Translational science in medicine: Implications for the pharmaceutical industry

Translational medicine has become a major tool for facilitating the transition of basic science results into clinical practice. Its main remit is the facilitation of the transition of drugs or medical devices/diagnostics between preclinical research and human applications. The major driver for development of this discipline in the pharmaceutical industry is the fact that the output of new drugs has been decreasing over past years, despite increased investment in research and development. The predictive value of preclinical and early clinical results has become a key issue as many drug candidates face late attrition because of adverse safety or insufficient efficacy data. It is likely this costly process would become more efficient if translation were properly addressed. Biomarkers describing disease-related processes and drug effects need to be developed to achieve this, and their predictive values must be assessed. A typical question is: 'If we measure an effect of the drug candidate X on Y in rats, dogs or mice, what will happen to the same parameter in humans?' Obviously, some biomarkers exist across all mammalian species (e.g. blood pressure) and are quite predictive, whereas others do not (e.g. rats, as opposed to humans, use corticosterone rather than cortisol as the main glucocorticoid). The important types of biomarkers are the serum-based markers (e.g. interleukin-6 for inflammatory processes) or imaging parameters, especially those obtained by magnetic resonance tomography. The particular importance of safety biomarkers is acknowledged, yet concerted actions to find and establish more sensitive, specific and predictive markers are rare. Early attempts to quantitatively assess the predictive value of biomarkers have been published and are under further development. However, the impact of attempts to improve translation has not become obvious and it is tempting to assume that the approach is based more on an empty phrase than on robust scientific methodology. This has to do with the fact that the standardisations and classifications of related processes, such as biomarker assessment or proof-of-principle study design are still lacking, and accompanying methodological research is as yet poor. It is still an area of 'gut feeling' assessments and guesswork resulting in the reluctance of regulatory authorities to rely more on surrogates or biomarkers in general. "Lost in translation" is a famous statement that describes an observation, but does not help to solve the problem. Only a structured, methodological, science-driven process will develop translational medicine into an efficient tool to better project clinical outcomes of drug interventions from preclinical and early clinical data. The direction is recognised, but there is still a long way to go. © 2006 Adis Data Information BV. All rights reserved.