I frequently hear that animal-based research informs scientists about diseases in humans or what a drug will do in humans. These statements take the form of the following. Basic science research informs scientists about biological phenomena. Animal models are then used to inform scientists about the phenomena in intact living animal systems and what the phenomena is like in humans. Based on the results from animal models, scientists then have an informed idea about the phenomena in humans and human-based research can begin more safely.
A press release from the University of Georgia discussed the research of Dr. Kate Creevy, an assistant pro
fessor in the UGA College of Veterinary Medicine, that was published in the Journal of Veterinary Internal Medicine. Creevy and colleagues studied the causes of death in 82 specific breeds of dogs and found large variation. Alluding to a canine version of personalized medicine Creevy stated: "If we can anticipate better how things can go wrong for dogs, we can manage their wellness to keep them as healthy as possible."
The press release then stated:
Promislow [a coauthor of the study] pointed out that because the building blocks of the dog genome and the human genome are the same, understanding the genetic basis of disease in dogs can inform human medicine. If specific genes are found to play a significant role in Cushing's disease in dogs, for example, scientists can assess whether the same process occurs in humans with the disease, with the ultimate goal of creating new strategies for early diagnosis and subsequent treatment . . . "There's potential to learn a lot about the genetics of disease in general using the dog as a model." (Emphasis added.)
I must admit that I do not really appreciate how the word inform is being used in these instances. It sounds to me like a weak version of predict. Weak, because the scientists that are speaking realize that animal models do not predict what a drug or disease will do in humans but must say something positive about using animals beyond the fact that they can be heuristic devices. Society is probably not going to allow scientists to insert electrodes into monkey brains, then water deprive them in order to get them to learn how to respond to visual stimuli, if all society is going to get in return is maybe an idea that might someday turn into a way to study humans. (This is essentially what basic research using animals is all about. See Is the use of sentient animals in basic research justifiable?)
If a scientist were to say that animal models per se are predictive for humans, then he would need to provide studies of numerous animal models revealing high positive and negative predictive values. Since studies do not show this, the animal modeler must use weasel words like inform. But upon closer examination these words convey no meaning. Let’s take toxicity testing for example. Compound A is tested on monkeys and the moneys apparently suffer no ill effects. Mice however are tested and found to suffer liver damage from the compound. Cats suffer no abnormalities nor do dogs. What does that tell us about the compound? All other factors being equal, it tells us nothing.
What about a drug that is a known irritant to the skin, eyes and lungs, and is a co-carcinogen in animals and that furthermore results in developmental defects in many animals? Think we should avoid it? That drug is aspirin.
How do animal tests and animal models inform us? They are not predictive for humans so what information are we getting from them. Granted, we can get ideas but this is not what we are lead to believe that we are getting. We are lead to believe that animal models allow scientists in drug companies to weed out bad drugs and develop good drugs. If that is indeed the case then animal models are being used as predictive models. But the evidence indicates that this is not in fact happening.
Caponigro and Sellers 2011:
Despite an improved understanding of the biology of cancer, and an unprecedented volume of new molecules in clinical trials, the number of highly efficacious drugs approved by the regulatory authorities remains disappointingly low. The significant attrition rate of drugs entering clinical trials comes at a high price. This price is paid primarily by the underserved patient and secondarily by the pharmaceutical and biotechnology community, which invests enormous resources perfecting a molecule only to watch it fail in humans . . . (Caponigro and Sellers 2011)
The Centre for Medicines Research International has noted that the average for the combined success rate at Phase III and submission has fallen to ~50% in recent years. To learn more about the causes for Phase III and submission failures, Thomson Reuters Life Science Consulting analysed the reasons for these failures between 2007 and 2010. This analysis included failures across all therapeutic areas, initial indications and major new indications. Re-formulations and new claims closely related to already approved indications were excluded from the analysis.
There were 83 Phase III and submission failures between 2007 and 2010. As shown in Fig. 1a, the therapeutic areas in which the largest proportions of these failures occurred were: cancer (28%); nervous system, which includes neurodegeneration (18%); alimentary and/or metabolism, which includes diabetes and obesity (13%); and anti-infectives (13%). Almost 90% of the failures across all therapeutic areas were attributable to either lack of efficacy (66%) or safety issues (21%) (Fig. 1b). (Arrowsmith 2011)
Animals are used to test both safety and efficacy. Many efficacious drugs are being weeded out by animal tests.
Robert Weinberg, of Massachusetts Institute of Technology, was quoted by Leaf in Fortune magazine as saying:
Weinberg explains. “And it’s been well known for more than a decade, maybe two decades, that many of these preclinical human cancer models have very little predictive power in terms of how actual human beings—actual human tumors inside patients—will respond . . . preclinical models of human cancer, in large part, stink . . . hundreds of millions of dollars are being wasted every year by drug companies using these [animal] models. (Leaf 2004)
Leaf also quotes Homer Pearce, “who once ran cancer research and clinical investigation at Eli Lilly and is now research fellow at the drug company” as saying:
. . . that mouse models are “woefully inadequate” for determining whether a drug will work in humans. “If you look at the millions and millions and millions of mice that have been cured, and you compare that to the relative success, or lack thereof, that we’ve achieved in the treatment of metastatic disease clinically,” he says, “you realize that there just has to be something wrong with those models.” (Leaf 2004)
The same can be said of disease research.
From 1996 to 1999, the U.S. food and Drug Administration approved 157 new drugs. In the comparable period a decade later—that is, from 2006 to 2009—the agency approved 74. Not among them were any cures, or even meaningfully effective treatments, for Alzheimer's disease, lung or pancreatic cancer, Parkinson's disease, Huntington's disease, or a host of other afflictions that destroy lives . . . the chance of FDA approval for a newly discovered molecule, targeting a newly discovered disease mechanism, is a dismal 0.6 percent. Diseases are complicated, and nature fights every human attempt to mess with what she has wrought. But frustration is growing with how few seemingly promising discoveries in basic biomedical science lead to something that helps patients, especially in what is supposed to be a golden age of genetics, neuroscience, and biomedical research in general. (Carmichael and Begley 2010)
In the recent past, nowhere was this inform notion more pronounced than in using chimpanzees to study HIV/AIDS. In the April 24, 2001 article Research request raises chimp debate, Chris Adams of McClatchy Newspapers, stated:
Eventually, at least 198 chimps were infected with HIV, according to a 1997 report by the National Research Council, a prestigious body affiliated with the National Academy of Sciences. But just one developed and died from an AIDS-like disease.
It should be noted that the chimpanzee that succumbed was “infected with three different isolates of type-1-HIV over a period of 10 years [that then lead to a] persistent decline in CD4+ T-lymphocytes that progressed to AIDS or an AIDS-like disease. Blood from this animal that was transfused into an uninfected chimpanzee induced a rapid depletion of CD4+ T-lymphocytes but did not cause clinical disease.” (Sibal and Samson 2001) So even this one case was highly questionable from the standpoint of proving that chimpanzees suffer from an illness that is like AIDS in humans.
The council said chimps could still be of value. But in its report, it concluded that “chimpanzees have not been a universally satisfactory model for human diseases” and “HIV infection of chimpanzees has not been an ideal model.”
That is science-speak for “chimpanzees did not work out as models for AIDS.” Adams:
Before that assessment, the U.S. government had bred about 400 chimps. The once-“critical model for understanding” HIV became “a surplus of chimpanzees and a substantial management problem,” the council report said . . . Despite their similarity to humans, chimps don’t react to infection the same way. In HIV research, for example, a possible vaccine protected the chimps but not people, NIH scientists wrote in the New England Journal of Medicine in 2007. Also in 2007, an article in the British Medical Journal concluded: “When it comes to testing HIV vaccines, only humans will do.”
Earlier this year, Ajit Varki of the University of California, San Diego, and his colleagues reviewed chimp research, disease by disease, and found that chimps and humans experienced disease differently. As they concluded in The Annual Review of Pathology: “Humans appear to have several surprising differences in the severity and/or incidence of diseases and pathologies that cannot be explained by environmental factors.”
It called into question chimps’ usefulness as human stand-ins.
“Chimps and humans are extremely close genetically, so there is this feeling that they should be a good model to study human diseases,” Varki said. “But most attention has been on what is similar. We should also pay more attention to the differences.” (Emphasis added.) (Adams 2011)
We should also pay more attention to the differences. Where have we heard that before?
Adams, Chris. 2011. Research request raises chimp debate. McClatchy Newspapers, April 24 2011 [cited April 24 2011]. Available from http://www.azcentral.com/arizonarepublic/news/articles/2011/04/24/20110424chimps-research0424.html.
Arrowsmith, John. 2011. Trial watch: Phase III and submission failures: 2007-2010. Nat Rev Drug Discov 10 (2):87-87.
Carmichael, Mary, and Sharon Begley. 2010. Despeately Seeking Cures. Newweek 2010 [cited May 31 2010]. Available from http://www.newsweek.com/id/238078?obref=obinsite.
Leaf, C. 2004. Why we are losing the war on cancer. Fortune (March 9):77-92.
Sibal, L. R., and K. J. Samson. 2001. Nonhuman primates: a critical role in current disease research. ILAR J 42 (2):74-84.