Advances in in vitro/ex vivo platforms for drug screening 13:30 - 15:25

Sander Basten¹, Nataliia Beztsinna¹, Bram Herpers¹, Mariusz Madej¹, Jara Garcia¹, Niels Meesters¹, Kuan Yan¹, Emma Spanjaard¹, and Leo Price¹

Crown Bioscience Netherlands, Leiden, The Netherlands

Extensive application of (immune)therapy to treat cancer patients has been hampered by low response rates in the clinic and high failure rates of clinical trials. One of the reasons is a lack of translational preclinical models that accurately replicate human disease complexity, leading to ineffective candidates entering clinical trials. The choice in advanced pre-clinical in vitro models that use cultures of patient-derived tissues is increasing, but their full potential in drug discovery has not been fully exploited. Assays often lack sensitivity or clinically relevant end-points or are not suitable for high throughput screening.

Here we present a 3D screening platform that combines complex biology with high content image (HCI)-based analysis to provide visualization and quantification of the effects of various targeted therapies on material derived from cancer tissues. Our proprietary, automated, high content 3D image analysis enables sensitive detection of treatment-induced and compound-specific morphological changes beyond conventional cell viability measurements, including those associated with tumor killing, cell cycle arrest, toxicity, and immune cells proliferation and tumor cell infiltration.

We present data of patient derived organoid cultures that are scalable and well controllable for larger throughput screening. In addition, we show data from a novel 3D Ex vivo Patient Tissue platform. Here, fresh patient tumor tissues are processed within 24 hours to preserve the tumor microenvironment (TME) and exposed to (immune-modulating) panels of drugs for 5-10 days prior to fixation and imaging. The image-based readouts can be complemented with extensive tissue characterization, including detection of secreted factors in supernatants (e.g. cytokines, chemokines), NGS based analyses, histology assessment and biomarker validation via IHC or FACS analysis.

Our approach offers a rapid, reliable and patient-relevant approach to test various clinically relevant (immune)therapies for different solid tumour types. It has the potential to significantly improve the preclinical evaluation of therapies and support the decision-making process during progression of drug candidates to the clinic.

Reinvigoration of tumor-specific T cells by cancer immunotherapies, in particular PD-1/PD-L1 blocking agents, has been the most important innovation in the treatment of patients with cancer. Nevertheless, durable clinical benefit is currently limited to a small number of patients. At present, the immunological alterations that occur in human cancers upon PD-1 blockade are not well understood. To investigate this, human ex vivo models are necessary that maintain tumor immune composition and architecture outside of the patient. By developing and using such patient-derived organotypic tumor models, we investigate how intratumoral and interpatient heterogeneity in immune infiltrates influence immunotherapy response and how distinct treatments can change immune activity in a tumor. The observed therapy-induced changes can then be linked to the inherent qualities of a tumor, thereby contributing to the identification of determinants for effective response to immunotherapy and to the development of novel treatment strategies.

Sanjiban Chakrabarty^1,7, William. F. Quiros-Solano^3,5 , Maayke Kuijten^1,2, Ben Haspels¹, Nikolas Gaio⁵, Jos Jonkers⁴, Ronald Dekker^3,6, Nitika Taneja¹, Roland Kanaar^1,2, Dik C. van Gent^1,2

¹ Department of Molecular Genetics, Erasmus MC, Rotterdam, The Netherlands
² Oncode Institute
³ Department of Microelectronics, Electronic Components, Technology and Materials, Delft University of Technology, Delft, The Netherlands
⁴ Mouse Clinic Intervention Unit, The Netherlands Cancer Institute, Amsterdam, The Netherlands
⁵ BIOND Solutions B.V., Delft, The Netherlands
⁶ Philips Research, Eindhoven, The Netherlands
⁷ Department of Cell and Molecular Biology, Manipal Academy of Higher Education, Manipal, India

Introduction

Breast Cancer (BrC) is the most common cancer in women worldwide. Chemotherapy remains a cornerstone treatment for a majority of BrC patients. Unfortunately, the response to these treatments is variable and biomarkers for therapy response are not sufficient to correctly predict patient responses to therapy. Therefore, better models to accurately predict chemotherapy response in BrC patients are dearly needed. For this purpose, we developed a novel microfluidic platform for chemotherapy testing, which we call a Cancer-on-chip (CoC) platform.

Materials and Methods

Patient Derived Xenographt (PDX) tumors were sliced into 300 µM slices using a Leica VT1200S vibratome. Tumor slices were cultured in a newly developed CoC platform, consisting of a microfluidic chip with microfluidic channels embedded in polydimethylsiloxane (PDMS). The tissue slice is placed such that flow is both on top and bottom, while the tissue slice is embedded in hydrogel. Media perfusion was provided through a pumping system with controlled flow. To assess cell proliferation EdU (3 µg/ml) was added to tissue slices 2 hours prior to fixation. Tissue slices were then Formalin-Fixed Paraffin-Embedded (PPFE), after which 4 µm sections were generated for immunostaining analysis.

Results and Discussion

We showed that CoC offers several advantages compared to the previously used ex vivo culture method. The response after cisplatin treatment in cisplatin-sensitive PDX tissue slices was stronger in CoC cultured slices comparing to the ex vivo culture, suggesting a better drug delivery in the CoC system. After 14 days of culture without treatment in the CoC device, more proliferation and less apoptosis was observed than in the ex vivo culture. Analogously, fewer transcriptional changes and less DNA damage was detected after CoC culture compared to the ex vivo cultured slices. We succeeded to run up to 90 experiments with tissue slices from one single PDX tumor.

Conclusions and Future Perspectives

The CoC is a more optimal culture method for organotypic BrC tumor tissue slices compared to the previously used ex vivo culture. In the future, the chemotherapy response assessment will be extended to other treatment such as: carboplatin, docetaxel, paclitaxel and eribulin. The CoC platform will be further optimized by e.g. introduction of endothelial cells to cover the microfluidic channel to mimic vasculature and considering different real time imaging and/or culture media analysis options.

Akira Inoue^1,2, John L. Rose¹, Sanjana Srinivasan¹, Rosalba Minelli¹, Fred S. Robinson¹, Giulio F. Draetta¹, Timothy P. Heffernan¹, Giannicola Genovese¹, Alessandro Carugo^1,3

¹ University of Texas MD Anderson Cancer Center, Houston, USA
² Osaka General Medical Center, Osaka, Japan
³ Istituto di Ricerche di Biologia Molecolare (IRBM), Rome, Italy

Introduction

As a scientific community we value animal studies for their high-predictive validity, but we are well aware of the low-throughput they can guarantee due to technical and logistic limitations. Therefore, it is critical we will rapidly reach an inflection point where relevant patient-derived models such as PDXs can converge into an organic discovery platform together with methodologies for comprehensive interrogation of gene-by-gene phenotypes and high-throughput drug responses.

Material and Methods

Here, we present PILOT (Patient-Based In Vivo Lethality to Optimize Treatment), a platform that enables the triangulation of patient-derived models with RNAi/CRISPR-based functional genomics and pharmacological drug screening. First, PILOT was deployed to identify clinically actionable targets for rapid drug repurposing in colorectal cancer (CRC) as we conducted a patient-centred RNAi screen paired with a high-throughput drug screen in four patient-derived xenografts. Second, by performing a CRISPR/Cas9 screen in a representative subset of pancreatic ductal adenocarcinoma PDX models we leveraged the PILOT platform to establish context specificity of individual gene targets within PDAC molecular subtypes.

Results and Discussion

Integration of RNAi and drug screening methods identified exportin 1 (XPO1) inhibitors as drivers of DNA damage-induced lethality in CRC. Molecular characterization of the cellular response to XPO1 inhibition uncovered an adaptive mechanism that limited the duration of response in TP53-mutated, but not in TP53-wild-type CRC models. These findings supported the clinical repositioning of XPO1 inhibitors, especially in combination with DNA damage checkpoint inhibitors, to elicit an enduring response in patients with CRC harbouring TP53 mutations. Similarly, by merging a PDX-based co-expression network with in vivo CRISPR screening we identified ILK, SMAD4, and ZEB1 as functional determinants of the highly aggressive basal-like subtype of PDAC. Then, by combining CRISPR knockout with single-cell transcriptomics we directly validated these three genetic dependencies and observed a significant clonal shift toward the classical-like signature of more indolent tumors. Our work enabled the identification of potential therapeutic targets to selectively eradicate dominant sub-clonal populations and possibly allowing to utilize changes in transcriptional heterogeneity as a measure of tumor response.

Conclusion

The biological and clinical insights highlighted by our PILOT platform are just scratching the surface of the potential that can be unleashed by blending together relevance of the pre-clinical models, advancements in functional genomics and compounds with pharmacological activity.

Annelies Van Hemelryk¹, Sigrun Erkens-Schulze¹, Lifani Lim¹, Corrina M. A. de Ridder¹, Debra C. Stuurman¹, Pim J. French^2,3, Martin E. van Royen⁴, Wytske M. van Weerden¹

¹ Department of Urology, Erasmus MC Cancer Institute, University Medical Center, Dr. Molewaterplein 40, 3015 GD Rotterdam, The Netherlands
² Cancer Treatment Screening Facility, Erasmus MC Cancer Institute, University Medical Center, Dr. Molewaterplein 40, 3015 GD Rotterdam, The Netherlands
³ Department of Neurology, Erasmus MC Cancer Institute, University Medical Center, Dr. Molewaterplein 40, 3015 GD Rotterdam, The Netherlands
⁴ Department of Pathology, Erasmus MC Cancer Institute, University Medical Center, Dr. Molewaterplein 40, 3015 GD Rotterdam, The Netherlands

Introduction

Organoid-based studies have revolutionized in vitro preclinical research and hold great promise for the cancer research field, including prostate cancer (PCa). However, standardized drug testing procedures are generally lacking, introducing experimental variability that complicates reproducibility, both within and between institutions. Moreover, common viability assays are restricted to endpoint measurements and viability as the sole read-out, losing important information on key biological aspects that might be exclusive to 3D organoids and that may be impacted by the administered drugs. In this study, we aimed to standardize a viability-based drug testing procedure for organoids derived from PCa patient-derived xenografts (PDXs). In addition, we developed and optimized an image-based live/dead assay using high-content fluorescence live-cell imaging in living PCa organoids.

Material and Methods

To initiate PCa PDX-derived organoids (PDXOs), PDX tumors were dissected and digested enzymatically, passed through a 100 µm cell strainer and resuspended in a synthetic, thermo-sensitive, hydrogel. Organoids were allowed to assemble in 24-well plates and harvested after 7 to 14 days by dissolving the hydrogel dome. Organoids were plated at a density of 5000 – 10000 organoids per 8 µL hydrogel domes in 96-well plates. Organoids were incubated with dose ranges of standard PCa chemotherapeutics (docetaxel and cabazitaxel) and of the anti-androgen enzalutamide. Organoid viability was measured with CellTiter-Glo (CTG) 3D (Promega). An optimized fluorescent dye combination of caspase 3/7, propidium iodide and Hoechst 33342 was used for imaging experiments. Confocal time-lapse imaging was performed with the Opera Phenix High Content Screening System (PerkinElmer). Staurosporine treatment was used as positive control, solvents as negative controls in both CTG assays and image-based drug tests.

Results and Discussion

We developed a standardized protocol for performing CTG-based drug tests in PCa PDXOs and found 3 factors to be crucial for reliable interpretation of viability assays in these PDXOs: presence of ROCK-inhibitor in organoid medium, organoid proliferation rate and treatment duration. Moreover, we generated an image-based live/dead assay and implemented a custom image analysis method to quantify the number of organoids, organoid sizes and the number of apoptotic, necrotic and viable cells. This image-based procedure provides valuable insights on treatment effects in living organoids.

Conclusion

We developed two protocols for PCa PDXO-based drug testing and provide additional read-outs with our image-based procedure. Both procedures might well be customized for PDX- and patient-derived organoids from various cancer types.