Mouse Marvels—The Brave New World of Genetically Modified Mice

Mouse models are a mainstay of biomedical research. Today, scientists are generating murine lines once thought to be impossible. From cancer to rare diseases, these models not only offer insight into pathophysiology, but also serve a critical role as avatars in precision medicine guiding the treatment of patients.

A short review of genetically modified (GM) mouse models

The mouse was the second mammal to be sequenced as part of the Human Genome Project. The genome of the C57/Black6 mouse was published in Nature in 2002 with over 200 scientists credited as having contributed to this historic achievement.

In the two decades since, the field has evolved in unimaginable ways. In 2006, the NIH initiated the Knockout Mouse Phenotyping Project (KOMP) with the goal of systematically knocking out every protein-coding gene in the mouse. Observing the effects of the switched-off gene offers clues to its function and potential link to diseases. The undertaking has since evolved into a global project, the International Knockout Mouse Consortium (IKMC), says Rob Taft, Senior Services Program Manager at The Jackson Laboratory (JAX), which “has generated thousands of new models and related phenotype data as a resource for the global research community. This has gone a long way toward accomplishing the NIH goal of reducing duplication of effort in generating new mouse models and greatly reduces the time and cost for investigators to access them, at least for knockouts of single alleles.”

More efficient means of genetic engineering would soon follow. “CRISPR would be the obvious answer for recent changes to making mouse models,” says Christopher Dowdy, Associate Director, Scientific Development Research Models and Services at Charles River Laboratories (CRL). Over 10 years ago, scientists first published on how to engineer GM mice using CRISPR technology.

This DNA-cutting ability was quickly adopted for editing genomes in mice (and humans), an alternative that is more precise and more efficient than other gene-editing methods. “What sets CRISPR apart is its remarkable compatibility with diverse genetic backgrounds without the need for extensive backcrossing over multiple generations, which makes the timeline to having a working model much faster. The previous set of approaches for making targeted mutations were only feasible in a handful of genetic backgrounds, and CRISPR opened a much more efficient path forward, eliminating the need for months to years of breeding,” explains Kevin Lominac, Scientific Content Manager at JAX.

Humanized mice and precision medicine

“Most non-lethal genes have an existing constitutive knockout already, and most have a conditional knockout that has been made by one of the many groups aiming to make a large library [like KOMP], and multiple commercial companies,” says Dowdy. CRL tends to get inquiries regarding conditional knockouts that don’t yet exist, single known integration of targeted transgenic lines, as opposed to random integration, and gene humanizing or knocking in a patient point-mutation. “As the technology improves, I would expect researchers trying to get as close as possible to the human gene(s) and systems to test therapy. This might be achieved through complex breeding of multiple lines, and/or further editing of engineered lines.”

Dowdy believes that the further editing of existing genetically modified lines will continue for some time. As an example, he references the evolution of growing human tumors. Xenograft models without an immune system are required to grow human cancers. Many groups have since modified these immunocompromised animals by adding additional human immune system genes to create a more complete, more “humanized” picture of how tumors behave and respond to potential treatments.

These humanized mice also allow for testing “modalities such as monoclonal and bispecific antibodies, and CAR-T cells, for safety and efficacy in ways that are otherwise not possible as those therapeutics recognize and bind to very specific sites, which differ between mice and humans,” adds Lominac.

Joshua Wood, Director of Research Services, also at JAX, says that mouse models are increasingly used to simulate human genetic diseases through two approaches: “Replacing the mouse gene with the human copy and inserting SNPs that are variants of unknown or suspected significance into the mouse or human version of the gene. The models reflect the exact mutation of a given patient or patient pool.”

This represents a big shift in the market from requesting knockouts and reporter mice in the early days to a more precise representation of human disease. These exciting models have the potential to validate novel gene therapies, which range from developmental disorders to cancers to rare and orphan diseases, according to Wood. “If the underlying cause is genetic, genetically engineered mouse models have a huge potential to impact medicine and patient care.”

Dowd points out that models of patient specific mutations for rare diseases are becoming more common, as well. Duchenne Muscular Dystrophy (DMD) is one example. “DMD patients can have many different mutations in the very large dystrophin gene. Researchers have been working with a mouse model expressing human DMD for some time, but only in the past few years have we been able to further modify the human DMD mouse, to edit their human gene to mimic patient mutations.”

JAX’s Rare Disease Translation Center has extensive partnerships with rare disease foundations to generate new mouse models with clinical mutations. So far, 57 such models have been created. “For monogenic diseases, the creation of a model carrying the same genetic mutation enables us to test numerous approaches to treatment, in particular gene therapies, allowing us to assess the efficacy of therapeutics against the underlying cause of the disease,” says Lominac.

Outsourcing considerations

As mentioned, many straight-forward mouse lines are already available thanks to initiatives like the IKMC. And Dowdy says some lab groups with the infrastructure to produce their own models, usually the ones “pushing the technical envelope,” likely have “many solid options for outsourcing all or part of the project.” He says it’s always worth checking repositories like the International Mouse Strain Resource (IMSR) to see if any existing models are a good fit.

Whether choosing to outsource a small part, or all, of a new mouse line, there are many commercial labs available. Taft says that most labs aren’t using CRISPR or other newer innovations to generate models in-house. While many labs have general capabilities to breed mice, most don’t have the advanced technology, space, or bandwidth to generate specific complex models. Further, it’s often the case that not enough models are being engineered to justify the expense.

This isn’t a bad thing. By working with a commercial lab, like JAX or CRL, investigators can take advantage of the most recent advances without having to be an expert, says Dowdy. Similarly, Taft thinks the most effective models combine the aptitude of the research lab with the technical know-how of GM mouse design and deep understanding of mouse strains of experienced providers.

There’s also the practical considerations. Dowdy cites several, including small internal cores at universities that may have a long queue, and animal health. Scientists may need extremely clean animals and if that’s the case, should work with vendors who have, for example, pathogen-free animal rooms. A lower cost provider that uses a “vivarium that can’t meet that final health standard could force the line to be rederived before studies. Suddenly the faster, cheaper option might no longer be fast or cheaper.”

Looking forward

The idea that “we don’t know what we don’t know” is a thrilling part of science, for Dowdy. “Ten years ago, some of the models we’re working on would have been hopelessly complicated,” he says. “New tools may be discovered, or more likely, existing tools will be combined in interesting new ways. While acknowledging the future is largely unknown, a central tenant is to, whenever possible, begin with the end in mind. It’s important to validate the current steps against the end goal, and make sure a potential shortcut now won’t cost you later.”