Mouse studies often don’t work, NIH admits after landmark 10-year study

From ANIMAL PEOPLE,  March 2013:

WASHINGTON D.C.–– Bluntly stated the headline in the February 11,  2013 edition of Proceedings of the National Academy of Sciences,  “Genomic responses in mouse models poorly mimic human inflammatory diseases.”

Echoing an article of faith prevailing among anti-vivisectionists for several centuries,  though not validated to biomedical researchers’ satisfaction by mouse studies,  the report in this instance came from 10 years of investigation by 39 leading biomedical researchers,  funded by the U.S. National Institutes of Health––long the world’s leading sponsor and at times the leading defender of animal-based research. Blogged NIH director Francis Collins,  “If it works in mice,  so we thought,  it should work in humans.  But,” Collins recounted,  “150 drugs that successfully treated sepsis in mice later failed in human clinical trials.  Sepsis,  a life-threatening systemic infection,   can be caused by a variety of pathogens,  including bacteria,  viruses, and fungi. Serious consequences occur when tissues damaged by infection produce proteins sometimes called ‘alarmins’ that send the immune system into overdrive.  Traumatic injuries involving extreme blood loss or burns can set off the same dangerous response.

“To probe the molecular response to all of these triggers,” Collins explained,  “the authors [of the study] took periodic blood samples from 167 trauma patients;  244 patients with burns over at least 20% of their body;  and four healthy volunteers who had been injected with a low-dose bacterial toxin.  Then they studied the activity of the genes in the white blood cells.  They found that of the 5,500 or so genes that responded to traumatic injury,  91% also played a role in burn response and recovery.  About 45% of these same genes were involved in recovery from the bacterial toxin exposure.

“Mice,  however,  apparently use distinct sets of genes to tackle trauma,  burns, and bacterial toxins,”  Collins continued.  “When the authors compared the activity of the human sepsis-trauma-burn genes with that of the equivalent mouse genes,  there was very little overlap.  No wonder drugs designed for the mice failed in humans: they were,  in fact,  treating different conditions!

“But that doesn’t mean studying mice is useless,”  Collins added. “Mice are more resilient to infection and mount a much more regulated immune response to pathogens than humans. Perhaps this is because mice nose around in some filthy places and can’t afford to overreact to every microbe.  If we knew how these rodents limit the drama of their immune response,  it might be useful for us humans.

“But this study’s implications may well go beyond mice and sepsis,”  Collins allowed.  “It provides more reason to develop better and more sophisticated models of human disease.  More than 30 percent of all drugs successfully tested in animals fail in human trials.  The NIH plans to commit $70 million over the next five years,”  Collins announced,  “to develop miniature 3-D organs made with living human cells to help predict drug safety and efficacy. Though this is high-risk research,”  Collins acknowledged,   “these ‘tissue chips’ may ultimately provide better models of human disease and biology than the use of animals.” Reported New York Times science writer Gina Kolata,  “Sepsis afflicts 750,000 patients a year in the United States,  kills one-fourth to one-half of them,  and costs the nation $17 billion a year.  It is the leading cause of death in intensive-care units.”


Despite the importance of the “Genomic responses” finding,  however,  co-author Ronald W. Davis of Stanford University told Kolata that the team tried for more than a year to publish their paper.  “They submitted it to the publications Science and Nature, hoping to reach a wide audience.  It was rejected from both,”  Kolata wrote.

Scientific journal peer reviewers,  said Davis,  “were so used to doing mouse studies that they thought that was how you validate things.  They are so ingrained in trying to cure mice that they forget we are trying to cure humans.”

“When I read the paper,  I was stunned by just how bad the mouse data are,”  said University of California at Los Angeles sepsis expert Mitchell Fink told Kolata.  “I think funding agencies are going to take note,”  Fink predicted. “Until now,  to get funding,  you had to propose experiments using the mouse model.” Two leading biomedical researchers told ANIMAL PEOPLE that they had already questioned the applicability of mouse studies to sepsis,  burns,  and trauma in human patients.

“I can only say that I would not consider the mouse as first choice model for the study of burns,  since the anatomy of the mouse skin,  and the gross physiology of the mouse have less similarities to humans then,  for instance, the pig,  and in particular the miniature hairless pig,”  offered Gad Simon,  past editor of the Israel Journal of Veterinary Medicine and for more than 20 years a member of the Animal Experimentation Review Committee at the Israel Institute for Biological Research.

“At least as far back as 1959 people were using mice for such studies,  dropping them in boiling water and then examining them for blisters and toxins and God knows what.  I saw such work with my own eyes,”  recalled Colorado State University at Fort Collins virologist Charles Calisher,  who said he had explosively disapproved of it even then, when animal studies were much less often questioned.

“As to what the findings will mean for the future of mouse studies in general,  I do not know,  of course.  Mice are small and inexpensive,  which I presume are the reasons anyone uses them as models in the first place,”  Calisher said.

Millions of mice

Even if the findings in the “Genomic responses” paper influence only studies of bacterial infection involving mice,  mice are used in experiments so often compared to all other animals that eliminating just 10% of mouse use would spare as many animals as if all use of other species stopped entirely.

“Clearly ‘Genomic responses’  is very significant,  and the New York Times was right to report it on the front page,”  American Anti-Vivisection Society and Alternatives Research & Development Foundation president Sue Leary told ANIMAL PEOPLE.  “This paper represents a trend,”  Leary continued.  “Similar studies have looked at animal research methods for particular diseases and have found problems almost as dramatic as the complete failure in inflammatory disease research.  The predominant mouse types used in Alzheimer’s research were shown to be useless to the point that leading researchers were sent ‘back to the drawing board.’  Comparable situations have been exposed recently in research on stroke,  multiple sclerosis,  and asthma.

“At the Eighth World Congress on Alternatives & Animal Use in the Life Sciences in 2011,”  Leary recalled, “nearly 1,000 attendees agreed to the ‘Montreal Declaration,’  which is designed to examine the validity of individual research proposals using animals and is gaining momentum.  The important lesson [from ‘Genomic responses’] is that everyone needs to challenge the assumptions behind routine approval of funding for research and testing that uses animals.  We saw that done successfully with the publication in 2007 of the National Academy of Sciences report ‘Toxicity Testing in the Twenty-first Century:  A Vision and A Strategy.’  That comprehensive examination of the problems of the current, animal-based methods in chemical safety assessment was commissioned by the Environmental Protection Agency.  Since then,”  Leary said,  “the EPA and other federal agencies have invested much more in alternative technologies to be used instead of animals.

“When AAVS and the Alternatives Research & Development Foundation conducted our push in 1999 to ban the use of mice to produce monoclonal antibodies,  which are widely used in all kinds of research,”  Leary recounted,   “we drew upon a growing scientific consensus about the benefits of alternative methods.  NIH ultimately declared that researchers should use in vitro methods to produce monoclonal antibodies,  unless they could provide a detailed justification.  Our understanding is that this development has prevented the use of up to one million mice a year.”

Other medical fields

Researchers in some areas relatively far removed from sepsis,  trauma,  and burns took immediate note of “Genomic responses.”

“A very interesting paper and I suspect sound. Of course it has to be tested and validated by other researchers, but it smells correct,”  Louisiana State University epidemiology professor emeritus Martin Hugh Jones told ANIMAL PEOPLE.

“Really mind-blowing!” said Jack Woodall,  a cofounder of the International Society for Infectious Diseases’ Program for Monitoring Emerging Diseases.  Woodall recalled that for decades virologists injected extracts of blood and tissues from humans,  livestock  and wildlife directly into the brains of baby mice to isolate viruses,  before tissue cultures were developed in the 1990s that produced faster and more accurate results.  “Baby mice were before that [believed to be] the most sensitive system––it had nothing to do with whether they reacted like people,”  Woodall explained.

The “Genomic responses” paper did not study mouse response to viruses,  but the much less sensitive response of mice to bacteria calls into question whether mice might also be less sensitive than humans to viral infection.

Mice are still used extensively in vaccine research,  including in findings reported on February 19,  2013 by University of Georgia professor Biao He which may lead to much more effective control of rabies.

Explained University of Georgia College of Veterinary Medicine publicist Kat Gilmore,  Biao He and team “used a common dog disease—canine parainfluenza—to build a new [anti-rabies] vaccine,”  which can be administered nasally, orally,  or as a conventional injection.  As detailed in the Journal of Virology,  Gilmore wrote,  “The study tested the efficacy of the drug on a mouse model.  When the mice were administered a lethal dose of rabies,  survival was 100% when they had received the vaccine nasally or into muscle.  Survival was 50% when the vaccine was administered orally.”

A rabies vaccine which can be administered nasally could potentially be sprayed into caves to immunize bats, the major reservoir of rabies in wildlife.


“The Humane Society of the U.S. and Humane Society International have urged similar critiques of other areas of animal research,”  said HSUS president Wayne Pacelle.  “For example,  we have argued that there are major problems with chimpanzee research and this was finally confirmed by the Institute of Medicine last year.  Leading research journals have published studies However,”  Pacelle added,  “it is important that we not simply focus on the failures of the current animal research paradigm,  but that we also encourage the development of non-animal alternatives.  We have launched the Human Toxicology Project Consortium to make this vision a reality,”  Pacelle said.

Commented Procter & Gamble toxicologist Harald Schlatter,  “The ‘Genomic responses’ finding is interesting, and needs to be verified over the coming years in terms of its relevance and impact––primarily for drug development. However,  I don’t think this impacts or relates directly to P&G,  as we do not test on sepsis,  trauma, and burns,” Schlattter added. “We only explore testing for very selected endpoints for which no alternatives exist yet.  Even for these,  we are working to develop non-animal alternatives.  We have spent the last few decades investing over $300 million in the development of over 50 alternative tests that are now used throughout the industry,  resulting in where we are today, with over 99% of our safety assessments worldwide being conducted without animals.”


Interniche representative Nick Jukes has for more than 10 years traveled the world introducing non-animal teaching methods.  The “Genomic responses” findings,  Jukes told ANIMAL PEOPLE,  are “about pure research and the testing of drugs,  not a pedagogical issue.  So it doesn’t directly correspond to our work.  However, there are of course connections between these fields,”  Jukes added.  “The findings are another reminder that within human medicine we should be focusing on human bodies and human tissue,  and that is also true within university-level education of medical students and within professional training of doctors.”

The findings also “makes very clear that a lazy acceptance of convention is far from scientific,”  Jukes added.

––Merritt Clifton

Print Friendly

Leave a Reply

Your email address will not be published.