Abstract Despite the rarity of each individual cancer type, about 200 different rare cancers constitute in total about 20% of all cancer cases, including pediatric cancers. In Europe, nearly half a million people live with a rare cancer. Like other rare diseases, rare cancers are particular challenging due to their low incidence, particularly for the identification of novel therapies that could improve patient survival. In spite of being the predominant type of pediatric liver malignances, hepatoblastoma (HB), with a world-wide incidence of 1 case per million persons per year, is a rare tumor. Differently from adult hepatocellular carcinoma (HCC) that develops on a cirrhotic or chronically-infected background, liver tumors in children and adolescents occur on apparently normal liver. The high rate (> 60 %) of β-catenin activating mutations places HB as the human tumor most tightly associated with activation of the Wnt/β-catenin pathway. Evidence for (epi)genetic origin of HB is provided by its association with congenital anomalies, Beckwith-Wiedemann syndrome, and familial adenomatous polyposis, a disorder caused by germline mutation of APC, involved in β-catenin degradation. HCC, fibrolamellar carcinoma (FLC), and transitional liver cell tumors (TLCT), which combines histological features of HB and HCC, also arise in children and adolescents, at a lower extent though. Sporadically, very rare forms of liver tumor likely of non-epithelial origin such as rhabdoid tumor or hepatic sarcoma also occur. In order to assist medical decision on the management of liver cancer in childhood and adolescence, we have launched a program aimed at the constitution of liver cancer patient-derived xenografts (PDXs). At present, 8/24 HBs, 2/2 TLCTs, 0/2 FLCs, 1/1 rhabdoid tumor and 0/1 hepatic sarcoma have been successfully grown in immunocompromised mice. HB, TLCT and rhabdoid PDXs maintain the histological features of primary human tumors, and upon treatment with different chemotherapy agents, these models show unique profiles of response, indicating a tumor-specific sensitivity. Given the relatively high number of HB models obtained, HB PDXs could be used as a preclinical cohort for phase II-like studies. This would allow the pre-screen of therapeutic solutions that would require years when not decades to be put in place via standard clinical assays. Moreover, as several HB PDXs harbour activating mutation of β-catenin, they could serve as very powerful tools for the development of efficacious Wnt/β-catenin inhibitors. In addition, for sporadic liver tumors like TLCT and rhabdoid tumor, as the creation of a preclinical cohort is hard to propose, a comprehensive drug screening in vivo could orientate adjuvant therapy in view of a personalized treatment choice, or contribute to accumulate evidence on the usefulness of the tested drugs on such types of liver malignancies. Citation Format: Stefano Cairo, Aurore Gorse, Delphine Nicolle, Erwan Selingue, Frédéric Gauthier, Christophe Chardot, Elie Fadel, Dominique Elias, Daniel Orbach, Catherine Guettier, Arnaud Beurdeley, Vanessa Yvonnet, Olivier Deas, Monique Fabre, Laurence Brugières, Sophie Branchereau, Jean-Gabriel Judde. Liver cancer patient-derived xenografts to improve disease management in childhood and adolescence: perspectives and challenges of personalized medicine. [abstract]. In: Proceedings of the 104th Annual Meeting of the American Association for Cancer Research; 2013 Apr 6-10; Washington, DC. Philadelphia (PA): AACR; Cancer Res 2013;73(8 Suppl):Abstract nr 2797. doi:10.1158/1538-7445.AM2013-2797

Abstract Triple negative breast cancer (TNBC) is a tumor subtype characterized by the absence of overexpressed estrogen receptor-alpha (ER), progesterone receptor (PR), and HER2 receptor, encoded by ERBB2, a known proto-oncogene. This type of tumors account for approximately 15–25% of breast cancers at diagnosis, and is one of the most aggressive subtypes, with 77% of patients that live free of disease 5 years post-diagnosis. One of the most reliable predictive markers of patient outcome is the pathological complete response (pCR), which indicates that the surgical specimen removed after neoadjuvant chemotherapy contains no viable tumor cells detectable at histopathological level. For patients with pCR, the probability of surviving the disease is very high, however, pCR is observed only in about 20–30% of TNBC. On the other hand, for patients with no pCR the probablity of developing recurrent disease at 5 years is 50%. As pCR is strongly correlated with treatment efficacy, it is mandatory to develop methods that allow to tell as quick as possible if the treatment chosen for a given patient is efficiently working or if it should be abandoned in favor of an alternative strategy that could prove more efficacious. XenTech collection of breast cancer patient-derived xenografts (PDXs) includes 25 models of TNBC that display heterogeneous response to different chemotherapy agents. We used our models to investigate if transcriptional changes could be detected in PDXs that responded well to genotoxic agents. To do this we analyzed the gene expression profile of laser-microdissected residual tumor nodules interspersed in the murine stroma upon very efficient response to Adriamycin/Cyclophosphamide (AC). When doing so, we identified several genes of the IFN/STAT1 pathway that were over-expressed when compared to untreated tumors. This activation seems to be a transient event, as it was lost in tumors relapsing after the residual tumor nodule stage. The finding that residual cells from tumors strongly responding to AC treatment over-expressed IFN/STAT1 pathway-related genes prompted us to investigate whether this effect could be detected as an early event upon tumor exposure to chemotherapy. All TNBC models tested that were good responders to AC treatment displayed over-expression of IFN/STAT1 pathway-related genes as early as 3 days post-treatment, most of them reaching a plateau of intensity at day 7 post-treatment. By contrast, TNBC insensitive or low responders to AC treatment failed to show over-expression of IFN/STAT1 pathway-related genes. To verify if the selective over-expression of these genes in TNBC models sensitive to AC was independent of the treatment administered, Irinotecan and Capecitabine were used to treat TNBC models with heterogeneous response to these drugs. Again, we found that overexpression of IFN/STAT1 pathway-related genes was specifically identified at early stages only in TNBC models that responded well to these drugs. These results suggest that genes of the IFN/STAT1 pathway could be early predictors of tumor response in patients receiving neoadjuvant chemotherapy. Prospective clinical validation studies are warranted to confirm the findings from this preclinical study. Citation Information: Cancer Res 2012;72(24 Suppl):Abstract nr P3-06-24.

Abstract Despite considerable efforts in understanding the biology and genetics of cancer, most currently available treatments fail to achieve tumor eradication in the majority of patients. Key to more effective therapies is adequate disease classification and subsequent patient stratification. In addition, it is important to understand the mechanisms of drug-response or resistance and identify novel targets amenable to therapeutic intervention. It is increasingly recognized that at the preclinical stage, testing therapeutic strategies and validating target relevance in more predictive models closely mimicking clinical disease such as patient derived xenografts (PDXs), may translate into improved clinical efficacy and lower rate of drug attrition. XenTech collection of over 120 runing PDX models is one of the largest in the world. PDX models were established by grafting post-surgery human tumor fragments in the interscapular region of immunodeficient mice. These deeply characterized PDX models can be used for in vivo preclinical assays. Such preclinical platform is a reliable surrogate of patient cohorts and can address several aims: 1. Evaluate tumor response to treatment. PDXs can be subjected to parallel evaluation of tumor response to various treatment protocols. Drug-response profile is linked to tumor histotype and molecular features in order to identify predictive markers of drug response to assist treatment choice. 2. Assess treatment-driven tumor eradication. The ability of a treatment to induce complete tumor response is assessed by monitoring tumor regression over a long period. Most tumors, despite complete macroscopic regression, are still present as latent microscopic nodular islands that may give rise to tumor recurrence. Molecular characterization of tumor foci responsible for tumor relapse may be performed to identify genes/pathways involved in residual tumor cell survival, which may provide new diagnostic and/or therapeutic targets for designing novel adjuvant treatment strategies. 3. Development of bioluminescent metastatic models to study the mechanisms of tumor invasion and to test anti-metastatic therapy. 4. Non-invasive molecular imaging technology to monitor tumor metabolism, vascularization and apoptosis. 5. Constitution of preclinical panels of rare malignancies to obtain phase II-like tumor cohorts. Development of new therapies for rare tumors is rendered difficult by the unavailability of patient cohorts wide enough to set up robust clinical trials. To assist the clinical need, these panels would allow the evaluation of new and more efficient therapies. We describe here in detail our PDX collection and illustrate how it represents a powerful tool to identify preferential therapeutic options for patients by exploring and improving anti-cancer therapeutic strategies. Citation Format: {Authors}. {Abstract title} [abstract]. In: Proceedings of the 103rd Annual Meeting of the American Association for Cancer Research; 2012 Mar 31-Apr 4; Chicago, IL. Philadelphia (PA): AACR; Cancer Res 2012;72(8 Suppl):Abstract nr 5273. doi:1538-7445.AM2012-5273