Consider the house mouse.

At maturity, Mus musculus weighs in at about an ounce. Yet over the last century, it’s become a heavyweight of biomedical research, helping scientists untangle everything from the role of the p53 tumor suppressor gene in human cancer to the cellular mechanisms by which cystic fibrosis wreaks havoc on the lungs. It has also yielded an array of insights into heart disease, diabetes and glaucoma, among other human maladies.

Polycystic ovarian syndrome—PCOS, for short—isn’t on that list. The most common cause of female infertility, PCOS also conveys heightened risk for miscarriage, gestational diabetes, preterm delivery, diabetes, heart disease, depression and anxiety, and endometrial cancer. Its hallmark is increased androgen production by the ovaries.

“Nobody really knows what causes PCOS,” says molecular endocrinologist Sheng Wu, [SD1]  has spent more than a decade investigating the hormones that regulate the reproductive system.

Unlike other human conditions, for which scientists often have a multitude of genetically engineered mice to choose among as they design studies to reveal pathology and identify treatment strategies, PCOS investigators have had to make do with simply injecting otherwise healthy mice with androgen. Trouble is, infusing the rodent circulatory system with a hormone is a particularly weak proxy for the complex physiology of a person with PCOS, whose ovaries overproduce the hormone, flooding the reproductive system and triggering a cascade of symptoms in other organs. 

“A good model can explain the mechanism of how excess androgen affects reproduction,” says Wu, “but injecting androgen doesn’t really reflect what happens in humans.”

This spring, scientists at the Johns Hopkins Transgenic Core Laboratory worked with Wu to develop precisely the mouse that she and molecular biologist Randall Reed had designed—one whose ovaries have been genetically modified to overproduce androgen in response to a dose of antibiotic. This gives the investigators control over when in its development a mouse begins experiencing PCOS symptoms—and allows Wu and her team to study how the syndrome affects such processes as sexual maturation, ovulation, pregnancy and metabolism.

The key to the effort was CRISPR, a tool that has transformed the world of gene modification by giving scientists the power to cut and paste strands of DNA at specific locations, all within the nucleus of living cells.

“Without CRISPR,” says Wu, “we would never even think about this goal because it would be too costly, take too much time and not even necessarily be attainable.”

Tailor-Made

Scientists have been messing around in the Mus musculus genome for more than a century. The earliest efforts used conventional breeding to amplify and isolate naturally occurring variations—a predisposition to mammary tumors, an inner-ear defect, black fur and so on. Later, as scientific understanding of DNA and its role in genetic inheritance accumulated, scientists began directly manipulating DNA itself, scrambling the order of nucleotides to discover how genotype informs phenotype.

Early methods were pure happenstance—the laboratory equivalent of a chimpanzee at a typewriter. Using X-rays or chemicals known to derange DNA replication, researchers were able to induce genetic mutations and study the resulting physiology. Technological innovations in recent decades—gene sequencing, cell cloning, RNA interference, zinc finger nuclease technology and transcription activator-like effector nucleases, for example—have given scientists ever-greater control of their tinkering, allowing them to turn on or off target genes to create “knock-in” or “knockout” animals.

CRISPR—pronounced like the refrigerator compartment where you stash your veggies—has unfurled a world of possibility for basic biomedical scientists who are intent on genetically engineering mice, zebra fish and other model organisms to replicate complex human conditions for use in the laboratory.

The resulting efficiencies have shifted gene editing into high gear, says physiologist Roger Reeves, a member of the McKusick-Nathans Institute of Genetic Medicine and director of Johns Hopkins’ Transgenic Core Laboratory. “This has changed the direction of our business quite a bit,” says Reeves, whose team has used CRISPR to design mice for more than 150 projects at Johns Hopkins since its first faculty collaboration, which launched in 2013.

Prior to 2013, Reeves and his team at the Transgenic Core Laboratory relied on embryonic stem cell modification, producing 100 or more strains of transgenic mice each year. It was a time-consuming, technically demanding strategy. To achieve the mutations they sought, investigators could count on spending about $60,000 over the course of 18 months—usually in the form of the salary for a postdoctoral research fellow who managed the tasks associated with inducing target mutations, then identified and cultured desirable cells slated for embryonic transfer.

“They didn’t work on it every day, but there were things going on all the time,” says Reeves, who used this method in his own lab, to swap out a couple of nucleotides on a single gene for his investigations into the array of symptoms associated with Down syndrome. Only after his team had achieved the desired modifications in a stem cell could they embark on the breeding steps necessary to stabilize the mutation and achieve a population adequate for the experimental design.

With CRISPR, investigators typically spend around $6,000 to achieve an embryo with the specific gene mutations they desire. “Now that postdoc could have a one-hour meeting with us, get on the computer and order the stuff they need in 10 minutes,” says Reeves. “We make the mice, and the first thing that scientist interacts with is a mouse in a cage in the mouse colony.”

In Wu’s case, it took just seven months from her first conversation with Reeves until she had enough PCOS mice to start running her experiment. “It’s a procedure that can be done extremely easily with CRISPR,” says Reeves, who notes that CRISPR has almost entirely supplanted embryonic stem cell modifications in the core facility. “The efficiency is very, very high.”

For early-career investigators on shoestring budgets, the economies are profound, says Jennifer Pluznick, a physiologist who worked with Reeves and his team to design a mouse in which to investigate the signaling pathways responsible for maintaining blood pressure.

“It’s an ethical advantage because we can get to the answer using fewer mice; it’s a time advantage for the researcher because we don’t have to wait for additional generations to be bred; and it’s a monetary benefit, because we’re funded by taxpayers, and we want to use that money efficiently,” she says.

Slicing and Dicing

Your smartphone owes its computing power to the zeros and ones that comprise its underlying code. The DNA that orchestrates the function of all living organisms derives from a quaternary code of adenine, cytosine, guanine and thymine. And like your phone’s strings of zeros and ones, the exact sequence of nucleotides has profound implications—setting the stage for an exoskeleton or a vertebral column, scales or fur, fecund offspring or the end of a lineage. And so nature has a wealth of redundant strategies for protecting the fidelity of DNA.

That’s what makes gene editing so challenging, especially in eukaryotes—the organisms, including humans, fungi, insects and the house mouse, that boast a cell nucleus to contain our DNA. Consider embryonic stem cell transfer, the technique Reeves and his team relied on before CRISPR came along. Using a combination of chemicals, electrical current, viruses and micropipettes, technicians had to break through the cellular membrane and into the cell nucleus to induce breaks in the DNA—all without killing the cell. Then, they took advantage of homologous repair, an innate quality control system that cells use to fix broken strands of DNA. (It’s a lot like patching a pair of jeans—so long as the patch and the hole correspond around the edges, the splice will hold.) “We might get one cell with the change we wanted out of 10 or 20 million cells,” says Reeves. “That’s why we had to do it in cell culture, where we could select for that rare result.”

It took a serendipitous discovery in prokaryotes—the single-celled bacteria and archaea from which eukaryotes evolved—to improve upon that approach. Unlike eukaryotes, prokaryotes lack a cell nucleus within which to contain their DNA. That means a prokaryote can’t afford to get sloppy about detecting and neutralizing the foreign DNA within its cell membrane that could prove its undoing. And so prokaryotes developed an innate DNA surveillance tool, the Clustered Regularly Interspaced Short Palindromic Repeats, from which the CRISPR acronym derives. Each time a bacterium vanquishes pathogenic DNA, it hangs onto a few characteristic snippets of the enemy’s DNA, compiling the genetic equivalent of a sheaf of wanted posters, at the ready for the next encounter. Imagine each clustered regularly interspaced short palindromic repeat, then, as a genetic night watchman, hoisting its lantern in the prokaryotic cytoplasm and bellowing, Who goes there?, skimming the wanted posters for a match. In the event of a positive ID, it uses a CRISPR-associated protein (Cas, for short) to mount a brisk and robust immune defense, snipping the offending DNA to pieces and stopping the invader in its tracks.

Compared to their more complex evolutionary successors, prokaryotes don’t get a lot of love in the lab, so it took nearly two decades from when scientists first identified those clustered regularly interspaced short palindromic repeats in 1993 to figure out their vital role in the adaptive immune system of an array of bacteria.  Then in 2011 or so, scientists started experimenting in the laboratory with methods to harness the CRISPR-Cas system to do their bidding. First, they created a CRISPR seeded with snippets of a target genetic sequence—a fake wanted poster, if you will—and then they injected it, along with a Cas enzyme, into the nucleus of a eukaryote. Voila! Cas induced double-stranded breaks precisely at the locations to which it had been directed. “The fact that the enzymes that work in prokaryotes also work in eukaryotes, which evolved more than 1 billion years later, is really amazing,” says Reeves.

Instead of culturing and manipulating tens of millions of cells, hoping to achieve one that incorporated the target mutation during repair, suddenly investigators could simply inject their tailored CRISPR-Cas compound into the nucleus of a single-celled mouse embryo, along with the genetic patch they wanted to introduce via homologous repair.

“Once the experiment is properly designed and we start injecting mouse embryos,” says Reeves, “we can get 30 to 50 percent of the embryos carrying the genetic mutation we want.”

A Work in Progress

In the five years since Science published the first reports of CRISPR as a gene-editing tool, scientists have published more than 5,000 papers tweaking and refining the system in everything from yeast to human embryonic cells.

The resulting buzz has fueled a steady stream of CRISPR hype, far beyond the staid world of peer review. Genetically engineered designer babies and villains with a knack for bioterrorism are an easy sell on the dystopian sci-fi market, and Hollywood screenwriters have made the most of that angle in Netflix’s Luke Cage and NBC’s upcoming C.R.I.S.P.R., starring Jennifer Lopez as a crime-fighting scientist. On Wall Street, venture capitalists have poured hundreds of millions of dollars into biotech and agricultural applications. The first CRISPR-engineered crop—a waxy strain of corn from DuPont Pioneer that’s ideal for paper adhesives and food thickeners—is slated to hit the market in 2020; CRISPR canola, cotton and tomatoes aren’t far behind.

While the fields of agriculture and basic biomedical research focus primarily on edits to the embryonic and germline cells that convey genetic information from one generation to the next, CRISPR’s promise in the clinic owes to the possibility of editing somatic cells—those in which the DNA on/off switches have specialized and converged to yield blood, muscle, skin and bones. By tweaking the genes that alter the shape of red blood cells in sickle cell disease, for example, or the genes responsible for generating dystrophin, the protein whose absence causes the symptoms of Duchenne muscular dystrophy, clinicians hope to develop therapies for everything from beta thalassemia to cancer immunotherapy. The FDA is currently evaluating more than a dozen clinical trial proposals, with several set to begin enrolling patients this year.

Even in the world of basic biomedical research, however, scientists have a lot of work to do refining their techniques. Among other things, CRISPR is an automated system, and it can get a little overzealous about its marching orders, clipping not only the target sequence of guanine, adenine, thymine and cytosine specified by the investigators, but also similar sequences nearby.

“It’s a little sloppy,” says Reeves. “You might not get exactly the change you want, even if you target the gene you were looking for, and there’s a reasonable probability you’ll make changes on other genes too, where they’re not easy to find.”

The onus is on investigators, says Reeves, to decide how much risk they can tolerate that a model organism isn’t precisely what they had in mind. “Because CRISPR is new and analysis of the positive outcomes vastly outpaces the search for possible problems, we don’t have an extremely strong idea of how often off-target problems appear,” he says. “There are papers suggesting they’re frequent, others suggesting they’re rare.”

One workaround is to run whole-genome sequencing to detect any off-target mutations before breeding the mutated animals; the other is to use conventional breeding techniques developed over the past two centuries to reveal and eliminate off-target mutations. “We know how to get rid of those problems,” says Reeves, “but it’s an extra year and a half of breeding to do it.”

Last year, Pluznick asked Reeves and his team to insert two strands of DNA in separate locations within the genome of a single embryo. Reeves and his team produced more than 20 mice, each with mutations to the target sequence. Just one had the combination of features Pluznick sought. That one is enough, she says, but such technical challenges give her pause when she reads headlines trumpeting the potential for using CRISPR to alter human germline cells.

“The fidelity is so much better than it was before,” she says, “but we’re nowhere near achieving the target outcome 100 percent of the time.”

Wu has been nibbling around the edges of PCOS for a few years now, using cell cultures to identify the molecular pathways affected by androgen in a quest to untangle how high levels of that hormone affect the combination of insulin resistance and intolerance that induce diabetes, which affects 50 to 70 percent of people with PCOS. Finally having a mouse model in which to further study the effects of androgen, from prenatal development to old age, promises a wealth of new insights to inform strategies for managing—and possibly even treating—PCOS.

“It’s a physiology question,” says Wu. “Cell lines are isolated, but in the whole organism, you see how organs connect, how they compensate for one another’s function.”

Watch a Video: Roger Reeves tells how research into Down syndrome can improve understanding of heart disease, Alzheimer’s and cancer.