Reports from the European Federation for Pharmaceutical Sciences (EUFEPS) and from other European organisations have repeatedly stressed the needs for a targeted effort to speed up the development of new and safe drugs. Stronger links between industry, academia, and regulatory authorities, more efficient use of modern information technologies, new methods of drug exploration, and tailored training programmes have been suggested as vital elements in a streamlining of the drug development process. Without such efforts, the European pharmaceutical industry is believed to be in danger of losing important grounds on the international market.
Introduction of new technologies such as high-throughput screening (HTS), computational chemistry, and combinatorial and automatic chemistry has made research and discovery in the pharmaceutical industry significantly more effective. These technologies enable a company to screen compound libraries for potential drugs at rates that are hundreds of times faster than the screening made by a skilled chemist. However, as the drug development process is performed today it still maintains many of the aspects of a trial-and-error process. A major improvement will require a more knowledge-oriented process that builds on a detailed and quantitative understanding of the biological and pathological processes associated with the functioning of the drug.
The current approach to drug discovery and development lacks efficient means to (i) define and control the conditions under which experiments and clinical tests are performed, (ii) extract the information available in the individual trial and validate it in terms of current knowledge in cell biology, medicine, etc., (iii) accumulate information from trial to trial and redesign the trial procedure to become an adaptive process where information acquired in one trial is immediately used to improve on the process, (iv) extrapolate results obtained from experiments on cell cultures and from animal experiments to apply to human patients, (v) predict the variation of drug efficacy and the occurrence of side effects in dependence of genetic modifications and of gender, age, or weight characteristics, and (vi) predict the likelihood that a particular chemical compound will function as a drug on the basis of knowledge about related compounds.
This situation may be viewed as a consequence of the extraordinary complexity of living systems. However, the more drug development resembles a trial-and-error process, and the less information one extracts from each individual test, the more tests must be performed before the drug can be approved. This makes the development process extremely lengthy and risky for the pharmaceutical industries and much too costly for the society. In view of the enormous progress that has been achieved over the last decade in information technology, systems biology, and complex systems theory, the situation is no longer acceptable. A completely new approach must be developed in which information is handled in a more rational manner and the enormous potential of more systematic approaches is taken into account. The BioSim Network integrates those aspects of biology, pharmacology, and modern information technologies that we consider most relevant to achieve a more rational drug development process.
A simulation model describes the temporal variation of a system in terms of the processes and interactions that are presumed to be at work. In connection with the development of a drug, the model will combine a pharmacokinetic description of the absorption, distribution, metabolism, and excretion of the drug with a detailed representation of the mechanisms responsible for its function and for the development of side effects and possible synergetic interactions with other drugs. To the extent that they are important, this description will include a representation of the drug's interaction with cellular receptors and intracellular reaction cascades, as well as of its effects on the intercellular communication. It may also be important to examine interactions with specific organs (the heart, liver, kidney, etc.) as well as with hormonal and immune regulations.
A simulation model so to speak translates our knowledge about the biological system into mathematical equations (usually a set of differential equations with a complementary set of algebraic equations). In the initial stage of the drug development process, one can use the simulation model to test any hypothesis one might have regarding the function of the drug vis-a-vis the established biological understanding. With approaches from bioinformatics, one can change the product in order to optimise its function, and even before the first molecule is produced one can estimate the likelihood that a given agent will function as a drug.
During the trial phase, the simulation model can be used as a vehicle to define an effective test protocol. The model can be used to check that the information obtained from the tests is consistent. Provided that it represents a proper representation of the underlying pharmacokinetic and biological mechanisms, the model can be used to predict the effect of a drug outside its normal physiological regime and under conditions not previously experienced. To the extent that these predictions are substantiated, our understanding of the drug's action is gradually extended. If the predictions are false we must examine both the hypotheses and the experimental procedures. The advantage is here that any discrepancy between hypotheses and experimental results appear right away and in a clear manner.
The gradual development of larger and larger models and the development of models that have been tested in more and more experiments represent an effective means of accumulation biological knowledge. Over the years this will allow the industry to perform the test procedure ever more effectively and to produce increasingly specific and advanced drugs. In the approval process, the regulatory agencies can use models to check that the tests are adequately performed. As previously mentioned, this is already to some extent a practise of the US Food and Drug Administration.
We presume that the development time for a drug can be reduced by 2-3 years as the industry adopts and becomes accustomed with the simulation approach. Hence, there is a strong economic motive for the industry in pursuing the use of simulation models. At the same time, the industry will be able to reduce its use of laboratory animals (and test persons) significantly. The main problem lies in the industry's lack of experience with this approach, and the associated lack of skilled modellers. In this connection it is important to emphasize that the process is only at its very beginning. The benefits to the industry will grow over time as the models become more and more detailed and realistic.
Part of the reduction in development time and cost may be turned into higher requirements to the test procedure and to the documentation of the test results. The reduced economic risks to the pharmaceutical industry may also allow it to engage into the development of drugs that would otherwise not be considered, e.g., individualised medicines and drugs for rare diseases. The continuous accumulation of coherent biological knowledge, that the simulation approach makes possible, exhibits no obvious saturation limits. This is by contrast to other approaches that have the character of technical fits. In the long run, a better and more systematic and quantitative understanding of the human biology is the most navigable route to new, safe and effective drugs.
At the present very few models exist that can describe a disease longitudinally, i.e. from its very start to the final cure (or the death of the patient). The advantage of simulation models is that they can represent available information about the relevant biological and pathological processes on any time scale. Hence, it is possible to describe how a disease develops over years in dependence of the performed treatment (the so-called disease life-cycle). This is of particular interest in connection with modern life-style diseases where active collaboration from the patient plays an important role. By means of a simulation model, health care providers can examine the long term consequences of possible adjustments in the treatment, and from simplified versions of the models the patients may learn to appreciate the significance of following a prescribed treatment and/or keeping a specific diet. Insights gained from such models can also be used to tailor health care policies, public campaigns, or educational programmes. An example could be the models developed some 15 years ago to design policies against the spread of AIDS.
By the time the Community support stops, the BioSim Network will have:
- Demonstrated to the European pharmaceutical industry that biosimulation is an important and effective approach in the drug development process
- Prepared the way for acceptance of simulation models as part of the documentation provided for the approval of a new drug
- Trained a group of highly motivated PhD's and postdocs with expertise in biosimulation and established a curriculum in biosimulation at several universities
- Developed into a major activity in EUFEPS with well-attended and lively annual conferences.
The success that the Network will have in achieving these goals depends on (i) the quality of the models developed by the Network, and the quality of its educational efforts, (ii) the success that the Network will have with respect to making computer simulations part of the normal operation of the industry, and (iii) the degree to which the regulatory agencies will accept simulation results as a complement to the normal trials. We are confident that the scientific qualifications of the participants suffice to guarantee fulfilment of the first condition, and that Community support to the Network will allow us to also fulfil the last two conditions.