Questioning the Mouse Model’s Legitimacy in Biomedical Research

A model organism is a non-human species that is studied to understand biological mechanisms that can be applied to humans or other organisms. While a wide range of species can be studied as model organisms, the cornerstone of biomedical research is the house mouse. Mice have been used as model organisms for many decades [1]. Scientists across the world have a detailed understanding of and a wide range of techniques to study mouse anatomy, physiology, and genetics. By 20th century, laboratories across the country were breeding mice with desired traits in order to explore the genetic aspects of human disease.  In the early 80s, researchers perfected a method of inserting select DNA into mouse embryos, establishing the first lines of genetically engineered mouse models (GEMM). A few years later, methods for removing genes were perfected, leading to the first “knockout mice.” GEMMs revolutionized biomedical research by allowing for greater specificity of disease models. Scientists could now essentially pick and choose which genes to keep on and which to turn off.

There are many reasons behind the widespread use and success of the mouse model and GEMMs. This animal shares 99% of its genome with humans, giving researchers confidence that their findings can be reproduced in similar human experiments. It also maximizes the number of experiments that can be done in mammals, since experiments involving the mice raise fewer ethical concerns than any other mammal model organism laboratory, mice communities are to maintain since they are small, inexpensive, and have short generation times. In addition, the wide array of anatomical, physiological, and genetic knowledge about one species is nearly unrivaled.

The mouse model is not perfect though. As our knowledge of disease mechanisms grows, our mouse models struggle to keep up. Ronald Tompkins, a professor of surgery at Harvard, stresses that “we need to recognize that simple models do not reproduce complex human diseases [2].”

At the beginning of 2013, Tompkins participated in a study that compared mouse and human genomic responses to inflammatory diseases. After as many as 150 clinical studies on potential anti-inflammatory drugs failed after suggesting promise in mouse studies, the co-authors examined temporal changes in mouse and human genomes and expression pathways after various types of inflammatory diseases took hold. Astoundingly, while the genomic responses among the various human diseases was highly similar, there was little to no correlation with the genomic reactions of mice to equivalent diseases. The mice did not even experience similar changes among the various types of inflammatory diseases[3].

The paper resurfaced a question that attacks a fundamental aspect of genetic research: is the mouse model capable of accurately mimicking human diseases or can it no longer produce clinically relevant data?

It is easy to assume that data obtained from a mammal that shares 99% of its genome with humans could be reliably reproduced in humans.  In reality, this is not the case. GEMMs are created by artificially removing or changing the expression of a single gene and its pathway, but human diseases are complex, involving multiple genes and pathways. Furthermore, genetic abnormalities caused by generations of inbreeding and anatomical differences cause mice to respond to potential treatment much differently than a human would, these changes are nearly impossible to predict.
There is a wide array of candidates for replacements of the mouse as research’s leading model organism, ranging from microscopic unicellular organisms such as E. coli to multicellular organisms as large as Rhesus monkeys. The decision to conduct research on any of these organisms involves consideration of background knowledge on the organism, the cost of maintenance, complexity of the condition under study, difficulty of experimental procedures, and biological and behavioral similarity to humans.

While numerous model organisms are smaller and easier to maintain than mice they often contain far fewer genes than humans do.  A significant number of these genes do not overlap with the human genome. A simpler cellular environment and tissue composition results in significant molecular differences in these organisms that cannot be reproduced in human cells and tissues.

Researchers may be able to produce more reliable and translatable data from larger mammals than from mice, but the use of these organisms garner substantial ethical and maintenance concerns. The similarities in cognitive ability between major primate species and humans prevent most scientists from conducting experiments on them. Some countries have banned several species from undergoing any testing at all. Furthermore, larger animals require higher and more frequent maintenance than mice, making their use in experiments financially unfeasible.

Another option is to study mechanisms of disease in cells of a specific organism grown in a controlled molecular environment, or in vitro studies. This technique allows a researcher to remove tissues or cells from a complex environment and study it in a simple, controlled medium. The simplicity of this model, however, prevents its application to humans. The number of environmental factors affecting the molecular interactions within a cell cannot be accounted for by an in vitro study that has very little outer-cell interactions.

Even with its unavoidable issues, the mouse model clearly provides the best balance of applicability to humans, cost, and simplicity for current experiments biomedical research. The best option moving forward, therefore, is to identify all possible bottlenecks and limitations with the model and correct them.

With so many opportunities to fine-tune the mouse model, several initiatives are already underway to address its problems. To better mimic human conditions on the anatomical and physiological level, researchers are integrating “humanized mice” into their experiments. These mice, which have been immuno-compromised and have been implanted with human hematopoietic cells to reproduce a human immune system, have been used to understand mechanisms of disease in areas that previously could not be understood in other mouse models.

Meanwhile, clinical trial labs are increasing the reliability of their data through some new methods and are improving older ones. In a co-clinical trial, for example, a drug is tested on a mouse under a “mouse clinical space,” which provides treatment as similarly as possible to a human clinic experience. In addition, researchers are tightening control over which trials can move on to humans by closely examining the lab environments of the mice, their behavior, and any stressors to which they are exposed.

Some scientists are attempting to modify lab mice to more closely mimic human anatomy and physiology. Others want to cut costs of obtaining mouse models by reducing the number of patents and increasing their availability to pharmaceutical companies. Better communication and exchange of mouse models and data would hasten and increase the efficiency of the clinical trial process. In addition, improved training in human and mouse pathology and further understanding of changes in knockout mice would allow researchers to plan experiments that would better mimic human conditions [4, 5].

Many other organizations are working to spread knowledge of the mechanisms behind the biochemical and behavioral differences in GEMMs. A better knowledge would save a lot of time in the preclinical research process, since a lab would be able to choose a mouse model with a phenotype that is closest to their disease of study. The European Conditional Mouse Mutagenesis Program is working to assess mouse model screening procedures and to standardize tests for easier use by mouse-based labs across the world [6].Stemming from EUMODIC is the International Mouse Phenotyping Consortium, which identifies differences in single-gene mutant mouse models from normal mice [7]. Scientists will be able to access this data and determine exactly what mechanistic and phenotypic changes occur in their chosen mouse model.

Although its problems have been exposed and improvements have only begun, scientists have confidence in the mouse model. With new insights and techniques that will improve the application of mice in the laboratory, this model isn’t going anywhere any time soon. 


  1. Spencer, Geoff. “Background on Mouse as a Model Organism.” National Human Genome Research Institute.
  2. Cossins, Dan. “Do Mice Make Bad Models?” The Scientist. February 11, 2013.
  3. Seok, Warren, and Others. “Genomic responses in mouse models poorly mimic human inflammatory diseases.” Proceedings of the National Academy of Sciences. 110, no. 9 (2013): 3507-3512. Accessed October 23rd, 2013.
  4. “Of mice and men – are mice relevant models for human disease?” European Commission.
  5. Shultz, Brehm, and Others. “Humanized mice for immune system investigation: progress, promises, and challenges.” National Reviews Immunology. 12, no. 11 (2012): 786-798. Accessed October 23rd, 2013.
  6. “EUMODIC Summary.” The European Mouse Disease Clinic.
  7. “Goals and Background.” International Mouse Phenotyping Consortium.
  8. Image credit (NIH): Bartlett, Maggie. “Knockout Mice.” Wikimedia Commons. Last modified June  7, 2012.

Itai Doron is a junior at Emory University majoring in Biology, with an interest in microbiology. Follow The Triple Helix Online on Twitter and join us on Facebook.