Since Congress mandated testing of new medicines and chemicals on animals back in 1937, billions of laboratory animals have been subjected to a vast number of chemicals and medicinal compounds — all under the entrenched belief that this “safety testing” is translatable to human animals; a concept known as the concordance assumption.
But a recent study by a consortium of medical researchers (Seok et al) and published on the Proceedings of the National Academy of Sciences website, throws the assumption of concordance into serious doubt. To be sure, there has been a scattering of previous studies* casting doubt on this concordance theory (and pharmaceutical companies have know for quite some time that most new drug candidates first tested on “animal models” fail to work when tested on humans), but this most recent study is the most comprehensive yet conducted.
Its wider implications cast serious doubt on seventy plus years of safety research upon which we base our notions of “safe” medicines and chemicals verses unsafe ones. The ramification of this most recent study are sure to have major impact on the current testing practices of the medical, chemical, and pharmaceutical industries.
The Study Results in More Detail
This particular research was prompted by an investigation of 150 clinical trials of drugs begin tested for treatment of inflammatory diseases (such as diabetes, arthritis, asthma). Investigators found (to date) a 100% failure rate of these anti-inflammatory drugs in these trials; all of the drugs were originally tested on mice.
To understand why this was happening, the consortium decided to conduct further studies of the effects of various treatments (that produced inflammation) and compare results between humans and mice.
Looking first at acute inflammation in mice resulting from three types of stimuli (bacterial toxins, burns, and trauma), the investigators quantified changes in gene activity (i.e., increased or decreased gene expression) in response to the stimuli in thousands of genes.
Using the concordance assumption (that mouse gene activity changes would be concordant with similar gene changes in humans), the investigators expected that the gene activity changes seen in mice would enable accurate predictions of similar changes in humans. Yet this was not what the investigators discovered. Instead, changes in a particular mouse gene, following drug treatment, failed to predict changes in the closest-related (homologue) gene in humans.
Additional experiments revealed another significant difference: in humans, similar patterns of gene activity in response to the stimuli were noted, but not in mice. Each type of treatment stimulus in mice resulted in distinct gene activity changes — confirming the earlier findings that similar treatments resulted in different (gene) responses between humans and mice.
These differences could not be attributed to experimental “noise”, according to the consortium, but rather, they were the result of fundamental physiological differences between mice and humans in how each responds to the various stimuli.
More Experimental Results
The consortium investigators next tested key biological signalling pathways (i.e., sequences of signals within cells that trigger various functions) after similar treatments. Once again, dissimilar cell signalling responses between mice and humans were observed. Further testing on other human/mouse models of inflammatory diseases also showed low correlations.
The experiments by Seok et al confirm that, physiologically, humans and mice have little in common and that mouse models (and presumably other animal models) are of little clinical relevance in testing new treatments for human inflammatory diseases, which are quite prevalent in humans (and include various allergies, celiac disease, asthma, RA and other autoimmune diseases).
The results would seem to be valid for other categories of human disease as well. However, for carcinogenicity studies, most researchers consider mice an appropriate test model.
Of Mice and Men…and other Animal Models
Many researchers have previously noted that there is a 120 million year difference in evolutionary (genetic) change between mice and humans. Consequently, mice have distinct immune systems and suffer from different diseases; they lack a gall bladder, menstrual cycles, and have different metabolic rates and lifespans (which are functions of their smaller size).
Thus, in most ways, these experimental results should not come as a great shock.
The General Medical Research Implications of the Findings
The general implication of these findings is that they most likely hold for other laboratory animals as well: rats, rabbits, and dogs being the most common examples. However, the experiments by Seok et al were not conducted on primates — monkeys and chimps — which are genetically and physiologically much more similar to humans.
The objection to using primates — and especially chimpanzees — in medical research is primarily one of ethics: should we subject our closest evolutionary relatives to what amounts to cruel and unusual treatment, that is, the confinement in cages and the introduction of experimental chemicals or disease-causing gene alterations?
While these results do not mean that medical testing as we know it is finished, it will force medical researchers to reevaluate their assumption about the validity of using animal models to test new medical treatments that will ultimately be used to treat human diseases.
It will also, no doubt, call into question the wisdom of spending billions of dollars every year on animal model testing of new drugs, and this will have a serious impact on future funding (and the careers of many laboratory scientists).
New chemical testing methodologies are emerging, but are yet to be fully validated. Although one cannot expect a complete ban on animal testing any time soon, on a more basic level, these findings may mean medical researchers will have to rely more on clinical observation, better gene analysis techniques, and better in vitro (human) cell models (cultures). These latter models are still undergoing validation studies, but there has been at least one recent advance here.
So, What About the Impact on Toxicology Testing?
Beyond the medical field, these results have serious implications for toxicology testing, for, if the concordance assumption is false — granted that the above experiments were not performed with synthetic chemicals (but see references, below) — then current toxicology testing on animals to determine human safety levels is probably invalid as well.
Our every day lives are permeated with household and environmental products containing a vast assortment of synthetic chemicals — most all of which have been tested on non-human animals prior to their approval for use by humans.
If non-human animals are poor models for predicting disease treatment effects in humans, can they be valid test models for toxicology safety?
If the failure of concordance in animal medical testing is applicable to general chemical safety testing (ironically, another form of concordance) then this would mean that our homes, food and environment are flooded with a dizzying assortment of essentially untested (or invalidly tested) chemicals.
Further, if the concordance assumption is generally false, meaning it is scientifically invalid, then the entire regulatory process upon which we base our confidence in chemical safety is seriously flawed.
Indeed, a recent National Research Council study supports the argument that many chemical safety tests are not relevant to humans.
It is hoped that these findings will spur investment and investigation into alternative forms of chemical safety testing that will ultimately, be less costly to all concerned.
The recent effort to develop a so-called ‘human on a chip’ (an expanded in silico variant of ‘organ on a chip’) technology holds some promise here, but is also awaiting full validation as a replacement for animal testing.
Main scientific reference for this post: Genomic responses in mouse models poorly mimic human inflammatory diseases (Seok et al, 2013)
Additional source material for this post came form the Truthout.org article (repost from Independent Science News): The Experiment Is on Us: Science of Animal Testing Thrown Into Doubt’ by Pat Dutt and Jonathan Latham, PhD.
* References for further academic reading:
Dressman HK et al, 2007. Gene expression signatures that predict radiation exposure in mice and humans. PLoS Med 4:4.
Greek CR, Swingle Greek, J (2003). Specious science: Why Experiments on Animals Harm Humans. The Continuum International Publishing Group, Ltd, London.
Knight A (2007) Systematic reviews of animal experiments demonstrate poor human clinical and toxicological utility. ATLA 35: 641-659.
Mestas, J and Hughes, CCW, (2004) Of mice and not men: differences between mouse and human immunology, The Journal of Immunology, 172: 5.
Stoloff L (1992) An analysis of the 1987 list of IARC-identified human carcinogens and the correlated animal studies. Regulatory Toxicology and Pharmacology 15: 10–13