Why mice may succeed in research when a single mouse falls short

What would happen if all clothes were made to fit only one person, or at most, that person and his or her identical twin? Whoever it was, this one person wouldn’t represent all people. I hope this is an obvious statement—we all have differences in every measurement possible, and certainly no manufacturer would make a line of clothing tailored only to one person’s size. 

But imagine taking this person and testing a new drug in her. Or him. Would you consider the drug fully tested for all people? No, it’s common sense that different people would respond differently, a concept borne out by the presence of side effects of varying severity for every significant pharmaceutical. But historically, that’s how most drugs have been selected for development until very late in the process. And that’s just one reason why it’s important to discuss the full story behind the recent New York Times article “Mice Fall Short as Test Subjects for Humans’ Deadly Ills.” 

Let’s move past the sweeping generalization of the article’s title, which is belied by the fourth sentence anyway: “The study’s findings do not mean that mice are useless models for all human diseases.” The main point of the article is valid, which is that a recent study in the journal Proceedings of the National Academy of Sciences (PNAS) shows using mice for research into response to sepsis, burns and trauma (collectively called “shock”) has not translated into useful medicines for humans. In fact, the researchers showed that the genetic response to the narrow spectrum of maladies under discussion had very little correlation at all between mouse and human. For many scientists, this is very old news.  

The New York Times article doesn’t address the fact that the studies it cites used the equivalent of one mouse—a single inbred strain, to be precise—to study the correlation (or the lack of correlation) between mouse outcomes and human outcomes in sepsis and shock. It is now well known that some inbred mouse strains, such as the C57BL/6J (B6 for short) strain used, are resistant to septic shock. Other strains, such as BALB and A/J, are much more susceptible, however. So use of a single strain will not provide representative results.

The strain in question, B6, is a reasonable starting point, but every B6 mouse is inbred to be an identical twin of any other B6 mouse. Characterizing the immune response in a single mouse strain is like doing so in a single person. Just like the analogy of the one-size clothing manufacturer, making a drug solely on the basis of one genetically isolated individual (especially a single mouse) is bound to fail. So it would have been far more accurate to use the title “A Single Mouse Falls Short” rather than “Mice Fall Short.”

Lenny Shultz, Ph.D., a professor and immunologist at The Jackson Laboratory who has made significant improvements to mouse models for human immune disease said, “. . . the mouse strain used in the study (C57BL/6) is representative of a single individual and doesn't cover the diversity in the mouse population. Use of diversity outbred cross or collaborative cross mice would provide additional diversity.” The diversity outbred cross (as previously discussed in this blog) and collaborative cross mice are mouse populations specifically developed to provide wide genetic variability, and both have been developed mainly within the past decade. Possibly, if this diversity outbred resource was used, an appropriate range of results more representative of human outcomes may have emerged.  

Elissa Chesler, Ph.D., a behavioral genomicist at The Jackson Laboratory, further commented: “For behavior and many other biomedically relevant fields we can’t simply generalize from “mouse” to “human”--we must ask which mice, and which human. Most studies involving mice are restricted to a small handful of strains. New genetic and genomic methods enable us to ask this question with improved efficiency and effectiveness. Learning how to grapple with genetic diversity and delivering experimental systems that make this genetic diversity readily accessible to those working on disease therapeutics is critical to improving the success rate of preclinical research.” Thus, genetic diversity should be accounted for in future pre-clinical tests, and researchers need to pay greater attention to selecting the right model system to mimic human disease. 

Now, largely through Lenny Shultz’s efforts, mice are also available that can host human cells. These so-called “humanized mice” have recently improved greatly in effectiveness and use, as Shultz himself documented in a recent Nature Reviews Immunology review. They are very useful for immune response studies, partly for the very reasons documented by the PNAS study authors—mouse and human immune responses differ. Engrafting human immune tissue into an experimental mouse system provides a much better platform for translational research: it tests a real human immune system in a whole organism rather than in a test tube. Therefore the mouse remains a pivotal model system for the human condition.  

Such improvement comes on top of the mouse’s already highly significant legacy, of course. I recently wrote about the work of George Snell, whose groundbreaking immunological research in mice led to the discovery of the major histocompatibility complex and, ultimately, successful organ transplants. A recent success is the multiple sclerosis therapeutic BG-12, which underwent testing in mice before showing dramatic success in clinical trials. The compound is still under review by the FDA, but its potential approval is highly anticipated.

There has been some thoughtful coverage of both the PNAS study and the New York Times article in publications such as The Scientist and Science News. Both publications speak mostly to those who are already scientifically inclined, however. It would be good to see more nuance in mainstream media outlets. It seems like there’s little middle ground between “hope for cure” articles from model organism studies that minimize the translational difficulties and “research debunked” articles like the current New York Times example. But in reality, almost all studies live in that middle ground.

Medical progress is hard-won, and few studies contribute directly to improvements in the clinic. But research adds to knowledge, some of which will eventually help doctors and their patients. Without it, we’ll have to live with the status quo, something very few will choose to accept. So read between the lines and learn about the roots of our medical “breakthroughs.” Chances are they started a while ago—in a mouse.