Preview

PDF, English - main document Download (31MB) | Terms of use

Abstract

The therapy of most cancers has greatly benefited from the use of targeted drugs. However, their effects are often short-lived since many tumors develop resistance against these drugs. Resistance of tumor cells against drugs can be adaptive or acquired and is often caused by genetic or non-genetic heterogeneity between tumor cells. A potential solution to overcome drug resistance is the use of drug combinations addressing multiple targets at once. Finding potent drug combinations against heterogeneous tumors is challenging. One reason is the high number of possible combinations. Another reason is the possibility of inter-patient heterogeneity in drug responses, making patient tailored treatments necessary. These require screens on patient material, which would drastically benefit from miniaturization, as it is the case in droplet-based microfluidics. However, drug screens in droplets against primary tumor cells have so far only been performed at a modest chemical complexity (55 treatment conditions) and with low content readouts. In this thesis we aimed at developing a droplet-based microfluidic workflow that allows the generation of high numbers of drug combinations in picolitre-sized droplets and their multiplexed analysis. To this end, we have established a pipeline to produce up to 420 drug combinations in droplets. We were able to significantly increase the number of possible combinations by building a microfluidic setup that comprises valve and micro-titer plate based injection of drugs into microfluidic devices for droplet generation Furthermore, we integrated a DNA-based barcoding approach to encode each treatment condition, enabling their multiplexed analyses since all droplets can be stored and processed together, which highly increases the throughput. With the established approach we can perform barcoding of each cells’ transcriptome according to the drugs it was exposed to in the droplet. Thereby, the effects of drug combinations on gene expression can be studied in a highly multiplexed way using RNA-Sequencing. We applied the developed approach to run combinatorial drug screens in droplets and analysed the effects of in total 630 drug combinations on gene expression in K562 cells. The low number of cells needed (max. 2 million cells) for such screens, could enable their application directly on tumor biopsies, thus paving the way for personalized therapy approaches. Since the established workflow is compatible with single cell readouts, we also envision its application to analyse drug resistances in heterogeneous tumor samples on the single cell level.