Background: Animal models closely resembling the etiopathogenesis of Alzheimer's disease (AD) are needed for research on disease mechanisms and for drug development. No natural model of AD is available, so big hopes arose from transgenic and knockout technology, expecting that modulation and expression of pathogenetically important proteins resemble human brain pathology and functional deficits in the expected morphological and temporal pattern.
Objective: The real usefulness of these models should be discussed from an objective point of view.
Results: Not a single one of the published transgenic rodent models fulfils this hope, and even complex multiple transgenic animals do not suffer from real AD. It is crucial to be aware that all of the commonly used mice and rats are just models, and therefore results from drug efficacy testing have to be interpreted with care. Repeated experience with failed trials of new treatments that previously had been published as successful in animals has led to the wrong conclusion that animal models are of low predictive value or even of no use. Often clinical trials replicate exactly what was shown in the animal proof-of-concept studies.
Conclusion: The value of animal models depends mainly on the careful experimentation and correct interpretation of results. Appropriate planning of experiments will help to increase the predictive value in drug development programs, though this may also increase negative findings. However, the early failure may enable a faster focus on more promising strategies.