This post is a part of our Bioethics in the News series. For more information, click here.
By Hilde Lindemann, PhD
According to a paper just published in the Proceedings of the National Academy of Sciences, using mouse models to study inflammatory disease caused by sepsis, burns, or trauma in humans is completely pointless. As the use of mice for this kind of research has long been the established practice, this means that years of study and billions of dollars have been squandered with nothing at all to show for it. There are several ethical issues here. For one thing, have we as a nation gotten so cavalier about nonhuman life that we are willing to injure lab animals routinely without even checking to see whether there’s a scientific reason for doing it? Can doctors trust the findings of medical research, and can patients, in turn, trust that their doctors’ recommendations are based on good science?
The new study, whose somewhat unwieldy title is “Genomic Responses in Mouse Models Poorly Mimic Human Inflammatory Diseases,” took ten years to complete and involved thirty-nine researchers from across the United States(1). It didn’t start out as a mouse study, though. The initial idea was to examine the white blood cells of hundreds of human patients to see what genes they were using to combat the ill effects of sepsis, burns, and trauma. When the results were in, interesting patterns emerged that seemed to suggest it would be possible to predict who would survive and who would end up in intensive care with a strong likelihood of dying.
However, when the group tried to publish their findings, no journal would accept them. Among other reasons the referees gave was that the researchers hadn’t shown that the genetic response was the same in mice. As the study’s contributing author, Ronald W. Davis put it, “They were so used to doing mouse studies that they thought that was how you validate things. They’re so ingrained in trying to cure mice that they forget we are trying to cure humans.”
Even so, the group figured there would be some similarities between mice and humans, so they took the referees up on their criticism. Because there are no studies that systematically evaluate, on a molecular basis, how well mouse models mimic human inflammatory disease, the group compared their own findings with what is known about genomic responses to sepsis, burns, and trauma in mice. The result? “Although acute inflammatory stresses from different etiologies result in highly similar genomic responses in humans, the responses in corresponding mouse models correlate poorly with the human conditions and also, one another. Among genes changed significantly in humans, the murine orthologs are close to random in matching their human counterparts.”
The reason appears to be one that even a non-scientist might have predicted: mice aren’t—get ready for it—human. For one thing, mice can eat rotten food and other garbage, because their bodies tolerate bacteria far better than human bodies do. It takes a million times more bacteria in the bloodstream to kill a mouse than it does to kill a person.
This should have been an obvious tip-off to researchers that there was something wrong with their assumption about the similarities between the species, as the likeliest explanation for mice’s high tolerance of bacteria would be that their immune systems are particularly powerful. And if that were established to be the correct explanation, it in turn should have been the tip-off to researchers that mice models for combating inflammatory disease would be no good, because the immune system is obviously implicated in that process as well.
As it turns out, inflammatory disease in mice that is caused by burns involve different groups of genes from that caused by sepsis, and if the disease is caused by trauma, still another group of genes is implicated. In humans, on the other hand, the etiology of the disease doesn’t matter—similar genes are used regardless of the cause. As a consequence, a drug that works in mice by disabling a particular gene that happens not to be involved in humans, or that actively helps combat the inflammatory disease in humans, could aggravate the disease or even kill the patient.
The implications of all this go far beyond the treatment of inflammatory disease. According to Mitchell Fink, an expert on sepsis at UCLA, funding agencies have required investigators to propose experiments using the mouse model. Obviously, peer-reviewed science journals have assumed the validity of that model as well. It will take a massive change in thinking, however, before better procedures can be established that shift research away from mice and toward the routine use of human tissues.
That shift may not happen. Bioethicists and other interested parties have been complaining for years, for example, against journals’ well-entrenched policy of not publishing negative results, even though these results, if widely known, might keep patients from ingesting useless drugs. Private IRBs answerable only to the pharmaceutical companies that have a vested interest in the drug trials they oversee are likewise well entrenched and unlikely to be dismantled any time soon. Large-scale federal funding for research sometimes goes to ill-designed studies that simply collect data, with no clear hypotheses that would give the data meaning. And these are only three of the many problems that plague medical research.
All the same, the “Genomic Responses” paper gives us a reason for optimism. It was, after all, the peer review process that goaded these researchers into exploding a basic research assumption. Never mind that the reviewers themselves not only bought the mistaken assumption, but insisted on it. The point is that, in doing so, they played an important, albeit unwitting, role in showing that it was false.
This post was written in response to the article “Mice Fall Short as Test Subjects for Humans’ Deadly Ills,” published on the New York Times website on February 11, 2013.
Hilde Lindemann, PhD, is a Professor of Philosophy and Associate in the Center for Ethics Humanities in the Life Sciences at Michigan State University.
Join the discussion! Your comments and responses to this commentary are welcomed. The author will respond to all comments made by Friday, March 8. With your participation, we hope to create discussions rich with insights from diverse perspectives.
You must provide your name and email address to leave a comment. Your email address will not be made public.
MSU Bioethics 
Thanks to Professor Lindemann for highlighting ethical issues we can summarize as “mice and-or-versus-men.” A worrisome public health corollary– rodent studies are the current standard for determining human toxicology danger threshholds. If it turns out that mice’s digestive and immunological fortitude extends to many toxins–that is, that they are more resistant to toxins than humans– the whole basis of current toxicology practice would be misguided. Of course, toxicologists add margins of safety, usually orders of magnitude, to account for so-called “uncertainty factors” in their risk analsyses . But certainly it is chastening to consider how uncertain the basic rodent testing model is. A critical question is whether that model could be distorted enough to offset standard uncertainty factors. In that case many of our regulatory threshholds for toxins would be set too low for public safety. There is room for a hero here: well -done observational epidemiological studies in humans. Such studies could become crucial comparisons for evaluating when rodent models work well and when they don’t. Despite the rhetoric, randomized clinical control trials are not always the “gold standard.” Ethically, we can’t randomly poison people to test toxicity threshholds! But we can turn misfortune to good by applying observational epidemiological techniques in the wake of known human contaminations, across as many cases as possible. Will the cumulative data show that “men” look like “mice?” We won’t know unless funders and researchers continue to do observational epidemiological studies in an age that has wrongly relegated them to 2nd class status in the scientific world.
Professor Mongoven makes a good point. Perhaps epidemiological studies should be preferred to randomized, double-blinded control trials more generally. For the reasons given, they are likely to be superior to mouse studies.