Poster session 2 (free drinks)

Chiara Agnoletto¹, Linda Minotti¹, Laura Brulle–Soumare², Marco Galasso¹, Fabio Corrà¹, Federica Baldassari¹, Jean-Gabriel Judde², Stefano Cairo², and Stefano Volinia¹

¹ Department of Morphology, Surgery and Experimental Medicine, LTTA Centre, University of Ferrara, 44121 Ferrara, Italy

² Xentech, Paris France

Introduction

Metastases from primary tumours account for the great majority of cancer-related deaths. This process is thought to involve a series of sequential steps, including the release of circulating tumour cells (CTCs) into the vasculature. CTCs have been detected in the peripheral blood of patients with advanced cancers. Recently, liquid biopsy has received enormous attention because of its clinical implications. Clinical applications of CTCs detection include early cancer diagnosis, prediction of the risk for metastatic relapse or progression, monitoring the effects of systemic therapies, and stratification of patients.

To date, only incomplete information is available on CTC molecular and biological heterogeneity, and significance. CTCs are extremely rare, even in patients with advanced metastatic cancers, posing a serious challenge for any analytical system. Thus, there is a critical need for increased detection sensitivity, for the capturing of rare tumour cells circulating in the blood. Most of the current CTC isolation techniques rely on antibody-based capture, and epithelial markers expressed on tumours have been frequently used to purify cancer cells. Thus, analysis of CTCs has been hampered by reliance on a restricted pool of markers, often leading to false-negative results.

Material and Methods

To address these limitations, we describe a novel approach that does not rely on the use of epithelial markers such as EPCAM for the analysis of CTCs in peripheral blood of patient-derived xenograft (PDX) models, independently of physical or immunological purification. A qRT-PCR assay was developed, using human-specific primers, to amplify the CTC markers EPCAM and keratins in a PDX cohort of breast cancer. The expression levels non-detected by the assay were imputed using the expectation–maximization (EM) algorithm. Normalized expression values for the human genes were calculated relative to murine Actb. The expression values for human target genes were further normalized on each human reference gene, to investigate the expression of epithelial markers within the human CTCs.

Results and Discussion

We demonstrated that the expression of human epithelial markers greatly varies, with no correlation between their expression and the content of CTCs. By using this highly specific and sensitive approach, without any CTC pre-isolation, we could identify and measure CTC mRNAs irrespectively of EPCAM content. Our innovative study holds unprecedented potential for the study of CTC heterogeneity and for the identification of novel CTC markers.

Conclusion

The identification and the relative quantification of the diverse spectrum of CTCs will have a great impact on personalized medicine: unrestricted CTCs characterization will allow the early detection of metastases in cancer patients and the assessment of ad-hoc therapies for personalised medicine.

Francesca Chiovaro¹, Irina Agarkova¹, Armin Maier², Simon Messner¹, JuliaSchueler², Patrick Guye¹

¹ InSphero AG, Wagistr. 27, 8952, Schlieren, Switzerland

² Charles River DRS Germany GmbH Am Flughafen 12, 79108 Freiburg im Breisgau, Deutschland

Introduction

A major issue for improving the overall success rate in drug development is the lack of accurate experimental human in-vitro models. While there is a significant and increasing need for such better ex-vivo cell-based models, the question on how closely these systems resemble and recapitulate the original tumors is of great importance. Patient-derived xenograft (PDX) models act as vehicle systems to propagate human tumor specimens, faithfully preserving the biological features and the genetic expression profile.

The retainment of such criteria in in vitro 3D InSight™ Tumor Microtissues derived from PDX lines is crucial to provide a relevant physiological environment and strategy to assess candidate drugs for novel therapeutic approaches. Here we aimed to develop and characterize in vitro 3D InSight™ Tumor Microtissues from patient-derived xenograft models.

Material and Methods

PDX cell suspensions of Lung, Breast and Melanoma origin were successfully used to assess 3D aggregation in 96 well format and characterized over 10 days in culture. After careful removal of mouse cell contaminants in each in vitro 3D PDX sample, PDX cell cultures were supplied with exogenous normal human dermal fibroblasts (nHDF). The morphology, biomarker phenotype (IHC) as well as cell proliferation were assessed by histological analysis. In addition to screening for standard diagnostic marker such as proliferating vs. dead cells (e.g. Ki67, ClCasp3) and stromal vs. epithelial-tumor cells (e.g. FAP, pan-CK, ECadherin), we also assessed the expression of cancer type-specific biomarkers. Moreover, to monitor the dynamics of cancer phenotypic alterations, epithelial-to-mesenchymal transition (EMT) marker were evaluated and scored.

Results and Discussion

Immunohistochemistry assessment of 3D microtumors validated the resemblance with their respective PDX tumor models. The viability and growth rate of PDX-derived microtumors were assessed by size analysis (cell scanner) and ATP assay. 3D tumor growth rate and cell behavior observations reflected the diversity of disease progression in vivo. Furthermore, we assessed the efficacy of specific targeted therapies which were selected on the basis of the distinct molecular signatures of PDX tumor cells.

Conclusion

Development and characterization of in vitro 3D InSight™ Tumor Microtissues from patient-derived xenograft (PDX) lines demonstrated that the morphological and molecular features of the parental tumors are well retained. We suggest that in vitro 3D PDX models offer a more suitable and robust approach to expedite faithful efficacy assessment and approval of optimal drug candidates.

Bo Franzén¹, Andrey Alexeyenko², Lena Kanter¹, Kristina Viktorsson¹, Jonas Kierkegaard³, Giuseppe Masucci^1,5, Ulf Landegren⁴ and Rolf Lewensohn^1,5

¹ Department of Oncology and Pathology, Cancer Center Karolinska, Karolinska Institutet, 171 76 Stockholm, Sweden

² Department of Microbiology, Tumor and Cell Biology (MTC), Karolinska Institutet, Stockholm, National Bioinformatics Infrastructure Sweden, Science for Life Laboratory, Box 1031, Solna, Sweden

³ BröstCentrum City, Drottninggatan 68, 111 21, Capio S:t Görans Sjukhus, 112 81 Stockholm, Sweden

⁴ Department of Immunology, Genetics and Pathology, Science for Life Laboratory, Uppsala University, Box 815, 751 08 Uppsala, Sweden

⁵ Theme Cancer, Patient area head and neck, lung, and skin, Karolinska University Hospital, Stockholm, Sweden

Introduction

One of the main challenges for therapy of solid tumors is to define methods that allow repeated and minimally invasive sampling with capacity to identify molecular signatures for treatment decisions. This is a requisite for treatment personalization with small molecules and immune therapy where several lines of treatment may be offered sequentially but will need repeated diagnostics. Fine needle aspiration (FNA) sampling is a well-established diagnostic procedure where cells, tissue fragments and/or fluid may be recovered from tumor tissue via a puncture using a very thin needle.

We have here explored a FNA-compatible technology for multiplex molecular profiling that meets requirements for sensitivity and reproducibility: proximity extension assay (PEA) [Assarsson E, et al., (2014) PLoS One 9, e95192]. On breast cancer patients we were able to multiplex assessment of ER, PGR, HER2 and Ki67 as well as many other molecular markers by PEA [Franzén B, et al., (2018) Mol. Oncol. In Press]. We have previously demonstrated that even “leftover” FNA sample material from BC can be analyzed by PEA with high sensitivity, and with results that correlated with routine assessments. We are now extending the work into assaying the microenvironment and immunodynamics in different human tumors.

Material and Methods

For FNA sampling 21-22 Gauge needles are used. Needles are rinsed with ice-cold medium, cells are pelleted, frozen and stored at -80°C. Samples are checked by cytological examination and lysed in RIPA buffer, and total protein concentrations determined. Only 0.5 µg total protein per multiplex panel is used for profiling by PEA.

Results and Discussion

Up to now we have applied proximity extension assays (PEA) for analyses of 167 proteins in FNA samples from patients with breast cancer (BC, n=25) and benign lesions (n=33). We demonstrate that FNA-based PEA profiling allows dividing cancer samples in two main hierarchical clusters that corresponded to some extent to BC subtypes. Our analysis also revealed several proteins that differed between BC and benign lesions (e.g. CA9, GZMB, IL6, VEGFA, CXCL11, PDL1 and PCD1) as well as several chemokines correlating with estrogen receptor (ER) and proliferation marker Ki67 status (e.g. CCL4, CCL8, CCL20, CXCL8, CXCL9 and CXCL17). Finally, we identified three signatures that could predict Ki67 status, ER status and tumor grade, respectively, based on a small subset of proteins which was dominated by chemokines.

Conclusion

Due to the minimally traumatic sampling and clinically important molecular information for therapeutic decisions, FNA methodology is promising for future diagnostics oriented on both immune therapeutics and regular treatments with e.g. small molecules. We now look for collaborations on treatment issues using PDXs including both immunocompetent and regular animal models.

Adam Sierakowiak¹, Petra Hååg¹, Ravi Saini¹, Ana Zovko¹, Vasiliki Arapi¹, Metka Novak¹, Dima Kovalerchick², Micha Ilan³, Shmuel Carmeli², Kristina Viktorsson¹, Rolf Lewensohn^1,4

¹ Karolinska Institutet, Dep. of Oncology and Pathology, Stockholm, Sweden

² Tel Aviv University, School of Chemistry, Raymond and Beverly Sackler Faculty of Exact Sciences, Tel Aviv, Israel

³ Tel Aviv University, Department of Zoology, George S Wise Faculty of Life Sciences, Tel Aviv, Israel

⁴ Theme Cancer, Patient area head and neck, lung, and skin, Karolinska University Hospital, Stockholm, Sweden

Introduction

Multiple anti-tumor pharmaceuticals have been reveal by screening marine organisms for bioactive compounds. Within the FP7 project SPECIAL (SPonge Enzymes and Cells for Innovative AppLications) we screened extracts from the marine sponge Cribrochalina vasculum and two acetylenic compounds (compound ¹⁄₂) which demonstrated anti-tumor activity was revealed (Zovko, Viktorsson et al., Mol Cancer Ther. 2014). We found that these compounds activated cell death and inhibited proliferation of non-small cell lung cancer (NSCLC) cells and of other origin e.g. breast cancer, ovarian carcinoma and small cell lung cancer without causing any effect PBMC and lung fibroblasts.

We also revealed that these compounds blocked insulin like growth factor receptor (IGF-1R) β phosphorylation and by cellular thermal shift assay (CETSA) we demonstrated binding of compound 1 to IGF-1R β (Zovko, Novak et al., Oncotarget. 2016). Importantly, compound 1 was also found to cause degradation of IGF-1R β degradation specifically in NSCLC cells.

Material and Methods

Compound 1-induced cytotoxicity was profiled by MTT cell viability assay in NSCLC or multiple myeloma (MM) cell lines and on patient derived NSCLC cells from pleural effusions (PE). Effects on IGF-1R signaling was analyzed by proximity ligation assay (PLA) and western blotting. Anti-tumor efficacy of compound 1 were analyzed in MM xenografts in SCID mice where treatments were given i.p. twice a week during a three-week period.

Results and Discussion

Primary NSCLC cells with different molecular alterations e.g. mutated EGFR, EML4-ALK fusion and K-RAS mutation were profiled for cytotoxicity to compound 1 and 2. In some but not all samples the compounds caused cell death and this was paralleled with effect on IGF-1R signaling. Compound 1 and 2 were also found to cause cell death in MM cells that are reported to be driven by IGF-1R with and effects on both expression and phosphorylation of IGF-1R β were observed. The in vivo anti-tumor activity of compound 1 was analyzed in SCID mice carrying MM xenografts where either compound 1 (12.5 mg/kg) or the clinically used proteasome inhibitor bortezomib (VELCADE®) (1 mg/kg) were given. Results demonstrated that both compound 1 and bortezomib attenuated tumor growth and prolonged mice survival.

Conclusion

Our results show that acetylenic compounds from marine sponge have anti-tumoral effect in NSCLC and MM. Our findings suggest that these marine-derived pharmaceuticals may allow for targeting IGF-1R driven tumors including NSCLC and indicate IGF-1R β phosphorylation as putative response biomarker.

Ankita Sati Batra¹, Oscar M Rueda¹, Wendy Greenwood¹, Alejandra Bruna¹, Carlos Caldas²

¹ Department of Oncology and Cancer Research UK Cambridge Institute, Li Ka Shing Centre, University of Cambridge, Cambridge CB2 0RE, UK

² Department of Oncology and Cancer Research UK Cambridge Institute, Li Ka Shing Centre, University of Cambridge, Cambridge CB2 0RE, UK; Cambridge Breast Unit, NIHR Cambridge Biomedical Research Centre and Cambridge Experimental Cancer Medicine Centre at Cambridge University Hospitals NHS Foundation Trust, Cambridge CB2 2QQ, UK

Introduction

Patient Derived Tumour Cells provide an indispensable tool for drug discovery providing us with the complexity of the primary tumour while being readily available in bulk to conduct high throughput drug screens. Although several experiments have been conducted, issues regarding quality control have arisen. These issues range from marked differences between technical replicates, minimal growth of cells to low cell viability. In order to combat these concerns, we began assessing each step of the high throughput drug screening process with an aim to improve reproducibility and efficiency.

Material and Methods

Patient Derived Tumour Xenografts (PDTXs) (Bruna et al. 2016, Cell) were dissociated into single cell suspensions (PDTCs) and subjected to high throughput drug screening using the Labcyte ECHO acoustic liquid dispenser. Miltenyi Biotech dissociators and human tumour dissociation kits were used for the semi-automated dissociation. Read out of final cell numbers was done using Cell Titer Glo 3D, Cyquant and Sytox assays.

Results and Discussion

1. The enzymatic dissociation process (Bruna et al. 2016, Cell) was compared to a semi-automated process involving both mechanical and enzymatic dissociation. This led to a marked increase in cell number and viability of cells. It also led to improved reproducibility and less hands-on time.
2. We also compared the utilisation of fresh versus frozen tissue for dissociation. Although frozen xenograft tissue resulted in fewer number of cells, it did not affect the growth characteristics of the cells over a period of 1 week. Use of frozen tissue also eradicated the issue of tumour availability on the day of dissociation, thus maximising the throughput.
3. Several concerns were also brought up regarding the use of Cell Titer Glo 3D (Luminescence based) as an acceptable method of read out of final cell numbers. The assay was compared with more robust but tedious DNA- binding dye assays like Cyquant and Sytox. All three assays showed very high correlation.
4. Over the course of optimisation, we also tested different media conditions and found RPMI1640+B27+growth factors resulted in better growth of cells. The dosing pattern of drugs was also adapted to remove any spatial differences between replicates. Keeping cells on ice while plating and not stacking plates in the incubators also helped in reducing differences between the replicates.

Conclusion

By testing various conditions of dissociation, growth, plating and read out we have optimised our high throughput drug screening protocol to maximise output and efficiency. The suggested steps also reduce variability and leads to more robust data for analysis.

Danilo Fiore¹, Filomena Di Giacomo¹, Jude M Phillip¹, Xujun Wang¹, Olivier Elemento¹, Shahin Rafii¹, Steven M. Horwitz², Leandro Cerchietti¹, David M. Weinstock³ and Giorgio Inghirami¹

¹ Weill Cornell Medicine, New York, NY, US

² Memorial Sloan Kattering Cancer Center, New York, NY, US

³ Dana Farber Cancer Instutute, Boston, MA, US

Introduction

Therapeutic improvements in T-NHL have been hindered by the lack of informative models. Patient-Derived-Xenografts (PDX), recapitulating many of the biological features of primary cancers, represent a powerful platform. Here, we constructed a library of T-NHL PDX to implement drug discoveries and patient-tailored approaches.

Material and Methods

Viable samples from primary T-NHL were implanted in NSG-B2m mice. Seeds from arising tumors were serially transplanted. The lymphoma phenotype was assessed by flow cytometry and IHC. TCR gene rearrangement, whole exome sequencing (WES), and total RNAseq were used to annotate primary and PDX. For in vitro studies, PDX were digested and stromal cells were excluded by panning. High-throughput (HTP) drug screening were performed using a library of 432 compounds (1 μM for 72hrs). Cell titer glo/blue, trypan blue cell count, and annexin-V/7AAD were used as readouts.

Results and Discussion

We implanted 172 T-NHL samples, resulting in the generation of 57 PDX lines. Successfully propagated PDX showed distinct immunophenotypic profiles, which closely mimiced primary samples. TCR gene rearrangement confirmed the clonal correspondence between donor and PDX samples. WES and total RNA-Seq yielded a rich landscape of mutations shared by primary tumors and PDX derivatives. Unique mutation burdens were defined along passages.

Since, Patient Derived Tumor Cells (PDTC) could survive in vitro for a short period of time (<7days), we envisioned two different approaches:

1. Exogenous supplementation of selected cytokines chosen accordingly on tumor transcriptome
2. Co-culture of T-cells with host cells (i.e. E4-Endothelial Cells, E4-EC or Patient-Derived Cancer Associated Fibroblast, PDCAF).

So far we have established 4 stable PDTC lines. Interestingly, we proved that PDCAF could rescued cell death induced by starvation, meanwhile E4-EC only partially ameliorate survival in selected PDX. Notably, host elements rescued cell death induced by selected compunds including Navitoclax®.

A total of 5 PDTC were also used to perform a HTP drug screening, demonstrating that each PDX exhibited a unique drug sensitivity profile, only in part dictated by its genomic landscape. In vivo studies have confirmed the predictive value of HTP screening.

Conclusion

We built a large PDX library from multiple T-NHL entities. Immunophenotyping and genomics showed a close correspondence between primary and correspondent PDX. We predict that our in vitro models will provide novel tools to dissect stromal:tumor cell interactions, modulating lymphoma survival and drug resistance. We anticipate that via HTP screening of PDX we will be able to provide patient drug response fingerprints which will foster the discovery of novel therapeutic compounds and the implementation of personalized medicine approaches.

Thea Kristin Våtsveen^1,2, Ulrika Warpman Berglund³, Baoyan Bai ^1,2, Helge Gad³, Therese Pham³, Idun Dale Rein⁴, Trond Stokke⁴, Erlend B. Smeland^1,2, June H. Myklebust^1,2, Thomas Helleday³, Morten P. Oksvold^1,2

¹ K.G Jebsen Center for B-cell malignancies, Oslo University Hospital (OUS), The Faculty of Medicine, University of Oslo, Norway

² Section for Cancer Immunology, Institute for Cancer Research, The Norwegian Radium Hospital, OUS, Oslo, Norway

³ Dept of Medical Biochemistry and Biophysics, Science for Life Laboratory, Div. Of Translational Medicine and Chemical Biology, Karolinska Institutet, Sweden

⁴ Section for Radiation Biology, OUS, Norway

Introduction

Certain B-cell lymphomas are still considered incurable with standard of care. This call for further development of novel therapeutics and to consider types of targets not previously examined. A novel approach is to manipulate the nucleotide metabolism, and karonudib (TH1579), recently developed to inhibit MTH1 (MutT homologue 1/NUDT1), is a potential new drug for lymphoma treatment. MTH1 belongs to the Nudix phosphohydrolase superfamily which converts oxidized nucleotide triphosphates (e.g. 8-oxo-dGTP and 2-OH-dATP) to the corresponding monophosphate forms. In this way incorporation of oxidized nucleotides into the DNA is avoided. Cancer cells, and particular lymphomas has a high level of MTH1 compared to normal healthy cells. Inhibiting MTH1 will increase the oxidized nucleotides in the cells to a level that makes the cells go into apoptosis. Our aim is to explore the potential of the MTH1 inhibitor karonudib in treatment of B-cell lymphoma.

Material and Methods

B-cell lymphoma cell lines representing most subtypes of B-cell lymphoma were treated with increasing concentrations of karonudib (0.06-1 µM) and viability measured as ATP levels was performed using CellTiterGlo, active caspase-3 was measured and TUNEL staining was performed using flow cytometry. Incorporation of 8-oxo-dGTP was measured by a modified comet assay. Cell cycle arrest was detected by Hoechst staining and flow cytometry (0.5 µM karonudib for all assays). BL-41-luc cells were inoculated subcutaneously in NSG mice to assess anti tumor activity in vivo. Karonudib (90 mg/kg) or vehicle was given by oral gavage b.i.d, three times a week. Tumor growth was monitored by In vivo imaging system. Diffuse Large B-cell lymphoma-ABC-subtype PDX model study is ongoing with same treatment strategy as for BL-41-luc. Cellular thermal shift assay (CETSA) was done on tumor material from the xenograft study.

Results and Discussion

Karonudib induces a strong anti-proliferative response in B-cell lymphoma cell lines, but not in healthy donor B-cells. An induction of apoptosis was seen by active caspase 3 and TUNEL staining after 24 h treatment. All cell lines gave a clear Metaphase arrest, whereas a TP53 WT cell line was arrests also in confirming that karonudib induces apoptosis independent of TP53 mutational status. The in vivo data show a significant decrease in both tumor size in karonudib treated mice and compared to vehicle. The modified comet assay verified incorporation of 8-oxo-dGTP in the cells treated with karonudib and CETSA showed that karonudib was binding and stabilizing MTH1 in cells from the in vivo study.

Conclusion

Karonudib shows potent anti-tumor response, coinciding with incorporation of 8-oxo-dGTP, cell cycle arrest and induction of apoptosis in B-cell lymphoma cells, making it a promising drug for B-cell lymphoma therapy.

Somaieh Hedayat, Georgios Vlachogiannis, Andrea Lampis, Khurum Khan, David Cunningham, Matteo Fassan, Ruwaida Begum, Marta Schirripa, Fotios Loupakis & Nicola Valeri

Institute of Cancer Resarch, London, United Kingdom

Introduction

PDOs represent robust pre-clinical models to recapitulate the complex three dimensional (3D) organization of cancer and potentially predict clinical outcomes. Regorafenib, a multi-tyrosine kinase inhibitor, has demonstrated efficacy in chemo-refractory mCRC patients by inhibiting tumour vasculature. Limited clinical benefit in unselected patient populations and early drug resistance highlights the unmet need for better patient selection and identification of mechanisms of resistance.

Material and Methods

We ran a translational phase II trial of regorafenib in chemo-refractory mCRC patients with biopsiable metastases. Tissue biopsies were obtained at baseline, after 2 months of treatment, and at disease progression. PDOs, PDO xenotransplants and PDO co-cultures with cancer associated fibroblasts (CAFs) and endothelial cells (EC) system were generated to study primary and acquired resistance to regorafenib. PDOs were developed from imaging-guided biopsies of liver and implanted orthotopically in the livers of NSG mice and treated with regorafenib.

Likewise, PDOs were cultured in 3D spheroid mono and co-cultures system to recapitulate patients’ tumor microenvironment. Response to regorafenib was tested by luminescence, functional MRI and histopathology scoring in vivo. Migration, viability, tumor volume and angiogenesis assays were performed ex vivo.

Results and Discussion

PDOs were successfully developed and retained genomic and transcriptomic features of parental biopsies over serial passages. PDOs implanted and metastasized within the liver of NSG mice: vascular density showed significant drop in CD-31 after regorafenib treatment in animals from responder patients (p=0.03, n=6) while no significant changes were observed in those from nonresponders (n=10). Similar findings were seen comparing response to regorafenib in pre- and post- treatment PDOs from a long-term responder patient. Co-culture of sequential PDOs obtained before and after regorafenib treatment recapitulated the micro-architecture and response to regorafenib observed in vivo. Use of conditioned media from responder and non-responder PDOs affected response to regorafenib.

Conclusion

PDOs recapitulate pathological, molecular and radiological features of matching metastases upon regorafenib treatment. In vivo and ex vivo co-culture data suggest vessel co-option and tumour microenvironment as a mechanism of acquired resistance to regorafenib. PDO co-cultures might resemble the metastatic niche predicting response to anti-cancer treatments to inform clinical decisions.

BT Preca¹, B Hamelin¹, M Diepenbruck¹, S Iftikhar¹, M Ritter¹, J Zeindler¹, R Mechera¹, R Okamoto¹, M Abanto¹, M Vetter¹, P Liberali³, K Volkmann³, T Vlajnic², WP Weber¹ and M Bentries-Alj¹

¹ Department of Biomedicine and Department of Surgery, University of Basel, University Hospital Basel, Basel, Switzerland

² Institute of Pathology, University Hospital Basel, Basel, Switzerland

³ Friedrich Miescher Institute for Biomedical Research; University Hospital Basel, Basel, Switzerland

Introduction

Although findings from the last decades have improved our understanding of some molecular and cellular mechanisms underlying cancer, we still lack effective therapies for the most aggressive cancers, and most oncology patients die because of drug resistant metastases. Identifying means to treat drug resistant metastases is urgently needed.

Material and Methods

To end this stalemate we are developing a translational platform to test drugs and combinations of drugs on patient breast cancer samples ex vivo in a bedside-to-bench- to-bedside manner. Based on patient derived 3D organoid cultures the platform will use automated high-throughput fluorescent microscopy and high-end single cell imaging to identify patient specific drug sensitivity.

Results and Discussion

Our ultimate goal is to beat refractory cancers and to provide treatment options for cancers that develop resistance to initial standard therapy.

Kristopher Frese¹, Kathryn Simpson¹, Alice Lallo¹, Melanie Galvin¹, Fiona Blackhall², Mark J O’Connor³, and Caroline Dive¹

¹ Cancer Research UK Manchester Institute, Manchester, UK

² Institute of Cancer Sciences, University of Manchester, Manchester, UK

³ AstraZeneca, Cambridge, UK

Introduction

The clinical management of small cell lung cancer (SCLC) has remained largely unchanged for over 30 years due in part to the lack of preclinical model systems that predict clinical outcomes. Despite numerous clinical trials, SCLC prognosis remains dismal with a 5 year survival rate of less than 5%. We have recently reported the generation of SCLC circulating tumour cell derived xenograft (CDX) mouse models that accurately reflect the histopathology, genetics, and biology of the donor patient.

Material and Methods

CTC’s were enriched from SCLC patient blood samples via RosetteSep and implanted into immunodeficient mice. Resulting CDX models were genetically, phenotypically, and functionally characterised. For pharmacology experiments, mice were randomised into cohorts, treated for 3 weeks, and tumour burden was regularly measured to assess therapeutic efficacy.

Results and Discussion

Here we report the generation of a burgeoning library of 45 SCLC CDX models that reflect the inter- and intra-tumoural heterogeneity of this disease. In addition to generating CDX models from both extensive stage chemosensitive and chemorefractory treatment-naive patients, we have generated seven paired models generated at different timepoints over the course of disease progression, as well as models representing limited stage disease and NSCLC-to-SCLC transdifferentiation. Utilizing these models we have begun to explore novel drug combinations including DNA damage repair and cell cycle checkpoint inhibitors. Toward this end we have examined the efficacy of the PARP inhibitor olaparib alone or in combination with the Wee1 kinase inhibitor AZD1775 in ten phenotypically-distinct SCLC CDX models in vivo and/or ex vivo.

Despite heterogeneous depth and duration of response to olaparib/AZD1775, this efficacy of this combination consistently exceeded that of cisplatin/etoposide including a cure in one ‘super-responder’ model. Molecular analyses revealed that defects in homologous recombination repair and oncogenes that induce replication stress predisposed this ‘super-responder’ to combined olaparib/AZD1775 sensitivity, and we are now extending these analyses to our broader panel of models to identify recurrent predictive biomarkers. These data suggest that olaparib/AZD1775 may prove efficacious in the proper setting and clinical trials investigating this combination in are currently underway.

Conclusion

CDX models represent a unique opportunity to sample the inter-tumoural heterogeneity of SCLC and can be exploited to examine the biology of this recalcitrant disease.

Rebecca Marlow¹, Luned Badder¹, Eleanor Knight², Stephen Pettitt², Bram Herpers³, Daniel Larcombe-Young¹, Erika Francesch-Domenech¹, Magdalena Lomzik-Borowik¹, Amelia Rushton², Daniela Novo², Leo Price³, Christopher Lord², Andrew Tutt^1,2

¹ Breast Cancer Now Unit, King’s College London, Division of Cancer Studies, London, SE1 9RT

² Breast Cancer Now Toby Robins Research Centre, The Institute of Cancer Research, London, SW3 6JB

³ OcellO B.V, Leiden BioPartner Center, 2333CH, Leiden, The Netherlands

It is widely recognised that Patient-derived Xenografts (PDXs) are gold-standard in vivo models of breast cancer, capable of faithfully recapitulating tumour biology. High throughput drug screening methods currently rely on in vitro clonal two-dimensional (2D) cell lines that lack such pathophysiological relevance to parental tumours, limiting their ability to predict patient responses to therapies. There is a need for in vitro models to bridge the gap between such model systems to reduce and refine the number of animal experiments for pre-clinical drug testing, whilst retaining accurate readouts of tumour responses.

Three-dimensional (3D) organoid models have shown promise as disease-relevant models, capable of providing a comparable genetic and phenotypic profile to parental tumours. However, robust protocols for organoid derivation are somewhat limited for breast cancer subtypes, including Triple negative breast cancers (TNBCs), which lack targeted therapies in the clinic. Furthermore, it is yet to be ascertained whether organoid treatment responses correlate with current gold standard PDX models of breast cancer.

In partnership with OcellO (Leiden, Netherlands), we have generated a cohort of patient-derived organoids from both our defined panel of TNBC PDX models, as well as directly from patient tissue. To systematically evaluate whether our models can recapitulate drug responses of companion in vivo PDX models, we have assessed and will present standard of care drug responses of matched organoid and in vivo pairs. This includes characterising BRCA-1 mutant models based on their sensitivities to Poly(ADP-ribose) polymerase (PARP) inhibitors. Thus far, we have found differential sensitivities to PARP inhibition within our BRCA-mutant- derived organoid models, reflective of comparative in vivo studies. We aim to interrogate drug sensitivities of our models further using high throughput phenotypic screening, which will enable us to rapidly test hypothesis-driven drug/target combinations, to establish whether organoid readouts can predict responses in the clinic.

Alexander Balhuizen¹, Christophe Côme¹, Anne-Katrine Frank¹, Kristian Helin¹, Kyoung Jae Won¹, Kim Theilgaard-Mönch¹, Kirsten Grønbæk^1,2, Kirster Wennerberg¹ and Bo Porse¹

¹ Biotech Research and Innovation Centre, Program of Translational Hematology, Copenhagen Biocenter, Copenhagen, Denmark

² Epigenomlaboratoriet, Rigshospitalet Dept. 3733, Bartholin Institute, Copenhagen Biocenter, Denmark

Introduction

Our program of translational hematology (PTH) aims to improve the prognosis for patients with acute myeloid leukemia (AML) and myelodysplastic syndrome (MDS). Indeed, the survival rates for AML/MDS have only increased marginally over the last decades. This shortcoming results from the current therapeutic regimens’ inefficiency to eliminate the resistant cancer stem cells (CSCs). Hence, novel therapies targeting CSCs are predicted to improve patient outcome. To develop such therapies, new patient-derived xenograft (PDX) mouse models are needed, as the current leukemic models do not recapitulate the primary cancer cells’ microenvironment and heterogeneity. The goal of the PTH PDX-platform is to generate a cohort of expanded samples from 70 AML and 70 MDS donors of different WHO-defined subclasses during the next 3 years and to characterize their CSC.

Material and Methods

To accomplish this, we are currently determining suitable methods to:

• Generate humanized bone marrow like scaffolds (hBMLS) in mice to improve the limited engraftment rate of low aggressive AML or MDS. hBMLS are generated by subcutaneously implanting human bone marrow mesenchymal stem cells (BM-MSCs) into the flanks of NSG mice to create a humanized BM niche.
• Select the optimal murine strain for hBMLS as these scaffolds have only been created in NSG mice, which requires myeloablation by irradiation before injecting AML cells. Therefore, we will test the NBSGW strain (NSG with c-kit^(W41) mutation, which allows for hematopoietic stem cell engraftment without irradiation) as an alternative recipient strain
• Define the origin of the CSCs: AML stem cells are generally defined as CD34+CD38- cells. However, 25% of human AML tumors consist mainly of CD34- cells, while CD34 is lowly expressed in bulk MDS cells. Therefore, based on CD34/CD38 expression, we will transplant the four lineage-negative subpopulations (CD34+/-CD38+/-) to identify fractions enriched in CSC activity, and characterize them by next-generation sequencing.

Results and Discussion

Our preliminary results confirm that the efficiency of BM-MSC to generate hBMLS is donor dependent. On the other hand, CD34 and CD38 expression is also variable between patients. Moreover, all AML/MDS patient samples tested so far present a T cell fraction of 4-10% indicating that T cell depletion by flow or magnetic sorting is necessary to avoid graft versus host reactions.

Conclusion

The AML/MDS PDXs will generate expanded primary AML/MDS patient populations useful for CSC characterization and pre-clinical evaluation of novel therapies identified by a drug screen platform within PTH. Moreover, hBMLS will allow us to study the impact of tumor microenvironment in these tumors. Finally, to our knowledge, the EuroPDX platform does not contain hematological PDX; therefore, our samples could bring additional models to the scientific community.

Magdalena Cybulska^1,2, Aleksandra Grochowska^1,2, Michał Kopczyński¹, Krzysztof Goryca¹, Agnieszka Paziewska², Jakub Karczmarski¹, Michalina Dąbrowska¹, Marek Woszczyński¹, Michał Mikula¹, Jerzy Ostrowski^1,2

¹ Maria Sklodowska-Curie Memorial Cancer Center and Institute of Oncology, Warsaw, Poland

² Medical Center for Postgraduate Education, Warsaw, Poland

Introduction

After 4 years of establishing and development, we present the first Polish PDX platform for preclinical studies with anticancer therapeutics. The panel of PDXs consists of 33 colorectal, 4 ovarian, 3 bladder and 2 stomach cancers along with 3 melanomas and a leiomyosarcoma. Thanks to access to next-generation DNA and RNA sequencing, histopathological and immunohistochemical evaluation together with well-equipped animal facility, we are able to create and assess a growing number of models and conduct preclinical experiments. So far all colorectal cancer (CRC) PDXs were subjected to thorough molecular characterization.

Material and Methods

Intrasurgically removed cancerous tissue from treatment naïve patients was subcutaneously grafted to nude or NSG mouse. Both primary tumors and PDXs were histologically evaluated and subjected to molecular survey using RNA-Seq application and sequencing of 409-cancer related genes on Ion Proton sequencer.

Results and Discussion

Histological assessment of CRC samples confirmed that PDX grading, stromal components, inflammation, and budding were consistent with those of the primary tumors. DNA sequencing identified an average of 0.14 variants per gene per sample. The percentage of mutated variants in PDXs increased with successive passages, indicating a decrease in clonal heterogeneity.

Gene Ontology analyses of 4180 differentially expressed transcripts (adj. p-value < 0.05) revealed over-representation of genes involved in cell division and catabolic processes among the transcripts upregulated in PDXs; downregulated transcripts were associated with GO terms related to extracellular matrix organization, immune responses, and angiogenesis. Neither a transcriptome-based consensus molecular subtype (CMS) classifier nor three other predictors reliably matched PDX molecular subtypes with those of the primary tumors. In sum, both genetic and transcriptomic profiles differed between donor tumors and PDXs, likely as a consequence of sub-clonal evolution at the early phase of xenograft development, making molecular stratification of PDXs challenging.

Conclusion

Our continuously growing collection of PDX models constitutes a valuable tool for preclinical and translational research. We are willing to join the consortium and contribute to the common goals.

Kralova Viziova P., Indrova M., Prochazka J., Sedlacek R.

Institute of Molecular Genetics of the ASCR, v. v. i., Prague, Czech Republic

Introduction

Standard procedures for delivery of carcinoma cells into the mouse mammary fat pad comprise blind injection through the skin or surgery, where the wound has to be sew (Kocatürk and Versteeg, 2015). In this work two new methods were developed for carcinoma cells delivery into the mouse mammary fat pad.

Material and Methods

The first method comprises mini invasive surgery technic under the general anesthesia. Two doses of mouse mammary gland carcinoma cells 4T1 labelled with Red-Fluc reporter (Perkin-Elmer) were delivered into the mammary fat pad of NSG mice (5x104 cells in 0.02 ml, left and right mammary fat pads). The caused wounds were less than 2.5 mm and therefore the suture was not necessary.

The second method was performed with just strong light and insulin syringe to deliver human carcinoma cells 4T1 (4T1-Red-FLuc) (5x104 cells in 0.02 ml, left and right mammary fat pads) through the skin into mammary fat pad. Shaved or bold white skin lean female mice kept under the optimal diet have to be used for this study. The growth of tumor transplants was checked weakly by luminescence and the sections of tumor and normal tissue were subjected to histology staining.

Results and Discussion

All mice developed tumor in mammary fat pad during 2 weeks. The first method is less time consuming, less expensive and less invasive than standard surgery technic, therefore the better welfare of mice is accomplished. Using a strong light to visualise the underskin structures apeared to be esential to distinguish mammary gland fat pad from vessels and subcutaneus fat and administer the tumor cells into precise position.

Conclusion

After application any of these methods, mice recovered very quickly without any unexpected side effects and the success rate for both methods was 100%. Therefore, these techniques have potential to be recognized as standard methods for cancer cells administration into mouse mammary fat pad.

Roni Oren, Alon Harmelin

Department of Veterinary Resources, Weizmann Institute of Science, Rehovot, Israel

In vivo mice model and particularly Patient-derived xenograft (PDX) models have recently emerged as a highly desirable platform for cancer research for basic and for pre-clinical studies. However obtaining PDX samples, maintaining them in mice and sample storage are complex process that require in depth planning and effort for infrastructure, funding and time. In the Weizmann institute we recently set to establish a new facility for in vivo PDX tumor model, combined with a tumor bio-bank. Establishing the new unit confronted us with multiple challenges from different perspectives.

The first challenge that needed to be addressed was sample collection. Weizmann scientists collaborate with Israeli hospitals and clinicians to collect samples with the proper authorization (Helsinki and IRB). Another pathway to obtain PDX samples is via scientific collaboration and from commercial companies. The PDX collection is gradually growing by additions of samples designated for specific defined projects.

For tracking and managing the PDX bearing mice, all experimental procedures and frozen samples we developed an internal web based tool named “Weizmann PDX”. All mice are tagged with a unique identifier RF chip (implanted subcutaneous). Tubes, Histology cassettes and slides are labeled with a unique barcode and tracked. Experimental procedure and histological H&E images are also recorded using the software. The developed tool allows for easy visualization of sample lineage, relevant clinical data, and basic aspects of analysis for tumor growth rate in volume charts.

Sample storage is a crucial part of tumor bio-bank. In order to maximize our storage capabilities we purchased a fully automated biobank at the constant temperature of -80ᵒC (Liconic). The biobank was designed to contain ~150,000 samples of 5 different tube sizes (ranging from 0.3 ml to 4.5 ml volume) and include different sample types (flash frozen and cryo preserved tumor samples, DNA, patient serum etc.). A backup for the PDX viable tissue is stored in liquid nitrogen. Managing samples in the biobank required the development of a management tool to allow researchers to deposit and collect samples in a controlled manner.

The tool was designed to fully integrate with the controlling system of the robotic freezer, thus allowing us to perform all necessary tasks via the platform we created. Each scientist only have access to samples from his research group. Tube barcode is validated entering the freezer and compared to existing record. The scientist can also select specific tubes to be exported upon request. Planning and executing the establishment of a new PDX and biobank facility is an ongoing challenging mission. In the future we would like to expand the services that we can offer for our researchers to include STR quality control, generation of cell lines, genomic analysis and humanized mice models.

Krausert Sonja^1,2, Brabetz Sebastian^1,2, Mack Norman^1,2, Schwalm Benjamin^1,2, Henkel Luisa², Rauscher Benedikt², Boutros Michael², Pfister Stefan^1,2, Kool Marcel^1,2

¹ Hopp-Children’s Cancer Center at the NCT Heidelberg (KiTZ), Heidelberg, Germany

² German Cancer Research Center, Heidelberg, Germany

Introduction

MYC and MYCN are transcription factors and known as proto-oncogenes. In 30% of human cancers, MYC genes are overexpressed and related to aggressive growth and poor clinical outcome. However, MYC also plays a central role in normal cells, which leads, in addition to other facts as the aggressive nature of MYC-related tumors and the low druggability of MYC, to an urgent need for development of better strategies to target MYC in tumor cells.

Pediatric brain tumors frequently harbour MYC- or MYCN-amplification. Examples for which current treatments are unsuccessful are for instance the medulloblastoma (MB) subgroups SHH (MYCN amplified), Group 3 (MYC amplified) and pediatric glioblastomas (MYCN amplified). To gather more information and insights in the biology of MYC(N)-driven pediatric brain tumors, we will use an in vivo CRISPR screen, which will also help to identify potential novel drug targets for therapy. A pooled sgRNA library will be transduced into cells of patient-derived xenografts (PDX) which will be transplanted orthotopically in mice. After tumor growth, genes necessary for survival of tumor cells can be analysed by next generation sequencing (NGS).

Material and Methods

For the screen we will use PDX cells growing in vivo and CRISPR technology. PDX cells of MYC(N)-driven tumors will in vitro be lentivirally transduced with TLCV2, which expresses Cas9 after induction with Doxycycline, and a pooled sgRNA library cloned into it. Afterwards, transduced cells will be selected with Puromycin and afterwards orthotopically injected in mice. Then, expression of Cas9 will be induced with Doxycycline. When solid tumors are formed, they will be dissected and tumor DNA will be isolated. With NGS, the quantity of each barcoded sgRNA can be analysed. We will screen for the “missing” barcodes as this means that the tumor cell couldn’t survive without the particular knocked-out gene.

Results and Discussion

For first experiments, a control sgRNA was cloned into TLCV2 and the transduction of two different PDX models, one for Group 3 MB and one for SHH, was tested. In repeating experiments, it was possible to confirm transduction with an MOI of 0.3 to ensure integration of just one plasmid in each cell. In addition, different libraries have been prepared including a library with subgroup specific lineage markers of MB and a library containing genes known as potential drug targets in pediatric tumors.

Conclusion

Our approach of an in vivo CRISPR screen is new and challenging but has also advantages. Besides less work for in vitro culturing of cells, a more sophisticated model will be used. Since PDX cells and in vivo growth mirror tumor growth and the situation in a patient more reliably, we hope that this approach will give us more information about MYC(N)-driven pediatric brain tumors, which may help to improve therapy for these patients.

Verena Kiver^1,2, Annika Wulf-Goldenberg⁶, Phillipp Jurmeister^1,5, Caroline Schweiger^1,3,4, Olga Gorea^1,2, Jens-Uwe Blohmer², Carsten Denkert^3,4,5, Jens Hoffmann⁶, Ulrich Keilholz^1,3,4

¹ Charité Comprehensive Cancer Center, Berlin, Germany

² Institute of Gynecology, Charité, Berlin, Germany

³ German Cancer Consortium (DKTK), Berlin, Germany

⁴ German Cancer Research Center (DKFZ), Heidelberg, Germany

⁵ Institute for Pathology, Charite, Berlin, Germany

⁶ Experimental Pharmacology & Oncology Berlin GmbH-Berlin-Buch, Berlin, Germany

Introduction

Preclinical models for breast cancer, especially hormone receptor positive (HR+), are difficult to establish. We outline our experience in establishing breast cancer patient derived xenografts (PDX).

Material and Methods

Since May 2017 samples were collected from treatment naïve or treatment-refractory breast cancer patients in the Department of Gynecology with Breast Center, Charité Campus Mitte, Berlin, Germany. The collective of samples comes from a variety of patients: TNBC, HR+ and Her2+ tumors, primary disease, recurrence or metastasis were included. Via biopsy or during surgery we collected ̴3 mm diameter tissue samples. To reduce ischemia time the samples were directly given into a transport medium and implanted into female immunodeficient NOG mice the same day. If the original tumor was hormone receptor positive, the mice received estradiol supplementation. To prove tumor stability during engraftment molecular profiling and immunhistopathology are performed. A drug panel, adapted to the histology of the tumor and the treatment regimen of the individual patient, is tested on established PDX.

Results and Discussion

Currently 34 samples have been processed. Out of 14 TNBC samples, eight were eliminated. One is engrafted and currently being tested with systemic therapy, two are in passage 3 (p3) and three are in p1. HR+, Her2 - breast cancer samples were taken from 19 patients. Ten are eliminated; one is in p2, one in p3, one in p4 and six in p1. The samples in p1 are mostly recently taken samples. One Her2+, HR- sample had to be eliminated. After altering collection and processing, we find that we could increase the initially low engraftment rate from 31% up to 60%.

Ki-67, HR status or recurrent vs primary vs metastatic tumor tissue have not shown any differences in engraftment rate so far. Immunohistopathology, molecular genetic testing and drug testing, adapted to the growth rate of the models, are currently underway.

Conclusion

From a diverse range of breast cancer subtypes, we are establishing PDX models. The familiar issues of growing HR+ breast cancer preclinical models was at least in part overcome. Slow growth rates can be expected from breast cancer samples. In addition, we are further improving our methods to increase the take-rate. Establishing a preclinical model for breast cancer is invaluable to further drug development and personalized treatment options.

Sebastian Brabetz^1,2, Susanne N. Gröbner^1,2, Till Milde^1,2,3, Sarah E. Leary⁴, Frank Braun⁵, Xiao-Nan Li^5,6, Robert J. Wechsler-Reya⁷, James M. Olson⁴, Stefan M. Pfister^1,2,3, and Marcel Kool^1,2

¹ Hopp Children’s Cancer Center at the NCT Heidelberg (KiTZ), Heidelberg, Germany

² German Cancer Research Center (DKFZ) and German Cancer Consortium (DKTK), Heidelberg, Germany

³ Center for Individualized Pediatric Oncology (ZIPO) and Pediatric Brain Tumors, Department of Pediatric Oncology, University Hospital and National Center for Tumor Diseases (NCT), Heidelberg, Germany

⁴ Fred Hutchinson Cancer Research Center and Seattle Children's Hospital, Seattle, WA, USA

⁵ Baylor College of Medicine, Houston, TX, USA

⁶ Ann & Robert H. Lurie Children's Hospital of Chicago and Northwestern University Feinberg School of Medicine, Chicago, IL, USA

⁷ Sanford Burnham Prebys Medical Discovery Institute, La Jolla, CA, USA

Introduction

Solid tumours of the nervous system are the most common childhood cancers after leukaemia. Even though we may be able to cure more and more patients, the intensive treatments cause survivors to suffer from severe sequelae. Therefore, new treatment strategies are urgently needed. Patient-derived orthotopic xenograft (PDOX) models are an excellent platform for biomarker and preclinical drug development. However, the rarity of paediatric brain tumours hindered the generation of a large collection of PDOX models covering the wide spectrum of the many different types of brain tumours, and makes it so far impossible to systematically assess efficacy of therapeutic options in a preclinical setting.

Material and Methods

PDOX models were established and maintained at multiple centers. PDOX models and matching patient tumors (if available) were analyzed by whole-exome and low-coverage whole-genome sequencing, as well as DNA methylation and gene expression profiling at the German Cancer Research Center (DKFZ).

Results and Discussion

Molecular subtyping by methylation profiling of in total 130 PDOX lines identified 62 medulloblastomas (MB) (4 WNT, 21 SHH, 24 Group 3, 13 Group 4), 24 high-grade gliomas (6 K27, 1 G34, 11 pedRTK1, 2 pedRTK2, 4 MYCN), 17 ependymomas (EPN) (11 PFA, 6 RELA), 11 atypical teratoid/rhabdoid tumors (8 SHH, 2 MYC, 1 TYR) and 26 models of other rare brain tumor entities. Although we are able to model 22 different molecular subgroups of pediatric brain tumors, there is a bias for more aggressive tumor subtypes and, consequently, we did not identify model systems yet for all tumor subgroups, such as EPN of the YAP subgroup.

In depth-molecular analysis within subgroups also identified a strong overrepresentation of the most aggressive tumors within subgroups, such TP53-mutant SHH MB. Comparisons of PDOX to their patient tumors showed high concordance on all levels of genomic data, but also continuous evolution. In pilot experiments, we confirmed sensitivity to targeted therapeutics in vivo based on tumor classification and the presence of specific biomarkers. In order to make the PDOX models and their associated molecular data available to the scientific community, we developed the interactive “PDX Explorer” within the online platform R2 to allow scientists to browse through the molecular data.

Conclusion

Our molecular characterization will allow researchers all over the world to find the correct models for specific scientific questions. The PDOX models generated and characterized thus far faithfully recapitulate the tumors from which they were derived from, but keep on continuously evolving over time. The bias for more aggressive tumor types within PDOX cohorts is a chance to study these deadly tumors better, but must also be taken into account when extrapolating results from the preclinical platform to clinical trials.

Alejandra Bruna, Wendy Greenwood, Martin O’Reilly, Elham Shirazi, Lisa Young, Steven Kupczak, Yi Cheng, Alasdair Russell, Oscar Rueda, Jean Abraham, Carlos Caldas

Cancer Research UK, Cambridge Institute, Cambridge, UK

Cancer’s heterogeneity and in particular cell functional heterogeneity has been historically underestimated and over-simplified, partly explaining the very disappointing overall survival benefits from new oncological therapeutic regimes. Heterogeneity of cancer needs to be adequately captured in pre-clinical models. With this in mind we recently published the derivation and validation of one of the largest and most extensively molecularly annotated breast cancer patient derived tumour xenograft (PDTX) collection.

Our work showed PDTXs preserved upon initial engraftment and through passaging in the mouse the morphological and molecular characteristics of the originating tumour. We further optimized the generation of short-term cultures of PDTX-cells (PDTCs) and observed the intra-tumour genomic clonal architecture present in the originating breast cancers was also mostly preserved in PDTXs and in PDTCs.

We next developed a robust, reproducible and selective high throughput drug-screening platform and showed most of the ex vivo drug responses were validated in vivo supporting this highly annotated PDTX/PDTC platform as a powerful resource for pre-clinical breast cancer pharmacogenomic studies, including identification of genomic biomarkers of response or resistance.

Our current focus is to:

1. Use these improved tools in drug discovery to unravel drug response-patient associations
2. Deepen our understanding of the genomic and functional mechanisms underlying cancer’s evolutionary processes
3. Explore the use of PDTXs in clinical decision making

Overall, we believe the use of tools that capture cancer’s most fundamental feature, heterogeneity, will help to get deeper genomic and functional understanding of tumour evolutionary processes and to better understand the complexity associated between cancer phenotypes and drug responses in an unprecedented manner.

Linda Xue, Annie An, Sheng Guo, Davy Ouyang, Henry Li

Crown Bioscience, San Diego, CA, USA

Introduction

Cancers are diverse genetic and immunological abnormalities, rendering a given treatment only effective for a small subset of patients. Discovery of therapeutics tailored to certain patient populations are thus essential. First, genetic heterogeneity is largely determined by tumor cells (e.g. oncogenic driver mutations), the foundation of targeted therapy. Second, heterogeneous tumor microenvironment (TME), e.g. TILs, plays critical role in cancer pathogenesis, prognosis and response to immuno-oncology (I/O) therapy. Targeting either or both is the future of cancer therapy. Patient-derived-xenograft (PDX) mimics patient (e.g. genetically) and is thus particularly powerful in test specific target therapy (hypothesis testing), and also, in discovery of predictive biomarkers, or hypothesis generation, if libraries of diverse PDXs were used in a population-based study. PDX has yet to be used for I/O research for lack of adequate immunity (Pharmacol Ther 2017;173: 34) and investigating TME-specific components is rather challenging for the difficulty to separate stroma from tumor cells physically or in silico.

Material and Methods

1. We recently created a large library of PDXs covering the most major cancer types and systematically genomic-profile them (> 2000x)
2. A unique acute myeloid leukemia (AML) PDX with IDH2 mutation was used to test a novel IDH2-mutant targeting AG-221, in collaboration with Agio Pharmaceuticals (Cancer discovery 2017;7: 478)
3. We performed mouse clinical trials using large cohorts of PDXs derived from patients of NSCLC, CRC and gastro-esophageal carcinoma to discover the gene signature predictive of response to cetuximab using machine learning
4. We separated the human and mouse contents of PDX transcriptomes in silico and aligned transcriptome reads to human and mouse genomes to identify both tumor and TME components respectively. We also use co-regulation analysis to identify inter-species interactions, or tumor-TME interaction.

Results and Discussion

1. We demonstrated for the first time that AG221’s antitumor activity using AM7577, proof of concept (POC) of IDH2 targeting of IHD2-mutated AML
2. The statistical analysis (machine learning) of large number of datasets of cetuximab mouse clinical trials together with the corresponding transcriptome sequences led to the gene signatures of predictive of specific cancer type responses to cetuximab
3. Our analysis identified different TME components of both mouse and human origins with variation among cancer types. We also identified the tumor-TME interactions likely playing roles in tumor growth, since their numbers are reversely correlated to the transplantation take-rate and negatively correlated to the presence of KRAS mutations

Conclusion

1. POC of AG221 in AM7577 ultimately guided clinical testing and supported accelerated approval of first-in-class Enasidenib
2. Predictive signature of cetuximab may eventually guide clinical testing and become companion diagnostics for treating these tumors
3. The identified tumor-TME interaction may potentially lead to the discovery of novel TME targeted drug intervention, including I/O

Vito Rebecca¹, Min Xiao¹, Katherine Nathanson², Meenhard Herlyn¹

¹ The Wistar Institute, Philadelphia, USA

² The University of Philadelphia, USA

Introduction

The therapeutic landscape of metastatic melanoma continues to evolve at a rapid pace. In parallel, the identification of novel therapeutic resistance mechanism continues to illuminate the breadth of melanoma biology and aggressiveness. The importance of tumour heterogeneity, plasticity, and the microenvironment is paramount to the escape of melanoma cells to existing targeted- and immune-based therapies.

Material and Methods

To facilitate further advancements in targeting therapy resistance through pre-clinical in vivo modelling, we have established >500 patient-derived xenografts (PDX) from 384 patients constituting the full spectrum of clinical, therapeutic, mutational, and biological heterogeneity of melanoma. The PDX repository has been extensively characterized using targeted sequencing of 108 key cancer genes and with reverse phase protein array (RPPA) analyses of >300 total- and phospho-proteins covering key cancer pathways ad processes.

Results and Discussion

Our PDX repository includes samples from treatment naïve, BRAFi, responder/resistant, BRAFi/MEKi responder/resistance, and immune checkpoint blockade responder/resistant patients. Our PDX repository also includes uveal, mucosal and acral subtypes, in addition to cutaneous.

Conclusion

We are actively performing pre-clinical trials to serve as proof-of-principle how the deep characterization and analysis of our PDX repository will facilitate the development of biomarkers, optimization of personalized therapy, and the generation of effective therapeutic strategies for rare genetic subgroups.

Dobrolecki LE^2,5, Wang T^2,5, Sallas CM^2,5, Lewis AN^2,5, Moore E², Hilsenbeck SG^2,3,5, Lewis MT^1,2,5

¹ Depts. of Molecular and Cellular Biology and Radiology

² Lester and Sue Smith Breast Center

³ Dept. of Medicine

⁴ Dept. of Molecular and Human Genetics

⁵ Dan L. Duncan Cancer Center. Baylor College of Medicine, One Baylor Plaza, Houston TX 77030, USA

Introduction

Clinically, breast cancers are divided into three groups: those that express the estrogen hormone receptor (ER+) (which typically also express the progesterone hormone receptor (PR+)), those that are amplified or overexpress the ErbB2 (HER2) oncogene (HER2+), and those that express none of these three markers (termed “triple negative” breast cancer (TNBC)).

Unlike ER+ and HER2+ breast cancers, there are currently no targeted therapies against TNBC. Treatment of TNBC entails surgery coupled with radio- or chemotherapy, or both. The most commonly used chemotherapies are Taxanes (e.g. Docetaxel, Paclitaxel) and more recently, platinum-based agents (e.g. Cisplatin, Carboplatin). However, other than BRCA1/2 mutation status correlating with increased efficacy of platinum-based agents, there are currently no clinically useful predictors of differential treatment response among these commonly used chemotherapeutics.

We hypothesize that a clinically useful molecular predictor of differential chemotherapy response can be developed using human breast cancer patient-derived xenografts as the discovery platform.

Material and Methods

We are testing this hypothesis in a series of “animal clinical trials” using a collection of 40+ PDX treated with docetaxel or carboplatin as single agents, and in selected models docetaxel plus carboplatin. We are also using a number of targeted agents both alone and in combination with one or more chemotherapies. Molecular correlates of response versus growth of untreated controls are being identified in DNAseq, RNAseq and proteomic data, and are being validated in a separate cohort of PDX, as well as in clinical datasets.

Results and Discussion

We have conducted both a single cycle and several multi-cycle chemotherapy trials and extend our previous analysis showing that treatment responses observed clinically are recapitulated in the PDX models, thereby establishing clinical relevance of observed responses.

We demonstrate that PDXs show four patterns of response:

1. Resistance to both agents
2. Response to both agents
3. Response to carboplatin only
4. Response to docetaxel only

We are using this distribution to develop distinct molecular signatures that delineate those PDX that respond to one given treatment from those that do not, and to identify candidate resistance mechanisms. Prospective prediction validation studies in newly developed PDX models are ongoing.

Conclusion

Together, these data suggest that PDX models should be useful for generating predictors of treatment response. Ultimately, if robust genetic, epigenetic, or molecular predictors of differential treatment response can be identified and applied clinically, it may be possible to avoid ineffective treatments, minimize exposure to cytotoxic agents, and ultimately increase survivorship.

Samuel Shelton, Anaid Diaz, Christopher Taylor, Richard Shaw, Robert Thong, Adrian Alexa, Fiona Nielsen

Repositive Ltd, Betjeman House, 104 Hills Road, Cambridge, CB2 1LQ

Introduction

A major challenge for researchers working in drug development is gaining access to suitable models to evaluate new targeted therapeutics. Data from commercial PDX models are typically not openly accessible and gaining access to relevant data for these models can be a lengthy process. In addition, a lack of standardisation for metadata and molecular characterisation makes it challenging to find, compare and evaluate models. Repositive is addressing these challenges through the Cancer Models Community and Cancer Models Platform.

Material and Methods

Repositive is assembling a community of researchers from biopharma, contract research organisations (CROs) and academia to help define best practices for the molecular characterisation of oncology models, starting with PDX models. In consultation with researchers from biopharma, we are establishing a set of guidelines for the types and technical specifications of molecular characterisation data. We have also developed a series of bioinformatic workflows which we are making openly available to the research community. Ultimately, all data on the platform will be processed through these workflows, providing greater transparency and improved comparability.

Results and Discussion

The Cancer Models Platform is a secure online resource for the translational oncology research community to share information about oncology models. It provides the flexibility for commercial CROs to present their model data, while giving them control over who has access to specific aspects of that data. Data presented on the platform is de-identified and compliant with current data protection laws. The Cancer Models Platform allows biopharma researchers to search for models securely and anonymously, while allowing researchers to engage with CROs who have suitable models and services.

Repositive is striving to improve the discoverability and comparability of oncology models through its platform and by engaging with the commercial oncology research community to define best practices for model characterisation.

Conclusion

Repositive has developed a commercial platform to provide a single, secure, data portal through which researchers gain access to a broad range of oncology research models and services. In doing this, Repositive is developing guidelines for best practices in data generation, analysis and protection. Collaboration between Repositive and other initiatives such as EuroPDX could allow synchronisation and improved compatibility between the initiatives, broadening the reach and impact of each individual initiative.

Trinidad EM, Ciscar M, Perez-Montoyo H, Cimas F, Gomez-Aleza C, Gomez-Miragaya J, Sanz-Moreno A, Morilla I, Petit A, Soler MT, Martinez-Aranda A, Sierra A, Rakha E, Green A, Quintela M, Arribas J, Serra V, Clarke R and Gonzalez-Suarez E

Introduction

In the last years RANK signalling pathway has been demonstrated to be essential for the mammary gland development. RANK signalling pathway mediates the major proliferative response of mouse and human mammary epithelium to progesterone. RANK loss or overexpression results in a disrupted mammary gland development during pregnancy and impairs lactation.

RANK pathway had also proved its essential role in tumor context, where it is the main mediator of the protumorigenic role of progesterone in the mammary gland. Inhibition of RANK pathway decreases the incidence of spontaneous preneoplastic lesions, increases tumor latency and impairs lung metastasis in oncogene driven mouse models. Importantly, RANKL inhibition also prevents mammary tumorigenesis in progestin/carcinogen induced tumors. Concomitantly RANK overexpression in human breast cancer cell lines increases stemness, tumorigenesis and metastasis.

Therefore, we hypothesize that in human breast adenocarcinomas RANK expression could increase the cancer stem cell population, leading to increased incidence of relapse, metastasis and resistance to chemotherapy.

Material and Methods

RANK and RANKL protein expression were analyzed in several independent Tissue Micro Arrays (TMAs) of breast cancer by immunohistochemistry (IHC). Through Patient Derived tumour Xenografts (PDXs) we evaluated whether the activation of RANK pathway could modify the tumor clinical parameters.

Results and Discussion

RANK and RANKL expression is found on tumor cells but also on stromal cells. Study of RANK in several collections in TMAs indicated that even though RANK is expressed in human breast adenocarcinomas, the expression levels are low. This expression is associated with ER-PR-, TNBC, big tumor size, advanced stage of the disease and poor prognosis.

RANK expression in breast cancer PDX resembles the pattern found in patients. Although low RANK expression levels were found in adenocarcinomas, RANKL treatment was able to activate NF-KB pathway. Three PDX models were treated with RANKL but only lymphoproliferative PDX, with higher levels of RANK, exhibited an increase in the number of tumor initiating cells (TICs) and tumor growth.

After RANKL treatment we analysed gene expression changes in the three different PDX model. This analysis showed that the number of genes and pathways regulated in each PDX model was proportional to the RANK expression levels.

Conclusion

RANK expression is associated with ER-PR-, TNBC, big tumor size, poor prognosis group and advanced stage of human breast cancer. Breast cancer PDX recapitulate the patterns of RANK expression found in human adenocarcinomas. RANKL treatment increases the number of TICs and tumor growth in certain PDX models, possibly related to the RANK expression levels. RANKL stimulation induces a common gene expression signature irrespectively of the tumor of origin.

Hyunsoo Kim, Pooja Kumar¹, Francesca Menghi¹, Javad Noorbakhsh¹, Eliza Cerveira¹, James Keck², Carol J. Bult³, Charles Lee¹, Edison T. Liu³, Jeffrey H. Chuang^1,4

¹ The Jackson Laboratory for Genomic Medicine, Farmington, CT 06030, USA

² In Vivo Services, JAX® Mice, Clinical & Research Services, The Jackson Laboratory, Sacramento, CA 95838, USA

³ The Jackson Laboratory, Bar Harbor, ME 04609, USA

⁴ UConn Health, Department of Genetics and Genome Sciences. Farmington, CT 06030, USA

Introduction

The processes by which tumors evolve are essential to the efficacy of treatment, but quantitative understanding of intratumoral dynamics has been limited. Although intratumoral heterogeneity is common, quantification of evolution is difficult from clinical samples because treatment replicates cannot be performed and because matched serial samples are infrequently available.

Material and Methods

To circumvent these problems we derived and assayed large sets of human triple-negative breast cancer xenografts and cell cultures from two patients, including 86 xenografts from cyclophosphamide, doxorubicin, cisplatin, docetaxel, or vehicle treatment cohorts as well as 45 related cell cultures. We assayed these samples via exome-seq and/or high-resolution droplet digital PCR, allowing us to distinguish complex therapy-induced selection and drift processes among endogenous cancer subclones with cellularity uncertainty < 3%.

Results and Discussion

For one patient, we discovered two predominant subclones that were granularly intermixed in all 48 co-derived xenograft samples. These two subclones exhibited differential chemotherapy sensitivity – when xenografts were treated with cisplatin for 3 weeks, the post-treatment volume change was proportional to the post-treatment ratio of subclones on a xenograft-to-xenograft basis. A subsequent cohort in which xenografts were treated with cisplatin, allowed a drug holiday, then treated a second time continued to exhibit this proportionality. In contrast, xenografts from other treatment cohorts, spatially dissected xenograft fragments, and cell cultures evolved unsystematically but with substantial population bottlenecks.

Conclusion

These results show that ecologies susceptible to successive retreatment can arise spontaneously in breast cancer in spite of a background of irregular subclonal bottlenecks, and our work provides to our knowledge the first quantification of the population genetics of such a system. Intriguingly, in such an ecology the ratio of common subclones is predictive of the state of treatment susceptibility, showing how measurements of subclonal heterogeneity could guide treatment for some patients.

J.M. Lubbers, N. van Rooij, M.J.C. Broekhuis, G.C. Huizinga, D.O. Warmerdam, S.A. Juranek, F. Foijer, H.W. Nijman, M. de Bruyn

University of Groningen, Groningen, the Netherlands

Introduction

Immunotherapy has emerged as a promising anti-cancer strategy, eliciting curation even in previously treatment-refractory diseases. However, patient-to-patient variation in response complicates patient stratification for specific immunotherapeutic modalities. Current studies have thus far failed to explain this variety as there is a shortage of volunteers for inclusion in clinical trials. In addition, no preclinical model exists that accurately resembles the patient for studies regarding cancer immunotherapy. Either the tumor is not represented well due to the use of homogeneous cell lines that additionally lack the tumor microenvironment, or, in the case of Patient Derived Xenograft (PDX) models, the immune system is lacking to prevent graft rejection. To address this issue, we aim to establish a novel, immune-competent PDX model.

Material and Methods

The approach to create this model is two-fold. The first step is immune-humanizing FOXN1-knock out mice via implantation of de novo thymus tissue that we generate from patient-derived induced Pluripotent Stem Cells. The second step is implanting patient-matched PDX tissue into these immune-humanized mice.

Results and Discussion

Thus far, we have been able to obtain patient-derived Stem Cells by reprogramming renal epithelial cells from the urine of ovarian cancer patients. Moreover, we have successfully redifferentiated these cells in vitro into progenitor thymus cells by directed differentiation that mimics organogenesis, Upon implantation of these cells, they should further differentiate into functional thymus tissue that serves to mature progenitor T cells into functional T cells.

Conclusion

As no current preclinical model is suitable to study cancer immunotherapy, we aim to create an immune-humanized PDX mouse model. Our strategy is two-fold, and we are currently first working on creating an sustainable immune-humanized mouse. At present, we have implanted the progenitor thymus cells and are awaiting data to analyse the efficacy of thymus maturation and T cell development in the mice.

J Gómez-Miragaya¹, S Morán1, L Palomero², R Tonda3, M Palafox¹, V Serra⁴, MA Pujana², M Esteller¹, XS Puente⁵ and E González-Suárez¹

¹ PEBC, IDIBELL, Barcelona, Spain

² ProCURE, IDIBELL, Barcelona, Spain

³ CNAG, Barcelona, Spain

⁴ VHIO, Barcelona, Spain

⁵ University of Oviedo, Oviedo, Spain

Introduction

Taxane-based regimens constitute the most common therapeutic option in patients with triple-negative breast cancer (TNBC). However, primary or acquired chemoresistance are common events and count on being the main cause of death in breast cancer patients. Breast cancer patient-derived xenografts(PDX) have emerged as powerful tools for the study of cancer biology and drug treatment response. Specific methylation patterns have been associated to clinical breast cancer subtypes but methylation remains unstudied in breast cancer PDX. We hypothesize that genetic and epigenetic changes contribute to the acquisition of chemoresistance to docetaxel in TNBC patients.

Material and Methods

Aiming to elucidate docetaxel-associated chemoresistance mechanisms, exome-sequencing, genome-wide DNA methylation and transcriptomic profiling have been performed in preclinical breast cancer patient-derived xenograft (PDX) models of different subtypes and with primary and acquired chemoresistance to docetaxel.

Results and Discussion

We found that DNA methylation patterns from breast cancer PDX are closer to breast cancer clinical samples than breast cancer cell lines (BCCLs) and they maintain subtype specific methylation patterns. Triple negative breast cancer (TNBC) PDX models show very stable methylation patterns accompanying chemoresistance acquisition but some critical genes/pathways were unraveled as differentially methylated.

Most genetic changes in the original human metastasis were maintained after serial passages in mice and long-term docetaxel treatment. We identified a chromosomal amplification present in the human metastatic sample and chemoresistant PDX of a BRCA1-mutant, but not in the docetaxel-sensitive PDX, which was further validated in another mutant model. Increased gene expression of multiple genes contained in that region was observed in residual disease and docetaxel resistant tumors of additional TNBC PDX. Clinical data confirmed that this amplification is associated with a subset of TNBC breast cancer patients, a BRCA1 mutational signature and poor survival after chemotherapy.

Conclusion

Our findings reveal tumor clonal dynamics during engraftment and drug treatments and identify differentially methylated genes and a specific amplification which may contribute to the acquisition of resistance to deocetaxel in TNBC.