorylated CNBP increases its association with Sufu, followed by CNBP stabilization, ODC translation and polyamine biosynthesis. Notably, CNBP, ODC and polyamines are hallmarks Hedgehog-dependent Medulloblastoma (SHH-MB) and genetic or pharmacological inhibition of this axis efficiently blocks Hedgehog-dependent proliferation of Medulloblastoma cells in preclinical settings. Conclusions: Together, these data illustrate a novel auxiliary mechanism of metabolic control by a morphogenic pathway with relevant implications in cancer. No conflict of interest. 357 The impact of Mycobacterium obuense on innate and adaptive immunity J. Crooks1 , S. Brown1 , C. Forss1 , A. Phythian-Adams1 , P. Cook1 , L. Rosa Brunet2 , A. MacDonald1 . 1 University of Manchester, MCCIR, Manchester, United Kingdom, 2 Immodulon Therapeutics Ltd, Immodulon Therapeutics Ltd, London, United Kingdom Background: IMM-101 is an immunomodulatory treatment, comprising of heat killed whole cell Mycobacterium obuense (NCTC13365), currently under clinical investigation in a variety of cancers, including pancreatic cancer. At present, available pancreatic cancer treatments only provide 5% survival rate over 5 years, with a median overall survival rate of 6−12 months, highlighting the need for new therapies. A phase II clinical trial (NCT01303172) has shown that treatment with the immunomodulator IMM-101 in parallel with gemcitabine increases median survival to 7.0 months in patients with metastatic pancreatic cancer, compared to 4.4 months following treatment with gemcitabine alone (Dalgleish, A.G. and The IMAGE I Trial investigators, 2015). In light of the promising results of this combination treatment, elucidation of how IMM101 induces its therapeutic effects is ongoing. Initial pre-clinical studies have suggested that antigen specific CD45RBlowCD44high cytotoxic CD8+ T cells are increased in IMM-101 treated animals and may play a key role.
Material and Methods: Dendritic Cell (DC) studies were conducted on IMM101 (or control) stimulated murine bone marrow derived GMCSF DCs or FLT-3 DCs, or human monocyte derived DCs. DC phenotypic activation was assessed by flow cytometry, and cytokine secretion assessed in culture supernatants by ELISA or cytokine bead array. CD4+ T cells used in murine coculture studies were isolated from the lymph nodes and spleens of transgenic OTII mice before being stained with CFSE to assess proliferation via flow cytometry. Results: We have investigated the impact of different concentrations of IMM101 on murine and human DC activation and function. GMCSF derived murine DCs displayed a dose dependent response to the immunomodulatory IMM101, with elevated expression of activation markers and increased secretion of IL-6, IL-12p40 and NO, compared to controls. This dose dependent response to IMM-101 was also replicated in murine FLT-3 generated DCs (pDCs, CD11b+ cDCs and CD24+ cDCs) and human monocyte derived DCs. Using murine DC:T cell co-cultures, we have demonstrated that IMM-101 influences the processing and presentation of antigen by DCs to CD4+ T cells. We have also shown that IMM-101 activated DCs instigate a T cell specific IFN-gamma, and a non-T cell specific IL-17, response following re-stimulation of draining lymph node cell preparations obtained 7 days after adoptive transfer of DCs into na¨ıve recipient mice. Conclusion: Our demonstration that IMM-101 activates primary DCs, both murine and human, and promotes IFN-gamma and IL-17 production in vivo helps build a better understanding of the fundamental mechanisms that may bridge innate and adaptive immune responses and may help elucidate the mode of action of this promising cancer treatment. No conflict of interest. 358 Novel methods for the isolation of tumor cells from human, mouse, and xenografted tumors D. Agorku1 , A. Langhammer1 , L. Willnow1 , K. Klingner2 , S. Tomiuk1 , 1 J. Kollet1 , S. Ruberg ¨ , J. Schuler ¨ 2 , A. Bosio1 , O. Hardt1 . 1 Miltenyi Biotec GmbH, Research and Development, Bergisch Gladbach, Germany, 2 Oncotest GmbH, In vivo Tumorbiology, Freiburg, Germany Solid tumors are vascularized and infiltrated by stromal cells such as leukocytes, endothelial cells and fibroblasts. The amount and composition of those non-tumor cells depends on various factors including tumor entity and stage, treatment history, status of the host organism and site of tumor growth. This results in a widely unpredictable contamination of non-tumor cells within tumor samples frequently hampering downstream applications like cultivation and molecular analysis. However, the direct isolation of target cells is difficult as there is a lack of markers exclusively expressed by tumor cells in many cases. To overcome these hurdles, we have developed methods allowing for the isolation of tumor cells from human, mouse, and xenografted tumors by specifically depleting non-tumor cells from the respective primary material by utilizing magnetic cell sorting. The combinations of antibodies specifically binding all non-tumor cells from human, mouse, and xenograft tumors have been identified by flow cytometrybased screening assays on multiple tissue samples as well as cell lines. The respective antibodies were coupled to superparamagnetic nanoparticles and optimized for efficient depletion utilizing MACS. Performance of nontumor cell depletion was assessed by flow cytometry and cell culture followed by immunocytochemistry of separated cells. Additionally, whole exome sequencing on bulk tumor and separated samples was performed to assess improvement in downstream analysis. Three different antibody cocktails were established for depletion of non-tumor cells from all common sources of tumor material. Using these optimized antibody combinations coupled to superparamagnetic nanoparticles allowed for the fast isolation of untouched tumor cells independent of tumor entity and origin. Typically, purities higher than 95% were achieved in less than 25 minutes. Even when starting with samples such as pleural effusions, containing only minor frequencies of tumor cells, purities of higher than 85% were reached by a single isolation step. By depletion of stromal cells, the downstream culture of tumor cells was significantly improved and standardized. In downstream analysis more SNPs were identified by next generation sequencing, and gene expression signatures of isolated tumor cells and cancer stem cells could be more reliably be interpreted upon removal of the bias caused by non-tumor cells. Taken together, we have developed easy and fast procedures to isolate tumor cells independent of tumor specific markers or tumor entity from human, mouse, and xenografted tumors. This allows for increased sensitivity, specificity and reproducibility in downstream analysis of tumor cells and subpopulations thereof. The “untouched” isolation of tumor cells allows for the subsequent sorting and analyses of minor tumor subpopulations, such as cancer stem cells. No conflict of interest.
S71
between expression of ER alpha and ER beta was found (rs = 0.61; 95% CI 0.38−0.74). Conclusion: It was shown that (1) ovarian cancer tumors are actually ER positive ones with higher level and frequency of ER beta expression as compared to ER alpha; (2) expression of both markers was correlated each other and (3) in 1/5 of the tumors co-expression high level of ER alpha and ER beta was revealed. We suppose that controversial information about contribution of ER in pathogenesis of ovarian cancer is because one type of ER − ER alpha only − is estimated in the tumor routinely. We believe that quantitative determination both ER alpha and ER beta in the tumor let distinguish the pts with high level of ER expression (ER alpha, ER beta or both), who needs adjuvant antiestrogen therapy similarly with the ER alpha positive breast cancer pts. Acknowledgement: Supported by RFBR grants (15-04-06991-a, 16−34-01049mol-a) and President of Russian Federation grant MK-7709.2016.7. No conflict of interest. 354 Colorectal adenocarcinoma stem cells show distinct growth patterns depending on the culture environment A. Borys1 , P. Wołkow1 . 1 Jagiellonian University Medical College, Center for Medical Genomics OMICRON, Krakow, Poland Introduction: Disease models for basic cancer research rely predominantly on established cell lines and animal studies. Lgr5, an R-spondin receptor, and a specific marker of gastrointestinal tract stem cells, enabled in vitro culture and observation of all steps of their differentiation and self-organization from single cells to “mini-organs”. Organoid cultures are currently the state-of-theart model for studies of colon, esophagus and small intestine pathophysiology. We used this method to produce cancer colonoids from stem cells, isolated from LoVo cell line. LoVo cell line, established from Dukes’ type C, grade IV, colorectal adenocarcinoma, contains, according to the literature, a significant fraction of cancer stem cells (CSCs). Materials and Methods: The characteristics of CSC derived cultures were assessed in two types of growth environments: 2D and 3D (laminin-rich extracellular matrix). Lgr5+ CSCs were isolated from amplified LoVo cells with anti-Lgr5+ magnetic beads. Proportion of living and dead cells, were measured with fluorescence/luminescence assay at several time points. Developed colonoids were treated with demethylation agent 5-azacytidine and with monoclonal antibody Cetuximab. Results and Discussion: The results show that established LoVo cell line contains small but relevant fraction of stem cells (1.5−3%), which can be isolated and cultured. CSC growth pattern and cell morphology are grossly affected by 2D or 3D culture environment. Cells kept in 2D, form cell aggregates, very similar to general LoVo population with elongated cells extending toward the inter-cellular space. Cells cultured in 3D matrix grow as compact ‘colonoids’ with almost invisible cell junctions and very smooth edges. Three-dimensional growth conditions support cell survival − there is a significantly higher proportion of living cells and lower of dead cells than in 2D. Colon cancer ‘colonoids’ show distinct response patterns to cytotoxic drugs treatment. Cells treated with 5-azacytidine show higher rate of viable cells in three-dimensional culture than in 2D conditions. LoVo cells were demonstrated to exhibit the high level of EGFR protein and therefore should be very sensitive to monoclonal EGFR targeted antibody, e.g. Cetuximab. Our study shows that colon cancer ‘colonoids’ are even more affected by Cetuximab in 3D conditions than in traditional monolayer. Conclusion: CSCs derived from LoVo cell line present distinct growth patterns with varying culture environment. In 2D, they form a monolayer of spindleshaped cells, in 3D they generate round, spherical structures with smooth edges. As organoids should develop all cell types, present in vivo in human colon, we plan to investigate cell subtypes present in LoVo ‘colonoids’. Also, patterns of demethylation induced changes in gene expression are under evaluation by microarrays. No conflict of interest. 355 Approaches to elucidate cancer metabolism using zebrafish models M.C. Mione1 , A. Kalkbrenner2 , V. Gourain1 , M. Mayrhofer1 , S. Burkart1 , C. Muhle-Goll2 , B. Luy2 . 1 Karlsruhe Institute of Technology, Institute for Toxicology and Genetics, Eggenstein-Leopoldshafen, Germany, 2 Karlsruhe Institute of Technology, Institute of Organic Chemistry, Karlsruhe, Germany Introduction: Metabolic reprogramming is a hallmark of cancer, as proliferating cells require a large amount of nutrients, energy, and biosynthetic activity to produce the macromolecular components required by the growing tumour cells. This leads to a switch from oxidative metabolism to glycolysis (Warburg 1925; 1956) to utilize metabolites, such as lactate and glutamine, as building blocks for macromolecule biosynthesis. Previously, changes in metabolism were mostly attributed to opportunistic utilization of alternative biochemical pathways; however, the systematic analysis of cancer genomes and their expression profiling have revealed a surprising abundance of genetic
S72
EACR24 Poster Sessions / European Journal of Cancer 61, Suppl. 1 (2016) S9–S218
alterations of key metabolic enzymes as early events in cancer. These are particularly interesting as enzymatic activities can be targeted with great specificity and success. To understand the consequences of altered enzymatic activities on cancer metabolism in vivo, analytical methods for quantifying metabolites, and system approaches to analyse the genetic and metabolome data in model systems are required. Materials and Methods: We have started to develop integrated maps of metabolic alterations at different stages of melanoma and glioma development using our zebrafish cancer models (Santoriello et al., 2010; Spitzner et al., 2014). We combined metabolome analysis (through NMR) and gene expression data (through RNA-Seq) at different stages of cancer development, in whole animals and dissected tumors. Results and Discussion: We identified metabolic reprogramming in the choline pathway leading to decreased activity of the trans-sulphuration pathway and S-adenosyl-methionine availability, and increased glutamine /acetate production in glioma due to extensive reprogramming of the TCA cycle. Conclusions: Changes in metabolites and enzyme expression leading to these metabolic reprogramming events were already detectable in transgenic larvae, thus suggesting that zebrafish cancer models may serve to study the evolution of metabolic reprogramming in cancer and provide an in vivo platform to screen for metabolic vulnerability in cancer. No conflict of interest. 356 Non-canonical Hedgehog/AMPK-mediated control of polyamine metabolism is required for medulloblastoma growth G. Canettieri1 , S. Coni1 , D.M. Laura1 , G. Sdruscia1 . 1 Sapienza University, Department of Molecular Medicine, Roma, Italy Introduction: Developmental Hedgehog signaling controls proliferation of cerebellar granule cell precursors (GCPs) and its aberrant activation is a leading cause of Medulloblastoma (SHH-MB, Hedgehog molecular subgroup). Treatment of SHH-MB patients with the FDA-approved Hh inhibitor Vismodegib (which targets the transducer Smoothened) has been disappointing because of the occurrence of resistance, attributed to Smo inactivating mutations or to activation of downstream effectors. Thus, it is now believed that the identification and targeting of novel downstream components represents a preferable option for this disease. Material and Methods: We used biochemical, cell biology and molecular biology approaches to characterize the molecular features of a novel mechanism involved in polyamine production upon Hh activation. The biological relevance of our findings has been evaluated in mice and patients with SHH MB. Results and Discussion: We show here that Hedgehog promotes polyamine biosynthesis in GCPs by engaging a non-canonical axis leading to ODC translation. This process is regulated by AMPK, which phosphorylates threonine 173 of CNBP in response to Hedgehog activation. Phosphorylated CNBP increases its association with Sufu, followed by CNBP stabilization, ODC translation and polyamine biosynthesis. Notably, CNBP, ODC and polyamines are hallmarks Hedgehog-dependent Medulloblastoma (SHH-MB) and genetic or pharmacological inhibition of this axis efficiently blocks Hedgehog-dependent proliferation of Medulloblastoma cells in preclinical settings. Conclusions: Together, these data illustrate a novel auxiliary mechanism of metabolic control by a morphogenic pathway with relevant implications in cancer. No conflict of interest. 357 The impact of Mycobacterium obuense on innate and adaptive immunity J. Crooks1 , S. Brown1 , C. Forss1 , A. Phythian-Adams1 , P. Cook1 , L. Rosa Brunet2 , A. MacDonald1 . 1 University of Manchester, MCCIR, Manchester, United Kingdom, 2 Immodulon Therapeutics Ltd, Immodulon Therapeutics Ltd, London, United Kingdom Background: IMM-101 is an immunomodulatory treatment, comprising of heat killed whole cell Mycobacterium obuense (NCTC13365), currently under clinical investigation in a variety of cancers, including pancreatic cancer. At present, available pancreatic cancer treatments only provide 5% survival rate over 5 years, with a median overall survival rate of 6−12 months, highlighting the need for new therapies. A phase II clinical trial (NCT01303172) has shown that treatment with the immunomodulator IMM-101 in parallel with gemcitabine increases median survival to 7.0 months in patients with metastatic pancreatic cancer, compared to 4.4 months following treatment with gemcitabine alone (Dalgleish, A.G. and The IMAGE I Trial investigators, 2015). In light of the promising results of this combination treatment, elucidation of how IMM101 induces its therapeutic effects is ongoing. Initial pre-clinical studies have suggested that antigen specific CD45RBlowCD44high cytotoxic CD8+ T cells are increased in IMM-101 treated animals and may play a key role.
Material and Methods: Dendritic Cell (DC) studies were conducted on IMM101 (or control) stimulated murine bone marrow derived GMCSF DCs or FLT-3 DCs, or human monocyte derived DCs. DC phenotypic activation was assessed by flow cytometry, and cytokine secretion assessed in culture supernatants by ELISA or cytokine bead array. CD4+ T cells used in murine coculture studies were isolated from the lymph nodes and spleens of transgenic OTII mice before being stained with CFSE to assess proliferation via flow cytometry. Results: We have investigated the impact of different concentrations of IMM101 on murine and human DC activation and function. GMCSF derived murine DCs displayed a dose dependent response to the immunomodulatory IMM101, with elevated expression of activation markers and increased secretion of IL-6, IL-12p40 and NO, compared to controls. This dose dependent response to IMM-101 was also replicated in murine FLT-3 generated DCs (pDCs, CD11b+ cDCs and CD24+ cDCs) and human monocyte derived DCs. Using murine DC:T cell co-cultures, we have demonstrated that IMM-101 influences the processing and presentation of antigen by DCs to CD4+ T cells. We have also shown that IMM-101 activated DCs instigate a T cell specific IFN-gamma, and a non-T cell specific IL-17, response following re-stimulation of draining lymph node cell preparations obtained 7 days after adoptive transfer of DCs into na¨ıve recipient mice. Conclusion: Our demonstration that IMM-101 activates primary DCs, both murine and human, and promotes IFN-gamma and IL-17 production in vivo helps build a better understanding of the fundamental mechanisms that may bridge innate and adaptive immune responses and may help elucidate the mode of action of this promising cancer treatment. No conflict of interest. 358 Novel methods for the isolation of tumor cells from human, mouse, and xenografted tumors D. Agorku1 , A. Langhammer1 , L. Willnow1 , K. Klingner2 , S. Tomiuk1 , 1 J. Kollet1 , S. Ruberg ¨ , J. Schuler ¨ 2 , A. Bosio1 , O. Hardt1 . 1 Miltenyi Biotec GmbH, Research and Development, Bergisch Gladbach, Germany, 2 Oncotest GmbH, In vivo Tumorbiology, Freiburg, Germany Solid tumors are vascularized and infiltrated by stromal cells such as leukocytes, endothelial cells and fibroblasts. The amount and composition of those non-tumor cells depends on various factors including tumor entity and stage, treatment history, status of the host organism and site of tumor growth. This results in a widely unpredictable contamination of non-tumor cells within tumor samples frequently hampering downstream applications like cultivation and molecular analysis. However, the direct isolation of target cells is difficult as there is a lack of markers exclusively expressed by tumor cells in many cases. To overcome these hurdles, we have developed methods allowing for the isolation of tumor cells from human, mouse, and xenografted tumors by specifically depleting non-tumor cells from the respective primary material by utilizing magnetic cell sorting. The combinations of antibodies specifically binding all non-tumor cells from human, mouse, and xenograft tumors have been identified by flow cytometrybased screening assays on multiple tissue samples as well as cell lines. The respective antibodies were coupled to superparamagnetic nanoparticles and optimized for efficient depletion utilizing MACS. Performance of nontumor cell depletion was assessed by flow cytometry and cell culture followed by immunocytochemistry of separated cells. Additionally, whole exome sequencing on bulk tumor and separated samples was performed to assess improvement in downstream analysis. Three different antibody cocktails were established for depletion of non-tumor cells from all common sources of tumor material. Using these optimized antibody combinations coupled to superparamagnetic nanoparticles allowed for the fast isolation of untouched tumor cells independent of tumor entity and origin. Typically, purities higher than 95% were achieved in less than 25 minutes. Even when starting with samples such as pleural effusions, containing only minor frequencies of tumor cells, purities of higher than 85% were reached by a single isolation step. By depletion of stromal cells, the downstream culture of tumor cells was significantly improved and standardized. In downstream analysis more SNPs were identified by next generation sequencing, and gene expression signatures of isolated tumor cells and cancer stem cells could be more reliably be interpreted upon removal of the bias caused by non-tumor cells. Taken together, we have developed easy and fast procedures to isolate tumor cells independent of tumor specific markers or tumor entity from human, mouse, and xenografted tumors. This allows for increased sensitivity, specificity and reproducibility in downstream analysis of tumor cells and subpopulations thereof. The “untouched” isolation of tumor cells allows for the subsequent sorting and analyses of minor tumor subpopulations, such as cancer stem cells. No conflict of interest.