Modifications in disease models that used to take months or years, can now be achieved in several weeks

In most areas of biology, the more we learn, the more we realize how complicated the systems actually are. Modelling complex human diseases requires complex genetics: one might need to remove the normal mouse gene, then add in multiple human disease genes. As the field discovers more genes affecting outcome, the models necessarily get even more complex. In areas with diverse genetic modifications, like cancer, the extent of genetics required to properly unravel a problem can quickly become overwhelming. Here at Charles River, we are looking at several approaches to attain better research models, and that is where new genome engineering tools like CRISPR become so powerful. Check out this webinar to learn more.

There are two main ways to edit a genome. The classical approach used homologous recombination to trick a cell into adopting something close to normal sequence. To knock in a couple 100 bases with your mutation, you would need to include thousands of bases that exactly match normal sequence, on both sides. Even then the efficiency was quite low (there’s a reason we used 384 well plates)! Building those large targeting vectors, coupled with the low efficiency makes the process relatively slow.

The newer approach is to cause specific DNA damage (double strand breaks or DSB) and let the cell repair them naturally. In principle, anything that could cause a DSB would work. In practice, getting breaks exactly where you want them was very challenging, and time consuming. Before CRISPR, tools available were cumbersome, expensive or difficult to readjust. You might need to re-engineer most of the protein to target a new sequence. 

CRISPR allows modifications that previously took months or years to be achieved in weeks. The endonuclease (Cas9) is conserved and a short RNA guide can specify targeting. In the past 10 years, the RNAi (RNA interference) field has taught us a great deal about targeting sequences with small pieces of RNA, and much of that knowledge can be applied to CRISPR.

Utilizing CRISPR, transgenic mice can be generated faster than ever before and with great precision. Through directly injecting a one-cell embryo CRISPR can modify the entire animal while it is only a single cell. This virtually guarantees germline transmission (your modification passing on to the next generation), which is frustratingly left to probability when you use embryonic stem cells. Knocking out a single gene is straightforward, and while knock in and conditional mutations are more complex, a single microinjection session should generate the mice needed to found your new line.

I had the privilege of attending the Keystone Symposium on Precision Genome Engineering to present a poster and hear about the leading edge of technology in the field. CRISPR has accelerated the rate of progress in genome engineering to a truly dizzying pace.

Investigators are using CRISPR-based screens to look for targets driving disease, and in some cases even to look for mutants of a disease gene that lessen severity. Re-breaking a broken gene to fix it wasn’t the only outside-the-box idea that is now readily accepted. CRISPR is so good at targeting a specific sequence that some groups are using a ‘dead’ nuclease (one that can target but no longer cuts) tethered to other proteins that can turn genes on or off. This has been termed CRISPRa and CRISPRi (to activate or inhibit, respectively) and is being used to tune gene expression. Other reports were on ways to further increase CRISPR’s accuracy, allow broader range of targeting and even going back to microbial genomes to mine for the next endonuclease we can use as a tool.

Our poster, a collaboration between Charles River’s Leiden and Wilmington sites, laid the proof of concept groundwork for offering our new Model Creation service. We knockout lines using CRISPR through both indel formation and by knocking in a small stop cassette. In one experiment we show we can utilize both main repair pathways. We’ve also replicated the work through direct embryo injection and by modifying ES cells to cover all our bases (some constructs are complicated enough you would still want to start in ES cells). And we will continuously look for ways to improve our process.

One interesting challenge with CRISPR is the efficiency of editing. Usually when talking about efficiency, the problem is with it being too low. While that is the case with knock in constructs requiring homology directed repair (HDR)—there were plenty of talks about increasing HDR efficiency at Keystone—I’m talking about mosaicism.

It turns out that CRISPR often works so well that trying to modify a target gene will actually modify both copies! Every cell gets two copies of each gene (one from each parent), so the concern is that when you inject an embryo with CRISPR your target gene will be modified twice (once on each copy). This is called mosaicism and can be a little troublesome when one copy is the intended variant and the other is not, especially when those variants could vary by a single nucleotide. Carefully screening animals is critical, and thankfully our Genetic Testing Service has assays sensitive enough to find those subtle variations. Once you know which mice might be your founder, one more round of breeding can segregate those mosaic variants and you have your new line!

When I left academia years ago, it was to work with tool builders, optimize techniques and find better methods. I knew my personal impact on medicine would be greater if I helped drive the research of others. I am glad to be at Charles River, where we leverage powerful tools like CRISPR to accelerate research.

If you are interested in learning more about CRISPR, check out these recent articles in Eureka on humanized mice and the need to be prudent in using CRISPR.