You are here


Hello Mr. de KERCHOVE, Director of Research at the FNRS,
I welcome the call for transparency by researchers who perform experiments on animals to write this open letter to you in response to an article published online by Le Soir on 12/02/2019 (“Nineteen research centers and Belgian companies undertake to communicate openly about their practices ”).


The purpose of this open letter is to initiate a debate among scientists about animal testing, as a prelude to a possible parliamentary inquiry. A basic principle in the life sciences is that no animal species is a biological model for another. Of course, animals resemble us. But it is important to distinguish between the word "similar" in everyday life and in the scientific study of living things.


The mouse is the most widely used animal in biomedical research. Although we share a majority of our genes with this species, there are important differences in gene regulation. For example, the gene for tail formation is activated in mice but not in humans. Aside specific  examples and empirical data, there is fundamental proof to demonstrate that no one animal species is a biological model for another: namely, the complexity of living systems (beings).


A complex system is composed of a large number of interacting entities that prevent the observer from predicting its feedback, behavior or future response by computational simulation. Chaos theory deals in particular with complex systems. Mice and humans are examples of complex evolutionary systems. Based on the definition of a complex system, it becomes evident that no one species can serve as a biological model for another (Greek & Hansen, 2013).


Recently, you refer to the article by Pulendran and Davis (published in the journal Science of September 25, 2020) which states: “Mice and humans are separated by 96 million years of evolution, and although the immune systems of the two species are broadly similar, they differ in many important details ”. This seems contradictory and lacks precision. On average, the regions encoding the proteins of the mouse and human genomes are 85% identical; some genes are 99% identical while others are only 60% (Zschaler et al, 2014).


In complex systems, the initial conditions are of primary importance. However, we observe a significant variation with respect to the “similarities” (regarding gene coding, which, in addition, does not guarantee a similarity of expression) between homologous murine and human genes. Moreover, since genes work in networks and not in isolation, the mouse is not a predictive model for humans. Empirical data in toxicology confirm this finding. One of the largest studies in this field yields a 43% correlation between rodents and humans (Leist & Hartung, 2013). If the rodent results were to be used to predict human outcomes, this correlation indicates that the prediction would be less reliable than tossing a coin toss.


According to the United States FDA, out of 10 new drugs that have passed animal testing, nine fail in clinical trials due to toxicity or lack of efficacy in healthy volunteers or patients (Akhtar, 2015).


Is the situation any better in universities where basic research (scientific curiosity) is practiced rather than regulatory toxicology? Here too, there is a serious "reproducibility crisis" of the experiments, namely the fact that most studies (involving animals) cannot be duplicated by other researchers, or even sometimes by the same team (Begley & Ioannidis, 2015). The many guidelines that call for improved reporting of animal experiments (ARRIVE, COMARADES, PREPARE, and ARRIVE 2.0) are doomed to fail as they are mostly voluntary. However, even if the researchers applied the correct methodology, it would not change the prediction rate for humans, for the reasons discussed above (Greek & Menache, 2013).


Researchers who experiment on animals regularly report dramatic advances in the media. But, in concrete terms, what is the cost-benefit balance of animal research? A flagship study published in 2003 covering 25,000 articles in basic research published between 1979 and 1983 in the best scientific journals, revealed that around 500 (2%) could potentially claim a future application in humans, around 100 (0,4%) resulted in a clinical trial and only one (0.004%) led to the development of a class of drugs useful in clinical medicine (angiotensin converting enzyme inhibitors) within 30 years the publication of their discovery in basic science (Crowley, 2003).


Considering all these facts, is it conceivable that animal experimentation constitutes one of the most grotesque mistakes made in the history of science?


Feedback and comments welcome.