Two reports about plants, which are complex systems, demonstrate why animal models cannot predict human response to drugs and disease.
Mukhtar et al. studied the mustard plant Arabidopsis and found 6,205 protein-protein interactions, among 2,774 individual proteins. This only represents approximately two percent of the total protein-protein interactome, which means that the total number of protein-protein interactions in the plant might be around 310,250. (Mukhtar et al. 2011)
Because animals (human and nonhuman) are also examples of complex systems, we expect them to have a large number of interacting parts and that these parts will be parts on one level and wholes on another. Further, there will be levels of organization where different parts interact. For example, genes interact with genes on one level while genes interact with proteins on another level, and on yet another level, proteins interact with proteins. There are more levels, but the fact that in the mustard plant there could be somewhere in the neighborhood of 300,000 protein-protein interactions gives us some idea about the difficulty involved in extrapolating outcomes between two complex systems. All those interaction probably have feedback loops and can be influenced by other factors.
In another study, scientists discovered that cloned trees, that were transported to different places after the cloning, reacted differently to drought. This is similar to monozygotic twins suffering from different diseases or reacting differently to medications. Both, twins and the trees, demonstrate the role that the environment and epigenetics play in determining the reaction to perturbations in the system.
Another example, this time from butterflies, reveals the importance of very small differences between species. Butterflies in the genus Heliconius were discovered to have evolved the same red spot on their wings by using the same gene. However, the subtle differences in the pattern are controlled by gene regulation.(Reed et al. 2011)
Another interesting principle seen with the wing colors is the fact that the gene controlling the color is also responsible for development of the eye in other species. The same gene, as I have pointed out, can result in different proteins and play different roles in different tissues. For example, by studying a human family from Pakistan, scientists discovered that a mutation in gene IL11RA caused craniosynostosis, maxillary hyperplasia, delayed tooth eruption and extra teeth. This complicates the picture when studying genes in mice in the effort to determine what they do in humans.
Reed, the lead author of the butterfly study stated: "This is in line with emerging evidence from various animal species that evolution generally is governed by a relatively small number of genes. Out of the tens of thousands in a typical genome, it seems that only a handful tend to drive major evolutionary change over and over again."
I should note here, that in order to gain this knowledge about butterflies, scientists experimented on the butterflies. As I have stated many times, animals can be used in science to discover interesting stuff. In this case, they discovered something about butterflies that is also known to occur in humans and reinforced a general principle in biology in addition to explaining a phenomenon in butterflies that had puzzled scientists for years. The fruitfly has been used to discover fascinating things about genes and some of these findings have informed scientists about humans.
Small differences among species in component interactions, genes, gene regulation and so forth explain the following. Kay, writing about gene therapy in 2011:
"Possibly the biggest hurdle is the inability to predict both innate and antigen-dependent immune responses in humans, some of which cannot currently be replicated in animal models. Clinical success more generally is also being hampered by the inability to accurately correlate animal and human studies: it is currently not possible to know whether vector-based gene transfer efficacy in humans will reflect that seen in non-human species. . . . Recombinant AAV vectors have also been administered systemically for liver-based treatment of factor IX deficiency (as occurs in haemophilia B). In mice and dogs, a single-dose of AAV2 vectors can successfully treat haemophilia B for many years81, 82, but similar success has not been achieved in humans. Therapeutic plasma levels of factor IX in a patient with haemophilia B lasted only a few months owing to a cellular immune response directed against the capsid peptides during degradation of the capsid in hepatocytes; this resulted in a transient, immune-based hepatitis and the loss of transduced hepatoctyes83. This response has not been observed in other mammalian species and is perhaps unique to humans because the AAV vector was derived from the human AAV2 virus — it is unclear whether a non-human AAV vector would give the same result. Even though it is a topic of great interest, it has not been possible to recapitulate this type of immune response in animal models. The diverse human polymorphic variations in genes affecting human immunity may make it difficult to predict which patients are most susceptible to these types of responses. This important issue may be resolved by a current haemophilia B clinical trial involving a vector packaged with a non-human primate-derived AAV8 capsid (which nonetheless shares 82% of its amino acids with the capsid of the human AAV2 virus)84. Transient administration of mild immunosuppressive agents provided at the same time as the vector may be required because the capsid peptides will have a finite lifetime, after which the transduced hepatocytes will no longer be a target for reactive T cells. "(Kay 2011)
Compare all of the above with the following from Lois Collins, of the Desert News:
"Watch a life-saving drug flow into the arm of a loved one who has cancer or reach for prescription medication to ease your allergies or lower your blood pressure and there's a good chance you owe some thanks to a fruit fly, a zebrafish or a mouse. Or maybe all of them. . . . It's what they have in common that has proven beneficial: protein-coding genes. For instance, of the fruit fly's 14,000, half are the same in humans. Zebrafish and humans both have about 25,000, many of them the same. Mice are even more similar genetically to humans, says Dr. Mary Beckerle, CEO and director of the Huntsman Cancer Institute. . . . Progress in individual diseases has hinged on studies that moved through a hierarchy of creatures, from simplest to more complex, until there was enough known and enough promise to justify human testing. . . . Once a gene that causes human disease is identified, scientists can search gene databases to see if that gene is present in a model organism. Next, they remove the gene from a worm, fish, fly or mouse or introduce a modified version of the gene and see what happens. . . . The gene that causes most colon cancers in humans is called APC. In model systems (where the equivalent of a human disease is created in an animal to learn about it and perhaps shed light on potential cures) the biochemical pathways APC controls were found, affording opportunities to study both how it should work and what can go awry. . . . Some model systems are also exceedingly helpful for "drug screens," where compounds are tested sometimes randomly to see what might have an effect on a particular gene mutation or disease."
The article pretty much ignores the relevant advances that science has made in the last decade or two and the poor state the pharmaceutical industry is in secondary to relying on animal models. According to PRNewswire, July 11, 2011:
"GBI Research, leading business intelligence provider, has released its latest research report, 'Top R&D Drug Failures - Toxicity and Serious Adverse Events in Late Stage Drug Development are the Major Causes of Drug Failure', which provides insights into major drug failures during 2005–2010. The 20 drugs included in the report belong to key pharmaceutical companies and were undergoing research for a major indication. . . . The main cause of drug failure in Phase III, when the efficacy of the drug is already evaluated, is either its safety measures or its ability to demonstrate any additional benefits in a larger population over a longer time period. . . . According to analysis conducted during 2005–2010, it was found that the number of drugs dropped from the pipeline increased from 2003 to 2007 and furthermore decreased from 2007 to 2010. Many drugs have been discontinued during the course of clinical trials, due to safety and efficacy issues [both are properties that animal models are used to evaluate]."
Alan Oliff, former executive director for cancer research at Merck Research Laboratories in West Point, Pennsylvania stated in 1997: “The fundamental problem in drug discovery for cancer is that the [animal] model systems are not predictive at all.”(Gura 1997) The so-called xenograft mice, mice who have cancer from human tumors, have not worked out for predicting the human effects of anti-cancer drugs either. Edward Sausville, associate director of the division of cancer treatment and diagnosis for the developmental therapeutics program at the NCI 1997: “We had basically discovered compounds that were good mouse drugs rather then good human drugs.” (Gura 1997)
In the April 1, 2010 issue of The Scientist "Mouse models that use transplants of human cancer have not had a great track record of predicting human responses to treatment in the clinic. It’s been estimated that cancer drugs that enter clinical testing have a 95 percent rate of failing to make it to market, in comparison to the 89 percent failure rate for all therapies . . . Indeed, “we had loads of models that were not predictive, that were [in fact] seriously misleading,” says NCI’s Marks, also head of the Mouse Models of Human Cancers Consortium . . ."(Zielinska 2010)
From Nature Biotechnology 2010: "The low predictive value of mouse cancer models for human disease is a major challenge for cancer research. Whereas human tumors develop from individual cells in the context of normal tissue, cancer research mostly relies on models employing xenografts or carrying oncogenic mutations throughout the whole animal or tissue." (M.E. 2010)
Dr. Richard Klausner, then-director of the National Cancer Institute: "The history of cancer research has been a history of curing cancer in the mouse . . . We have cured mice of cancer for decades—and it simply didn't work in humans." (Cimons, Getlin, and Maugh II 1998)
On January 12, 2006, then U.S. Secretary of Health and Human Services Mike Leavitt stated: “Currently, nine out of ten experimental drugs fail in clinical studies because we cannot accurately predict how they will behave in people based on laboratory and animal studies.” (FDA 2006)
Even Francis Collins of the NIH stated in 2011: "The use of small and large animals to predict safety in humans is a long-standing but not always reliable practice in translational science. . . . The use of animal models for therapeutic development and target validation is time consuming, costly, and may not accurately predict efficacy in humans." (Collins 2011)
We could learn a lot from botanists and lepidopterists.
Cimons, Marlene, Josh Getlin, and Thomas H. Maugh II. 2010. Cancer Drugs Face Long Road From Mice to Men 1998 [cited Nov 8 2010]. Available from http://articles.latimes.com/1998/may/06/news/mn-46795.
Collins, Francis S. 2011. Reengineering Translational Science: The Time Is Right. Science Translational Medicine 3 (90):90cm17.
FDA. 2010. FDA Issues Advice to Make Earliest Stages Of Clinical Drug Development More Efficient. FDA, June 18, 2009 2006 [cited March 7 2010]. Available from http://www.fda.gov/NewsEvents/Newsroom/PressAnnouncements/2006/ucm108576.htm.
Gura, T. 1997. Cancer Models: Systems for identifying new drugs are often faulty. Science 278 (5340):1041-2.
Kay, Mark A. 2011. State-of-the-art gene-based therapies: the road ahead. Nat Rev Genet 12 (5):316-328.
M.E. 2010. In This Issue. Models that better mimic human cancer. Nature Biotechnology 28 (1):vii.
Mukhtar, M. Shahid, Anne-Ruxandra Carvunis, Matija Dreze, et al. 2011. Independently Evolved Virulence Effectors Converge onto Hubs in a Plant Immune System Network. Science 333 (6042):596-601.
Reed, Robert D., Riccardo Papa, Arnaud Martin, et al. 2011. Optix Drives the Repeated Convergent Evolution of Butterfly Wing Pattern Mimicry. Science.
Zielinska, Edyta. 2010. Building a better mouse. The Scientist 24 (4):34-38.