Many of the mouse models used in laboratories show inconsistencies between their names and their genetic makeup, according to an analysis

Is the article of good quality?

“The article is solid, extremely timely, and has significant practical implications. Biomedical research has always taken it for granted that laboratory mice are genetically what the label says they are. However, this study reveals, for the first time, that in many cases the animal’s actual genetic background differs completely from the information provided by the source laboratory. This affects not only the animal’s response to a disease or a drug, but also the conclusions of a study.

The publication is not radically innovative, but it quantifies a problem that the scientific community has suspected for decades. Its acceptance in one of the world’s most important scientific journals underscores that the problem is not the animal model itself, but the lack of rigor in its use. It is a message directed not only at the scientific community, but also at institutions, at a time when political pressure to reduce the use of laboratory animals is growing, especially in the United States.”

How does this fit with previously known evidence, and what implications might it have?

“A mouse’s genetic background is not a minor detail; in fact, it is probably the most important experimental variable and, at the same time, the least controlled. Numerous studies have shown that the same genetic mutation can produce radically different phenotypes depending on the animal’s genetic background. In other words, when two laboratories use ‘the same strain’ but with different genetic backgrounds, the results can be inconsistent without any apparent explanation.

This does not merely mean that the mice have genetic errors, but it raises the question of how many published studies might be affected by this problem and how much time and money has been wasted as a result. This research suggests that part of the answer may lie in something as basic as not knowing exactly which mouse is being used.

On the other hand, the article offers a solution that has already been implemented: a genetic quality control system designed to be interpretable even by non-experts. This is MiniMUGA, a commercial product already available, with public reference data and documentation detailed enough to allow for its reproduction by third parties.”

Are there any significant limitations to consider?

“Among the study’s main limitations is that the analysis of mouse strains is based on a U.S. bank, so we do not know if the problem is just as common or even more severe in other countries. Furthermore, the article does not discuss the economic cost of creating a ‘genetic ID’ for each mouse nor does it address the question of whether all laboratories worldwide could afford that additional expense. In fact, the MiniMUGA system is currently available only for strains from suppliers in the United States and Europe, which could limit its global implementation.

Finally, although the article demonstrates the existence of genetic inconsistencies in mice, the direct link between these and experimental errors is implied but not objectively proven. It is also unknown whether the rates of inconsistency would be comparable in strains maintained internally in academic laboratories, which could yield better or worse results depending on the rigor of each group.”

This is an interesting topic. In laboratories, we crossbreed and conduct experiments with strains of inbred mice that have been produced by backcrossing parents and offspring over successive generations until the two copies of each gene tend to be identical, and thus the offspring are as genetically identical as possible. Genetic variation occurs, which creates problems in experimental repetitions and replicates across different laboratories. The authors provide evidence and analysis showing that these problems occur within the scientific community and argue that they affect the reproducibility of results across laboratories.

They propose standardizing genetic quality control processes to specifically identify the identity of the mouse strain used in the experiments reported in a given publication. This is clearly an initiative that will likely be implemented and that will ensure these genetic characterizations of mouse strains are performed systematically and reported alongside publications of experimental results. In my opinion, this is an interesting initiative that, in the field of cancer research—and specifically in cancer immunology and immunotherapy—can prevent problems, but at the same time, it will add some additional costs to the final price of experimentation.

Humans, in general, are not consanguineous; we do not share all the same genetic variants. We know that, in humans, consanguinity, inbreeding, and the birth of children to closely related relatives (such as siblings) is strongly discouraged, as it increases the risk of disease due to the presence of mutations in the same genes inherited from both the father and the mother in a single individual. Let’s remember that we share 99.9% of our genome with any other human being; in other words, we differ by 0.1%. But this seemingly small amount conceals a significant number of nucleotides—the letters of our genome—with no fewer than three to six million positions differing when we compare one person to another. And this makes us, fortunately, unique individuals who differ in detail, even though we share the vast majority of the genome in substance.

With mice, the situation is different. Rodents, and mice in particular, do not face the same issues of inbreeding and tolerate inbreeding well. For this reason, researchers have become accustomed to using inbred lines—strains of mice that are practically genetically identical to one another, the result of many consecutive crosses between siblings—until a specific genome is established and maintained that is essentially the same in all individuals within the same colony. Since all individuals in the colony are genetically so similar, the benefit is that experiments have less variability (we almost eliminate individual differences), which means we can use fewer animals in experiments to detect, if any, statistically significant differences in the parameter we are studying.

The problem is that there is not a single strain of mice, but rather hundreds of inbred strains. Within each strain, all individuals are extremely similar, but if we compare the genome of one mouse strain with that of another, we will find many genetic differences. To complicate matters further, we must also account for the spontaneous mutations that constantly arise in any living organism, including mice (this is called genetic drift). If we take a group of mice from a specific strain, split it into two groups to create two different colonies at two different research centers, and keep them by breeding them separately, they will eventually become slightly different, as they will accumulate genetic mutations that are not the same. These differences will increase the longer the two groups of mice from the same strain remain separated. Thus, two researchers may believe they are using mice from the same strain, but if they have kept the colonies separate for a long time and have not taken care to refresh the colony with original individuals from the same supplier, the truth is that the mice will be genetically distinct, even though the researchers name them the same way and (erroneously) believe they are equivalent.

All of the above affects the genetic characteristics of each mouse model of a disease of biomedical interest. For the conclusions reached by two researchers using the same mouse strain for their respective experiments to be comparable, the animals they use must also be comparable. If, on the other hand, the colonies have been separated for some time and have accumulated various mutations, then it is very likely that the conclusions reached by the researchers will differ, even if they believe they are using the same strain of mice.

A study recently published in the journal Science by mouse geneticists in the U.S. highlights the underlying, and often unknown, of the mice used by researchers in biomedicine and calls into question the genetic purity of many mice stored in mouse repositories (in the form of cryopreserved sperm or embryos) under strain names and with the presence of certain genetic constructs that frequently do not correspond to the reality reflected in the descriptive name of those mice. The researchers analyzed the genomes of 611 individuals derived from 341 mouse strains deposited in the American MMRRC mouse repository and found that only 20% of them accurately correspond to the genetic characteristics indicated in the strain name. In the rest, they found additional genetic modifications, or the absence of modifications that should be present, or genetic variants indicating strain mixing.

This is a well-known fact that those of us responsible for mouse repositories continue to grapple with, trying to convince our colleagues to detail the genetic characteristics of the mice they report in any scientific article, explaining the genetic variants they possess, the genetic modifications introduced via transgenesis, mutation, or gene editing, so that those who use them later are not surprised to discover that they do not contain the genetic variants they should and, perhaps, contain others that should not be present. In Europe, through the European infrastructure INFRAFRONTIER, we have recently published recommendations for reporting in detail and with precision all the genetic characteristics of a mouse strain so that those who use them later know exactly what type of mice they are experimenting with.

The origin of the problem likely stems from the multitude of crosses of all kinds that have been carried out with the thousands of mouse strains created by the scientific community. If a mouse from strain A carries genetic modification 1 and we are interested in seeing the effect of another genetic modification 2 present in a mouse from strain B, then the standard practice has been to cross the two mice until, after several crosses and generations, genetic modifications 1 and 2 are present in the same individual. However, the strain of that mouse will then no longer be either A or B, but rather a mixture of the two genomes. And if we now send our mouse with the double genetic modification to a collaborator who wants to investigate the effect of a third genetic modification 3 present in strain C, then, after the appropriate crossings, we will end up with a mouse carrying the triple genetic modification but whose genetic variants in the genome are neither A, nor B, nor C, but a mixture of the three genomes. And researchers might expand their colony by crossing the mice with other A individuals and describe that mouse as A in their scientific publication, when in reality, if we were to investigate its genetic purity, we would find variants of B and C that were not in its name but are still present. This is an all-too-common error that contributes to variability in results and a lack of reproducibility, as the mice used may have other genetic variants unknown to the researcher.

This study in Science once again highlights a well-known problem that mouse geneticists strive to combat by every means possible, because it has a solution, recommending that periodic genetic tests be performed to verify at all times that we are truly working with the X strain mouse we believe we are working with, and that other mutations and genetic variants—which have not been reported but are present in the mice we are using—have not slipped in, contributing to genetic noise and variability. American researchers recommend the use of a genetic test they themselves helped develop, the MiniMUGA genetic quality test, but there are other ways to genetically validate a mouse’s purity, such as through whole-genome sequencing.