Researching CRISPR

The CRISPR Revolution

In the beginning of 2013, a new tool, ingeniously adapted from an immune strategy that bacteria and archaea use to protect themselves from viruses, was being evaluated for the potential to revolutionize the way engineered mice were made. The name of the new technology was CRISPR (Clustered Regularly Interspaced Short Palindromic Repeats) and most mouse geneticists had never heard of it before.

Before 2013, engineering of mice was a laborious multistep process that involved genetically altering mouse embryonic stem (ES) cells, injecting them into an embryo, and breeding multiple generations of animals. Even the best researchers could take up to 2 years to engineer a mouse using this technology. CRISPR replaces all of this with a molecular complex that can do targeted genetic surgery on a fertilized egg, producing a strain of transformed mice in 6 months.

Most investigators get their engineered mice from colleagues or by purchasing them from commercial outfits or academic-based repositories. Custom-making a genetically engineered mouse through industry can run anywhere from $40,000 to $60,000 or more using the older technology. By making the engineering of mice far simpler and cheaper, CRISPR opens the way for more labs to make custom models themselves. “When you made knockout mice before, you needed some skills,” says Rudolf Jaenisch at the Massachusetts Institute of Technology (MIT) in Cambridge. “Now, you don’t need them anymore. Any idiot can do it.”

But can any idiot do it?

Some mouse engineers have second thoughts about the rush to completely replace ES cell technologies with CRISPR. Few doubt its potential, but the technique is still a work in progress, and its ability to alter genomes has one big gap. Although CRISPR knocks out genes with ease, it is less efficient at inserting, or knocking in, new DNA. That’s important not just for giving an animal a novel function, but also for creating a knockin known as a “conditional” knockout, an animal model in which researchers can turn off a target gene at specific times of life or in specific tissues.

Because CRISPR is less adept at making conditional knockouts, William Skarnes, who led a team making mutant mouse ES cells at the Wellcome Trust Sanger Institute in Hinxton, U.K., worries that researchers are overemphasizing the new approach. “The decision to abandon the ES resource in favor of making simple knockouts is a mistake,” Skarnes says. “You still want to make conditionals through the ES route.”

CRISPR researchers are now refining the technique to do knockins with greater efficiency. But that entails tinkering with the mechanisms that cells use to repair broken DNA, which are critical to their health. Some are cautious about overmanipulation of biology to increase efficiency.

MIT’s Jaenisch was the first to show the power of CRISPR for producing mouse knockouts. He and co-workers reported that the technique successfully disrupted five genes in a single set of mouse ES cells, something that was not possible before. More important, they showed that they could bypass ES cells altogether and simultaneously knock out two genes in single-celled mouse zygotes, or fertilized eggs. No longer would researchers have to modify ES cells and painstakingly breed several generations of mice to produce an animal that carried the mutant gene in its egg or sperm cells. And researchers who wanted mice with two mutations would no longer have to interbreed single mutants and go through a similarly time-consuming, cumbersome process to arrive at progeny with the altered germ line. Since then, more than 500 papers have detailed how CRISPR can both knock out and knock in genes in mice.
CRISPR’s impact is measured in more than savings. The ease and speed of the technique makes it possible to engineer mice on the fly, to solve new and specific targeting puzzles, and also easily modify previously genetically altered mice in other, new ways.

On one front, however, the CRISPR revolution is faltering. Three months after his lab’s first CRISPR report, Jaenisch and co-workers published a second paper in Cell that suggested CRISPR could easily perform more complex genetic surgery, knocking in chunks of DNA rather than simply disabling genes. As a demonstration, they used CRISPR to knock fluorescent tags into mouse zygotes, which lighted up whenever a specific gene was turned on. They also created conditional mutants, which are key to many research efforts. Using CRISPR to insert loxP’s into zygotes, Jaenisch’s team reported making conditional mice with relatively “high efficiency”—about 16% of the zygotes led to mouse pups with the correct mutations.
Skarnes is one of many researchers bowled over by Jaenisch’s initial reports, but he was disappointed when he tried to take the technique into his own lab. “It looked from his papers that this was going to be straightforward and I was quite confident this would make ES obsolete,” Skarnes says. “What was disappointing is none of us could reproduce at the efficiencies reported by Jaenisch. … It works at 1% or 2% and a lot of projects are failing. It’s really not proven to be a robust method.”

Skarnes calls the shift to CRISPR for conditional knockouts premature. But he concedes that researchers will “eventually” figure out how to tweak CRISPR so that it makes conditional mutant mice with high efficiency. Whatever CRISPR’s shortcomings appear to be at this point, its potential for engineering mice should not be underestimated. There are things CRISPR can do that people are just beginning to understand. It’s at the very early phases of development and the tool has infinite possibilities.