Replacement of animals in cancer drug development by using 3D in vitro functional assays for increased predictive power

Before any drug can be used in patients, it must be tested in animals for safety and activity. However, in cancer in particular, many drugs fail to work in the clinic because the preclinical testing is inadequate. Brain tumours in both adults and children are particularly devastating because there are currently no curative treatments. We have therefore chosen to focus our attention on streamlining the testing of new drugs in brain cancers. Currently, drugs are initially tested in simple systems where cancer cells are grown in a single 2-dimensional (2D) layer on plastic plates, but this does not in any way match the complexity of tissues where the tumours develop in the body in 3-dimensional (3D) clusters and where drugs may not easily penetrate. However, there is scope to test drugs more effectively before they are tested in vivo. Our view is that the 2D culture systems are not sufficiently predictive of in vivo responses and so many animals are used for subsequent tests without legitimate validation of potential drugs. Our objective is to bridge the gap between the standard current methods and animal testing. We are developing a panel of reproducible, 3D sytems which will measure not only cancer growth, but also additional devastating aspects of cancer behaviour: their ability to invade tissues and to make the body provide them with a new blood supply ? a process called angiogenesis. It is this which not only feeds the cancer but also enables it to potentially spread to other parts of the body. We have engineered the tests so that we measure effects on almost 100 mini cancers on one small plate the size of a small paperback book. By developing a more sophisticated way of identifying compounds more likely to work in the body, we will be able to replace a significant proportion of the animals currently required to bring new drugs to the clinic.

Animal testing is mandatory for any drugs entering clinical trials. However, in cancer in particular there is a high attrition rate prior to clinical deployment. Simple 2-dimensional monolayer cultures of tumour cells are not sufficiently predictive of in vivo responses and thus many animals are used for subsequent tests without legitimate in vitro validation of drug candidates. Our objective is to bridge the gap between standard in vitro cell cultures and in vivo efficacy studies. We focus on malignant brain tumours because they are almost universally fatal in both adults and children. Using a well-characterised panel of human adult and paediatric gliomas and medulloblastomas, we will optimise 96-well plate 3D spheroid assays of growth, invasion and angiogenesis. We have already shown proof of principle of the assays and that they are amenable to high throughput drug evaluation. Spheroids are established in suspension culture and their growth kinetics measured over time by quantitative image analysis. Endpoint cell viability assays enable GI50 values to be determined for direct comparison with 2D cultures. By addition of matrigel, we convert the assay into an invasion assay which can be similarly quantified. Finally spheroids co-cultured with endothelial cell monolayers provide a model of perivascular invasion and confrontation cultures between embryoid bodies and tumour spheroids mimic angiogenesis. We will test in these models inhibitors of PI3Kinase and the HSP90 chaperone (validated targets in these diseases and where we have extensive preclinical and clinical experience). We will then compare results with those already obtained in 2D assays and in vivo as a benchmark for qualification of our proposed alternative in vitro 3D systems. Parallel mechanistic studies will evaluate the effects of hypoxia and stem cell populations (prevalent in 3D cultures but lacking in 2D systems) for their impact on drug responses. In addition, we will undertake detailed pharmacodynamic biomarker assays (e.g. phosphoproteins levels for PI3K inhibitors; client protein expression for HSP90 inhibitors) to understand the basis for differences in response in the three systems. By developing a more sophisticated triage system for experimental compounds we will be able to replace a significant proportion of the animals currently required to bring new drugs to the clinic. Although we will exemplify the assays in brain tumours, we have already shown that they are applicable to many other tumour types and hence will have wide applicability in academic and pharmaceutical drug discovery.