Predicting patient responses to anticancer drugs is a major challenge both at the drug development stage and during cancer treatment. Tumor explant culture platforms (TECPs) preserve the native tissue architecture and are well-suited for drug response assays. However, tissue longevity in these models is relatively low. Several methodologies have been developed to address this issue, although no study has compared their efficacy in a controlled fashion. We investigated the effect of two variables in TECPs, specifically, the tissue size and culture vessel on tissue survival using micro-dissected tumor tissue (MDT) and tissue slices which were cultured in microfluidic chips and plastic well plates. Tumor models were produced from ovarian and prostate cancer cell line xenografts and were matched in terms of the specimen, total volume of tissue, and respective volume of medium in each culture system. We examined morphology, viability, and hypoxia in the various tumor models. Our observations suggest that the viability and proliferative capacity of MDTs were not affected during the time course of the experiments. In contrast, tissue slices had reduced proliferation and showed increased cell death and hypoxia under both culture conditions. Tissue slices cultured in microfluidic devices had a lower degree of hypoxia compared to those in 96-well plates. Globally, our results show that tissue slices have lower survival rates compared to MDTs due to inherent diffusion limitations, and that microfluidic devices may decrease hypoxia in tumor models.

Anticancer drugs have the lowest success rate of approval in drug development programs. Thus, preclinical assays that closely predict the clinical responses to drugs are of utmost importance in both clinical oncology and pharmaceutical research. 3D tumour models preserve the tumoral architecture and are cost- and time-efficient. However, the short-term longevity, limited throughput, and limitations of live imaging of these models have so far driven researchers towards less realistic tumour models such as monolayer cell cultures. Here, we present an open-space microfluidic drug screening platform that enables the formation, culture, and multiplexed delivery of several reagents to various 3D tumour models, namely cancer cell line spheroids and ex vivo primary tumour fragments. Our platform utilizes a microfluidic pixelated chemical display that creates isolated adjacent flow sub-units of reagents, which we refer to as fluidic ‘pixels’, over tumour models in a contact-free fashion. Up to nine different treatment conditions can be tested over 144 samples in a single experiment. We provide a proof-of-concept application by staining fixed and live tumour models with multiple cellular dyes. Furthermore, we demonstrate that the response of the tumour models to biological stimuli can be assessed using the platform. Upscaling the microfluidic platform to larger areas can lead to higher throughputs, and thus will have a significant impact on developing treatments for cancer.

Anti-cancer drugs have the lowest success rate of approval in drug development programs. Thus, preclinical assays that closely predict the clinical responses to drugs are of utmost importance in both clinical oncology and pharmaceutical research. 3D tumour models preserve the tumoural architecture and are cost-, labour-, and time-efficient. However, the short-term longevity, limited throughput, and limitations to live imaging of these models have so far driven researchers towards simpler, less realistic tumour models such as monolayer cell cultures. Here, we present a static open-space microfluidic drug screening platform that enables the formation, culture, and multiplexed delivery of several reagents to various 3D tumour models, namely cancer cell line spheroids and ex vivo primary tumour fragments. Our platform utilizes an open-space microfluidic technology, a pixelated chemical display, which creates fluidic “pixels” of biochemical reagents that stream over tumour models in a contact-free fashion. Up to 9 different treatment conditions can be tested over 144 samples in a single experiment. We provide a proof-of-concept application by staining fixed and live tumour models with multiple cellular dyes. Furthermore, we demonstrate that the various responses of the tumour models to biological stimuli can be assessed using the proposed drug screening platform. The platform is amenable to various 3D tumour models, such as tumour organoids. Upscaling of the microfluidic platform to larger areas can lead to higher throughputs, and thus will have a significant impact on developing treatments for cancer.