The pharmaceutical industry urgently needs reliable pre-clinical models to evaluate the efficacy and safety of new chemical entities before they enter the clinical trials. Development of in vitro model systems that emulate the functions of the human liver organ has been an elusive task. Cell lines exhibit a low drug-metabolizing capacity and primary liver cells rapidly dedifferentiate in culture, which restrict their usefulness substantially. Recently, the development of hepatocyte spheroid cultures has shown promising results. The proteome and transcriptome in the spheroids were similar to the liver tissue, and hepatotoxicity of selected substances was detected at in vivo-relevant concentrations.
The Drug Discovery DilemmaThe drug discovery process is suffering from poor productivity as depicted by the low number of new chemical entities that survives the clinical drug development phases. The termination of compounds in late pre-clinical and early clinical phases is mainly due to safety and poor efficacy [1,2]. The lessons learned from the AstraZeneca pipeline are that 82% of projects that are closed during the pre-clinical phase is due to unacceptable safety profile. In the clinical phases I and II, the closure of projects due to efficacy is increasing and in Phase IIb, and the number is as high as 88% of all projects that were terminated at this stage [1]. The poor translation of animal to human being is illustrated by the fact that animal testing identified only 70% of compounds that are toxic in human beings [3]. Another interesting finding is that 37% of first in class drugs discovered between 1999 and 2008 was based mainly on phenotypic pre-clinical screening (animal or in vitro models where a pharmacological response is measured but the target is not known), while only 23% of the first in class drugs was based on target-based screening [4]. This is in an era where the majority of the drug discovery is focused on target-based screening. Target-based screening is easily set up in a highthroughput format, and thus, a large number of substances can be screened. High-throughput format usually also delivers very potent and selective substances for the specific target nominated. However, the poor clinical outcome clearly indicates that the pre-clinical animal and in vitro models that are used to select efficacious and safe compounds do not translate into the human being.