Poster session 1 (free drinks)

Antonia Graja^1,2, Sarah Meneceur^1,3, Sandra Hering⁴, Michael Baumann^1,2,3,5,6,7, Mechthild Krause^1,2,3,5,6 and Cläre von Neubeck^1,2

¹ OncoRay – National Center for Radiation Research in Oncology, Faculty of Medicine and University Hospital Carl Gustav Carus, Technische Universität Dresden, Helmholtz-Zentrum Dresden – Rossendorf, Germany

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

³ Helmholtz-Zentrum Dresden – Rossendorf, Institute of Radiooncology – OncoRay, Germany

⁴ Institute for Legal Medicine, Faculty of Medicine, University Hospital Carl Gustav Carus, Technische Universität Dresden, Germany

⁵ Department of Radiotherapy and Radiation Oncology, Faculty of Medicine and University Hospital Carl Gustav Carus, Technische Universität Dresden, Germany

⁶ National Center for Tumor Diseases (NCT), Partner Site Dresden, Germany: German Cancer Research Center (DKFZ), Heidelberg, Germany; Faculty of Medicine and University Hospital Carl Gustav Carus, Technische Universität Dresden, Dresden, Germany, and; Helmholtz Association/Helmholtz-Zentrum Dresden – Rossendorf (HZDR), Germany

⁷ German Cancer Research Center (DKFZ), Heidelberg, Germany

Introduction

Clinical cancer research is mainly based on the use of cell line derived xenografts. However, those xenografts often fail to recapitulate the heterogeneity of human malignancies which leads to misjudgment of drug and radiation sensitivities as well as potentially to inadequate treatment decisions. To include the patient’s individual biological status, patient derived xenograft (PDX) models have emerged as a powerful tool in cancer research.

Material and Methods

Primary tumor material (P) derived from patients undergoing surgery for head and neck squamous cell carcinoma (HNSCC) or glioblastoma multiforme (GBM) was transplanted subcutaneously into NMRI Foxn1^nu/nu mice. At 4 cm³ tumor volume (F1) animals were sacrificed and tumors were taken for analysis and sequential transplantation (F2, F3). Analysis for immunogenic potential was performed by subcutaneously transplanting tumor pieces on the hind-leg of whole-body irradiated (4 Gy) or non-irradiated NMRI Foxn1^nu/nu mice. Afterwards, tumors at a volume of 0,12 cm³ were randomly allocated in the treatment groups of graded single doses under clamp blood flow conditions inducing homogenous hypoxia. Local tumor control was determined over 240 days.

Results and Discussion

Successful engraftment for 34 HNSCC-PDX (32%) and 22 GBM-PDX (27%) was achieved. Engraftment rates as well as tumor volume doubling times stabilized in F2 and F3, suggesting a clonal selection due to sequentially passaging. Morphologic characteristics in P- and F-samples were comparable but genetic alterations were evident in microsatellite analysis strengthening the assumption of a clonal selection. Preliminary results of irradiated PDX indicate longer latency times and lower take rates if the host did not receive a whole body irradiation.

Conclusion

PDX models are an important tool to develop novel strategies against drug-and/or radio-resistant tumors in preclinical trials. However, only non-immunogenic xenografts seem to be suitable for radio-oncologic questions leading to a personalized and successful cancer therapy.

Alberto Villanueva^1,2,#, Aroa Soriano³, Gabriela Guillén⁴, Marta Garrido⁴, Josep Roma³, Juana Fernández-Rodriguez⁵, Conxi Lazaro⁵, Carol Armengol⁶, Miguel F. Segura³, Josep Sánchez de Toledo^3,4 and Soledad Gallego^3,4

^# Correspondence: avillanueva@iconcologia.net

¹ Group of Chemoresistance and Predictive Factors, Program Against Cancer Therapeutic Resistance (ProCURE), Catalan Institute of Oncology (ICO-IDIBELL), L’Hospitalet de Llobregat, Barcelona, Spain

² Xenopat S.L., Business Bioincubator, Bellvitge Health Science Campus, Barcelona, Spain

³ Group of Translational Research in Child and Adolescent Cancer, Vall d’Hebron Research Institute, Passeig Vall d’Hebron 119, 08035 Barcelona, Spain

⁴ Vall d’Hebron Hospital, Passeig Vall d’Hebron 119, 08035 Barcelona, Spain

⁵ Hereditary Cancer Program, Catalan Institute of Oncology (ICO-IDIBELL), L’Hospitalet de Llobregat, Barcelona, Spain

⁶ Childhood Liver Oncology Group (c-LOG)-CIBEREHD, Program for Predictive and Personalized Medicine of Cancer (PMPPC), Germans Trias i Pujol Research Institute (IGTP) I Campus Can Ruti, Badalona, Spain

Introduction

The prognosis for children with high-risk/metastatic tumors is often poor and survivors can suffer from severe side effects. Predictive advanced preclinical models and novel therapeutic strategies for high-risk disease are therefore a clinical imperative. Despite the improvement in patient survival during the last 50 years, 20% of all pediatric cancer remains incurable. Herein, we show the feasibility of developing children-derived orthotopic xenografts (named Childox) from different tumor-types, locations and samples types, and its usefulness for deciding therapeutic interventions in real-time.

Material and Methods

In the last three years we have orthotopically implanted 55 pediatric tumors mainly from the Vall d’Hebrón Hospital (Pediatric Unit), although some also came from other Spanish hospitals. Both, primary tumors and metastases from different locations were implanted in mice in their original locations, and maintaining their three-dimensional tissue structure. Additionally, in several cases, Childox was used as a tool to integrate in real-time the genetic alterations of the tumor and the response to selected drugs.

Results and Discussion

To data 28 children-derived orthoxenografts have been generated including: Rhabdomyosarcomas, Ewing sarcomas, Malignant peripheral nerve sheath tumors (MPNST), Neuroblatomas, Osteosarcomas, Wilm’s tumors, Hepatoblastomas, Malignant rhabdoid tumor, Hepatocarcinomas and other sarcomas. For several tumor types, metastatic models were developed in their original locations (brain metastases of neuroblastoma, breast metastasis from alveolar rhabdomyosarcoma, lung metastases of osteosarcoma). The implantation take rates have been improving over time, being 14% in 2015, 22% in 2016 and 43% in 2017. Two patients (urologic rhabdomiosarcoma; MPNST) were treated based on the drug-response results obtained from the treatment of their own Childox (some drugs selected according genomic profiling).

Conclusion

Children-derived orthoxenografts may be outstandingly valuable preclinical tools to improve the treatment of bad prognosis/metastatic pediatric tumors and for the development of personalized treatments.

Puliga Elisabetta¹, Capeloa Tania Isabel^2,#, Apicella Maria¹, Migliore Cristina², Petrelli Annalisa¹, Ughetto Stefano², Durando Stefania¹, GEA Consortium, Corso Simona^1,2 and Giordano Silvia^1,2

¹ Candiolo Cancer Institute, FPO IRCCS, Candiolo, Italy

² University of Torino, Department of Oncology, Candiolo, Italy

^# Current Affiliation: Pole of Pharmacology, Institut de Recherche Expérimentale et Clinique (IREC), Universitè catholique de Louvain (UCL), Brussels, Belgium

Introduction

Gastric cancer (GC) is the third leading cause of cancer-related death worldwide. The overall clinical outcome for patients with advanced disease is poor (5-20% 5-year survival). Recently, whole genome molecular analysis revealed that KRAS is frequently altered, not only due to point mutations but also to gene amplification (9% of the patients). Since the oncogenic role of KRAS mutations and amplifications has been poorly studied in GC, taking advantage of a proprietary, molecularly annotated colony of gastric cancer PDXs (Patient-derived xenografts) we aim to i) search for strategies effectively targeting KRAS in GC, and ii) unravel differences between KRAS amplification and mutation in terms of response to therapy.

Material and Methods

To address our analysis we took advantage of a multi-level platform of GC models, comprising more than 211 GC PDXs, primary cell lines and organoids. Our routinely RT-PCR analyses identified 12 KRAS amplified PDXs (12 to>40 gene copies), while Sanger sequencing confirmed the presence of KRAS mutations in 11 PDXs. Cell lines and organoids were generated from these models.

Results and Discussion

To evaluate the activation status of KRAS in amplified and mutated cells we first performed a RAS-GTP pull down. Mutated cells showed a higher RAS-GTP/RAS ratio than KRAS amplified cells. However, the total amount of active KRAS was greater in KRAS amplified cells compared to KRAS mutated ones. These results demonstrate that, in spite of the fact that the specific activity of KRAS is lower, the increased KRAS expression in amplified cells makes the total KRAS activity higher in KRAS amplified vs mutated cells. To investigate whether KRAS can be targeted in our mutated or amplified models, we evaluated cell addiction to this oncogene. Silencing experiments showed that indeed both mutated and amplified cells rely on KRAS activity for survival, suggesting that its inhibition might lead to a therapeutic response.

Recent studies demonstrated that KRAS mutated tumors present specific metabolic changes required for their maintenance and aggressiveness. To explore if these changes can also occur in gastric cancer bearing KRAS mutations or amplifications we started studying cell metabolism in our models. Western blot analysis showed a significant increase in glycolytic markers such as the glucose transporter 1, monocarboxylate transporter 4, hexokinase II and pyruvate kinase M2 in both KRAS mutated and amplified cells. Biochemical analysis revealed that KRAS amplified models as well as KRAS mutated ones showed similar level of glycolytic markers and lactate production/uptake.

Conclusion

TNBC PDX reproduce the molecular heterogeneity and diversity of TNBC patients. This large collection of PDX is a clinically relevant platform for drug testing, biomarker discovery and translational research.

Florence Coussy^1,2, Ahmed Dahmani², Elodie Montaudon², Virginie Bernard¹, Marion Lavigne³, Leanne De Koning⁴, Sophie Chateau-Joubert⁵, David Gentien⁶, Ivan Bieche^1,7 Elisabetta Marangoni²

¹ Candiolo Cancer Institute, FPO IRCCS, Candiolo, Italy

² University of Torino, Department of Oncology, Candiolo, Italy

^# Current Affiliation: Pole of Pharmacology, Institut de Recherche Expérimentale et Clinique (IREC), Universitè catholique de Louvain (UCL), Brussels, Belgium

Introduction

Triple negative breast cancer (TNBC) is a very heterogeneous disease. Patients diagnosed with TNCB have a poor prognosis, and identification of new biomarkers and therapeutic agents is a high priority. Patient Derived Xenograft (PDX) are clinically relevant models that have emerged as important tool for the analysis of drug activity and predictive biomarker discovery. The purpose of this work was to analyse the molecular heterogeneity of a large panel of TNBC PDX with the perspective to test targeted therapy and to identify biomarkers of response.

Material and Methods

PDX of early stage TNBC established from 2003 to 2016 (N=61) were molecularly characterized for gene expression (Human Genome U133 Plus 2.0 array) and somatic copy number analysis (Cytoscan HD Arrays). Subtypes were identified with the tool TNBCtype (Chen et al. 2012) based on transcriptomic data. A targeted next-generation sequencing (NGS) of 100 genes (the top frequently mutated genes in breast cancer) was performed on Illumina HiSeq2500 sequencer. COSMIC, Tumorportal and cBioportal databases were used for the interpretation of genomic variants. Frequencies of genomic alterations were then compared to the TGCA breast cancer data. Immunochemistry (IHC) and morphological analysis of PDX were performed as compared to the corresponding patients’ tumors. PDXs carrying targetable genomic alterations in the PI3K/AKT/mTOR and MAPK pathways were treated by specific inhibitors (selumetinib, BAY80-6946 and PF-04691502).

Results and Discussion

At the gene expression level, TNBC PDXs represent all the different TNBC subtypes identified by the Lehmann classification. The frequency of the different TNBC subtypes was similar to the TGCA TNBC, expect for the immunomodulatory subtype, underrepresented in PDX. Somatic pathological mutations and copy number alteration were similar in PDX and TCGA TNBC patients. Among the top altered genes are TP53 and oncogenes and tumor suppressors of the PI3K/AKT/mTOR and MAPK pathways (including PIK3CA, AKT1, NF1 and NRAS/KRAS). At the histological level, TNBC PDX were mainly composed of invasive ductal carcinoma of no special type, with some tumors being classified as apocrine or metaplastic carcinomas. Comparison with the original tumors show similar pattern (based on IHC analysis of CK5, CK8/18, CK14 and AR, FOXA1, EGFR and Ki67). In vivo efficacy experiments with PI3K and MAPK pathways inhibitors showed marked anti-tumor activity in PDX carrying genomic alterations of PIK3CA, AKT1 and NRAS, NF1 genes. Combination of PI3K and MAPK inhibitors in PDX with simultaneous genomic alterations of PI3KCA and MAPK related genes resulted in marked anti-tumor activity.

Conclusion

All together these results suggest KRAS as driver and a therapeutic target in GC. Moreover, increasing evidences of the metabolic changes in tumor cells with KRAS deregulation prompts the use of metabolism-targeting drugs.

Apicella Maria¹, Migliore Cristina², Durando Stefania¹, Ughetto Stefano², D’Errico Laura², Cargnelutti Marilisa¹, Bellomo Sara Erika¹, Moya Daniel Rull², Corso Simona^1,2 and Giordano Silvia^1,2

¹ Candiolo Cancer Institute, FPO IRCCS, Candiolo, Italy

² University of Torino, Department of Oncology, Candiolo, Italy

Introduction

EGFR is amplified in around 5% of gastric cancers, characterized by poor prognosis. Two phase III studies (EXPAND and REAL3) have investigated the efficacy of Cetuximab in association with chemotherapy in patients with previously untreated tumors, but they did not observe any benefit with the addition of EGFR monoclonal antibodies. However, in these two studies no patient selection was done on the basis of EGFR amplification/overexpression. While the EGFR status is not usually analysed in colon cancer, as the wild type receptor activity is considered critical for tumor growth, no such data are available in Gastic Cancer. Recently, it was reported that 57% of patients with EGFR amplification (54-167 gene copies) treated with anti-EGFR mAbs benefitted from anti-EGFR treatment.

Material and Methods

At the moment, the best preclinical model to validate targets and positive/negative response predictors is represented by Patient-Derived Xenografts (PDXs), an experimental model that retains the principal histologic and genetic characteristics of the donor tumor, is predictive of clinical outcome and is a valuable tool for personalized medicine strategies. We have recently generated a molecularly annotated colony of gastro-oesophageal PDXs (at the moment, >200 PDXs). The platform also comprises primary cell lines and 3Dcoltured organoids derived from gastric cancer PDXs. EGFR amplified gastric cancers were selected choosing the PDXs with the highest levels of EGFR amplification and overexpression (a minimum of 8 copies has been established as the inferior limit).

Results and Discussion

Three models presenting between 10 and 450 copies of EGFR were enrolled in three different in vivo preclinical trials. As there is no knowledge of which anti- EGFR drug/s could be more effective in this context, we treated our PDXs with different drugs such as anti-EGFR TKIs (Erlotinib, Lapatinib and Afatinib) and mAbs (Cetuximab and Panitumumab), as single agents or in combination. RECIST criteria were used to evaluate treatment responses and identify clinically relevant outcomes. We observe a clinical response in ²⁄₃ models with the mAbs.

Moreover, the combined treatment with mAbs and TKIs was more active than the mAbs mono-treatment, not only in inducing a complete but also a durable response. Finally, in the non-responsive case we discovered a novel negative biomarker of response to anti-EGFR therapy, suggesting a possible therapeutic option for these patients.

Conclusion

We demonstrated that EGFR can be a good target in GC patients displaying high level of EGFR amplification (at least more than 8 copies).

Bellomo Sara Erika¹, Durando Stefania¹, D’Errico Laura², Bass Adam⁴, Isella Claudio¹, Medico Enzo^1,2, Sapino Anna^1,3, GEA Consortium, Giordano Silvia^1,2 and Corso Simona^1,2

¹ Candiolo Cancer Institute, FPO IRCCS, Candiolo, Italy

² University of Torino, Department of Oncology, Candiolo, Italy

³ Department of Medical Science, University of Torino, Torino, Italy

⁴ Department of Medical Oncology, Dana-Farber Cancer Institute, Boston, MA

Introduction

Gastric cancer is the third leading cause of cancer-related deaths worldwide. Despite significant therapeutic improvements, advanced gastric cancers continue to have a poor outcome, with a median overall survival of about 10 months. Several molecular therapies entered clinical trials during the last 5 years but, with the exception of Trastuzumab and Ramucirumab, all of them failed, indicating the need for a broader understanding of the genomic landscape of this tumor before going to clinic.

Material and Methods

To identify and validate novel therapeutic approaches for gastric cancer we generated a large and “multi-level” platform of gastric cancer preclinical models annotated for molecular alterations. To complement the suite of mouse models, we have also developed a unique collection of PDX-derived organoids (>50) and primary cell lines (>50).

Each PDX has been molecularly annotated for the following:

• Copy number variation (CNV) for the most frequently amplified receptor tyrosine kinases (MET, HER2, EGFR, FGFR2) and KRAS (qReal Time PCR)
• Sequencing of a panel of 250 gastric-specific genes
• Expression profiling (Illumina microarrays)
• EBV status
• MSI status

CNV analysis, global expression profile, MSI and EBV status have been evaluated also in primary tumors, as well as in primary cell lines and organoids.

Results and Discussion

The platform is based on more than 211 PDXs generated by subcutaneous implant of human gastric tumors (576 samples) in NOD SCID mice (engraftment rate around 40%). Around 30% of the mice developed human-derived lymphomas that were monoclonal and EBV positive, with a mutational burden and an expression profile distinct from gastric adenocarcinomas. Regarding potential factors influencing PDX generation, we did not observe any statistically significant correlation between engraftment and patient characteristics or pathological features of the tumors such as age, gender, history of neoadjuvant chemotherapy, tumor site or EBV status

On the contrary, intestinal tumor histology, TNM III/IV stage, MSI+ status or the amplification of receptor tyrosine kinases positively affected engraftment rate. Taking advantage from PDXs transcriptional profiles we identified a MSI signature associated with a prognostic value independent from MSI status. Remarkably, all the molecular analyses confirmed a very good correspondence among primary tumors, PDXs, cell lines and organoids.

Conclusion

To the best of our knowledge, this collection of gastric PDXs is the largest available in an academic institution and one with the most accurate annotation of molecular characteristics. As a whole, this work will hopefully provide a scientific basis for future clinical applications and guide the rational design of molecularly-oriented clinical trials for gastric cancer.

Lidia Moserle, Roser Pons, Mar Martínez-Lozano, Gabriela A. Jiménez-Valerio and Oriol Casanovas

Tumor Angiogenesis Group, ProCURE Research Program, Catalan Institute of Oncology – IDIBELL, Barcelona, Spain

Introduction

Antiangiogenic therapies are currently used to treat several types of cancer, including renal cell carcinoma (RCC). However, these treatments fail to produce long-term durable effects in many patients, due to tumor adaptation and resistance to therapy. Furthermore, preclinical data have shown that antiangiogenic therapies can switch on an invasive and metastatic phenotype as an adaptation to the inhibition of the VEGF pathway. But this is still controversial in the field.

Material and Methods

We have developed an extensive series of patient-derived orthotopic xenograft models (Ren-PDOX) from RCC patients. Tumor samples from 56 untreated clear-cell RCC patients have been implanted in the kidney of mice and 27 Ren-PDOXs models have been stably generated, exhaustively reproducing the inter-patient heterogeneity. These models not only maintain the original patient’s tumor architecture, histological and genetic characteristics, but also recapitulate human neoplastic progression from primary tumor to spontaneous metastasis development.

Results and Discussion

By evaluation of Ren-PDOX tumor margins and local invasion at tumor-kidney interface, we have observed that in specific tumors blockade of VEGF signaling results in increased invasiveness and metastatic dissemination, while in others aggressiveness remains unchanged. Furthermore, using a novel tumor-specific RNA-sequencing analysis we have identified several pathways as important axes implicated in the increased malignant phenotype of renal tumors after anti-angiogenic therapy.

Conclusion

Our results confirm that the effects of antiangiogenic drugs on tumor invasion and metastatization are heterogeneous and in specific patients they may profoundly affect the natural progression of tumor and promote malignancy. Additionally, this reinforces the importance of appropriate models for preclinical research to identify markers and tumor features for the selection of patients that could really benefit from treatment.

Shang Li¹, Joseph J. Caumanns², Gert Jan Meersma¹, Ate. G.J. van der Zee², G. Bea A. Wisman², Steven de Jong¹

¹ Department of Medical Oncology, University Medical Center Groningen, Groningen, The Netherlands

² Department of Gynecologic Oncology, University Medical Center Groningen, Groningen, The Netherlands

Introduction

Ovarian clear cell carcinoma (OCCC) is a subtype of ovarian cancer with high chemoresistance and a poor prognosis in advanced stage, which emphasizes that novel therapies for OCCC are warranted. Patient-derived xenograft (PDX) models are an in vivo platform of increasing interest to test therapy strategies as they more accurately represent patient tumors compared to conventional in vivo models. In this study, we compared the protein expression and mutational features between paired primary OCCC tumors and PDX tumors, and tested a novel treatment based on common mutational features.

Material and Methods

To establish PDX models of OCCC, tumor tissues from 14 OCCC patients were implanted subcutaneously in NSG mice. The PDXs and paired primary tumor tissues were characterized by sequencing frequent OCCC mutations. To further assess the morphological and genomic effects of engraftment, immunohistochemistry was used to analyse Ki67 and SNP array analysis was used to determine copy number alterations (CNAs). Efficacy of PI3K/AKT/mTOR and MAPK kinase inhibitors (AZD8055, GDC0941 and selumetinib) as single treatment and in combination were tested in a panel of seven OCCC cell lines with different genetic makeup (n=7). The established PDXs were treated with a combination of AZD8055, GDC0941 and selumetinib to test their potential in vivo in relevant OCCC preclinical models.

Results and Discussion

OCCC primary tumors had a 50% engraftment success rate as PDX. In primary implantation (F1), using fresh tumor tissue instead of vitrified ones led to higher success rate but with similar growth latency. For F2 implantation, higher success rate, around 50% shorter latency time and faster growth were mostly observed, in line with Ki67 staining results. Mutations of OCCC-related genes, ARID1A and PI3K/Akt signaling (e.g. PIK3CA, TP53, PTEN, BRCA1) retained mutation during engraftment.

Consistent with the similar morphological features between paired tumor and PDXs by H&E staining, only 3% of CNAs were either gained or lost in PDXs compared to patient tumors. CNAs were observed in 67% of the genes. IC20 combinations of AZD8055, GDC0941 and selumetinib effectively inhibited proliferation in all seven cell lines. Finally, the antitumor activity of low-dose triple combinations of PI3K/AKT/mTOR and MAPK kinase inhibitors in OCCC PDX significantly reduced tumor growth and induced apoptosis, while weight loss in mice was not observed.

Conclusion

Seven OCCC PDX models were established that generally mimicked their patient tumor at the genomic and protein level. The application of PDX models in combining PI3K/AKT/mTOR and MAPK kinase inhibitors showed the power of PDXs for future translational research in OCCC.

Irene Rius Ruiz^1,4,9, Rocío Vicario^1,9, Beatriz Morancho^1,4, Enrique Arenas¹, Johannes Sam⁵, Jeff Sperinde⁶, Cristina Saura^2,3, Christian Klein⁵, Cristina Bernadó Morales^1,4,10 and Joaquín Arribas^1,4,7,8,11

¹ Preclinical and ² Clinical Research Programs, Vall d’Hebron Institute of Oncology (VHIO), ³ Vall d’Hebron University Hospital (HUVH) and ⁴ CIBERONC, Barcelona. 08035, Spain

⁵ Roche Innovation Center Zurich, Roche Pharmaceutical Research & Early Development, Wagistrasse 18, 8952 Schlieren, Switzerland

⁶ Monogram Biosciences, Laboratory Corporation of America® Holdings, South San Francisco, CA, USA

⁷ Department of Biochemistry and Molecular Biology, Universitat Autónoma de Barcelona, Campus de la UAB, 08193, Bellaterra, Spain

⁸ Institució Catalana de Recerca i Estudis Avançats (ICREA), 08010, Barcelona, Spain

⁹ Equal contributors

¹⁰ Presenter

¹¹ Corresponding author

Introduction

T cell bispecific antibodies (TCBs) are engineered molecules that include, within a single entity, binding sites to the T cell receptor and to tumor-specific or tumor-associated antigens. The receptor tyrosine kinase HER2 is a tumor-associated antigen in ~25% of breast cancers. TCBs targeting HER2 may result in severe toxicities, likely due to the expression of HER2 in normal epithelia. Approximately 40% of HER2-positive tumors express detectable levels of p95HER2, a carboxy-terminal fragment of HER2.

Material and Methods

In vitro co-culture assays to assess T cell mediated cytotoxicity against normal cells, tumor cell lines and Patient-Derived Xenografts (PDX)-derived cultures. In vivo Breast-Cancer PDX models grafted in both PBMCs (patient-matched and non-matched) and CD34+ Humanized mice to assess TCB effect. Intracraneal tumor graft to assess, by in vivo bioluminescence imaging, TCB effect on brain metastasis using CD34+ Humanized mice.

Results and Discussion

Using specific antibodies, here we show that p95HER2 is not expressed in normal tissues. We developed a p95HER2-TCB and show that it has a potent anti-tumor effect on p95HER2-expressing breast primary cancers and brain lesions. In contrast with a TCB targeting HER2, the p95HER2-TCB has no effect on non-transformed cells expressing normal levels of HER2.

Conclusion

These data pave the way for the safe targeting of a subtype of HER2-positive tumors with a novel tumor-specific TCB.

Ramos-Medina Rocío¹, López-Tarruella Sara², Bernat Rebeca¹, Jeréz Yolanda², García Saenz Jose Angel³, del Monte-Millán María², Gayarre Javier¹, Moreno Fernando³, Asensio Fernando⁴, Martin Miguel²

¹ Laboratory of Translational Oncology, Instituto de Investigación Sanitaria Gregorio Marañón, Madrid, Spain

² Medical Oncology Service, Hospital General Universitario Gregorio Marañón, Universidad Complutense, Madrid, Spain. CiberOnc

³ Medical Oncology Service, Hospital Clínico San Carlos, Madrid, Spain

⁴ Unidad de Medicina y Cirugía Experimental, Instituto de Investigación Sanitaria Gregorio Marañón, Madrid, Spain

Introduction

Breast cancer is a heterogeneous disease, TNBC account for approximately 10-17% of all breast cancers and is associated with a poorer clinical outcome with respect to other BC subtypes. In absence of targeted therapies, conventional chemotherapy remains the therapeutic backbone of TNBC. Neoadjuvant chemotherapy for TNBC always includes combinations of agents, making difficult the identification of biomarkers of response to a particular drug. We engrafted several chemotherapy-naïve human TNBC samples from primary tumors to NSG mice (PDXs) and tested in this model several chemotherapy drugs in order to assess antitumor activity of single drugs and combinations of drugs, to correlate the results with the response of these drugs in the patients from whom the PDX were derived.

Material and Methods

NOD.Cg-Prkdcscid IL2rgtm1Wjl (NSG) mice (female, 8 weeks old) were purchased from Jackson Lab (USA). TNBC biopsy samples from therapy naïve patients were implanted into mammary fat pads. PDXs grew until a median size of 250 mm3 and randomized in 5 groups of treatment (n=10): (a) control; (b) D: 8.10 mg/kg i.v. once a week (qw.); © Cb: 48.64 mg/kg i.p. qw. (d) D+Cb: keeping the doses, qw; (e) Dx: 8.10 mg/ml i.p. qw. PDXs were treated during 4 weeks. Tumor size was evaluated three times a week using the following formula: tumor volume (TV) = [length x width2]/2. The statistical analysis was performed using the t-test. p<0.05 was considered to be significant.

Results and Discussion

From 2015 to 2018 a total of 63 TNBC patients were recruited for the study. After biopsy and tumor implant in the mice, patients were uniformly treated with D+Cb. We selected PDXs from 4 germline BRCA-1 wild type patients and from 1 patient with a germline BRCA-1 mutation. Histopathologic characterization showed a concordance between PDX and the corresponding patient´s tumor.

The evaluation of the volume of the PDX tumors showed that in all PDX models coming from patients who had a good pathological response with D+Cb (Symmans RCB-0&1), mice treated with D+Cb had significantly lower tumor growth than controls (p˂0.0001), with a median of tumor growth inhibition (TGI) of 99%. In the case of models coming from patients that had a poor response to D+Cb (Symmans RCB-3) the TGI in the group of mice treated with D+Cb was only 69%. D and Cb as single agents produce a lower TGI than the combination of both in either group. To validate these in vivo results, we analyse the expression of Ki67 in representative tumors from each of the treatment groups. We observed a clear reduction of Ki67 expression with D+Cb in PDXs whose tumors came from patients with good response, but not in those coming from poor responders.

Conclusion

In our preclinical model of PDX, the activity of the combination of D+Cb in mice mimics the actual response seen in patients treated with the same combination.

Author list: Olivier Déas¹, Ludovic Bigot², Guillaume Lang¹, Yohann Loriot², Benjamin Besse², Stefano Cairo¹, Marie Tavernier¹, Katell Mevel¹, Jean-Gabriel Judde¹, Luc Friboulet²

¹ XenTech, France

² Institut de Cancérologie Gustave Roussy, Villejuif, France

The last 20 years have witnessed the identification of an increasing number of druggable oncogenic drivers and the development and clinical use of specific inhibitors against these targets. Unfortunately, patients treated with targeted therapies consistently develop resistance and progression under treatment. Hence, important scientific, pharmaceutical and medical research efforts are directed towards understanding the mechanisms of acquired resistance to explore new therapeutic pathways.

The MATCH-R clinical trial enrolls patients with oncogene-driven cancer who have had previous clinical response to targeted therapy and subsequently experienced disease progression. In the framework of this project, Gustave Roussy and XenTech are joining forces to develop a panel of patient-derived xenografts (PDXs) derived from biopsies collected from these patients at the stage of acquired resistance. These PDX models will be used to improve knowledge on the mechanisms underlying resistance to treatment and to evaluate response to new treatments.

In this perspective, the development of 75 PDX-AR (Active Resistance) models is planned over 3 years. All the models are maintained under the same therapeutic pressure the parental tumor was submitted to at the time of biopsy, and will be subjected to extensive phenotypic and genotypic characterization.

Among the models established so far are:

• ENDx-MR-004-AR (endometrial): resistant to the combination of MEK and MDM2 inhibitors
• LCx-MR-007-AR: (NSCLC): resistant to a second generation EGFR inhibitor
• UREx-MR-015A-PD-AR (ureter) and VEx-MR-086A-PD-AR (bladder): resistant to a FGFR inhibitor
• PARx-MR-010-AR (parotid): resistant to a NOTCH Inhibitor
• TCx-MR-122-AR (colon): resistant to an ATR inhibitor

To favor successful xenograft establishment, the first two passages were performed without drug treatment, which was applied from the third passage on. When doing so, some models showed resistance from the first passage under treatment, whereas others showed stabilization under treatment at the first passages and rapidly acquired resistance over passages. These different behaviors might underlie different mechanisms of resistance, irreversible (monoclonal) for the former, reversible (polyclonal) for the latter.

Parallel to the development of UREx-MR-015A-PD-AR, we developed the UREx-MR-015B-SD (stable disease) model from a biopsy collected from a different metastasis in the same patient, but stabilized by the therapy. Comparative analysis of these two models will provide important insights into the mechanisms of resistance to FGFR inhibitors. The MatchR PDX project will provide a unique preclinical platform for identifying resistance mechanisms to current targeted therapies and developing next generation therapeutic strategies.

Anne Marthe Fosdahl, Helga Bergholtz, Anna Ploec, Therese Sørlie, Jens Henrik Norum

Department of Cancer Genetics, Institute for Cancer Research, The Norwegian Radium Hospital, Norway

Introduction

Breast cancer (BC) typically develops through multiple stages starting from premalignant atypical hyperplastic lesions, to carcinoma in situ, invasive carcinoma, and in some cases, to metastatic disease. Ductal Carcinoma In Situ (DCIS) is a non-invasive, non-obligate precursor of invasive BC characterised by proliferation of cancerous cells within the lumen of the mammary duct without invasion into surrounding tissue. DCIS lesions are histologically, molecularly and clinically heterogeneous and maybe as few as half of the DCIS cases might progress to invasive disease. In this project we aim to elucidate mechanisms underlying the earliest stages of BC progression.

Material and Methods

Based on previously published protocols, we have established a technique for injecting cells directly into the mouse mammary gland ducts using a minimally invasive method, the so called Mouse INtraDuctal (MIND) method. We inject human DCIS and BC cell lines, as well as dissociated cells from PDX tumours in NSG mice. Tumour growth is monitored by MRI scanning prior to the presence of palpable tumours. Upon euthanizing the mice, samples are collected for histological and molecular analyses to determine tumour growth pattern, invasive potential and molecular characteristics of invasive and non-invasive cells.

Results and Discussion

We show that BC and DCIS cell lines as well as dissociated PDX tumour cells can successfully grow and generate tumours using the MIND method. The injected cells display different growth patterns. Pre-palpable lesions are detectable by MRI scanning. Mammary gland wholemount staining with Carmine Alum Stain showed the presence of enlarged mammary ducts in mice injected with DCIS or BC cells. Immunohistological analyses showed that MRI detectable lesions included areas with intraductal and invasive growth. Next we seek to determine the BC molecular subtype of the injected cells and whether the molecular subtype of the injected cells change upon intraductal and invasive growth.

Conclusion

We have successfully established a pipeline for studying the early events in BC progression, including intraductal growth and the transition to invasive behaviour, monitored by MRI scanning in the early phases. Cells with assumed different phenotypic chrematistics behave differently after intraductal injections, suggesting that the method is suitable to identify the molecular traits of invasive and non-invasive cells.

Introduction

In the past decade, Patient derived Xenografts (PDX) emerged as a reference platform for preclinical cancer research, also providing new opportunities for the characterization of tumor heterogeneity and tumor stroma interactions. PDXs represent a viable in-vivo model for preclinical pharmacological trials that can recapitulate the clinical response spectrum of human tumors; therefore they represent a unique opportunity to design and test innovative therapeutic approaches.

Material and Methods

Collection of PDX from metastatic colorectal cancer. molecular annotation: gene expression microarray and mutational hotspot for KRAS, BRAF, PIK3Ca. Pharmacological annotation: response to anti-EGFR treatment.

Results and Discussion

He PDX platform was recently exploited for different approaches in colorectal cancer (CRC), including: (i) Identification of HER2 amplification as a therapeutically actionable mechanism of resistance to anti-EGFR therapy, validated in patients through the HERACLES clinical trial; (ii) characterization of the stromal heterogeneity of CRC, and description of a novel subtype of tumors accruing high amounts of cancer-associated fibroblasts and endowed with poor prognosis; (iii) identification of five colorectal cancer cell-intrinsic transcriptional subtypes (CRIS A-E), characterized by distinct biological properties, and associated with disease outcomes and response to treatment.

Conclusion

At the moment, the lack of adaptive immune cells in the model represents the major drawback for the application of PDXs in immuno-oncology. However, the development of mouse humanization strategies may overcome this limitation in the near future. Altogether, these data consolidate PDXs as a useful preclinical platform to improve characterization, diagnosis and treatment of human cancer.

Caroline Mignard¹, Jean-François Mirjolet¹, Josselin Caradec¹, Valentin Derangere², François Ghiringhelli³, Laurent Arnould², Séverine Tabone-Eglinger³, Pauline Wajda^3,4, Bertrand Dubois⁴, Fabrice Lavial⁴, Isabelle Treilleux³, Olivier Duchamp¹

¹ Oncodesign, Dijon, France

² Centre Georges François Leclerc, Dijon, France

³ Centre Léon Bérard, Lyon, France

⁴ Cancer Research Center of Lyon, France

Immunotherapy is one of the most exiting recent breakthroughs in the field of cancer treatment. Different approaches are developed such as cancer vaccines, adoptive cellular immunotherapy or immune checkpoint blockade, and a number have been regulatory approved or are currently investigated in clinic.

Effective immunity against cancer involves complex interaction between the tumor, the host and the environment. The assessment of cancer immunotherapy requires the use of appropriate preclinical animal models that sufficiently reflect the physiological situation in humans and that must be chosen carefully to address a specific scientific question.

Patient-derived xenograft (PDX) models have been developed from different tumor types and are used for the pre-clinical drug evaluation and for the predictive results of clinical outcomes because they conserve original tumor characteristics such as heterogeneity and molecular diversity. Nevertheless, they have limited utility for therapies that target the human immune system. This limitation has been partially overcome by the development of humanized mice meaning co-transplantation of human tumor and human immune or hematopoietic progenitor cells.

Humanized models were then used for in vivo evaluation of immunomodulatory agents with mice xenografted with cell lines or patient-derived xenografts. Therapeutic efficiency was assessed by following mice survival and tumor growth. The impact of therapeutics on tumors and immune cells was also assessed by flow cytometry and immunohistochemistry analyses.

We developed a large panel of humanized models that are available for the investigation of the mechanism of action of novel immuno-oncology therapies. Nevertheless challenges are still remaining, for the development of immune checkpoint-based therapy, in particular, to identify and validate predictive models and biomarkers.

Anna Golebiewska¹, Ann-Christin Hau¹, Anaïs Oudin¹, Linsey Houben^1,2, Daniel Stieber³, Francisco Azuaje⁴, Tony Kaoma⁴, Arnaud Müller⁴, Frank Hertel⁵, Michel Mittelbronn^1,6, Rolf Bjerkvig^1,7 and Simone P. Niclou^1,7

¹ NorLux Neuro-Oncology Laboratory, Department of Oncology, Luxembourg Institute of Health, Luxembourg, Luxembourg

² Maastricht Radiation Oncology Laboratory, Department of Radiation Oncology, Universiteit Maastricht, Maastricht, the Netherlands

³ Department of Genetics, Laboratoire National de Santé, Dudelange, Luxembourg

⁴ Genomics and Proteomics Research Unit, Department of Oncology, Luxembourg Institute of Health, Luxembourg, Luxembourg

⁵ Neurosurgical National Department of Luxembourg, Centre Hospitalier de Luxembourg, Luxembourg, Luxembourg

⁶ Department of Anatomic and Molecular Pathology, Laboratoire National de Santé, Dudelange, Luxembourg

⁷ KG Jebsen Brain Tumour Research Centre, Department of Biomedicine, University of Bergen, Bergen, Norway

It is well recognized that long term cell cultures are poor models to study human cancer, largely because of loss of clonal heterogeneity, accumulation or loss of genomic alterations and adaptation to a highly artificial environment. Patient-derived orthotopic xenografts (PDOX) based on organotypic three-dimensional tumor spheroids from human glioma samples are proposed to represent a reliable and clinically-relevant animal model.

We have generated a living biobank of PDOX models from 34 glioma patients (grade III and IV), including longitudinal patient samples with matched recurrent tumors. Using an efficient orthotopic xenografting procedure we obtain an overall tumor take-rate of close to 80%. We show that our glioma PDOX retain the genetic and epigenetic profiles of primary patient biopsies throughout several generations of xenotransplantation. In particular they not only faithfully recapitulate gene amplification and expression of EGFR and EGFRvIII variant in a reproducible manner, also amplification and expression of rarer patient-specific EGFR variants is maintained.

Overall genome-wide transcriptomic profiles of PDOXs remain very similar to patient biopsies and correlate better with the GBM cohort of TCGA (538 GBM samples) than conventional cell lines. Observed differences at the transcriptomic level are largely based on the replacement of human to mouse stromal cells, which impacts on the molecular sub-classification of GBM. We conclude that glioma PDOX models reflect patient heterogeneity and treatment response, and represent appropriate avatars for reproducible pre-clinical trials.

Furthermore, by combining profiling of the somatic mutational landscape with large-scale drug screening, PDOX-derived tumor organoids can elucidate druggable targets and tumor response profiles in a personalized patient-specific manner.

Kralova Viziova P., Indrova M., Spoutil F., Michalcikova T., Makovicky P., Kucera J., Kucera L., Prochazka J., Sedlacek R.

Institute of Molecular Genetics of the ASCR, v. v. i., Czech Centre for Phenogenomics, Videnska 1083, Prague, Czech Republic

Introduction

The aim of this study was to verify tumor cell line viability in situ and investigate if carcinoma cells from small piece of tumor adherent to colon adventicia will infiltrate into the deeper structures of large intestine.

Material and Methods

Tumor tissue was obtained from donor - NSG mice (subcutaneously growing tumor HCT 116 (HCT 116-Red Fluc, day 43)). Two pieces 2 mm3 of tumor tissue were attached to the colon wall of the acceptor NSG mouse under the general anesthesia. The first allocation for transplantation was chosen on colorectal adventitia and second between ceacum and colon ascendens. The attachment of tumor pieces was done by combination of fibrin tissue glue and stitches.

Growth of tumors was monitored weekly using luminescence imaging. At the day 17, the tumors were isolated and sampled for histology, Maldi Imaging analysis and Micro Computer tomography.

Results and Discussion

The mouse recovery run quickly and without any issues due to skin subdermal stitches, which would prevent unwanted licking of wound. Carcinoma cells from the tumor placed between caecum and colon ascendens infiltrated large intestine serosa. Metastasis of poorly differentiated carcinoma in liver and in vesiculae seminales were observed for the cell line HCT 116 and set up of experiment. The results for the tumor on colorectal adventitia are less conclusive and have to be verified in additional models.

Conclusion

We have observed infiltration of the tumor cells from the transplanted tumor tissue into deeper structure of large intestine and corresponding metastasis have been developed.

Elodie Montaudon¹, Laura Sourd¹, Ahmed Dahmani¹, Ludivine Morriset¹, Fariba Nemati¹, Sophie Chateau-Joubert², Rania El Botty¹, Guillaume Dutretre¹, Paul Cottu¹, Ivan Bieche¹, Elisabetta Marangoni¹

¹ Institut Curie, Paris, France

² Alfort Veterinary School, Maisons-Alfort, France

Introduction

Bone is a common site for metastasis formation in patients with advanced breast cancer. Bone metastases are considered incurable and the mechanisms underlying their formation remain to be elucidated in order to identity new therapeutic targets. Unfortunately, efforts towards this have been slowed down by the lack of relevant preclinical models. Moreover, obtaining human bone biopsies is technically challenging as bone remains one of the more difficult areas to biopsy.

The aim of this study was to establish Patient-Derived Xenografts (PDX) from hormone-resistant breast cancer bone metastasis suitable for identification of new drug combinations.

Material and Methods

One hundred eleven biopsies of breast cancer spine bone metastases, obtained from patients surgically treated to relieve pain and restore neurological functions, were implanted into the fat pad of swiss nude mice. PDX were morphologically and genetically characterized using histopathology (HES staining), immunohistochemistry, high-resolution array CGH and a targeted NGS of 90 genes (the most frequently mutated in breast cancers). A transcriptomic analysis was performed on primary tumors and bone metastases derived PDX. The in vivo response to hormone-therapies and targeted therapies was evaluated and PDX carrying targetable genomic alterations were treated by specific inhibitors.

Results and Discussion

Ten stable PDX have been established (engraftment rate 15%), 8 were ER+ and 2 triple-negative. The histology and the IHC profiles were conserved in PDX. Targeted NGS analysis of PDX identified targetable genomic alterations including mutations in AKT1, PI3KCA, BRCA2. aCGH and IHC analysis identified focal amplifications of FGFR1 and FGFR2 associated with strong protein over-expression. Interestingly, several models display RB gene deletion associated with RB protein loss and are predicted to be resistant to the recently approved CDK4/6 inhibitor palbociclib.

The transcriptomic analysis of bone metastases derived PDX as compared to primary breast tumors identified up-regulation of several targetable genes involved in proliferation and more specially in the mitotic phase of cell cycle (such as PLK1, TKK and AURKA). In vivo efficacy experiments are currently ongoing in PDX and show a difference of sensitivity to hormone-therapies between the different models. Combination of PI3KCA inhibitors and hormone-therapy was highly efficient and greater than palbociclib associated with hormone-therapy combination in PI3KCA mutated PDX. Drug combination including FGFR1, AKT1, PARP and PLK1 inhibitors are being tested and results will be presented at the meeting.

Conclusion

We have developed for the first time a panel of human breast cancer bone metastasis PDX that provides precious tools to assess innovative therapeutic approaches.

Sofia Karkampouna¹, Joël Grosjean¹, Irena Klima¹, Andrea Sboner², Charlotte K. Y. Ng^3,4, Salvatore Piscuoglio⁴, Martin Spahn⁵, Mark A. Rubin⁶, George N. Thalmann^1,7, Marianna Kruithof-de Julio^1,7

¹ Department for BioMedical Research, Urology Research Laboratory, University of Bern, Bern, Switzerland

² Department of Pathology and Laboratory Medicine, Institute for Computational Biomedicine, Englander Institute for Precision Medicine, Weill Cornell Medicine, New York, USA

³ Department of Biomedicine, University Hospital Basel, University of Basel, Basel, Switzerland

⁴ Institute of Pathology, University Hospital Basel, University of Basel, Basel, Switzerland

⁵ Centre for Urology Hirslanden clinic, Zürich, Switzerland

⁶ Department for BioMedical Research, Precision Oncology Laboratory, University of Bern, Bern, Switzerland

⁷ Department of Urology, Inselspital, Bern University Hospital, Bern, Switzerland

Introduction

Understanding tumor initiation and metastatic processes in prostate cancer (PCa) is crucial for prognosis of high risk patient groups and requires well-characterized experimental models that closely mirrors the stages of the human disease. We aim to establish patient derived xenograft (PDX) and other preclinical models (e.g. human organoids and ex vivo culture of human tumor tissue) to provide insights into tumor growth and metastasis and response to therapy.

Material and Methods

Patient material was obtained from radical prostatectomy or metastasis resection procedures. Our standard operating protocol includes: (i) PDX establishment by subcutaneous implantation of needle biopsies, (ii) ex vivo maintenance of PDX derived tissue in culture, and (iii) organoid derivation from resected or PDX tumor tissue. The response of ex vivo tumor tissue following 5-day drug exposure was assessed by histological evaluation of morphology, viability and apoptosis. We tested drug responses of organoids to standard-of-care compounds by Viability CellTiter Glo 3D assay.

Results and Discussion

We have established a PDX model from penile metastasis biopsies derived from a primary PCa (PNPCa). Tumor growth kinetic (~8 weeks) and luminal epithelial morphology is maintained through the passages of the PDX and compared to the resected tumor morphology. This model is classified androgen-sensitive as the PDX tumors regressed upon castration and re-grew after testosterone administration.

Drug screens on PDX tumor specimens cultured ex vivo showed that docetaxel had cytotoxic effects, while enzalutamide induced downregulation of luminal markers and reduced nuclear translocation of androgen receptor. We have also established PDX PNPCa-derived organoids and are currently assessing their molecular and genetic composition versus the originating primary PCa tissue. PDX PNPCa organoids are insensitive to hormone inhibitors, suggesting that they may represent an androgen-resistant, cancer stem cell population.

Conclusion

We have successfully established a novel hormone-sensitive PCa PDX model which is suitable to provide further understanding of the transition from androgen-dependency to androgen-independency, therapy response and resistance mechanisms.

Roberta Sommaggio¹, Elisa Cappuzzello², Anna Dalla Pietಠand Antonio Rosato^1,2

¹ Veneto Institute of Oncology IOV – IRCCS, Padua, Italy

² Department of Surgery, Oncology and Gastroenterology, Immunology and Oncology Section, University of Padua, Padua, Italy

Introduction

Cytokine-Induced Killer (CIK) cells are an attractive approach for cellular immunotherapy, as they are capable of recognizing tumor cells without the need of antigen-specific priming and can be efficiently and rapidly expanded in vitro. We previously demonstrated that CIK cells can exert ADCC when combined with clinical-grade monoclonal antibodies. In this study, we evaluated the efficacy of this combined therapy in vivo in triple negative breast cancer (TNBC), an aggressive tumor that occurs in approximately 15% of all breast cancer patients. Women with TNBC cannot be treated with hormonal treatments or molecular targeted therapies, as they lack the appropriate targets for these drugs, and thus new therapeutic options are urgently needed.

Material and Methods

Different breast cancer mouse models were established in NSG mice, injecting MDA-MB-231 cells either intravenously (i.v.) (0.1x10⁶ cells) or in mammary fat pad (10⁶ cells), or implanting patient-derived tumor Xenografts (PDX) in mice mammary fat pad. The combined treatment consisted in i.v. or intratumoral (i.t.) injection of 1.5 mg Cetuximab (CTX) together with 10x10⁶ CIK cells, while CTX or CIK cells administered individually served as control treatments. Tumor growth and metastases were monitored by bioluminescence or immunohistochemistry.

Results and Discussion

CTX induced a significant enhancement of CIK cell cytotoxicity in vitro against breast cancer cell lines, as previously reported [Cappuzzello E., et al, 2016]; the efficacy of this combined therapy was then assessed in vivo. Both in mice injected i.v. with MDA-MB-231 receiving i.v. therapy, and in mice with fat pad PDX receiving i.t. or i.v. therapy, the combined therapy induced a significant delay in tumor growth, compared to the control treatments. To better simulate a metastatic human cancerous disease, an orthotropic model was established injecting MDA-MB-231 cells in mice mammary fat pad and treating the mice with i.v. injections; also in this setting, the combined therapy was more efficient than control treatments in inhibiting tumor growth in the primary site.

Remarkably, the combined treatment almost completely abolished metastatic spreading to the lungs. We confirmed this relevant finding in a mouse model where the primary tumor was surgically removed before the i.v. treatment, as it occurs for the treatment of patients. The growth of metastases in the lungs and lymph nodes was significantly delayed by the treatment of CIK cells with CTX compared to the control group, and mice had a significant better survival.

Conclusion

These data confirm CIK cells combined with tumor-specific antibodies as a promising treatment for TNBC metastases. Moreover, the combination of CIK cells and specific monoclonal antibodies represents a new perspective for adoptive immunotherapy to treat solid tumors.

Rita Cascão¹, Carlos Custódia¹, Eunice Paisana¹, Tânia Carvalho¹, Pedro Pereira², José Pimentel², João T. Barata¹ and Claudia C. Faria^1,3

¹ Instituto de Medicina Molecular João Lobo Antunes, Faculdade de Medicina, Universidade de Lisboa, Lisboa, Portugal

² Laboratory of Neuropathology, Neurology Department, Hospital de Santa Maria, Centro Hospitalar Lisboa Norte, Lisboa, Portugal

³ Department of Neurosurgery, Hospital de Santa Maria, Centro Hospitalar Lisboa Norte, Lisboa, Portugal

Introduction

Brain metastases (BMs) affect approximately 40% of patients with any given cancer. Despite current available treatments, including surgery and radiation therapy, BMs are incurable and patients have a dismal outcome. There is an urgent need for personalized therapeutic approaches. Therefore, the development of appropriate animal models of human BMs is crucial to validate novel treatment strategies.

Material and Methods

Thirty-five BMs with different primary origins were collected during surgery in the Department of Neurosurgery at Hospital de Santa Maria (Lisbon, Portugal). Immunocompromised mice were used to develop patient-derived xenograft (PDX) models of BMs. Human samples were implanted subcutaneously in the flank of NSG mice, and serially passaged. Engrafted tumors were harvested, dissociated into a single cell suspension and injected in the heart of NSG mice.

Animals were euthanized upon reaching the humane endpoint and the central nervous system (CNS) along with other organs were collected for histological analysis, to assess the pattern of cancer cell dissemination. Patient derived cultures were also established and used for drug testing. Response to drugs was assessed through proliferation assays and western blot.

Results and Discussion

The overall take rate in the subcutaneous model was 66% and it correlated with patient survival (p=0.0032). Among the 23 engrafted flank tumors, 61% (n=14) originated spontaneous metastases to diverse organs including the CNS (21%). Of notice, samples from lung BMs developed spontaneous metastases in the lung, suggesting “homing” of these tumor cells tothe primary organ. The intracardiac injection of tumor-isolated cells increased their metastatic potential, particularly to the CNS. Remarkably, mouse xenografts and matched patients shared at least two metastatic sites in 6 samples (46%).

Finally, we assessed the efficacy of targeted antineoplastic agents, currently in use in the clinic, in the established patient-derived cultures of BMs. Buparlisib (pan-PI3K inhibitor) and everolimus (mTOR inhibitor) induced specific pathway inhibition and successfully reduced cancer cell proliferation.

Conclusion

We conclude from our work that patient-derived models of human BMs mimic patient’s disease and can be a powerful tool in the development of novel targeted therapies to treat metastatic cancer.

Federica Marocchi¹, Jole Costanza^2,3, Fernando Palluzzi¹, Laura Riva^2,4, Luca Mazzarella¹, Pier Giuseppe Pelicci¹, Daniela Bossi^1,5, Luisa Lanfrancone¹

¹ Department of Experimental Oncology, European Institute of Oncology, Milan 20141, Italy

² Center for Genomic Science of IIT@SEMM, Fondazione Istituto Italiano di Tecnologia, Milan 20139, Italy

³ Bioinformatics and Computational Biology Unit - Fondazione IRCCS Ca' Granda - Ospedale Maggiore Policlinico, Milan 20122, Italy

⁴ Wellcome Trust Sanger Institute, Cambridge, United Kingdom

⁵ Institute of Oncology Research, Oncology Institute of Southern Switzerland, Bellinzona

Introduction

Metastatic melanoma exhibits increasing incidence rate, with highly aggressive and heterogeneous traits. Despite improvement in response rates and overall survival, the majority of patients experiences recurrence and resistance, as well as severe immunotherapy toxicities. In metastatic melanoma patient-derived xenografts (PDXs) we have previously shown that numerous and patient-specific chromatin modifiers are essential for tumor growth, even when not-mutated. In this setting, we aim to identify actionable genes and pathways, for which drugs are already available and that can be repurposed for melanoma treatment

Material and Methods

Double screen is carried out using a shRNAs library targeting 77 actionable genes. Transduced PDX cells are injected in NSG mice and tumors analyzed by NGS. Depleted genes are identified as candidate hits. In parallel, transduced cells are seeded in modified Boyden chambers, migrated/not-migrated populations collected and analyzed by NGS. A preliminary screen is performed with a library carrying non targeting barcodes (BCs) to assess feasibility. Actionable essential genes are then identified scoring their differential expression in the two populations. PDX metastatic potential is assessed by intradermal injection and primary tumor resection. Metastasis drop-out screen feasibility is tested counting BCs represented at metastatic sites.

Results and Discussion

The in vivo actionable screen was performed on two PDXs and identified both mutated and not-mutated genes as hits, showing heterogeneous vulnerabilities among patients, but also a pivotal role of common pathways. We are exploring combinatorial therapeutic testing in the two PDXs and perform mechanistic studies. In parallel, using a non-targeting, barcoded library, we proved that the entire molecular repertoire of BCs is recovered, with comparable relative representation of individual BCs in the migrated and not-migrated populations in an in vitro migration screen. Metastasis screen is a feasible approach in our system, since a sufficient number of BCs can be recovered at the metastatic sites, to fully represent the actionable library. Therefore, we performed screen and analysed lymph node metastases to identify actionable genes essential for the dissemination process.

Conclusion

We set up a powerful screen system that allows to rapidly convert tumor vulnerabilities into translational opportunities for patient care. Based on deregulated pathways identified, it is possible to test different combinatorial therapeutic strategies. These results will lead to the definition of molecular subtypes and to the identification of new therapeutic strategies in melanoma.

Anna Portela^1,#, Rosa Bosch¹, August Vidal^1,3 and Alberto Villanueva^1,2,4

¹ Xenopat S.L., Business Bioincubator, Bellvitge Health Science Campus, Barcelona, Spain

² Group of Chemoresistance and Predictive Factors, Program Against Cancer Therapeutic Resistance (ProCURE), Catalan Institute of Oncology (ICO-IDIBELL), L’Hospitalet de Llobregat, Barcelona, Spain

³ Department of Pathology, Hospital Universitari de Bellvitge (IDIBELL), CIBERONC, Barcelona, Spain

⁴ EurOPDX consortium

^# Correspondence: anna.portela@xenopat.com

Introduction

The main cause of cancer patient’s death is the appearance of metastases. Usually metastases develop as a consequence of acquired resistance to the different oncologic therapies. As a biological process, metastasis is quite complex as it reflects the diverse barriers that cancer cells that leave the primary tumor must overcome to generate aggressive secondary lesions. Nowadays, it is a priority to generate good tumor models where new therapeutic strategies to attack metastasis in different stages of development can be tested.

Aim

To generate advanced tumor models that mimic in mice tumor distal dissemination and to generate models directly from patient metastatic tissue.

Material and Methods

Xenopat, in close collaboration with ICO/IDIBELL, has developed two approaches for the generation of metastatic-related tumor models:

• Orthotopic implantation of primary human tumors for more than 12 different human tumors. Thus, an important collection of orthotopic models (300-350 different models named orthoxenografts®), including: colorectal, lung, ovarian, endometrial, cervix, pancreas, melanoma, breast, glioma, sarcoma, head and neck, testicular germ cell tumors, rare tumors (Fanconi Anemia, NUDD, ovarian hypercalcemic tumors, etc.), pediatric tumors, prostate and bladder tumors has been developed. Presence of synchronic metastases (micro or macro metastasis) to different organs has been identified in several orthoxenografts® for different tumor types, reproducing human dissemination patterns. To investigate metachronous metastases, the tumor was removed in different tumor types by surgery.
• Orthotopic implantation of human metastasis obtained both by surgery and radiological techniques. Using this approach, Xenopat has generated orthoxenografts® from metastases of liver, lung, brain or peritoneal implants by implanting them in the same organ of origin. Biopsies of lymph nodes have been implanted in the organ of origin of their primary tumor.

Results and Discussion

Both approaches to metastatic models are useful to evaluate both the effect of a drug in the dissemination process, as well as in the metastasis itself once settled. In addition, these models are very useful tools to study the mechanisms of drug resistance acquisition since they summarize the patient’s therapeutic process.

Conclusion

Orthoxenografts® derived from primary tumors or directly from metastatic lesions may be outstandingly preclinical tools not only to develop new oncologic drugs but also to assess personalized treatment of bad prognosis/metastatic patients.