MD Biosciences Blog

It's a question that comes up regularly, from scientists who contact us to publications to conference lectures. One of the challenges in preclinical drug discovery is how translatable is the preclinical data from animal studies to the human situation?

Rodent models of pain such as nerve injury models are important to understand the mechanisms that may contribute to human neuropathic pain. Imaging studies in human have identified cortical regions specifically involved in the subjective, conscious perception of pain. Although laboratory animals process painful stimuli using similar mechanisms and thresholds of awareness as humans, it is much harder to assess the subjective pain experienced by animals as they can not self-report. This has led researchers to rely on objective measures of pain-related behaviors such as evoked responses to noxious stimuli. Humans, however, are able to voice discomfort, which provides rapid and direct access to the subjective experience. 

If you work in drug discovery and development, you are well aware of the failure rate at clinical trials. Industry estimates are that clinical candidates have a 85-90% chance of failure during clinical trials, the most costly stage of evaluation. A report in Nature Biotechnology 32, 40–51 (2014) breaks this success/failure rate down between phases as well as the likelihood of approval from the start of clinical trials. For candidates that are suspended during clinical stages, 83% of these reported efficacy or safety as the reason for suspension. 

This is costly and time consuming for drug developers. So is there a way to increase the predictability from preclinical phases to clinical phases?  We have been evaluating this question for a number of years in our Research Group at MD Biosciences. Animal models used in preclinical development phases are pivotal for understanding mechanisms that contribute to human disease conditions and effective therapies. Rodent models are commonly employed due to their reproducibility and simplicity, however the predictability to the clinic is often times lacking.