Validating microfluidics-based personalized cancer therapy in mouse models

• Applicants: Professor Dr. Thorsten Cramer; Professor Dr. Christoph A. Merten; Professor Dr.-Ing. Julio Saez-Rodriguez
• Subject Area: Gastroenterology ; Bioinformatics and Theoretical Biology ; Hematology, Oncology
• Term: from 2018 to 2023
• Project identifier: Deutsche Forschungsgemeinschaft (DFG) - Project number 411368829

Cancers of the hepato-pancreatico-biliary (HPB) tract arecharacterized by a very poor prognosis. which is largely explained byrobust therapy resistance of HPB cancers. Even though theimpressive advances of recent years have resulted in a much betterunderstanding of the molecular pathogenesis of cancer in general andHPB tumours in particular, these developments have not yettranslated in an improved prognosis of patients with this malignancy.In principle, strategies aiming at therapy individualization are able tosignificantly improve clinical care. However, until today only fewresults from genetic analyses have successfully been translated intothe clinics, e.g. for the treatment of leukemias, breast and coloncancer. Phenotypical analyses allowing for sensitivity testing oftumour cells of individual patients against a wide battery ofantiproliferative substances (and combinations thereof) could be analternative to define personalized therapies, but they require cellnumbers that are normally not available from biopsies. In preliminarystudies our research groups could show that a newly developedmicrofluidics platform is able to predict therapy sensitivity ofestablished human cell lines and small biopsies from patients withpancreatic cancer. We used it to identify particularly potent drugcombinations whose synergistic effects had not been describedbefore. For example, we could show that a combined application ofPHT-427 and Selumetinib is highly efficient by blocking twoindependent pathways downstream of KRAS. Our platform allows forrapid and automated testing of systematic combinations ofantiproliferative drugs on patient tumour material, which is usuallyavailable only in very small amounts. Due to the low number of cellsper sample (ca. 100) we are able to analyse one to two orders ofmagnitude more treatment conditions than with existing methods. Ourplatform has additional unique advantages compared to existingtechnologies, such as the avoidance of cell cultivating steps, whichcan potentially confound drug effects. Furthermore, our method israpid (results are available within 48h) and cost effective (~150 eurosconsumables costs per patient sample). We now aim for a functionalin vivo validation of this approach, which has not been possible previously using human samples, exploiting different HPB tumourmouse models. In comprehensive experiments with expandedreadouts we will a) predict the optimal therapy ex vivo, b) performsignalling pathway modelling to understand resistance mechanismsand identify potentially new drug sensitizers and c) validate thepredicted best treatment option in vivo. We are convinced that thiswork will represent the last missing step to initiate clinical studies withour microfluidics platform.

• DFG Programme: Research Grants