Organoid and single cell models of bladder cancer

Project background

Recent advances in bladder cancer research have identified and defined subtypes of bladder cancer based on the presence of specific molecular markers. In light of this discovery, it has become apparent that current disease models do not represent the full range of bladder cancer subtypes. Cells can be isolated from patient tumour tissue and used to generate patient-derived xenograft (PDX) models in immunocompromised mice. PDX mice can then be used to provide a comprehensive overview of the sensitivity of the tumour and the tumour microenvironment to cancer therapeutics. However, the available tumour cell lines and the majority of PDX models are biased towards more aggressive tumour subtypes. As such, new models are required for both functional investigations and preclinical drug assessments.

Why we funded it

This studentship aims to replace the need for PDX mice in the study of bladder cancer with the development of a bank of in vitro models which represent the full range of bladder cancer subtypes defined.

297 publications using PDX mice in urinary bladder studies have been published, with 112 of these published since 2014. These publications typically required a total of ~60 animals per study, with 4-7 animals used per condition, resulting in ~3,360 animals needed per annum. Professor Knowles and colleagues estimate 75% of such experiments could be replaced with the organoids developed in this studentship. These estimations exclude animal use in industry.

Research methods

Human cells derived from tumours, or normal tissues, can be maintained in in vitro 3D organoids and banks of organoids have previously been established for colorectal and prostate cancers. This studentship aims to develop culture conditions suitable for long-term in vitro organoid culture of patient-derived bladder cancer tissues, and create a bank of organoids representing the full range of molecular subtypes now known to exist. Cancer subtypes show variable sensitivity to therapeutics but PDX models are not suitable for rapid drug sensitivity assessment for individual patients. After the bank of organoids has been established they will be used in combination with a novel microfluidics platform for drug sensitivity studies. Initial experiments will look to demonstrate comparability to pre-existing PDX models before the organoids are used for rapid evaluation of patient samples in drug sensitivity studies.

Recent genome sequencing and expression analyses have defined multiple molecular subtypes of bladder cancer. Available cell lines that are currently used for functional studies of specific genes and for preclinical drug development have limited representation of these subtypes.

Several groups and commercial providers are developing panels of patient-derived xenograft (PDX) models, which provide more faithful representation of human bladder tumours than cell line models. However these use many animals, are costly, time-consuming and not suitable for high throughput or rapid assessment of drug sensitivity for individual patients. The urgent need for new therapies for bladder cancer combined with increased molecular understanding has led to a rapid increase in animal use despite the limitations of these models. Of a total of 297 publications on bladder cancer that have used cell line or PDX models, 112 were published since 2014.

Our aim is to develop a bank of tumour-derived organoids that can replace these models and ultimately provide faithful representation of all molecular subtypes of bladder cancers, including rare variants. In parallel, we will develop microfluidics-based platforms that can assay drug sensitivity at high throughput using these models. We will use existing PDX-derived tissues with known drug sensitivities to validate this approach. 

These models will be suitable to replace animal use for an estimated 75% of animal studies on bladder cancer that are aimed at understanding specific gene function or development and testing of drugs, particularly in the pharmaceutical industry. Such models have low cost, are highly scalable and provide rapid read-outs. We also hypothesise that single cells or organoids derived directly from clinical samples will allow immediate personalised assessment of drug sensitivity.