PREDECT – Seeking an effective solution for the lab to trial conundrum (Guest blog)

The disparity between positive results of drug testing in the lab and negative results with the same drugs in clinical trials represents a significant dilemma for the pharmaceutical industry and healthcare in general. For example, a plethora of anticancer drugs brim with promise in the laboratory but deliver either a marginal or no effect in patients.

Enter the Innovative Medicines Initiative PREDECT, which, since 2010, has been moving towards a potential panacea for this situation. The PREDECT consortium is a collaborative effort between nine pharmaceutical companies, three biotechnology companies and nine academic laboratories across Europe and in the USA (

It has long been accepted that the problem stems from the use of laboratory models of cancer that fail to mimic the highly complex biology and heterogeneity of human cancer. PREDECT looks to square the circle by scrutinising models that better represent this complexity.

Brainstorming meetings have brought together, for the first time, people from competing companies that previously had not collaborated together. Their experience of drug discovery, and its challenges, was then shared with the academic and biotechnology groups.

Already, groups at the EPFL in Lausanne in Switzerland and the University of Tartu, Estonia have witnessed a milestone moment. Cathrin Brisken and her EPFL team created a breakthrough model of oestrogen positive breast cancer, published in Cancer Cell1.

One commentary describes the model as “a potential game-changer for breast cancer research” that could lead to new drugs2. This is because it highlights in particular the importance of the native endocrine tumour microenvironment in successfully establishing graft models of patient-derived tumours. What’s more, recently, the Erasmus University in the Netherlands published a tissue histopathology study of murine prostate cancer in PlosOne3, showing pronounced tumour heterogeneity – underscoring that all tumours are not the same.

Both studies highlight the complexity of tumour biology. Moreover, they also emphasise the importance of localised tissue environments in creating appropriate mouse models to mimic human cancer.

Moving along to the ex vivo laboratory bench. Traditionally, in this area, it has been quite a challenge to create models that represent the architecture and complexity of human cancer by using isolated cancer cells. PREDECT attacked this from three different angles.

The first involves a model generated at Boehringer-Ingelheim, developed together with a group of academic collaborators in Germany and Israel. This model is the subject of a publication in PlosOne4 where it highlights how cancer cells change their behaviour when grown alongside their surrounding connective tissue-derived cells – particularly where messages they send to the immune system are concerned.

The second angle culminated in an important publication in Biomaterials5, detailing the creation of a three-dimensional (3D) model of cancer, which was grown in a bioreactor to control its nutrient supply. This portrayed clearly the mimicry of the motile and invasive behaviour of breast cancer when grown close to connective tissue cells. Here Abbvie and Boehringer were able to harness the impressive expertise in miniature bioreactor technology of the biotech iBET in Oeiras, Portugal.

Finally, let’s shine a spotlight on a striking industrial-academic, collaborative PREDECT team, and its recent publication of a landmark study in Nature Scientific Reports6. The study showed that very thin slices of cancer tissue that capture the disease’s native microenvironment, could be used to study tumour complexity and heterogeneity. It has presented protocols for the preparation and cultivation of the slices, and reported on the limitations, on which we elaborate in ongoing work. AstraZeneca and Servier are now incorporating precision-cut slices into their battery of ex vivo tests.

What IMI’s PREDECT consortium has managed to do is to create a variety of more appropriate models of disease. These can now be incorporated into the panoply of tests and hurdles that novel drugs must overcome before they can be considered for clinical trial.

Over the next few decades, clinical results should be able to validate the new models – which, with any luck, will then lead to an increase in clinical successes. By sharing the challenges and frustrations of drug discovery between different companies and with adequate funding secured for academic groups from FP7 and matching EFPIA funds (IMI), PREDECT is bearing fruit. The harvest will continue and new collaborative seeds will be planted as PREDECT draws to a close in April this year.

Emmy Verschuren

Dr. Emmy Verschuren received her PhD from UCL in 2003, following graduate studies with Prof. Nic Jones at the Imperial...
Read Morechevron_right

John Hickman

Professor John Hickman trained in the UK in Pharmacy, has a PhD in organic chemistry and a Masters in Biochemistry....
Read Morechevron_right
  1. Sflomos, G. et al (2016) Cancer Cell 29: 407-422
  2. Haricharan, S. et al (2016) Cancer Cell 29: 249-250.
  3. Korsten, H. et al (2016) PloSOne 10: doi: 10.1371/journal.pone.0147500.
  4. Rudisch, A. et al (2015) PloSOne 10: doi: 10. 1371/journal.pone.0124283.
  5. Estrada, M. et al (2016) Biomaterials 78: 50-61.
  6. Davies, E.J. et al (2015) Nature Scientific Reports 5: doi: 10.1038/srep17187.